Book - Gears - Volume 1 Geometric and Kinematic Design
Book - Gears - Volume 1 Geometric and Kinematic Design
Book - Gears - Volume 1 Geometric and Kinematic Design
Vincenzo Vullo
Gears
Volume 1: Geometric and Kinematic Design
Springer Series in Solid and Structural
Mechanics
Volume 10
Series Editors
Michel Frémond, Rome, Italy
Franco Maceri, Department of Civil Engineering and Computer Science,
University of Rome “Tor Vergata”, Rome, Italy
The Springer Series in Solid and Structural Mechanics (SSSSM) publishes new
developments and advances dealing with any aspect of mechanics of materials and
structures, with a high quality. It features original works dealing with mechanical,
mathematical, numerical and experimental analysis of structures and structural
materials, both taken in the broadest sense. The series covers multi-scale, multi-field
and multiple-media problems, including static and dynamic interaction. It also
illustrates advanced and innovative applications to structural problems from science
and engineering, including aerospace, civil, materials, mechanical engineering and
living materials and structures. Within the scope of the series are monographs,
lectures notes, references, textbooks and selected contributions from specialized
conferences and workshops.
Gears
Volume 1: Geometric and Kinematic Design
123
Vincenzo Vullo
University of Rome “Tor Vergata”
Rome, Italy
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Aphorism
v
To my wife Maria Giovanna,
my sons Luca and Alberto,
my nephew Nicolò
and my students
Preface
Gears and gear power transmissions used in most mechanical applications and,
especially, in the automotive and aerospace industry constitute one of the classical
topics of Machine Design Theory and Methodology. In the opinion of many
scholars, including the author of this monothematic textbook, the gears are, by far,
one of the most complicated but fascinating subjects of the mechanical engineering,
because it is extremely polyhedral, as it calls into question a remarkable variety of
disciplines, of which the main ones are: plane and space geometry; various bran-
ches of theoretical and applied mechanics, such as kinematics, dynamics, elasto-
hydrodynamics, vibrations and noise; continuum mechanics and machine design
theory and methodology; static and dynamic material strength; structural opti-
mization; surface contact and lubrication theories and tribology; materials science
and metallurgy; cutting processes and, more generally, technological processes of
manufacture; maintenance, etc.
Despite the fact that, from a historical perspective (see Gears—Vol. 3: A Concise
History), the gears constitute, after the potter’s wheel, the oldest mechanism
invented by Homo Sapiens Sapiens, we can say that a unified scientific theory
of the gears, able to consider all the aforementioned disciplines, amalgamating them
in a “unicum” does not yet exist at present. Currently, according to the most
widespread thinking that is shared by historians of science and scientists, a science,
to be such, must have at least the following essential features:
• All that science claims should not involve real objects, but specific theoretical
entities.
• The theory on which a science is based must have a strictly deductive structure,
i.e., it must be characterized by a few fundamental statements (called axioms or
postulates or principles), and by a unified method, universally accepted, to
deduce from them an unlimited number of consequential properties.
• Its applications to real objects must be based on rules of correspondence
between the theoretical entities and real objects.
ix
x Preface
Well, a unified and comprehensive science of the gears, which meets all these
requirements with reference to multiplicity of the disciplines mentioned above, does
not exist and, perhaps, it will not exist for a long time yet. The way to achieve this
important goal is still long and fraught with difficulties, since it is necessary to agree
the different scientific principles that form the bases of the various disciplines
involved. Based on the current state of knowledge, we can only say that there are
different scientific theories of the gears, as many as the disciplines that contribute to
defining a gear design.
Gear design scholars and experts are well aware that, with a design based on the
fundamental statements related to only one of the disciplines involved, there is a
reasonable certainty of not meeting the postulates related to other disciplines. The
difficulties to make a theoretically perfect design of a gear power transmission, i.e.,
respectful of the countless design requirements related to the various disciplines
involved, are actually insurmountable, so the designer must fall back to a com-
promise design (the so-called design optimized, but often very far from the theo-
retical one), able to reconcile as best as possible the different and almost always
conflicting requirements, dictated by these disciplines.
The scholars of vaunted credit, who claim to know everything and to be able to
speak or write about any subject and who boast or attribute themselves knowledge
in any field, consider the gears as a synonym of obsolescence, a symbol of the past
or, when they are benevolent, a nineteenth-century old stuff. This way of thinking
of those we have benevolently referred to as scholars of vaunted credit (no one can
therefore accuse this author of not being kindly gentlemen) implies at least the
ignorance of the historical evidence that the theory of the gears, considered as
mechanisms, is much older, having it characterized the birth of science, in the first
Hellenism (see Gears—Vol. 3, Chap. 3). But the unjustified conviction of these
so-called scholars hides a far more serious ignorance. In fact, they prove not to
know that the most significant contributions for the calculation of the load carrying
capacity of the gears were brought gradually in the entire twentieth century (with
the exception of Lewis, 1982—see references in Gears—Vol. 2, Chap. 3) and that
in this beginning of the third millennium new and equally significant contributions
of high scientific value appeared and continued to appear at the horizon (see, e.g.,
Vol. 2, Chaps. 10 and 11).
On the contrary, the true scholars and experts of gear power transmission sys-
tems are well aware that the gears were yesterday, continue to be today and for a
long time yet will continue to be an ongoing scientific and technological challenge.
Even today, surely it is worth investing significant financial resources, in terms of
manpower, tools and means, in R&S activities on this important area, as all the
technologically advanced countries continue to do. This depends on the fact that the
gears are a very complex multidisciplinary field, as few in mechanical engineering,
and knowledge still be acquired is numerous. We can affirm, without the fear of
being denied, that gears are the result of ancient knowledge in continuous updating.
Knowledge still to be acquired concerns not only those specific to each of the
disciplines involved (geometry and kinematics; static and dynamic loads, including
those due to impact; friction and efficiency; dynamic response and noise emission;
Preface xi
static and fatigue tooth root strength; contact stresses and surface fatigue durability;
nucleation of fractures of any kind and their propagation until breakage; full film,
mixed and boundary lubrication; scuffing and wear; materials and heat treatments;
new materials; cutting processes and other manufacturing processes, etc.), but also
those arising from their mutual interactions. Other important challenges are those
related to the new fields of application (see, e.g., those related to the helicopter and
aerospace industry and wind power generators), which require the introduction and
design of new types of gear drives that are not reflected in the current technological
landscape.
Researchers working in this field are well aware of being in front of an important
goal: to develop a comprehensive and unified scientific theory of the gears. Really,
at present, a scientific theory of gears having a general horizon, capable of con-
sidering and balancing all the aspects involved in their design (i.e., geometric and
kinematic aspects, mechanical strength aspects and technological and production
aspects) does not yet exist. Only attempts of partial scientific theories of gearing
exist, which are limited only to the first of the three aforementioned aspects. At the
moment, partial scientific theories concerning the other two aspects mentioned
above cannot unfortunately be reported, as non-existent.
Among the partial scientific theories concerning the geometric and kinematic
aspects, it is however worth mentioning the laudable attempt by Radzevich (2013,
2018: see references in this volume, Chaps. 1 and 11) who, for the first time, tried
to develop a scientific theory of gearing. In reality, Radzevich is heavily indebted to
Litvin (1994: see references in this volume, Chap. 12) and his followers, inasmuch
as he treasures the substantial contributions on the subject in terms of differential
geometry brought by them, presenting already known knowledge concerning the
above aspects in the systematic form of a scientific theory. However, Radzevich has
the great merit of having set up a general scientific theory of gearing that, although
limited to geometric and kinematic aspects, is capable of treating all types of gears,
including hypoid gears that include, as special cases, all other types of gears.
Notoriously, gears constitute a classic multidisciplinary machine design topic. In
this multidisciplinary framework, numerous and different areas can be identified,
which however are between them interdependent, since the variation of a parameter
or quantity relative to a given area, inexorably, affects a parameter or quantity of
other areas involved. Usually, three areas (it is to be noted that each of these areas is
composed of two or more disciplines) are introduced, which identify three char-
acteristic aspects of the gears. They are as follows:
• Geometric–kinematic area, which considers the gears as ideal theoretical enti-
ties, and that, in these ideal conditions, tries to optimize their geometric and
kinematic quantities.
• Stress analysis area, which aims to stretch as much as possible the lifetime of the
gear, considered as an actual structural machine element, working under real
operating conditions.
xii Preface
is based on a set of postulates from which we can deduce the entire body of
knowledge of a given area concerning the gears.
Even this monograph on the gears is substantially aligned with those published
previously. However, unlike the latter, this monograph favors the basic concepts
of the calculation of the gear strength, inclusive of the rating of their load carrying
capacity in terms of tooth root fatigue strength, surface durability (pitting),
micropitting, tooth flank fatigue fracture, tooth interior fatigue fracture, scuffing,
wear, etc. However, these calculations cannot be performed accurately without
in-depth knowledge of the geometric–kinematic aspects of the mating tooth flank
surfaces as well as those concerning the tooth cutting processes. In this perspective,
a large part of this textbook (the first volume) is dedicated to the fundamental
geometric–kinematic aspects, while the second volume of the same textbook is
reserved to the basic concepts of strength and durability design, which constitute
the basis of international standards on the topic. The technological aspects of the
tooth cutting are not however discussed in a specific part of this monograph,
comparable by extension and deepening with the parties reserved to the two
aforementioned aspects; they are just called up and described here and there, in the
framework of the first two aspects, just enough because the designer has a picture as
complete and comprehensive as possible to be fulfilled for a good gear design.
This textbook mainly collects the lectures held by the author, in the four uni-
versities where his academic career took place, namely in chronological order:
Polytechnic of Turin, first as an assistant professor and later as an associate pro-
fessor, from 1971 to 1980; University of Naples “Federico II”, as a full professor,
from 1980 to 1986; University of Rome, “Tor Vergata”, as a full professor from
1986 to 2013 and as a full professor retired from 2014 to 2017; Cusano Telematic
University of Rome, as a contract professor from 2018 to date.
However, it is noteworthy that this textbook takes into account the lectures held
in these four universities, in the framework of many courses of Bachelor’s degree,
Master’s degree and Ph.D./D.Phil. terminal degree in Mechanical Engineering
Science, but also lectures for the training of researchers and research technicians,
specialized in gear design, which the author held for Costamasnaga S.p.A. The
textbook also collects and builds on many years’ experience (over fifteen years)
gained from the author at this company, as a consultant, especially for industrial
research programs and projects of technological innovation regarding special and
custom-made gear systems.
This monothematic textbook has the twofold aim of meeting the needs of uni-
versity education and those of the engineering profession. In the first area, the goal
is to provide a link between the matters covered in the most basic textbooks on
gears intended for students in three-year first-cycle degree programs and those
addressed in the more advanced monographs on gear theory to be used in
second-cycle and third-cycle or doctoral programs. In the second area, the purpose
is to provide practitioners and professional engineers working in research and
industry with an advanced knowledge on the subject that can serve as a basis for
designing gear systems efficient and technologically sophisticated, or for devel-
oping new and innovative applications of the gears.
xiv Preface
Anyone who has ever worked with the analysis and design of gears is well aware
that their actual engineering applications, with no exception, are characterized by
the fact that a gear or a gear train is part of a complex mechanical system, which
must be analyzed in its entirety, that is considering all the possible interactions with
other mechanical parts or mechanical units that make up the entire system. These
interactions between the components or groups of the system enhance the resulting
stress and strain states, including those that arise under dynamic operating condi-
tions, such as impact loading, vibration, high- and low-cycle fatigue and so forth.
The higher the transmission error, the greater the deviations between the theoretical
conditions and real operating conditions.
The analysis of such complex systems requires the use of sophisticated and
refined mathematical models. To save time and costs for the calculations to be
made, these mathematical models can use simultaneously numerical models, based
both on the discretization of the continuum, such as finite element method
(FEM) and boundary element method (BEM), and on the discretization of the
governing differential equations, such as finite differences (FD) and analytical
models, such as step-by-step integration methods of the differential equations
governing the dynamic behavior of the system to be examined and modal analysis
methods for the study of the system’s response to the dynamic loads applied to it.
All the problems presented in this textbook on the gears are approached and
solved preferably using, whenever possible, theoretical methods, and attention is
mainly focused on analytical and methodological aspects. The analytical definition
of the tooth flank surface geometry, including that of the tip and fillet with all or
part of the rim of the gear members to be analyzed, has another great advantage
compared to traditional methods that do not use the differential analytical geometry.
In fact, the equations that analytically define the aforementioned surfaces including
their intentional modifications allow to automatically generate refined FE models or
BE models, with which to perform the contact analysis as well as the stress analysis
of the gear pair or gear drive under consideration, thus avoiding the loss of accuracy
due to the development of solid models by computer-aided design (CAD) computer
programs.
Some topic is discussed using different methods, especially when each of the
considered methods allows to examine it from different points of view that allow to
better evaluate significant design aspects. In this framework, also some method that
may appear a little dated is described, especially when it allows us to better
understand the physical phenomenon underlying the topic under discussion.
Moreover, this choice has also the advantage of keeping alive the historical memory
of the pioneers who have traced new paths that have facilitated the task of devel-
oping the most sophisticated current calculation methods.
When necessary for the understanding of the phenomena of interest, hints are
made to methods of experimental analysis, and to numerical methods. However,
generally, these are not extended to the whole mechanical system, but only limited
to the single gear pair. This is because the attention wants to be called on the
understanding of the specific phenomena, which are proper of the same gear pair,
without the risk that they are more or less substantially altered or overshadowed by
Preface xv
the above-mentioned interaction effects with other components and groups of the
mechanical system to be examined.
The analytical and numerical solutions proposed here are formulated so as to be
of interest not only to academics, but also to the designer who deals with actual
engineering problems concerning the gears. This is because such solutions, though
sophisticated and complex (see, e.g., the solution of the matrix equations that define
the contact surfaces between the mating tooth flanks), have become immediately
usable by practitioners in the gear field thanks to today’s computers, which can
readily solve demanding equations. For the reader, moreover, these solutions pro-
vide the grounding needed to achieve a full understanding of the requirements laid
out in the standards applying in this area. In most cases, the analytical relationships
developed for use in the design analysis and/or in the response analysis (the latter
analysis, also called verification analysis, is used mainly by the standards) are also
presented in graphical form. This gives the reader an immediate grasp of the
underlying physical phenomena that these formulae explain and clarifies exactly
which major quantities must be borne in mind by the designer.
In dealing with certain topics, including for example the geometric–kinematic
aspects, fully developed exercises have been included to draw the reader’s attention
to the problems that are of greatest interest to the designer, as well as to clarify the
calculation procedure. The author has not considered it necessary to provide such
exercises in cases where the problems could be solved immediately with the rela-
tionships presented in each chapter, as it was felt that they would add nothing to an
understanding of the text.
Each topic is addressed from a theoretical standpoint, but in such a way as not to
lose sight of the physical phenomena that characterize various types of gears
gradually examined, up to hypoid gears, which are those with a more complex
geometry. The study of the gears proceeds in steps, starting from the geometric–
kinematic aspects, which are related to the cutting conditions of the teeth, and
gradually coming to the analysis of the loads and, subsequently, to the analysis of
stress and strain states, which allow to evaluate the tooth bending strength, surface
durability in terms of macropitting (pitting) and micropitting, scuffing load capacity,
tooth flank fracture load capacity, tooth interior fatigue fracture load capacity, wear
strength and service lifetime even under variable loading. The material is thus
organized so that the knowledge gained in the beginning chapters provides the
grounding needed to understand the topics covered in the chapters that follow.
This monothematic textbook, entitled Gears, is also intended for the students in
the course on Machine Design Theory and Methodology (Progettazione Meccanica
e Costruzione di Macchine) at the Università degli Studi di Roma “Tor Vergata”. It
consists of the following three volumes:
• Gears—Vol. 1: Geometric and Kinematic Design. This volume is organized in
14 chapters, which tackle the geometric and kinematic design of various types of
gears most commonly used in practical applications, also considering the
problems concerning their cutting processes.
xvi Preface
methods is done not only because, as we have said before, the memory of the latter
is not lost, but also and above all to demonstrate that the results obtained with their
use, although approximate, are not far from those obtained with the aforementioned
differential geometry and matrix methods, and therefore, have full design validity.
These stunning results obtained with these traditional mixed methods give credit to
their creators, who also knew the differential geometry and matrix methods equally
well, but could not use them because in their time the current calculation tools were
not available.
About the formulae, diagrams and figures taken or derived from the ISO stan-
dards, it should be pointed out that, in the form in which they are discussed and
presented here, they do not guarantee the accuracy of the results obtained. They are
to be considered as important points of reference for the calculations concerning the
determination of the geometric and kinematic parameters and the assessment of the
load carrying capacity of the gears as well as a clear demonstration of the usefulness
of the theoretical bases previously discussed. The guarantee of reliability of the
results with formulae, diagrams and figures drawn from ISO standards is in any case
given by the use of the original ISO standards, which moreover almost always add
to the standard specification, technical specification or technical report concerning a
certain topic an equally standard specification, technical specification or technical
report regarding interesting and clarifying examples of calculation for the same
topic, to which the user can directly draw.
The first two volumes of this monothematic textbook draw their origin from the
continuous and incessant stimulation of the numerous students of the author in the
three public universities where he taught. The thirst for knowledge led them to
treasure the following famous maxim of Aristotle, quoted by Diogenes Laërtius
“sῆ1 paideίa1 sὰ1 lὲm ῥίfa1 eἶmai pijqά1, sὸm dὲ jaqpὸm cktjύm”, i.e., “The
roots of education are bitter but the fruit is sweet.” It has stimulated the author’s
treatment of unusual and even little difficult subjects, which the students have
always shown to appreciable for their growth in view of their inclusion in the world
of profession or research. The author’s greatest satisfaction is to feel their gratitude
when, for some occasional reason, they come back to visit him. They have never let
the author miss their attachment and affection, and they still do not miss him. The
deepest and heartfelt thanks of the author go to them, who represent the most
authentic, sincere and vital component of the university.
Last but not least, the author wishes to express his affectionate, warm and
heartfelt thanks to Dr. Ing. Alberto Maria Vullo, his son, who in addition to
showing enthusiasm and meticulous care in the drafting of the drawings and graphs
of this textbook has validly collaborated in the preparation of its format. The author
then wishes to express his equally affectionate thanks to his wife, for her great
willingness to transfer in print format the textbook manuscript and formulae as well
as for her uncommon patience shown in having the author taken away from the
family needs, given the commitment made in almost four-year work. Finally, the
author’s thanks are also due to his first-born son, Dr. Ing. Luca Vullo, who even
xviii Preface
from a distance has been able to give a small but valuable contribution. These
heartfelt thanks are a small gesture of gratitude from the author to ask for an ocean
of excuses for having stolen time and affection from the whole family, to which,
however, he has the pleasure of expressing the deepest movement of his soul.
xix
xx Contents
xxvii
xxviii Symbols, Notations and Units
qFn Root fillet radius at point of contact of 30° tangent in normal section
(mm)
qP0 Offset or crown gear to cutter center (mm)
qaP Distance between the center of curvature of rounded tip corner and
cutter tip line (mm)
qb Radius of epicycloid base circle (mm)
qc Radius of centrode circle (mm)
qfP Radius of the rounded corner at the tooth tip of pinion-type cutter
(mm)
qfP Root fillet radius of basic rack for cylindrical gears (mm)
qlim Limit radius of curvature (mm)
qmb Tooth mean radius of curvature in lengthwise direction (mm)
qn;min Minimum principal radius of curvature in the normal plane (mm)
qt Radius of curvature in the transverse plane (mm)
q Radius of curvature (mm)
q Ratio of pitch radii of pinion and wheel (–)
q Position vector
r1;2 Tooth flank surface of pinion or wheel
r2 Tooth flank surface of face gear
rFlim Nominal stress number for bending (N/mm2)
rs Tooth flank surface of shaper
r Curve of pointed teeth for worm gear pairs
r Surface in three-dimensional space
R#f Sum of dedendum angles (°)
R#fC Sum of dedendum angles for constant slot width taper (°)
R#fM Sum of dedendum angles for modified slot width taper (°)
R#fS Sum of dedendum angles for standard taper (°)
R#fU Sum of dedendum angles for uniform depth taper (°)
Rs Supplementary angle of shaft angle (°)
R Plane tangent to a helicoid surface at any point
R Shaft angle, crossing angle of virtual crossed axes helical gear (°)
R Summation notation
RP Sum of powers (kW)
RT Sum of torques (Nm)
s0 Curve for which two branches of lines of contact occur for worm
gear pairs
s1;2 Angular pitch of pinion or wheel (° or rad)
sv1;2 Angular pitch of virtual pinion or wheel (° or rad)
siði¼1;2Þ Unit vector
s Angular pitch (° or rad)
s Curve of interference for worm gear pairs
u1;2 Angular displacement or angle of rotation of the pinion or wheel
about their axes (° or rad)
u2 Intermediate angle (°)
Symbols, Notations and Units xli
Abstract In this chapter, the main types of gearing used in practical applications
for transferring power between parallel, intersecting and crossed axes are described.
The tooth parts and some quantities that characterize the toothing are also described
and defined. The first law of gearing to be met to have a constant transmission ratio
is also recalled, and the main geometric and kinematic parameters of gear drives,
gear pairs, and gear wheels are defined. The geometrical parameters concerning the
toothing and its sizing are expressed in terms of transverse module, but indications
are given to express them in terms of transverse diametrical pitch. Moreover, the
ranges of the reference values of efficiency obtainable with the various types of
gears are provided, with indications for reaching the maximum values. Finally,
attention is focused on the parameters that define the precision of the teeth and the
accuracy grade of the gears, highlighting their influence on the transmission error,
under static and dynamic conditions.
1.1 Introduction
power transmissions must convert the input power characteristics (torque and
angular velocity at the input), adapting their to the output power characteristics
(torque and angular velocity at the output).
In the manufacturing industry, gear units are widely used for several reasons:
versatility of use, which makes them suitable to meet most of the requirements
posed by the vast majority of practical applications, even those with strong envi-
ronmental requirements (with the exception of non-low acoustic emission, which
can still be remedied by means of active and passive design solutions); inexpen-
siveness from the point of view of the costs of purchase and from that of running
costs; robustness and ease of maintenance; positive transmission of power; ability
to cover almost all application fields, from small miniature instrument installations
to huge powerful gears, like those used in gear drives for turbines, mining skips,
astronomical telescopes, cement kilns, etc.; widespread interchangeability thanks to
the standardization in size and shape of the gears.
Gear design therefore requires that two closely interdependent aspects be con-
sidered and analyzed: the geometric-kinematic aspect, and the aspect of the
mechanical strength. Both these issues are very complex and interconnected, but we
need to address them simultaneously if we want to assure the performance of
assembled drive gear systems. The first aspect involves the deepening of the
gearing geometry and kinematics (module, tooth profile, pressure angle, line of
action, length of path of contact, contact ratio, tip and root relief, crowning, end
relief, profile shift or addendum modification, meshing interference, cutting inter-
ference or undercutting, specific sliding, absolute and relative speeds, etc.) in order
to achieve predetermined design goals. The second aspect involves the analysis of
the static and fatigue strength as well as the scuffing and wear resistance, with the
purpose of ensuring the predetermined durability under the given operating con-
ditions. On these two aspects, we refer the reader well to traditional textbooks, such
as Buckingham [4], Merritt [26], Dudley [9], Giovannozzi [13], Pollone [34],
Henriot [15], Niemann and Winter [29, 30], Townsend [39], Maitra [25], Dooner
and Seireg [7], Jelaska [23], Litvin and Fuentes [24], and Radzevich [36].
The two above-mentioned aspects are deeply interconnected, constituting de
facto two sides of the same coin. In the description that follows, they are analyzed
separately, but only to make immediate and understandable basic concepts. In this
context, it should be kept in mind the fact that the choice of any quantity of interest
carried out by geometric-kinematic considerations inevitably has an impact on the
quantities correlated to the mechanical strength, and vice versa. Therefore, the
separate analysis of the two aspects must be continuously revised and governed by
a more general point of view, which allows to evaluate the mutual interdepen-
dencies, with the aim of being able to make the optimum choice of all the quantities
of interest in relation to the design goals to be achieved.
Of course, the gear designer must necessarily take into account the technological
and manufacturing aspects, which include not only the cutting technologies and,
more generally, the manufacturing process of the gears, but also the choice of
materials and heat treatments. These important issues are not directly addressed, but
only briefly recalled to memory in the framework of the two design aspects
1.1 Introduction 3
where xi ¼ 2pni =60 and x0 ¼ 2pn0 =60; ni and n0 are the rotation speeds of the
first driving shaft and respectively of the last driven shaft (xi and x0 in rad/s, ni and
n0 in min−1, i.e. rpm).
Figure 1.1 shows schematically a mechanical system where a gear unit, R,
carries out its design function. In this figure, M and U represent, in order, the
driving machine or prime mover and the driven machine, which are connected to
the gear unit, or gearbox, by means of suitable couplings, G, and shafts. Generally,
as design data, the rotational speed n0 and starting torque T0 of the driven machine
are know; they depend on the working requirements of this machine.
In this regard, we may be faced with very different cases as: (i), very high and
constant rotational speed, and torque also constant, but relatively low, as in the case
of a rotary compressor, for reasons of economy; (ii), very low and constant rota-
tional speed, and torque also constant but relatively high, as in the cases of belt and
screw conveyors, escalators, moving walkways, rotary kilns and mills for cement,
crushers, machine tools, etc.; (iii), low rotational speed and high torque for a vehicle
in the start-up phase, and alternately high rotational speed and low torque for the
same vehicle during the running regime (we have operating conditions similar to
the latter in the rectilinear advancement motion at different speeds, as occur, for
example, in some machine tools and rolling carpets).
In most cases, the driving machine, which is characterized by a given rotational
speed ni and by a given torque Ti , is not able to meet the above-mentioned
requirements. For example, a three-phase asynchronous electric motor has its own
operating rotational speed, which depends on the number of its poles and the
frequency of the network; an internal combustion engine, in condition of maximum
efficiency, must operate within a narrow range of variability of the rotational speed;
small and medium power turbines work well at very high rotational speeds; etc.
Therefore the gear unit, which is interposed between the driving and driven
machines, performs the function of changing, according to the fixed transmission
ratios, the rotational speed and torque available to the output shaft of the driving
machine, adapting their to the needs of the driven machine, represented by the
rotational speed and torque that must be available to the input shaft of the latter.
In cases where the driven machine must operate at variable rotational speed, the
use of a gear unit with fixed transmission ratio would be unsuitable. In these cases,
it is necessary to use variable speed drives, which allow to obtain a continuous
variation of the transmission ratio, within a given variability range. To meet the
same requirements, a static frequency converter can be used that, through the
energy distribution network, supplies the three-phase electric motor (this solution is,
however, very expensive). Alternatively, a gear unit with fixed transmission ratio,
interposed between the driving machine and the driven machine, and a static fre-
quency converter upstream of the electric motor can be used simultaneously (this
solution is cheaper than the previous one). In some cases, the optimal technical
solution is just the simultaneous use of a static frequency converter and a gear unit;
in these cases, regardless of the type of the driving machine, it is certainly
preferable to make the continuous variation of the rotational speed on the input
shaft, on which a lower torque is applied, and thus obtain higher torque, with a
corresponding low rotation speed, by means of a gear drive.
With reference to the Fig. 1.1, we consider separately the gear unit R. Since in a
mechanical system in steady-state equilibrium the sum RP of the mechanical
powers that come into play must be equal to zero, we can write the following power
balance equation:
RP ¼ Pi þ P0 þ Pd ¼ Ti xi þ T0 x0 þ Pd ¼ 0; ð1:2Þ
where Pi , P0 and Pd are respectively the input power, output power and total power
loss, that is the power dissipated for the various causes of loss, while Ti and T0 are
respectively the input and output torques. By convention, we assume the input
power as positive, and the output power and dissipated power as negative. The total
power dissipated is then expressed as the sum of three main contribution, i.e. as:
1.1 Introduction 5
Pd ¼ Pz þ Pb þ Pv ; ð1:3Þ
where Pz is the power dissipated at the contact between the mating teeth, Pb is the
power dissipated in the bearings (rolling bearings, plain bearings, etc.), and Pv is the
power dissipated for all other losses (ventilating effect, splash lubrication, contact
seals, idling, etc.).
Therefore, the total efficiency of the gear unit is equal to:
P0 P0 Pi Pd Pd
gt ¼ ¼ ¼ ¼1 : ð1:4Þ
Pi P0 þ Pd Pi Pi
RT ¼ Ti þ T0 þ Tr ¼ 0; ð1:5Þ
where, remaining the same the meaning of the quantities already introduced, Tr is
the reaction torque applied to the gear unit housing by the structure to which it is
connected. In dependence on the functional requirements of the driven machine, it
will be possible to have T0 greater than, less than or equal to Ti (the difference
between T0 and Ti is compensated by the reaction torque Tr ), and, consequently, n0
less than, greater than or equal to ni .
When the absolute value of the transmission ratio is grater then unity, that is
when jij [ 1; jni j [ jn0 j and jxi j [ jx0 j, we have a speed reduction ratio, and the
gear unit is a speed reducing gear drive. Vice versa, when the absolute value of the
transmission ratio is less than unity, that is when jij\1; jni j\jn0 j and jxi j\jx0 j,
we have a speed increasing ratio, and the gear unit is a speed increasing gear drive.
The ratio in ¼ n0 =ni between the output speed n0 and the input speed ni in a gear
power train is called the speed conversion factor; the ratio iT ¼ T0 =Ti between the
output torque T0 and the input torque Ti is called the torque conversion factor.
A transmission ratio jij 6¼ 1 should only arise when there is both speed and
torque conversion. In the ideal case of power losses equal to zero, the product in iT
be unitary, and therefore we will have in \1 only if iT [ 1 and vice versa. In the
actual case, which is always characterized by power losses, we have the following
relationship that correlates the quantities in , iT and i (for the sign conventions, see
below):
T0 g ni g
iT ¼ ¼ t ¼ t ¼ gt i: ð1:6Þ
Ti n0 in
6 1 Gears: General Concepts, Definitions and Some Basic Quantities
It should be noted that, in the literature concerning this topic, the reciprocal in of
i, that is in ¼ 1=i ¼ x0 =xi ¼ n0 =ni , is better known as transmission function, and is
often used improperly as transmission ratio.
Gear units or gear drives, widely used in practical applications, are combinations of
gear pairs arranged in a wide variety of ways in order to form a gear train or
powertrain. We can have two types of gear train: a simple train (in this type of gear
train, one or more idle gears are introduced between the driving and the driven
member of the gear system, in such a way that each idle gear is able to reverse the
direction a rotation, without changing the transmission ratio), and a compound train
(in this type of gear train, the intermediate shafts carry at least two separate gear
wheels or compound gear wheels having different numbers of teeth, and each stage
of reduction provides its own transmission ratio). A powertrain is conceived so as to
meet particular requirements, such as: gear ratios, total transmission ratio, direction
of rotation, position of the input, intermediate and output shafts, etc.
The gear trains composed of three coaxial members and precisely by two end
gear wheels (the sun gear and the ring gear, also called annulus gear or simply
annulus) with fixed axis and a frame (the planet carrier, also called carrier, spider
or arm) free to rotate around to the common axis of these end gear wheels, are
called simple epicyclic gear trains or simple planetary gear trains. The planet
carrier supports the movable axis (or axes) of one (or more) gear wheels called
planet gear(s), which mesh simultaneously with the sun gear and annulus gear. The
connection between this planetary gear train and any external mechanism is
accomplished only through these three coaxial members. We have a compound
epicyclic gear train or compound planetary gear train in the two cases of a
planetary gear train with compound planet gears (the planet gears themselves are
compound, instead of simple gears) or a combination of two or more simple
planetary gear trains. All the epicyclic gear trains (simple or compound) are studied
with reference to the basic train, that is the gear train obtained by helding at rest the
planet carrier, while the other two coaxial members rotate. By fixing the planet
carrier, the simple epicyclic gear train becomes a simple gear train, while the
compound epicyclic gear train becomes a compound gear train.
According to ISO 1122-1:1998 [18], a planetary gear train (also called planetary
gear, epicyclic gear or epicyclic gear train) is a combination of coaxial elements, of
which one or more are annulus gears, and one or more are planet carries, which
rotate around the common axes and support one or more planet gears which mesh
with the annulus gears and one or more sun gears. This definition is quite general,
as it includes both the simple planetary gear trains and the compound planetary gear
trains.
1.2 Gear Units and Gears 7
Whatever its configuration, each gear train can be thought of as made up of pairs
of toothed wheels (gear pair) which mesh mutually. The term toothed wheel
indicates a tooted member which, through the action of teeth which mesh in suc-
cession, rotates another or is driven in rotation by another. Generally, the term gear
indicates an elementary mechanism consisting of two toothed wheels, one which
drags the other and vice versa, and rotating around axes in invariable relative
position; it is therefore synonymous with gear pair. But it should be noted that,
often the term gear is used to indicate one of the two toothed wheels of the gear
pair, the one with larger number of teeth. In any case, we call pinion and wheel
respectively the smallest and the largest of the toothed wheels of the gear pair. To
avoid misunderstandings, for the member with larger number of teeth it would be
good to use the term gear wheel.
One of the gear design data consists of the specification of the relative position
of the shafts connected by the two toothed members of a gear pair. With reference
to the relative position of the two shafts (Fig. 1.2), their axes may be parallel
(coaxial gears constitute a particular case), intersecting or crossed (or skew), i.e.
neither parallel nor intersecting (axes at right angles constitute further particular
cases of intersecting and crossed axes). Correspondingly, the gear pair connected to
the two shafts is a parallel gear pair (parallel gears, Fig. 1.3a), an intersecting gear
pair (intersecting gears, Fig. 1.3b) or a crossed gear pair (crossed gears or skew
gears, Fig. 1.3c).
Parallel gears are generally cylindrical gears, consisting of a pair of cylindrical
toothed gear wheels. However, it should be noted that bevel gears with parallel axes
exist as well (see Niemann and Winter [30]). These cylindrical gears can be spur
gears, helical gears and double-helical gears. Spur gears may be external spur
gears (the term spur gears without qualification implies an external spur gear),
internal spur gears, and spur rack-pinion pair. The term spur gears implies also
Fig. 1.2 Relative position of the axes: a parallel axes; b intersecting axes; c crossed or skew axes
(a and R are respectively the center distance and shaft angle)
8 1 Gears: General Concepts, Definitions and Some Basic Quantities
Fig. 1.3 Types of gears and quantities defining the relative position: a parallel gear; b intersecting
gear; c crossed gear; d worm gear; e center distance in a parallel cylindrical gear; f shaft angle in an
intersecting gear; g shaft angle in a crossed gear
that the teeth are parallel to the axes, so that their profiles in the transverse sections
(i.e. sections with planes perpendicular to the axes) do not change in the longitu-
dinal direction from end to end.
Helical gears can also be external helical gears (also here the term helical gear
without qualification implies an external helical gear), internal helical gears and
helical rack-pinion pair. Helical gears have the same use of the spur gears, but their
teeth have helical or screw shape; therefore transverse profiles of teeth are the same
in the various transverse sections, but change in the angular position along the
longitudinal direction from end to end.
Double-helical gears are normally external, even if internal double helical gears
are possible but not usual. In the two component members of these types of gear,
toothing is composed of two portions side by side, of opposite inclination. The teeth
may be continuous or separated by a gap. Double helical gears work like two
single-helical gears having helix of opposite direction, and running side by side. In
the gear pairs formed by a double-helical pinion and a double-helical rack, the first
meshes with the second, which can be thought of as an external or internal wheel of
infinite radius. Note that triple-helical gears have been made, but they are not used
1.2 Gear Units and Gears 9
crossed shafts, usually, but not necessarily, perpendicular. Rotation of the worm,
which simulates a motion of retrograde linear advancement of a rack, determines
the consequent rotation of the worm gear. Generally, the profiles of the worm
threads have the same geometry as those of helical and spiral gears, and the
geometry of the worm-wheel, also called worm gear, is similar to that of a
helical gear, except that the teeth are curved and concave to envelop the worm.
The contact between the worm threads and the worm-wheel teeth is a line
contact, but between them there is a sliding motion much higher compared to
what occurs in other types of gears. Sometimes the worm-wheel or both worm
and worm-wheel are modified to better envelop the mating member, and we can
have (see ISO 1122-2: 1999(E/F) [19]): a single enveloping worm gear pair,
where an enveloping worm is in meshing with a cylindrical worm wheel; a
single enveloping worm gear pair, where a cylindrical worm is in meshing with
an enveloping worm-wheel; a double-enveloping worm gear pair, where an
enveloping worm is in meshing with an enveloping worm-wheel. This last case,
which is the most demanding from the technological point of view, allows for a
greater area of contact, but cutting processes and mounting techniques more
precise are required. The term globoidal worm gear pair is reserved for a worm
gearing in which both worm and worm-wheel have modified form, and are
concave on their respective axial sections.
– Hypoid gears: they are spiral bevel gears, the axes of which, however, do not
intersect. This type of gear is a special case of hyperboloid gear, to which it
resembles; it is generated using cutting processes somewhat similar to those of
the spiral bevel gears. Hypoid gears are usually used in motor vehicles (cars and
trucks), with the axis of the pinion disposed below of the crown wheel. Note
that, from the theoretical point of view, hyperboloid gears represent the general
case of which cylindrical and bevel gears are particular cases; in these hyper-
boloid gears, teeth are formed on hyperboloids of revolution.
These type of gears do not exhaust the possible cases, but cover a substantial
portion of the most significant practical applications. Gears for special purpose,
such as pump gears, timing gears, instrument gears, sprocket wheels, and so on,
remain outside of the classification above. These special gears must have specific
characteristics and must meet given design requirements. They include non-circular
gears, which are used for the generation of a prescribed transmission function or as
a generating driving mechanism to modify the displacement function or the velocity
function. These types of gears are not discussed in this monographic book.
Now let’s go back to the specification of the relative position of the shafts
connected by the gear pair, and see its interdependence with respect to the direction
of rotation of the shafts, which is another of the design date of the gears. For parallel
cylindrical gears, the relative position of the axes is completely defined only by the
center distance between the shafts (Figs. 1.2a and 1.3e), and for an external gear
there is only one possible direction of relative rotation, once fixed the direction of
rotation of one of the two members of the gear pair. Since their axes are parallel,
the shaft angle R, that is the other of the two quantities that, in the general case,
1.2 Gear Units and Gears 11
The designer of gears and gear transmission systems must have a clear and precise
knowledge of efficiency achievable with various types of gears that he must design.
Among other influences, with the other variables that remain the same, the effi-
ciency of the gears and gear systems is influenced by the value of the transmission
ratio actually achieved. The selection of the appropriate gears or the appropriate
gear systems for different practical applications can be made, so as accurately and
precisely by the designer, after reasoned comparison of the efficiency of the dif-
ferent gear types, taking account of the transmission ratio to be realized.
The analytic determination of efficiency of the gears having involute tooth
profiles (these are the gears that interest in this textbook) is a problem that is
generally not easy to be solved (see: Panetti [32]; Ferrari and Romiti [11]; Scotto
Lavina [38]). We will focus on this issue with more detail in the chapters on the
various types of gears used in most current practical applications. Here we want to
give the reader a general guideline on the variability ranges of efficiency values
achievable with the various types of gears. The Eq. (1.4) shows that the total
efficiency gt of a gear transmission system can be simply calculated as the per-
centage value of the quotient of the output shaft power divided by the input shaft
power, i.e. gt ¼ ðP0 =Pi Þ100%: We have then P0 ¼ ðPi Pd Þ, i.e. the output power
is the difference between the input power, Pi , and the power total losses, Pd .
The Eq. (1.3) then shows that the power total losses in gear transmission systems
mainly depend on the tooth contact friction losses, Pz , bearing losses, Pb , and
lubrication churning losses, Pv . Friction losses depend on the gear design and their
manufacturing process, gear size, transmission ratio, pressure angle, and coefficient
of friction. For the main types of gear pairs usually used in practical applications, a
rough estimate of the efficiency range associated with the tooth friction can be made
using the guideline data shown in Table 1.1, which also show the usual variability
range of the transmission ratio, i, and the usual pitch line velocity, v (in m/s).
12 1 Gears: General Concepts, Definitions and Some Basic Quantities
It is to be noted that the values of efficiency due to tooth friction losses shown in
Table 1.1 are valid only for single tooth meshes. For gear transmission systems
consisting of a given number of gear pairs of the same type, the efficiency related to
tooth friction is obtained by multiplying the efficiency of the single gear pair by the
number of gear pairs. In the case in which the gear unit is constituted by different
gear pairs, the efficiency is obtained by multiplying the various efficiencies of the
individual gears pairs.
Also it is to be noted that the theoretical efficiency ranges of the various gear
types shown in Table 1.1 do not include bearing and lubricant losses. Furthermore
they assume ideal mounting and assembly conditions in regard to axis orientation
and center distance. Deviations from these ideal conditions will downgrade the
efficiency values.
Bearing losses, Pb , and churning losses, Pv , are almost independent of the type
of gear and transmission ratio. The determination of these two losses is not easy, so
at least in the initial stage of a new gear drive design, the designer would do well to
base the design on the experience gained in similar cases. Very often these two
losses are being combined into a single loss, and then they are evaluated globally.
In any case they are respectively influenced by the types of bearings used and the
tangential velocity of the gears passing through the lubricant contained in the oil
sump.
It is not the place to delve into this interesting topic. To get a rough idea of what
happens in the rolling bearings, we refer the reader to [31]. Always to have a rough
idea about what is happening in the plain bearings (radial and axial hydrodynamic
bearings, radial and axial hydrostatic bearings, pneumostatic and aerostatic bearings
and magnetic bearings) we refer the reader to Niemann et al. [31], and Chirone and
Vullo [6].
1.3 Efficiency of the Gears 13
Now we return to focus our attention only on friction losses, Pz , which occur at
the contact between the mating teeth. From data shown in Table 1.1, we can deduce
the following (see also Dubbel [8]):
– Cylindrical spur gears can guarantee much higher efficiencies compared to other
type of gears; their efficiency varies between 98 and 99%. Table 1.1 shows the
same efficiency values for cylindrical, helical and double helical gears. In this
regard, however, it is to be noted that, with the same accuracy grade and the
same operating and assembly conditions, the efficiencies of these two types of
gears are slightly smaller than the efficiency of cylindrical spur gears. For the
cylindrical spur gears, the highest values of efficiency are related to lower gear
transmission ratios, while the cylindrical helical and double helical gears can
work with very high pitch line velocities and can reach their higher efficiency
with maximum transmission ratios up to 10:1.
– Straight bevel gears have efficiencies a little lower than those of cylindrical spur
gears and, as for the latter gears, the highest efficiency values are correlated to
the lowest values of the transmission ratio. With the same accuracy grade and
the same operating and assembly conditions, spiral bevel gears have efficiencies
greater than those of the equivalent straight bevel gears; this is due to tooth
spiral shape, from which even less noise and vibration are the result.
– The efficiencies of the crossed cylindrical helical gears, worm gears, and hypoid
gears are much lower than those of the gears examined above, having the same
accuracy grade and the same operating and assembly conditions. In Chaps. 10,
11 and 12, where these gear types are respectively analyzed in detail, we will
deepen the subject concerning their efficiencies, and we will provide the
information necessary to obtain, if possible, their maximum values, through
optimization of the design choices. We will provide also reliable relationships to
calculate their values as a function of the quantities on which they depend.
However, we must already emphasize the high transmission ratios attainable
with hypoid gears (up to 200:1), with efficiencies that can be considered suffi-
ciently satisfactory.
– Although in this textbook we do not speak of cycloidal gears, we considered it
appropriate to provide the data relating to these gears, in order to have the terms
of comparison with the gears having involute tooth profiles. As Table 1.1
shows, under normal working conditions and with the same accuracy grade and
the same operating and assembly conditions, cycloidal gears are characterized
by lower efficiencies than those of the gears with involute tooth profiles. In
addition, cycloidal gears allow to obtain much higher transmission ratios, and
can work in very high efficiencies at relatively high values of the transmission
ratio, above 30:1.
Finally it should be stressed the fact that, for the various type of gears taken into
consideration, the data summarized in Table 1.1 constitute the guideline values of
the efficiencies obtainable for a single gear pair of these gears. In addition, these
guideline values only consider the friction losses at the contact between mating gear
14 1 Gears: General Concepts, Definitions and Some Basic Quantities
Fig. 1.4 Pitch cylinders and common pitch plane; 1 and 2, axes of pinion and wheel
1.4 Basic Law of Mating Gear Teeth 15
Quantities shown in Fig. 1.4 are: r1¼d1 =2 and r2¼d2 =2, the pitch radii and pitch
diameters of the two cylinders; x1 and x2 their angular velocities (x1¼2pn1 =60 and
x2¼2pn2 =60, where n1 and n2 are the rotational speeds); vt¼x1 r1¼x2 r2 , the pitch
line velocities or tangential velocities; a¼r1þr1 = ðd1þd Þ=2, the center distance; z1
and z2 their number of teeth. Note that the subscripts 1 and 2 indicate, respectively,
quantities referring to the pinion (the smaller wheel) and the wheel (the larger
wheel). Note also that for external gears a, z1 and z2 are considered positive,
whereas for internal gears, a and z2 have a negative sign, and z1 has a positive sign.
The transverse sections of the two pitch cylinders with a transverse plane, i.e. with a
plane normal to their axes, are the pitch circles of two flat gear wheels by which the
kinematics of their relative motion is described. Using these pitch circles we also
study the shape of the teeth profiles which realize a geometric coupling with a
predetermined transmission ratio between the two members of the gear pair.
The basic requirement that the geometry of the gear-tooth profile needs to satisfy
is the guarantee of a transmission ratio exactly constant. It is noteworthy that a
geared mechanical transmission may also be made with a specified variable
transmission ratio, but this type of geared drive is not among those considered in
this textbook. Notoriously this requirement is satisfied by two mating profiles that
perform a conjugated action. The basic law of conjugated action between two
mating profiles states that as the profiles rotate, the common normal to the profiles
at the point of contact must always intersect the line of centers at the same point C,
called the pitch point (Fig. 1.5). This basic law of conjugate action is known as the
first law of gearing.
From a more general point of view, mating gear teeth acting against each other to
produce a rotational motion are similar to cams. The profiles of cams or teeth must be
designed so as to get a constant transmission ratio during meshing, and then to have a
conjugate action. In theory, it is possible to select any arbitrary profile for one cam or
Fig. 1.5 Basic law of conjugate profiles, normal of contact, and velocities
16 1 Gears: General Concepts, Definitions and Some Basic Quantities
tooth and then to find the correspondent theoretical profile for the meshing cam or
tooth which will give a conjugate action. Among the possible pairs of profiles,
involute profiles are those of greatest interest; in fact, these involute profiles, with few
exceptions, are used universally for gear teeth. Our attention is focused on these
profiles, even if cycloidal profiles, circular-arc profiles or Wildhaber-Novikov pro-
files where used in the past and in some specific cases are still used (see: Giovannozzi
[13]; Niemann and Winter [29]; Litvin and Fuentes [24]).
Regarding the discussion that follows, we assume the teeth (or cams) to be
perfectly shaped, perfectly smooth and absolutely rigid. Of course, these assump-
tions are unsuitable in practice, due to errors in cutting and finishing operations of
the gears, mounting and assembly errors as well as to elastic deformations related to
workloads.
Let us now consider two flat mating profiles of two conjugate members 1 and 2,
that rotate about their axes O1 and O2 (Fig. 1.5), with angular velocities x1 and x2 .
When one profile pushes against the other, the point of contact P occurs where the
two conjugate profiles are tangent to each other, and the forces exchanged at every
instant are direct along the common normal n to the two profiles at point P. The line
n is the line of action, that is the line representing the direction of action of the
forces exchanged between the two mating profiles, under the hypothesis that there
are no friction losses.
This line of action will intersect the line of centers O1 O2 at the same point
C. The transmission ratio i ¼ x1 =x2 between the two conjugate members is
inversely proportional to their radii to the point C. Circles drawn through point
C from each center O1 and O2 are the loci of the instantaneous center of rotation in
relative motion between the two conjugate members; they are called centrodes or
pitch circles, which are synonymous. The radius of each pitch circle is the pitch
radius, while point C is the instantaneous center of rotation or pitch point.
Therefore, the rotational motion of the two conjugate members about their
respective axes O1 and O2 can be replaced by a motion cinematically equivalent:
the rolling motion without sliding of the pitch circle related to the member 2 to the
pitch circle related to the member 1. Obviously, during rolling of the pitch circles
on each other, at the current point of contact P a sliding also occurs, unless
P coincides with C. It should be noted that the centrodes or pitch circles are the
transverse sections of the axodes or pitch cylinders previously defined.
In order to have a constant transmission ratio, the pitch point C must always
remain the same, that is all the instantaneous lines of action for every instantaneous
point of contact must pass through the same point C. Obviously, manufacturing
errors and tooth deformations will cause slight deviations in transmission ratio, but
the analysis of tooth profiles is based on theoretical curves that meet these
requirements. In the case of involute profiles, all points of contact occur on the same
straight-line n, that is all straight-lines normal to the profiles at the point of contact
coincide with the line n, and thus we have a transmission with uniform rotational
motion and constant direction of the forces exchanged.
1.4 Basic Law of Mating Gear Teeth 17
Since the angular velocities x1 and x2 can be expressed as the ratio of the
tangential velocity at any point of the two conjugate members 1 and 2 and the
distance of this point from the centers of rotation O1 and O2 of the same members,
with the symbol shown in Fig. 1.5 we have:
vP1 v1 vn1
x1 ¼ ¼ ¼ ð1:7Þ
rP1 r1 rb1
vP2 v2 vn2
x2 ¼ ¼ ¼ : ð1:8Þ
rP2 r2 rb2
By the similarity of the triangles CT1 O1 and CT2 O2 of Fig. 1.5, we obtain:
rb1 rb2
cos a ¼ ¼ ð1:9Þ
r1 r2
and then
Since the mating profiles must be in contact without backlash and no penetra-
tion, vectors vn1 and vn2 must be equal, i.e. they should have the same direction and
the same absolute value. It is worth nothing that vectors are indicated in bold letters.
From equality vn1 ¼ vn2 , taking into account the Eqs. (1.10) and (1.11), it follows
that also the tangential velocities must be the same on the line of centers, i.e.
v1 ¼ v2 ¼ vt . This is only possible at the operating pitch point C, where the pitch
circles touch and are tangent with each other. Then we will have:
x1 vt vt r2
¼ = ¼ : ð1:12Þ
x2 r1 r2 r1
In this more general case, all transverse sections are similar in so far as tooth
profile is concerned (this profile may be identical with that of spur gear), but change
in the longitudinal position of the transverse plane of section, producing a corre-
sponding change in the angular position of the individual tooth sections. The
position of the point of contact of the teeth on any transverse section will vary from
one end of the gears to the other, and the line of contact is no longer parallel to the
teeth axes, but inclined. However, there is no longitudinal sliding, and the motion is
transmitted uniformly even if the aforementioned law of the mating profiles is
satisfied only for one position of the point of contact corresponding to a given
transverse section. As shown in Fig. 1.6, the profiles obtained by sectioning the
tooth with transverse planes 1,2,3 and 4 come into contact in succession. The point
of contact P moves with velocity vx ¼ vt =tg b along the face-width, where b is the
helix angle (see Chap. 8). Note that the position of the point of contact P in each
transverse section remains constant.
Whatever the geometry of the profile (involute profile, cycloidal profile,
circular-arc profile, etc.), we have:
• Transmission ratio, i: the transmission ratio of a gear pair is the ratio between
the angular velocity of the driving wheel and the angular velocity of the driven
wheel, i.e.:
x1 n1 r2 z2
i¼ ¼ ¼ ¼ : ð1:13Þ
x2 n2 r1 z1
It should be noted that, from here on, unless we say otherwise, we assume that
members 1 and 2 of a gear pair are respectively the driving and driven members.
For an external gear pair, the directions of rotation of the pinion and wheel are
opposite, whereby, z1 and z2 being both positive, the transmission ratio is negative;
obviously also the quantities x2 , n2 and r2 are to be considered as negative
quantities (conventionally, clockwise rotation is regarded as positive, and coun-
terclockwise rotation as negative). Vice versa, in an internal gear pair, the direction
of rotation of both gear wheels is the same, for which the transmission ratio is
positive, as z2 is conventionally negative. More generally, in a gear train, the
transmission ratio is the ratio between the angular velocity of the first driving wheel
and the angular velocity of the last driven wheel. Therefore, the transmission ratio
of a gear train is the product of individual transmission ratios of component gears
pairs. However, it should be remembered that, especially in cases where misun-
derstandings are not possible, only the absolute value of the above mentioned
quantities are considered, without taking into account their conventional sign.
Speed reducing gear pairs (or speed reducing gear trains) and speed increasing
gear pairs (or speed increasing gear trains) are those for which the angular velocity
of the second (or last) driven wheel is respectively less than or greater than that
of the first driving wheel. With the exception of special cases (for example,
1.4 Basic Law of Mating Gear Teeth 19
the elliptical gears), the cases of most interest are those for which the transmission
ratio is constant.
• Gear ratio, u: the gear ratio of a gear pair is the ratio between the number of
teeth of the wheel and that of the pinion, i.e.:
20 1 Gears: General Concepts, Definitions and Some Basic Quantities
z2 r2
u¼ ¼ : ð1:14Þ
z1 r1
This gear ratio is positive for external gear pairs, while it is negative for internal
gear pairs. In absolute value, in the case of driving pinion we have u ¼ i, while in
the case of driving wheel we have u ¼ 1=i:
• Isogonality factor, gx : this factor is an index of the uniformity of the trans-
mission of motion; it is defined as:
i x 1 z1
gx ¼ ¼ : ð1:15Þ
u x 2 z2
Under ideal conditions, i.e. for gears without errors or other deviations from the
theoretical conditions, and with profiles of the teeth in accordance with the laws of
the mating profiles, gx ¼ 1. Any deviation of this factor from the unit is an indi-
cation of non-uniformity of the transmission of motion.
• Constance torque factor, gT : this is an index of the constancy of the torque
transmitted; it is defined as:
iT T2 z1
gT ¼ ¼ ; ð1:16Þ
u T1 z2
where iT ¼ T2 =T1 is the already defined torque conversion factor, i.e. the ratio
between the output torque T2 and input torque T1 ; thus iT is the specialization of the
Eq. (1.6), valid for a gear train, to a gear pair. If during the gear meshing the
direction and intensity of the friction force change, the output torque fluctuates
accordingly, even if the input torque remains constant. Consequently also iT fluc-
tuates, and the deviation of the factor gT from the unit gives the measure of the
irregularity of the torque transmission.
• Transmission function, in ¼ 1=i: it is advisable to introduce this function, pre-
viously defined, in the case of gears for special purpose, such as for example
non-circular gears, which may have parallel or intersecting axes and are char-
acterized by transmission ratio that is not constant. Being in the Eq. (1.13)
x1 ¼ du1 =dt and x2 ¼ du2 =dt, where du1 and du2 are respectively the dif-
ferential angular displacements of the driving and driven members about their
axes, this transmission function is given by the following relationship:
1 du2
in ¼ ¼ : ð1:17Þ
i du1
It therefore defines the relationship that correlates the angular position of the
driving member with the corresponding angular position of the drive member.
1.5 Tooth Parts and Some Quantities of the Toothing 21
Figure 1.7 shows the main parts of a tooth having any profile, and some quantities
characterizing a cylindrical spur gear. We define axial and transverse planes
respectively planes containing the axis and perpendicular to the axis of the gear;
instead we define normal planes the planes perpendicular to the axis of the tooth.
Obviously transverse and normal planes coincide only in the case of straight teeth,
while they are different in the case of curved teeth. We call blank the member of
gear pair (pinion or wheel) before the cutting of the teeth. Top land or crest of the
tooth is the portion of tip surface (also called addendum surface or outside surface)
between the opposite tooth flank surfaces of the same tooth, while the bottom land
is the portion of root surface or dedendum surface between the roots of two
adjacent teeth. Rim is the portion of a gear wheel that provides full support for the
tooth roots. Tooth flanks are the portions of the surfaces of a tooth between the
addendum surface and the dedendum surface.
The reference surfaces are imaginary conventional surfaces with respect to
which the dimensions of the teeth are defined. Instead, as we have seen in the
previous section, the pitch surfaces of a given gear pair are the geometrical surfaces
described by the instantaneous axis of rotation of the relative motion of the mating
gear member, in relation to the gear member under consideration. We also said that
pitch surfaces and axodes are synonymous.
Fig. 1.7 Main parts of a tooth, and some quantities of a spur gear
22 1 Gears: General Concepts, Definitions and Some Basic Quantities
The pitch surfaces of parallel or intersecting gears are those ideal surfaces of
revolution, having the same axes of the two gears, which touch each other along a
common line of contact and roll on each other without sliding when the two wheels
to which they are rigidly connected rotate with angular velocities inversely pro-
portion to their pitch diameters. For parallel cylindrical gears (spur or helical gears)
and intersecting gears (straight or curved-teeth bevel gears) these pitch surfaces are
respectively cylinders and cones, and are therefore termed pitch cylinders and pitch
cones. Instead, crossed gears have no pitch surfaces understood with the above
defined meaning, i.e. ideal surfaces which rotate with the gears and roll together
without slide; for each of these crossed gears the term pitch surface has a different
meaning, which will be explained case by case in the next chapters.
The pitch surface divides the tooth flank (left flank and right flank) in two
portions: the addendum flank and dedendum flank respectively between the tip
surface (or addendum surface) and pitch surface, and between the pitch surface and
root surface (or dedendum surface). The tip is the edge where the flank meets
theoretically the crest; in practice, however, the flank and crest surface are con-
nected by a connecting surface, for which we still use the term tip, but with a larger
meaning, which indicates the portion of the tooth surface adjacent to that edge. In
this respect, often the tip edges are rounded or chamfered, in order to facilitate the
engagement of the mating gear members. These two types of modification of the tip
edge, the first characterized by a given radius of curvature, and the second by a
given depth, are called respectively rounding and chamfering. Instead the fillet is
the portion of the flank surface that connect the usable flank surface to the root
surface. The geometry of the fillets of teeth of one member of the gear pair can have
any shape, with the only condition to have no interference with the tip edges of
teeth of the mating gear during the meshing.
The pitch surface intersects the flank surface along the pitch line, which is a
straight line segment parallel to the axis in the spur gears, a straight line segment
inclined with respect to the axis in the helical gears, and a segment of a planar curve
in the spiral gears. The latter curve is also called tooth spiral. Instead, in the case of
gear having helical teeth or threads the intersection line above is called tooth helix
or thread helix. We define tooth trace the line of intersection of the tooth flank
surface with the reference surface. Instead, the flank line is the line of intersection of
the tooth flank surface with a surface of revolution coaxial with the reference
surface.
The helix angle (or spiral angle) is the angle between a tangent to a tooth helix
(or tooth spiral) at any point and a pitch-surface generatrix passing thought that
point. A normal helix (or normal spiral) is a helix (or spiral) lying on the pitch
cylinder and having a helix angle (or spiral angle) complementary to that of the
helix (or spiral) which form the tooth helix (or tooth spiral). The lead angle of a
helix (or spiral) is the angle between a tangent to the helix (or spiral) and a
transverse plane. The lead of a helix (or spiral) is the axial advance per revolution,
i.e. the axial distance between similar points in successive convolution.
The intersections of the pitch surface, tip surface and root surface with a
transverse plane, i.e. a plane perpendicular to the gear wheel axis, are respectively
1.5 Tooth Parts and Some Quantities of the Toothing 23
the pitch circle, addendum circle and dedendum circle. Tooth profiles are the
intersections of the tooth flank surface with any defined surface which also cuts the
reference surface, that is the conventional surface with respect to which the
dimensions of the teeth are defined (for a cylindrical gear it is the pitch surface of
the gear wheel meshing with the basic rack). The intersections of the tooth flank
surface with a transverse plane, normal plane (plane perpendicular to the tooth axis)
and axial plane (plane containing the gear wheel axis), are respectively the trans-
verse profile, normal profile and axial profile.
It is also necessary to keep in mind other possible intentional modifications of
the tooth shape, among which those described hereunder deserve specific mention.
The tip relief or tip-easing and root relief or root-easing are intentional modifica-
tions of the tooth profile at the expense of the material thickness (a small amount of
material is removed from the flank towards the crest or the root), whose purpose is
to soften the beginning of the contact between the mating flanks, during the
meshing.
The undercut is another intentional modification: it concerns the fillet and results
in a removal of material (this can be obtained for example with a cutting tool fitted
with protrusion), aimed to facilitate any subsequent operation to the cut. It is also
important to remember the barrelling or crowning, that is the modification by
which the teeth are slightly and progressively thinned from the centerline towards
their ends, in order to avoid concentration of loading at the ends. The end relief
serves the same purpose: it consist of a progressive reduction of the teeth thickness
at their ends, on a narrow portion of the face width.
The main quantities which characterize a toothing and a gear are the following
(other quantities will be introduced at the appropriate time):
• Circular pitch or transverse circular pitch, p: it is the distance between corre-
sponding profiles (both right profiles or both left profiles) of two successive
teeth measured along the pitch circle. Indicating with d, z and m, respectively,
the reference diameter (i.e., the diameter of the reference circle), number of
teeth and module, we will have (p and m are given in mm):
pd
p¼ ¼ pm: ð1:18Þ
z
dedendum, thickness and so on. It should be remembered that the term module,
without qualifying adjective, indicates the transverse module.
It is noteworthy that, in the English-speaking word, the reference quantity is not
the transverse module, but the transverse diametral pitch, P, which is defined as the
ratio between the number p and transverse pitch, expressed in inches, or as the ratio
between the number of teeth and the reference diameter, expressed in inches.
Therefore, m, p and P are related by the relationship: pP ¼ p or m ¼ 25:4=P. It also
to be noted that, for the involute toothing, the pitch circle and reference circle may
be different. Finally, it is to keep in mind that sometimes we use the angular pitch s
(in rad), which is the angle subtended by the circular pitch; thus it is defined as the
ratio between the round angle and the number of teeth, i.e. s ¼ 2p=z.
• Reference diameters, d1 and d2 , and center distance, a: of course, to have a
theoretically correct engagement, the pitches of the pinion and wheel must
coincide. Therefore, from Eq. (1.18) we obtain the following relationship.
where da1 and da2 are the addendum (or outside or tip) diameters of the pinion and
wheel. For the involute toothing this diameters are given by (Fig. 1.8):
where c1 and c2 are the bottom clearances or simply clearances between pinion
and wheel, and between wheel and pinion: in other words, c1 is the minimum radial
distance between the tip surface of the pinion and root surface of the mating wheel,
26 1 Gears: General Concepts, Definitions and Some Basic Quantities
while c2 is the minimum radial distance between the tip surface of the wheel and the
root surface of the mating pinion. Clearances c1 and c2 , which usually are included
in the range ð0:1 0:3Þm, can be expressed by:
c1 ¼ h1 hw ¼ a da1 þ df 2 =2 ð1:25Þ
c2 ¼ h2 hw ¼ a da2 þ df 1 =2: ð1:26Þ
s þ e ¼ p: ð1:28Þ
For involute toothing, with reference to the basic rack and in the theoretical case
of meshing without circumferential backlash and penetration of material, s and
e will have the same nominal value, equal to half transverse pitch, i.e. s ¼ e ¼ p=2.
We define the tooth thickness half angle as the half of the angle between the tooth
traces of a tooth, and the space-width-half angle as the half of the angle between the
tooth traces of a space-width.
In the actual operating conditions, the transverse thicknesses s1 and s2 will be
less than p=2, while the space-widths e1 and e2 will be greater than p=2, for which
there will be a circumferential backlash, jt ; of course, all these quantities are
measured as lengths of the arcs of the pitch circle. The circumferential backlash
represents the possible movement of one gear, measured along the pitch circle,
relative to the other, i.e. the length of the arc of pitch circle of which a wheel can
rotate so that its non-operating flanks are in contact with the non-operating flanks of
the mating gear remained in its position.
Deviations of the values of s1 ; s2 ; e1 and e2 from the nominal values, respec-
tively less for the tooth thicknesses and in addition for the space-widths, have the
purpose of allowing a correct lubrication and avoid the interlocking danger of a
tooth into the space-width of the other, due to effects of thermal expansions and
assembling inaccuracies. This circumferential backlash is given by:
jt ¼ p ðs1 þ s2 Þ: ð1:29Þ
1.5 Tooth Parts and Some Quantities of the Toothing 27
more general meaning, which includes not only the true tip relief, but also any other
intentional modification in the gear tooth profile close to its tip. These tip relief
circles are the circles that determine, with their intersections with the line of action,
the extreme points of the useful portion of this line, and hence the effective path of
contact and its length.
Beyond these circles, the teeth extend up to the addendum circles, but with
profiles (whether rounding or chamfering or tip relief) which theoretically do not
take part to the contact and have the purpose of facilitating the engagement and
avoid scratches and scuffing. The portions of the profiles between the tip relief
circles and addendum circles can have any geometry, as long as this is compatible
with the need to avoid the interference, only if the cut of the teeth is performed by
means of milling cutter designed in order to produce the desired geometry of the
profiles near to the tips. If instead the cutting of the teeth is performed by envelope
cutting (generation by rack-type cutters, pinion-type cutters or shapers and hobs),
the geometric shapes of the tip profiles is automatically defined by the shape of fillet
curves of the cutting tools, i.e. rack-type cutters, shapers and hobs.
Figure 1.8 refers to the theoretical case in which the involute profile is extended
to the addendum circle. It shows that ga ¼ gf 1 þ ga1 ¼ ga2 þ gf 2 , where ga2 ¼ gf 1
and ga1 ¼ gf 2 are respectively the length of recess path and length of approach
path. The lengths of approach path and recess path are the lengths of those parts of
the path of contact along which approach contact and respectively recess contact
occur. With driving pinion, in the approach path of contact, a point of the deden-
dum flank of pinion is in contact with a corresponding point of the addendum flank
of the driven wheel, while during the recess path of contact, a point of the
addendum flank of pinion is in contact with a corresponding point of the dedendum
flank of the wheel. This fact is highlighted by a double subscript, the first of which
(a or f) refers to the addendum or dedendum, while the second (1 or 2) refers to
driving pinion or driven wheel.
The path of contact starts at the point A, where the dedendum flank of one tooth
of the driving pinion meets for the first time the tip of the driven wheel, and ends at
the point E, where the tip of the same tooth of the pinion leaves the contact with the
dedendum flank of the wheel. Points A and E on the line of action match on the
pitch circle of the pinion points A and E . The arc Ad E is called transverse arc of
action or simply arc of action, and its length q is called length of the transverse arc
of action or length of path of rotation, while the corresponding angle is called
transverse angle of action and its length is called length of the transverse angle of
action. It should be noted that the arc of action, q, is the arc measured on the pitch
circle, during which two conjugated profiles remain in contact. Therefore it is the
arc of pitch circle through which a tooth profile moves from the beginning to the
end of contact with the mating profile.
We can write q ¼ ðqa þ qr Þ, i.e. the arc of action is the sum of the arc of
approach, qa , and the arc of recess, qr . The arc of approach, qa , and arc of recess,
qr , are therefore the arcs of the pitch circle through which a tooth profile moves
from the beginning of contact to the pitch point and, respectively, from the pitch
1.5 Tooth Parts and Some Quantities of the Toothing 29
point until contact ends. The generally used convention is to measure the arc of
approach on the driving wheel pitch circle (usually, the pinion), and the arc of
recess on the driven wheel pitch circle. However, it is obviously equivalent to
measuring these arcs of approach and recess on one or the other of the two pitch
circles, as these roll one on the other, without sliding. In order to ensure continuity
of transmission of motion, the following condition must be met: q ¼ ðqa þ qr Þ p.
In this way a new pair of mating teeth will be in engagement before the previous
one leaves the contact.
Figure 1.8 also highlights the active portions of the two profiles, i.e. the active
profiles AKc 1 and EK c 2 . It is noteworthy that the profiles of the dedendum flanks
must be shaped so that the tips of the two gear members are in no way therein to
collide; in other words, the trajectory of the crest of the mating gear must develop at
the outside of the profile of the dedendum flanks. Finally is to be observed that
points T1 and T2 intercepted on the line of action by the straight lines perpendicular
to it conducted for the centers O1 and O2 are the points of interference, which
cannot be exceeded by the addendum circles of the two mating gears (note that
T1 ; A; C; E; T2 are points on line of action; they derive only from geometrical
consideration and are related to the meshing evolution from pinion root to pinion
tip, regardless of whether pinion or wheel drives).
Pressure angle, a, is the acute angle, measured in a chosen plane, between the
common pitch plane and the common normal to the profiles of the two gear teeth at
a point of contact. In transverse, normal and axial planes, we have respectively the
transverse pressure angle, at , the normal pressure angle, an ; and the axial pressure
angle, ax . Without subscript and when there is no possibility of misunderstanding, a
indicates the transverse pressure angle. The value of the pressure angle influences
the shape of the tooth profiles as well as the components of the load exchanged
between the teeth. When a increases, the tooth profiles tend to flatten out, up to
incurring the risk of the pointed tooth, but the tooth thickness at the bottom
increases. On the contrary, when a decreases, the tooth profile tends to curve, the
thickness in addendum grows and that in dedendum decreases. Moreover, when a
increases, the radial component of the force exchanged between the teeth increases
well; this component does not participate in the transmission of power, but over-
loads bearings.
• Transverse contact ratio, ea : this is the ratio between the length of path of
rotation q and circular pitch p or, what is the same, the ratio between the length
of path of contact ga and base pitch pb (see the following chapter). If q [ p or
ga [ pb , before a pair of meshing teeth comes out from the contact, the next pair
of teeth is already engaged in the contact. Just so the rolling motion of the pitch
circles is a uniform motion. Then the following inequality must be necessarily
verified, otherwise irregularities of motion occur:
30 1 Gears: General Concepts, Definitions and Some Basic Quantities
q ga
ea ¼ ¼ [ 1: ð1:30Þ
p pb
It is however necessary to consider that, with ea [ 1, when the tip of the tooth
flank of the driving pinion comes out of contact at the point E, it moves towards the
crest of the tooth flank of mating wheel. This actual behavior under load is due to
the elastic springback which occurs when the load instantly vanishes. If the teeth
were not loaded, the contact at the point A between the tip of the wheel tooth and
dedendum flank of the pinion would be a regular contact, i.e. without any impact,
provided that the teeth do not show errors. However, since the teeth are loaded, and
thus are subjected the bending, we have an inlet impact at the point A, because the
edge tip of the tooth flank of the wheel tends to penetrate into the dedendum flank
of the pinion, shaving the latter towards its root. Then a relative sliding is added to
impact: consequently we have irregular motion, heavy wear of the addendum flank
of the wheel and dedendum flank of the pinion, and considerable noise. At least
within certain limits, we can remedy this behavior with an adequate tip relief of the
teeth of the driven wheel or with a suitable root relief of the teeth of the driving
member.
• Root fillet radius, qF : it is the radius of the root fillet, i.e. the radius of the
portion of the clearance curve joining the flank profile to the dedendum circle.
As the profile portion between the tip relief circle and addendum circle, also the
root fillet (for involute profiles, the fillet develops from the fillet circle until the
dedendum circle) can have any geometry, as long as this is compatible with the
need to avoid the interference, only if the cut of toothing is performed by means
of a milling cutter designed in order to produce the desired geometry of the
profile adjacent to the root. When the teeth cutting is instead performed by
envelope cutting, the geometric shape of the root fillet is automatically defined
by the shape of the tip curve of the cutting tool (for example, the basic rack
profiles which can be with and without undercut). In the design of the gears, the
study of the root fillet geometry has great importance. Indeed a root fillet well
designed strengthens the tooth at its root and decreases considerably the notch
effect at the interface tooth/rim.
The load capacity and regular operating conditions of the gears considerably
depend on the accuracy of the production process by which the two members of a
gear pair are obtained. As for any mechanical member or engineering component, it
is practically impossible to produce a gear wheel without manufacture errors, which
represent all the deviations of dimensions and geometry of the real gear wheel with
respect to those theoretically perfect of the same gear wheel. To ensure good
1.6 Precision and Accuracy Grade of the Gears 31
operating conditions for the gears, these deviations must be contained within
appropriate limits, which are known as tolerances or allowable errors.
For economic reasons, these permissible errors must have values as high as
possible (i.e., tolerances must be as wide as possible), as the selection of too narrow
tolerances in cases where precision is not necessary would make the product too
expensive or even unsaleable. On the other end, it is risky to choose too large
tolerances, which could compromise the same functionality of the gear to be
designed. Therefore, to make a reasoned choice of allowable errors, the gear
designer must make an in-depth analysis, taking into account the related conse-
quences to every design choice made by him.
In this regard, the factors to be considered are numerous and differentiated, and
often each of them plays a contrasting role than the other. Among this factors, some
deserve special attention, such as: the type of gear unit to be designed, and its
working conditions; the expected lifetime in hours of work; the maintenance
conditions inclusive of the possibility of interchangeability of gear wheels; the
available manufacturing process, inclusive of cutting machines and shaving or
grinding machines; the control and testing procedures in relation to the available
measuring instruments; and above all the techno-economic aspects regarding the
economic feasibility and viability of the gear design. It would, therefore, most
appropriate that the gear designer made use and treasure of previous knowledge and
experiences, when these were available.
The unavoidable manufacturing errors affecting both the body of gear wheels
and the teeth are due mainly to: deviation in pitch, tooth profile, and flank line;
center distance deviations; deviations from parallelism or misalignment of the axes
of the two members of a gear pair with respect to their direction or position
theoretically correct; radial and axial runouts of blank or body of a gear wheel,
which represent the rotation accuracy in the radial and respectively axial directions,
and greatly influence the precision attainable by means of the technological process
by which the toothing is obtained; etc. This errors may result in: increased operating
noise, torsional vibration, and loss of rotational accuracy of the driven gear wheel
with respect to the driving gear wheel; uneven distribution of the pressure of contact
on the active flanks, with consequent non-uniform wear of the same flank; irregular
operating conditions; uneven distribution of the load along the tooth flank line;
increased dynamic loads; alterations in the load distribution in the single and double
contact areas of path of contact in comparison with the theoretical distribution, and
other detrimental effects.
However, it is to be noted that not all the deviations and errors that may take
place are equally important to ensure the correct operating conditions of a gear unit.
Only the dimensional sizes of a gear wheel which are essential for the aforemen-
tioned correct operation of the gear transmission system must be subject to toler-
ance, since tolerance have a considerable cost, both because it is necessary to
provide an adequate process of machining for they can be obtained, and because all
the dimensional sizes subject to tolerance must be checked and tested by means of
suitable measuring instruments. To give a rough idea of the cost of tolerance, just
remember that, for accuracy grades between 5 and 8, the production cost of a gear
32 1 Gears: General Concepts, Definitions and Some Basic Quantities
wheel increases by about (60 80)% from an accuracy grade to that immediately
finer [40].
It is also be noted that deviations of the single dimensional sizes are all important
to ensure the proper operating conditions of the gear unit, but their importance may
be higher for some quantities, and less for the others. In this regard, for a reasoned
judgment on the members of a gear transmission system, it is necessary to check
how individual deviations affect together the regularity of motion and load trans-
mission. To this end, appropriate quantities that allow to evaluate this global effect
are introduced. The determination of the tolerances to be assigned to the individual
dimensional sizes of gears is much more complex compared to what happens for fits
or other mechanical members coupled together, as kinematic considerations come
into play in the case of the gears, which greatly complicate the problem. The
introduction of the comprehensive quantities above seeks to facilitate, at least a
little, the solution of this problem. In the absence of specific experience gained in
the design of gear units similar to the one of interest, standards codified in this
regard may be a valid and reliable guideline for the gear designer.
The ISO Standard provide a system of accuracy for cylindrical (ISO 1328-1:
2013 [20], and ISO 1328-2: 1997 [21]) and bevel gears (ISO 17485: 2006 [22]),
and allow to determine the tolerances to be assigned to the dimensional sizes and
comprehensive quantities of these gears. We refer the reader directly to these ISO
standards that, for obvious reasons of brevity, are summarized below only on their
essential directives. Only for particular aspects not expressly covered in the ISO
Standards we will briefly refer to other national standards.
These ISO Standards provide 12 accuracy grades (or quality grades) which, in
decreasing order of accuracy, are designed by the numbers from 1 to 12. The
accuracy grade 1 and 2 refer to master gear wheels (or reference test gear wheels),
which are produced by means of highly sophisticated machines. The accuracy
grades 3 and 4 refer respectively to high-precision gears and medium-high precision
gears, used for measuring instruments. The accuracy grade 5 is mainly used in fine
mechanics. The accuracy grades from 6 to 12 are those which are interest in
practical applications of the mechanical industry. However, it is to be born in mind
that, for the production of usual gears, the accuracy grades 6, 7 and 8 are those
normally used; they correspond respectively to: high-accuracy gears that work at
high speed and high load; normal-accuracy gears that operate at high speed and
moderate load or at moderate speed and high load; low-accuracy gears that operate
at low speed and low load. The accuracy grades from 9 to 12 relate to gears of
progressively decreasing qualitative characteristics, so that the accuracy grade 12
corresponds to coarse-quality and low-speed gears, namely to the poorest quality
gears in the considered scale.
Usually, for a pair of mating gears, both members are manufactured with the
same accuracy grade. This custom, however, is not a rule to be respected strictly; it
may be waived by agreement between the manufacturer and user.
To complete the subject, we consider appropriate to briefly describe below the
main individual errors and cumulative errors. We recall once again that the accu-
racy grade of a gear depends on the tolerance limits for the circular pitch, tooth
1.6 Precision and Accuracy Grade of the Gears 33
profile and flank line (this last gives a measure of the tooth alignment). We recall
also that the deviations of parameters of interest from their theoretical values can be
accurately measured, by means of suitable measuring instruments, but they do not
allow to evaluate properly and exhaustively what are their effects on the regularity
of the motion of running gears, because the running quality depends mainly on the
sum of these errors as well as on the radial and axial runouts.
The main types of individual errors are as follows (here we give symbol and unit
of each of them, while in regard to the choice of their values we refer the reader to
the aforementioned standards):
– single pitch deviation or pitch error, fp ðlmÞ: it is the maximum absolute value
of all the individual single pitch deviation, fpi , which are observed; fpi is the
algebraic difference between the actual pitch and the corresponding theoretical
pitch in the transverse plane,
detected on the measurement circle of the gear
wheel. Then fp ¼ maxfpi . We consider also the total cumulative pitch deviation
or total index deviation, fp ðlmÞ, defined as the largest algebraic difference
between the individual cumulative deviation values for a specified flank
obtained for all the teeth of a gear, Fpi ; this last is also called individual index
deviation, and is defined as the algebraic difference, over a sector of n adjacent
pitches, between
the actual length
and the theoretical length of the relevant arc.
Then Fp ¼ max Fpi min Fpi . The transverse base pitch deviation or simply
base pitch deviation, fpb ðlmÞ, is the difference between the actual and the ideal
base pitch; it is related to the single pitch deviation by means of relationship
fpb ¼ fp cos a. The base pitch deviation is essential for uniformity of transmis-
sion of motion and distribution of load over all tooth pairs in meshing. It is
crucial that the base pitches of pinion and mating wheel coincide, while equal
base pitch deviations of pinion and mating wheel cancel one another during the
engagement.
– Total profile deviation, fa ðlmÞ: it is defined as the distance between two fac-
similes of the design profile, which enclose the measured profile over the profile
evaluation range. The total profile deviation is the result of the profile form
deviation, ff a ðlmÞ, and the profile slope deviation, fHa ðlmÞ. Profile form
deviation is defined as the distance between two facsimiles of the mean profile
line, which enclose the measured profile over the profile evaluation range, while
the profile slope deviation is defined as the distance between two facsimiles of
the design profile, which intersect the extrapolated mean profile line at the
profile control diameter, dcf ðmmÞ, and the tip diameter, da ðmmÞ. For further
detail on this subject, we refer the reader directly to ISO 1328-1: 2013, in which
five interesting examples of profile deviations are shown, namely: profile
deviations with unmodified involute; profile deviations with pressure angle
modified; profile deviations with profile crowning modification; profile devia-
tions with profile modified with tip relief; and profile deviations with profile
modified with tip and root relief. All five examples show the influences of the
profile form and profile slope deviations on the total profile deviation.
34 1 Gears: General Concepts, Definitions and Some Basic Quantities
– Total helix deviation or total flank line deviation, Fb ðlmÞ: it is defined as the
distance between two facsimiles of the design helix which enclose the measured
helix over the helix evaluation range. The total helix deviation is the result of the
helix form deviation, ff b ðlmÞ, and the helix slope deviation, fHb ðlmÞ. Helix form
deviation is defined as the distance between two facsimiles of the mean helix line,
which enclose the measured helix over the helix evaluation range, while the helix
slope deviation is defined as the distance between two facsimiles of the design
helix, which intersect the extrapolated mean helix line at the end points of the
face-width, bðmmÞ. Again, for the further details on this subject, we refer the
reader directly to ISO 1328-1: 2013, in which five other interesting examples of
helix deviations are represented, namely: helix deviations with unmodified helix;
helix deviations with helix angle modification; helix deviations with helix
crowning modification; helix deviations with helix end relief; helix deviations
with modified helix angle with end relief. All five examples show the influences
of helix form and helix slope deviations on the total helix deviation.
– Runout, Fr ðlmÞ: it should first be noted that the term runout, without any
specification, indicates the radial runout. The runout (or radial runout) is the
difference between the maximum and the minimum individual radial measure-
ment, ri ðlmÞ; this is the radial distance from the gear wheel axis to the center or
other defined location of a probe (cylinder, ball or anvil), which is placed
successively in each tooth space. During each measurement, the probe contacts
both the right and left flanks at approximately mid-tooth depth. It can also be
calculated from pitch measurements. The runout is a measure of the eccentricity
of the gear wheel teeth, i.e. the measure of the off-center error due to the fact that
the gear wheel is not exactly round.
– Axial runout: this deviation is usually measured in the testing stage of a gear
wheel, and gives a measure of the wobble of the same gear wheel. The mea-
surement is performed by placing a dial gauge, the axis of which is held at a
specific distance and parallel to the axis of rotation of the gear wheel under
consideration.
For a judgment on the operation of the toothed members of a gear unit it is
necessary to check how the various individual deviations jointly affect the regularity
of the motion and load transmissions. The overall effect due to two or more indi-
vidual errors, acting simultaneously, can be verified by means of suitable composite
deviation texts. These type of tests approximate the operation of the gear examined
in conditions of service. The composite deviation tests are of two types: single flank
composite testing and double flank composite testing.
In both the two types of tests, the gear wheel to be tested is rotated through at
least one full revolution in close contact with a master gear wheel (or reference
gear wheel) of know accuracy grade. However, with the single flank composite
deviation test, the gear wheel pair simulates the actual meshing conditions,
including the backlash, but with a reduced load; on the contrary, with the double
flank composite deviation test, the two gear wheels are made to rotate without
backlash, radially pressing them one against the other. It is to be noted that the first
1.6 Precision and Accuracy Grade of the Gears 35
type of test, which simulates the actual meshing conditions, allows to evaluate the
errors in angular transmission, whereas the second type of test allows to evaluate
the deviations in center distance.
With the single flank composite deviation testing, the master gear wheel is
rotated at an angular velocity strictly kept constant, and the center distance and
alignment are adjusted so as to ensure the design backlash; only the rolling on any
one operating flank of the gear wheel to be tested is examined, and any deviation in
the angular motion is measured. The conclusions on the operating behavior of a
gear wheel obtained with this testing method are more realistic and unique com-
pared to those allowed by the double flank composite deviation testing, because
with this second testing method also the deviations of the non-operating flank come
into play. However, the equipment and measuring instruments to carry out the
checks with the first testing method (i.e. the single flank composite testing) are more
complex, sensitive and expensive compared to those of the second testing method.
The single flank composite deviation testing is that used for the evaluation of the
transmission error of a gear. Transmission error is defined as the deviation of the
angular position of the driven gear wheel, for a given angular position of the driving
gear wheel, from the position that the driven gear wheel should theoretically have if
the gear was geometrically perfect. In even more general terms, the transmission
error can be defined as the difference between the actual position of the output gear
wheel and the position it would occupy if the gear drive was perfectly conjugated.
For gears with theoretically perfect involutes and an infinite stiffness, the rotation of
the output gear wheel would be a function of the input rotation and the transmission
ratio. Therefore, in these ideal conditions, a constant rotation of the output shaft
would also correspond to a constant rotation of the input shaft. Instead, in the actual
operating conditions, a motion error of the output gear wheel with respect to the
input gear wheel will occur, due to both intentional shape modifications and
unintentional modifications, as well as assembly errors and defects in stiffness
[17, 39, 33].
The transmission error can be measured statically or dynamically, under load or
without load. For a gear pair, this error is measured with an apparatus that analyzes
the two component gear wheels to be tested, in their mutual engagement, but the
same apparatus is used, as mentioned above, to test the individual gear wheels of
the gear pair, each of which run against a master gear wheel, to measure the
individual contributions to the transmission error.
Test for evaluation of the transmission error are normally carried out at very light
torques, so as to eliminate or minimize the contribution of deflections of the same
test apparatus, which may affect the measurement results. Tests of the same type,
but under heavy loads, such as those in actual practical applications, are normally
carry out in the actual gearbox or in a very rigid special text box, using other types
of equipment.
All deviations of the gear pair detected with the above-mentioned testing method
are useful to control gear functional characteristics; also nicks and burrs can be
detected. Given the above-mentioned test procedures, the results obtained allow to
calculate the total no-load transmission error, in terms of total single flank
36 1 Gears: General Concepts, Definitions and Some Basic Quantities
composite deviation, Fis ðlmÞ, and the tooth-to-tooth error, in terms of tooth-to-
tooth single flank composite deviation, fis ðlmÞ. In order to have correct operating
conditions for the gear in question, these two errors must fall within specific limits,
which are respectively defined by the total single flank composite tolerance,
FisT ðlmÞ, and the tooth-to-tooth single flank composite tolerance, fisT ðlmÞ. About
the calculation method and the values of these two tolerances, to be taken in the
design stage, we refer the reader to the afore-mentioned ISO Standard.
It should be remembered that the transmission error, together with the mesh
stiffness variation, are considered to be the primary source of excitation of vibra-
tions and noise, which are radiated outwards through the gearbox. The character-
istics of low or high vibration and noise level of a gearbox depend on the
instantaneous meshing conditions that arise between the various meshing tooth
pairs. Under load, but at very low speeds, we have the static transmission error,
which mainly depends on tooth deflections and manufacturing and assembly errors.
Under operating conditions where both loads and speeds are high, the mesh stiff-
ness variation (it is due to changes in the length of contact line and tooth deflec-
tions) and the excitation, which is localized at the instantaneous point of contact,
generate dynamic mesh forces, which are transmitted to the housing through shafts
and bearings; in this case, we are talking about dynamic transmission error. Noise
radiated by the gearbox is closely related to the vibratory level of the housing.
It is to be noted that the tooth-to-tooth single flank composite deviation is a
parameter of great importance in reference to the control of vibration and noise, and
smoothness of operation. The primary source of this deviation is due to errors in
tooth profiles with respect to their theoretical geometry, which alter the correct
conjugacy of mating teeth. Even the modified tooth shape (tip and root relief, profile
crowning, helix modification, etc.) can greatly influence the tooth-to-tooth single
flank composite deviation, as the gears are tested at low load, while the teeth are
designed so as to have the right conjugacy only to the specific design load, which is
very high. It is also to be noted that the primary source of the no-load total
transmission error is due to the accumulated pitch error, i.e. the already defined total
cumulative pitch deviation, FP .
The double flank composite deviation testing is used to evaluate the total radial
00
composite deviation, Fi ðlmÞ; and the tooth-to-tooth radial composite deviation,
00 00
fi ðlmÞ. The total radial composite deviation, Fi , is defined as the difference
between the maximum and minimum values of center distance, which occur in the
test when the gear wheel to be tested, with its right and left flanks simultaneously in
contact with those of the master gear wheel, is rotated through one complete rev-
00
olution. Instead, the tooth-to-tooth radial composite deviation, fi , is defined as the
value of the radial composite deviation corresponding to one angular pitch,
s ¼ ð360 =zÞ, during the complete cycle of engagement of all the teeth of the gear
wheel to be tested.
These two radial composite deviations are affected by the accuracy of the master
gear wheel and the total contact ratio of the gear pair composed by the master gear
wheel and the gear wheel to be tested. Both these deviations show the combined
1.6 Precision and Accuracy Grade of the Gears 37
effect of different errors, such as profile error, pitch error and variation in tooth
thickness. The total radial composite deviation includes, in addition to these
deviations, even runout and wobble. In the usual types of test equipment, errors
may be recorded in the form of a circular trace or a linear trace; these traces can be
processed using statistical methods, such as shown in Murari et al. [28].
Finally, it should be noted that the topic covered in this textbook assumes the
reader has the basic knowledge regarding the dimensional tolerances (including ISO
system of limits, fits and tolerances), and geometrical tolerances and deviations.
These are very important issues which, for reasons of space, cannot be recalled
here. Only when the clarity requires a memory recall, a brief mention will be made,
in their essential concepts. For basic concepts, we refer the reader to ISO specifi-
cations, as regards the general aspects, or to specialized textbooks, as regards
particular aspects, such as the computer aided toleracing (see: Björke [1]; Chévalier
[5]; ElMaraghy [10]).
References
18. ISO 1122-1: 1998 Vocabulary of gear terms—Part 1: definition related to geometry
19. ISO 1122-2: 1999(E/F) Vocabulary of gear terms—Part 1: definition related to worm gear
geometry
20. ISO 1328-1: 2013 Cylindrical gears-ISO system of flank tolerance classification-Part 1:
definitions and allowable values of deviations relevant to flank of gear teeth
21. ISO 1328-2: 1997 Cylindrical gear-ISO system of flank tolerance classification-Part 2:
definitions and allowable values of deviations relevant to radial composite and runout
information
22. ISO 17485: 2006 Bevel gears-ISO system of accuracy
23. Jelaska DT (2012) Gears and gear drives. Wiley, Chichester
24. Litvin FL, Fuentes A (2004) Gear geometry and applied theory, 2nd edn. Cambridge
University Press, Cambridge
25. Maitra GM (1994) Handbook of gear design, 2nd edn. Tata McGraw-Hill Publishing
Company Ltd, New Delhi
26. Merritt HE (1954) Gears, 3th edn. Sir Isaac Pitman & Sons, Ltd., London
27. Micheletti GF (1977) Tecnologia Meccanica, 2nd edn. Unione Tipografica – Editrice Torinese
(UTET), Torino
28. Murari G, Stroppiana B, Vullo V (1981) Approccio statistico per la descrizione di
caratteristiche strutturali di superfici lavorate. ATA 5:351–355
29. Niemann G, Winter H (1983) Maschinen-Elemente, Band II: Getriebe allgemein,
Zahnradgetriebe-Grundlagen, Stirnradgetriebe. Springer, Berlin, Heidelberg
30. Niemann G, Winter H (1983) Maschinen-Elemente, Band III: Schraubrad-, Kegelrad-,
Schnecken-, Ketten-, Rienem-, Reibradgetriebe, Kupplungen, Bremsen, Freiläufe. Springer,
Berlin, Heidelberg
31. Niemann G, Winter H, Höln B-R (2005) Maschinenelemente: Konstruktion und Berechung
von Verbindungen, Lagern, Wellen, Band 1. - 4. Auflage. Springer, Berlin, Heidelberg
32. Panetti M (1937) Lezioni di Meccanica Applicata alle Macchine, Parte IIa, Ruote – Roteggi –
Macchine Funicolari – Cingoli. Arti Grafiche Pozzo, Torino
33. Podzharov E, Syromyatnikov V, Ponce Navarro JP, Ponce Navarro R (2008) Static and
dynamic transmission error in spur gears. Open Ind Manuf Eng J 1:37–41
34. Pollone G (1970) Il Veicolo. Libreria Editrice Universitaria Levrotto&Bella, Torino
35. Radzevich SP (2017) Gear cutting tools: science and engineering, 2nd edn. CRC Press,
Taylor&Frencis Group, Boca Raton
36. Radzevich SP (2018) Theory of gearing: kinematics, geometry and synthesis, 2nd edn. CRC
Press, Taylor&Francis Group, Boca Raton
37. Rossi M (1965) Macchine Utensili Moderne. Editore Ulrico Hoepli, Milano
38. Scotto Lavina G (1990) Riassunto delle Lezioni di Meccanica Applicata alle Macchine:
Cinematica Applicata, Dinamica Applicata - Resistenze Passive - Coppie Inferiori, Coppie
Superiori Ingranaggi – Flessibili – Freni. Edizioni Scientifiche SIDEREA, Roma
39. Townsend DP (1991) Dudley’s gear handbook. McGraw-Hill, New York
40. Vullo V (1983) Calcolo delle Tolleranze con Metodi Probabilistici e Criteri di Minimo Costo.
Cooperativa Libraria Universitaria Torinese, Torino
41. Woodbury RW (1958) History of the gear-cutting machines: a historical study in geometry
and machines. M.I.T. Technology Press, Cambridge, Massachusetts
Chapter 2
The Geometry of Involute Spur Gears
Abstract In this chapter, the fundamentals of involute spur gears geometry are
given. Once the way of generating an involute of base circle has been described, the
polar coordinates of any of its current points are given. The parametric equations
that describe the position vector as a function of both the involute roll angle and the
involute polar angle are also obtained, also defining the tangent and normal vectors
as well as their unit vectors. The discussion is then generalized to include the
ordinary, extended and shortened involute curves. The fundamental properties of
the involute curves are subsequently described, with special reference to their use as
gearing teeth profiles. Particular attention is given to the geometric sizing of the
gears, in terms of modular sizing, without however neglecting the one based on the
diametral pitch. The main quantities defining the teeth geometry are given, par-
ticularly those that ensure the appropriate kinematic operation of the gears and those
that play a fundamental role for the addendum modification in terms of profile shift.
Finally, a reference is made to standard and no-standard basic rack tooth profiles.
are to be found in the many and significant advantages associated with it, among
which the following advantages deserve specific mention: the conjugate action is
independent of the variation of center distance; it allows to obtain a high accuracy
grade of the gears, since the standard basic rack teeth have straight-sided profiles (as
well as those of the cutting tools derived from them), so they can be made as
accurately as possible; a single cutting tool can generate gear wheels of a given
module, with any number of teeth.
By definition, an involute is the path described by any point on a taut inex-
tensible thin strip as it unwinds from a given curve, called evolute. Therefore, given
a planar curve (evolute) and a tangent straight-line to it at a given point, the curve
generated by this tracing point of the tangent straight-line (or any other tracing
point on this tangent straight-line), when it rolls without sliding on the given curve,
is the involute of it (see Gray [13], McCleary [21], Smirnov [30]) . Any convex
curve can be used for this purpose, but in gear design the term involute implies the
involute of a circle. The evolute, that is the circle from which the involute of the
circle is generated, is called the base circle.
With few exceptions, the involute of a circle is, nowadays, the curve charac-
terizing the tooth profiles of cylindrical gears with parallel and crossed axes and the
involute worm. With its generalization in three-dimensional space, which leads to
the spherical involute, it also forms the basis of the bevel and spiral bevel gears. The
knowledge of the geometry of the involute of a circle, which is a planar curve, and
of the spatial or spherical involute, which is a three-dimensional curve, is of fun-
damental importance for the understanding of the gears. In this Chapter, we sum-
marize the basic concepts of the involute of a circle, while in Chap. 9 we will
summarize the basic concepts of the spherical involute. For any further details, we
refer the reader to Tuplin [31], Merritt [22], Colbourne [6], Phillips [26], Litvin and
Fuentes [20], and Dooner [8].
Figure 2.1 shows the involute of a circle with fixed center at point O, base radius
rb and base diameter db ¼ 2rb ; generated as the trace of the tracing point P0 of the
tangent straight-line t-t, when the latter rolls round the base circle in a clockwise
direction without sliding. A symmetrical involute with respect to the radial
straight-line OP0 (not shown) could be generated as the trace of the same tracing
point P0 , when the tangent straight-line t-t rotates in a counterclockwise direction.
The involute clearly cannot extend inside its base circle, and can be developed as
far as desired outside the same base circle. All geometric relationships of the
involute derive from the fact that, due to the rolling without sliding of the tangent
straight-line t-t round the base circle, whatever the tracing point P, the segment PT
of tangent line t-t is equal in length to the corresponding arc Pd0 T of the base circle.
It is to be noted that, at every its point P, the involute is normal to the taut
inextensible thin strip, which materializes the generating straight-line. In the
position of this generating line represented by Pi Ti in Fig. 2.1, the point Ti on the
base circle is the instantaneous center of rotation of the generating straight-line.
Hence an infinitesimal length of involute arc at point Pi is indistinguishable from a
circular arc having its center at point Ti and radius Pi Ti ; then Pi Ti is the radius of
2.1 Generation of the Involute and Its Geometry 41
We have seen that the point of initial tangency P0 (Fig. 2.1) of any rolling
straight-line can generate two branches of an involute curve symmetrical with
respect to the radial straight-line OP0 , when the generating straight-line rolls
without sliding over the base circle clockwise and counterclockwise, respectively.
From a geometrical point of view, each of these two branches represents the profiles
of the right and left flanks of teeth of an involute spur gear. If the involute is used as
profile of the teeth flanks, two families of involute equally spaced along the base
circle have to be constructed. In other words, the symmetry of the tooth requires the
tracing of as many profiles on the same base circle obtained by unwinding the taut
inextensible thin strip in the opposite rotational directions.
From the point of view of the conditions of contact between the mating profiles,
which will be discussed soon, it is interesting to consider the correlation between
the arcs (and the corresponding angles) of the base circle described by the gener-
ating straight-line and the corresponding arcs of the involute described by the
tracing point during the involute generation. Figure 2.3 highlights the fact that, at
equal arcs of the base circle described by the generating straight-line during its
rolling without sliding round the base circle, progressively increasing arcs of the
involute correspond.
To define analytically the geometry of the involute, we return to consider the
Fig. 2.1, where P is any current point on the involute with origin at point P0 . At
point P, the angle of rotation during the rolling motion of the generating
straight-line t-t from the initial position (point of tangency with the base circle at P0 )
is #. This is the involute roll angle through which the taut inextensible thin strip has
been unwound (according to [16], it is also the angle whose arc on a circle of unit
radius is equal to the tangent of the pressure angle at a given point in an involute to
that circle). With # measured in radians, the involute arc, P d0 T , and the segment,
2.1 Generation of the Involute and Its Geometry 43
d
PT, will be both equal to #rb , that is P0 T ¼ PT ¼ #rb . If we denote respectively
with u and a the involute polar angle and the difference between the involute roll
angle and the involute polar angle, we can write the following relationship:
d
P0T PT
u¼#a¼ a¼ a ¼ tan a a; ð2:1Þ
rb rb
where r ¼ OP is the radius vector to any current point P of the involute curve, and
u is the involute polar angle or involute vectorial angle, i.e. the angle between the
radius vector to the current point under consideration and the radial straight-line
through the origin of the involute (see ISO 1122-1 [16]).
With reference to Fig. 2.1, we draw the circle having radius r passing through
the current point P on the involute curve. The angle that the generating straight-line,
normal to the involute at point P, forms with the tangent to this circle at the same
point P is equal to a, and is called the pressure angle of the involute on the circle of
radius r. Therefore, given an involute of base circle, for each value of r we have the
corresponding pressure angle, a.
44 2 The Geometry of Involute Spur Gears
which can be obtained by numerical methods (see Démidovitch and Maron [7],
Moré et al. [24]) or with an approximate relationship, but sufficiently precise,
proposed by Cheng [5].
or
(a) (b)
Fig. 2.4 Parameters for parametric representation of an ordinary involute curve: a, parameter a; b,
parameter #
2.2 Parametric Representation of the Involute Curve 45
to be used as an alternative. In this equations, i and j are the unit vectors of the x-
axis and respectively y-axis, while xðaÞ and yðaÞ are continuous functions of the
variable parameter, a, which can vary within the open interval ða\a\bÞ; in a
similar way, xð#Þ and yð#Þ are continuous functions of the variable parameter, #,
which can vary within the open interval ðc\#\d Þ.
We leave aside the cases of the extended involute curve and shortened involute
curve, which are self-intersecting curves, and focus our attention on an ordinary
involute curve or conventional involute curve, which does not have points of self-
intersection. Therefore, this last curve is a simple curve, inasmuch as it is charac-
terized by one-to-one correspondence between the current point on the involute
curve and the selected parameter, what does not happen for extended and shortened
involute curves. In addition, the ordinary involute curve is a regular curve, because
the following conditions are satisfied:
• Functions xðaÞ and yðaÞ as well as functions xð#Þ and yð#Þ have continuous
derivative to the first order at least;
• In terms of variable parameters a, it must be:
drðaÞ
ra ¼ ¼ xa i þ ya j 6¼ 0; ð2:8Þ
da
drð#Þ
rh ¼ ¼ x# i þ y# j 6¼ 0; ð2:10Þ
d#
d 2 r ð aÞ
ra ¼ 0 and raa ¼ 6¼ 0 ð2:12Þ
da2
46 2 The Geometry of Involute Spur Gears
or
d 2 rð# Þ
r# ¼ 0 and r## ¼ 6¼ 0: ð2:13Þ
d#2
where inva ¼ ðtana aÞ in accordance with Eq. (2.2), with a that varies within the
open interval ðp=2\a\p=2Þ;
drð#Þ
T ¼ r# ¼ ¼ x# i þ y# j ¼ rb #ðsin #i þ cos #jÞ ¼ Tx i þ Ty j; ð2:16Þ
d#
2.2 Parametric Representation of the Involute Curve 47
T T
t¼ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ sin #i þ cos #j; ð2:17Þ
jT j T þ T2
2
x y
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where jT j ¼ Tx2 þ Ty2 is the absolute value of T, with the condition that it is valid
provided # 6¼ 0, i.e. at any current point on the involute curve, with the exception of
its point of regression;
2 38 9
i j k Ty < i =
N ¼ T k ¼ Tx Ty 0 ¼ 4 Tx 5 j ¼ r# #ðcos #i sin #jÞ; ð2:18Þ
0 : ;
0 1 0 k
which highlights the fact that the direction of normal N at any current point
P coincides with the direction of tangent PB to the base circle (Fig. 2.4b);
N Ty i Tx j Ty i Tx j
n¼ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ cos #i sin #j; ð2:19Þ
jN j N þN2 2 T2 þ T2
x y x y
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
where jN j ¼ Nx2 þ Ny2 ¼ Tx2 þ Ty2 is the absolute value of N, with the condition
that it is valid provided # 6¼ 0, i.e. at any current point on the involute curve, with
the exception of its singular point. This singular point, i.e. the point of regression of
the involute curve, is determined by imposing the two conditions
r# ¼ drð#Þ=d# ¼ 0, and r## ¼ d 2 rð#Þ=d#2 6¼ 0, where
d 2 rð# Þ
r## ¼ ¼ rb ½ðsin # þ # cos #Þi þ ðcos # # sin #Þj: ð2:20Þ
d#2
(a) (b)
Fig. 2.5 a extended involute curve; b Archimedean spiral as special case of an extended involute
curve
generating line, but a point that is on the straight-line OP0 , at the distance from P0 ,
l ¼ P0 P00 , called offset, and rigidly connected with the same generating line. If P00 is
between O and P0 , the generated curve is an extended involute curve (see
Fig. 2.5a), while if P00 is external to P0 , the generated curve is a shortened involute
curve.
Using an approach similar to the previous one, and taking # as a variable
parameter, we get the following equations:
where the upper sign corresponds to the extended involute curve, and the lower sign
corresponds to the shortened involute curve. Even in these two cases, two branches
of extended and shortened involute curves are obtained, which are symmetric with
respect to the y-axis. However, contrary to what happens for an ordinary involute
curve, where the point P0 which is common to its two branches is a singular point,
in the cases of extended and shortened involute curves the point P00 , which is
common to their two branches, is a regular point.
Taking into account Eqs. (2.21), we can generalize Eq. (2.15), which becomes:
rð#Þ ¼ ½ðrb lÞ sin # rb # cos #i þ ½ðrb lÞ cos # þ rb # sin #j: ð2:22Þ
For l ¼ 0, Eqs. (2.22) and (2.21) are reduced to those of an ordinary involute
curve, obtained previously. Another particular case is interesting, that for which
l ¼ rb , whereby P00 coincides with the origin O of the Cartesian reference system
(see Fig. 2.5b); in this case, the extended involute curve turns out into an
Archimedean spiral, defined by equation r ð#Þ ¼ r b #.
2.2 Parametric Representation of the Involute Curve 49
The components along the coordinate axes of the tangent T and the normal
N ¼ T k for extended and shortened involute curves are given by the
relationship:
Tx ¼ l cos # rb # sin #
ð2:23Þ
Ty ¼ l sin # þ rb # cos #;
Nx ¼ l sin # þ rb # cos #
ð2:24Þ
Ny ¼ l cos # rb # sin #;
Let us now consider the involute profile to examine how it satisfies the requirement
for the transmission of the uniform motion between two parallel axes by means of
an external spur gear. Figure 2.6 shows a spur gear consisting of two gear wheels
with fixed centers O1 and O2 (these are the traces of the two axes on the drawing
plane), and having base circles whose respective radii are rb1 ¼ O1 T1 and
rb2 ¼ O2 T2 , where T1 and T2 are the points of tangency of the straight-line t-t with
Fig. 2.6 Involutes in contact and their mutual action for an external gear pair
50 2 The Geometry of Involute Spur Gears
these circles. This straight-line t-t is one of the two common tangents to the two
base circles, the one compatible with the directions of rotation of the two gear
wheels, obviously with the opposite direction relative to one another.
The contact between any two mating profiles during meshing takes place at a
given point P, located on the above-mentioned common tangent, which is also the
common normal to the two mating profiles at the same point P. When the driving
wheel 1 rotates in the clockwise direction, as indicated by arrow, the point of
contact P moves from T1 to T2 . The two base circles undergo the same peripheral
displacement, moving as if they were drag by the straight-line t-t, which moves
parallel to itself along its direction, by rolling on the base circles without sliding.
Therefore, the two base circles have equal tangential velocities, and thus their
angular velocities x1 and x2 , and their rotational speeds n1 and n2 (x1 and x2 in
rad/s, and n1 and n2 in rpm) will be inversely proportional to their radii.
Point C, that is the point of intersection of the common tangent t-t to the center
distance, a, of the gear wheels, is distant from the center O1 and O2 of lengths
proportional to the radii of the base circles, and therefore we have:
x1 n1 rb2 CO2
¼ ¼ ¼ : ð2:25Þ
x2 n2 rb1 CO1
The two circles having radii r1 ¼ CO1 and r2 ¼ CO2 ; supposed rigidly con-
nected with the two wheels, also have the same tangential velocity, and then roll
against each other without slide. They are the pitch circles of the gear pair under
consideration. Direction of rotation of the two members of the gear pair are
opposite.
We now imagine that pitch circles represent two cylindrical friction wheels
pressed together. In the absence of sliding, rotation of one friction wheel (i.e. pitch
circle) causes rotation of the other in the opposite direction, at an angular velocity
ratio inversely proportional to their radii or diameters. But the friction wheels
transmit notoriously very low torque (and then low power for a given angular
velocity), which is closely related to the coefficient of friction. In order to transmit
more torque than is possible with the friction wheels alone, we now introduce, in
addition to the friction wheels, a driving crossed belt (or a taut inextensible thin
strip) that drives the two base circle used as two cylindrical pulleys. If the smaller
pulley rotates clockwise, the belt (or strip) will cause the larger pulley to rotate
counterclockwise in accordance with the equation:
x1 rb2 db2 r2 d2
¼ ¼ ¼ ¼ ; ð2:26Þ
x2 rb1 db1 r1 d1
where the minus sign indicates that the two pulleys rotates in opposite directions.
In order to transmit even more torque, the third and the last step is to replace the
combined action of the friction wheels and driving crossed belt (or strip) with that
of teeth having involute profiles that exchange their mutual actions along the tan-
gent straight-line t-t to the base circles, and therefore behave kinematically as a
2.3 Involute Properties and Fundamentals 51
driving belt. We can imagine that these profiles are generated from the tracing point
P, when the belt is cut at this point, and the two branches of the belt are winded and
un-winded on the two base circles. In this way the tracing point, which coincides
with the two ends of the belt which has been cut, describes the two involutes drawn
with solid line in Fig. 2.6. These involutes are thus generated simultaneously by the
tracing point P, which is the point of contact, while the portion of the belt T1 T2 is
the generation straight-line.
This generation straight-line is always tangent to the base circles, and does not
change position, but moves parallel to itself from T1 to T2 causing the rolling
without sliding of the two base circles about their axes. The point of contact P also
moves along the generating straight-line, and the latter is always normal to the
involutes at the point of contact. Then the requirements for uniform motion and so
the law of gearing are satisfied.
If the traces O1 and O2 , of the axes of the two wheels in the drawing plane are
arranged on the same side with respect to the area where the teeth touch, the larger
wheel turns in an internal gear or annulus gear or ring gear, having fixed center
O2 , radius of the base circle rb2 , and teeth protruding inward. Both wheels rotate in
the same direction (Fig. 2.7). When the driving wheel having fixed center O1
Fig. 2.7 Involutes in contact and their mutual action for an internal gear pair
52 2 The Geometry of Involute Spur Gears
(the pinion) rotates in the clockwise direction, as shown in Fig. 2.7, the contact of
the profiles is still at a point P of the common tangent to the two base circles.
During the rotation, the point of contact moves along this common tangent from
T1 to the right. Pitch circles of the two wheels are still those with fixed centers O1
and O2 , passing through point C, that is the intersection of this tangent with the line
of centers O1 O2 . For this internal gear pair still apply Eqs. (2.25) and (2.26), with
the caveat that the radius of the ring gear is to be assumed negative. In this way, the
ratio ðx1 =x2 Þ in Eq. (2.26) becomes positive, namely the two gear wheels rotate in
the same direction. If r1 and r2 are the radii of the pitch circles, a is the pressure
angle of the toothing with respect to these pitch circles, a is the center distance and
a0 is the sum or the difference of the radii of the base circles, for the case of an
external gear pair we have the relations:
a ¼ r1 þ r2 ð2:27Þ
Both from Figs. 2.6 and 2.7 and equations above, we can infer that, by varying
the center distance a, the mating profiles are always touching at a point of the
common tangent to the two base circles, but the radii of the pitch circles vary in
proportion to the center distance, and at the same time also the pressure angle a of
the profiles varies. We also infer that the radii of the base circles are invariable
elements of the two wheels (in order to transmit uniform motion between two
rotating shafts with angular velocities x1 and x2 , the base circles must satisfy the
2.3 Involute Properties and Fundamentals 53
relationships (2.25 and 2.26), while the radii of the pitch circles and the pressure
angle are elements that vary with the center distance of the gear pair.
With reference to Fig. 2.6, suppose that the member of the gear pair having
center O1 is the driving wheel and that it rotates clockwise. During the motion, the
mating profiles in mutual contact are sliding over one another, but, if the friction is
negligible, the line of application of the force F that they transmit coincides with the
common normal, i.e. the tangent straight-line to the base circles. This interior
common tangent to the base circles is the line of action (more exactly the transverse
line of action) of the teeth, i.e. the line along which contact may take place. This
line meets the line of centers O1 O2 at the pitch point C. Since then for a given
center distance, a, the position of the pitch point C is independent of the position of
the point of contact P, the profiles would ensure a uniform angular velocity ratio.
The line of action forms with the normal to the line of centers the pressure angle
a. As far as we said earlier, we infer that, given the radius rb1 of the base circle of
the wheel 1 and the direction of rotation of the latter, and chosen the radius r1 of the
pitch circle, which must operate coupled to the wheel 2, and the direction of the line
of centers, the pitch point C and direction of the line of action are determined
(Fig. 2.6). Then also the pressure angle a is determined by means of Eq. (2.29).
Now consider the meshing between any two involutes of two base circles
constrained to rotate about their fixed centers O1 and O2 (Fig. 2.6). The point of
contact, P, between the two involute mating profiles will move along the line of
action. If the two base circles, and then the two involutes related to them, rotate
uniformly with respect to their centers, the length of the segment T1 P will increase
uniformly, while the length of segment T2 P will decrease uniformly, since the
length of the line of action, T1 T2 , remains constant.
On the other hand, the ratio ðx1 =x2 Þ which is closely linked to radii of the two
base circles in accordance with Eq. (2.25), does not change whatever the center
distance, a. Moreover, whatever the center distance, the two involute mating pro-
files will touch however along the line of action; of course, this occurs within the
limits related to the radial development of the two involutes or, what is the same,
within limits related to the tooth depths of the two mating teeth. The variation of the
center distance will determine instead a change of the pressure angle, a, and
diameters of the pitch circles, d1 and d2 .
The line of action is closely related to the radii of the two base circles and center
distance. Considering an actual parallel involute gear pair in the three-dimensional
space, we can define the plane of action as the plane containing the infinity of lines
of action related to any transverse section. Then we define the transverse path of
contact, or simply path of contact, as the locus of successive points of contact
between mating tooth profiles on any transverse section. For involute gear pair with
parallel axes, the transverse path of contact is the portion of the line of action lying
between their tip circles. Therefore, the path of contact depends on tooth depths.
Therefore, the line of contact must not be confused, as sometimes happens, with
the path of contact. The latter is in fact a portion of the line of action along which
contact actually takes place. In the case of unmodified involute profiles free from
undercutting, the path of contact is the portion of the line of action intercepted
54 2 The Geometry of Involute Spur Gears
between the addendum circles of the two members of the gear pair. However, the
profile of the cutting tool (rack-cutter, hob, or shaper) may be symmetrically or
unsymmetrically modified, and the path of contact will no longer be that corre-
sponding to the theoretically conditions. For example, in the case in which the
profile of the cutting tool is modified in order to obtain the tip-easing (or tip relief),
the theoretical path of contact will end at the point at which tip relief begins.
If now, with a = const, we grow progressively the radius r2 of the pitch circle of
the wheel 2 (see Figs. 2.6 or 2.7), the radius rb2 of the base circle, center distance a,
and distance CT2 , i.e. the distance of point C from the point of tangency T2 of the
line of action with the base circle, will increase progressively. But if the radius r2
becomes infinitely large, the pitch circle of the wheel 2 degenerates into the normal
straight-line to O1 C through the point C, while point T2 moves at point at infinity of
the line of action, and base circle is transformed into the line at infinity of the plane.
Therefore, the involute tooth profile of this wheel having infinite pitch radius
becomes the straight-line normal to the line of action.
Therefore, the rack, i.e. the wheel with pitch circle degenerated into a
straight-line that has as conjugate the pitch circle chosen for the wheel 1, has teeth
with straight flanks, inclined of the angle a (pressure angle) with respect to O1 C,
that is with the normal to the pitch line. In this way, starting from an external or
internal gear pair, we have obtained a pinion-rack pair. Therefore, invariable ele-
ments of the involute rack are the pitch line, the active straight flank of the tooth,
and the pressure angle a, that is the angle formed by this straight flank with the
normal to the pitch line. Then, given a rack as the one shown in Fig. 2.8, which
transmit the motion in the direction indicated by the arrow, and the radius r1 of the
pitch circle of a gear wheel intended to meshing with it, the direction of the line of
action, which is normal to the tooth flank at point C, and radius of its base circle
rb1 ¼ r1 cos a are determined. Contrary to what happens for an external gear pair, a
variation of the center distance of the pinion-rack pair does not cause any change of
the pressure angle, since this is an invariable quantity of the involute rack.
First it is necessary to examine what might be the maximum length of the path of
contact. We have already seen above that, in the case of unmodified involute
profiles, the path of contact is the portion of line of action intercepted between the
addendum circles of the two members of the gear pair. However, both for external
and internal gear pairs, the addendum circles of the two gear wheels which make up
the gear pair cannot be arbitrarily large, but must satisfy well defined limits.
For an external gear pair, such as that shown in Fig. 2.6, the addendum circles of
the pinion and wheel can have a radius at most equal to O1 T2 and O2 T1 respec-
tively. In fact, it should be noted that un-conjugate behavior exists if the initial and
final engagement between two teeth occurs outside of the two points T1 and T2
respectively, where the line of action is tangent to the two base circles. These points
2.4 Characteristic Quantities of the Involute Gears 55
are the limit points beyond which, as shown below, the so-called theoretical
interference or involute interference occurs.
Figure 2.9 shows the limit condition above, with the addendum circles that pass
through the points T1 and T2 . The profiles p1 and p2 of the two teeth in contact at
point T1 are the two involutes generated by the same tracing point T1 of the line of
action, when it rolls without sliding on the two base circles. With the tooth depths
shown in figure, whereas the driving wheel 1 (the pinion) rotates clockwise, the
contact can extend from point T1 up to point T2 . When the point of contact P is
located in an intermediate position between T1 and T2 , the centers of curvature of
the two mating profiles p1 and p2 , which coincide with points T1 and T2 , are located
one on one side and the other on the other side with respect to the point
P. Therefore, until the point P is in an intermediate position between the points T1
and T2 , the transmission of motion is kinematically correct.
Suppose now that the addendum circle of the pinion has a radius larger than
O1 T2 . When the point of contact P is located beyond the point T2 , the interference
occurs. In fact, if we consider a point P on the line of action positioned beyond T2 ,
the profiles generated from it are those drawn with a dashed line and indicated with
p02 and p1 (Fig. 2.9). We see immediately that the driving profile p1 instead of push,
should pull the driven profile p02 . In the passage of the contact through the point T2 ,
the profile of the driven tooth should change and assume a curvature symmetrical
56 2 The Geometry of Involute Spur Gears
Fig. 2.9 Path of contact of an external gear pair in the limit condition to avoid interference
with respect to that of the profile p02 . We can also see that at point P, the radius of
curvature of the profile p1 is PT1 , while the radius of curvature of the profile p02 is
PT2 . The two centers of curvature T1 and T2 of the two profiles p1 and p02 at point
P are therefore located on the same side with respect to point P. Thus, the profile p02
has a curvature in the same direction of that of the profile p1 ; therefore, it is nothing
other than the fictitious branch of the profile p2 symmetrical with respect to the
radius O2 R, where R is the origin of the two branches p02 and p2 of the involute
curve.
However, as we can see from Fig. 2.9, the origin R of the involute p02 on the base
circle having radius O2 T2 is inside the involute p1 . Therefore, the profile p1 cuts the
actual branch of the involute of the profile p2 at point Q. This is the theoretical
interference or involute interference or primary interference. Then teeth cannot
properly transmit motion beyond the point T2 .
We can achieve the same conclusions in the case where the addendum circle of
the driven wheel has a radius greater than O2 T1 , and the point of contact P is to the
left of the point T1 . Therefore, to avoid incorrect contacts due to the theoretical
interference, it is necessary that the radii of the addendum circles of the pinion and
wheel do not exceed respectively O1 T2 and O2 T1 . Thus, the maximum possible
length of the path of contact is equal to T1 T2 .
2.4 Characteristic Quantities of the Involute Gears 57
When instead we analyze the contact problem between the teeth of an internal
gear pair, such as that shown in Fig. 2.7, we deduce that correct contacts cannot
occur at points located on the line of action to the left of point T1 , while at first sight
limitations would not exist to the extension of the contacts on the line of action to
the right of this point. When, however, we will examine more thoroughly the
problem of interference in an internal gear pair (see Chap. 5), we will see that, for
other reasons, also in this case there are limitations on the length of the path of
contact.
In the limit condition to avoid interference shown in Fig. 2.9, we note that,
during the rotation of the pinion (the driving wheel 1), while the point of contact
between the teeth moves from T1 to T2 , the point T1 of the wheel 1 moves in T10 , and
the point T 0 of the wheel 2 moves in T2 . The arcs Td
2
0 d0
1 T and T2 T , equal to each
1 2
other because both equal to the maximum length of the path of contact T1 T2 , are the
arcs of action or transverse arcs of transmission on the base circles of the two teeth;
the angles T1dO1 T 0 and T2d
1 O2 T 0 corresponding to these arcs are instead different
2
from each other, as the radii rb1 and rb2 of the base circles are generally different.
Therefore, the arc of action is the arc through which one tooth moves along the base
circles (and then along the path of contact), from the beginning until the end of
contact with its mating tooth. It is to be noted that those shown in Fig. 2.9 are the
maximum possible lengths of arcs of action on the base circles.
If each gear member was fitted with a single tooth, the pinion 1 could transmit
the motion to wheel 2 for a maximum angle of rotation equal to the angle T1d O1 T 0 1
(it is the angle of action or transverse angle of transmission), after which, lacking
the contact between the same pair of teeth profiles, there would be no more
transmission of motion. If the teeth, as always happens, have depths smaller than
those shown in Fig. 2.9, being truncated, for example, for the pinion by the
addendum circle passing through the point E, and for the wheel by the addendum
circle passing through point A, their transverse arcs of transmission on the base
circles will have only the length AE, and the transmission of motion can be only
carried out for a corresponding rotation to an arc of such length.
It is noteworthy that all the aforementioned arcs and angles above on the base
circles have their counterparts on the corresponding arcs and angles on the pitch
circles. Therefore, the transverse arcs (angles) of transmission on the pitch circles
are the arcs (angles) of rotation of the pitch circles from the beginning until the end
of meshing for the same pair of profiles.
Now consider an external gear pair such as that shown in Fig. 2.10, character-
ized by the addendum circles of the pinion and wheel respectively having radii
equal to ra1 ¼ da1 =2 ¼ O1 A and ra2 ¼ da2 =2 ¼ O2 E. The length of the arc of action
on the base circles, equal to the length of path of contact g ¼ ga ¼ AE, is uniquely
determined. Points A and E indicate the points of contact on the line of action at the
beginning and at the end of meshing for the same pair of teeth profiles respectively.
We define meshing cycle or cycle of engagement the total angular displacement that
exists between the instant a pair of teeth come into contact until they become
separated.
58 2 The Geometry of Involute Spur Gears
The initial contact will take place when the dedendum flank of the driving tooth
comes into contact with the tip of the driven tooth. This occurs at point A where the
addendum circle of the driven gear crosses the line of action. If we now depict tooth
profiles through point A and, in this position, we draw lines from the intersection of
the tooth profiles with the pitch circles to the centers O1 and O2 , we obtain the
angles of approach for each gear. From this initial point of contact onwards there is
a mixed contact between the meshing profiles, that is a contact composed of a
2.4 Characteristic Quantities of the Involute Gears 59
rolling contact combined with a sliding contact. Furthermore, while developing the
meshing, the point of contact, which moves on the line of action from A to E, will
move along the same profiles, from the root to the tip on the driving profile, and
from the tip to the root on the driven profile. We will have pure rolling without
sliding only in the instant in which the point of contact passes through the pitch
point C. From this instant onwards, the point of contact continues to move over the
tooth profiles in the same direction until contact finally ceases when the same point
of contact, moving on the line of action, reaches point E. This final point of contact
is located where the addendum circle of the driving gear crosses the line of action.
Following a similar procedure as before, that is depicting tooth profiles through
point E and drawing lines from the intersection of the tooth profiles with the pitch
circles to the centers O1 and O2 , we obtain the angles of recess for each gear. In
order not to complicate Fig. 2.10, we have not traced the profiles at the beginning
and at the end of the contact, for which the angles of approach and recess therein
shown are approximate. In any case the sum of these two angles is called the angle
of action. These three angles have their corresponding arcs, i.e. arc of approach,
arc of recess and arc of action on the pitch circles. Note that the arcs of approach
and recess are normally measured on the pitch circle of the driving gear and on the
pitch circle of the driven gear, respectively. To be more precise, the arcs of
approach is the arc through which any tooth moves along the path of contact, from
the beginning of contact with its mating tooth until the contact arrives at the pitch
point, while the recess arc is the arc through which the same tooth moves along the
path of contact, from time in which the contact coincides with the pitch point until
the contact with its mating tooth ceases.
In order to ensure the continuity of motion, the two members of a gear must be
characterized by teeth that, on the base circles, follow one another at a distance
measured on the same base circles smaller than the length of path of contact AE
(Fig. 2.10). In this way, before the end of the contact between one pair of teeth, the
next pair is already in contact, and therefore the continuity of motion is guaranteed.
Obviously the arcs of action on the base circles (not shown in the figure) have
lengths equal to that of the path of contact, while, on the same circles, the distance
measured between corresponding flanks (or homologues flanks) of two successive
teeth is equal to the length of the segment PN, which is the transverse base pitch or
simply base pitch, pbt ¼ pb , of the toothing. Therefore, we infer that two toothed
gear wheels meshing properly between them have teeth of equal base pitch, pb .
Thus, if z1 and z2 are the numbers of teeth of the pinion and wheel, we will have:
2prb1 2prb2
z1 ¼ z2 ¼ : ð2:37Þ
pb pb
The circular pitch (or transverse pitch or simply pitch), p, is the distance
measured on the pitch circle between corresponding profiles of two adjacent teeth,
and then it is the sum of the tooth thickness and space width, both measured on the
pitch circle. Therefore, the circular pitch is the quotient of the length of circum-
ference of the pitch circle divided by the number of teeth, i.e.:
60 2 The Geometry of Involute Spur Gears
2pr1 2pr2
p¼ ¼ : ð2:38Þ
z1 z2
360 2p
s¼ ¼ : ð2:39Þ
z z
The transverse module (or simply module), m, is defined as the ratio between the
circular pitch in millimeters and the number p, or in equivalent terms as the ratio
between the pitch diameter in millimeters and the number of teeth (in other words,
number of millimeters of pitch diameter per teeth), i.e.
p d1 d2 d
m¼ ¼ ¼ ¼ : ð2:40Þ
p z1 z2 z
Therefore, since the diametral pitch is the ratio between the number of teeth and
pitch diameter in inches, or in equivalent terms the ratio between the number p and
the circular pitch in inches, the two indices of gear-tooth size are related by the
relationships:
pP ¼ p ð2:42Þ
25:4
m¼ : ð2:43Þ
P
We have already seen above that, in the design of a gear pair, it is necessary to
ensure the continuity of motion. To do this, the proportions of the involute profiles
of the mating teeth must be chosen so that a new pair of mating teeth comes into
contact before the previous pair is out of contact. In this regard, the most significant
parameter is the transverse contact ratio, ea (or simply the contact ratio, e, for
cylindrical spur gear). This parameter is defined as the quotient of the transverse
angle of transmission (or angle of action, i.e. the angle through which the gear
rotates from the beginning to the ending of contact on the transverse profile) divided
by the angular pitch. If the angle of action be less than the angular pitch, there
would be a time interval during which no tooth pair would be in contact.
We can express the transverse contact ratio, ea , in various way, of course using
homogenous quantities. With reference to Figs. 2.10 and 1.8, we can write:
AE ga A E q
ea ¼ ¼ ¼ ¼ : ð2:44Þ
pbt pbt AE p
In order to ensure the continuity of the motion, it is obviously necessary that the
tooth profiles be sized so that a second pair of mating teeth come into contact before
the first pair is out of contact. In other words, the contact ratio must be greater than
unity. However, to have good working condition of the gear pair, it is desirable that
this contact ratio is at least equal to (1.4 1.7). It is to remember that the contact
ratio represents the average number of teeth in contact as the gears rotate together.
A contact ratio of 1 or 2 means that one pair or two pairs of teeth are engaged at all
times during meshing, while a contact ratio of 1.5 indicates that two pairs of teeth
are engaged for 50% of the time, and only one pair is engaged for the remaining
50% of the time. AGMA (American Gear Manufactures Association) Standards
state that for satisfactory performance, the contact ratio should never be less than
1.2. However, in the gear design, under no circumstance the contact ratio should
drop below 1.1, when it is calculated with all tolerances at their worst values.
From Eqs. (2.29, 2.37 and 2.38) we can deduce the following relationship that
relates the circular pitch, p, base pitch, pb , and the pressure angle, a:
pb ¼ p cos a: ð2:45Þ
From the previous relationship we infer that between the module on the pitch
circle, m, and the base module (i.e. the module on the base circle), mb , we have a
relationship similar to Eq. (2.45), that is:
mb ¼ m cos a: ð2:46Þ
If the center distance and the radii of the base and addendum circles are given,
and AE and pb are respectively the arc of action of the base circle and the base pitch,
the arc of action on the pitch circle, q, and circular pitch, p, will be given by the
following relationships:
62 2 The Geometry of Involute Spur Gears
r1 r2
q ¼ AE ¼ AE ð2:47Þ
rb1 rb2
r1 r2
p ¼ pb ¼ pb : ð2:48Þ
rb1 rb2
The gear ratio, u, is the quotient of the number of teeth of the wheel, z2 , divided
by the number of teeth of the pinion, z1 . For a single gear pair it coincides with the
transmission ratio, i, defined as the quotient of the angular velocity of the driving
gear, x1 , divided by the angular velocity of the driven wheel, x2 ; therefore, for a
single gear pair, we have:
z2 x1 n1 rb2 r2
u¼ ¼i¼ ¼ ¼ ¼ : ð2:50Þ
z1 x2 n2 rb1 r1
It should be noted that the aforesaid coincidence between the gear ratio, u, and
the transmission ratio, i, no longer exists for a train of gears, since in this case the
gear ratio applies to the individual gear pairs that compose it, while the transmission
ratio regards the whole gear train. In fact, the transmission ratio of the latter is
defined as the quotient of the angular velocity xi of the first driving wheel, which is
the power input wheel, divided by the angular velocity x0 of the last driven wheel,
which is the power output wheel; it is given by Eq. (1.1).
From Eqs. (2.37) we infer that, for two gear wheels meshing with each other, the
condition that they have the same base pitch may be replaced by the one for which
the radii (or diameters) of the base circles are proportional to the numbers of teeth.
For these meshing wheels it is preferable that the usable tooth depth to be divided
into two equal portions, inside and outside the pitch circle. For a given tooth depth,
this condition corresponds to the minimum relative sliding between the mating
profiles (this topic will be discussed later). Then we need to introduce the reference
circles in relation to which we define the main quantities concerning the toothing
and the gear blank, which include the thicknesses of the teeth, space widths, and
pressure angle. It is however necessary to keep in mind that the reference circle (the
one shown on the design drawings) is often different from the pitch circle (the one
correlated to the actual operating conditions).
2.4 Characteristic Quantities of the Involute Gears 63
Now we denote with the subscript, 0, the quantities referred to this reference
circle, named nominal quantities. The nominal tooth thickness, s0 , of the tooth and
the nominal space width, e0 , between two adjacent teeth are theoretically equal
between them and equal to half the nominal pitch, p0 , that is:
p0 pm0
s 0 ¼ e0 ¼ ¼ ; ð2:51Þ
2 2
where m0 is the nominal module. The nominal diameter of this reference circle will
be d0 ¼ 2r0 ¼ m0 z (r0 is the nominal radius), while the nominal pressure angle on
the same reference circle is a0 . Once these nominal quantities are known, also the
diameter db of the base circle and the base module mb can be determined by means
of the relationships:
mb ¼ m0 cos a: ð2:53Þ
These quantities related to the base circle allow to calculate the pressure angle
and the module on any circle of the gear wheel. Therefore, the pressure angle and
the module vary with the radius of the circle under consideration. It is also worth
noting that many of the aforementioned nominal quantities come into play in the
generation processes of the teeth (see Woodbury [32], Galassini [11], Rossi [28],
Henriot [15], Micheletti [23], Boothroyd and Knight [2], and Björke [1]).
If instead we consider a rack (it is a wheel having infinite radius), a shift of the
pitch line parallel to itself corresponds to a change of the radius. However, since in
this case the involute is a straight-line (the teeth have trapezoidal shape), the shift of
the pitch line parallel to itself does not cause any change of the pitch, module and
pressure angle, which are therefore invariable quantities of the rack. This shift
instead determines a variation of the thickness of the tooth and space width, for
which it is necessary to give their nominal values with reference to the nominal or
reference pitch line, the position of which is uniquely determined by its distance
from the dedendum line (or root line) or from the addendum line (or tip line).
Since in the rack the quantities p, m and a are invariable irrespective of the pitch
line choice on it, a toothed gear having z1 teeth which must mesh with this rack will
be tangent to the said pitch line (Fig. 2.11). Therefore, this toothed gear wheel will
have a pitch circle and a base circle whose radii are respectively given by:
z1
r1 ¼ m ð2:54Þ
2
z
1
rb ¼ r1 cos a ¼ m cos a: ð2:55Þ
2
Module and pressure angle become then the module and the pressure angle of
the toothed gear wheel on the pitch circle having radius r1 .
64 2 The Geometry of Involute Spur Gears
Fig. 2.1, which correspond to points P and P0 (u and u0 are the involute polar
angles related to point P and P0 , while a ¼ ð# uÞ and a0 ¼ ð#0 u0 Þ, where #
and #0 are the corresponding involute roll angles), we have:
d 0 ¼ u0 u ¼ inva0 inva:
POP ð2:56Þ
Since then:
rb rb
OP ¼ r ¼ OP0 ¼ r 0 ¼ ; ð2:57Þ
cos a cos a0
where rb is the radius of the base circle of the gear, we infer that thicknesses s and s0
on the circles having radii r and r 0 are related by the relationship:
66 2 The Geometry of Involute Spur Gears
s h i
s0 ¼ r 0 d 0 ¼ r 0 s þ 2ðinva inva0 Þ :
2 POP ð2:58Þ
r r
If we take as a point P0 the apex V (Fig. 2.12) in which the profiles of the two
opposite flanks of tooth meet (at this apex the thickness s0 is zero), and denote invc
the corresponding angle u0 which subtends the tooth half-thickness on the base
circle, we obtain:
If we know the tooth thickness sb on the base circle and the radius rb of the latter,
the angle c is defined by the relationship:
ha ¼ m 0
hf ¼ 1:25 m0 ¼ ð5=4Þm0 ð2:61Þ
h ¼ ha þ hf ¼ 2:25 m0 ¼ ð9=4Þm0 :
The fillet radius q at the root of the tooth is included in the range
ð0:2m0 \q\0:4m0 Þ, and it is commonly assumed equal to ðm0 =3Þ. It is noteworthy
that the dedendum can be even greater than the standard value when we want a
larger radius of curvature between the flank of the tooth and its root. For external
gears, the nominal pitch diameter, and diameters of the addendum and dedendum
circles are given respectively by:
2.5 Gear-Tooth Sizing 67
d0 ¼ 2r0 ¼ m0 z
da ¼ d0 þ 2m0 ¼ m0 ðz þ 2Þ
ð2:62Þ
10
df ¼ d0 2:5 m0 ¼ m0 z ;
4
while for internal gears the same quantities are given by:
d0 ¼ 2r0 ¼ m0 z
da ¼ d0 2m0 ¼ m0 ðz 2Þ
ð2:63Þ
10
df ¼ d0 þ 2:5m0 ¼ m0 z þ :
4
The nominal center distance, a0 , of an external gear pair with z1 and z2 teeth is
given by:
z1 þ z2
a0 ¼ m0 ; ð2:64Þ
2
while the same quantity for an internal gear pair with z1 and z2 teeth is given by:
z2 z1
a0 ¼ m0 : ð2:65Þ
2
68 2 The Geometry of Involute Spur Gears
In Countries that use Imperial System, the geometric sizing of the teeth of spur
gears is carried out with reference to the diametral pitch P, which is the quotient of
the number p divided by the pitch expressed in inches, or the quotient of the
number of teeth divided by the reference diameter, expressed in inches [see
Eq. (2.41)]. In the United States the standard system of the nominal diametral
pitches is the one established by the AGMA Standards and ANSI Standards. The
nominal sizing most widely used provides a standard addendum equal to 1=P0 (in
inches) and standard dedendum equal to 1.25 times the addendum. It is to be
observed that, for fine pitch gears of P0
20, the standard dedendum is
½ð1:20=P0 Þ þ 0:002 in. With this nominal standard sizing we have:
ha ¼ 1=P0
hf ¼ 1:25=P0 ¼ ð5=4Þ=P0 ð2:66Þ
h ¼ ha þ hf ¼ 2:25=P0 ¼ ð9=4Þ=P0 :
The fillet radius q at the root of the tooth is commonly about 0:35=P0 . The
standard dedendum can be increased in the same proportion indicated above for the
modular sizing. For external gears the nominal pitch diameter, and diameters of
the addendum and dedendum circle are respectively given by:
d0 ¼ 2r0 ¼ z=P0
da ¼ d0 þ 2=P0 ¼ ðz þ 2Þ=P0
ð2:67Þ
10
df ¼ d0 2:5=P0 ¼ z =P0 ;
4
while for internal gears the same quantities are given by:
d0 ¼ 2r0 ¼ z=P0
da ¼ d0 2=P0 ¼ ðz 2Þ=P0
ð2:68Þ
10
df ¼ d0 þ 2:5=P0 ¼ z þ =P0 :
4
The nominal center distance of an external gear pair with z1 and z2 teeth is given
by:
z1 þ z2
a0 ¼ ; ð2:69Þ
2P0
while the same quantity for an internal gear pair with z1 and z2 teeth is given by:
z2 z1
a0 ¼ : ð2:70Þ
2P0
2.5 Gear-Tooth Sizing 69
From the comparison of Eqs. (2.61–2.65), valid for SI units, with Eqs. (2.66–
2.70), valid for Imperial System units, we see immediately that the two sizing
systems are perfectly equivalent. If this is so, we can ask what the reason for which
one system of sizing is not internationally used. Unfortunately, the large number of
existing cutters and designs, based on the two systems using SI and Imperial
System, makes it unlikely that any one of them will become the sole standard.
In addition to the two above-mentioned nominal sizings, special sizings are
sometimes used. The reason why these special sizings are used can be many, such
as for example:
• the necessity to have pinions with few teeth avoiding the cutting interference
and, especially, the interference in the working condition between the gears that
mesh with each other;
• the reduction of the relative sliding between the teeth;
• the need to reduce the tooth depth in gears very stressed under bending loads;
etc.
Therefore, gears are realized with teeth of different depth from that given by the
Eqs. (2.61 and 2.66), or with equal depths to those given by these equations, but
divided in different proportions between the addendum and dedendum, and with
different division of the pitch between the tooth thicknesses and space widths.
An example of special sizing is that represented by the stub system, in which the
stub teeth have depth smaller than those given by Eqs. (2.61 and 2.66). Both for SI
and Imperial System, a standard stub system was the one with pressure angle of
20°, and addendum and dedendum shortened according to the relationships:
ha ¼ 0:8m0 ¼ 0:8=P0
ð2:71Þ
hf ¼ ð1:0 1:1Þm0 ¼ ð1:0 1:1Þ=P0 :
Stub teeth have a higher strength than teeth having nominal depth, but the
contact ratio is lower. Gears with stub teeth have been produced in the past, and still
continue to be produced, having tooth depth different from the one given by
Eq. (2.71), and pressure angle different from 20°.
Another example of special sizing is represented by gears with high tooth depth,
i.e. with tooth depth deeper than that of a tooth having nominal depth. These gears
have the advantage of a higher contact ratio, but the strength of the teeth under
bending loads is considerably lower. When, however, the high tooth depth is
obtained also choosing an addendum greater than that of the standard basic rack, the
risk can appear of exceed the interference limiting points T1 and T2 (Fig. 1.8). In
fact, for zero-shifted tooth profiles, any tooth addendum that extends beyond these
interference points is not only without usefulness, but also interferes with the root
fillet area of the mating tooth. This results in a typical undercut tooth, that deter-
mines not only a weakening of the tooth to its root, but also the removal of a
remarkable part of useful involute profile adjacent to the base circle.
70 2 The Geometry of Involute Spur Gears
The most commonly used pressure angle, a, with both SI and Imperial System, is
20°. In the United States, a ¼ 25 is also a standard value. However, the following
pressure angles are still commonly used: a ¼ 14 1=2; a ¼ 15 ; a ¼ 22 1=2. In
certain cases, for internal gear pairs, a pressure angle of 30° is used. The pressure
angle of 20° commonly used allows to have gear pairs able to solve most of the
design problems concerning the gear transmissions.
The actual pressure angles that we have in the working conditions may differ
significantly from the nominal pressure angles, as we shall see when we analyze the
topic about the addendum modification in terms of profile shift. It is however
necessary to keep in mind that, in the case of large values of the pressure angle, it
may have drawbacks when the shafts on which gears are mounted are too flexible.
Finally, it must be remembered that even the face width b, i.e. the width of the
toothed part of a gear, measured along a generatrix of the reference cylinder, is
given as a function of the module or diametral pitch. This face width is not stan-
dardized, but it is generally included within the following ranges:
9m0 \b\14 m0
ð2:72Þ
9=P0 \b\14=P0 :
The wider the face width, the more difficult is to manufacture and mount the
gears so that contact is uniformly distributed across the entire face width.
As we already said, the involute tooth profile, analytically defined by Eq. (2.2),
cannot extend inside the base circle. On the other hand, if we choose the gear-tooth
sizing defined by the relationships (2.61), or the equivalent relationships (2.46)
when we use the gear-tooth sizing based on the diametral pitch, we see immediately
that only for very high number of teeth the profile is all outside the base circle.
Indeed, in the extreme case in which the profile from tip to root is all active, the
nominal quantities involved must satisfy the following inequality:
ðr0 rb Þ ¼ r0 ð1 cos a0 Þ
hf ¼ 1:25m0 ð2:73Þ
From this inequality, taking into account that m0 ¼ d0 =z ¼ 2r0 =z; we obtain:
5
z
: ð2:74Þ
2ð1 cos a0 Þ
Tooth profile will then be all outside the base circle only if: for a0 ¼ 15 ; z
74;
for a0 ¼ 18 ; z
52; for a0 ¼ 20 ; z
42; for a0 ¼ 22 ; z
35; for a0 ¼ 24 ; z
29;
for a0 ¼ 26 ; z
25; for a0 ¼ 28 ; z
22; for a0 ¼ 30 ; z
19. In cases where the
correspondences are not fulfilled, for which z is lower than the value given by the
ratio to the right side of the inequality (2.74), it is necessary to extend the profile of
the tooth inside the base circle. The portion of the tooth profile inside the base circle
can have any geometric shape provided that interference will not occur; this shape
will be the one related to the profile of the tool head used in the cutting process of
the gear.
2.5 Gear-Tooth Sizing 71
It is then preferable that the portion of the involute profile in the vicinity of the
base circle is not used. This is because the involute profile immediately close to the
base circle has a radius of curvature which decrease to zero in correspondence of
the same base circle. The involute profile in this region is therefore very difficult to
obtain accurately using current manufacturing processes. Furthermore, in this area,
the relative sliding, wear of the tooth and local compression contact stress tend to
grow strongly.
Fig. 2.14 Standard basic rack tooth profile and mating standard basic rack profile according to
ISO 53
subscript, specifying its meaning). Addendum haP and dedendum hfP of this profile
are equal to m and 1.25m respectively, but the straight portion of the standard basic
rack tooth dedendum (i.e. the portion of dedendum really usable) has depth equal to
hFfP ¼ m. The tooth of the standard basic rack is bounded by the tip line at the top
and by the root line at the bottom. Of course, datum line, tip line and root line are
parallel lines. The working straight portion of the basic rack tooth profile is
interconnected with the root line by the fillet, which has the shape of a circular arc
with a radius equal to qfP .
The working portions, that is the involute portions of the standard basic rack
tooth profiles, are the straight lines passing through the equally-spaced points
located symmetrically with respect to the axis of symmetry of the tooth, which
bisects the thickness of the tooth. These straight lines are inclined at the pressure
angle aP ¼ 20 with respect to the axis of symmetry of the tooth. On the datum line,
the tooth thickness sP is equal to the space width eP , and both are equal to one-half
the pitch, i.e.: sP ¼ eP ¼ p=2 ¼ pm=2.
The mating standard rack tooth profile is the rack tooth profile symmetrical to
the standard basic rack tooth profile with respect to the datum line P-P, and dis-
placed by half a pitch relative to it. Figure 2.14 also shows this mating standard
rack tooth profile drawn with a dashed line. Furthermore, Fig. 2.14 shows the
following main characteristics of the standard basic rack tooth profile having
module, m, and pitch p ¼ pm:
• the flanks are straight for the portion hwP ¼ haP þ hFfP , and are inclined at the
pressure angle aP ¼ 20 to a line normal to the datum line P-P;
2.6 Standard Basic Rack Tooth Profile 73
• the tip and the root lines are parallel to the datum line and respectively at
distances of haP ¼ m and hfP ¼ 1:25 m from it;
• all dimensions of the standard basic rack tooth profile and the mating standard
basic rack tooth profile use line P-P as the base datum, and the active tooth
depth, hwP , of both profiles is equal to 2 haP ;
• with certain exceptions (see below), the standard value of the bottom clearance,
cP , between the standard basisrack tooth and mating standard basic rack tooth is
equal to 0:25m, that is cP ¼ hfP hFfP ¼ 0:25m (this clearance is measured
along a line normal to the datum line);
• the fillet radius, qFP , of the standard basic rack is determined by the standard
clearance, cP .
It is obvious that the higher the bottom clearance (and then the greater the ratio
cP =m), the larger can be the fillet radius of the basic rack, with corresponding
improvement of the bending strength of the tooth. Figure 2.14 shows that the center
of the fillet radius is not on the center of the rack space. The maximum fillet radius
of the basic rack, qfP max , which instead is centered on the rack space, can be
calculated with the following relationships:
cP
qfP max ¼ ð2:75Þ
1 sin aP
ðpmÞ=4 hfP tan aP
qfP max ¼ ; ð2:76Þ
tan½ð90 aP Þ=2
which are valid for a basic rack where aP ¼ 20 ; cP 0:295m and hFfP ¼ 1 m (hFfP
is the straight portion of the standard basic rack tooth dedendum), and respectively
for a basic rack where aP ¼ 20 and 0:295m\cP 0:396 m.
It is to be noted that the actual root fillet, which is outside the active profile, can
vary depending on various influences such as the profile shift, number of teeth, and
manufacturing method. Four types of basic rack tooth profiles are provided,
respectively called A, B, C and D. Table 2.1 summarizes the main characteristics of
these four types of profiles, to be used as an alternative depending on the practical
application requirements. As a guideline, the standard basic rack tooth profile type
A is recommended for gears transmitting high torques, while types B and C are
Table 2.1 Basic rack tooth Symbol Types of basic rack tooth profiles
profiles according to ISO 53:
A B C D
1998 (E)
aP 20° 20° 20° 20°
haP 1m 1m 1m 1m
cP 0.25 m 0.25 m 0.25 m 0.40 m
hfP 1.25 m 1.25 m 1.25 m 1.40 m
qfP 0.38 m 0.30 m 0.25 m 0.39 m
74 2 The Geometry of Involute Spur Gears
recommended for normal service (type C is also applied for manufacturing with
some standard hobs). Profile type D is equivalent to a full radius shape for the fillet;
the high value of dedendum, hfP ¼ 1:40 m, with the associated fillet radius,
qfP ¼ 0:39 m, and the bottom clearance, cP ¼ 0:40 m, permit the finishing tool to
work without interference, while maintaining the maximum fillet radius. It is rec-
ommended for high precision gears with tooth flanks finished by shaving and
grinding, and transmitting high torques. During finishing of these gears care should
be taken to avoid notches of the fillet, which would generate stress concentrations.
The standard basic rack tooth profiles described above are without undercut. For
gears cut by a protuberance tool and finished by shaving or grinding, a basic rack
tooth profile with a given undercut, UFP , and a given angle of undercut, aFP , is
used. Figure 2.15 shows the basic rack tooth profile with a given undercut. The
specific values of the undercut, UFP , and the angle of undercut, aFP , depend on
various influences, the main of which is the method of manufacturing.
Figure 2.16a and b show the normal sections of addendums of two cutting tools,
the first with protuberance, and the second without protuberance. They are used for
cutting teeth with undercut and, respectively, without undercut. The same figures
show the main quantities that define the geometry of these two types of cutting
tools. Among the quantities we have not yet introduced, we have: qpr , protuberance
of the tool; q, material allowance for finish machining; spr ¼ qpr q, residual fillet
undercut. These three quantities are given in mm. The other quantities that char-
acterize the tooth sizing in the normal section will be defined in next chapters.
The standard profile of the rack type cutter, with which the teeth of an inter-
mating series of gears can be cut, is the profile with which a rack, that is a toothed
wheel having infinite diameter, is generated. With the exception of precision gears,
in general the tip cylindrical surface of a toothed gear wheel is not manufactured by
Fig. 2.15 Basic rack tooth profile with given undercut and angle of undercut
2.6 Standard Basic Rack Tooth Profile 75
(a) (b)
Fig. 2.16 Normal sections of a tooth of two cutting tools: a with protuberance, and finishing
allowance for each tooth flank; b without protuberance
the cutter, but it is commonly manufactured by using another machine tool. In fact,
the starting product in gear machining is the gear blank, which is obtained by
means of the so-called blanking operations carried out using machine tools other
than gear cutting machines.
To obtain a suitable circumferential backlash, the standard profile of the
rack-type cutter must cut more deeply than the theoretical condition provides. In
gashing operations instead, the cutter must cut less than the theoretical condition
provides, in order to leave the suitable finish stock. In the latter case, a greater
diameter of the root circles is obtained, but during the finishing it is necessary to
operate in such a way that the root circle is not cut, and that the bottom clearance is
not reduced. For this reason, the roughing cutters are characterized by an adequate
machining allowance, q, so as to obtain the desired root circle with subsequent
machining operations. Even the finishing cutters used to prepare super-finishing
machining, such as shaving and grinding, are characterized by allowance which can
be less than or greater than the allowance of the roughing cutters. For other con-
siderations in this regard, we refer the reader to Niemann and Winter [25].
76 2 The Geometry of Involute Spur Gears
With the basic rack described above (and corresponding rack cutter), and with a
suitable addendum modification, well-balanced toothings can be obtained, able to
satisfy most of the requirements for practical applications. In special case, the
design process involves the choice of quantities which differ from those of the basic
rack. When a large transverse contact ratio is required (this happens, e.g., in printing
machines, in which the distance between the pressure cylinders varies), or when a
gear particularly silent is prescribed as a design requirement, we choose a smaller
angle of pressure (for example, a ¼ 15 or a ¼ 14 1=2 once used in the United
States).
For gears particularly stressed, which must have higher bending strength and
surface durability (pitting), sometimes higher pressure angles are chosen (for
example, a ¼ 22 1=2 or a ¼ 25 ). In this regard, it is noteworthy that in the United
States a ¼ 25 is also a standard value. The increase of the pressure angle with
respect to the standard value ða ¼ 20 Þ causes an increase of the bending strength
and surface durability of the teeth, especially when the number of teeth is high, but
determines, as a counterpart, a minor transverse contact ratio and a lower fillet
radius at the tooth root, and puts limitations on the addendum modification.
Furthermore, with increasing pressure angle, the thickness of the tooth at the top
land decreases (for a ffi 33 we have s ¼ 0, and then a pointed tooth) and, therefore,
the risk of a tooth chipping increases appreciably, particularly when the teeth have
been case-hardened.
Figure 2.17 shows the great influence that the pressure angle has on the shape of
the profile of the tooth; it relates to a toothed gear wheel with number of teeth
z ¼ 20 and a given base circle, but the consequences of the change of the pressure
angle are quite general. The figure shows that, when the pressure angle increases,
assuming values greater than the standard basic rack value ðaP ¼ 20 Þ, the tooth
thickness at the bottom land, the pitch diameter and, as we will show later, the
radial component of the transmitted force and loads on the bearings increase. On the
contrary, when the pressure angle decreases, assuming values lower than the stan-
dard basic rack value, the portion of the tooth dedendum affected from the
undercutting grows. The four profiles A, B, C and D are related respectively to the
pressure angles a ¼ 15 ; a ¼ 20 ; a ¼ 25 and a ¼ 30 .
Figure 2.18, which refers to the same toothed gear wheel of the Fig. 2.17, shows
instead the change, as a function of pressure angle a, of the ratio ðqa =q20 Þ, where
qa and q20 are the radii of curvature on the pitch circles related respectively to the
pressure angle a and to the standard basic rack pressure angle aP ¼ 20 , and to the
same base circle. As we will see in Vol. 2, Chap. 2 regarding the calculation of
surface durability (pitting), the radii of curvature are very important, because they
control the value of the Hertzian contact stress.
The actual
tooth depth may be higher or lower than the tooth depth of standard
basic rack hP ¼ haP þ hfP ¼ 2:25 m : We can get gears particularly silent with
teeth having tooth depth greater than the standard tooth depth. These are the
2.7 No-Standard Basic Rack Tooth Profiles 77
Fig. 2.17 Correlation between the pressure angle and the tooth profile shape, for a given base
circle
Fig. 2.18 Variation of the ratio qa =q20 as a function of the pressure angle a
78 2 The Geometry of Involute Spur Gears
Fig. 2.19 Particular no-standard basic rack tooth profile with tip and root modification with
respect to the standard basic rack tooth profile
so-called high depth teeth, that can have working depth hwP ¼ 2:25 m or even
hwP ¼ 2:50 m rather than hwP ¼ 2 m. These thoothings are however characterized
by a higher sliding velocity, so that the scuffing danger increases. Therefore, ad-
dendum modifications in terms of profile shift thoroughly studied are necessary.
The tooth thickness at the top land is also reduced, and therefore it must be checked
carefully.
With intentional modifications compared to the trapezoidal shape of the standard
basic rack tooth profile, we can obtain corresponding intentional modifications with
respect to the involute profile (straight-line profile), for the toothed gear wheels we
want to achieve. The field of use of particular standard cutters based on this concept
(see Fig. 2.19), however, is limited, and the position and amount of the profile
modification that we want to obtain for the toothed wheel also depend on the
number of its teeth and addendum modification. Standard profiles with full radius
shape for the root fillet (on the tool qa0 ¼ 0:38m) lead to a higher bending strength
(both static and fatigue strength) of the tooth root. Sometimes they are used when
the flanks of the tooth are hardened by induction, in order to compensate for the less
of mechanical strength due to hardening of the area corresponding to the length of
path of recess.
Standard profiles of the tool with protuberance (Fig. 2.16b) allow to realize a root
relief, for which the notches in the tooth root are avoided in the subsequent grinding
operations. In the case of toothed gear wheels with a high number of teeth it is,
however, necessary to control the extent to which the active portion of the dedendum
2.7 No-Standard Basic Rack Tooth Profiles 79
References
Abstract In this chapter, the minimum number of teeth to avoid interference in the
operating conditions of rack-pinion, external and internal spur gear pairs is first
determined, and useful considerations on this topic are made. Other important
characteristic quantities of these types of gears are then determined, such as the
lengths of the path of contact, path of approach and path of recess as well as the
lengths of the corresponding arcs and values of the angles correlated to the latter.
The relationships for calculating the transverse contact ratio and the instantaneous
radii of curvature of the involute mating profiles are obtained, and the generalized
laws of gearing are described. Subsequently, the main kinematic quantities related
to the rolling and sliding motions of the mating teeth flanks are determined, with
particular attention to relative sliding and specific sliding to which wear damage
and gear efficiency are correlated. In particular, the relationships that express the
instantaneous and average efficiencies as well as those that define the total contact
efficiency and effective driven torque along the path of contact are determined.
Finally, short notes on the main cutting processes of cylindrical spur gears are
given.
We have already seen that, for an external gear pair with unmodified involute
profiles (Fig. 2.9), to prevent the theoretical interference or involute interference,
the maximum radii ra1 max and ra2 max of the addendum circles of the pinion and gear
wheel can be equal to O1 T2 and O2 T1 respectively. In these limit conditions, the
addendum circles pass through the points of interference T1 and T2 previously
defined (see Sect. 1.5, and Figs. 1.8 and 2.10).
Now let’s go back to examine the external gear pair shown in Fig. 2.10, where
the condition of non-interference is met, since the two addendum circles intersect
the line of action at points E and A, located internally with respect to the inter-
ference points T2 and T1 . The condition of non-interference can be expressed
analytically by the following two inequalities, which must be both satisfied (see
Colbourne [13], Ferrari and Romiti [21]).
By the same Fig. 2.10, however, we deduce immediately that the more
restrictive inequality is that related to the addendum circle of the gear wheel having
a larger diameter. Therefore, it is sufficient to consider the inequality corresponding
to it, i.e. the second of the inequality Eqs. (3.1). Then, since CT1 ¼ r1 sin a, the
latter can be written in the form:
CA r1 sin a: ð3:2Þ
Evidently (see Fig. 2.10) the risk of theoretical interference is increased when
the point A is approaching the point T1 or when the point T1 approaches the point
A. This occurs, respectively, when the gear wheel diameter increases, while that of
the pinion remains unchanged, or when the pinion diameter decreases, while that of
the wheel remains unchanged. Therefore, the risk of interference is increased when
the number of teeth z2 of the gear wheel increases, and the number of teeth z1 of the
pinion decreases.
This case corresponds to the limit case of an external cylindrical spur gear of
infinite pitch radius r2 ðr2 ! 1Þ, while r1 remains unchanged. From Eq. (3.6) we
can deduce that, when r2 increases, x also grows. Therefore, for an external gear
pair, the maximum value of x is that corresponding to r2 ! 1, i.e. to the case of
the rack. In this case ðr2 ! 1Þ, from Eq. (3.6) we get:
3.1 Minimum Number of Teeth to Avoid Interference 83
ha2
x¼ ð3:7Þ
sin a
and, since x ¼ CA; from this, taking into account the inequality Eq. (3.2), we
obtain:
ha2
r1 sin a; ð3:8Þ
sin a
where k1 and k2 are the addendum factors, i.e. the proportionality coefficients that
correlate the addendum of the two members of the gear pair to the module. It is to
remember that:
• ha1 ¼ ha2 ¼ m; for the usual standard sizing;
• ha1 ¼ ha2 ¼ 0:8m; for the usual stub-sizing;
• ha1 6¼ m and ha2 6¼ ha1 ; for particular sizings, that we will discuss in the next
chapters.
From the inequality Eq. (3.8), taking account of the second of Eq. (3.9), and
recalling that m ¼ ð2r1 =z1 Þ; we obtain:
2k2
z1 ¼ zmin : ð3:10Þ
sin2 a
This is the minimum number of teeth of the toothed gear wheel, which must
mesh with the rack of the corresponding series. If we assume zmin in accordance
with the inequality (3.10), we avoid the theoretical interference during the meshing
between the wheel under consideration and the rack having the same circular pitch.
Therefore, for the considerations made above, we also avoid this type of interfer-
ence between the same wheel and any other external wheel drawn in accordance
with the modular sizing.
Of course, with the same inequality (3.10) we can calculate the minimum
number of teeth that can be cut without interference, when instead of considering
the meshing conditions between a pinion and a rack, we consider the cutting
operation, that is, we replace the rack-type cutter to the rack gear.
84 3 Characteristic Quantities of Cylindrical Spur Gears …
Now let’s examine how to avoid theoretical interference during the meshing of the
two members of an external cylindrical spur gear characterized by a given gear
ratio u ¼ ðz2 =z1 Þ ¼ ðr2 =r1 Þ. As we said above, in this case we find a value of the
minimum number of the teeth lower than the value obtained for the rack-pinion
pair.
From Fig. 2.10, we infer that, for a given value of the pressure angle a, the
maximum value of x ¼ CA corresponds to the maximum value of ha2 . Therefore,
the limit value of ha2 to avoid the theoretical interference is the one obtained
from Eq. (2.5) when we substitute in it the maximum value xmax ¼ CAmax ¼
CT1 ¼ r1 sin a; given by the second of Eq. (3.1). So we have:
ha2 ð2r2 þ ha2 Þ ¼ r1 sin að2r2 sin a þ r1 sin aÞ ¼ r1 sin2 að2r2 þ r1 Þ: ð3:11Þ
Introducing in this equation the absolute value u ¼ jr2 =r1 j of the gear ratio, and
simplifying it, we get the following algebraic equation of the second degree:
2
ha2 ha2
þ 2u ð2u þ 1Þ sin2 a ¼ 0; ð3:12Þ
r1 r1
This equation provides the limit value of z1 , as ha2 ¼ k2 m and r1 ¼ ðmz1 =2Þ.
Therefore, we obtain the following final relationship:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u þ u2 þ ð2u þ 1Þ sin2 a
2k2
z1 ¼ zmin qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 2k2 :
ð2u þ 1Þ sin2 a
u þ u2 þ ð2u þ 1Þ sin2 a
ð3:14Þ
In the case of an internal cylindrical spur gear pair, the value of the minimum
number of teeth to avoid theoretical interference is greater than that given by
3.1 Minimum Number of Teeth to Avoid Interference 85
Eq. (3.10). Mutatis mutandis, proceeding in a similar way to that described in the
previous section, we obtain the following relationship:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uþ u2 ð2u 1Þ sin2 a
2k2
z1 ¼ zmin qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 2k2 :
ð2u 1Þ sin2 a
u u2 ð2u 1Þ sin2 a
ð3:15Þ
The chart shown in Fig. 3.1 allows calculating the minimum number of teeth, z1 min ,
for external and internal gears without profile shift (see Chap. 6), for different
values of the pressure angle, a; and the standard values of k2 (i.e., k2 ¼ 1), as a
function of the ratio ð1=uÞ ¼ ðz1 =z2 Þ: In this figure, we preferred to use the ratio
ð1=uÞ ¼ ðz1 =z2 Þ, instead of the gear ratio, u, that appears in the inequalities
Eqs. (3.10, 3.14 and 3.15). This in order to have an absolute value of the input data
on the ordinate axis (positive for external gears, and negative for internal gears)
included in the range between 0 and 1, instead in the range between 1 and 1.
The use of the chart is very simple. First, we enter it, on the ordinate axis, with a
horizontal line corresponding to the design value of the ratio ð1=uÞ ¼ ðz1 =z2 Þ: We
then extend this line to the right, until it intersects the curve of the value of the
pressure angle referred to the nominal center distance. Finally, we exit with a
vertical line up to the abscissa, where we read z1 min . For sizing other than the usual
standard, i.e. for k2 6¼ 1, the minimum number of teeth, z1 min , is obtained by
multiplying the value derived of the chart for the value of k2 actually used.
Both from Eqs. (3.10, 3.14 and 3.15), and from the curves shown in Fig. 3.1, the
strong influence of the pressure angle a on the minimum number of teeth z1min is
very clear. Equations (3.14) and (3.15) also show that, for an external or internal
gear pair, the minimum number of teeth z1 min of the wheel with smaller diameter
(the pinion) is proportional to the addendum factor k2 of the mating gear.
For the reasons that we will describe later, often the gear designer is forced to
choose a pinion with a number of teeth, z1 min , less than that obtained using Eqs. (3.14
or 3.15). To avoid the theoretical interference that would occur in this case, a
possible solution is to reduce k2 , multiplying it with the ratio between the desired
number of teeth and the minimum number of teeth obtained by using the above
equations for the given values of the pressure angle a and gear ratio u. The negative
aspect of this solution consists in the fact that the path of contact is reduced, due to
the consequent decrease of the addendum. In addition, if we want to keep the same
bottom clearance that we have in the usual standard sizing ðc ¼ 0:25mÞ; once the
value of k2 has been calculated in accordance with the procedure
described above,
the dedendum must be obtained using the relationship hf =m ¼ k2 þ 0:25:
86 3 Characteristic Quantities of Cylindrical Spur Gears …
Fig. 3.1 Minimum number of teeth z1min as a function of the ratio ð1=uÞ ¼ ðz1 =z2 Þ, for k 2 ¼ 1
and different pressure angles (a ¼ 15 ; a ¼ 17 300 ; a ¼ 20 ; a ¼ 22 300 ; a ¼ 25 ; a ¼ 27 300 ;
a ¼ 30 Þ
The problem of determining the minimum number of teeth z1 min to prevent the
theoretical interference must be considered not only in the operating conditions of
the gear, but also during the related cutting operations. From this second point of
view, it is noteworthy that the minimum number of teeth z1min defined by the
Eqs. (3.14 and 3.15) can be cut without interference only if the cutting is performed
by milling, with form milling cutters (see Sect. 3.11.1).
If instead the teeth are cut with generation process, by means of rack-type
cutters, hobs, and pinion-type cutters (Fellows gear shapers or simply Fellows
shapers), the minimum numbers of teeth that we can cut without interference are:
3.2 Considerations on the Minimum Number of Teeth 87
generation cutting process of a gear wheel a light interference between the cutter
profile and profile generated by it is not only perfectly permissible, but it is indeed
often desirable. This is because such a light interference undoubtedly determines a
light reduction of the active profile, but in return offers the significant advantage of
eliminating the part of involute curve immediately near the base circle, to which, as
we will see, different problems are related.
Therefore, in this framework, it is admitted that, in practice, the minimum
number of teeth that can be cut by a generation process, using a rack-type cutter, is
equal to 5/6 of the theoretical value given by Eq. (3.10). Figure 3.2, valid for
Fig. 3.2 Theoretical and practical number of teeth, z1min , as a function of the pressure angle, a, for
cylindrical spur gears cut with rack-type cutter
3.2 Considerations on the Minimum Number of Teeth 89
cylindrical spur gears obtained by cutting with a rack-type cutter, shows the vari-
ation of z1 min as a function of the pressure angle a, both in the cutting theoretical
condition and in cutting practical condition.
Once the geometric quantities of the two members of the gear pair under consid-
eration have been determined, it is necessary to assess the length g ¼ ga of the path
of contact (see Figs. 1.8 and 2.10). Considering the case in which the pinion drives
the wheel or rack (mutatis mutandis, in the same way we can proceed when the
pinion is driven by the wheel or rack), the total length of the path of contact is equal
to thesum of the
lengths
of the
path of approach, AC and path of recess, CE i.e.,
ga ¼ gf 1 þ ga1 ¼ ga2 þ gf 2 : The three possible cases are those of external gear
pairs, pinion-rack pairs and internal gear pairs.
For an external gear pair (Fig. 3.3a), we have:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi
gf 1 ¼ ga2 ¼ AC ¼ AT2 CT2 ¼ ra2 rb2 2 r sin a
2 ð3:16Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ga1 ¼ gf 2 ¼ CE ¼ T1 E T1 C ¼ 2 r 2 r sin a:
ra1 b1 1 ð3:17Þ
So, if we consider that ra2 ¼ ðr2 þ k2 mÞ, ra1 ¼ ðr1 þ k1 mÞ, rb2 ¼ r2 cos a,
rb1 ¼ r1 cos a, r2 ¼ ðmz2 =2Þ and r1 ¼ ðmz1 =2Þ, we get:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
gf 1 ¼ ga2 ¼ z22 sin2 a þ 4k2 ðz2 þ k2 Þ z2 sin a ð3:18Þ
2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
ga1 ¼ gf 2 ¼ z21 sin2 a þ 4k1 ðz1 þ k1 Þ z1 sin a : ð3:19Þ
2
To the lengths gf 1 ¼ ga2 and ga1 ¼ gf 2 , measured along the line of action, the
lengths of the arc of approach, e1 ¼ ea2 , and arc of recess, e2 ¼ ea1 , correspond.
They are given respectively by:
AC
e1 ¼ ea2 ¼ ð3:20Þ
cos a
CE
e2 ¼ ea1 ¼ : ð3:21Þ
cos a
It should be noted that the symbol, e, used here to indicate arcs of contact, should
not be confused with the same symbol, also used to indicate space width; here the
90 3 Characteristic Quantities of Cylindrical Spur Gears …
(a) (b)
(c)
Fig. 3.3 Schematic diagrams for evaluation of lengths gf 1 ¼ ga2 , ga1 ¼ gf 2 , and
ga ¼ gf 1 þ ga1 , for: a external gear pair; b pinion-rack gear pair; c internal gear pair
symbols e, e1 and e2 indicate respectively the total arc of contact, arc of approach,
and arc of recess. It should also be noted that the most precise double subscript calls
only the addendum flanks of the two members of the gear pair. It indicates that, in
the path of approach, the addendum flank of the driven wheel is in contact, while in
the path of recess the addendum flank of the driving pinion is in contact. For
brevity, from here on, we will use only one subscript, as misunderstandings are not
possible.
These arcs of approach and recess are the lengths of addendum contact
respectively of the wheel and pinion, for which they are measured along the pitch
3.3 Lengths of the Path and Arc of Contact, and Angles of Contact 91
circles of the wheel and, respectively, of the pinion. Usually however, according to
the most commonly used convention, the arc of approach and arc of recess are
respectively measured on the pitch circles of the driving and driven members of the
gear pair under consideration. However, the complete equivalence of measuring
these arcs on one or the other of the pitch circles of the two members of the gear
pair is obvious, given that they roll over each other without sliding.
The correlated angles of approach and recess are respectively found by dividing
the path of approach, AC, and path of recess, CE, by the radius of the base circle of
the driving (or driven) wheel; they are expressed in circular measure or radians. In
completely equivalent terms, using quantities referred to the pitch circle of the
pinion, they are given by the following relationships:
e1
n1 ¼ ð3:22Þ
r1
e2
n2 ¼ : ð3:23Þ
r1
Thus, the total lengths of the path of contact and arc of contact, and the total
angle of contact are given by the following final relationships:
ga ¼ gf 1 þ ga1
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
¼ z22 sin2 a þ 4k2 ðz2 þ k2 Þ
2 ð3:24Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
þ z1 sin a þ 4k1 ðz1 þ k1 Þ ðz1 þ z2 Þ sin a
2 2
ga
e ¼ e1 þ e2 ¼
qcos a
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
¼ z2 sin2 a þ 4k2 ðz2 þ k2 Þ þ z21 sin2 a þ 4k1 ðz1 þ k1 Þ ðz1 þ z2 Þ sin a
2
2 cos a
ð3:25Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
n ¼ n1 þ n 2 ¼ z22 sin2 a þ 4k2 ðz2 þ k2 Þ
z1 cos a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð3:26Þ
þ z1 sin a þ 4k1 ðz1 þ k1 Þ ðz1 þ z2 Þ sin a :
2 2
therefore, if we consider that ha2 ¼ k2 m, ra1 ¼ ðr1 þ k1 mÞ, rb1 ¼ r1 sin a, and
r1 ¼ ðmz1 =2Þ, we get:
To the lengths gf 1 ¼ ga2 and ga1 ¼ gf 2 measured along the line of action, the
lengths of the arc of approach, e1 ; and arc of recess, e2 , correspond; they are given
by Eqs. (3.20) and (3.21). For this type of gear pair, these quantities are also the
lengths of addendum contact of the rack and pinion, but with the difference that
they are both measured along the pitch circle of the pinion. Thus, the related angles
of approach and recess are given by:
e1 AC
n1 ¼ ¼ ð3:31Þ
r1 r1 cos a
e2 CE
n2 ¼ ¼ : ð3:32Þ
r1 r1 cos a
Thus, the total lengths of the path of contact and arc of contact, and the total
angle of contact are given by:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m 2k2
ga ¼ gf 1 þ ga1 ¼ z21 sin2 a þ 4k1 ðz1 þ k1 Þ z1 sin a þ ð3:33Þ
2 sin a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ga m 2k2
e ¼ e1 þ e 2 ¼ ¼ z21 sin2 a þ 4k1 ðz1 þ k1 Þ z1 sin a þ
cos a 2 cos a sin a
ð3:34Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
e 1 2k2
n ¼ n1 þ n 2 ¼ ¼ z21 sin2 a þ 4k1 ðz1 þ k1 Þ z1 sin a þ :
r1 z1 cos a sin a
ð3:35Þ
ga1 ¼ gf 2 ¼ CE ¼ T2 E T2 C ¼ T2 T1 þ T1 E T2 C
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð3:37Þ
¼ a sin a þ 2 r 2 r sin a:
ra1 b1 2
3.3 Lengths of the Path and Arc of Contact, and Angles of Contact 93
So, if we consider that ra2 ¼ ðr2 k2 mÞ, ra1 ¼ ðr1 þ k1 mÞ, rb2 ¼ r2 cos a,
rb1 ¼ r1 cos a, r2 ¼ ðmz2 =2Þ and r1 ¼ ðmz1 =2Þ, we get:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
gf 1 ¼ ga2 ¼ z22 sin2 a 4k2 ðz2 k2 Þ þ z2 sin a ð3:38Þ
2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
ga1 ¼ gf 2 ¼ z21 sin2 a þ 4k1 ðz1 þ k1 Þ z1 sin a : ð3:39Þ
2
The lengths at the arcs of approach and recess e1 and e2 related to the lengths
gf 1 ¼ ga2 and ga1 ¼ gf 2 are given by Eqs. (3.20) and (3.21). They are the lengths of
addendum contact of the annulus (the ring gear) and pinion, measured along the pitch
circles of the annulus and pinion respectively. The related angles of the approach and
recess are given, also for this type of gear pair, by Eqs. (3.22 and 3.23).
Thus, the total lengths of the path of contact and arc of contact, and the total
angle of contact are as follows:
ga ¼ gf 1 þ ga1
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
¼ z22 sin2 a 4k2 ðz2 k2 Þ
2 ð3:40Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
þ z21 sin2 a þ 4k1 ðz1 þ k1 Þ þ ðz2 z1 Þ sin a
ga
e ¼ e1 þ e2 ¼
cos a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m
¼ z22 sin2 a 4k2 ðz2 k2 Þ ð3:41Þ
2 cos a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
þ z21 sin2 a þ 4k1 ðz1 þ k1 Þ þ ðz2 z1 Þ sin a
n ¼ n1 þ n2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
¼ z22 sin2 a 4k2 ðz2 k2 Þ
z1 cos a ð3:42Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
þ z21 sin2 a þ 4k1 ðz1 þ k1 Þ þ ðz2 z1 Þ sin a :
By definition, the transverse contact ratio ea is the ratio between the length of the
arc of contact and circular pitch or, in equivalent terms, the ratio between the length
of the path of contact and transverse base pitch, i.e.:
94 3 Characteristic Quantities of Cylindrical Spur Gears …
e ga
ea ¼ ¼ : ð3:43Þ
p pbt
Since pbt ¼ p cos a and p ¼ pm; considering the pairs of Eqs. (3.24) and (3.25),
(3.33) and (3.34), (3.40) and (3.41), respectively valid for external gear pairs,
pinion-rack gear pairs, and internal gear pairs, we get:
• For an external gear pair:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
ea ¼ z22 sin2 a þ 4k2 ðz2 þ k2 Þ
2p cos a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð3:44Þ
þ z21 sin2 a þ 4k1 ðz1 þ k1 Þ ðz1 þ z2 Þ sin a ;
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 2k2
ea ¼ z21 sin2 a þ 4k1 ðz1 þ k1 Þ z1 sin a þ ð3:45Þ
2p cos a sin a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
ea ¼ z22 sin2 a 4k2 ðz2 k2 Þ
2p cos a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð3:46Þ
þ z1 sin a þ 4k1 ðz1 þ k1 Þ þ ðz2 z1 Þ sin a :
2 2
We have already said that the meshing cycle of one tooth pair begins when the
teeth first make contact, and ends when the contact is broken. In order the trans-
mission of motion between the two meshing gear wheels is continuous, there must
clearly be at least one tooth pair in contact at all times. However, it is well known
that a smooth working condition is only possible when the contact between one
tooth pair continues until sometime after the contact between the next pair of teeth
is started.
Essentially, there must be parts of the meshing cycle during which two pairs of
teeth are in contact simultaneously. The transverse contact ratio is then a measure of
the amount of this overlap. For example, ea ¼ 1:4 means that one pair of teeth is in
contact for 60% of the meshing cycle, and two pairs of teeth are in meshing for the
3.4 Transverse Contact Ratio 95
remaining 40%. Indeed, ea ¼ 2:2 means that three pairs of teeth are in contact for
20% of the meshing cycle and two pairs of teeth are in contact for the remaining
80% of the meshing cycle. Thus, in this last case, at least two pairs of teeth are
theoretically in contact at all times; whether or not they are actually in contact
depends on the tooth stiffness, applied load, and precision of manufacture and
assembly.
In general, the greater the transverse contact ratio the smoother and quieter the
working condition of the gear pair. From the point of view of the gear design, it is
necessary to bear in mind that, for gears with low speed of rotation, it is not
appropriate that the value of ea falls below (1.2 1.3), while for fast gears the
value of ea must be greater than (1.4 1.5). In any case, the value of ea must not
fall below the unit, otherwise the continuity of motion is lost.
If the addenda of the two members of an external gear pair (Fig. 3.3a) assume
their maximum values k1 m and k2 m, compatible with non-interference (in this case,
the point A in the figure coincides with the point T1 , while the point E coincides
with the point T2 ), we will have (see also Giovannozzi [24]):
From this equation, we obtain the following minimum value of the sum of the
numbers of teeth of wheel and pinion compatible with the limiting condition of
continuity of the transmission of motion ðea ¼ 1Þ:
Therefore, in this case, to ensure the continuity of motion, the minimum sum
ðz1 þ z2 Þmin must not fall below: 24 teeth for a ¼ 15 ; 18 teeth for a ¼ 20 ; and 11
teeth for a ¼ 30 .
Both from Fig. 3.3a and Eqs. (3.18, 3.19, 3.20, 3.21 and 3.44), we deduce that
an increase in the addendum factors k1 and k2 increases the arc of approach and the
arc of recess, and consequently, the transverse contact ratio between external gears
in meshing. The limiting values for the arc of approach and arc of recess are when
either one of them exceeds the pressure angle a. The transverse contact ratio for an
external gear pair also increases when the radius of the pitch circle of the gear
members increases. By contrast, an increase of the pressure angle causes a decrease
of the transverse contact ratio. Figure 3.4 shows the effect of the numbers of teeth z1
and z2 on the transverse contact ratio, for an external gear pair with k1 ¼ k2 ¼ 1 and
a ¼ 20 .
Mutatis mutandis, starting from Fig. 3.3b and Eqs. (3.29, 3.30, 3.20, 3.21 and
3.45), entirely similar considerations can be made for the pinion-rack gear pair, as it
can be considered the limit case of an external gear pair for z2 ¼ 1
ðu ¼ z2 =z1 ¼ 1Þ. The same Fig. 3.4 is valid for this gear pair, characterized by
k1 ¼ k2 ¼ 1 and a ¼ 20 .
96 3 Characteristic Quantities of Cylindrical Spur Gears …
Fig. 3.4 Effect of the numbers of teeth z1 and z2 on the transverse contact ratio ea , for external
gear pairs and pinion-rack gear pairs, with k 1 ¼ k 2 ¼ 1 and a ¼ 20
Both from Fig. 3.3c and Eqs. (3.38), (3.39), (3.20), (3.21) and (3.46), we infer
that also for an internal gear pair an increase in the addendum factors k1 and k2
increases the arc of approach and the arc of recess, and consequently, the transverse
contact ratio. We also infer that the transverse contact ratio increases when the
radius of the pitch circle of the pinion increases, while it decreases when the radius
of the pitch circle of the annulus increases. In addition, in this case, an increase of
the pressure angle determines a decrease of the transverse contact ratio.
Let consider a current point P between two mating profiles of the external gear pair
shown in Fig. 2.6. This point, as Fig. 2.10 shows, can be located between the points
A (start of contact) and E (end of contact). The determination of the radii of
curvature of the two involute profiles is very simple, because we can make use of
one of the special properties of the involute curve [25].
In Sect. 2.1 we found that at every point P the involute of a base circle is normal
to the generating line, and in any position of this generating line (see position Pi Ti
in Fig. 2.1) its point of tangency with the base circle is the instantaneous center of
rotation. Therefore, an infinitesimal length of involute arc at point Pi (Fig. 2.1) is
indistinguishable from a circular arc having its center at point Ti and radius Pi Ti ; in
other words, Pi Ti is the radius of curvature of the involute at point Pi .
3.5 Radius of Curvature of Involute Tooth Profiles … 97
For an external gear pair, whatever the position of the point of contact P on the
line of action (see Fig. 2.6), the instantaneous centers of curvature of the two
conjugate profiles are then the points T1 and T2 , while the instantaneous radii of
curvature are q1 ¼ PT1 for the driving profile, and q2 ¼ PT2 for the driven profile.
We can achieve the same result (i.e., T1 and T2 instantaneous centers of curvature,
and q1 ¼ PT1 and q2 ¼ PT2 instantaneous radii of curvature of the two mating
profiles) using the well-known Euler-Savary equation for planar motion, as the
involutes considered here are planar curves (see Belfiore et al. [5]).
Then, if we denote by gP the algebraic value of the distance between the point of
contact P and the instantaneous center of rotation C (Fig. 2.6), measured along the
line of action, the instantaneous radii of curvature of the two profiles will be
expressed by the following equations:
If for an internal gear pair (Fig. 2.7) we repeat, mutatis mutandis, the same
considerations made for an external gear pair, we still find that the points T1 and T2
are the instantaneous centers of curvature, while the instantaneous radii of curvature
of the two mating profiles are given by:
Now we analyze the kinematics of contact between two involute conjugate profiles
of an external gear pair. The general validity of the discussion is not compromised,
as the considerations and deductions made for the external gear pairs are provided,
with the variations of the case, even for the internal gear pairs. The choice of
considering the external gear pairs is motivated by the fact that these gears are the
most widely used in practical applications.
Here we assume that the meshing gears and mechanical elements for supporting
loads, which globally constitute the system for power transmission, are rigid. In
fact, all of these system components will deform as a function of the transmitted
loads. These deformations, due to deflection of the teeth relative to the gear blanks,
bending and torsional displacements of the shafts and gear blanks, deflections of the
bearing supports, and compliance of the housing used to support the bearings, will
be introduced in a later stage of deepening.
In Fig. 2.3 we highlighted the fact that progressively increasing arcs of the
involute correspond to equal arcs of the base circle described by the generating line
during its rolling without sliding about the base circle. It follows that the point of
contact between two mating involute profiles, associated with their base circles,
describes on the two involute curves two arcs of different lengths in the same time
interval.
Consequently, during the meshing cycle, the contact between the two mating
profiles is not a pure rolling contact, but a contact of rolling and sliding together.
Figure 3.5, which refers to two mating involute profiles related to two base circles
having the same diameters, show that, whatever the direction of rotation, the lengths
of the involute arcs described by the point of contact on the two profiles are not equal.
It is therefore necessary to examine the relative motion between the two mating
profiles, to which a continuously variable relative sliding between the two involute
curves is associated. Even in the simplest case of planar gears (in this case, the flank
surfaces are reduced to their transverse sections, which are the involute profiles),
different types of velocities come into play. These velocities are: angular velocity;
tangential velocity at pitch circle, also called pitch line velocity; tangential velocity
at the base circle; absolute velocity; sliding velocity; entrainment velocity; spinning
velocity; etc. The evaluation of these velocities is based on an idealized center of
contact, which coincides with the theoretical point of contact, and neglects the
effects of micro-slip (see Cattaneo [10], Johnson [31], Kalker [33], Mindlin [38],
Polach [40]).
With these velocities, it is possible to determine the tribological conditions at the
contact zone. As matter of the fact, the mechanisms of wear as well as the analysis
of the rheological phenomena between two surfaces in contact depend on the
relative displacements between the two surfaces. Only the magnitudes of these
velocities are needed to predict such tribological conditions and related effects
during the meshing.
100 3 Characteristic Quantities of Cylindrical Spur Gears …
Fig. 3.5 Meshing of two involute profiles related to two base circles having the same diameters
With these premises, we examine in details the kinematics of the external gear
pair shown in Fig. 2.10, and considered as a mechanism. It is a three-link mech-
anism consisting of three elements: the two-toothed gears and the fixed frame. This
mechanism, according to the Grübler’s mobility criterion (see Ambekar [1], Di
Benedetto and Pennestrì [16], Hain [26]), has mobility equal to unity, i.e. it is a
1-dof mechanism. In other terms, as one of the two gears rotates, the other gear
must rotate according to the transmission function, in defined by Eq. (1.17).
In order to analyze the kinematic behavior of gears characterized by a
non-constant transmission ratio (for example, non-circular gears, elliptical gears,
harmonic gears, and similar), the concept of transmission function must also be
generalized. For these types of gears, which are not discussed in this monographic
textbook, the transmission function is not constant. However, it is always given by
Eq. (1.17) and, as we have already said, helps to define the first law of gearing,
according to which for a given position of the axes and a given transmission
function, a unique relationship exists between the instantaneous displacement of the
driven member and the instantaneous displacement of the driving member.
In general terms, the transmission function is the reciprocal of the instantaneous
transmission ratio. Therefore, for a gear train it is defined as the quotient of the
instantaneous angular displacement of the driven member, du0 , divided by the
instantaneous angular displacement of the driving member, dui , i.e.
in ¼ ð1=iÞ ¼ ðdu0 =dui Þ. This function has a constant value for circular gears, while
for non-circular gears and similar it is not constant. To generalize the problem,
taking into account both circular and non-circular gears, a further constraint
equation must be added to the aforementioned first law of gearing. This constraint
equation is given by the following relationship:
3.6 Kinematics of Gearing: Rolling and Sliding Motions of the Teeth Flanks 101
2p
Z
in dui ¼ rational function; ð3:54Þ
0
this function must always be a rational function because, otherwise, the driven
member could not sustain an indefinite number of cycles with the desired trans-
mission function [22].
To minimize dynamic effects, the transmission function must be as smooth as
possible. The actual graph of this function assumes R decisive Rimportance in the
design of non-circular gears. Since the integrals du0 and in dui define the
angular position of the driven member, u0 , for a given angular position of the
driving member, ui , it is evident that any discontinuity of the transmission function
would result in a not one to one correspondence between these two angular posi-
tions, with the risk of having two or more simultaneous angular positions of the
driven member for a given angular position of the driving member.
Since the transmission function expresses the change du0 in the angular position
u0 of the driven member with respect to the change dui in the angular position ui
of the driving member, it is also known as velocity ratio. The derivative i0n ¼
d 2 u0 =d 2 ui is called acceleration ratio; it is related to the curvature of the two
centrodes. If this derivative is too high, the centrodes become pointed, so their
manufacture is difficult. Furthermore, as the required output torque increases as the
acceleration ratio increases, the loads acting on the mating teeth will also increase
accordingly. The next derivative i00n ¼ d 3 u0 =d 3 ui is called jerk ratio, while the two
subsequent derivatives i000 4 4 0000
n ¼ d u0 =d ui and in ¼ d u0 =d ui are called, respec-
5 5
du1 du2
x1 ¼ x2 ¼ ; ð3:55Þ
dt dt
v ¼ v1 ¼ r1 x1 ¼ v2 ¼ r2 x2 : ð3:56Þ
In fact, the rolling of pitch circles relative to one another is a pure rolling, i.e.
without sliding.
102 3 Characteristic Quantities of Cylindrical Spur Gears …
As shown in Fig. 1.5 and, on a larger scale, in Fig. 3.6, the point of contact P is
any current point of path of contact in any transverse section. This point can be
considered belonging to both the driving and driven profiles. The absolute velocity
vector vP1 of the point P belonging to the driving profile is a vector perpendicular to
the line connecting P and the fixed center O1 . Similarly, the absolute velocity vP2 of
the same point P belonging to the driven profile is a vector perpendicular to the line
connecting P and the fixed center O2 . These vectors are respectively given by the
two cross products (it is still to be noted that vectors are indicated by bold letters):
Obviously, these two vectors are not collinear. They become collinear only
when, during meshing, the point P lies on the line connecting the two fixed centers
O1 and O2 ; that is when it comes to coincide with the pitch point C.
Vector vP1 can be resolved into two components: one of these components, vn1 ,
is parallel to the line of action, while the other component, vt1 , is perpendicular to it.
Similarly, vector vP2 can also be resolved into two components: one of these
components,vn2 , is parallel to the line of action, while the other component vt2 is
perpendicular to it (Fig. 3.6).
In order to satisfy the basic law of conjugate profiles, and the profiles themselves
remain in meshing, the components vn1 and vn2 along the line of action must be
equal, i.e. vn1 ¼ vn2 ¼ vn ; otherwise, the two profiles would no longer be in contact,
(a) (b)
Fig. 3.6 a Velocity vectors associated to any position of the current point of contact P along the
line of action; b Detail on a larger scale
3.6 Kinematics of Gearing: Rolling and Sliding Motions of the Teeth Flanks 103
but separated. These velocities can be conceived as the absolute velocities of points
of the base circles about their fixed centers. Thus, they are nothing more than the
tangential velocities at base circles, so we can write in terms of their absolute
values:
In order to calculate the components vt1 and vt2 , we can follow two alter-
natives, both related to the motion of the point P, whether belonging to the driving
or driven profiles. According to the first alternative, the motion of point P is
regarded as compounded of a motion along the line of action, directed from T1 to
T2 ; and a rotation around T1 and T2 . According the second alternative, kinemati-
cally equivalent to the first, the motion of point P is regarded as compounded of a
motion along the straight-line tangent to the pitch circle at the pitch point, and a
rotation around the pitch point. Here it is convenient to follow the first alternative,
as we have already calculated the absolute values of the components vn1 and vn2 :
According to this alternative, taking into account Eqs. (3.49) and (3.50), we have,
in terms of their absolute values:
gP
vt1 ¼ x1 q1 ¼ x1 ðr1 sin a þ gP Þ ¼ v sin a þ ð3:59Þ
r1
gP
vt2 ¼ x2 q2 ¼ x2 ðr2 sin a gP Þ ¼ v sin a : ð3:60Þ
r2
With pinion driving, the term gP is regarded as negative during the approach
path of contact (contact between the dedendum flank of the pinion and addendum
flank of the wheel), and positive during the recess path of contact (contact between
the addendum flank of the pinion and dedendum flank of the wheel).
Notoriously, two virtual equivalent cylinders, with the appropriate rolling and
sliding motions along their contact line, may represent the conditions of teeth
engagement. In the case of spur gears, all transverse sections are similar, and the
contact line between the two cylinders is perpendicular to the plane of section, and
thus parallel to the axes of the gears. In any transverse section, the two cylinders
become two circles, also virtual, touching at point P (this is the intersection of the
contact line between the cylinders with the transverse section under consideration),
and having centers at fixed points T1 and T2 , and radii q1 and q2 , the latter variable
without solution of continuity during the meshing (Fig. 3.6a). So, it is evident
that vt1 and vt2 are nothing more than the instantaneous tangential velocities or
absolute velocities at these circles; these velocities are known in the scientific
literature as the rolling velocities.
As Eqs. (3.59) and (3.60) show, the magnitudes of the absolute values of the
rolling velocities vt1 and vt2 vary linearly as a function of the position of the
point of contact P on the line of action, i.e. as a function of gP (note that, in the case
of non-involute profiles, these velocities differ in both magnitude and direction), but
104 3 Characteristic Quantities of Cylindrical Spur Gears …
vt1 increases and vt2 decreases during meshing. These velocities have different
magnitudes, but they are collinear, as their straight line of application is the com-
mon tangent to the mating profiles at point of contact, that is the perpendicular
straight line at point P to the line of action; however, they become equal only when,
during the meshing, the point of contact P comes to coincide with the pitch point C.
Equation (3.59) shows that the rolling velocity vt1 is always positive (it would be
equal to zero only in the case were the point of contact P were to coincide with the
point of interference T1 of the pinion), and acts outwards of the pinion, with the
pinion driving. In addition, the rolling velocity vt2 given by Eq. (3.60) is positive
(this would also be equal to zero in the case in which the point of contact P were to
coincide with the point of interference T2 of the wheel), and acts radially inwards
over the wheel teeth.
During recess (gP positive), with the pinion driving, the absolute value of vt1 is
greater than the one of vt2 , and the rolling velocity ratio, defined as the ratio of the
smaller to the larger rolling velocity, is given by:
vt2 r2 sin a gp
¼ : ð3:61Þ
vt1 u r1 sin a þ gp
Conversely, during approach (gP negative), the absolute value of vt1 is less than
the one of vt2 , and thus the rolling velocity ratio is given by:
vt1 u r1 sin a þ gp
¼ : ð3:62Þ
vt2 r2 sin a gp
to be used with the sign convention described above. This cumulative velocity is a
very important quantity, because it is related to the pressure that is generated within
the lubricant film between the flanks of mating teeth, and therefore to the actually
gear capacity to withstand the working loads. From Eq. (3.63) we deduce the
following dimensionless ratio, KR , between the absolute values of the velocities, vR
and v, called the cumulative factor:
vR gP 1
KR ¼ ¼ 2 sin a þ 1þ ; ð3:64Þ
v r1 u
rolling velocity, which affects the fluid film thickness at the mesh. From this point
of view, it would be appropriate to call this average rolling velocity the entrainment
velocity.
However, we must keep in mind that the entrainment of lubricant into the contact
zone between the teeth depends on the surface topology of the tooth flanks and their
angular displacements about the line of action at point of contact, as well as the
average rolling velocity vw . This velocity and the sliding velocity vg are parallel
vectors only for cylindrical spur gears. In the more general case of crossed or skew
gears, these vectors are no longer parallel; therefore, in this case, the angle between
them, and its variability during the meshing come into play.
The spinning velocity is defined as the change in this angle per unit time. Since
the sliding velocity vector vg , rolling velocity vector vw , and angle between vg and
vw are strongly affected by the pressure angle a, spiral angle b, and shaft angle R,
tribologists studying the rheological conditions of gearing in the presence of a
lubricant introduce an entrainment velocity more complex than that given by the
cumulative semi-velocity vector vw . On this subject we refer the reader to spe-
cialized textbooks and the relevant scientific literature (see Dooner [18], Dooner
and Seireg [19]).
Of course, since vt1 is generally different from vt2 , the two profiles slide relative
to one another. We define as sliding velocity vector vg between the two conjugate
profiles at point P the difference between the related rolling velocities vt1 and vt2 ,
that is, respectively:
where x ¼ ðx1 þ x2 Þ is the absolute value of the relative angular velocity, and the
upper and lower signs apply for vg1 and vg2 , respectively; the sign convention for gP
is that previously indicated.
The sliding velocity is therefore equal to zero at pitch point C, where gP ¼ 0,
while its absolute value increases proportionally to the distance gP from the pitch
point, reaching its maximum absolute value at points A and E (Fig. 2.10), where the
contact respectively begins and ends. This sliding velocity is a quantity extremely
important in the gear design. Indeed, it has a great influence on the power losses by
friction and the related heating, as well as on the wear conditions and scuffing load
carrying capacity.
Figure 3.7 shows the qualitative distribution curves of the values of the com-
ponents vn , vt1 and vt2 of the absolute velocities vP1 and vP2 of point of contact
106 3 Characteristic Quantities of Cylindrical Spur Gears …
P about the fixed centers O1 and O2 , as well as the ones of the cumulative velocity
vR and sliding velocities vg1 and vg2 , for a given external gear pair. Ictu oculi, we
see that vn is a constant, vt1 and vt2 have the same direction, while vg1 and vg2
change their direction as the point of contact crosses the pitch circles. More
specifically, vg1 is negative (i.e., it is directed from the pitch circle to the dedendum
circle) during the approach path of contact, and positive (i.e., it is directed from the
pitch circle to the addendum circle) during the recess path of contact. The sliding
velocity vg2 instead behaves in the opposite way to the sliding velocity vg1 .
This reversal of the direction of the sliding velocities at the pitch point causes, in
a contact inevitably characterized by friction, a cyclic variation of the contact forces
that come into play; it is undoubtedly a source of fatigue and noise.
In the optimal condition, the maximum values of vg1 and vg2 at points A and E,
where the contact begins and ends, must be equal (the reason of this is described in
the next section). In this condition, a fictitious average sliding velocity, vgm , is
introduced [39]; it is given by the following relationship:
this velocity, that is constant along the path of contact, is used for a simplified
calculation of the power losses due to friction in the contact between the teeth.
Finally, as dimensionless characteristic of the toothing geometry, a sliding factor
Kg is sometimes used, defined as the ratio between the absolute values of the sliding
velocity vg and tangential velocity v at pitch circle, that is:
vg 1 1 gP 1
Kg1;2 ¼ ¼ gP þ ¼ 1þ ; ð3:68Þ
v r1 r2 r1 u
on this factor, we can do the same considerations made for vg , and use the same sign
conventions.
The kinematic quantities concerning the internal gear pairs can be obtained with
considerations similar to those described above for the external gear pairs. This
work, however, is left to the reader. In this textbook, we are only going to provide
the relationships relating the internal gears. Equations (3.56, 3.58 and 3.59) do not
change. Equations (3.60 and 3.66) change and become, respectively:
gP
vt2 ¼ v sin a þ ð3:69Þ
r2
1 1
vg1;2 ¼ gP v ¼ gP ðx1 x2 Þ ¼ gP x; ð3:70Þ
r1 r2
where x ¼ ðx1 x2 Þ is the absolute value of the relative angular velocity of the
internal gear pair.
3.6 Kinematics of Gearing: Rolling and Sliding Motions of the Teeth Flanks 107
If we take into consideration Fig. 2.7 and on it we built diagrams of vt1 , vt2 and
vg corresponding to those shown in Fig. 3.7, we would notice that:
• contrary to what occurs for the external gear pairs, for the internal gear pairs
both vt1 and vt2 , during the meshing, increase linearly as a function of the
position of the point of contact P on the line of action, i.e., as a function of gP ;
108 3 Characteristic Quantities of Cylindrical Spur Gears …
• all other conditions being equal, for the internal gear pairs the absolute value of
vg is less than that occurs for the external gear pairs, because it is proportional to
the relative angular velocity x ¼ ðx1 x2 Þ:
In the previous section, we saw that the conditions of relative motion between two
spur-gear mating teeth, whatever the position of the point of contact along the path
of contact, correspond to those between two virtual equivalent cylinders having
tangential velocities equal to the respective rolling velocities vt1 and vt2 . In order to
understand better the nature of this relative motion, it is appropriate to introduce the
concepts of relative sliding and specific sliding.
Consider the condition of relative motion related to any point of contact
P located between points A and E (Fig. 2.10). In the position of contact so crys-
tallized, the two equivalent cylinders have instantaneous radii q1 and q2 (Fig. 3.6).
In the time interval dt, the two pitch circles roll describing different angles du1 ¼
x1 dt and du2 ¼ x2 dt, but equal arcs r1 du1 ¼ r2 du2 : In the same time interval dt,
the point of contact P moves, on the two conjugate profiles, describing on them two
differential involute arcs ds1 and ds2 of different lengths. The lengths of these two
arcs are equal only when the point of contact P, during the meshing, comes to
coincide with the pitch point C. Since vt1 and vt2 can also be conceived as absolute
speeds of the point P about the centers T1 and T2 , with point P thought of as
belonging to the driving profile once, and then to the driven profile, we have:
different definition, in fact coincide; for this reason, we designed them with the
same symbols.
We define as specific sliding of one profile the ratio between the relative sliding
of the profile itself with respect to the other and the arc along which, on the same
profile, the contact is moved, and therefore, for the two profiles we have
respectively:
ðds1 ds2 Þ
f1 ¼ f2 ¼ ðds2dsds 1Þ
: ð3:72Þ
ds1 2
Instead we define as slide-roll ratio of a given profile the ratio between the
sliding velocity of the profile itself with respect to the other at the point of contact
and the rolling velocity at the same point considered belonging to the same profile,
and therefore, for the two profiles we have respectively:
Taking into account Eq. (3.71), the Eqs. (3.72) and (3.73) obviously coincide.
Thus, recalling Eqs. (3.61 and 3.62), from Eq. (3.73) we obtain:
which are respectively valid for the driving and driven gear wheels. The sign
conventions to be used for the latter equations are obviously those previously
described.
Equations (3.74) and (3.75) show that the specific slidings and equivalent
slide-roll ratios are continuously variable along the path of contact. They are equal
to zero at the pitch point and reach their absolute maximum values in the extreme
points A and E of the path of contact (Fig. 2.10). These absolute maximum values
are negative (negative specific slidings) and occur during the approach path of
contact for the driving gear wheel, and during the recess path of contact for the
driven gear wheel. In general, therefore, for the usual progressive contact that
occurs in spur, helical and bevel gears, the specific slidings and slide-roll ratios are
positive for all points on the addendum flanks of the teeth, and negative for all
points on the dedendum flanks of the teeth.
Experimental evidence shows that the destructive effects of pressure and friction
are greater on the flank surfaces having negative specific slidings and slide-roll ratios.
What is more, the absolute values of these negative quantities are numerically the
greater. A high negative value of these quantities, which occurs during the dedendum
contact for both toothed wheels of the gear pair, means that a small arc of the
110 3 Characteristic Quantities of Cylindrical Spur Gears …
dedendum profile of a tooth flank meshes with a large arc of the addendum profile of
the conjugate tooth flank. Therefore, the work of the friction forces acts on the small
arc of the dedendum profile of a tooth, while operates on a great arc of the addendum
profile of the conjugate tooth. The resulting damages due to the wear are therefore
much more pronounced on the dedendum portions of the tooth flank profiles.
If the path of contact was used until the limiting interference points T1 and T2 (in
this case, as Fig. 2.10 shows, the radii of the addendum circles of the two gears
would be respectively equal to O2 T1 and O1 T2 ), we would have gP ¼ r1 sin a in
Eq. (3.74), and gP ¼ r2 sin a in Eq. (3.75). Thus, we would have f1 ¼ 1 at
point T1 , and f2 ¼ 1 at point T2 . Based on this result, the opportunity to exclude
from the contact the portions of the involute profiles near the base circles is evident.
Figure 3.7 shows also the qualitative diagrams of the quantities f1 and f2 , which
characterize the external gear pair referred to therein. The figure shows the maxi-
mum negative values of these two quantities at point A and E, as well as their
asymptotic values tending to minus infinity at both points T1 and T2 . Scientists and
industry experts attach great importance to the distribution of these quantities along
the path of contact, as they have a deep effect on the gear design quality (see
Giovannozzi [24], Henriot [27], Niemann and Winter [39]). The requirements
considered optimum for the distributions of these quantities along the path of
contact are essentially two:
• the absolute maximum values of f1 and f2 in the extreme points A and E of the
path of contact must not be too high;
• their distribution curves should not be too different for the two members of the
gear pair.
The way to achieve both of these important design goals is practically obliged: to
use gears with addendum modification in terms of profile shift, which offers to the
designer many opportunities of intervention.
Unlike external gears, for an internal gear pair both the rolling velocities vt1 and
vt2 (given respectively by Eqs. (3.59) and (3.69)) increase as gP increases.
During recess (gP positive), with the pinion driving, vt1 is greater than vt2 , and
the rolling velocity ratio is given by:
vt2 r2 sin a þ gP
¼ ; ð3:76Þ
vt1 uðr1 sin a þ gP Þ
instead, during approach (gP negative), vt1 is less than vt2 , and thus the rolling
velocity ratio is:
vt1 uðr1 sin a þ gP Þ
¼ : ð3:77Þ
vt2 r2 sin a þ gP
In the previous section, we have seen that the specific sliding is one of the main
causes of wear of the tooth flank surface. From the theoretical point of view, there
would be no changes in the tooth profile if the abrasive wear [32] of the profile itself
proceed in such a way that the ratio ðsw =rb Þ was constant. This is the ratio between
the tooth thickness worm away at a given point, sw ; measured along the normal to
the profile, and the moment arm, rb ; of the normal to the profile at the same point
with respect to the wheel axis. In this way disfigurements and changes of profile due
to wear correspond to a virtual rotation of the same profile about the fixed center of
the wheel [24].
However, it is to remember that the hypothesis of Reye [45], referring to the
wear, states that the volume of material removed in a given time interval due to the
wear is proportional to the work done by the friction forces in the same time
interval. Consequently, the so-called wear equation can be written as:
sw K
¼ pvg ; ð3:80Þ
t H
where:
• sw =t is the wear velocity, that is the quotient of the wear depth sw (in mm)
divided by the time t (in s);
• K is the dimensionless wear coefficient, which depends on the wear modes
(abrasive wear, adhesive wear, fretting, etc.), material combinations (the softer
of the two rubbing metals conditions the combination of the materials involved),
and presence or absence of lubricant;
• H (in MPa) is the surface hardness;
• p (in MPa) is the surface interface pressure;
• vg (in mm/s) is the sliding velocity.
For two rubbing surfaces, which are the flanks of two mating teeth, this equation
implies that the rate of wear of one of the two flanks, with respect to the conjugate
flank, is proportional to the wear coefficient, that of material of the tooth under
consideration in contact with the material of the mating tooth. The same rate of wear
is inversely proportional to the surface hardness of the tooth analyzed, and directly
proportional to the rate of friction work, if we assume a constant coefficient of friction.
112 3 Characteristic Quantities of Cylindrical Spur Gears …
For a given compressive force between the flank surfaces, the volume of material
worm away is independent of the area of contact. Therefore, multiplying both sides
of Eq. (3.80) for the area of the contact surface, we infer that the material depth sw
removed by wear is proportional to the specific sliding f multiplied by the normal
force ðFt = cos aÞ acting on the tooth (Ft is the nominal transverse tangential force at
reference cylinder). It should be noted that here we have used the normal force
Ft = cos a (Fig. 3.9a), which refers to the ideal case of frictionless. Due to the
non-eliminable friction, the problem becomes more complex, as we will see in the
next section.
However, for involute gearing, we have:
Ft
¼ const rb ¼ r cos a ¼ const: ð3:81Þ
cos a
Table 1.1 shows the indicative values of efficiency of the various types of gears,
used in practical applications. In this section, we want to deepen this important
topic, with special reference to cylindrical spur gears. The general analytical bases
to calculate the efficiency of gears were developed starting from the second half of
the 19th century, with insights and generalizations gradually processed in the first
half of the 20th century. In this regard, contributions specifically to be mentioned
are those of [[9], 12, 34, 44, 50]. These contributions led to the definition of
analytical relationships that, with little significant changes, coincide with the
equations now universally used, and described below (see also Ambekar [1], Ferrari
and Romiti [21], Scotto Lavina [47]).
Consider the cylindrical spur gear shown in Fig. 2.10, and provisionally assume
that the transverse contact ratio is equal to unity ðea ¼ 1Þ, whereby the path of
contact, ga , coincides with the transverse base pitch, pbt . Therefore, we have a spur
gear operating in the limiting condition of motion continuity, and only one mating
tooth pair is in meshing along the entire path of contact. Suppose also that pinion
and gear wheel are, respectively, the driving and driven wheels, and that the
directions of rotation are as indicated in the same figure. We assume finally that the
involute profiles of the teeth are perfectly shaped, and equally spaced along the base
circle (Fig. 2.2).
Let x1 , x2 , T1 and T2 , the angular velocities and nominal torques (as usual,
subscript 1 and 2 refer respectively to the driving pinion, or simply driver, and the
114 3 Characteristic Quantities of Cylindrical Spur Gears …
driven gear wheel or simply follower), and suppose that the point of contact
between the two mating profiles is placed at any position along the path of
approach, namely between points A and C in Fig. 2.10. In the ideal case of zero
friction, the force ðFt = cos aÞ that the driving profile transmits to the driven profile,
which touch each other in the above current point, is directed according to the
common normal of contact that, for the involute profiles here considered, coincides
with the line of action, passing through the pitch point, C.
Instead, in the real case, due to the friction, the line of application of the total
force F 6¼ Ft = cos a, exchanged between the mating profiles, is deviated with
respect to the straight line T1 T2 of an angle equal to the angle of friction, u. This
deviation is in direction such as to perform a negative work (and so to counter the
motion), during the relative motion of the driven wheel with respect to the driving
wheel. Therefore, the line of application of the instantaneous force, actually
exchanged between the two mating profiles, will intersect the center distance at a
point which is shifted, with respect to the pitch point, C, towards the center of the
driven wheel, O2 .
It is to be noted that, for external gears, this occurs whatever the position of the
point of contact along the path of contact. In other words, the point of contact may
be a point of the path of approach or a point of the path of recess. This behavior is
justified by the fact that, to perform a negative work, the total force F must have,
with respect to the center of instantaneous rotation of the relative motion (i.e., the
pitch point, C), a torque of opposite sign with respect to that of the relative angular
velocity, x ¼ ðx2 x1 Þ; of the driven wheel with respect to the driving wheel.
It is also to be noted that l ¼ tan u is the coefficient of sliding friction that, for
dry surfaces, satisfies approximately the Coulomb’s law of friction (see Caubet [11],
Coulomb [15]), and the two Amontons’ laws [2]. According to these laws, the
kinetic friction is independent of the sliding velocity (Coulomb’s law), and the force
of friction is directly proportional to the applied load (Amontons’ first law), and
independent of the apparent area of contact (Amontons’ second law). As we will
point out in the “Gears - Vol.3: A concise history” of this monothematic textbook,
to honor the historical reality, the two Amontons’ Laws should be attributed to
Leonardo da Vinci, who was the first to formulate them.
For internal gears, the torque of the total force, F, that the driving profile
transmits to the driven profile must always be of opposite sign with respect to that
of the relative angular velocity, x ¼ ðx2 x1 Þ of the driven wheel with respect to
the driving wheel. Therefore, this torque must have the same sign as that of ðx1 Þ,
since jx1 j [ jx2 j. Consequently, the line of application of the exchanged force, F,
must intersect the straight line O1 C, passing through the centers O1 and O2 of the
two members of the gear under consideration, at a point located at the outside of the
pitch point, C, i.e. at a point shifted towards the body of the annulus gear.
We determine now the instantaneous efficiency, gi , of the external gear pair,
corresponding to any position of the point of contact along the path of approach
(Fig. 3.9a), assuming that the pinion is the driver. In the ideal case, if Ft is the
nominal transverse tangential load, we will have T1 ¼ Ft r1 and T2 ¼ Ft r2 , whereby
the ratio between the driving torque, T1 (i.e., the nominal torque at the pinion), and
3.9 Efficiency of Cylindrical Spur Gears 115
(a) (b)
Fig. 3.9 Geometric quantities for determination of the total force exchanged between the teeth,
when the point of contact coincides with: a A point of the path of approach; b A point of the path
of recess
the useful resisting torque, T2 (i.e., the nominal torque at the wheel), is equal to the
ratio between the radii r1 and r2 of the reference pitch circles, i.e., ðT1 =T2 Þ ¼ r1 =r2 .
It should be noted that T1 and T2 are no longer the interference points, but rather the
driving and driven torques.
However, in the real case (Fig. 3.9a), we are facing an actual force
F 6¼ Ft = cos a, whose line of application intersects the center distance at point C 0 .
The component of this force applied at point C0 , in the direction normal to the
straight line O1 O2 , is ½F cosða þ uÞ 6¼ Ft . In addition, the position of point C 0
varies, depending of the position of the current point of contact P along the path of
approach. Thus, the actual instantaneous driving and driven torques will be given
respectively by: T10 ¼ ½F cosða þ uÞ ðr1 þ CC 0 Þ; and T20 ¼ ½F cosða þ uÞ
ðr2 CC 0 Þ. Thus, we get:
T10 r1 þ CC 0
0 ¼ : ð3:82Þ
T2 r2 CC 0
Easily, it can be shown that the relationship Eq. (3.83) is also valid when the
point of contact coincides with any point of the path of recess (Fig. 3.9b). The same
116 3 Characteristic Quantities of Cylindrical Spur Gears …
sin u
CC 0 ¼ CP ; ð3:84Þ
cosða uÞ
1 ðCC 0 =r1 Þ
gi ¼ ; ð3:85Þ
1 ðCC 0 =r2 Þ
where the plus sign is to be taken for external gears, and the minus sign is to be
taken for internal gears.
The instantaneous efficiency is continuously variable as the point of contact
moves along the path of contact. Therefore, it is necessary to calculate the average
efficiency during a meshing cycle, which corresponds to the time interval in which
the pitch circles roll on one another, without sliding, of an arc equal to the circular
pitch, after which the operating conditions of the gear pair under consideration
repeat itself cyclically; in this regard, we must remember that we have assumed
ea ¼ 1. To this end, let us consider an external gear pair with a driving pinion;
during the path of approach (Fig. 3.9a), the differential energy (or work) lost by
friction, dWl , in the infinitesimal time interval dt is given by:
3.9 Efficiency of Cylindrical Spur Gears 117
dWl
¼ F jx2 x1 jd; ð3:86Þ
dt
where d0 ¼ CC 0 :
Therefore, taking account of Eqs. (3.84) and (3.87), Eq. (3.86) becomes:
dWl
¼ F jx2 x1 jðCPÞ sin u: ð3:88Þ
dt
On the other hand, if F is the instantaneous force exerted by the driving profile
on the driven profile, an equal and opposite force, i.e. ðF Þ; will act on the driving
profile. Therefore, assuming a steady-state operating condition, which is charac-
terized by a constant angular velocity, the torque T1 applied by this force to the
pinion will be T1 ¼ Fb, where b is the moment arm (or lever arm), that is the
distance of the line of application of the force ðF Þ from the center O1 . From
Fig. 3.9a, we obtain:
T1 T1
F¼ ¼ : ð3:90Þ
b r1 cosða þ uÞ þ ðCPÞ sin u
dWl T1
¼ jx2 x1 j ðCPÞ sin u: ð3:91Þ
dt ½r1 cosða þ uÞ þ ðCPÞ sin u
Since the point of contact P moves along the path of contact with a constant
velocity, given by Eq. (3.58), the distance CP of the point of contact from the pitch
point is expressible as:
CP ¼ u1 r1 cos a; ð3:92Þ
where u1 is the angle of rotation of the driving gear wheel (the pinion), corre-
sponding to the displacement of point of contact from P to C.
It follows that, for contact at a current point of the path of approach, the quotient
of the differential work lost by friction divided by the differential work at the same
time done by the driving torque T1 can be expressed by:
118 3 Characteristic Quantities of Cylindrical Spur Gears …
dWl x 2 u1 cos a sin u
¼ 1
T1 x1 dt x1 ½cosða þ uÞ þ u1 cos a sin u
ð3:93Þ
r1 u1 cos a sin u
¼ 1 ;
r ½cosða þ uÞ þ u cos a sin u
2 1
with
sin u
CC 0 ¼ CP : ð3:95Þ
cosða uÞ
Dividing by cos a sin u the numerator and denominator of the fractions to the
right side of Eqs. (3.93) and (3.96), and bearing in mind that x1 dt ¼ du1 and
l ¼ tan u, we can write:
dWl r1 l u1
¼ 1 ; ð3:97Þ
T1 du1 r2 ½1 þ lðu1
tan aÞ
where, at the denominator of the last fraction, we must take the minus sign or plus
sign, depending on whether the point of contact is within the path of approach or
within the path of recess.
Assuming then that the driving torque T1 is a constant, and remembering that
ea ¼ 1; we infer that the work dissipated by friction, in the time corresponding to
duration of contact between the two teeth, i.e. during the entire meshing cycle, is the
summation of all differential works lost by friction during the paths of approach and
recess. Thus, it is given by the following equation:
(e =r )
r ðl=l0 Þu1 du1 e2Z=r1 ðl=l0 Þu1 du1
Wl ¼ T1 1 l0
1Z 1
1
þ
r2 0
½ 1 þ lðu1 tan aÞ 0
½1 þ lðu1 þ tan aÞ
r1
¼ T1 1 l0 D;
r2
ð3:98Þ
3.9 Efficiency of Cylindrical Spur Gears 119
where e1 and e2 are respectively the arcs of approach and recess, given by
Eqs. (3.20) and (3.21), D is a parameter that briefly indicates all that the bracket
contains, and l0 is the value of coefficient of friction l corresponding to a given
point of path of contact, which can be determined conventionally.
Taking into account that, during the same meshing cycle considered for calcu-
lation of Wl , the work done by the driving torque, Wm , is given by:
e1 þ e2
Wm ¼ T1 ; ð3:99Þ
r1
we infer that the loss of efficiency of the cylindrical spur gear pair examined here
can be expressed as:
Wl r1 r1
1g¼ ¼ 1 l0 D: ð3:100Þ
W m e1 þ e2 r2
D⋇ will be of the order of magnitude of unity. Finally, since ðr1 =r2 Þ ¼ ðz1 =z2 Þ,
Eq. (3.100), can be written in the form:
p z1 1
1
1 g ¼ 1 l0 D ¼ 2p l0 D⋇ :
⋇
ð3:102Þ
r1 z2 z1 z2
If we assume that the coefficient of the friction, l, varies little during the
meshing cycle, for which it can be considered as a constant, the calculation of D⋇
becomes very simple. If then we neglect ½lðu1 tan aÞ with respect to unity (see
Eq. 3.98), we obtain:
2
⋇ r1 r 2 e2 þ e2 1
D ¼D ¼ 12 1 2 2 ¼ e21 þ e22 ; ð3:104Þ
p p 2r1 2
120 3 Characteristic Quantities of Cylindrical Spur Gears …
where e1 ¼ ðe1 =pÞ ¼ ðga2 =pbt Þ and e2 ¼ ðe2 =pÞ ¼ ðga1 =pbt Þ are respectively the
addendum contact ratios of the wheel and pinion. Therefore, Eq. (3.103) becomes:
1 1
g ¼ 1 pl0 e21 þ e22 : ð3:105Þ
z1 z2
This equation is known as the formula of Poncelet (see Ferrari and Romiti [21],
Poncelet [41]). Then keeping in mind that the relationship that correlates circular
pitch, p, radius of the pitch circle, r, and number of teeth z (see Eq. 2.38), it can be
expressed also in the form:
l0 1 1 2
g¼1 e1 þ e22 : ð3:106Þ
2p r1 r2
Equations (3.105) and (3.106) are also valid for pinion-rack gear pairs; to do
this, just put in them z2 ¼ 1 and, respectively, r2 ¼ 1. It should be noted that the
same Eqs. (3.105) and (3.106) do not show a difference in efficiency between a spur
gear pair with driving pinion and driven wheel, and the same gear pair with driving
wheel and driven pinion. However, an in-depth analysis of this problem, which is
left to the reader as an exercise, would highlight the fact that friction power losses in
a given cylindrical spur gear pair are different depending on whether the driving
member is the pinion or the gear wheel. In fact, for spur gear pairs whose members
have different diameters, we would find that efficiency is greater when the smaller
member is the driving wheel. This behavior justifies the fact that, for high trans-
mission ratios, spur gear pairs are used mainly as speed reducing gears rather than
as speed increasing gears.
It is to be noted that, in the conditions above described, the quantity D, i.e. the
sum of the two integral functions appearing into the bracket of Eq. (3.98), is that
shown in Fig. 3.10. On abscissa of this figure, the arc of approach, e1 , is drawn on
the left of the pitch point,
C, while
2 the arc
of recess, e2 , is drawn on the right. Since
then D is proportional to ð1=2Þ e1 þ e2 , proportionality factor being the constant
2
ratio ðl=l0 Þ, it is obvious that from the end points, A⋇ and E⋇ , two vertical
segments must to be drawn, respectively equal to e1 and e2 . Therefore, the two
straight lines CA0 and CE 0 represent the above two integral functions. The sum of
the two integrals is then proportional to the sum of the areas of two triangles CA⋇ A0
and CE ⋇ E 0 , and this sum is proportional to the work lost by friction, coefficient of
proportionality being the ratio ðl=l0 Þ. The upper boundaries of these two triangular
areas are highlighted with dashed lines in Fig. 3.10.
It should be noted that the end-points A⋇ and E ⋇ that delimit the arc of action
measured over any of the two pitch circles correspond respectively to the end points
A and E of the path of contact, shown in Figs. 1.8 and 2.10. Also, based on the
correlation between quantities measured on pitch circles and quantities measured on
the line of action (or in equivalent way, on the base circles), any segment on the
3.9 Efficiency of Cylindrical Spur Gears 121
abscissa in Fig. 3.10 is equal to the quotient of the corresponding segment on the
path of contact divided by cos a:
However, usually the arc of action ðe1 þ e2 Þ is greater than the circular pitch and,
in most practical applications, it is included in the range p\ðe1 þ e2 Þ\2p. In this
case, the calculation of the work dissipated by friction in the infinitesimal time
interval dt (and therefore the one dissipated during an angle of rotation corre-
sponding to the circular pitch) requires that the forces exchanged between the
various mating tooth profiles are known.
The problem of determination of these forces is very complex (see Vol. 2,
Sect. 1.7), because the sharing of the total load between the various tooth pairs in
simultaneous meshing depends on several types of deformability, i.e.: deformability
of the teeth; deformability of the gear wheel body; deformability of the shafts;
deformability of the bearings; deformability of the gearbox; etc. It also depends on
the local deformations due to the Hertzian pressure of contact, and the accuracy
grade of the tooth profiles. In addition, it depends on the friction type, which is not a
friction between dry surfaces, as hitherto assumed implicitly, but friction between
lubricated surfaces.
For an approximate calculation, we assume here that the total load is distributed
in equal parts between the two tooth pairs in simultaneous meshing. With this
simplifying assumption, the work lost by friction will be half of what it would in the
case of a single tooth pair in meshing. Figure 3.11 shows what happens in this case:
on abscissa, in addition to the pitch point C, the four marked points A⋇ , B⋇ , D⋇ , and
E ⋇ are reported. The points B⋇ and D⋇ are defined univocally, since A⋇ D⋇ ¼ p and
B⋇ E⋇ ¼ p. In accordance with the afore-mentioned simplifying assumption, the
points A00 and E 00 are defined by the relationships A⋇ A00 ¼ A⋇ A0 =2 and
E ⋇ E 00 ¼ E⋇ E0 =2:
In the region of single contact, between marked points B⋇ and D⋇ , only one
tooth pair is in meshing, for which all the load is transmitted by means of this tooth
pair. Therefore, the work lost by friction is proportional (as usual, the propor-
tionality factor is given by the ratio l=l0 ) to the sum of areas of the two triangles
B⋇ B0 C and D⋇ D0 C (Fig. 3.11). In the regions of double contact, between the
122 3 Characteristic Quantities of Cylindrical Spur Gears …
Fig. 3.11 Distribution curve of D, for ½p\ðe1 þ e2 Þ\2p ; unless the proportionality constant
ðl=l0 Þ
marked points A⋇ and B⋇ , and D⋇ and E ⋇ , two tooth pairs are simultaneously in
meshing, whereby each of them bears half of the transmitted load. Therefore, the
work lost by friction in these regions is equal to the sum of the left trapezoid area
A⋇ A00 B⋇ B00 , which is half of the left trapezoid area A⋇ A0 B⋇ B0 , and right trapezoid
area D⋇ D00 E⋇ E00 , which is half of the right trapezoid area D⋇ D0 E ⋇ E 0 .
The sum RA of these areas is therefore given by:
1h 2 i
RA ¼ e1 þ e22 þ ðp e1 Þ2 þ ðp e2 Þ2
4 ð3:107Þ
1
¼ e21 þ e22 þ p2 pðe1 þ e2 Þ ;
2
p2
2
RA ¼ e1 þ e22 þ 1 ðe1 þ e2 Þ : ð3:108Þ
2
For high enough values of e1 and e2 that, in the usual practical applications, are
almost next to unity, the sum e21 þ e22 is not very different from the sum ðe1 þ e2 Þ.
Consequently, the value of the square bracket that appears in Eq. (3.110) is almost
unitary; thus, under these conditions, Eq. (3.110) is simplified, and becomes:
1 1
g ¼ 1 pl0 : ð3:111Þ
z1 z2
From this last equation, we can deduce that the average efficiency of a cylin-
drical spur gear varies as follows: (i), it improves with increasing the number of
teeth, z1 and z2 ; (ii), with the same number of teeth, it is higher for the internal gear
pairs. In addition, both from Eqs. (3.105) and (3.110), and from Figs. 3.10 and
3.11, we can deduce that the average efficiency of these gears is much higher, as
much as the sum ðe1 þ e2 Þ is low, i.e. as much as the transverse contact ratio is low;
in fact, the values of e1 and e2 are generally smaller than the unity. Low values of
the sum ðe1 þ e2 Þ mean equivalent low values of sum ðe1 þ e2 Þ, i.e. smaller areas
under the diagrams shown in Figs. 3.10 and 3.11.
For ðe1 þ e2 Þ [ 2p, three tooth pairs are simultaneously in meshing. To obtain
the distribution curve of the work lost by friction along the path of contact, we can
use the same procedure described for the case where ðe1 þ e2 Þ ¼ p, and
p\ðe1 þ e2 Þ\2p. To this end, as Fig. 3.12 shows, first we will identify the six
marked points that characterize this special case. In fact, in addition to the usual end
points of path of contact, A⋇ and E⋇ , we will have the points B⋇ and F ⋇ , defined by
E ⋇ B⋇ ¼ B⋇ F ⋇ ¼ p, and the points D⋇ and G⋇ , defined by A⋇ D⋇ ¼ D⋇ G⋇ ¼ p. It is
quite evident that, in this case, there is no region of single contact; instead, three
Fig. 3.12 Distribution curve of D, for ðe1 þ e2 Þ [ 2p; unless the proportionality constant ðl=l0 Þ
124 3 Characteristic Quantities of Cylindrical Spur Gears …
regions of triple contact exist (those comprised between the marked points A⋇ and
F ⋇ ; D⋇ and B⋇ ; G⋇ and E⋇ ), and two regions of double contact (those comprised
between the marked points F ⋇ and D⋇ ; B⋇ and G⋇ ).
Here too, for an approximate calculation, we assume that the total load is dis-
tributed in equal parts between the three or two tooth pairs in simultaneous
meshing. With this simplifying assumption, in the triple and double contact regions
the work lost by friction will be respectively equal to (1/3) and (1/2) of what would
occur in a hypothetical region of single contact. Thus, with reference to Fig. 3.12,
we will have: A⋇ A000 ¼ A⋇ A0 =3; F ⋇ F 000 ¼ F ⋇ F 0 =3; F ⋇ F 00 ¼ F ⋇ F 0 =2; D⋇ D000 ¼
D⋇ D0 =3; D⋇ D00 ¼ D⋇ D0 =2; B⋇ B000 ¼ B⋇ B0 =3; B⋇ B00 ¼ B⋇ B0 =2; G⋇ G000 ¼ G⋇ G0 =3;
G⋇ G00 ¼ G⋇ G0 =2; and E ⋇ E 000 ¼ E ⋇ E 0 =3: The area between the abscissa axis and
dashed segments shown in Fig. 3.12 is proportional to the work lost by friction, the
proportionality factor being the ratio ðl=l0 Þ. This area can be expressed as a
function of e1 , e2 , and p; for which the use of the previous procedure would lead to
write the average efficiency in explicit form. This exercise is left to the reader.
Finally, it is to be noted that the average efficiency hitherto considered is that
related only to the work lost by friction due to contact conditions between the
various pairs of mating teeth that do not include the micro-slip effects. To determine
the total contact efficiency, gt , we would need to consider the losses due to the
rolling resistance caused by the micro-slip phenomena we mentioned in Sect. 3.6.
Here, however, we will continue to address the topic in terms of
macro-phenomenon, rather than micro-phenomenon, using the same procedure that
allowed us to deride Eqs. (3.82) and (3.83). As we have already pointed out, this
procedure reduces the micro-slip effects to resisting friction forces, which instead of
passing through the centers of the gear wheels are tangent to the friction circles, and
act in the opposite direction to the motion. Anyway, to define the total loss with a
sufficient approximation, it is convenient first to express, in a different way, the
average efficiency.
In this regard, we should return to consider Fig. 3.9a, assuming that the force
F is passing through a particular point C 0 , such that CC 0 ¼ d0m , where d0m indicates
the particular distance between C and C 0 , which is representative of the average
conditions, and therefore no longer variable instant by instant, but constant. Thus,
using Eq. (3.83) which gives the instantaneous efficiency, and writing it for the
afore-mentioned average conditions, i.e. for CC 0 ¼ d0 ¼ d0m , and then for gi ¼ gm ,
we get the following equation in terms of average efficiency, gm :
1 d0m =r2
gm ¼ : ð3:112Þ
1 þ d0m =r1
3.9 Efficiency of Cylindrical Spur Gears 125
1 gm
d0m ¼ : ð3:113Þ
ðgm =r1 Þ þ ð1=r2 Þ
Under ideal operating conditions, the force Ft = cos a passes through the pitch
point, C (Fig. 3.9a), and the passive forces, i.e. the reaction forces exerted by the
gearbox on the shaft pins, pass through the centers O1 and O2 . Instead, under actual
operating conditions, the force F, which represents the average conditions, passes
through the particular point C 0 , defined by CC 0 ¼ d0 ¼ d0m . The reaction forces
exerted by the gearbox on the shaft pins are parallel to the force F, but no longer
pass through the centers O1 and O2 . They are tangent to the circles of friction, the
radii of which are respectively q1 and q2 (the symbols introduced here to indicate
these radii are the same used for the radii of curvature of the involute profiles, but
the letters of the Greek alphabet are only twenty-four!). In addition, the points of
tangency of these reaction forces with circles of friction are directed in such a way
that these passive forces are opposite to the relative motions of the individual gear
wheels with respect to the gearbox.
Based on these considerations, we can calculate the total contact efficiency that
takes into account the losses due to contact between the mating teeth, including the
rolling resistance due to the micro-slip effects. The active and passive forces
described above can be considered as average forces having a constant value
throughout the meshing cycle, and equivalent to the actual ones, which are variable
instant by instant. Therefore, using the same procedure that allowed us to derive the
relationship (3.83), we get:
r2 d0m q2 = cos a =r2 1 d0m =r2 ðq2 =r2 Þ
gt ¼
ffi : ð3:114Þ
r1 þ d0m þ q1 = cos a =r1 1 þ d0m =r1 ðq1 =r1 Þ
The approximation in this relationship is due to the fact that, for the usual values
of the pressure angle a, we can take cos a ffi 1:
Since the four ratios that appear in the above relationship are very small, we can
easily demonstrate that it can be written in the form (see Scotto Lavina [47]):
1 d0m =r2 q q q q
gt ffi 0
1 2 ¼ gm 1 2 : ð3:115Þ
1 þ dm =r1 r 1 r 2 r 1 r2
From this last relationship we deduce that the total contact efficiency, gt , differs
little from the average efficiency, as q1 and q2 are small amounts with respect to r1
and r2 respectively.
126 3 Characteristic Quantities of Cylindrical Spur Gears …
If the pinion is the driving member of the gear pair under consideration, in the ideal
case where the friction losses are equal to zero, the nominal driving torque at the
pinion, T1 ¼ Ft r1 , and the nominal driven torque (or resisting torque) at the wheel,
T2 ¼ Ft r2 , must be the same, so we have:
r1 x2 T2
T1 ¼ T2 ¼ T2 ¼ ; ð3:116Þ
r2 x1 jij
where jij is the absolute value of the transmission ratio, given by Eq. (1.13).
Instead, in the actual case, which is characterized by the total contact efficiency
gt , the effective driving torque T1e will be given by:
T1 T2
T1e ¼ ¼ : ð3:117Þ
gt jijgt
This relationship allows to calculate the effective driving torque (and thus the
effective driving power) to be applied to the driving shaft to win an assigned
resisting torque, with a given transmission ratio, when the allowable value of total
contact efficiency can be provided with sufficient precision. In other cases, where
the usable driving power and the resisting torque to be won are known, the same
relationship allows to obtain the required transmission ratio, by a trial procedure.
We have already pointed out that appropriate design of any type of gear must take
into account the cutting technologies used in its production process. Therefore, we
believe that it is necessary here to describe briefly the main methods employed for
the manufacturing of cylindrical spur gears. These methods can be summarized in
the two following categories: form cutting methods, and generation cutting meth-
ods. Anyway, for the insights that may be need in this regard, we refer the reader to
known traditional textbooks (see Björke [6], Boothroyd and Knight [7], Galassini
[23], Henriot [28], Jelaska [30], Micheletti [37], Radzevich [42], Rossi [46]).
This is the easiest cutting method, which is done with a milling machine, using a
form-milling cutter, shaped exactly like the tooth space of the gear wheel to be
manufactured. Therefore, the cutting edges of this form-milling cutter are shaped so
3.11 Short Notes on Cutting Methods of Cylindrical Spur Gears 127
that all of its sections, made with a plane passing through its axis, reproduce the
tooth space between tooth and tooth of the gear wheel to be cut. In this way, it is
possible to regrind the cutting edges according to a radial plane without any
alteration of the shape of the same cutting edges.
To meet the above conditions, and to have a convenient relief angle, c, the cutter
edges of the form-milling cutter are generated by the motion of the cutter profile
according to an Archimedes spiral, using special lathes. This motion ensures that
the relief angle is almost exactly constant (it is to be noted that, to have the
theoretical exact geometric shape, this motion should follow a logarithmic spiral).
A section of the cutter’s tooth with a plane perpendicular to the rake face, which
delimit it in front (section A A in Fig. 3.13), has a trapezoidal shape, and thus it is
characterized by a side rake angle, c0 , which is defined by means of the following
considerations. If a is the angle between the tangent at any point of the involute
cutting edge of the cutter and the centerline of the form-milling cutter tooth, a
differential radial shift, dr, of the trace of section A A corresponds to a decrease in
half-thickness equal to ðdr tan aÞ. Because of the relief angle of the cutting edge, c,
a differential shift, dr, corresponds to a differential displacement, dx, in the direction
of the cutting-edge section, such that dr ¼ ðdx tan cÞ, so we will have:
From this relationship, we infer that the side rake angle, c0 , depends on tan a.
Thus, at points where angle a is very small (this angle is equal to zero on the base
circle of teeth with involute profiles), c0 is also very small, so over the base circle
and near it unfavorable conditions to an easy and precise cutting occur. However,
for teeth with involute profiles, this disadvantage has relatively modest effects
because, as we have already pointed out, the portion of profile near the base circle is
not normally used. It should be noted, however, that this disadvantage would be
more serious for teeth with cycloid profile, for which the portion of profile near the
pitch circle, where a is equal to zero, is generally the most important. However, this
topic does not interest here, since in this textbook only teeth with involute profiles
are considered.
With this cutting technology, the gear blank is held stationary, while the milling
cutter is fed slowly in the direction of the axis of the same gear blank, as long as a
tooth space is cut throughout its depth. The gear blank is then indexed to an angle
exactly corresponding to one angular pitch, and a second tooth space is cut, and so
on, i.e. the process is repeated until all tooth spaces are cut and the gear wheel is
completed. In order to reduce the wear of the cutting edges, and thus limit the
number of re-grindings, the finishing milling cutter having the appropriate geo-
metric profile is coupled often to a roughing milling cutter (see Fig. 3.14). This
roughing milling cutter has the function of cut the tooth spaces in a broadly
approximate way, leaving to the finishing milling cutter the task of cutting the
theoretically exact profile.
128 3 Characteristic Quantities of Cylindrical Spur Gears …
This cutting method of cylindrical spur gears has the great advantage of allowing
the use of an ordinary milling machine, which is always available in small machine
shops, together with the required angular indexing mechanism. However, normally,
special milling machines with automatic indexing devices are used. Another
advantage of this cutting method is that milling cutters are generally cheaper than
the cutting tools used for other cutting methods, such as the generation cutting
methods.
Faced with the above advantages, the form cutting method has many disad-
vantages. It should first be noted that, from the theoretical point of view, a different
milling cutter is required for each combination of values of the module, m, pressure
angle, a, number of teeth, z, and tooth thickness, s. In fact, it is well known that the
shape of tooth space depends on all four of the above parameters. Further, this
disadvantage is increased in the case of profile-shifted toothings, since the shape of
the tooth space will also depend on the value of the profile shift coefficient (see
Chap. 6). This disadvantage is particularly relevant if a small number of identical
gear wheels are to be made, while it tends to diminish as the number of gear wheels
to be cut increases, since it is then inexpensive to buy a particular cutter to perform
the job.
Another disadvantage of this cutting method is when different gear wheels are to
be made, using cutters that are already in stock. In order to minimize the number of
cutters in stock, keeping it within acceptable limits, it is usually assumed that the
cutters are used to cut gear wheels operating at the standard center distance.
Therefore, the cutters are designed in such a way that the tooth thickness of each
3.11 Short Notes on Cutting Methods of Cylindrical Spur Gears 129
Fig. 3.14 Finishing milling cutter coupled with roughing milling cutter
gear wheel is equal to half the circular pitch, but with adequate tolerance, com-
patible with the necessary backlash. In addition, since the tooth space of a gear
wheel with z1 teeth is similar in shape to that of a gear wheel with z2 teeth, provided
z1 and z2 are sufficiently close, each cutter is usually used for a defined range of z-
values. Thus, a set of eight cutters (or fifteen cutters, for very large modules) is
generally considered to be adequate for each combination of values of m and a. It
follows that the cutter shape is only correct for a particular z-value, while for the
other z-values within the range of possible utilization, inevitable errors in the gear
tooth profile are to be expected (see Galassini [23], Micheletti [37], Rossi [46]).
Other disadvantages of this cutting method of cylindrical spur gears should not
go under silence. In fact, this cutting method is a very rigid method in that it allows
130 3 Characteristic Quantities of Cylindrical Spur Gears …
Due to its many practical applications, the generation cutting method is the cutting
method of cylindrical spur and helical gears, which has the highest overall appli-
cation and allows for maximum accuracy. This method is now used to cut most of
these gears having teeth with involute profiles. The tooth generation process is
carried out by shaping, with a rack-type cutter or a pinion-type cutter, or by hob-
bing, with a hob. During this cutting process, the cutter of whatever type it is, and
the gear blank are equipped with a relative motion that simulates their engagement,
as if they were a meshing gear pair.
Some gear shaping machines use, as a cutting tool, a rack-type cutter whose
geometry is defined with reference to the standard basic rack, shown in Fig. 2.14.
This rack-type cutter is provided with a reciprocating cutting motion parallel to the
axis of the gear wheel to be generated. At the same time, the gear wheel is provided
with a motion that is composed of a translation in the direction of the pitch line of
the rack-type cutter, and a rotation about its axis. This compound motion corre-
sponds to the rolling without sliding of the gear wheel’s pitch circle on the
rack-cutter’s pitch line. Therefore, this motion simulates the relative engagement
motion of the gear wheel with the rack-type cutter. In other types of gear shaping
machines that use this technology, the rack-type cutter is provided not only with the
reciprocating cutting motion, but also with the translation motion in the direction of
its pitch line, while the gear wheel to be generated only rotates about its axis.
Anyway, the relative motion between the rack-type cutter and gear wheel is that of
a pure rolling between the pitch surfaces involved.
With this cutting technology, it is possible to vary the radius of the cutting pitch
circle of the gear wheel to be generated. To this end, it is sufficient to vary the ratio
between the translation motion in the direction of the pitch line of the rack-type
cutter, regardless of whether it is given to the rack-type cutter or to the gear wheel,
and the rotation of the gear wheel about its axis. This allows cutting with great easy
and economy, and only with one cutter of defined sizing, gear wheels with
profile-shifted toothings, both without variation of center distance, and with
3.11 Short Notes on Cutting Methods of Cylindrical Spur Gears 131
variation of center distance. This technology has in fact a unique limitation: that of
not allowing the cutting of internal gear wheels.
However, it should be noted that this cutting method has only one disadvantage:
that of not being able to get the gear wheel with a continuous generating motion,
because to do this the rack-type cutter should have the same number of teeth as the
gear wheel to be generated. It is not possible to meet this condition, especially when
the number of teeth of the gear wheel to be generated is very large, because in this
case we would have rack-type cutters of such dimensions as to be incompatible with
the usual generating shaping machines. Most rack-type cutters are made with a
small number of teeth (generally, six or eight teeth). Therefore, in order to be able to
cut the total number of teeth of any gear wheel, the cutting process is stopped each
time the gear wheel is rotated one or more angular pitches, after which the rack-type
cutter is moved back the same number of angular pitches. This way of working of
the gear-cutting machine not only results in an elongation of cutting time, but also
results in tooth profile errors and pitch deviations.
Other gear cutting machines use, as a cutting tool, a pinion-type cutter having the
shape shown in Fig. 3.15. This pinion-type cutter, also called Fellows gear shaper,
in honor of its inventor, American entrepreneur Edwin R. Fellows. He introduced it
in 1896, together with Fellows shaping machines, which gave a vital contribution to
the mass production of effective and reliable gear transmissions for the nascent
automotive industry. The Fellows gear shaper has a number of teeth a little less than
the minimum number of teeth corresponding to the non-interference with the rack
gear, so as to decrease the portion of tooth profile lost due to interference, when a
gear wheel with a small number of teeth must be cut. For example, for a pressure
angle of 15°, the pinion-type cutter has 24 teeth instead of 30 teeth, as the diagram
shown in Fig. 3.1 indicates.
The pinion-type cutter is provided with a reciprocating cutting motion parallel to
the axis of the gear wheel to be generated, and a rotation about its axis. At the same
time, the gear blank is provided with a rotation about its axis. Pinion-type cutter and
gear blank are driven at constant angular velocities, as if they were a meshing gear
pair. In other words, the two rotations above simulate the pure rolling without
sliding of the cutting pitch cylinder of pinion-type cutter over the cutting pitch
cylinder of the gear wheel to be generated, and vice versa.
After the cut has been completed, the tooth shape of the gear wheel is the result
of envelope of the successive positions of the pinion-type cutter. It can be easily
shown that this shape is not exactly that of an involute curve, but consists of a series
of arcs, whose sizes depend on the number of strokes needed to cut each tooth.
However, the above-mentioned generating motion begins only when the two cut-
ting pitch surfaces come to mutual contact. This results from the fact that the
amount of material removed with each cutting stroke is limited, so it is not possible
to cut the first tooth of the gear blank to its full depth. Therefore, before the
generation cutting process starts, the gear blank is fed radially towards the
pinion-type cutter, until the chip removal allows the cutting pitch surfaces to come
in mutual tangency. At this point, the radial feed of the gear blank is stopped, and
the generation cutting process starts. In some cases, a single cut is sufficient. Indeed,
132 3 Characteristic Quantities of Cylindrical Spur Gears …
in other cases, one or more roughing cuts are necessary before the finishing cut. In
other case, in addition to finishing cut, other more precise machining may be
required (e.g., shaving or grinding).
This cutting method of cylindrical spur gears can be used indifferently both for
external gears and for internal gears. For a pinion-type cutter having a standard
sizing (addendum ha0 ¼ ð5=4Þm, and dedendum hf 0 ¼ m, corresponding respec-
tively to the dedendum, hf , and addendum, ha , of the gear wheel to be generated),
the extension of the cutting tooth inside the base circle is done in the radial
direction. With this radial shape of the cutting edge, cutting interference occurs
when gear wheels with over 48 teeth have to be cut. Consequently, in this case, a
portion of the tooth profile close to the tip circle is removed. This removal can be
deliberately intended to obtain teeth with adequate tip relief, allowing us to achieve
interesting design goals. Instead, for usual gear wheels, which have fewer teeth than
48, the cutting interferences are much smaller than those that would be, ceteris
paribus, in the case of cutting with rack-type cutter.
A single pinion-type cutter is required for a given combination of values of the
module, m, and pressure angle, a. It can be used to cut gear wheels with any number
of teeth, as well as profile-shifted toothings both without variation of center dis-
tance, and with variation of center distance. Of course, the pinion-type cutter is a bit
different from an ordinary pinion, not only for the above-highlighted sizing, but also
3.11 Short Notes on Cutting Methods of Cylindrical Spur Gears 133
because its teeth are not parallel to the cutter axis. In fact, they are relieved, in order
to provide clearance behind the cutting edges. Due to this relief, the shape of the
cutter teeth will change slightly after each sharpening.
Despite these variations in the cutting tooth shape, geometrically exact tooth
profiles can still be cut in the gear blank. This is due to the particular configuration
of the teeth of the pinion-type cutter (Fig. 3.15), which are designed in such a way
that the projection of the cutting edges onto a plane perpendicular to the cutter axis
is always an involute curve of the same base circle, even after each sharpening of
the cutting tool. As a result, teeth on the gear blank will always be cut with the same
base pitch, although after each sharpening the tooth thickness will be slightly
reduced.
The most commonly used generation method for cutting cylindrical spur and
helical gears is by hobbing. To this end, gear hobbing machines are used, while the
cutting tool is a hob whose geometry is shown in Fig. 3.16. As this figure high-
lights, a hob is simply a screw with one or two threads. Each thread is cut by a
number of gashes forming the cutting faces, which are similar to those of the
form-milling cutter shown in Fig. 3.13. These cutting faces can either be at right
angles to the threads or parallel to the hob axis. In any case, the section of a hob
with a plane passing through its axis is a rack with teeth having straight profiles. We
can imagine that the hob is generated by this profile when it is equipped with a
screw motion, according to a mean helix having lead angle, c0 , and belonging to the
pitch cylinder (see Collins et al. [14]).
The shape of a hob can be specified by the lead angle,c0 , or the helix angle, b0 .
Usually, the lead angle is used. For a right-hand hob, the lead angle is defined as the
complement of helix angle, i.e. c0 ¼ ð90 b0 Þ. Generally, the lead angle is always
small, especially when the hob has only one thread. For a left-hand hob, whose
helix angle is negative, the lead angle is defined as c0 ¼ ð90 þ b0 Þ, and this in
order to have a negative lead angle. However, usually, to define the geometry of a
hob, the absolute value of the lead angle is given, with the addition of the speci-
fication that the hob is a right-hand or left-hand hob.
The cutting action of a hob is very similar to that of a rack-type cutter. The hob
and the gear blank rotate about their axes with angular velocities x0 and x2 meeting
the condition:
x0 z2
¼ ; ð3:119Þ
x2 z0
where z2 is the number of teeth of the gear wheel to be generated, and z0 is the
number of threads of the hob (usually, z0 ¼ 1).
The meshing of the hob with the gear wheel being generated can be considered
as a rack-gear wheel meshing, because the rotation of the hob simulates the
translation of an imaginary rack-type cutter. This is evident by observing that, due
to its screw shape, when the hob rotates about its axis, the threads appear to move in
the direction of the same axis. In addition to rotating about its axis, the hob moves
parallel to the gear wheel axis (Fig. 3.17); this is the feed motion of the hob.
However, despite the similarity between the cutting action of a hob and that of a
rack-type cutter, the hobbing process has at least the following two advantages: (i),
the motions of the cutting faces are continuous in both tangential and axial direc-
tions, so there is no need to move the cutting tool back after one or two teeth have
been cut, as in the shaping process with a rack-type cutter; (ii), the cutting action is
obtained without any reciprocating motion of the cutting tool. As Fig. 3.17 shows,
the hob begins the cutting action at one end of the gear wheel, and deepens it
progressively over the entire face width, seamlessly. This results in a great finishing
of tooth flanks, and the practical elimination of pitch deviations.
The similarity between hob and rack-type cutter goes beyond the above con-
siderations. In fact, since the cutting faces of the hob simulate the teeth of the
imaginary rack-type cutter, each of them meet the gear blank at the same angle as a
rack-type cutter. Therefore, to cut the teeth of a cylindrical spur gear wheel, a
right-hand hob must be inclined at a small angle with respect to a plane perpen-
dicular to the axis of the gear wheel to be generated. This angle is called swivel
angle, and is equal to the lead angle of the mean helix, c0 (Fig. 3.18). In more
general terms, the hob axis must be inclined with respect to the axis of the gear
wheel to be generated by an angle ð90
c0 Þ, where the minus and plus signs are
respectively valid for right-hand hob and left-hand hob.
It should be noted that the hob would be completely the same as the rack-type
cutter (and hence it would cut exact involute profiles) only if it had an infinite pitch
radius, i.e. r0 ¼ 1: In reality this does not happen. Therefore, the thread being not
straight, but helicoidal, the profiles cut with this cutting process are approximate
involutes. The approximation is bigger, the smaller the tooth depth compared to r0
(i.e., the smaller the hob axial pitch, as this is proportional to the tooth depth,
depending on the sizing adopted). Since the hob axial pitch is given by:
it follows that the approximation is bigger, the smaller the lead angle, c0 .
In practice, c0 does not exceed ð5 6Þ , and deviations from the theoretical
involute profiles are very small, so much so that this cutting process is used for
gears of high accuracy grade. The gear wheel cut with this cutting method has
circular pitch, p, equal to the normal pitch, pn0 , of the hob, that is:
Fig. 3.18 Working position of a right-hand hob with respect to the gear blank
136 3 Characteristic Quantities of Cylindrical Spur Gears …
In addition, the same gear wheel has a pressure angle, a, equal to the pressure
angle, an0 , in the normal section of the hob, that is:
where aa0 is the pressure angle of the hob in an axial section, i.e. a section of the
hob with a plane passing through its axis.
Also, for this cutting process, a single hob is required for a given combination of
values of the module, m, and pressure angle, a: It can be used to cut gear wheels
with any number of teeth, as well as profile-shifted toothings both without variation
of center distance, and with variation of center distance.
In the chapters that follow, when we consider it necessary, further details on the
three generation cutting processes described above will be given. However, for the
more detailed analysis of the hobbing cutting process as well as of the two
afore-mentioned shaping cutting processes, we refer the reader to more specialized
textbooks (see Jain and Chitale [29], Rajput 43], Sharma [48]).
References
1. Ambekar AG (2007) Mechanism and machine theory. Prentice-Hall of India, New Delhi
2. Amontons G (1699) De la resistance causée dans les machines, tant par les frottements des
parties qui les composent, que par roideur des cordes qu’on y employe, & la maniere de
calculer l’un & l’autre, Mémoires de l’Académie Royale des Sciences. Histoire de l’Académie
Royale des Sciences, Paris, pp 206–222
3. Ball RS (1876) The theory of screws: a study in the dynamics of a rigid body. Hodges, Foster,
and Co., Grafton-street, Dublin
4. Ball RS (1900) A treatise on the theory of screws. Cambridge University Press, Cambridge
5. Belfiore NP, Di Benedetto A, Pennestrì E (2005) Fondamenti di meccanica applicata alle
macchine. Casa Editrice Ambrosiana, Milano
6. Björke Ö (1995) Manufacturing systems theory. Tapir Publishers, Trondheim
7. Boothroyd G, Knight WA (1989) Fundamentals of machining and machine tools, 2nd edn.
Marcell Decker, New York
8. Bottema O, Roth B (1979) Theoretical kinematics. North-Holland, Amsterdam
9. Buckingham E (1949) Analytical mechanics of gears. McGraw-Hill Book Company Inc,
New York
10. Cattaneo C (1938) Sul contatto di due corpi elastici: distribuzione locale degli sforzi.
Rendiconti dell’Accademia Nazionale dei Lincei, 27, 242–248, 434–436, and 474–478
11. Caubet JJ (1964) Théorie et pratique industrielle du frottement. Dunod, Paris
12. Clapp WH, Clark DS (1942) Engineering materials and processes. International Textbook
Company, Scranton, PA
13. Colbourne JR (1987) The geometry of involute gears. Springer, New York Inc, New York
14. Collins JA, Busby HR, Staab GH (2009) Mechanical design of machine elements and
machines: a failure prevention perspective, 2nd edn. Wiley, New York
15. Coulomb CA (1785) Théorie des machines simples en ayant égard au frottement de leurs
parties. Acad Roy Sci Memo Math Phys X:16
References 137
16. Di Benedetto A, Pennestrì E (1993) Introduzione alla cinematica dei meccanismi, vol 1. Casa
Editrice Ambrosiana, Milano
17. Disteli M (1914) Über des analogen der Savary schen formel und konstruktion in der
kinematischen geometrie des raumes. Zeitschrift für Mathematic und Physik 62:261–309
18. Dooner DB (2012) Kinematic geometry of gearing, 2nd edn. Wiley, Chichester, UK
19. Dooner DB, Seireg AA (1995) The kinematic geometry of gearing: a concurrent engineering
approach. Wiley, New York
20. Dowson D, Higginson GR (1966) Elasto-hydrodynamic lubrication. Pergamon Press, Oxford
21. Ferrari C, Romiti A (1966) Meccanica applicata alle macchine. Unione Tipografico-Editrice
Torinese (UTET), Torino
22. Freudenstein F (1962) On the variety of motion generated by mechanism. ASME J Eng Ind,
pp 156–160
23. Galassini A (1962) Elementi di Tecnologia Meccanica, Macchine Utensili, Principi
Funzionali e Costruttivi, loro Impiego e Controllo della loro Precisione, reprint 9th edn.
Editore Ulrico Hoepli, Milano
24. Giovannozzi R (1965) Costruzione di Macchine, vol II. Casa Editrice Prof. Riccardo Pàtron,
Bologna
25. Gray A (1997) Modern differential geometry of curves and surfaces with mathematica, 2nd
edn. CRC Press, Boca Raton, FL
26. Hain K (1967) Applied kinematics. McGraw-Hill, New York
27. Henriot G (1979) Traité théorique and pratique des engrenages, vol 1, 6th edn. Bordas, Paris
28. Henriot G (1972) Traité théorique et pratique des engrenages. Fabrication, contrôle,
lubrification, traitement thermique, vol 2. Dunod, Paris
29. Jain KC, Chitale AK (2014) Textbook of production engineering, 2nd edn. PHI Learning
Private Limited, New Delhi
30. Jelaska DT (2012) Gears and gear drives. Wiley, Chichester, U.K
31. Johnson KL (1985) Contact mechanics. Cambridge University Press, Cambridge
32. Juvinall RC, Marshek KM (2012) Fondamentals of machine component design, 5th edn.
Wiley, New York
33. Kalker JJ (1990) Three-dimensional elastic bodies in rolling contact. Kluwer Academic
Publishers, Dordrecht
34. Leutwiler OA (1917) Elements of machine design. McGraw-Hill, New York
35. Levi-Civita T, Amaldi U (1929) Lezioni di Meccanica Razionale—Vol. 1: Cinematica—
Principi e Statica, 2nd edn. Reprint 1974, Zanichelli, Bologna
36. Litvin FL, Fuentes A (2004) Gear geometry and applied theory, 2nd edn. Cambridge
University Press, Cambridge
37. Micheletti GF (1977) Tecnologia meccanica, 2nd edn. Unione Tipografico–Editrice Torinese
(UTET), Torino
38. Mindlin RD (1949) Compliance of elastic bodies in contact. J Appl Mech 16(3):259–268
39. Niemann G, Winter H (1983) Maschinen-elemente, Band II: Getriebe allgemein,
Zahnradgetriebe-Grundlagen, Stirnradgetriebe. Springer, Berlin
40. Polach O (2005) Creep forces in simulations of traction vehicles running on adhesion limit.
Wear 258:992–1000
41. Poncelet JV (1874) Course de Mécanique Appliquée aux Machines. In: Posthumous (ed) vols
1, 2. Gauthier-Villars, Imprimeur-Libraire de l’École Polytechique, Paris
42. Radzevich SP (2017) Gear cutting tools: science and engineering, 2nd edn. CRC Press, Taylor
& Frencis Group, Boca Raton
43. Rajput RK (2007) A textbook of manufacturing technology: manufacturing processes, 1st
edn. Laxmi Publications (P) Ltd, Bangalore
44. Reuleaux F (1893) The constructor. A handbook of machine design. German edition: first
published in (1861). H.H. Suplee, Philadelphia
45. Reye T (1860) Zur theorie der zapfenreibung. Der Civilingenieur 4:235–255
46. Rossi M (1965) Macchine utensili moderne. Ulrico Hoepli Editore, Milano
138 3 Characteristic Quantities of Cylindrical Spur Gears …
47. Scotto Lavina G (1990) Riassunto delle Lezioni di Meccanica Applicata alle Macchine:
Cinematica Applicata, Dinamica Applicata—Resistenze Passive—Coppie Inferiori, Coppie
Superiori (Ingranaggi—Flessibili—Freni). Edizioni Scientifiche SIDEREA, Roma
48. Sharma PC (1999) A textbook of production engineering. S. Chaud & Company Ltd,
New Delhi
49. Veldkamp GR (1967) Canonical systems and instantaneous invariants in spatial kinematics.
ASME J Mech 2:329–388
50. Weisbach J (1894) Mechanics of engineering and machinary. In: Hermann G (ed) German
edition: first published in (1875). Wiley, New York
Chapter 4
Interference Between External Spur
Gears
Abstract In this chapter, the problems of interference between external spur gears
in both cutting and working conditions are first described and discussed. The dis-
cussion is quite general and therefore includes involute or primary interference and
fillet interference. The relationships of the main geometric quantities that allow to
avoid the involute interference in working conditions are obtained and design
indications in this regard are provided. The reductions of the path of contact due to
primary interference both in cutting and operating conditions are then quantified,
and the beneficial effects of the cutting interference on the intentional modifications
at the tip and root of the tooth profile are discussed. The geometry of the fillet
profile generated by means of a rack-type cutter (or hob) and a pinion-type cutter is
subsequently defined, and the interference effects of the rounded tip of the cutter
teeth are discussed. Finally, suggestions are given to avoid the dangers of the fillet
interference in operating conditions.
4.1 Introduction
In Sects. 2.4 and 3.1 we have already introduced the topic of the interference, when
we defined the maximum radii of the addendum circles needed to avoid it. In this
Chapter we will deepen further this interesting topic, because it is an important step
of the gear design. In fact, interference results in a typical undercut tooth, which not
only weakens the tooth to its root (it takes shape as a wasp-like waist), but also
removes part of the useful involute profile near the base circle (see Buckingham [2],
Dudley [6], Giovannozzi [11], Henriot [15], Litvin and Fuentes [20], Maitra [21],
Merritt [23], Niemann and Winter [25], Pollone [26], Radzevich [29]).
It is first necessary to keep in mind that two profiles, although conjugate, cannot
always be actually usable profiles of a kinematic pair. In fact, in some cases, to have
the certainty that the kinematic coupling can take place, the members of a kinematic
pair should be characterized by penetration of their solid parts. In other words, the
members of the kinematic pair should allow that physics does not allow, that is this
penetration or interference between solid bodies [8].
Fig. 4.1 Interference between involute profiles of two centrodes touching at point C
This happens, for example (Fig. 4.1), between the involutes p1 and p2 generated
by the tracing point P on a straight-line t, which rolls without sliding on two pitch
curves or centrodes (respectively, the fixed and moving centrodes, pf and pm ,
touching at point C, which is the instantaneous center of rotation through which the
straight-line t passes. Remember, incidentally, that centrode in kinematics is the
path traced by the instantaneous center of rotation of a rigid plane figure moving in
a plane [7]. Then, in Fig. 4.1, centrodes are the loci of the pitch points for each
angular position of the straight-line t when the latter rolls without sliding on the
pitch curves (these are the evolute curves), pf and pm .
However, the two involutes, namely the profiles p1 and p2 so generated, are each
ðp1 Þ the envelope of the successive positions taken by the other ðp2 Þ during the pure
rolling motion of the centrode pm , which is rigidly connected with profile p2 , on the
other centrode. However, these involute profiles cannot be taken as conjugate
profiles because, having in correspondence of the point of contact P the same center
of curvature (the point C), the type of contact at point P is a contact of second order,
for which the two involutes intersect each other. However, we must consider the
involute profiles p1 and p2 not as geometric curves, but as edges that delimit real
physical bodies, i.e. as boundaries of solid body sections. Figure 4.1 highlights the
fact that, wherever the solid parts of the two profiles are positioned, there is always
a certain region in correspondence of which the solid parts are placed at the same
side; thus, the interference occurs, and the appropriate kinematic coupling between
the two members is impossible.
An interference of the same type occurs all the times that two profiles have, in
their tangent point, a contact of the second order, that is a point of contact where
they have not only the same tangent, but also the same center of curvature and,
therefore, the same curvature. However, the interference can also occur when at
point of tangency between two profiles there is a simple contact of first order,
because the parts of the profiles, the most distant from the point of contact, are to
intersect. This condition, which occur quite often, forces to limit the extension of
the profiles actually usable.
The interference, of whatever kind it may be, is never allowed during the
meshing of two toothed gear wheels in their actual working conditions. In fact,
4.1 Introduction 141
if the interference would take place during operation, an unfavorable condition for
transmission of motion would occur, due to jamming between the teeth pairs
successively in meshing.
On the contrary, as we will specify in more detail below, during the cutting
process of the teeth a certain interference between the gear wheel to be cut and the
cutting tool is not only permissible, but in certain cases it is also desirable and
desired. This why, by means of it, those parts of the involute profile of the teeth,
which, for various reasons, may generate problems of transmission of motion, can
be removed [11].
For the discussion of this important subject, we distinguish the case of external
gear pairs from that of internal gear pairs. In the first case, two types of interference
may occur: the theoretical interference (also called involute interference or primary
interference), and the fillet interference. In the second case, five types of interfer-
ence may occur (see Colbourne [3], Radzevich [28], Yu [37]):
• the theoretical interference;
• the secondary interference;
• two fillet interferences (the first, between the tip of the pinion and root of the
annulus, and the second between the tip of the annulus and root of the pinion);
• the trimming interference, i.e. the interference of the radial approach of the
cutter during the generation cutting process, or of the pinion during the assembly
of the internal gear pair.
In Sect. 2.4 we have already described in detail what this type of interference
consists and we have seen that, to avoid it, the outside (or tip or addendum) circles
of the pinion and wheel must have radii at most equal to O1 T2 and O2 T1 respec-
tively (see Fig. 2.9). In particular, we have shown that, if the tooth involute profiles
are too extended in the radial direction (i.e., if the tip circles of the two members of
the gear pair go beyond the point T1 and T2 ), the parts of the same profiles extended
beyond these points are no longer able to transmit correctly the motion. This instead
does not happen when the same profiles are in contact over the entire line of action,
T1 T2 .
If the pinion teeth were higher (and therefore the radius of its addendum circle
was greater than O1 T2 ), by rotating the two wheels of the gear pair as if meshing
properly, we would see that, beyond the point T2 , the tooth tip of the wheel 1
penetrates the tooth flank of the wheel 2. Obviously, in a case like this, when the
contact reaches point T2 , the motion could no longer continue, due to the jamming
of the tooth tip of the pinion 1 against the dedendum flank of the tooth of the wheel
2. In this case, therefore, the interference would occur as the impossibility of
operation of the gear pair (see also Pollone [26]).
142 4 Interference Between External Spur Gears
The maximum values of the addendum of the wheel and pinion must therefore
satisfy respectively the inequalities:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m 2 2
ha2 max z2 þ z1 þ 2z1 z2 sin a z2
2 ð4:4Þ
2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m 2 2
ha1 max z1 þ z2 þ 2z1 z2 sin a z1 :
2 ð4:5Þ
2
From these two inequalities, for given values of a, z1 and z2 , with z2 [ z1 , ha2 max
is less than ha1 max . Therefore, if the two wheels of the gear have equal addendum,
when the value of this increases, the interference of the teeth of the wheel 2 with
those of the wheel 1 occurs before.
4.2 Theoretical Interference Between External Spur Gears 143
From the inequality (4.4), solved with respect to z1 , we get the following min-
imum number of teeth of the pinion, z1 min , that meshes, without interference, with
the wheel having z2 teeth:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
h a2 max z 2 þ ha2 max
z1 min z2 þ z22 þ 4 m
; ð4:6Þ
m sin2 a
which, by introducing, in accordance with the second of Eqs. (3.9), the maximum
addendum factor of the wheel 2, k2 max ¼ ðha2 max =mÞ, can be written in the more
compact form:
144 4 Interference Between External Spur Gears
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðz2 þ k2 max Þ
z1 min z2 þ z22 þ 4k2 max : ð4:7Þ
sin2 a
by introducing, in accordance with the first of Eqs. (3.9), the maximum addendum
factor of the pinion 1, k1 max ¼ ðha1 max =mÞ, this inequality can be written in the
form:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðz1 þ k1 max Þ
z2 min z1 þ z21 þ 4k1 max : ð4:9Þ
sin2 a
The chart of Fig. 4.3, in which the number of the teeth z1 of one of the two
wheels of the gear pair is shown on the ordinate, while the number of the teeth z2 of
the other wheel is shown on the abscissa, shows two families of curves, each
corresponding to a given value of the pressure angle a. The two families of curves
are separated by means of the dotted straight line defined by the equation z1 ¼ z2 ,
from which the two branches of the same curve relating to a given value of the
pressure angle bifurcate.
Curves at the right side of this straight line refer to gears for which z1 \z2 , and
represent the function given by Eq. (4.6) in the case in which the center distance is
the nominal one, and ha2 max =m ¼ 1 ðk2 max ¼ 1Þ. Curves at the left side of this
straight line instead refer to gears for which z1 [ z2 , and represent the function
given by Eq. (4.8), also here in the case in which the center distance is the nominal
one, and ha1 max =m ¼ 1 ðk1 max ¼ 1Þ.
Example The following example clarifies how the chart of Fig. 4.3 must be used.
Consider a toothed wheel having z1 ¼ 15 teeth and pressure angle a ¼ 20 . We
want to know for which value of z2 the interference occurs. The horizontal line
z1 ¼ 15, parallel to the abscissa axis, intersects the two branches of the curve
relating to the pressure angle a ¼ 20 at the points E and F. From the chart we can
deduce that: for wheels with a number of teeth z2 which is included in the range
between points D and E of this horizontal line ðz2 \13Þ, interference occurs
between the tooth tip of the addendum flank of the wheel with z1 ¼ 15 and the
dedendum flank of the wheel with z2 \13 teeth. For wheels with a number of teeth
4.2 Theoretical Interference Between External Spur Gears 145
Fig. 4.3 Chart to determine z1 min as a function of z2 , for z1 \z2 ; and z2 min as a function of z1 , for
z2 \z1 ; ðha1 max =m ¼ ha2 max =m ¼ 1, and nominal center distance)
z2 which is included in the range between points E and F ð13\z2 \45Þ, interfer-
ence does not occur; for z2 [ 45 interference occurs between the tooth tip of the
addendum flank of the wheel with z2 teeth and the dedendum flank of the wheel
with z1 ¼ 15 teeth.
However, the case that occurs more often is that in which a given gear ratio
u ¼ z2 =z1 must be obtained by means of a gear pair having normal sizing, and with
the pinion with a minimum number of teeth z1 as possible, without the interference
between the wheel and pinion occurs. In this case, processing Eq. (4.4), we obtain
the following inequality:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
1 þ 1 þ z2 z2 þ 2 sin a
z1 z1 2
2ha2 max
z1 min ; ð4:10Þ
m
z2 þ 2 sin a
z1 2
146 4 Interference Between External Spur Gears
Fig. 4.4 Areas with different conditions of interference for a spur gear pair
which, for ha2 max =m ¼ k2 max ¼ k2 , is nothing more than the relationship (3.14),
already obtained by other considerations in Sect. 3.1.2.
The conditions of theoretical interference between the two members of a spur
gear pair depend not only on their number of teeth, z1 and z2 , and then on the ratio
1=i ¼ 1=u ¼ z1 =z2 , but also on the transverse contact ratio ea ¼ ga =pbt (see
Eq. 3.43). Taking into account all these quantities ðz1 ; z2 ; 1=i; and ea Þ, it is possible
to identify four areas, each of which corresponds to different operating conditions.
With reference to Fig. 4.4, these areas are characterized by:
• area A, with dangerous interference on both members of the gear pair;
• area B, with dangerous interference on the gear wheel 1 (the pinion);
• area C, with dangerous interference on the gear wheel 2;
• area D, without any interference on both members of the gear pair.
The interference should also be seen in comparison with the backlash between
the teeth, as both influence the operating conditions of the gear. Remember that the
backlash is the amount by which the space width exceeds the tooth width thickness.
It is first necessary to prevent that the non-operating flanks of gear teeth come into
contact. Backlash is also needed to accommodate different thermal expansions of
the gear pair, tooth deflections, foreign material in the lubricant, as well as errors in
the assembly, manufacture, and operation of gears in mesh. Anti-backlash gears,
i.e. gears with zero backlash, are sometimes used in gear trains if the driving gear
changes direction of rotation frequently.
4.2 Theoretical Interference Between External Spur Gears 147
Now, if considerable backlash exists between the teeth and at the same time the
gear is characterized by theoretical interference, the transmission of motion is not
interrupted certainly, but the contact between the teeth takes place in very bad
conditions, causing variations in angular velocities, high noise, strong vibrations
and very fast wear. Instead, if the backlash between the teeth does not exist or it is
too small, the transmission of motion is interrupted because of the jamming.
In Sect. 3.2 we made some considerations on the theoretical interference during the
cutting processes, both without generation (for example by form milling cutters)
and with generation (e.g., by means of rack-type cutters, hobs, and pinion-type
cutters). However, the problem of the theoretical interference in the cutting pro-
cesses cannot be considered concluded with the observations made in that section.
In respect of this issue, it is necessary that we make further considerations to be kept
in mind in the gear design.
The cutting process without generation is performed, more frequently, by means
of a single process, with milling machines using form cutters (a disk-type gear
cutter, end mill) and indexing head, but also with an entire process, with a profiled
cutter (an involute gear cutter), by broaching, stamping and cold drawing. Whatever
the cutting process, the cutters used have profiled section as the space width
between the successive teeth of the wheel to be profiled. With reference to the
milling process (the same happens with the other process without generation), if the
involute gear cutter has a theoretical profile that extends to the whole tooth flank
depth, and the wheel has many teeth, we might have interference when it is coupled
with a pinion having few teeth.
In the continuous generation process (also called generation for envelope),
which is made with all purpose gear shapers, type Fellows, pinion-type cutters are
used, which are cutting tools equivalent to involute toothed wheels. These
gear-cutting machines, in addition to the slotting motion, also have a rolling motion
relative to the wheel to be cut, with a transmission ratio defined by the ratio between
the number of teeth of the pinion-type cutter and number of teeth of the wheel. If
the number of teeth of the pinion-type cutter is not suitably chosen in relation to that
of the teeth of the toothed wheel to be cut, two type of interference may occur,
namely:
• interference between the addendum flank towards the tip of the pinion-type
cutter and the dedendum flank towards the root of the teeth of the wheel,
whereby the teeth of the latter are weakened at the root due to the root relief, in
the case of interference not very high, or to the undercut, in the case of very high
interference;
148 4 Interference Between External Spur Gears
Fig. 4.5 Effects of the interference in the cutting conditions with pinion-type cutter
• interference between the addendum flank towards the tip of the wheel and the
dedendum flank towards the root of the teeth of the pinion-type cutter, whereby
the teeth of the wheel will have a smaller thickness towards the crest, due to the
tip relief [15].
Figure 4.5 shows the effects of the interference in the cutting conditions, in the
case where the cut is made with a pinion-type cutter (remember that here the
subscript 0 identifies the cutting tool and, more generally, the cutting conditions).
From the figure it is evident that, due to a too high addendum of the wheel to be cut
ðha2 [ ha2 max Þ, which brings its addendum circle beyond the interference point T0
of the pinion-type cutter, the interference causes a pronounced tip relief of the
wheel. The same figure also shows that the pinion-type cutter has an addendum
greater than the maximum value to avoid the interference ðha0 [ ha0 max Þ, so also its
addendum circle goes beyond the interference point of the wheel; because of this,
the interference determines a root relief of the wheel.
Example now we do an example of cutting interference, when the cutting is done
with a pinion-type cutter, using for this purpose the chart shown in Fig. 4.3.
Suppose to use a pinion-type cutter with z0 ¼ 24 teeth, pressure angle a0 ¼ 15 ,
and normal sizing ðh0 ¼ 2:25m0 ; ha0 ¼ m0 ; hf 0 ¼ 1:25m0 Þ. From Fig. 4.3 we infer
that, using such a cutting tool, we can cut wheels with z2 teeth in the range
ð22 z2 44Þ without the cutting interference takes place. Instead, we will have
the cutting interference with the dedendum flank toward the root of the wheel for z2
less than or equal to 21 teeth, and with the addendum flank towards the tip of the
wheel for z2 greater than 44 teeth. To cut without cutting interference a wheel with
4.3 Possibility to Realize Gear Pairs with Pinion Having … 149
21 teeth, the pinion-type cutter to be use, still remaining its other characteristics,
must have 21 teeth instead of 24 teeth.
In the semi-continuous generation process for envelope, which is made with
gear rack cutting machines, straight-sided rack-type cutters are used. With these
machines, after each cutting stroke, the gear blank and rack type cutter roll slightly
on their pitch circles and, when the blank and cutter have rolled a length equal to
the pitch, the cutter is returned to the starting point, and the process is continued
until all the teeth have been cut.
In the continuous generation process for envelope, which is made with gear
hobbling machines (in these machines, the hob and blank rotate seamlessly at the
proper angular velocity ratio, but the hob is also fed slowly across the face width of
the blank until all the teeth have been cut), hobs are used, which are cutting tools
shaped like worms. The hob teeth have straight flanks, as in the rack-type cutter.
However, the hob axis must be turned through the lead angle in order to cut spur
gear teeth. For this reason, the teeth generated by a hob have a slightly different
shape than those generated by a rack-type cutter (this despite the fact that each
longitudinal section of hob has the profile of the rack cutter). For more insights on
these cutting processes, we refer the reader to Micheletti [24], Galassini [9], Rossi
[31], Henriot [14], Dooner and Seireg [5], and Marinov [22].
From the point of view of the cutting interference, we can consider completely
equivalent the two above-mentioned processes (the semi-continuous and the con-
tinuous one), since, in this regard, the differences are not significant. For both
processes, with a normal standard sizing of the teeth, the minimum number of teeth
below which we do not to go down, if we want to avoid the cutting interference, is
to be deduced from the chart of Fig. 3.1 for 1=u ¼ 0.
We have already shown (see Sect. 3.1) that the risk of interference increases
when the number of teeth z2 of the wheel increases, and the number of teeth z1 of
the pinion decreases. The risk of the cutting interference is then maximum when we
wanted to cut a pinion with a few teeth using the rack-type cutter, which can be
considered as a wheel tool of infinite radius, and thus with number of teeth even
infinite. Figure 4.6 shows the very high cutting interference that occurs in the case
of a pinion with eight teeth, cut by means of a rack-type cutter having teeth with
normal standard sizing and pressure angle a0 ¼ 15 . The tooth is strongly carved at
its root, with a very pronounced undercut, which determines the removal of
material of the dedendum flank toward the root even within the fillet. This undercut
occurs because the addendum line of the rack-type cutter intersects the line of action
beyond the interference point T1 of the pinion to be cut, so that ha0 ¼ ha0 max [15].
The undercut often constitutes a design choice, especially in cases where we
want to remove the part of the involute near the base circle [11]. In fact, in cor-
respondence of the base circle, the involute has a radius of curvature equal to zero,
for which it is extremely difficult (if not impossible) to obtain accurately with the
generation cutting of conventional processes, which, moreover, are the most
technologically advanced. The contact at the point of regression of the involute on
the base circle, which is a particular singular point, is also to be avoided due the
150 4 Interference Between External Spur Gears
Fig. 4.6 Effect of the interference (undercut) on a pinion with eight teeth, in cutting conditions by
means of a rack-type cutter
radius of curvature equal to zero. As already we have seen, this would determine
specific sliding tending to infinity, and, as we shall see later, equally infinite values
of the Hertz stresses (see Chap. 2 of Vol. 2). Lastly, the undercut may be desirable
to accommodate post-processing, such as shaving, honing, or burnishing. The
material removal in these post-processes is called shaving stock [29].
A safe way to get undercut is to use cutting tools with a protuberance (see
Fig. 2.16a). The undercut determines a reduction in tooth thickness in the fillet
region, for which it, based on considerations of stress concentrations and bending
and fatigue strength, is not desirable. Moreover, teeth that have been undercut can
determine additional discontinuities in the transverse contact ratio. However, it is
equally likely that, even with cutting tools without protuberance, if interference
conditions are met, one can get involute profiles characterized by cusps or slope
discontinuities, with all the ensuing problems.
In conclusion, it is to be observed that, with the generation cutting of the teeth,
whatever the process used (with rack-type cutters, pinion-type cutters or hobs), the
tool cuts in the wheel blank space widths corresponding to their free relative
motion, and therefore we will never have jamming. This is due to the relative
motion between the generation cutting tool and the wheel to be cut. However, in the
case in which the theoretical interference occurs, the involute profile of the tooth
flank of the wheel will be altered.
With the single cutting process by milling, with form cutter, the effects of the
interference during the cutting operations are not to be feared, as there is no a
relative motion between the cutting tool and the wheel to be cut. However, it may
possible have jamming between the two gear wheels generated with this process,
when they are coupled to form a gear pair.
4.3 Possibility to Realize Gear Pairs with Pinion Having … 151
With the exception of the case of the generation process by means of a rack-type
cutter, we must always verify that the risk of interference does not have, both during
the generation of the two wheels of the gear pair, and during their mutual meshing.
Example We want to realize a gear pair having: ha1 max ¼ ha2 max ¼ m; a ¼ 15 ;
z1 ¼ 22 teeth and z2 ¼ 27 teeth. This pair can be generated without interference
only by milling, with a profiled form cutter, or by generation using a pinion-type
cutter having between 22 and 27 teeth. The two wheels of the gear pair cannot
instead be generated without interference using a rack-type cutter.
In the design of spur gear pairs characterized by pinions with small number of teeth,
in order to prevent the theoretical interference, we can implement different methods
of solution, which emerge clearly from the foregoing considerations in the previous
sections. In this regard, the most frequently used methods (they can be used not
only individually, but also simultaneously, in various combinations, in order to
enhance their effects), are as follows [35]:
• increase in the nominal pressure angle;
• use of a stub system (see Sect. 2.5), with teeth having depths smaller than those
defined by the nominal sizing system, and given by Eq. (2.71);
• use, for a given nominal pressure angle, of teeth having nominal standard
depths, but with profile shift, i.e. with a different distribution of the total depth of
the tooth between the addendum and dedendum, and therefore with displace-
ment of the addendum and dedendum circles.
With reference to the first method, from the inequalities (4.6) and (4.7) as well as
from the chart shown in Fig. 4.3, it follows that the minimum number of teeth of the
pinion, which meshes without interference with a given wheel, decreases when the
nominal pressure angle of the toothing increases. This is referred to wheels having
nominal standard sizing, and operating with the nominal center distance.
Example Wheel with z2 ¼ 50, having nominal standard sizing and operating with
the nominal center distance. From inequalities (4.6) and (4.7) and chart of Fig. 4.3,
we obtain: z1 min ¼ 25, for a ¼ 15 ; z1 min ¼ 19, for a ¼ 17 300 ; z1 min ¼ 16, for
a ¼ 20 ; z1 min ¼ 13, for a ¼ 22 300 .
We obtain very similar results when the ratio 1=i ¼ z1 =z2 is an input data of the
design; obviously in this case we use the inequality (4.10) and the chart shown in
Fig. 3.1.
Example External gear pair with ratio 1=i ¼ z1 =z2 ¼ 1=4, having nominal stan-
dard sizing and operating with the nominal center distance. From inequality (4.10)
and chart of Fig. 3.1, we obtain: z1 min ¼ 27, for a ¼ 15 ; z1 min ¼ 16, for a ¼ 20 .
152 4 Interference Between External Spur Gears
greater resistance to bending of the teeth. The most negative aspect of this method
is also the economic one, since the usual modular cutters cannot be used (they are
standardized types of cutters corresponding to some fixed value of the stub factor,
defined as the ratio between the stub addendum and the nominal standard adden-
dum, equal to 0.9 or 0.8).
The third method consists of a decrease of addendum of the wheel (recall in this
regard that the more restrictive condition to avoid interference is that related to the
addendum circle of the wheel having a larger diameter, see Sect. 3.1), obtained
approaching its addendum circle at nominal pitch circle and moving away from the
latter the dedendum circle. In doing so the total depth of the wheel tooth does not
change; only its sizing change, since its addendum is less than the nominal one,
while its dedendum is greater than the nominal dedendum by an amount equal to
the decrease of addendum. In this way, as the inequality (4.10) shows, the minimum
number of teeth, z1 min , of the pinion decreases proportionally to ha2 max . Of course,
the addendum of the pinion increases and its dedendum decreases to the same
extent as, respectively, the dedendum of the wheel increases and addendum of the
latter decreases.
With this method, the path of contact AE (Fig. 2.10) moves along the line of
action, in the direction that goes from T1 to T2 , but its length remains substantially
unchanged. With this displacement of the path of contact from left to right
(Fig. 2.10), the danger of interference is removed during the approach (point
A moves away from T1 ), but it should be checked that it is not present during the
recess (point E approaches T2 ). Another drawback that can occur with this solution
is that of an excessive reduction of the tooth thickness of the pinion to its crest; this
drawback manifests itself especially in cases in which the two wheels of the gear
pair must work with nominal center distance.
In order to remedy, at least partially, to the above drawbacks, a solution a little
more complex than that described above is used. It combines the aforementioned
addendum modifications with a change of the center distance in the operating
conditions, compared to the center distance in the cutting conditions. It follows that
the pressure angle under actual operating conditions is greater than that which
occurs in the cutting conditions, with all the resulting benefits. This combined
method is of considerable design importance, because it enables the improvement
of certain important operating quantities of the gear pair (in particular the decrease
of the specific sliding), and therefore it is used widely, even in cases where the
interference is not to be feared.
All three of the aforementioned methods to avoid interference require to have
generation tools with different pressure angles and with various depths of the teeth.
In addition, according to the type of cutting tool used, they require to verify that,
during the process of generation, there is no interference between the generation
tool and the gear wheel to be cut.
The processes most widely used in the cylindrical spur gear production are the
generation processes, that employ rack-type cutters, hobs and pinion-type cutters as
generation tools. These cutting tools are very expensive. It is therefore interesting to see
how, with an usual tool characterized by certain nominal quantities (module and
154 4 Interference Between External Spur Gears
pressure angle), gears with the same nominal quantities of the tool can be obtained,
with even small number of teeth, avoiding problems due to interference. So, we get
toothing with addendum modification generated with rack-type cutters (hob, Maag
cutter, Sunderland cutter), and those generated with pinion-type cutters (Fellows cutter).
Let us now examine the generation cutting process, by a rack-type cutter (the same
considerations are substantially valid in the case in which hobs are used for cutting),
of toothed wheels having a number of teeth z less than the minimum number of
teeth zmin that can be cut without interference. In these conditions, it is obvious that
a part of the involute profile of the teeth of the gear wheel to be cut will be removed
as a result of the cutting action of the addendum edge of the rack-type cutter; it
fellows clearly a decrease in the path of contact.
For the analysis of the problem, we assume that (see Sect. 3.2), as regards the
risk of cutting interference, things are as if the cutter addendum profile ended at the
beginning of tip fillet. Consequently, we assume that this part of the addendum
profile is not dangerous in respect to the interference (see Giovannozzi [11],
Schiebel [33], Schiebel and Lindner [34]). In other words, we consider nonexistent
the part of the rack-type cutter profile, which connects the straight-line involute
from the point where it ends up to the edge on the tip line. Therefore, with reference
to the effects of the cutting interference, things go as if the tooth of the rack-type
cutter was truncated, sharp-edged, in that point in which the straight-line involute
finishes. In these conditions, the active dedendum ha0 of the rack-type cutter will be
equal to the difference between the dedendum hf ¼ 1:25=m of the teeth to be cut
and the dimension normal to the tip line of its curved profile at tip edge; the latter is
equal to the root fillet radius, qfP , of the basic rack. Then we have ha0 ¼ hf qfP .
In order to quantify the decrease of the path of contact due to interference, we
must first identify the relative position between the cutting tool and the profile of the
wheel to be cut, in the condition of incipient cutting interference. This is the
position in which the virtual tip (the one defined above) with a sharp edge of the
tool is for the first time in contact with the profile of the wheel, a part of which will
subsequently be removed. In Fig. 4.7 this position is clearly identified by the point
P, where the virtual tip line of the rack cutter intersects the involute profile of the
wheel, whose point of regression on the base circle having radius rb is the point P0 .
In the aforementioned position of incipient cutting interference, the straight-line
profile of the tool is tangent at point A to the other branch of the involute (the one
shown in the figure with a dashes line, which we might call the virtual involute),
also with the point of regression at the point P0 and symmetrical compared to the
first branch with reference to the radial straight line OP0 . The fact that point P is
precisely the point of incipient cutting interference fellows from the fact that, during
4.5 Reduction of the Path of Contact Due to Cutting Interference 155
Fig. 4.7 Relative position between the rack cutter profile and profile of the wheel to be cut, in the
condition of incipient cutting interference
cutting, the tool profile must be always tangent to an involute profile having the
center of the curvature on the normal to the straight profile of the tool. Therefore,
due to known properties to the involute, if two profiles are touching at a point A0
between points T and C (i.e., within the theoretical length of the path of contact), a
branch of involute exists that touches the next profile of the tool at point A. This last
point is located on the line of action at a distance from the point A0 equal to the base
pitch pb (see Gray [12], Lawrence [19]). The distance AA0 is therefore equal to the
length of the arc of base circle between the origins P0 and P00 of two successive
involute profiles.
Figure 4.7, which shows the condition crystallized precisely in the instant in
which the cutting interference begins, highlights the fact that, due to the interfer-
ence, the portion of the profile corresponding to the arc PP0 is lost. For known
geometric properties of the involute, this arc corresponds to a loss of the path of
contact equal to the distance PK between the points P and K, the latter being the
point of tangency of the normal at point P to the involute profile with the base
circle. However, the length of the segment PK is equal to that of the arc P0 K, i.e.
PK ¼ P0 K ¼ rb u; where u is the angle subtended by the arc P0 K; these lengths are
also equal to the length of the segment TN, where N is the point of intersection of
the line of action with the circle having fixed center O and radius OP (Fig. 4.7).
156 4 Interference Between External Spur Gears
With the notations shown in Fig. 4.7, where, for convenience of calculation, the
auxiliary angle h was introduced (h is the angle subtended by the arc P0 T), pro-
jecting the broken line OKPA along the directions of the straight line CH and OT,
we obtain the following pair of trigonometric equations:
sinðh þ uÞ u cosðh þ uÞ ¼ h
ð4:11Þ
cosðh þ uÞ þ u sinðh þ uÞ þ ðk hÞ tan a ¼ 1;
c sin c
u¼ ; ð4:12Þ
1 cos c
Fig. 4.8 Distribution curves of the function u ¼ uðkÞ, for three different values of the pressure
angle a
4.5 Reduction of the Path of Contact Due to Cutting Interference 157
Once the value of the parameter k has been so determined, we will derive u by
means of the curves shown in Fig. 4.8. For values of the pressure angle a different
from those considered in Fig. 4.8, it is necessary to draw on it the corresponding
curves, and then to proceed in the same way, or use directly the analytical equations
described above, which are valid for any value of the angle a.
If the profiles of the teeth of both wheels of an external gear pair are cut by a
rack-type cutter with a certain interference, according to what we said above, due to
the removal of the corresponding profiles, two parts of the path of contact,
respectively having lengths T1 N1 ¼ rb1 u1 and T2 N2 ¼ rb2 u2 , are unusable
(Fig. 4.9). The quantities u1 and u2 can be obtained using the curves of Fig. 4.7 or
with the analytical relationships above mentioned.
The loss of these two portions of the theoretical path of contact affects the use of
the relationship (3.44), which is used to calculate the transverse contact ratio of an
external spur gear. In this respect, in fact, this relationship can be used without
further considerations only in the case in which the values of the addendum factors
k1 and k2 are of such magnitude that the addendum circles related to them intersect
the line of action at points located within the segment whose end points are N1 and
N2 (Fig. 4.9). In this case, the risk of interference between the two meshing wheels
of the gear pair, under real operating conditions, is excluded. Instead, in the case in
which the addendum factors k1 and k2 have values such that the addendum circles
intersect the line of action at external points of the segment whose end points are N1
and N2 , the values of k1 and k2 to be introduced in the relationship (3.44) are those
that it would have if the addendum circles would pass through the points N1 and N2 .
We must keep in mind that the cut of toothed wheels having small numbers of
teeth as desired (for example, less than 8 teeth for a ¼ 15 , and less than 7 teeth for
a ¼ 20 Þ is not possible, because below such numbers it would have pointed teeth,
that is, with zero tooth thickness at top land. In gear design it is always advisable to
check the tooth thickness at top land, especially in the case where a profile shift is
made; we will return later on how this checking must be done.
158 4 Interference Between External Spur Gears
Fig. 4.9 Unusable parts of the path of contact due to the cutting interference
connection between the involute profile part and dedendum circle is as smooth as
possible, that is made without abrupt variations of the radius of curvature.
The fillet profile, which is not an involute curve, does not take part to the contact
between the two meshing wheels of the gear. If a fillet contact occurs, the tooth tips of
a wheel would dig material from the root fillets of the other, and a smooth meshing
between the two wheels of the gear pair would be impossible. This phenomenon is
known as fillet interference (it is also called trochoid interference), and gear pairs
must be designed in such a way that this type of interference does not occur.
If we want to better understand the problem, so to have the verification methods
of the gears that allow us to dispel definitely the risk of fillet interference, it is
necessary to deepen our understanding of the geometry of the fillet, in relation to
the cutting processes. In this regard, we consider a transverse section of one
member of the spur gear pair under consideration (Fig. 4.10), and call, as fillet
circle, the circle at which the bottom point of the involute profile is connect with the
top point of the fillet profile. Above this point (indicated by Pf in Fig. 4.10), where
the fillet circle intersects the tooth profile, we have the involute profile, while below
it we have the fillet profile, the geometry of which is closely correlated with that of
tip edge of the cutting tool and the associated cutting process. The fillet circle has
radius rf ¼ OPf .
From the point of view of a greater bending strength of the tooth near its root, it
would be appropriate for the fillet profile to be outside of the extension of the
involute inside the fillet circle, but the cutting process is not always compatible with
this design goal. Anyhow, the radius rf depends on the cutting process and cutter
type used and, generally, it is larger than the radius rb of the base circle, because it
is impossible for the involute to extend inside the base circle. In exceptional cases,
rf may be equal to rb , but it is never smaller, that is rf rb .
160 4 Interference Between External Spur Gears
Whatever the generation cutting process (with rack-type cutter and, in equivalent
terms, with hob, or with pinion-type cutter), we are faced with three possibilities
(see also Henriot [15]):
• There is no a cutting interference, and then the bottom point of the involute
profile is external to the base circle ðrf [ rb Þ. The involute part of the tooth
profile near the root circle is free from a singular point in the cusp shape, the
fillet and involute profile have a common tangent at point Pf , and the wheel
tooth is not undercut.
• We are at the limit of the cutting interference, and then the bottom point of the
involute profile reaches the base circle ðrf ¼ rb Þ, and Pf coincides with the point
of regression, P0 , of the involute. Also, in this case, the involute part of the tooth
profile near the root circle is free from a singular point in the cusp shape, the
fillet and involute profile have a common tangent at point Pf ¼ P0 , and the
wheel tooth is not undercut.
• There is a cutting interference, for which the profile is undercut at point Pf
where the involute ends and the fillet begins. This point is a singular point in the
shape of a cusp, due to the discontinuity of the local tangent, which therein
undergoes a sharp change; in this case it is also placed at the outside of the base
circle, for which rf [ rb .
Figure 4.11 shows examples of generation of spur involute teeth by a rack-type
cutter. It show two different shapes of the tooth profile, respectively without
(Fig. 4.11a) and with undercut (Fig. 4.11b), both obtained as the envelope of the
successive positions assumed by the cutting tool during the relative motion between
the cutting tool and the wheel to be cut. The involute part of the tooth profile shown
in Fig. 4.11a is free from a singular point, the tooth is not undercut, and the fillet
profile and involute profile have a common tangent at their connection point.
Figure 4.11b instead shows a tooth with undercut, at point where the fillet profile
and involute profile intersect one another, so that at this connection point the
tangent undergoes an abrupt change, resulting in a singular point, a cusp.
Fig. 4.11 Generation of tooth profiles by a rack-type cutter: a Without undercut; b With undercut
4.6 The General Problem of the Interference … 161
For the moment, we leave aside the determination of the radius rf of the fillet
circle, which refers to the cutting process, and focus our attention on the operating
conditions, for which no interference (theoretical interference and fillet interference)
is allowed. The gear pairs to be considered in their working conditions are those
shown in Figs. 3.3a, b, which respectively show the meshing diagrams of an
external gear pair and a rack-pinion pair. In both cases, the first condition to be met
in order to avoid the interference is that concerning the theoretical interference.
In order to prevent the theoretical interference for an external gear pair, it is first
necessary that the limit circles or contact circles intersect the line of action at point
A and E (Fig. 3.3a) located inside the segment T1 T2 ; in other words, the path of
contact AE must be less than the line of action, T1 T2 , which represents its limit
value. It is noteworthy that the limit circles coincide with the tip or addendum
circles when the involute profile reaches the addendum circle; this is the hypothesis
that here we do. Therefore, the necessary condition for no interference in working
condition, AT2 \T1 T2 ; can be expressed in the following form (we assume that the
pinion is the smaller wheel, whereby the position of point A with respect to point T1
is the most restrictive):
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 r 2 \ðr þ r Þ tan a:
ra2 ð4:16Þ
b2 b1 b2
Following the same procedure, we can deduce that, to prevent the theoretical
interference for a rack-pinion pair (Fig. 3.3b), the first necessary condition that
must be met ðAC\T1 CÞ can be expressed in the form:
ha2
\rb1 tan a: ð4:17Þ
sin a
If the inequalities (4.16) and (4.17), respectively valid for external gear pairs and
rack-pinion pairs, are not satisfied, the theoretical interference will take place.
However, even when these inequalities are satisfied, they are not always sufficient
to prevent interference. In fact, the interference can still be present as fillet
interference.
We defined above the fillet circle, which intersects the tooth transverse profile at
point Pf , and has a radius rf , yet to be determined. We also said that point Pf is
located outside the base circle, and it is the top point of the fillet profile. We must
now ensure that the top of the fillet profile does not come into contact with the teeth
of the meshing wheel. To this end, the necessary and sufficient condition is that of
imposing that the point where the contact on the gear tooth profile begins is outside
with respect to the point Pf , namely that the radius of the limit circle, rc , is greater
than the radius of fillet circle, rf , and therefore:
rc [ rf : ð4:18Þ
162 4 Interference Between External Spur Gears
The radius of the limit circle, rc ¼ O1 A (see Figs. 3.3a, b), is determined on the
basis of simple geometric considerations. In the case of an external gear pair
(Fig. 3.3a), we have:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2
rc2 ¼ O1 A2 ¼ O1 T12 þ T1 A2 ¼ 2
rb1 þ ðrb1 þ rb2 Þ tan a ra2
2 r2
b2 : ð4:19Þ
2 ha2 2
rc2 ¼ O1 A2 ¼ O1 T12 þ T1 C AC ¼ rb1
2
þ rb1 tan a : ð4:20Þ
sin a
rc 0:025m rf : ð4:21Þ
The inequality (4.16) and (4.21) for external gears, and (4.17) and (4.21) for
rack-pinion pairs, are sufficient conditions to ensure that no interference will occur
at the top of fillet profile, near the point Pf . However, there is still the possibility of
interference at points of the fillet profile between the fillet circle and dedendum
circle. In order to avoid this risk, which will be better detailed in the following
sections, we must design the gear pair with adequate bottom clearances at each
dedendum circle (or root circle).
The bottom clearance at the root circle of the wheel 1 is defined as the difference
between the dedendum of the same wheel 1 and the addendum of the wheel 2 (and
vice versa for the wheel 2). Then, for the wheel 1, we have c1 ¼ hf 1 ha2 , and
therefore:
c1 ¼ r1 rf 1 ðra2 r2 Þ ¼ ðr1 þ r2 Þ rf 1 þ ra2 ð4:22Þ
where rf 1 and ra2 are respectively the radii of the dedendum and addendum circles
of the members 1 and 2 of the gear pair under consideration.
The recommended minimum value of the bottom clearance for both wheels of a
gear pair is equal to 0:25 m, and one of the reasons for this recommendation is to
help remove the risk of fillet interference. Notoriously (see Sects. 4.7 and 4.8), this
type of interference is more likely to occur at the pinion fillets than at those of the
wheel. Therefore, the gear designer, who must check that the above conditions of
non-interference are satisfied and, simultaneously, ensure that the bottom clearances
are adequate, performs both verifications for the pinion, while, for the wheel, he
circumscribes his attention to verification of such bottom clearance.
4.7 Fillet Profile Generated by a Rack-Type … 163
The basic rack is the fictitious rack that has, in its transverse cross section, the
reference standard profile, and to which the dimensions of a series of toothings
capable of meshing without distinction between them are referred. Then this basic
rack forms part of an intermating series. The counterpart rack, that is the rack
which can be inserted into the basic rack so that the teeth of one fit perfectly in the
space widths of the other, serves as the basis of the shape of a generating cutter
which will produce an intermating series of gears. It is the rack-type cutter, and is
characterized by the so-called axis of symmetry, that is the reference straight-line or
datum line on which the tooth thicknesses are equal to the space widths, the one and
the other theoretically equal to one-half of the pitch (Fig. 2.14).
The shape of the rack-type cutter consists of two straight lines passing through
the equally-spaced points along the axis of symmetry, and inclined at the pressure
angle a, that generate the involute curves of the toothed wheel, an addendum
straight line that generates the dedendum circle of the wheel, and the rack tip fillets
that generate the wheel fillets.
In order to specify the whole profile of the rack-type cutter, additional special
features are required. The addendum of the rack-type cutter must be equal to the
dedendum of the wheel to be generated, and since this exceeds the addendum of the
mating gear, the corners (or tips) of the tool teeth can with advantage be rounded by
a radius qfP . Incidentally, it may also be useful to ease the tips of the generated teeth
by providing a corresponding thickening near the root of the rack-type cutter tooth.
However, since the tip easing or tip relief does not affect the issue of interest here,
we omit the discussion relating thereto.
As we mentioned in Sect. 3.2, the rounded corner of the rack-type cutter tooth,
having radius of curvature qfP , does not give rise to cutting interference. Therefore,
everything works as if the addendum line of the rack-type cutter was the straight
line parallel to the axis of symmetry passing through the point, L, of connection
between the straight involute of the flank profile and the rounded profile of this
corner, i.e. as if the rack-type cutter ended with a sharp edge at point L.
Figure 4.12 shows the tip geometry of the teeth of a rack-type cutter without
undercut, in accordance with the standard basic rack of ISO 53: 1998 (E). The
rounded corner of the tooth begins at point L, whose distance from the cutter
datum-line defines the cutter active addendum ha0 of the rack-type cutter, and ends
at point M, which is located on the cutter tip line. The center of curvature, I, of this
rounded corner is located at distances qfP and E respectively from the cutter tip line
and axis of symmetry of the tooth, which bisects the thickness of the latter. The
distance E is defined by (see ISO 6336-3 [17]):
pm0 qfP
E¼ hfP tan a0 ð1 sin a0 Þ : ð4:23Þ
4 cos a0
164 4 Interference Between External Spur Gears
The position of point L is then uniquely identified, once that m0 , a0 , hfP and qfP
are known. Consequently, also the effective addendum line or effective tip line of the
rack-type cutter is identified. It is noteworthy that, for most rack-type cutters and
hobs, the values of hfP and qfP are chosen so that the active addendum ha0 is slightly
larger than m0 . However, here, we assume that ha0 is substantially equal to m0 .
Now we consider the relative motion between the rack-type cutter and the wheel
to be generated, and assume that, in the cutting conditions, the cutting pitch line
coincides with the cutter datum line. When this pitch line rolls without sliding on
the pitch circle of the wheel to be cut, the point L, which is rigidly connected to this
pitch line, but outside of it, generates the fillet profile of the wheel. This profile is a
planar curve, a trochoid (so it was called by Gilles Personne de Roberval, see Auger
[1], Jullien [18], Walker [36]) or cycloid.
We remember that a trochoid is the curve described by a point rigidly connected
to a circle as it rolls without sliding along a planar curve, and becomes a cycloid
when this planar curve is a straight line. Then the trochoid family is more general
than the cycloid family. In our case, however, trochoid and cycloid coincide, so we
will use the terms as synonyms.
To define the parametric equations of that particular trochoid, which is the
cycloid, consider the circle of radius r and center O1 , in the initial contact at point
O with a straight-line tangent to it, and that rolls without sliding on the same
straight line (Fig. 4.13). Assume, as Cartesian coordinate system, the system
Oðx; yÞ, with origin at point O, the x-axis coinciding with the straight line on which
4.7 Fillet Profile Generated by a Rack-Type … 165
the circle is rolling, and positive in the direction from left to right, and y-axis
perpendicular to the x-axis, and positive in the direction that goes from the origin
O to the center O1 .
Then consider a point A, located on the y-axis, in an intermediate position
between points O1 and O, or coincident with point O, or beyond point O, that is, on
the negative y-semiaxis, and denote by k its distance from the center O1 of the
circle. In the chosen coordinate system, the initial coordinates of point A are thus
½O; ðr kÞ, with k taken as algebraic value.
During the rolling of the circle on the x-axis, starting from the initial position
detected by the point of contact O, depending on whether point A lies inside the
circle ðk\r Þ, on its circumference ðk ¼ r Þ, or outside the circle ðk [ r Þ, the tro-
choid (alias cycloid) is described as being shortened (or curtate, or contracted),
ordinary (or conventional or common), or extended (or prolate) respectively.
The trochoid, described by the tracing point A, is a periodic and transcendent
planar curve, made up of an infinite number of arcs between them equal to one
another, each corresponding to one complete revolution of the circle rolling on the
x-axis. At the end points of each complete arc, i.e. the points of abscissa 2pn, with
n an integer, the ordinary cycloid has many cusps, while the extended cycloid has
many loops, but the one and the other have no inflection points. Instead, the
shortened cycloid has no multiple points, as the extended cycloid, but presents
infinite inflection points at points of abscissa ð2n þ 1Þp=2 (Fig. 4.13). The y-axis
can be considered as an axis of symmetry of the two branches of the cycloid in the
right and left of it, whatever the type of cycloid (shortened, ordinary, or extended).
From the aforementioned considerations, we infer the following parametric
equations of the trochoid/cycloid.
166 4 Interference Between External Spur Gears
x ¼ ru k sin u
ð4:24Þ
y ¼ r k cos u;
where u indicates the angle by which the initial radius OO1 ¼ r is rotated in the
clockwise direction (we recall that C is the instantaneous center of rotation, O is the
instantaneous center of rotation at the beginning of the rolling motion, and center
O1 moves parallel to x-axis), regardless of the actual position of the circle rolling on
the x-axis. A complete arc of cycloid is obtained from Eq. (4.24) by varying u from
O to 2p.
By eliminating u from the parametric Eq. (4.24), we obtain the following
Cartesian equation of trochoid:
r y pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
x ¼ r arccos ð k þ r yÞ ð k r þ yÞ : ð4:25Þ
k
Fig. 4.14 Polar reference system, and quantities to draw, point by point, the fillet profile
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
OT T 00 T PT 0 rP2 ðr ha0 Þ2
u¼ ¼ ¼ ¼ : ð4:28Þ
r r r r
From these relationships, we obtain the following equation, which expresses the
angle dP (in rad), as a function of the radius rP :
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rP2 ðr ha0 Þ2 rP2 ðr ha0 Þ2
dP ¼ arctan : ð4:29Þ
ðr ha0 Þ r
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
0:5 k0
rP ¼ : ð4:31Þ
2
It remains to be determined the radius rfil in the general condition in which the
cutting pitch line of the rack-type cutter does not coincide with its datum line
(Fig. 4.15). In this condition, the distance between the latter and the center O1 of
the wheel blank to be cut is equal to ðr1 þ xm0 Þ, where x is the addendum modi-
fication coefficient which will be discussed in Chap. 6. Therefore, the distance
between point L (Fig. 4.12) and the cutting pitch line of the rack-type cutter is equal
to ðha0 xm0 Þ.
The involute part of the tooth profile of the wheel to be generated is cut by the
straight profile part of the cutter tooth. Then the end point Pf (Fig. 4.10) of the
wheel tooth involute is cut by point L on the cutter profile. This happens when
point L, during the relative motion of cutting process (the path followed by point
L of the cutter during this motion coincides with the active addendum line), is to
coincide with point L0 , where the line of action in the cutting condition and active
addendum line of the cutter intersect. Thus point L0 is the start point of the path of
contact during the cutting meshing. Then, the fillet circle of the wheel is the circle
having center O1 and radius rf ¼ rfil ¼ O1 L0 (Fig. 4.15); therefore, we have (see
also Colbourne [3]):
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2ffi
ð h a0 rm 0 Þ
rfil ¼ rb1
2 þ r tan a
b1 0 : ð4:32Þ
sin a0
Fig. 4.15 Meshing diagram between the rack-type cutter and wheel to be cut
4.8 Fillet Profile Generated by a Pinion-Type Cutter 169
The generation process of gears with pinion-type cutter, first conceived by E.R.
Fellows (see Galassini [9], Radzevich [30]), is a gear cutting method by shaping,
whereby a reciprocating gear-like tool (this is the pinion-type cutter or shaper,
which is itself a toothed wheel with hardened cutting edges), and a gear blank rotate
together as if they were a gear pair. Of course, with respect to a common gear pair,
the pinion-type cutter of such a gear pair is equipped with a cutting reciprocating
motion in the direction of the axes, which is typical of the all-purpose gear shapers.
Suitably combining together the cutting reciprocating motion of the pinion-type
cutter, and the rotation motions of the gear blank and pinion-type cutter about their
axes, after a certain number of strokes for which the cutting operation is completed,
we get a toothing geometry of the gear wheel so cut, which is the envelope of the
successive positions of the pinion-type cutter in the said relative motion. Here we
are interested in studying the tooth fillet profile of the wheel, that is, the profile at
the tooth root, the one that connects the involute part of the tooth profile with the
dedendum circle. Since the tooth tip of the pinion-type cutter cuts this tooth fillet, it
is first necessary to define the tooth tip geometry of the cutting tool.
Figure 4.16 shows the tip geometry of a pinion-type cutter with a rounded corner
at the tooth tip. The transverse section of this rounded corner is a circular arc, with
radius qfP and center I. The rounded corner of the tooth begins at point L (this point is
the upper end of the involute part of the cutter tooth profile), whose distance from the
cutter reference circle or datum circle defines the active addendum ha0 of the
pinion-type cutter, and ends at point M, which is located on the cutter tip circle (or
cutter addendum circle), having radius ra0 . The circular arc LM has a common
tangent with the involute profile at point L, and with the addendum circle at point M.
The polar angle uI between the straight line O0 I and the x-axis is given by the
difference
u I ¼ aI u R ; ð4:34Þ
d
where aI is the angle IO 0 T (not shown in Fig. 4.16) and uR is the angle defining the
polar axis position, i.e. the angle between the x-axis and the straight line O0 T.
Considering the triangle IO0 T, we can write the following relationship which
allows to calculate the angle aI :
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rb0 tan aI ¼ rI2 rb0 2 ; ð4:35Þ
where rb0 is the radius of the base circle of the pinion-type cutter.
To uniquely define the polar coordinate of point I, it remains to calculate the
angle uR . To do this, however, it is convenient to first define the polar coordinates
of point L. The normal to the profile at point L passes through the center of
curvature I and is tangent to the base circle at point T. The length of segment TL is
equal to rb0 tan aL , where aL is the pressure angle at point L, that is the angle
between the straight line O0 L and the tangent to the profile at point L. Therefore,
considering the triangle LO0 T and noting that LT ¼ LI þ IT, we get the following
relationship that allows us to calculate aL :
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rb0 tan aL ¼ qfP þ rI2 rb0 2 : ð4:36Þ
The radius of the effective addendum circle of the pinion-type cutter, i.e. the
distance rL from the origin O0 to point L, is given by:
rb0
rL ¼ : ð4:37Þ
cos aL
4.8 Fillet Profile Generated by a Pinion-Type Cutter 171
The angle uL between the straight line O0 L and the x-axis (Fig. 4.16), according
to the considerations made in Sect. 2.5 (see also Fig. 2.12), can be written in the
form:
s0
uL ¼ ðinvaL inva0 Þ; ð4:38Þ
2r0
where s0 and a0 are the nominal standard values of the tooth thickness and pressure
angle at the standard pitch circle having radius r0 . The angle uL is then uniquely
defined, when the thickness s0 is known.
The last unknown quantities, uR , which appears in Eq. (4.34), is immediately
defined as the difference:
uR ¼ aL uL : ð4:39Þ
Therefore, instead of using Eq. (4.34), we can calculate uI with the following
relationship:
yI
uI ¼ arctan : ð4:42Þ
xI
In some cases, the tooth profile of the pinion-type cutter does not have a rounded
corner at the tip, but the involute profile develops up to the tip circle, so connection
between the tooth profile and addendum circle configures a sharp corner. In these
cases, we can use the same previous relationships, with the caveat to put in them qfP
equal to zero, and rI ¼ rL ¼ ra0 :
The pinion-type cutter shown in Fig. 4.16 is designed in such a way that, in the
Cartesian coordinate system indicated therein, point I has a positive ordinate, i.e.
yI [ 0. With this measure, the circular corner at the tip of profile smoothly connects
with the tip circle. Otherwise, when yI \0, the tooth tip of the pinion-type cutter
would be slightly pointed, as a consequence of a too large value of the radius qfP .
In the design of a pinion-type cutter, sometimes we have advantages if we take
the values of qfP as high as possible, especially when the pinion-type cutter is to be
used for cutting internal gears. In fact, a high value of qfP leads to a larger radius of
fillet of the wheel, and therefore to a lower stress concentration to the tooth root.
The highest value of qfP that can be used is that for which, in Eq. (4.41), yI ¼ 0
(then I is located on the tooth centerline, i.e. on the x-axis), but the calculation
method of this value is not simple. Generally, we proceed by trial, choosing a value
of qfP , and then verifying that yI is positive.
172 4 Interference Between External Spur Gears
We thus uniquely defined the geometry of the tooth tip of the pinion-type cutter,
including its active addendum ha0 . Now we consider the relative motion between
this pinion-type cutter and the wheel to be generated, and assume that, in the cutting
conditions, the cutting pitch circle coincides with the cutter standard datum circle.
When this pitch circle rolls without sliding on the pitch circle of the wheel to be cut,
the point L, which is rigidly connected to this pitch circle and is outside of it,
generates the fillet profile of the wheel, which is still a trochoid, in the most general
meaning of the term, and, more specifically, an extended (or prolate) epicycloid.
We define as epicycloid or hypocycloid the planar curves described by a tracing
point rigidly connected to a mobile circle with radius r2 as it rolls without sliding on
a fixed circle having radius r1 , depending on whether the mobile circle is tangent
externally or internally to the fixed circle. The epicycloids and hypocycloids are
described as being shortened (or curtate), ordinary (or conventional or common), or
extended (or prolate), depending on whether the tracing point L lies inside the
mobile circle ðk\r2 Þ, on its circumference ðk ¼ r2 Þ, or outside the mobile circle
ðk [ r2 Þ. The quantities k and r2 are respectively the distance of the tracing point
from the center of the mobile circle, and the radius of the latter.
In addition, we remember that the distinction between epicycloid and hypocy-
cloid is, in general, only formal as the Bernoulli-Euler-Goldbach double generation
theorem (see Hall [13]) clearly shows that any epitrochoid (and then any epicy-
cloid) can be expressed or generated as a hypotrochoid (and thus as a hypocycloid)
and vice versa, but with some limitation, a hypocycloid can be generated as an
epicycloid (in this second case, if and only if its rolling circle is larger than the fixed
circle).
To define the parametric equations of the epicycloid, consider a rolling circle of
radius r2 and center O2 , in contact at a point C with a fixed circle of radius r1 and
center O1 , with O1 , C and O2 aligned along a vertical line, and on it located one
over the other, in such a way that we have O1 C ¼ r1 , and CO2 ¼ r2 (Fig. 4.17).
Assume, as Cartesian coordinate system, the system O1 ðx; yÞ, with origin at point
O1 , the y-axis coinciding with the straight line O1 O2 and positive in the direction
from point O1 to point O2 , and the x-axis normal to the y-axis and positive in the
direction from left to right. Then consider a point P0 , located on the y-axis at
distance k [ r2 from point O2 , and rigidly connected to the rolling circle; in this
chosen coordinate system, the initial coordinates of point P ¼ P0 are thus
½0; ðr1 þ r2 k Þ.
During the rolling without sliding of the rolling circle on the fixed circle from the
initial position detected by the point of contact C, the tracing point P describes an
extended epicycloid (we focus only on this, because it is the one that interests us
here), which consists of two symmetrical branches, with respect to the y-axis
(Fig. 4.17). Generalizing the problem, in order to extend it to the case of internal
gear pairs (in this case, in fact, it is interesting to consider the hypocycloid),
according to the considerations mentioned above, we infer the following parametric
equations of the epicycloid or hypocycloid (see Gibson [10], Rutter [32]):
4.8 Fillet Profile Generated by a Pinion-Type Cutter 173
r1 r2
x ¼ ðr1 r2 Þ sin u k sin ur2
ð4:43Þ
r1 r2
y ¼ ðr1 r2 Þ cos u k cos r2 u;
involute part of the cutter tooth profile. Then the end point Pf of the wheel tooth
involute is cut by point L on the pinion cutter profile. This happens when point L,
during the relative motion of the cutting process (the path followed by point L of the
pinion-type cutter during this motion coincides with the active addendum circle), is
to coincide with point L0 , where the line of action in the cutting condition and active
addendum circle of the cutter intersect. Point L0 is then the start point of the path of
contact during the cutting meshing.
Therefore, the fillet circle of the wheel is the circle having center O1 and radius
rf ¼ rfil ¼ O1 L0 (Fig. 4.18) given by:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2
rfil ¼ rb1
2 þ ðr þ r Þ tan a
b0 b1 0 rL2 rb0 2 : ð4:44Þ
Of course, in this equation the radius rL must be replaced with the radius ra0 (see
Fig. 4.16) in the case for which the pinion-type cutter does not have rounding at the
tooth tips.
4.9 Interference Effects of the Rounded Tip of the Cutter … 175
The tooth profile of most rack-type cutters, hobs and pinion-type cutters is rounded
at the tip. The question is therefore legitimate to ask in relation to the effects of such
rounded tips of the teeth of the cutting tools on interference in the cutting and
operating conditions (see Polder and Broekhuisen [27]). In this respect, the cut with
rack-type cutter (and the equivalent thereto with hob), and cutting with pinion-type
cutter are to be evaluated in a different way.
When the cut is made with a rack-type cutter (the same considerations, with
adequate explanations that follow, are basically valid also when the cut is made
with hobs), we can show that the rounded tip of the cutter tooth does not cause a
cutting interference. To prove this fact, just compare the trochoid described by the
tracing point L (see Fig. 4.12), and the curve obtained as the envelope of infinite
circles of radius qfP and centers on the trochoid described by the tracing point
I (Fig. 4.12), which is the center of curvature of the rounded tip of the cutter tooth.
Thus, we demonstrate, unequivocally, that the trochoid described by point L and the
envelope curve associated with the aforementioned infinite circles of radius qfP
and trochoid described by point I intersect the involute generated from the straight
part of the rack-type cutter profile at the same point, and then determine the same
fillet circle having radius rfil .
Therefore, in conclusion, as regards the problem of interference during the
cutting with rack-type cutter, we will consider only the point L (Fig. 4.12), where
the involute straight profile ends and the rounded tip begins, and proceed as if the
cutting tool were truncated with a sharp edge at point L. Of course, all other
conditions being equal, the cutting interference is much higher, the lower the
pressure angle. Figure 4.19, taken from Henriot [15], but reworked in the form that
is of interest here, shows the tooth profiles of a pinion with standard sizing, and
without profile shift, with 15 teeth, cut with four rack-type cutters, with rounded
tips all characterized by the same geometry, but with pressure angles respectively
equal to 15°, 20°, 25° and 30°. The four trochoids traced by point L are virtually
overlapped, and then the four profiles of the root fillets practically coincide, but
with a0 ¼ 25 and a0 ¼ 30 the cutting interference does not occur, while for
a0 ¼ 15 and a0 ¼ 20 cutting interference and undercut occur. The same figure
highlights the fact that the trochoid TL traced by point L and the envelope curve
T generated from the rounded tip of the four cutting tools intersect in the same point
V the involute profile of the tooth that is generated.
Quite similar considerations can be made when hobs are used for the gear wheel
cutting. As we have already mentioned in Sect. 3.1.3, few additional problem can
occur in cases where, for special achievements (see, for example, the cutting by
hobs for large wheels), hobs with rounded tips having radii of curvature more
extensive than those corresponding to the rack-type cutter are used. In these cases,
176 4 Interference Between External Spur Gears
Fig. 4.19 Tooth profiles of a pinion with 15 teeth and standard sizing, generated by four rack-type
cutters having different pressure angles, but rounded tips with the same geometry
effects of cutting interference of these rounded tips with the tooth profiles of the
wheels to be generated cannot be excluded. In these cases, it is necessary to resort to
the procedure scribed in the aforementioned Sect. 3.1.3.
We have already said that the interference in the actual operating conditions
cannot be never accepted, and must be categorically excluded. The condition to be
met to exclude the interference in the operating conditions is given by the inequality
(4.18), or better still by the inequality (4.21). Well, when the cutting of the two
toothed wheels of a gear pair is made using rack-type cutters or hobs, these
inequalities are always verified.
These inequalities instead should be carefully verified in the case of a wheel cut
using a pinion-type cutter, the latter with a low number of teeth, intended to mesh,
under real operating conditions, with another wheel having a high number of teeth.
In practice of design, the pinions generated in condition of cutting interference are
very few. Then the bottom point Pf (Fig. 4.10) of the involute tooth profile is
external to the base circle ðrf [ rb Þ.
However, if a pinion like this (that is, without cutting interference), under real
operating conditions, meshes with a toothed gear wheel with a high number of teeth
(the limit case is obviously that of a rack, and therefore a rack-pinion pair), the tooth
tip of the member with high number of teeth interferes with the root fillet of the
member with less number of teeth, even before the theoretical interference occurs.
Figure 4.20, taken from Henriot [15], but also reworked in the form that is of
interest here, shows this state of things for a pinion with 40 teeth and standard
sizing, cut using a pinion-type cutter with 20 teeth, ha0 ¼ 1:25m0 , hf 0 ¼ m0 , and
4.9 Interference Effects of the Rounded Tip of the Cutter … 177
pressure angle in the cutting conditions a0 ¼ 20 , and meshing, in the operating
conditions, with a rack, that is, with a wheel having infinite number of teeth. The
figure highlights the fact that rc \rf , albeit slightly. The inequality (4.18) is
therefore not satisfied, and it follows a small interference of the tooth tip of the rack
with the root fillet of the pinion, also highlighted in Fig. 4.20, where the trochoid
Tr , described by the tooth tip A0 of the rack, penetrates inside the trochoid TP ,
described by the tooth tip A00 of the pinion-type cutter.
A fillet profile of a given wheel of center O1 , defined with geometric law which
ensures non-interference with the addendum flank profile to the mating gear, having
center O2 , is that proposed for the first time by Schiebel in 1912 (see also Schiebel
and Lindner [34]). This fillet profile is defined as an extended epicycloid of point
P (Fig. 4.21), in which the involute tooth profile ends, and the fillet profile begins,
when an auxiliary circle, of radius CO0 and center O0 , rolls without sliding on the
pitch circle of the wheel to be generated. The center O0 of this auxiliary circle (it is
nothing else then the pitch circle of the pinion-type cutter to be used for cutting) is
located on the line of centers O1 O2 . It is identified by the intersection between this
line of centers and the axis of the segment AB, where A is the end point of the path
of contact, and B is the intersection between the line of centers O1 O2 and the
dedendum circle (Fig. 4.21).
To demonstrate that this extended epicycloid does not cause interference, just to
show that the path described by point P, supposed rigidly connected with the aux-
iliary circle, during rolling without sliding of the latter on the pitch circle of the
wheel with center O1 , is at the outside of the path described by the same point P, this
178 4 Interference Between External Spur Gears
Fig. 4.21 Fillet profile in the shape of prolate epicycloid, according Schiebel
time supposed rigidly connected with the pitch circle of the wheel with center O2 ,
when the latter rolls without sliding on the pitch circle of the wheel with center O1 .
Since both the epicycloids are extended epicycloids described by the same point
P, during pure rolling of two different circles on the same fixed circle, due to well
know properties of epicycloids, it is sufficient to prove that CO0 \CO2 , i.e. that the
radius CO0 of the pitch circle of the pinion-type cutter is smaller than the radius
CO2 of the pitch circle of the wheel that will mesh with the one considered here. In
particular, by the Euler-Savary equation (see Dooner [4]), it is easy to verify that, at
point of contact P between the two epicycloids, the condition CO0 \CO2 deter-
mines, at point P, a radius of curvature of the epicycloid related to the pitch circle of
the pinion-type cutter smaller than the radius of curvature of the epicycloid related
to the pitch circle of the mating wheel.
4.9 Interference Effects of the Rounded Tip of the Cutter … 179
However, the inequality CO0 \CO2 is immediately verified, by noting that the
distance of point B from the center O2 is greater than the distance of point A0 (this
point is obtained as the intersection of the line of center O1 O2 with the circle of
radius AO2 and center O2 , which is the addendum circle of the mating wheel) from
the same center O2 . This is because a sufficient bottom clearance A0 B must exist
between the dedendum circle of the wheel under consideration, and the addendum
circle of mating wheel. Therefore, the axis of the segment AB intersects the line of
centers O1 O2 at point O0 , the distance of which from point C will be smaller than
the distance, from the same point C, of point O2 where the axis of the circular arc
AA0 intersects the same line of centers O1 O2 .
If we draw the tooth profile in the position in which it passes through point
C (Fig. 4.21), and we want the epicycloid may link up at point P with the involute
tooth profile, with a common tangent, its center of curvature should be clearly on
the normal PN to the tooth profile at point P. However, also the center of instan-
taneous rotation, during rolling of the auxiliary circle on the pitch circle of the
wheel, must be on the same normal. Consequently, the auxiliary circle must be
tangent to the pitch circle at point N, when point P is located in the position shown
in Fig. 4.21.
The apex of the epicycloid, i.e. its point of tangency V with the dedendum circle,
will have when the auxiliary circle will have rotated, with respect to the position for
which the contact was at point N, of a circular arc NM such that point P, rigidly
connected with it, is located on the straight line joining point O1 to the center of the
rolling auxiliary circle. Since the segment AC is none other than the segment PN,
when the point P of the latter has moved to point A, the circular arc MN we seek is
equal to the circular arc KC shown in Fig. 4.21.
Therefore, reporting from the point N a circular arc NM ¼ KC, we identify the
point M and, then, the apex V of the epicycloid as the intersection of the straight line
O1 M with the dedendum circle of the wheel. Finally, the epicycloidal profile
between the end points P and V is then completed using the analytical relations
given in the previous section.
References
Abstract In this chapter, the problems of interference between internal spur gears
in both cutting and working conditions are first described and discussed. The dis-
cussion is quite general and therefore includes the five types of interference that
may occur in this case, i.e.: theoretical or primary interference; secondary inter-
ference; trimming interference during the radial approach of the pinion when the
internal gear pair is assembled, and of the pinion-type cutter in the conditions of the
generation cutting process; fillet interference between the tip of the pinion and
the root fillet of the annulus; fillet interference between the tip of the annulus and
the root fillet of the pinion. For this purpose, analytical methods are used, without
forgetting well-known traditional methods which, although approximate, provide
reliable results of great engineering-design value. Finally, the condition for
avoiding rubbing during the annulus cutting process is discussed and defined.
5.1 Introduction
To transmit the motion between two parallel shafts, which rotate in the same
direction, internal gear pairs are used. An internal spur gear pair is constituted by a
cylindrical pinion with external toothing and a cylindrical internal gear (also called
ring gear, annulus gear or more simply annulus) with internal toothing. The
geometry of an internal gear is very similar to that of an external gear. Therefore,
mutatis mutandis, we can extend to the internal spur gears the same concepts
described in the previous chapter, for external spur gears (see Buckingham [3],
Dudley [7], Merritt [14], Radzevich [20], Schreier [22]). For this reason, we will
focus our attention only on those aspects that differentiate the internal gears from
external gears (see Tuplin [23], Polder [15–17], Litvin et al. [12], Litvin and
Fuentes [11]).
Since one of the two members of an internal gear pair is an external gear wheel
(the pinion), from the geometrical point of view, the first difference between an
internal gear pair and an external gear pair concerns only the annulus. In fact, the
teeth of this member of the internal gear pair lie outside the profiles, while those of
an external gear wheel lie inside them. In other words, the teeth of an annulus have
exactly the same shape of the tooth spaces in an external gear wheel. Furthermore,
the addendum circle of an annulus lies inside the reference pitch circle, while the
dedendum circle lies outside it (see Pollone [19], Henriot [9]).
Consider now a ring gear with involute tooth profiles and nominal or standard
sizing (i.e., with ha2 ¼ m, and hf 2 ¼ 1:25 m), having module m, pressure angle a,
and number of teeth z2 (as usual, the subscript 2 denotes the ring gear, and the
subscript 1 denotes the pinion). The radius rb2 of its base circle is given by:
mz2
rb2 ¼ cos a; ð5:1Þ
2
while its tooth profiles will have the shape of the space-widths between the teeth of
the pinion, whose number of teeth is z1 .
The nominal or standard values of the diameter of the pitch circle and center distance
are the ones given respectively by the first of Eqs. (2.63) and (2.65), which here
are rewritten with the specific subscripts to standard symbols (see ISO 6336-1 [10]):
d2 ¼ 2r2 ¼ z2 m; ð5:2Þ
z2 z1
a ¼ r2 r1 ¼ m: ð5:3Þ
2
In the case of standard sizing, the diameters of the addendum circle (also called
tip circle or outside circle), and dedendum circle (or root circle) are those respec-
tively given by the second and third of Eqs. (2.63), which here are also rewritten
with the specific subscripts to standard symbols:
da2 ¼ d2 2m df 2 ¼ d2 þ 2; 5m: ð5:4Þ
However, it is to be noted that, for the internal spur gears, special sizings are
often adopted. We will make mention of these special sizings, other than the
standard sizing, as the opportunity presents itself.
For an internal gear pair, we have the following relationships, with the usual
meaning of the symbols:
x1 r1 ¼ x2 r2
r1
r2 ¼ zz12 ¼ x
x1 ¼ rb2
2 rb1
r1 ¼ z2zz
1a
1
r2 ¼ z2zz
2a
ð5:5Þ
1
pb1 ¼ 2pr
z1 ¼ pb2 ¼ z2
b1 2prb2
p1 ¼ 2pr
z1 ¼ p2 ¼ z2 ¼ pm
1 2pr2
rb1 ¼ r1 cosa
rb2 ¼ r2 cosa:
5.1 Introduction 183
Fig. 5.1 Contact between profiles of an internal spur gear pair within the path of contact, and
outside of the path of contact
The two pitch circles are tangent at the instantaneous center of rotation or pitch
point C, and the line of action is the tangent to the two base circles at points T1 and
T2 . As Fig. 5.1 shows, even the two profiles p1 and p2 are touching at pitch point
C. At this point, the center of curvature of the profile p1 is T1 , while the center of
curvature of the profile p2 is T2 . The two centers of curvature T1 and T2 are located
on the line of action on the same side with respect to pitch point C. The angle a
between this line of action and the common tangent to the pitch circles at pitch
point C is the operating pressure angle (i.e. the pressure angle at the pitch cylinders)
of the internal gear pair. Obviously, the lines O1 T1 and O2 T2 are parallel and both
perpendicular to the line of action. Therefore, each of them makes the angle a with
the line of centers O1 O2 (outside of Fig. 5.1, downward).
By extension of the concepts that we have described for the external spur gears,
we can say that we will have a conjugate contact only if the point of contact
between the profiles p1 and p2 will be between points E and T1 . Point E is the point
of intersection of the line of action with the addendum circle of the pinion, located
in the upper right out of the figure. Instead, there will be non-conjugate contact if
the point of contact will be between points T1 and T2 , which are the points of
interference of the pinion and ring gear, respectively.
Therefore, contact cannot go beyond the point T1 , in the direction from point T1
to point T2 . Instead, in the direction from point T1 to point C, contact can go, at least
from a theoretical point of view, up to infinity. This is why, in principle, the
involute profiles of an annulus may extend to any radius, and conjugate contact is
5.2 Theoretical Interference in the Internal Spur Gears 185
theoretically possible, however large is the radius of the addendum circle of the
pinion. Actually, beyond the point C, the path of contact is limited by point E,
where the addendum circle of the pinion intersects the line of action.
During the engagement, the point of contact between the conjugate profiles
moves on the line of action. As long as the point of contact is located between
points E and T1 , we have a conjugate contact, and the meshing has a correct
kinematics. When the point of contact reaches the point T1 , it comes to coincide
with the point of regression P01 of the involute profile p1 . Point T1 , i.e. the point of
interference of the pinion, represents the limit position of the point of contact,
beyond which the contact is non-conjugate, and theoretical interference occurs.
Consider now a point of contact P on the line of action, between points T1 and
T2 . In this point, which is a point of non-conjugate contact, the center of curvature
of the profile p2 is T2 . The center of curvature of the profile which is conjugate of
the profile p2 is always the point T1 . Thus, this conjugate profile has a curvature in
the opposite direction compared to the one of the profile, p2 . Therefore, the profile
that is conjugate of the profile p2 ; is no longer the profile p1 , but the profile p01 , i.e.
the fictitious branch of the profile p1 symmetric with respect to the line O1 P01 ,
where P01 is the origin of the two branches p01 and p1 of the involute curve on the
base circle of the pinion.
Therefore, the two actual profiles p1 and p2 intersect, and thus interference
occurs, because the pinion dedendum profile near its root goes to undermine the
annulus addendum profile near its tip. To avoid this type of interference, i.e. the
theoretical interference, it is sufficient that the value of the radius ra2 of the
addendum circle of the annulus is greater or equal to O2 T1 , namely ra2 O2 T1 .
Therefore, considering the triangle O2 T1 T2 (Fig. 5.1), to avoid the primary inter-
ference, the following inequality must be satisfied:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2 2 þ ½ðr r Þsina2 ;
ra2 O2 T1 ¼ O2 T2 þ T1 T2 ¼ rb2 2 1 ð5:6Þ
Fig. 5.2 Meshing diagram and contact between two conjugate profiles of an internal gear pair
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m 2ffi
h2amax z2 z2 2z1 z2 z1 sin a ;
2 2 ð5:7Þ
2
from which we obtain that, for a given addendum h2a of the annulus, and for a
given ratio z1 =z2 , the following relationship must be satisfied:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
1 þ 1 z2 2 z2 sin a
z1 z1 2
addendum h2a ¼ k2 m of the annulus are given. In the case in which the teeth have
nominal standard sizing, i.e. when k2 ¼ 1 and h2a ¼ m, the inequality (5.8) takes
the form:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uþ u2 ð2u 1Þsin2 a
z1 ¼ z1min 2 : ð5:9Þ
ð2u 1Þsin2 a
The minimum values of z1min are shown in the lower part of the diagram of
Fig. 3.1. From this diagram we infer that the value of z1min , for given values of the
operating pressure angle a and ratio ðz1 =z2 Þ ¼ ð1=uÞ; is much greater for internal
gear pairs than the one for external gear pairs.
In order to use pinions with a limited number of teeth, it is necessary to choose
pressure angles a higher than those chosen for external gear pairs, or to use stub
teeth since, as the inequality (5.8) shows, the minimum number of teeth z1min varies
proportionally to the addendum factor of the annulus k2 ¼ h2a =m:
It should be noted that, generally, the solution that provides addendum modi-
fication (with consequent shifting of the addendum and dedendum circles) is not
convenient because, due to the addendum modification, it is very likely that the
secondary interference can manifest (see next section).
In the case where the solution would be to limit the addendum h2a of the
annulus, and the pressure angle a is equal to 20°, the diagram shown in Fig. 5.3 can
Fig. 5.3 Addendum factor k2 of the annulus as a function of z1 and z2 , for a ¼ 20
188 5 Interference Between Internal Spur Gears
be used (see also Henriot [9]). It allows to obtain the value of the addendum factor
k2 of the annulus as a function of the numbers of teeth z1 and z2 , when a ¼ 20 .
Example With z2 ¼ 60 and z1 ¼ 20, we can choose k2 ¼ 1; i.e. a nominal standard
sizing. Instead, with z2 ¼ 100 and z1 ¼ 15, the value of k2 ¼ 0:80 cannot be
overcome. In this last case, it is necessary to adopt a stub gearing or an addendum
modification factor x2 ¼ 0:20 (see next Chapter).
From inequality (5.7) we infer that, to avoid primary interference between an
annulus having z2 teeth and a pinion with z1 teeth, the following relationship must
be satisfied (see also Pollone [19]):
which, in the case of nominal standard sizing, i.e. for k2 ¼ 1 and h2a ¼ m, becomes:
z21 sin2 a 4
z2 ¼ z2min : ð5:11Þ
2z1 sin2 a 4
The diagram shown in Fig. 5.4 allows to calculate, for a nominal standard sizing,
the minimum number of teeth z2min of the annulus as a function of the number of
teeth z1 of the pinion, for five values of the pressure angle a, in the range between
a ¼ 20 and a ¼ 30 .
Example For z1 ¼ 15 and a ¼ 25 , we have z2min ¼ 27.
The different curves of this diagram are limited, in the lower right, to values of
z2min for which the difference ðz2 z1 Þ is the minimum sufficient to avoid the
secondary interference, which constitutes the subject of the next section.
In the case where the diameter of the pinion is quite close to that of the annulus, in
addition to the primary interference between the active flank profiles as those shown
in Fig. 5.2, another type of interference can occur in the internal spur gears. This
very special interference, typical of the internal gears, is called secondary inter-
ference, and does not occur in the external gears. In fact, when the numbers of teeth
of the annulus, z2 , and pinion, z1 , do not differ much from each other, the tips of the
teeth can interfere in the surroundings of the point I (see Fig. 5.2), giving rise to the
secondary interference, which is also called fouling [13].
To understand how this type of interference occurs, let’s examine what happens
when the point of contact passes the other side of the pitch point C compared to the
one where the interference points T1 and T2 are located. On this side of the line of
action T1 T2 , the limit position of the end point of the path of contact is identified by
5.3 Secondary Interference in the Internal Spur Gears 189
Fig. 5.4 Minimum number of teeth z2min of the annulus as a function of the number of teeth z1 of
the pinion, for five values of the pressure angle a
190 5 Interference Between Internal Spur Gears
point E, obtained as the intersection of the pinion addendum circle with the line of
action. Beyond this point, the flank profiles of the teeth are in contact only along
their extensions, at the point R on the line of action (see Fig. 5.5).
Let’s consider now, in addition to the line of action T1 T2 which is tangent to the
base circles at point T1 and T2 , the second line of action T10 T20 , which is tangent to
the same base circles at points T10 and T20 . These two lines of action are obviously
symmetrical with respect to the common tangent to the pitch circles at the pitch
point C. The fact that the second line of action T10 T20 is normal to the two profiles p1
and p2 at the points P1 and P2 is equally evident.
As it can be shown (see Henriot [9]), the distance P1 P2 , given by the following
relationship:
is a constant, for which the profiles p1 and p2 can never touch on this line of action
or above it.
However, during the meshing cycle, it may happen that the tip V1 of the
addendum profile of the pinion matches the tooth flank profile of the annulus at a
certain point, as Fig. 5.5 shows. Evidently, in this point the two profiles p1 and p2
cannot be tangent. Starting from this instant, the tooth tip V1 of the pinion notches
that of the annulus, with consequent removal of a tooth portion of the latter. This
removed portion of the annulus tooth is bounded by the extended hypocycloid
described by the tip V1 of the pinion tooth during the relative motion of rolling
without sliding of two pitch circles on each other, as Fig. 5.5 highlights by a dashed
line. This is the secondary interference or fouling.
Fig. 5.5 Secondary interference or fouling: contact between the tip V 1 of the pinion tooth and the
tooth flank of the annulus
5.3 Secondary Interference in the Internal Spur Gears 191
Figure 5.6 shows how and to what extent this secondary interference alters the
tip profile of the annulus. The portion of the annulus profile corresponding to
triangles ABC is removed by the action exerted by the tip of the pinion tooth.
To remove the risk of secondary interference, it is necessary to verify that the
extended hypocycloid described by point V1 during this relative motion does not
intersect the profile of the tooth flank of the annulus. To this end, known
geometric-analytical methods may be used, based on the parametric equations of
the hypocycloid given by relationships (4.43), or the Cartesian equation that is
derived by them. In the next section, we will describe one of these methods, the one
proposed by Colbourne [4].
Traditional methods developed by Buckingham [3], and Henriot [9] can also be
used, as an alternative to such geometric-analytical methods. It is not the case to
bring full the analytical developments made by these two researchers. Here we give
only the results obtained by Henriot [9], as they can easily be used to verify the fact
that this type of interference does not occur.
Henriot considers an internal gear pair in the limiting condition of secondary
interference, which occurs when the tip V1 of the addendum profile p1 of the pinion
touches, at a point P geometrically well-defined, the tip V2 of the addendum profile
p2 of the annulus (with reference to Fig. 5.5, this limiting condition would come to
have point P and V1 both coinciding with point V2 ). Based on geometric consid-
erations, Henriot leads to a system of algebraic equations, where the following
quantities appear:
• the numbers of teeth z1 and z2 ;
• the addendum factors k1 and k2 ;
• the angles d1 and d2 between the line of centers O1 O2 and the lines O1 P and
O2 P;
• the operating pressure angle a;
• the involutes invaa1 and invaa2 , where aa1 is the involute polar angle between
the line O1 P and O1 P01 , and aa2 is the involute polar angle between the line O2 P
and O2 P02 , where P01 and P02 are the origins of the involutes on the related base
circles.
Fig. 5.6 Alteration of the annulus tip profile due to secondary interference
192 5 Interference Between Internal Spur Gears
Fig. 5.7 Minimum difference ðz2 z1 Þmin to avoid the secondary interference, as a function of the
addendum factor k and operating pressure angle a
194 5 Interference Between Internal Spur Gears
Very often, to reduce the size of the gear drives for power transmission mechanical
systems, and thus their weight and their cost, the design conditions impose stringent
limits on the number of teeth of the pinion. In these cases, the design choices of the
internal gear pair, able to satisfy the input data, must take account of the need to
dispel the double risk of primary and secondary interferences. The following
examples show how to achieve the design requirements.
Example 1 We want to realize an internal gear pair with a gear ratio u ¼ z2 =z1 ¼
2; a pinion with z1 ¼ 15 teeth, and pinion and annulus having teeth with nominal
standard sizing, i.e. with addendum factor k ¼ k1 ¼ k2 ¼ 1:
From Fig. 5.4 we infer that, to avoid the primary interference, we need to choose
a pressure angle a ¼ 25 . Since this internal gear pair consists of a pinion with
z1 ¼ 15 teeth, and the annulus with z2 ¼ 30 teeth, the secondary interference does
not occur. In fact, Fig. 5.7 provides, for k ¼ ha1 =m ¼ ha2 =m ¼ 1 and a ¼ 25 , a
limiting value of difference ðz2 z1 Þmin equal to 5, when actually we have
ðz2 z1 Þ ¼ 15.
The secondary interference occurs for transmission ratios close to unity, that is
when the number of teeth z1 of the pinion becomes close to that of the annulus, z2 .
Example 2 We want to realize an internal gear pair with numbers of teeth z1 ¼ 55
and z2 ¼ 60 (therefore with gear ratio u ¼ z2 =z1 ¼ 60=55 ¼ 1:0909), and nominal
standard sizing of the teeth.
From Fig. 5.4 we infer that, to avoid the primary interference, it would be
enough to choose a pressure angle a ¼ 15 : However, since ðz2 z1 Þ ¼ 5, to avoid
the secondary interference, it is necessary to choose a pressure angle a ¼ 27 300
(see Fig. 5.7). If we want to avoid the primary and secondary interferences during
the generation cutting, to be made with pinion-type cutter, it is necessary to proceed
as the following example shows.
Example 3 We want to avoid the primary and secondary interferences in the
cutting conditions, with pinion-type cutter having number of teeth z0 ¼ 25, and
pressure angle a ¼ 20 .
From Fig. 5.4 we infer that the minimum number of teeth z2min of the annulus,
which can be generated without primary interference, is equal to 38 teeth. From
Fig. 5.7 we infer that, for a ¼ 20 and k ¼ k1 ¼ k2 ¼ 1; the minimum difference
ðz2 z1 Þmin is equal to 8. Therefore, since ðz2 z0 Þ ¼ 13, no secondary interfer-
ence occurs in the cutting conditions. If the pinion-type cutter had number of teeth
z0 ¼ 27, pressure angle a0 ¼ 20 and nominal standard sizing, the minimum
number of teeth of the annulus z2 , which can be generated without any of the two
types of interference (primary and secondary), is z2min ¼ 35. If, for a given pressure
angle, stub teeth are used, two goals may be simultaneously achieved. The first of
5.4 Possibility to Realize Internal Gear … 195
these goals is the reduction of the number of teeth of the pinion, z1 , suitable to avoid
the primary interference. The second goal is the decrease of the minimum limiting
value of the difference ðz2 z1 Þmin ; below which the secondary interference
manifests.
In Sect. 5.3 we described the general concepts regarding the secondary interference;
in addition, we summarized the approximate method proposed by Henriot [9] to
avoid it, highlighting its great design importance. In this section, we describe the
geometric-analytical method for checking of the secondary interference, proposed
by Colbourne [4], also using general concept concerning the kinematic geometry of
gearing (see Dooner [5], Dooner and Seireg [6]).
For this purpose, we must first define the angular positions of the pinion and
annulus with respect to the centerline O2 O1 of the internal gear pair, by means of a
constraint equation that links the various quantities involved. To this end, we
denote by:
• x1 and x2 , the tooth centerlines of the pinion and annulus;
• b1 and b2 , the angles that define the angular positions of the pinion and annulus,
both measured from the centerline O2 O1 clockwise to the x1 -axis and x2 -axis;
• s1 and s2 , the tooth thicknesses of the pinion and annulus at the pitch circles,
having radii r1 and r2 ; respectively.
Figure 5.8 shows an internal gear pair in the position where the contact point
coincides with the pitch point. Obviously, in this initial position of the point of
contact (the second subscript, i, refers to this initial position), b1;i is negative, while
b2;i is positive. These quantities are given respectively by:
From the first of Eqs. (5.5), as x1 ¼ db1 =dt and x2 ¼ db2 =dt, we get the fol-
lowing relationship:
196 5 Interference Between Internal Spur Gears
Fig. 5.8 Initial position of the two members of an internal gear pair, with contact at the pitch point
Therefore, multiplying both sides of Eqs. (5.14) and (5.15) for r1 and r2
respectively, with replacement, in Eqs. (5.14), of ðr2 =r1 Þdb2 instead of db1
resulting from Eq. (5.16), and subtracting side by side the Eq. (5.15) from
Eq. (5.14) so obtained, we get the following equation that correlates b1 with b2 :
1
r1 b1 r2 b2 þ ðs1 þ s2 Þ ¼ 0: ð5:17Þ
2
In Sect. 5.3 we have already highlighted the condition that, to avoid the sec-
ondary interference, it is necessary that the extended hypocycloid described by the
tooth tip V1 of the pinion during the meshing cycle must not intersect the profile of
the tooth flank of the annulus. This hypocycloid (also called hypotrochoid), which
is a convex curve, touches the annulus tooth profile at its limit circle (see Sect. 5.7)
and, in a well-designed internal gear, must lie within the tooth space of the annulus,
without intersecting its profile.
The path of the tooth tip V1 of the pinion, during the meshing cycle, intersects
the annulus addendum circle at the point V10 . This position V10 of the point V1 should
be compared with that of the tooth tip V2 of the annulus. To avoid the secondary
interference, it is necessary that the distance V2 V10 , measured along the annulus
addendum circle, is greater than or at least equal to zero. In this regard, as a specific
5.5 A Geometric-Analytical Method … 197
value, depending on the size and accuracy of the gears, we impose an adequate
margin, i.e. that the arc length V2 V10 is greater than or at least equal to 0:05 m.
In order to calculate this arc length, we determine first the polar coordinates of
the point V1 in the local coordinate system of the pinion, which has its origin at the
center O1 and its tooth centerline x1 as reference axis (Fig. 5.9). The distance
between V1 and O1 is the radius ra1 of the pinion addendum circle. The other polar
coordinate is the angle #a1 between the straight line O1 V1 and the x1 -axis. It can be
expressed as (see Figs. 5.9 and 2.10):
s1
#a1 ¼ þ inva1 invaa1 ; ð5:18Þ
2r1
where s1 , r1 , and a1 are the standard values of the tooth thickness, pitch circle radius
and pressure angle of the pinion, and aa1 is the profile angle or pressure angle at
point V1 , which is given by the following relationship:
rb1
cosaa1 ¼ : ð5:19Þ
ra1
We determine now the polar coordinates of the point V2 in the local coordinate
system of the annulus, which has its origin in the center O2 and its tooth centerline
x2 as reference axis (Fig. 5.9). The distance between V2 and O2 is the radius ra2 of
the annulus addendum circle. The other polar coordinate is the angle #a2 between
the straight line O2 V2 and the x2 -axis. To determine #a2 as well as other interesting
Fig. 5.9 Relative position of the pinion and annulus for checking of the secondary interference
198 5 Interference Between Internal Spur Gears
where a and ar are the profile angles at points S and P, respectively related with u
and ur (see also Fig. 2.1); therefore, a is the pressure angle.
The tooth thickness sr on the circle of radius r is equal to
s2
sr ¼ 2r#r ¼ r 2ðinva invar Þ : ð5:21Þ
r2
Therefore, for P coincident with V2 , and bearing in mind that #a2 is negative
(Fig. 5.9), because point V2 is located on the upper face of the tooth, and the polar
angle is defined as positive when it is clockwise, we get:
s2
#a2 ¼ inva2 þ invaa2 : ð5:22Þ
2r2
Figure 5.9 shows the instantaneous position that is when V1 V10 , i.e. the
position that is when the tip point V1 on the pinion profile comes to coincide with
the point V10 , determined as the intersection of the path of the same point V1 with the
annulus addendum circle. In this position, the distance between V10 and O2 is the
radius ra2 of the addendum circle of the annulus. The angles b1 and b2 define
the angular positions of the pinion and annulus, and #2 is the polar angle which
identifies the position of the straight line O2 V1 O2 V10 with respect to the x2 -axis.
From triangle O2 O1 V1 O2 O1 V10 , we infer the following two equations:
Finally, from Eq. (5.23), we can calculate #2 , which is given by the following
relationship:
ra1
#2 ¼ arcsin sinðb1 þ #a1 Þ b2 : ð5:27Þ
ra2
To avoid the secondary interference, the condition that the arc length V2 V10 is
greater than or at least equal to 0:05 m, leads to writing the following inequality,
which must be satisfied:
200 5 Interference Between Internal Spur Gears
Generally, this inequality is satisfied for internal gear pairs for which the min-
imum difference ðz2 z1 Þmin is 8 or more (it is to be noted that, in some case, this
difference drops to 7 or 6). Therefore, using the inequality (5.28), we get results in
full agreement with those obtained with the approximate method of Henriot [9].
Anyway, it is to keep in mind that this checking type must be carried out necessarily
when the difference ðz2 z1 Þ is small, and that the amount of clearance depends on
several quantities, such as the center distance and radii of the addendum circles.
In the case of an internal gear pair, the pinion and annulus can be assembled in
their meshing position by assembling the pinion in the axial direction (Fig. 5.11a).
This axial assembling is not physically possible in the initial cutting conditions,
when the pinion is the pinion-type cutter, and the ring gear is an internal gear blank
or ring gear blank. In this case, before the motion of generation begins, the
pinion-type cutter is pulled over and penetrates radially into the ring gear blank
being cut, until the two cutting pitch circles are tangent at the pitch point
(Fig. 5.11b). Starting from this position, the cutting motion in the axial direction is
associated with the motion of generation, consisting of the pure rolling of the pitch
circle of the pinion-type cutter on the pitch circle of the annulus in the cutting
conditions.
The check that the tertiary interference does not take place is based on geometric
considerations, in both of the following conditions:
• The centerline of a tooth of the pinion coincides with the centerline of a tooth
space of the gear ring, and both these centerlines coincide with the centerline of
the internal gear pair.
(a)
(b)
Fig. 5.11 a axial assembly, and axial motion of shaping; b radial assembly, and radial approach
motion of the pinion-type cutter
202 5 Interference Between Internal Spur Gears
• The centerline of a tooth space of the pinion coincides with the centerline of a
tooth of the ring gear, and both these centerlines coincide with the centerline of
the internal gear pair.
Obviously, for pinion we mean both the actual pinion, and the pinion- type
cutter, while for internal gear pair we mean both the actual internal gear pair, and
the internal gear pair in the cutting condition, depending on whether we consider
the operating condition or the cutting condition.
In order to make such a check, first an orthogonal Cartesian coordinate system is
chosen, with origin at the center O2 of the ring gear, x-axis coinciding with the
centerline O2 O1 of the internal gear pair, and y-axis normal to x-axis. Then we
consider the successive tip points V2;i (with i ¼ 1; 2; 3; . . .) of each tooth of the ring
gear 2, identified on the addendum circle from the one of the two tip points that is
closest to the x-axis. We consider also the successive tip points V1;i (with
i ¼ 1; 2; 3; . . .) of each tooth of the pinion 1, which are furthest from the x-axis. For
the identification of the latter it is to be observed that, for the pinion teeth closest to
the x-axis, the points V1;i furthest from the x-axis are not the furthest tip points.
They indeed correspond to the points obtained as the intersection of the tooth
profiles with the tangent to the base circle, perpendicular to the x-axis; their distance
y from the x-axis is equal to half the span-gauging chords, measured over three
teeth. For the remaining teeth, they are identified on the addendum circle from the
one of the two tip points that is the furthest from the x-axis (see also Colbourne [4]).
It should be remembered that the span gauging is a method of measuring tooth
thickness by means of the dimension between opposite flanks of teeth several pitch
apart, measured along a chord tangential to the base circle. In this case, the mea-
surement is done by considering three teeth, the central one whose centerline
coincides with the x-axis, and the two adjacent teeth. Therefore, the total number of
teeth spanned by the chord is z ¼ 3. Therefore, we will have s ¼ ðz 1Þpb þ sb ¼
2pb þ sb ; where sb is the tooth thickness on the base circle, given by Eq. (2.60).
The check must be made in both the afore-mentioned conditions. In each of
these conditions, we calculate the y-coordinates of the labelled points V2;i on the
ring gear, and verify that, in any case, they are greater than the y coordinates of the
corresponding points V1;i on the pinion. If these conditions are satisfied for each
pair of corresponding points labelled on the ring gear and pinion, the radial
assembly can be carried out, and simultaneously the tertiary interference in cutting
condition is avoided.
This procedure, very simple, but a little long, leads to the conclusion that the
conditions to dispel the risk of tertiary interference are more restrictive than those
are necessary to avoid the secondary interference. In other words, the minimum
value of the difference ðz2 z1 Þmin , necessary to avoid the tertiary interference (and
therefore to allow also the radial assembly), is greater than the minimum value of
the same difference, necessary to avoid the secondary interference. This minimum
value of the difference ðz2 z1 Þmin , for which radial assembly is possible and
tertiary interference does not occur, depends on the pressure angle. For the cutting
of a ring gear, it is necessary to use pinion-type cutters having numbers of teeth
5.6 Tertiary Interference in the Internal Spur Gears 203
very different from those of the annulus to be cut. For this reason, the pinion-type
cutters for cutting of internal gears are characterized by a very small number of
teeth.
In order to avoid this long procedure, Henriot [9] suggests adopting the fol-
lowing simple rule: the minimum value of the difference ðz2 z1 Þmin obtained using
the diagram of Fig. 5.7 must be increased by six teeth. This approximate rule
provides sufficiently reliable results.
For now, we focus our attention on what happens when the point of contact,
moving on the line of action in the direction from point A to point E, arrives in the
vicinity of point E. In this region of the path of contact, we must take into account
the fact that the involute portion of the annulus tooth profile does not arrive until its
dedendum circle, but ends at the fillet circle, having radius rfil;2 . In addition, we
must consider that the end point of the active profile of the annulus tooth, which is
closest to the root, is the so-called limit point, and that the circle with center O2 and
passing through this point is the limit circle of the annulus, having radius
rl2 ¼ O2 E. We remember that, for both members of an external or internal spur
204 5 Interference Between Internal Spur Gears
Fig. 5.12 Meshing diagram of an internal gear pair, consisting of a driving pinion and a driven
annulus
gear, the active profile of a tooth is the part that actually makes contact between
mating profiles.
From Fig. 5.13, which shows the meshing diagram of an internal gear pair, we
infer that the radius rl2 of this limit circle is given by:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2 2 2
rl2 ¼ O2 E ¼ O2 T2 þ T2 E ¼ O2 T2 þ T1 E þ T2 T1
( qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 )1=2 ð5:31Þ
¼ rb2
2
þ 2 r 2 þ ðr r Þtana
ra1 b1 b2 b1 :
To ensure that the fillet interference between the tip of the pinion, and root fillet
of the annulus does not occur, since the teeth of the annulus are facing inward, the
radius of its limit circle must be smaller than the radius of its fillet circle, i.e.:
Here also it is to be noted that, to neutralize the effects of inevitable errors in the
center distance, it is appropriate to design the internal spur gear pair so that the fillet
circle of the annulus is greater than the limit circle by a certain margin, usually
5.7 Fillet Interference Between the Tip … 205
The radius rfil;2 of the annulus fillet circle appearing in the inequities described
above remains to be calculated. This will be done in Sect. 5.9.
In Sect. 5.2, we found that the contact is non-conjugate when the point of contact is
located beyond the interference point T1 , in the direction from point T1 to point T2
(see Figs. 5.1 and 5.13). In addition, we have shown that, to avoid the theoretical or
primary interference, the minimum value of the radius ra2 of the addendum circle of
the annulus must be greater than or equal to O2 T1 , i.e. ra2 O2 T1 (Figs. 5.1 and
5.13). This amounts to saying that the end point A of the path of contact must lie
206 5 Interference Between Internal Spur Gears
beyond the interference point T1 , in the direction from T2 to T1 , that is, T2 A must be
greater than T2 T1 (Fig. 5.13). Therefore, the following inequality must be satisfied:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 r 2 [ ðr r Þtana:
ra2 ð5:34Þ
b2 b2 b1
Now, we focus our attention on what happens when the point of contact is
located near the point A, where the path of contact begins. In this region of the path
of contact, we must take into account the fact that the involute portion of the pinion
tooth profile does not arrive until its dedendum circle, but ends at the fillet circle,
having radius rfil;1 . We must also consider that the end point of the active profile of
the pinion tooth, which is nearest to its root, is the so-called limit point, and that the
circle with center O1 and passing through this point is the limit circle of the pinion,
having radius rl1 ¼ O1 A. From Fig. 5.13, we infer that the radius rl1 of this limit
circle is given by:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2 2 2
rl1 ¼ O1 A ¼ O1 T1 þ T1 A ¼ O1 T1 þ T2 A T2 T1
( qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 )1=2 ð5:35Þ
¼ rb1 þ
2
ra2 rb2 þ ðrb2 rb1 Þtana
2 2 :
To ensure that the fillet interference between the tip of the annulus, and root fillet
of the pinion does not occur, the radius of the limit circle of the pinion must be
larger than the radius of its fillet circle, i.e.:
Here also, to neutralize the effects of inevitable errors in the center distance, it is
appropriate to design the internal gear pair so that the limit circle of the pinion is
greater than its fillet circle by a certain margin, usually assumed to be 0:025 m.
Therefore, the theoretical inequality (5.36) changes, and assumes the form of the
following design inequality:
The quantities that appear in the theoretical inequalities (5.32) and (5.36), and in
the design inequalities (5.33) and (5.37) are all known, with the exception of the
fillet circle radius rfil;2 of the annulus. Instead, rl1 and rl2 are given respectively by
Eqs. (5.35) and (5.31), while rfil;1 is given by Eq. (4.44), or by Eq. (4.32),
depending on whether the pinion cutting process is made with a pinion-type cut-
ter, or with a rack-type cutter or a hob.
5.9 A Design Consideration on the Interference … 207
Fig. 5.14 Limit value if the addendum, a2lim , of the annulus to prevent the primary interference
The shaping process with a pinion-type cutter is the cutting method almost always
used to cut a ring gear. The funtamentals of this cutting method have been recalled
at the beginning of Sect. 4.8, to which we refer the reader. The use of a pinion-type
cutter with rounded tips is preferable since, otherwise, at the root fillet of the
5.10 Annulus Fillet Profile Generated by a Pinion-Type Cutter 209
Fig. 5.15 Trimming factor of the annulus addendum as a function of z2 and k 2 , for a ¼ 20
annulus, stress concentrations too high would generate, resulting from radii of
curvature of the fillets too small.
The tip geometry of a pinion-type cutter with rounded corner at the tooth tip is
that shown in Fig. 4.16. The end point L of the involute portion of the tooth profile
has polar coordinates ðrL0 ; uL0 Þ, given by Eqs. (4.37) and (4.38), while the center
I of the circular portion at the tooth tip has polar coordinates ðrI0 ; uI0 Þ, given by
Eqs. (4.33) and (4.34).
To determine the radius rfil;2 of the annulus fillet circle, which intersects the
annulus tooth profile at point Pf where the involute portion of the same profile ends
and the fillet begins, we consider the cutting meshing diagram between the
pinion-type cutter and annulus to be generated (Fig. 5.16). The involute portion of
the cutter tooth profile, which ends at point L (Fig. 4.16), cuts the involute portion
of the annulus tooth profile. Thus, the end point Pf of the annulus tooth involute
profile is cut by the point L on the pinion-type cutter profile. This happens when the
point L, during the cutting relative motion, is to coincide with point E 0 , where the
210 5 Interference Between Internal Spur Gears
line of action in the cutting condition and active addendum circle of the pinion-type
cutter intersect. Point E 0 is therefore the end point of the path of contact during the
cutting meshing. It is to be noted that the path followed by point L of the
pinion-type cutter during this motion coincides with the active addendum circle.
The fillet circle of the annulus is the circle having center O2 and radius
rfil;2 ¼ O2 E 0 . This radius can be read from the meshing diagram shown in Fig. 5.16,
and is given by the following relationship:
( qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 )1=2
rfil;2 ¼ 2
rb2 þ ra0 rb0 þ ðrb2 rb0 Þtana0
2 2 ; ð5:38Þ
Fig. 5.16 Meshing diagram between pinion-type cutter and annulus to be generated
5.10 Annulus Fillet Profile Generated by a Pinion-Type Cutter 211
With reference to the meshing diagram shown in Fig. 5.16, we can define other
quantities relating to the cutting condition. To this end, we denote by: z2 and z0 , the
numbers of teeth of the annulus and pinion-type cutter; a0 , the cutting pressure
angle, which is the same for the pinion-type cutter and annulus to be generated; rb2
and rb0 , the radii of the base circles of the annulus and pinion-type cutter; a0 , the
center distance in the cutting condition. Note that, in cases in which a possibility of
misunderstanding does not exist, only the subscript 0 is used to indicate quantities
that relate both to the cutting tool, and to the cutting condition. Instead, in cases in
which a possibility of ambiguity exists, to avoid misunderstandings, the symbol 0 is
used as the second subscript, where deemed necessary, to indicate quantities that
relate to the cutting condition.
Recalling equations already shown in previous chapters, and referring to
Fig. 5.16, we obtain the following relationships:
z2 z0
a0 ¼ m0 ð5:39Þ
2
a0 cosa0 ¼ ðrb2 rb0 Þ ð5:40Þ
z2
r2;0 ¼ a0 ð5:41Þ
ðz2 z0 Þ
z0
r0;0 ¼ a0 ð5:42Þ
ðz2 z0 Þ
2p
p0 ¼ a0 : ð5:43Þ
ðz2 z0 Þ
In these equations, m0 is the module in the cutting condition (or cutting module),
p0 is the cutting circular pitch, and r2;0 and r0;0 are the radii of the cutting pitch
circles of the annulus and pinion-type cutter. Of course, m0 and p0 are the same for
the pinion-type cutter and annulus to be generated.
During the cutting process, the meshing diagram (Fig. 5.16) between the
pinion-type cutter and annulus to be generated is equal to that of an internal gear
pair (see Fig. 5.13), with the only feature that it is without backlash. Therefore, the
tooth thickness s2;0 of the annulus is equal to the space width of the pinion-type
cutter, both measured on the cutting pitch circles; thus, we will have:
s2;0 ¼ p0 s0;0 ; ð5:44Þ
where s0;0 is the tooth thickness of the pinion-type cutter, also measured on the
cutting pitch circle.
The tooth thickness of the annulus, s2;0 , at the cutting pitch circle of radius r2;0 , is
related with the tooth thickness s2 at the standard pitch circle of radius r2 . The
relationship that correlates these two tooth thicknesses is obtained from Eq. (5.21),
so it can be written as follows:
212 5 Interference Between Internal Spur Gears
s2
s2;0 ¼ r2;0 2ðinva inva0 Þ ; ð5:45Þ
r2
this is because ar ¼ a0 :
Therefore, using the general Eq. (2.58), we infer that the tooth thickness of the
pinion-type cutter, s0;0 , at the cutting pitch circle of radius r0;0 , is related with the
tooth thickness s0 at the reference pitch circle of radius r0 by the following
equation:
s0
s0;0 ¼ r0;0 þ 2ðinva inva0 Þ : ð5:46Þ
r0
It is noteworthy that the reference pressure angle, a, and the cutting pressure
angle, a0 , of the pinion-type cutter are equal to the corresponding ones of the
internal gear to be generated.
The cutting center distance a0 and the reference cutting center distance a0;s are
given respectively by:
a0 ¼ r2;0 r0;0 ð5:47Þ
In order to obtain the tooth thickness of the annulus, s2 , at its standard pitch
circle of radius r2 as a function of a0;s , we first substitute Eqs. (5.45) and (5.46) into
Eq. (5.44); then, multiplying by the ratio a0;s =a0 both members of the equality so
obtained, and bearing in mind that both the products a0;s =a0 r2;0 =r2 and
a0;s =a0 r0;0 =r0 are equal to unity, we obtain:
where p is the standard or reference circular pitch. From this equation, we obtain:
1
inva0 ¼ inva þ ðp s0 s2 Þ: ð5:50Þ
2a0;s
Once the value of inva0 is known, the angle a0 can be derived using, for
example, the following approximate procedure, in two steps. As a first step, we set:
1
¼ 1:0 þ 1:04004q þ 0:32451q2 0:00321q3 0:00894q4 þ 0:00319q5
cos a0
þ 0:00048q6 :
ð5:52Þ
ðrb2 rb0 Þ
a0 ¼ : ð5:53Þ
cosa0
As Fig. 5.17 shows, the fillet circle of the pinion-type cutter intersects the line of
action at point F2 localized between T0 and C. Since the involute portion of the
cutter tooth profile begins from this fillet circle, and then develops at the outside
thereof, it is clear that the cutting of the involute portion of the annulus tooth profile
can take place only outside this circle. Thus, to avoid undercut, the path of contact
during the cutting meshing must begin at a point of the line of action close to the
point F2 , but located beyond this, i.e. between F2 and C (or at least coincident with
F2 ). Therefore, the following inequality must be satisfied:
hqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi i2
2
ra2 rb2
2
þ 2 r 2 þ ðr r Þtana
rfil;0 b0 b2 b0 0 : ð5:54Þ
If the required radius ra2 of the annulus addendum circle is smaller than the
minimum value given by this inequality, it is necessary to use a pinion-type cutter
with more teeth.
The tip interference between the pinion-type cutter and the annulus during the
cutting process may cause another type of undercut of the annulus to be generated.
This undercut can occur during the generating motion (in this case, the cutting
center distance is equal to a0 ), or during the initial phase of radial penetration of the
pinion-type cutter inside the ring gear blank. During this initial radial penetration,
the cutting center distance is less than a0 , and its instantaneous value ac is variable
with continuity, until reaching the maximum value equal to a0 . Actually, with the
usual gear cutting machines, this center distance varies step by step. In this
framework, we consider any position of the pinion-type cutter related to a given
step of the initial phase of the cutting tool radial penetration, which is characterized
by the defined value of the cutting center distance ac , corresponding to the step
under consideration.
So far, we have considered a pinion-type cutter with teeth rounded at the tips
(Fig. 4.16). Here, to simplify the calculations related to the case of tip interference
that we want to examine, we assume instead that the involute profile of the cutter
tooth extends until the addendum circle, which is intersected by the involute curve
at the point V0 (Fig. 5.18). The path of this point V0 intersects the annulus
addendum circle at the point V00 . Then we denote with V2 the tooth tip corner of the
annulus, i.e. the point in which the annulus tooth profile intersects the annulus
addendum circle. To avoid this tip interference, and related undercut, it is necessary
that the distance V2 V00 , measured along the annulus addendum circle, is greater than
or at least equal to zero at the limiting condition; as a specified value, typically we
impose that the arc length V2 V00 is greater than or at least equal to 0:02 m.
Fig. 5.18 Relative position of the pinion-type cutter and annulus to be generated, for checking of
the tip interference and related undercut
216 5 Interference Between Internal Spur Gears
In order to calculate the arc length V2 V00 , we first determine the polar coordinates
of the point V0 in the local coordinate system of the pinion-type cutter, which has its
origin at the center O0 and as reference x0 -axis its tooth centerline (Fig. 5.18). The
distance between V00 and O0 is not more than the radius ra0 of the addendum circle
of the pinion-type cutter. The other polar coordinate is the angle #V00 between the
straight line O0 V00 and the x0 -axis; it can be expressed as (see Figs. 5.18 and 2.12):
s0
#V00 ¼ þ inva0 invaa ; ð5:55Þ
2r0
where aa is the pressure angle or profile angle at point V00 , which is given by the
relationship:
rb0
cosaa ¼ : ð5:56Þ
ra0
Figure 5.18 shows the instantaneous position that is when V0 V00 , i.e. when the
tip point V0 of the pinion-type cutter comes to coincide with the point V00 , deter-
mined as the intersection of the path of the same point V0 with the annulus
addendum circle. The distance between V00 and O2 is the radius ra2 of the addendum
circle of the annulus. The angle #2 is the polar coordinate of point V00 relative to the
x2 -axis, the latter coinciding with the tooth centerline of the annulus. From triangle
O2 O0 V00 we infer the following two equations:
sin b0 þ #V00 sinðb2 þ #2 Þ
¼ ð5:57Þ
ra2 ra0
2
ac þ ra0
2
ra2
2
cos b0 þ #V00 ¼ : ð5:58Þ
2ac ra0
1 p
b2 ¼ r0 b0 ; ð5:60Þ
r2 2
which is obtained from Eqs. (5.17) or (5.26), where we replace r1 with r0 , b1 with
b0 , s1 with s0 , and we remember that s0 þ s2 ¼ p:
Finally, from Eq. (5.57), we infer #2 , i.e. the polar coordinate of the point
V0 ¼ V00 , which is given by the following relationship:
5.11 Undercut or Cutting Interference 217
ra0
#2 ¼ arcsin sin b0 þ #V00 b2 : ð5:61Þ
ra2
To avoid the tip interference and related undercut, the condition that the arc
length V2 V00 is greater than or at least equal to 0:02 m, leads to write the following
inequality, which must be satisfied:
where #a2 is the polar coordinate of point V2 on the tooth tip of the annulus, given
by Eq. (5.22).
During the initial phase of radial penetration of the pinion-type cutter inside the
ring gear blank, i.e. before the generating motion begins, the center O0 of the
pinion-type cutter moves along the straight line O2 O0 , and the center distance ac ,
initially equal to the difference ðra2 ra0 Þ, increases step by step up to the final
value a0 . Therefore, to ensure that the tip interference, and related undercut, do not
occur, we must carry out the checking for different values of ac , starting from the
minimum value equal to ðra2 ra0 Þ, and ending with ac ¼ a0 . If such checking
shows the evidence of this tip interference for any value of ac , between ðra2 ra0 Þ
and a0 , most likely the annulus would be not usable, since such interference would
have removed a substantial portion of its involute profile.
2
ra2 a2 rL2
a ¼ arccos aL ; ð5:64Þ
2arL
5.12 Condition to Avoid Rubbing During the Annulus Cutting Process 219
which allows to calculate the pressure angle a at the upper end point L of the
involute part of the tooth cutter profile.
Obviously, the angle a thus found represents the maximum value that the
pressure angle at a point can assume, since point L is the upper end point where the
involute curve of the cutting edge is truncated. As the pinion-type cutter rotates, and
the cutting edge penetrates further into the ring gear blank, the pressure angles at
subsequent point of contact (i.e. the cutting points) gradually decrease. Therefore,
to prevent rubbing by the cutter tooth trailing profile, it is necessary that in its return
stroke the pinion-type cutter is moved from the cutting zone in a direction that
forms an angle greater than a with the line of centers.
We now consider the leading profile of the cutter tooth, which is in contact with
a tooth of the ring gear blank. The angles between the tangents at points of contact
and the line of centers reach their maximum values when the point of contact is the
one corresponding to the lower end point of the cutting edge; in this case, this angle
becomes equal to the cutting pressure angle, a0 . Therefore, to prevent rubbing by
the cutter tooth leading profile, it is necessary that in its return stroke the pinion
cutter is moved from the cutting zone in a direction that forms an angle less than a0
with the line of centers.
Both of the above-mentioned requirements must be met to avoid rubbing. First,
the designer should check that the pressure angle a at point L, given by Eq. (5.64),
is less than the cutting pressure angle a0 by at least a few degrees. Then the designer
has to define the working displacement angle, a w , which must be between a and
a0 , i.e.:
220 5 Interference Between Internal Spur Gears
a \a w \a0 : ð5:65Þ
For the best possible determination of the value of this working displacement
angle, it should be noted that rubbing, for a displacement angle equal to a0 is more
severe than it is for a displacement angle equal to a ; this is apparent from
Fig. 5.19. Therefore, it is more appropriate to choose a working displacement angle
closer to a than to a0 . In case the Eq. (5.64) would give a value of a not less than
a0 with a sufficient margin, then a pinion cutter with fewer teeth should be used.
References
21. Radzevich SP (2017) Gear cutting tools: science and engineering, 2nd edn. CRC Press,
Taylor&Francis Group, Boca Raton, Florida
22. Schreier G (1961) Stirnrad-Verzahnung: Berechnung, Werkstoffe, Fertigung. VEB Verlag
Technik, Berlin
23. Tuplin WA (1967) Tip interference in internal gears. The Engineer 224(5839):827
Chapter 6
Profile Shift of Spur Gear Involute
Toothing Generated by Rack-Type
Cutter
Abstract In this chapter, the fundamentals of the profile shift of spur gear involute
toothing generated by rack-type cutter are first described, highlighting also the
technical-applicative interest of the use of this type of profile-shifted toothing The
problem is analyzed in its maximum generality, so the cases of zero profile-shifted
toothing and of profile-shifted toothing without and with variation of center dis-
tance are discussed, both for external and internal spur gears In this general
framework, the problems related to the minimum profile shift coefficient to avoid
interference and the risk of pointed teeth are also discussed. The problem of how to
split the sum or difference of the profile shift coefficients between the pinion and
mating gear wheel is then dealt with, under the double aspect of direct problem and
inverse problem also with reference to the different criteria that can be used to
achieve specific design goals. Finally, the problem of how to ensure the optimal
value of backlash between the mating teeth of cylindrical spur gears is discussed.
6.1 Introduction
In Sect. 4.4 we have saw that, to avoid the primary interference of external spur
gear pairs with pinions having small numbers of teeth, we can choose three different
design solutions, namely: the increase of the standard pressure angle; the use of
stub-toothed gears; the use of toothing with profile shift. We also reported that the
first two solution methods have many disadvantages and drawbacks, and that the
only method used is actually the third method, that is to employ involute
profile-shifted toothing (see Buckingham [3], Merritt [21], Dudley [8], Giovannozzi
[12], Ferrari e Romiti [11]; Pollone [25], Henriot [13], Niemann and Winter [22],
Colbourne [5], Scotto Lavina [29], Maitra [20]) .
Profiles shift is closely related to the involute tooth profile. In scientific and
technical literature, the term profile shift, exactly corresponding to the German word
Profilverschiebung, has many synonyms, such as: addendum modification, cor-
rection, cutter offset and hob offset. None of the five synonyms can be considered
without any objection.
Fewer objections are registered for the use of term profile shift. It is to be
preferred to the term addendum modification, however still very widely used, but
misleading, because with it we could understand a stub toothed gear, which is
something quite different from a profile-shifted gear. In fact, the profile shift is a
measure of the tooth thickness, rather than of the addendum length. In other word,
the profile shift implies a change of the addendum length because of the tooth
thickness variation, with the tooth depth that remains constant. The only change in
the addendum length, without variation of the tooth thickness, like the one we have
in the stub-toothed gears, does not fall within the definition of profile shift.
The term correction is an abbreviation of the older term correction for undercut.
It is therefore limiting, because it would imply that the profile shift is motivated
only by the need to avoid undercut. The terms hob offset and cutter offset are even
more restrictive, as well as ambiguous. The reader tries to wonder what meaning the
term hob offset might have when the gear is not cut by a hob. The term cutter offset
is certainly an improvement over the term hob offset, but it is not suitable for a gear
cut by a pinion-type cutter, since in this case the profile shift is not equal to the
cutter offset (see Chap. 7).
Notoriously, the profile shift of involute toothing is the most striking stage of
their geometric- kinematic design. In addition to the reduction of the minimum
number of teeth to prevent the primary interference and undercut (and therefore, the
size, weight and cost of the gear pair, with the same gear ratio), the profile shift
allows also to achieve other desirable design goals. The main goals are as follows
(see Oda et al. [23], Simon [30], Li [18], Chauhan [4]):
• shapes of teeth having greater strength, both for bending fatigue loads and
surface fatigue loads;
• more uniform distribution curves of the specific sliding and slide-roll ratios for
the driving and driven gears along the path of contact;
• length of path of contact as much as possible extended, for noise control;
• flanks of small pinions without undercut that, as it is well known, causes a
reduction of the contact ratio, and weakens the tooth to its root; etc.
Actually, the profile shift is not only performed in cases where it is necessary to
prevent the primary interference and undercut of the flanks of pinions with small
number of teeth, but also in cases where it is not strictly necessary. Among these
cases are, for example, those in which the design requirements impose to overcome
the well-known limits related to the part of the involute curve near the base circle,
where the capacity to resist surface loading is little, due to the comparatively small
radius of curvature, while the sliding velocity tend to be high, and rolling velocity
low. In these cases, the profile shift, although not necessary, is advantageously
carried out in order to achieve better conditions of surface contact and relative
motion, and a better balance of the sliding conditions between the pinion and wheel.
The profile shift is also carried out in order to achieve an imposed center distance.
The previously mentioned strength of profile-shifted toothing, together with its
manufacturing cost, identical to that of the zero-profile-shifted toothing (or standard
6.1 Introduction 225
toothing), fully justify the fact that it is used in most of the usual technical appli-
cations, even when the problem of interference does not exist.
By definition (see ISO 1122-1: [4]) the profile shift is the distance, xmn , mea-
sured along a common normal between the reference cylinder of the gear and the
datum plane of the basis rack, when the rack and the gear are superposed so that the
flanks of a tooth of one are tangent to those of the other (Fig. 6.1). By convention,
which is valid for both external and internal gears, the profile shift is positive when
the datum plane of the basic rack is external to the gear reference cylinder
(Fig. 6.1), and negative when this datum plane cuts the gear reference cylinder. For
internal gear, tooth profiles are considered those of tooth spaces. The profile shift
coefficient, x, is the quotient of the profile shift, expressed in millimeters, divided by
the normal module, mn (in Imperial system, the profile shift coefficient is the
product of the profile shift, expressed in inches, for the normal diametral pitch).
From the point of view of the history of technology, it is to be considered that the
knowledge related to the profile-shifted gears were known from some time, and so, at
least partially, the benefits that could be obtained with their use. However, as long as
the only technology of gear cutting was the non-generation process, as the milling
cutting, the profile-shifted gears could not actually find wide application, as, for the
teeth cutting, it was necessary to construct two special milling cutters for each gear
pair. The profile-shifted gears entered instead in the current technological practice
when the gear generation cutting with rack-type cutter was introduced [9, 10, 31].
However, even when the gear generation cutting began to spread, a lot of time
was needed to understand the full potential of this technology in relation to the
optimal design of the gears. In fact, initially, the profile shift was used for involute
gears designed to mesh at the standard center distance, and thus emulating what
happened in the cycloidal gear pairs, which notoriously may mesh only at their
standard center distance. In this way, the standardization, and thus the inter-
changeability and production of the catalogue gears, were privileged. Only in a
second time the designer was able to understand the enormous benefits that could
be drawn from the flexibility of use the profile shift, which, though within certain
limits, allows to choose the values of the tooth thicknesses, center distance, and
gear blank diameters so as to satisfy as much as possible the design requirements.
The only possible restriction, consisting in the use of standard cutters for cutting
profile-shifted gears, is not valid for the gear generation cutting, as one of the main
advantages of the involute profile is to allow the cutting of non-standard gears by
standard cutters.
The gear cutting technology with profile shift make it possible to move the
position of the cutting pitch circle with respect to the teeth. Therefore, the tooth
retains its overall tooth depth, h, by varying the distance of the rack-type cutter from
the axis of the gear to be generated, and leaving unaltered the speeds (translational
and rotational) of the rack-type cutter and the wheel. Of course, the addendum and
dedendum values change with respect to those that characterize the normal standard
sizing.
A gear without profile shift will differ under many aspects compared to a gear
with a positive or negative profile shift. Figure 6.2 compares two gears with the
same number of teeth, z, and cut with the same rack-type cutter, and therefore with
the same pitch circle (this circle coincides with the cutting pitch circle), having
radius r ¼ d=2 ¼ mz=2, but the first (Fig. 6.2a) with zero-profile shift and the
second (Fig. 6.2b) with positive profile shift. Since both gears have the same pitch
circle and the same pressure angle, a (this does not change regardless of the
straight-line parallel to the datum line of the rack-type cutter that is considered),
they will have the same base circles, whereby the teeth profiles of each gear are
constituted by different parts of the same involute curve. The tip and the root
circles, the tooth thicknesses on the pith circles, and addendum and dedendum of
the two gears are different. More particularly, the tip circle radius and tooth
thickness on the pitch circle of the positive profile-shifted gear are larger compared
to those of the gear with a zero-profile shift.
In addition, Fig. 6.2 shows that, in a gear with positive profile shift, the involute
curves forming the opposite flanks of each tooth are more spaced than those of a
gear with zero-profile shift, but have the same base and standard pitch circles. In
other word, the tooth profiles of the second gear are shifted relative to those of the
first gear. Then we define as profile-shifted gear, i.e. as gear cut with profile shift,
any gear whose tooth thickness measured along the cutting pitch circle is not equal
to p=2 ¼ pm=2.
(a) (b)
Fig. 6.2 Gears with the same base and pitch circles; a zero-profile shift gear; b positive profile
shift gears
6.1 Introduction 227
Either or both of the two members of a gear pair may have a profile-shifted
toothing, and the profile shift coefficient may be positive or negative. According to
this, we have two methods of applying profile shift:
(1) With no change in the center distance, which remains the same as for
zero-profile-shifted gears. The profile shift coefficients of the two members of
the gear pair are then equal and opposite.
(2) With a change in the center distance (usually an increase, but in some cases also
a decrease). The profile shift coefficients of the two members of the gear pair
are not equal and opposite, but different from each other.
Usually, the profile shift with no change in the center distance is used when the
tooth-sum is large, and the center distance may be chosen to suit the tooth-sum and
module. Usually, a positive profile shift coefficient is chosen for the pinion and, for
the wheel, a negative profile shift coefficient, equal and opposite to that of the
pinion. That is because the positive profile shift coefficient for the pinion improves
its shape and strength, whilst the negative profile shift coefficient for the wheel is
not so large as to impair its strength. With this choice, the resistance of the gear pair
to surface loading is increased.
Instead, the profile shift with a change in their center distance is applied when
the tooth-sum is comparatively small (i.e. when a negative profile shift coefficient
would impair the quality of the wheel teeth). It is also used when the gear
dimensions must be handled in order to agree with an imposed center distance, not
strictly complied with the one related to the module and number of teeth.
These two methods together provide an infinite number of possible ways of
applying profile shift, or of diving either an imposed or an arbitrary total sum of the
profile shift coefficients between the pinion and mating gear wheel. With current
knowledge, however, it is not possible to have precise rules that allow to achieve
optimum tooth shapes, and, this is because many phenomena relating to the
meshing between the teeth, especially those of contact loading, are still not com-
pletely know. For this reason, the judgment of the gear designer is still essential.
Figure 6.3 shows the tooth profile of generating rack-type cutter (or the axial
section of a hob), consistent with the standard basis rack tooth profile. Teeth are
shaped trapezoidal, with rounded tips and flanks inclined at the pressure angle a0 to
the tooth axis, which is normal to the datum line P-P; a0 is equal to the real value of
the pressure angle aP of the standard basis rack. The datum line (or reference line)
of the rack-type cutter bisects the tooth depth h0 ¼ ð10=4Þm0 , for which addendum
and dedendum are equal between them and equal to ð5=4Þm0 .
With reference to the cutting conditions, it should be noted that the addendum of
the generating rack-type cutter has to dig the tooth space width below the pitch
228 6 Profile Shift of Spur Gear Involute Toothing …
circle of the gear to be generated. Furthermore, during the tooth generation motion,
the top land of the gear must not touch the bottomland of the tooth space width
between two adjacent teeth of the rack-cutter. Since then the rounded tips of the
rack-type cutter have a depth equal to ð1=4Þm0 , in relation to the problem of the
cutting interference, the rack-type cutter shown in Fig. 6.3 will behave as if its
addendum was equal to m0 .
The tooth thickness s0 and space width e0 on the nominal datum line are equal to
half the pitch, i.e. s0 ¼ e0 ¼ p0 =2 ¼ pm0 =2. Given the geometrical shape of the
rack-cutter, the module m0 (and hence the pitch p0 ) and pressure angle a0 do not
vary whatever the straight line, parallel to the nominal datum line, which is chosen
to generate a gear wheel with z teeth. Regardless of this choice, the gear will have a
generation pitch circle (or cutting or nominal pitch circle) and a base circle whose
diameters are given by:
d ¼ 2r ¼ zm0 ð6:1Þ
Figure 6.4 shows the geometrical quantities characterizing both the generating
standard rack-type cutter, and a gear with zero-profile shift, generated by it. In this
case, the cutting pitch line of the rack-type cutter coincides with its datum line, and
we obtain a zero profile-shifted gear wheel, i.e. a wheel with normal standard
toothing. On the cutting pitch circle of the gear, having radius r ¼ ðz=2Þm0 , we
have: tooth thickness equal to half the pitch, i.e., s ¼ e ¼ p0 =2 ¼ pm0 =2; adden-
dum ha ¼ m0 ; dedendum hf ¼ 1:25 m0 ; tooth depth h ¼ ha þ hf ¼ 2:25 m0 .
6.2 Fundamentals of Profile Shift 229
This distance is also called cutter offset. Therefore, for a gear generated by a
rack-type cutter, the profile shift is equal to the cutter offset. Since then, in common
practice, we assume that a gear generated by a hob is identical to that generated by a
rack-type cutter (this, of course, without prejudice to the considerations described in
Sect. 3.11.2), it is equally clear that the profile shift is equal to the hob offset, in the
cut with a hob.
The tooth shape of the gear to be generated, characterized by a given radius
r ¼ ðz=2Þm0 of the cutting pitch circle, depends markedly on the position of the
cutting pitch line of the rack-type cutter with respect to its datum line. Figure 6.6
shows the tooth shape of two gear wheels with the same number of teeth z (and
therefore with the same cutting pitch circle), both cut with the same rack-type
cutter, having pressure angle a0 and module m0 , but the one on the right with
positive profile shift ðxm0 [ 0Þ, and the one on the left with negative profile shift
ðxm0 \0Þ.
Figure 6.7 refers to a gear wheel with z ¼ 20 teeth, cut by means of a rack-type
cutter having a module m0 and a pressure angle a0 ¼ 20 . It highlights the tooth
shapes due to the profile shift, with profile shift equal and opposite ðx ¼ 0:5Þ,
which are compared with the tooth shape related to a zero-profile shift coefficient
ðx ¼ 0Þ. Figure 6.8, which refers to the same gear of Fig. 6.7, shows the change of
the ratio qx =qx¼0 , as a function of the profile shift coefficient x; qx and qx¼0 are the
radii of curvature of the tooth profiles on the cutting pitch circle related respectively
to any value, x, of the profile shift coefficient and to a profile coefficient equal to
zero.
Both Figs. 6.7 and 6.8 show that a positive profile shift causes an increase of
tooth thickness to its bottom land (but also a decrease of tooth thickness to its top
land), while a negative profile shift causes a reduction in the tooth thickness to its
bottom land. Therefore, with large positive profile shift we can fall into the danger
of the pointed tooth, while with a large negative profile shift we can meet the danger
Fig. 6.6 Tooth shapes of two gear wheels with the same number of teeth, the one on the right
with positive profile shift, and the one on the left with negative profile shift
6.2 Fundamentals of Profile Shift 231
Fig. 6.7 Correlation between tooth shape and the profile shift coefficient
Fig. 6.8 Correlation between ratio qx =qx¼0 and profile shift coefficient
232 6 Profile Shift of Spur Gear Involute Toothing …
of undercut. Figure 6.8 shows that the radius of curvature of the tooth profile on the
cutting pitch circle increases when the profile shift coefficient increases, and
decreases with decreasing of it.
Generally, rebus sic stantibus, we adopt a positive profile shift in the following
cases:
• to avoid cutting interference, which weakens the tooth to its root and reduces the
contact ratio;
• to improve the overall condition of engagement and sliding;
• to have good load carrying capacity in terms of surface durability (pitting) and
root bending strength for both the members of the gear pair;
• to achieve the appropriate center distance.
The negative profile shift, which causes a deterioration of load carrying capacity
in terms of surface durability and root bending strength, is used when it is necessary
to obtain a predetermined center distance.
We return now to consider the cutting condition characterized by any value of
the profile shift, xm0 , such as that shown in Fig. 6.5. We have already seen that,
whatever the profile shift, xm0 , the nominal pitch circle diameter of the gear to be
cut (this, during the generation motion, will roll without sliding on the cutting pith
line that was selected on the rack-type cutter) is defined by Eq. (6.1), and that the
module m0 and the pressure angle a0 of the rack-type cutter are the nominal ones of
the gear to be generated. On the cutting pitch circle, having the same radius r ¼
ðz=2Þm0 as that of the gear generated with zero-profile shift, we have: tooth
thickness different from the space width; sum of the tooth thickness and space width
equal to pm0 ; addendum ha 6¼ m0 ; dedendum hf 6¼ 1:25m0 ; tooth depth
h ¼ 2:25m0 .
On the cutting pitch line of the rack-type cutter, the tooth thickness, s0 and space
width, e0 , are no longer equal to half of the pitch, but will be given respectively by:
pm0 p
s0 ¼ 2xm0 tan a0 ¼ m0 2x tan a0 ð6:3Þ
2 2
pm0 p
e0 ¼ þ 2xm0 tan a0 ¼ m0 þ 2x tan a0 ; ð6:4Þ
2 2
where the subscript 0 refers to the cutting tool (ISO 6336-1: [15]).
Correspondingly, the nominal value of the tooth thickness, s, and the space
width, e, of a gear with z teeth, generated on a cutting pitch line of the rack-type
cutter characterized by a profile shift equal to xm0 , will be given by:
pm0 p
s¼ þ 2xm0 tan a0 ¼ m0 þ 2x tan a0 ð6:5Þ
2 2
pm0 p
e¼ 2xm0 tan a0 ¼ m0 2x tan a0 : ð6:6Þ
2 2
6.2 Fundamentals of Profile Shift 233
The diameter of the base circle of the gear generated is given by Eq. (6.2). To
calculate the tip and root circle diameters of the gear, we must keep in mind that a
rack-type cutter, which is cutting a gear with a profile shift xm0 , behaves as if it had
addendum ha0 and dedendum hf 0 given respectively by:
ha0 ¼ m0 ð1 xÞ ð6:7Þ
5
hf 0 ¼ m 0 þx : ð6:8Þ
4
Since a clearance of ð5=4Þm0 should remain between the tip circle of the gear
and the root line of the rack-type cutter, the diameters of the tip circle and root circle
of the gear will be given by:
Considering Eq. (1.18), the reference center distance, a0 (also called standard or
nominal center distance), i.e. the sum of the radii of the cutting pitch circles of the
two members of an external gear pair, having respectively z1 and z2 teeth, is given
by:
z þ z
1 2
a0 ¼ m0 : ð6:11Þ
2
When the working center distance, a0 , of a gear is equal to the reference center
distance, a0 , the working pitch circles (or pitch circles) coincide with the cutting
pitch circles. Evidently, to have the correct kinematic operating conditions when the
center distance is the reference one, the sum of the tooth thicknesses of the two
members of the gear pair, measured on the respective cutting pitch circles, must be
equal to pm0 , that is:
where s10 and s20 are the thicknesses on the cutting pitch circles, and s01 and s02 are
the thicknesses on the pitch circles (it is to be noted that, in agreement with the ISO
standards, the prime indicates quantities concerning the operating conditions; the
same standards use the subscript, w, as equivalent to the prime).
Whenever ðs01 þ s02 Þ will be different from pm0 , the operating center distance, a0 ,
of the gear will differ from the reference center distance, a0 . If the two members of
the gear have normal toothing, i.e. zero-profile-shifted toothing, the working center
distance will be apparently equal to the reference center distance. Instead, if the two
wheels have profile-shifted toothing, the two following cases may occur:
234 6 Profile Shift of Spur Gear Involute Toothing …
The easiest method to obtain profile-shifted toothings for an external spur gear pair
consists of the choice of a positive profile shift coefficient, x1 , for cutting of the
pinion teeth, and, for cutting of the wheel teeth, a negative profile shift coefficient
equal and opposite to that of the pinion, i.e. x2 ¼ x1 . So, with this method, we
have (see Giovannozzi [12], Pollone [25], Henriot [13]):
ðx1 þ x2 Þ ¼ 0: ð6:13Þ
External gear pairs obtained with this method, which is also known as long-short
addendum system [19], are nonstandard gears. With this method, the cutting pitch
circles, having respectively diameters equal to z1 m0 and z2 m0 , can be taken as
working pitch circles, that is, as pitch circles of the two wheels in their mutual
engagement. This because the pressure angle of the involute curves on the cutting
pitch circles has the same value, a0 , for both the gear members, and the sum of the
teeth thicknesses, ðs10 þ s20 Þ, on these pitch circles is equal to the pitch, p0 ¼ pm0 .
In fact, just this has to happen, so that we can have a backlash-free contact, that is a
meshing without backlash and no penetration between teeth.
The tooth thickness (of course, the circular tooth thickness) of the pinion,
measured along the cutting pitch circle, according to the Eq. (6.5), is given by:
pm0
s10 ¼ þ 2x1 m0 tan a0 : ð6:14Þ
2
Similarly, the tooth thickness of the wheel, measured along its cutting pitch
circle, also considering the Eq. (6.13), is given by:
pm0 pm0
s20 ¼ þ 2x2 m0 tan a0 ¼ 2x1 m0 tan a0 : ð6:15Þ
2 2
6.3 Profile-Shifted Toothing Without Variation of the Center … 235
Adding member to member the Eqs. (6.14) and (6.15), and considering
Eq. (6.13), we get Eq. (6.12), which therefore is satisfied. Therefore, we infer that,
if the two members of an external gear pair have profile shift coefficients equal to x1
and x2 ¼ x1 respectively, and tooth thicknesses respectively equal to the nominal
values corresponding to the profile shifts x1 m0 and x2 m0 ¼ x1 m0 , they will engage
without backlash at the reference center distance, a0 , and with the pressure angle of
generation, a0 .
With this method, compared with the case of zero-profile-shifted toothing, the
pinion and the wheel retain unaltered the module, pitch circles and center distance.
Table 6.1 summarizes, for an external spur gear pair, the characteristic quantities of
the profile-shifted toothing without variation of center distance, and compares them
with the corresponding quantities of the zero-profile-shifted toothing.
These characteristic quantities can also be deduced regarding the two members
of the gear pair (Fig. 6.9) as if they worked on the opposite flanks of the profile of
the common basic rack, and considering that all three members have a common
pitch point, and common points of contact P1 and P2 .
Table 6.1 Quantities of the profile-shifted toothing compared with the corresponding ones of the
zero-profile-shifted toothing, for an external spur gear pair
Quantity Profile-shifted toothing without variation of the Zero-profile-shifted
center distance toothing
Pinion Wheel Pinion Wheel
Addendum ha1 ¼ m0 ð1 þ x1 Þ ha2 ¼ m0 ð1 x1 Þ ha1 ¼ ha2 ¼ m0
Dedendum hf 1 ¼ m0 ð1:25 x1 Þ hf 2 ¼ m0 ð1:25 þ x1 Þ hf 1 ¼ hf 2 ¼ 1:25m0
Tooth depth h1 ¼ 2:25 m0 h2 ¼ 2:25 m0 h1 ¼ h2 ¼ 2:25 m0
Profile shift coefficient x1 x2 ¼ x1 x1 ¼ x2 ¼ 0
z þ z2
Center distance a0 ¼ a0 ¼ m0 1
2 a0 ¼ m0 z1 þ2 z2
Pressure angle a0 ¼ a0 a0
Fig. 6.9 External gear pair with profile-shifted toothings without variation of center distance
236 6 Profile Shift of Spur Gear Involute Toothing …
In the gears with profile-shifted toothing without variation of the center distance,
the length ga of the path of contact remains almost unchanged compared to that of
the corresponding gears with zero-profile-shifted toothing. The path of contact AE
(Fig. 6.10) moves on the line of action in the direction from T 1 to T 2 . Now we
know that the primary interference that occurs first is that related to the point T 1 .
This interference occurs in the approach stage of the meshing cycle when the pinion
is driving, and is due to an excessive value of the wheel addendum, which brings
the point A beyond the point T 1 , in the direction from C to T 1 (Fig. 6.10).
Figure 6.10 shows, to the right, an external spur gear pair with
zero-profile-shifted toothing. The tip circle of the wheel intersects the line of action
at the point A, which is located well beyond the interference point T 1 , at the outside
of the segment T 1 T 2 (T 2 is not shown in the figure). Therefore, primary interference
during the approach stage of meshing cycle occurs. To avoid this interference, it is
necessary to reduce the addendum of the wheel, so that its tip circle intersects, at
most, the line of action at point T 1 , increasing the dedendum of the same wheel
without altering the tooth depth, and using, for the pinion, a profile shift equal and
opposite to that of the wheel.
The same Fig. 6.10 shows, to the left, the same gear pair with profile-shifted
toothing without variation of the center distance, with profile shift corresponding to
the aforementioned limit condition. Therefore, the point A comes to coincide with
the point T 1 , and profile shift coefficients of pinion and wheel are equal to those
usually mostly recommended for this type of profile shift, i.e. x1 ¼ 0:50 and
x2 ¼ x1 ¼ 0:50. The figure shows at least two of the effects of this type of
profile shift, such as the strengthening of the pinion tooth to the root, and the
elimination of the undercut.
6.3 Profile-Shifted Toothing Without Variation of the Center … 237
The profile-shifted toothing without variation of the center distance has other
interesting features that we will point out every time the opportunity will present
itself. We have already said that, from the point of view of the cutting technology, it
is obtained with the same simple generation processes and with the same cutting
tools used for the zero-profile-shifted toothing. We also said that the tooth thickness
measured along the cutting pitch circles varies by an amount equal to 2xm0 tan a0 .
This thickness variation, measured along the normal to the profile, is equal to
ð2xm0 tan a0 Þ cos a0 ¼ 2xm0 sin a0 , which also represents the thickness variation on
the base circle.
Generally, with this type of profile shift, the continuity of the motion is not
compromised, since the length of the path of contact remains substantially
unchanged. It also allows, according to what we have seen above, to avoid primary
interference in the stage of approach of the meshing cycle. However, since this type
of profile-shift causes a displacement of the path of contact towards the interference
point T 2 , we may encounter the danger of interference in the stage of recess of the
meshing cycle. Therefore, this type of profile shift cannot be used when an inter-
ference in the stage of recess was initially present.
Consequentially, before proceeding with the definition of the profile shift to be
made, it is necessary to ensure that interference in the stage of recess is not present,
verifying that the minimum number of teeth, z2min , of the driven wheel is greater
than that given by Eq. (4.9) for k1max ¼ 1. In addition, once the profile shift to do is
defined, it is necessary to ensure that the interference does not appear in the recess
stage. In this regard, the minimum number of teeth, z02min , of the driven wheel
should be greater than that given by the same Eq. (4.9), but for k1max ¼ ð1 þ x1 Þ.
Another drawback that can occur with this type of profile shift is the excessive
reduction of the tip thickness of the pinion tooth, measured on the addendum circle.
We must therefore ensure that this thickness does not fall below the well-defined
limit values, set by current standard. We will see later how to do this type of
checking (see Sect. 6.5).
If we denote by z0 the minimum number of teeth that we can cut, without
interference, using a rack-type cutter having standard sizing and predetermined
module, m0 , and pressure angle, a0 , for cutting without interference the teeth of the
pinion and wheel characterized by profile-shifted coefficients x1 and x2 ¼ x1 , the
latter must have minimum numbers of teeth respectively equal to z0 ð1 x1 Þ and
z0 ð1 x2 Þ ¼ z0 ð1 þ x1 Þ. The pinion will therefore have, at minimum, z0 ð1 x1 Þ
teeth, while the wheel will have, at minimum, z0 ð1 þ x1 Þ teeth. Therefore, when this
method is used, the sum of the number of the teeth of the pinion and wheel cannot
fall below the value 2z0 . In practice, however, admitting a slight cutting interference
according to what previously said (see Sect. 3.2, and Fig. 3.2), the sum above
cannot be below of ð5=6Þ2z0 ¼ ð5=3Þz0 .
Moreover, for a given value of the profile shift coefficient, ðx1 ¼ x2 Þ, the gear
ratio u ¼ z2 =z1 cannot fall below the value given by:
238 6 Profile Shift of Spur Gear Involute Toothing …
z0 ð1 þ x1 Þ 1 þ x1
u¼ ¼ : ð6:16Þ
z0 ð1 x1 Þ 1 x1
The value of the transverse contact ratio, ea , is obtained from Eq. (3.44), with
k1 ¼ ð1 þ x1 Þ and k 2 ¼ ð1 þ x2 Þ ¼ ð1 x1 Þ. According to Eq. (6.16), the profile
shift usually most recommended, which is the one proposed by Lasche [17],
characterized by x1 ¼ 0:50 and x2 ¼ x1 ¼ 0:50, cannot be used for a gear ratio
less than 3. In practice, this toothing is convenient for gear ratios somewhat greater
than 3, and typically higher than 4.
We must also keep in mind that, with positive profile shifts, the tooth thickness
at its root increases with the increase of the positive value of x, but the radius of the
fillet decreases. Consequently, the tooth capacity to withstand bending loads and
fatigue stress increases as long as the first effect (increase of the thickness) overrides
the second effect (reduction of the fillet radius), and this occurs for small numbers
of teeth. As Fig. 6.8 shows, with positive profile shifts, the radii of curvature of the
profiles increases, and then the capacity to withstand surface loads grows propor-
tionally. For negative profile shifts, the opposite effects to those described just now
occur.
With appropriate choice of profile shift coefficients, we can still obtain satis-
factory distribution curves of the specific sliding and sliding velocity, as well as a
good load carrying capacity for the pinion and wheel. With positive profile shift, the
pinion is strengthened, while the wheel with a sufficiently large numbers of teeth is
only slightly weakened by the negative profile shift, and in certain circumstances
even strengthened, because of the positive increase in the fillet radius. With values
about the unit of the gear ratio ðu ffi 1Þ, gears with positive-shifted toothing without
variation of the center distance are however not rational, and they are therefore to be
avoided.
Finally, it is to be noted that the gears with profile-shifted toothing without
variation of the center distance have not characteristics of catalogue gears; they are
in fact non-standard gears. Since then the center distance is equal to that of the gear
with zero-profile-shifted toothing, any fixed center distance cannot be maintained
for spur gears, using standard modules.
If the number of teeth of the pinion of an internal spur gear pair is less than the
minimum number of teeth to avoid interference, it is necessary to use gear with
profile-shifted toothing, without or with variation of center distance. In usual
technical applications, the use of internal gears with profile-shifted toothings is
usually advantageous, for which the zero-profile-shifted toothing constitutes, in
fact, not the norm, but rather a rarity. However, for an internal gear pairs, in contrast
to what happens for an external gear pair, it is not always possible to freely share
6.3 Profile-Shifted Toothing Without Variation of the Center … 239
the profile shift coefficients, x1 and x2 , and simultaneously obtain optimal values of
the specific slidings and contact ratios. In this respect, in fact, the tendency of the
internal gear teeth to interfere between them imposes considerable limitations (see
Chap. 5, and Schreier [28], Dudley [8], Polder [24]).
For internal gears, the determination of the profile shift coefficients is not so
easy. Generally, it is necessary to proceed by trial and error, case by case, since the
interference problem constitutes a conditioning priority. In this regards, it should be
remembered that, for external gear with standard sizing, the minimum number of
teeth of the pinion to avoid interference with the mating wheel decreases when
ð1=uÞ ¼ z1 =z2 increases, and then u decreases. Instead, for internal gears, the
opposite happens: the minimum number of teeth of the pinion to avoid interference
with the annulus increases when ð1=uÞ ¼ z1 =z2 increases, and then u decreases. For
example, for a pressure angle a ¼ 20 and standard sizing ðk2 ¼ 1Þ, from Fig. 3.1
we obtain, with u ¼ z2 =z1 ¼ 1:5 and then ð1=uÞ ¼ 0:667, z1min ¼ 14, for the
external gear pair, and z1min ¼ 26, for the internal gear pair. Instead, for u ¼
z2 =z1 ¼ 8:0 and then ð1=uÞ ¼ 0:125, we obtain z1min ¼ 16, for the external gear
pair, and z1min ¼ 18, for the internal gear pair.
In Sect. 5.3 we have shown that, to avoid the risk of secondary interference in an
internal spur gear pair, the difference ðz2 z1 Þ between the number of teeth of the
annulus and pinion must not fall below a minimum difference value ðz2 z1 Þmin ,
which depends on the addendum factor, k, and the operating pressure angle, a0 .
However, in many practical applications, it can be useful to fall below of this limit
difference, and, in this case, it is necessary to resort to profile-shifted toothing. The
determination of the profile shift coefficients must therefore take into account both
types of interference, primary and secondary. In this section, we circumscribe the
problem to the profile-shifted toothing without variation of center distance.
We have already mentioned in Sect. 6.1 that the sign convention of the profile
shift of an internal gear is the same as that of an external gear. Obviously, for
internal gears, the generating basic rack may be just an imaginary rack. With a
profile shift equal to x1 m0 , the tooth thickness of the pinion on its cutting pitch
circle is given by Eq. (6.14). The tooth thickness of the wheel measured along its
cutting pitch circle is instead given by:
p
s20 ¼ m0 2x2 tan a0 : ð6:17Þ
2
x2 x1 ¼ 0; ð6:18Þ
z z
2 1
a0 ¼ m0 : ð6:19Þ
2
Figure 6.11 shows an internal gear with profile-shifted toothing without varia-
tion of the center distance, and profile shift coefficients, x1 and x2 , both positive and
equal to 0.50. For a discussion on the choice of the profile shift coefficients, we
refer to Sects. 6.8 and 6.9. Table 6.2 summarizes, for an internal spur gear pair,
the characteristic quantities of the profile-shifted toothing without variation of the
center distance, and compares them with the corresponding quantities of the
zero-profile-shifted toothing.
Fig. 6.11 Internal gear with profile-shiffted toothing without variation of center distance
ðx1 ¼ x2 ¼ 0:50Þ
Table 6.2 Quantities of the profile-shifted toothing compared with the corresponding ones of the
zero-profile-shifted toothing, for an internal spur gear pair
Quantity Profile-shifted toothing without variation of Zero-profile-shifted
center distance toothing
Pinion Annulus Pinion and annulus
Addendum ha1 ¼ m0 ð1 þ x1 Þ ha2 ¼ m0 ð1 þ x1 Þ ha1 ¼ ha2 ¼ m0
Dedendum hf 1 ¼ m0 ð1:25 x1 Þ hf 2 ¼ m0 ð1:25 x1 Þ hf 1 ¼ hf 2 ¼ 1:25m0
Tooth depth h1 ¼ 2:25m0 h2 ¼ 2:25m0 h1 ¼ h2 ¼ 2:25m0
Profile shift x1 x2 ¼ x1 x2 ¼ x1 ¼ 0
coefficient
z z
2 z1 2 z1
Center distance a0 ¼ a0 ¼ m0 2 a0 ¼ m 0 2
Pressure angle a0 ¼ a0 a0 .
6.4 Minimum Profile Shift Coefficient to Avoid Interference 241
In Sect. 3.1 we derived the Eq. (3.10), which allows to determine the minimum
number of teeth of a gear wheel, which must mesh with the rack of the corre-
sponding series, without operating primary interference. This equation, when
applied to the cutting condition of a gear with zero-profile-shifted toothing, having
z teeth and cut by means of a rack-type cutter having standard sizing, becomes:
k0
zmin 2 ; ð6:20Þ
sin2 a0
where k 0 and a0 are respectively the addendum factor and pressure angle of the
rack-type cutter.
About the effects of the cutting interference and consequent undercut, a
rack-type cutter, which is cutting a gear wheel with a profile shift xm0 , behaves as if
it had an addendum, ha0 , and a dedendum, hf 0 , given by:
ha0 ¼ m0 ð1 xÞ ð6:21Þ
5
hf 0 ¼ m 0 þx ð6:22Þ
4
and, therefore, as if it had k 0 ¼ ð1 xÞ. Thus, from Eq. (6.20), we infer that the
values of zmin , for which the cutting interference with the generating rack-type
cutter is avoided, are proportional to ð1 xÞ. Therefore, the values of zmin can be
taken from Fig. 3.1, by multiplying by ð1 xÞ those obtained from the same figure
for a ratio ðz1 =z2 Þ ¼ 0: This is because Fig. 3.1 refers to an addendum factor
k2 ¼ k0 ¼ 1, while in this case we have an addendum factor k 0 ¼ ha0 =m0 ¼
ð1 xÞ. Thus, we have:
where z0 is the value derived from Fig. 3.1 for a given value of the pressure angle,
a0 , and ðz1 =z2 Þ ¼ 0: Thus, we find that, compared to the values which are obtained
from Fig. 3.1 for a given a0 and ðz1 =z2 Þ ¼ 0, zmin decreases for positive values of
the profile shift coefficient ðx [ 0Þ, and increases for negative values of it ðx\0Þ.
From Eq. (6.23) we derive that the minimum profile shift coefficient, xmin , and
therefore the minimum profile shift, ðxmin m0 Þ, of the cutting pitch line of the
rack-type cutter with respect to its datum line, which is necessary to avoid the
cutting interference when we have to cut a gear having z teeth, is given by:
242 6 Profile Shift of Spur Gear Involute Toothing …
z z0 z
xmin ¼ 1 ¼ : ð6:24Þ
z0 z0
We can obtain directly this equation remembering that, generally, the cutting
interference of a gear with z teeth, cut by means of a given rack-type cutter having
active addendum m0 , will be avoided if the following condition is satisfied. The tip
line of the rack-type cutter, i.e. the line parallel to the datum line passing through
the point where the involute ends and the rounded tip connecting the involute with
the top line starts, must pass through the interference point T 1 on the line of action.
In this limit condition, the cutting pitch line of the rack-type cutter is positioned at a
distance from the point T 1 , measured along a line perpendicular to the datum line
(Fig. 6.12), equal to [12]:
zm0 2
CH ¼ T 1 C sin a0 ¼ rsin2 a0 ¼ sin a0 : ð6:25Þ
2
In this position, the profile shift, xm0 , of the cutting pitch line with respect to the
datum line is equal to:
zm0 2
xm0 ¼ m0 sin a0 : ð6:26Þ
2
Fig. 6.12 Cutting conditions of the gear with z teeth with a rack-type cutter having active
addendum m0
6.4 Minimum Profile Shift Coefficient to Avoid Interference 243
negligible decreases in the contact ratio and tooth thickness at its root, and, on the
other hand, it has the advantage of eliminating the involute part immediately near
the base circle to which different problem are related. For this reason, instead of
Eq. (6.24), the following relationship is used:
ð5=6Þz0 z
xmin ¼ : ð6:27Þ
z0
The lower limit, the negative one, is related to the need to avoid undercut.
Actually, although gears with undercut are sometimes used, it is best to design gears
which have not undercut teeth. The upper limit, the positive one, is related to the
need to avoid too pointed teeth. In fact, when the profile shift coefficient increases,
the corresponding increase in the addendum determines a reduction in the tooth
thickness at the tip circles. This reduction can become unsustainable when it falls
below the limit values fixed by the standards (in the usual practice of gear design,
the minimum value of the tooth thickness at the tip circle below which we must not
fall is equal to 0:25m).
In addition to the above-mentioned requirements, which are related to the cutting
conditions, other numerous and important requirements must be met, related to the
assembly and operating conditions. The main ones of these requirements, which
depend primarily on the tooth thickness and addendum values, are: the need to
avoid operating interference, that never can be tolerated; the need to have an
adequate amount of backlash; the need to ensure a suitable operating depth and
contact ratio, and, simultaneously, a sufficient clearance at the root circles of each
gear.
The determination of the optimum values of the profile shift coefficient based on
the above requirements is of a certain complexity, but represents a striking stage of
the gear design. We must however point out that, although most of the gears are
designed using profile shift coefficients that fall within the limits defined by the
inequality (6.26), many exceptions exist for which the aforementioned limits are not
satisfied, and the number of these exceptional cases is growing in recent years.
In the previous Section, we said that the upper positive limit of the profile shift
coefficient is related to the need to avoid pointed teeth. In fact, with increase of
positive profile shift coefficient, the crest of pinions with small number of teeth
tends to run to a point. Therefore, if the wheel to be generated has few teeth and, to
avoid the cutting interference, we choose a positive profile shift coefficient too high,
the tooth could may have pointed end (a cusp), that is, the tooth crest-width at tip
circle may drop to zero. This pointed termination of the tooth should be avoided
because it can cause breakage (especially for case-hardened gear wheels), and
malfunctioning of the gear. We have therefore the need to find the relationship
between x1 and z1 (remember that, generally, the pinion is characterized by a
positive profile shift coefficient) which defines the condition of the pointed tooth.
The tooth of the gear wheel with z1 teeth, generated with a profile shift coeffi-
cient x1 , on the cutting pitch circle has a thickness given by Eq. (6.14), to which a
tooth thickness angle corresponds, given by:
s10 p x1
2#s0 ¼ ¼ þ 4 tan a0 ; ð6:29Þ
r 10 z1 z1
where #s0 is the tooth thickness half angle on the cutting pitch circle (Fig. 2.12),
and r 10 ¼ z1 m0 =2. Remembering that the angular pitch, s1 (in radians), of the
pinion (i.e. the quotient of the angular units in a circle divided by the number of
teeth of a gear) is given by
s1 ¼ 2p=z1 ; ð6:30Þ
we obtain the following relationship which defines the space width angle:
p x1
2#e0 ¼ 4 tan a0 ; ð6:31Þ
z1 z1
where #e0 is the space width half angle (note the use of double subscript: the first
subscript, s or e, refers respectively to tooth thickness and space width, while the
second subscript, 0, refers to cutting condition).
The tooth thickness angle, 2#b1 , on the base circle of the same gear wheel is
obtained (Fig. 2.12) by adding twice the angle u0 ¼ inv a0 (see Eq. 2.2) to the
angle given by Eq. (6.29). Then we obtain:
p x1
2#b1 ¼ 2#s0 þ 2inv a0 ¼ þ 4 tan a0 þ 2inv a0 : ð6:32Þ
z1 z1
On the tip circle, whose radius r a1 ¼ d a1 =2 is given by Eq. (6.9), the pressure
angle, aa1 , at the tip point is given by (see Eq. 2.3):
6.5 Pointed Teeth and Tooth Thickness 245
r b1 z1
cos aa1 ¼ ¼ cos a0 : ð6:33Þ
r a1 z1 þ 2ð1 þ x1 Þ
The condition of pointed tooth on this tip circle is equal to impose the equality to
zero of the tooth thickness angle 2#a1 on the same tip circle, that is:
p x1
2#a1 ¼ 2#b1 2 Pd
0 OV ¼ þ 4 tan a0 þ 2inv a0 2inv aa1 ¼ 0; ð6:34Þ
z1 z1
where Pd0 OV is the tooth thickness half angle at the tooth apex V (Fig. 2.12). From
this equation, we get:
p x1
inv aa1 ¼ inv a0 þ þ 2 tan a0 : ð6:35Þ
2z1 z1
Using Eq. (6.36) and (6.37), it is possible to obtain the curve that gives the
values of the profile shift coefficient x1 for which a gear with z1 teeth, generated by
a rack-type cutter having pressure angle a0 , has pointed teeth. Figure 6.13 shows,
left at the top, the two limit curves which give, as a function of z1 , the profile shift
coefficients for which the teeth are pointed, for pressure angles a0 ¼ 15 and a0 ¼
20 : They represent the maximum theoretical values, obviously not to be exceeded.
In the same figure, other four curves are shown, and precisely:
• the curve 1 and 10 that give respectively the theoretical and practical profile shift
coefficients to avoid the cutting interference, when a rack-type cutter with
pressure angle a0 ¼ 15 is used;
• the curve 2 and 20 that give respectively the theoretical and practical profile shift
coefficients to avoid the cutting interference, when a rack-type cutter with
pressure angle a0 ¼ 20 is used.
The tooth thickness sa1 and sf 1 on the tip and root circles of the gear can be
determined using the general Eq. (2.59), taking into account Eq. (6.56) which
defines inv c1 (see Sect. 6.6.1). On the tip circle, the tooth thickness should not fall
below the limit values set by international standards. Figure 6.14 shows, in quan-
titative terms, as the profile shift affects both aforesaid tooth thicknesses, sa and sf ,
246 6 Profile Shift of Spur Gear Involute Toothing …
Fig. 6.13 Limit curves of pointed teeth, and values of their profile shift coefficients to avoid the
cutting interference, with rack-type cutters having pressure angle a0 ¼ 15 and a0 ¼ 20
for values of the profile shift coefficients in the range of variability given by
inequality (6.28).
As Eq. (6.32) shows, the maximum value of the tooth thickness angle is the one
we have on the base circle, given by the same Eq. (6.32). When the root circle of
the gear is inside the base circle, the parts of the tooth between the base circle and
the root circle, the radius of which is then smaller than that of the first, have a
decreasing thickness. Theoretically, the tooth flanks of these parts inside the base
circle are made with radial profile.
In some cases, one of the requirements of the design is to get the maximum
strength of the pinion teeth. In these cases, the designer must set the
geometrical-kinematic design of the gear in such a way that the root circle coincides
with the base circle, i.e. d b ¼ d f . To this end, just match the diameters of these two
circles, given respectively by Eqs. (6.2) and (6.10), i.e.:
6.5 Pointed Teeth and Tooth Thickness 247
Fig. 6.14 Ratios sa =m and sf =m as a function of the profile shift coefficient, x, and number of
teeth, z; limit curves of pointed teeth, and cutting interference
5
z1 m0 cos a0 ¼ m0 z1 2 x1 : ð6:38Þ
4
5 z1
x1 ¼ ð1 cos a0 Þ: ð6:39Þ
4 2
The most general method to obtain profile-shifted toothing for an external gear pair
is to assume values of the profile shift coefficient, x1 and x2 , which are not related
by Eq. (6.13), but determined based on the considerations described below, and
other observations that will be described in the following sections. In this case, by
248 6 Profile Shift of Spur Gear Involute Toothing …
cutting the two wheels of the gear pair for which profile shift coefficients x1 6¼ x2
were chosen, the thicknesses of the teeth measured on the cutting pitch circles,
having respectively radii r 10 ¼ z1 m0 =2 and r 20 ¼ z2 m0 =2, will be different and their
sum will not be equal to the pitch p0 ¼ pm0 of the rack-type cutter.
Therefore, if we place the wheels axes to the reference center distance, a0 , given
by Eq. (6.11), so that the cutting pitch circles are tangent, we get a correct kine-
matic behavior, but the sum of the tooth thickness is not equal to the pitch. The
operating condition of backlash-free contact between the teeth is not satisfied, while
the condition for a correct kinematic operation is satisfied. In fact, the kinematics is
correct whatever the distance between the axes, as the tooth profiles are arcs of
involute curves. Therefore, we conclude that the operating pitch circles to be
determined (as we already stated elsewhere, just pitch circles), will necessarily be
different from the cutting pitch circles.
We denote by r10 and r20 the unknown radii of these operating or working pitch
circles. The center O1 and O2 of these two pitch circles must be positioned at an
operating center distance, a0 , this also to be determined. From Eq. (6.5), we obtain
that the teeth thicknesses of the two gear wheels on the cutting pitch circles are related
to the profile shift coefficients, x1 and x2 , by the following relationships (note that, in
accordance with ISO 6336-1:2006, in this textbook we indifferently use the prime, or
subscript, w, to indicate quantities that relate to operating or working conditions):
p
s10 ¼ m0 þ 2x1 tan a0 ð6:40Þ
2
p
s20 ¼ m0 þ 2x2 tan a0 : ð6:41Þ
2
in this case, the operating center distance, a0 , operating module, m0 , and oper-
ating pressure angle, a0 , will necessarily be greater than the corresponding
reference values, that is: a0 [ a0 ; m0 [ m0 ; a0 [ a0 .
• The sum of the two profile shift coefficients is less than zero ðx1 þ x2 \0Þ, for
which:
in this case, the operating center distance, a0 , operating module, m0 , and oper-
ating pressure angle, a0 , will necessarily be less than the corresponding reference
values, that is: a0 \a0 ; m0 \m0 ; a0 \a0 .
• The sum of the two profiles shift coefficients is equal to zero ðx1 þ x2 ¼ 0Þ. This
case is the one discussed in Sect. 6.3; it therefore constitutes a special case of the
general discussion that follows.
In the first two cases, those for which the sum ðx1 þ x2 Þ is different from zero, we
have to determine the operating center distance a0 6¼ a0 for which we have a
meshing with backlash-free contact between the teeth, or, in equivalent terms, the
operating module m0 6¼ m0 or the operating pressure angle a0 6¼ a0 . Figure 6.15
shows the meshing diagram in an external spur gear pair with ðx1 þ x2 Þ [ 0. It
highlights the two operating pitch circles that are touching at pitch point C, and all
other circles and quantities that affect the solution of our problem. Note that, from
now on, in accordance with the vocabulary of gear terms of the ISO 1122-1: [14],
we will omit to premise the adjective “operating” for all the quantities of interest,
but in doing so we always mean quantities related to the operating conditions; in
this section, this quantities are indicated with the prime.
For the determination of these quantities that are interdependent between them,
since they are closely related to each other (therefore, just determine one on them to
obtain the other), we note first that the radii r b1 and r b2 of the base circles are given by:
where all known quantities appear. The radii of the pitch circles are given by:
where a0 is the pressure angle, which is unknown. The circular pitch, p0 , also
unknown, is given by:
where
cos a0
m0 ¼ m0 ð6:49Þ
cos a0
From Eq. (2.59), written for both the wheels in the operating meshing condi-
tions, in which the radii r10 and r20 , and the pressure angle, a0 , come into play, we get:
The tooth thickness of the two wheels, measured on their cutting pitch circles,
are given by the relationships (6.40) and (6.41). According to the general Eq. (2.59
), these relationships may also be written in the form:
6.6 Profile-Shifted Toothing with Variation of Center … 251
By equating the two relationships which express in two different but equivalent
ways the thicknesses on their cutting pitch circles, i.e. equating Eq. (6.40) with
Eq. (6.54), and Eq. (6.41) with Eq. (6.55), we obtain the following equations which
give inv c1 and inv c2 as a function of quantities all known:
2 p
inv c1 ¼ inv a0 þ þ x1 tan a0 ð6:56Þ
z1 4
2 p
inv c2 ¼ inv a0 þ þ x2 tan a0 : ð6:57Þ
z2 4
x1 þ x2
inv a0 ¼ inv a0 þ 2 tan a0 : ð6:58Þ
z1 þ z2
So, once calculated the pressure angle a0 , we have also determined the module,
m , and the radii r10 and r20 of the pitch circles, given by Eqs. (6.49), (6.46) and
0
1 z1 þ z2 cos a0 cos a0
a0 ¼ r10 þ r20 ¼ ðz1 þ z2 Þm0 ¼ m0 0
¼ a0 : ð6:59Þ
2 2 cos a cos a0
p x2
2#b2 ¼ þ 4 tan a0 þ 2inv a0 ; ð6:60Þ
z2 z2
r10 z1
¼ ð6:61Þ
r20 z2
Since the angular pitches, s1 and s2 , of the two gears are given by
s1 ¼ 2p=z1 ð6:64Þ
s2 ¼ 2p=z2 ; ð6:65Þ
the tooth thickness angles 2#0s;1 and 2#0s;2 , and space width angles 2#0e;1 and 2#0e;2 of
the same two gears, on their operating pitch circles, can be written in the form:
p x1
2#0s;1 ¼ þ 4 tan a0 þ 2ðinv a0 inv a0 Þ ð6:66Þ
z1 z1
p x2
2#0s;2 ¼ þ 4 tan a0 þ 2ðinv a0 inv a0 Þ ð6:67Þ
z2 z2
2p
2#0e;1 ¼ s1 2#0s;1 ¼ 2#0s;1 ð6:68Þ
z1
2p
2#0e;2 ¼ s2 2#0s;2 ¼ 2#0s;2 : ð6:69Þ
z2
The Eqs. (6.66) and (6.67) are obtained by using, for each of the two mating
gears, the same procedure that allowed us to derive Eq. (6.34), while Eqs. (6.68)
and (6.69) are obtained, for each gear wheel, as the difference between the angular
pitches and tooth thickness angles.
6.6 Profile-Shifted Toothing with Variation of Center … 253
To have the tooth profiles of the mating gear wheels in backlash-free contact, on
the operating pitch circles the space width between the teeth of the first gear must be
equal to the tooth thickness of the second gear. Therefore, it must be:
r20 0 z2
#0e;1 ¼ # ¼ #0 : ð6:71Þ
r10 s;2 z1 s;2
From this relationship, developed taking into account Eqs. (6.66) to (6.68), we
obtain the fundamental Eq. (6.58), previously deduced by another procedure.
We now analyze, for example, what happens for ðx1 þ x2 Þ [ 0, with x1 and x2
both greater than zero, when we place the two gears, cut with these profile shift
coefficients, so that their center distance, a, is equal to the sum of the radii r 1 and r 2
given by (see also Henriot [13]):
z
1
r 1 ¼ r 10 þ x1 m0 ¼ m0 þ x1 ð6:72Þ
2
z
2
r 2 ¼ r 20 þ x2 m0 ¼ m0 þ x2 : ð6:73Þ
2
ðz1 þ z2 Þ
a ¼ r 1 þ r 2 ¼ m0 þ m0 ðx1 þ x2 Þ ¼ a0 þ m0 ðx1 þ x2 Þ: ð6:74Þ
2
Figure 6.16 shows the two gear wheels, generated by the same rack-type cutter
with profile shift coefficients x1 [ 0 and x2 [ 0, and placed at center distance
O1 O2 ¼ a ¼ ðr 1 þ r 2 Þ. The figure highlights that, with this center distance, a, there
is no contact between the tooth profiles of the pinion and wheel. Therefore, it is
necessary to approach the two members of the gear pair up to the operating center
distance, a0 (hereafter, the center distance), which is the theoretical one for which
the backlash becomes equal to zero. From the aforementioned considerations, we
infer the following important property: the variation of the center distance, Da0 ¼
ða0 a0 Þ i.e. the difference between the center distance, a0 , and the reference center
distance, a0 , is always less than the sum of the profile shifts, ðx1 þ x2 Þm0 , that
characterize pinion and wheel.
Therefore, we have:
Fig. 6.16 External gear with profile shift coefficients x1 [ 0 and x2 [ 0, and gear wheels placed
at center distance a ¼ r 1 þ r 2
2ðx1 þ x2 Þ
Da ¼ ða a0 Þ ¼ a0 ¼ a0 X ð6:77Þ
ðz1 þ z2 Þ
where
2ðx1 þ x2 Þ
X¼ : ð6:78Þ
ðz1 þ z2 Þ
a ¼ a0 ð1 þ XÞ: ð6:79Þ
Taking into account the Eqs. (6.59) and (6.11), the variation of center distance
Da0 can be written in the following form:
6.6 Profile-Shifted Toothing with Variation of Center … 255
cos a
Da0 ¼ ða0 a0 Þ ¼ a0
0
1 ¼ a0 Y ð6:80Þ
cos a0
where
cos a
0
Y¼ 0
1 : ð6:81Þ
cos a
a0 ¼ a0 ð1 þ YÞ: ð6:82Þ
a a0 ¼ a0 ðX YÞ ¼ Km0 ; ð6:83Þ
where
z1 þ z2 h z1 þ z2 cos a0 i
K¼ ðX YÞ ¼ ðx1 þ x2 Þ 1 : ð6:84Þ
2 2 cos a0
Factor K is the truncation coefficient, i.e. the quotient of the truncation divided
by the module m0 , while Km0 is the truncation, i.e. the reduction of the addendum,
considering the addendum defined by the standard basic rack tooth profile.
Truncation is also known as topping, and represents the addendum reduction of the
two members of the gear, which is necessary if we want to have the standard
clearance, c, equal to 0:25 m. Therefore, the tooth depth will be given by:
while the tip circle diameters of the two wheels are respectively
Suppose, for a moment, that the truncation is not made. In this case, the clear-
ance, c, i.e. the distance, along the line of centers, between the tip surface of the first
gear and the root surface of its mating gear, can be calculated by subtracting from
the operating center distance, a0 , given by Eq. (6.59), the tip radius of the first of the
two wheels, r a1 , given by:
m0
r a1 ¼ ½z1 þ 2ð1 þ x1 Þ; ð6:90Þ
2
m0 h z1 þ z2 cos a0 i
c ¼ a0 r a1 r f 2 ¼ m0 ðx1 þ x2 Þ 1
4 2 cos a0 ð6:92Þ
m0 m
½m0 ðx1 þ x2 Þ ða0 a0 Þ ¼
0
¼ Km0 :
4 4
Therefore, if Km0 [ 0, the clearance, c, would be less than the standard value,
equal to ðm0 =4Þ. In this case, the clearance is reset to the standard value with the
above-described truncation. However, when the designer feels abundant this
clearance standard value, at his discretion, the amount of topping can be reduced.
This reduction is allowed, provided that other crest-width limitations do not con-
dition the choice of the clearance value to be adopted. In any case, the chosen
clearance value is to be checked.
The truncation Km0 can be achieved by another method, which allow us to
define other quantities of the profile-shifted gears with variation of center distance
(see Giovannozzi [12]). As Fig. 6.16 shows, in the cutting conditions, the deden-
dum of the two members of the gear pair are respectively given by (note that the
third subscript, 0, refers as usually to cutting conditions):
5
hf 1;0 ¼ x1 m 0 ð6:93Þ
4
5
hf 2;0 ¼ x2 m 0 : ð6:94Þ
4
z1 0 z1 cos a0
y1 ¼ g1 m0 ¼ ðr10 r 10 Þ ¼
ðm m0 Þ ¼ m0 1 ð6:95Þ
2 2 cos a0
z2 z2 cos a0
y2 ¼ g2 m0 ¼ r20 r 20 ¼ ðm0 m0 Þ ¼ m0 1 : ð6:96Þ
2 2 cos a0
When the pitch circles are in contact, the radial distance, y, between the cutting
pitch circles is equal to:
z1 þ z2 cos a0
y ¼ y1 þ y2 ¼ ðg1 þ g2 Þm0 ¼ m0 1 ¼ nm0 : ð6:97Þ
2 cos a0
The radial distances hf 1 and hf 2 between the pitch circles and root circles of the
two members of the gear pair are:
5
hf 1 ¼ x 1 þ g1 m 0 ð6:98Þ
4
5
hf 2 ¼ x 2 þ g2 m 0 : ð6:99Þ
4
Therefore, if between the tip circle of one member and the root circle of the other
member of the gear pair we want to have a clearance c ¼ ðm0 =4Þ, i.e. a clearance
value equal to the nominal one, the radial protrusions of the teeth beyond the pitch
circles having radii r10 and r20 must be equal to:
Instead, according to the generation method, these radial protrusions are equal to:
Therefore, it is necessary to cut the teeth in gear blanks having reduced diam-
eters, so that they have less depth, compared with the nominal ones, by an amount
given by:
The transverse contact ratio is calculated with Eq. (3.44), taking into account
that, due to the profile shifts, x1 and x2 , the addendum factors, k 1 and k2 , are
respectively given by:
cos a0
k 1 ¼ ð 1 x 2 þ g2 Þ ð6:105Þ
cos a0
cos a0
k 2 ¼ ð 1 x 1 þ g1 Þ : ð6:106Þ
cos a0
However, we must note that external gear pairs having profile-shifted toothing
with variation of center distance are mainly characterized by positive profile shift
coefficients (i.e. x1 [ 0 and x2 [ 0). In principle, with these types of gears, we use
the most favourable properties of the profile shift for both members of the gear, the
pinion and wheel. The limitations of the profile-shifted toothing without variation of
center distance are eliminated and, for a given transmission ratio, the predetermined
center distance can be maintained using standard modules, but choosing the cor-
responding sum ðx1 þ x2 Þ of profile shift coefficients. It is also possible to obtain
gear wheels having characteristics of catalogue wheels. The calculations are how-
ever more complex. Compared to an external gear with zero-profile-shifted tooth-
ing, if ðx1 þ x2 Þ [ 0, the contact ratio and radial component of the load are higher.
On the contrary, if ðx1 þ x2 Þ\0, the contact ratio and radial component of the load
are lower.
In Sect. 5.3 we have shown that, to avoid the risk of secondary interference in an
internal spur gear pair, the difference ðz2 z1 Þ between the numbers of teeth of the
annulus and pinion must not fall below a minimum difference value ðz2 z1 Þmin ,
which depends on the addendum factor, k, and the operating pressure angle, a0 .
However, in many practical applications, it can be useful to fall below of this limit
difference, and in this case, it is necessary to resort to profile-shifted toothing.
For an internal gear pair with profile shift coefficients x1 [ 0 and x2 [ 0, the
tooth thicknesses, s10 and s20 , on the two cutting pitch circles, having radii r 10 ¼
ðz1 =2Þm0 and r 20 ¼ ðz2 =2Þm0 , are given respectively by Eqs. (6.14) and (6.17).
Considering the two circles with radii r 1 and r 2 given by Eqs. (6.72) and (6.73), and
locating their axes at a center distance, a, equal to the difference ðr 2 r 1 Þ, we have:
ðz2 z1 Þ
a ¼ r 2 r 1 ¼ m0 þ m0 ðx2 x1 Þ ¼ a0 þ m0 ðx2 x1 Þ: ð6:107Þ
2
Figure 6.17, which shows the two gear wheels placed at center distance
O1 O2 ¼ a ¼ ðr 2 r 1 Þ, highlights that, with this center distance, a, the teeth of the
pinion and wheel intersect. Therefore, it is necessary to reduce this center distance.
Mutatis mutandis, the property already set out for external gears here becomes: the
variation of center distance Da0 ¼ ða0 a0 Þ, i.e. the difference between the center
distance and reference center distance, is always less than the difference of the
profile shifts ðx2 x1 Þm0 . Thus, we have:
Fig. 6.17 Internal gear with profile shift coefficients x1 [ 0 and x2 [ 0, and gear wheels placed at
center distance a ¼ ðr 2 r 1 Þ
Always changing what needs to be changed, by applying to the internal gear the
same method that allowed us to infer the Eq. (6.58), we find that the latter is
transformed into the following equation:
x2 x1
inv a0 ¼ inv a0 þ 2 tan a0 : ð6:109Þ
z2 z1
ðx2 x1 Þ
Da ¼ ða a0 Þ ¼ ðx2 x1 Þm0 ¼ 2a0 ¼ a0 X ð6:110Þ
ðz2 z1 Þ
where
6.6 Profile-Shifted Toothing with Variation of Center … 261
ðx2 x1 Þ
X¼2 : ð6:111Þ
ðz2 z1 Þ
and then
a ¼ a0 ð1 þ XÞ: ð6:112Þ
where Y is given by Eq. (6.81). From the latter equation, we obtain yet the
Eq. (6.82).
Comparing the Eq. (6.112) with Eq. (6.82), it is evident that a0 \a. Therefore,
the truncation is not necessary, i.e. it is not necessary to decrease the addendum of
the two members of the internal gear pair, as in the case of an external gear pair.
The tooth depth does not change and, moreover, the clearance is slightly increased,
and this is certainly an advantage.
In Sect. 6.6 we saw that Eqs. (6.58) and (6.109), respectively valid for external and
internal gear pairs, correlate three unknown quantities, a0 , x1 , and x2 , and that, to
determine these quantities, it is necessary to write, in addition to these equations,
two other equations forming, together with Eq. (6.58) or Eq. (6.109), a system of
three independent algebraic equations, all in the same unknowns. However,
whatever the design criterion, we must in any case use the Eq. (6.58) or
Eq. (6.109). Conversely, the other two equations can be written as a function of the
goals to be achieved, and the resulting design criteria related to them. In Sect. 6.6.1
we mentioned the two Schiebel’s criteria which, of course, are not the only pos-
sible. Other researchers have proposed various criteria other than Schiebel’s, aimed
at determining the optimal shape of the teeth according to the goals to be achieved.
Even the manufacturers of generating gear cutting machines have proposed their
criteria.
Therefore, the gear designer has available various formulations, which allow him
to determine the profile shift coefficients, x1 and x2 , with an accuracy more or less
262 6 Profile Shift of Spur Gear Involute Toothing …
high, but still sufficient for this type of calculation. Some of these formulations have
also been translated into schedules, tables, and diagrams, which have been devel-
oped based on different criteria and points of view. In the following, we will
describe some of these formulations. For the moment, however, temporarily we put
aside this very complex subject, and we focus our attention on so-called direct and
inverse problems. However, it is first necessary to point out that, once the profile
shift coefficients have been determined, we use Eq. (6.58) or Eq. (6.109) to
determine the last unknown, i.e. the operating pitch angle, a0 .
In the direct problem, the profile shift coefficients, x1 , and x2 , are assumed
known, as determined preliminarily by one of the optimization criteria that we have
mentioned above. The geometry of the rack-type cutter, defined by the module, m0 ,
and pressure angle, a0 , is also supposed known. The nominal center distance, a0 , of
the gear, given by Eq. (6.11) or Eq. (6.19) depending on whether the gear is
external or internal, is therefore uniquely defined, but we need to determine the
operating center distance, a0 .
For an external gear pair, sequentially the following quantities are then calcu-
lated: X, by means of Eq. (6.78); a0 , by means of Eq. (6.58); Y, by means of
Eq. (6.81); a0 , by means of Eq. (6.59); K, by means of Eq. (6.84); h, by means of
Eq. (6.85); d a1 and d a2 , by means of Eqs. (6.86) and (6.87).
Instead, for an internal gear pair, sequentially the following quantities are cal-
culated: X, by means of Eq. (6.111); a0 , by means of Eq. (6.109); Y, by means of
Eq. (6.81); a0 , by means of Eq. (6.113); K, by means of Eq. (6.84), in which we
replace ðx1 þ x2 Þ and ðz1 þ z2 Þ with ðx2 x1 Þ and ðz2 z1 Þ respectively; h, by
means of Eq. (6.85); d a1 , by means of Eq. (6.86); d a2 ¼ m0 ½z0 2ð1 x2 KÞ,
which is obtained from Eq. (6.87) written for an internal gear.
The following examples clarify how to use the above-described procedure of the
sequential calculation.
Example 1 External gear pair. Input data: z1 ¼ 10; z2 ¼ 15; a0 ¼ 15 ; m0 ¼ 10
mm. Profile shift coefficients, x1 ¼ þ ð1=2Þ and x2 ¼ þ ð1=3Þ, also known as
previously chosen. We will have: a0 ¼ 125 mm; X ¼ 0:067; a0 ¼ 23 200 ;
Y ¼ 0:052; a0 ¼ 131:46 mm; K ¼ 0:187; h ¼ 20:63 mm; d a1 ¼ 126:26 mm; d a2 ¼
172:93 mm.
Example 2 Internal gear pair. Input data: z1 ¼ 10; z2 ¼ 5 0; a0 ¼ 20 ; m0 ¼ 5 mm.
Profile shift coefficients, x1 ¼ þ ð1=2Þ and x2 ¼ 0 (i.e. wheel with
zero-profile-shifted toothing), also known as previously chosen. We will have:
a0 ¼ 100 mm; X ¼ 0:025; a0 ¼ 14 490 ; Y ¼ 0:028; a0 ¼ 97:20 mm; K ¼ 0:06
(Km0 ¼ 0:3 mm, for which the reduction of the tooth depth can be neglected);
h ¼ 11:25 mm; d a1 ¼ 65 mm; d a2 ¼ 240 mm.
Instead, in the inverse problem, we start from a value imposed of the center
distance, a0 , and a range of standard rack-type cutters, and we must first to deter-
mine the sum ðx1 þ x2 Þ or the difference ðx2 x1 Þ of the profile shift coefficients,
depending on whether the gear is external or internal. Subsequently, we must
determine the distribution of this sum or difference between pinion and wheel. The
6.7 The Sum (or Difference) of the Profile … 263
initial data are therefore the numbers of teeth, z1 and z2 , the center distance, a0 , and
the range of standard rack-type cutters with which the two members of the gear pair
must be generated. Each of these cutting tools is defined by the module, m0 , and the
pressure angle, a0 . The center distance, a0 , to be met, which is different from the
nominal center distance, a0 , given by Eq. (6.11) or Eq. (6.19), is that given by
Eq. (6.59) or Eq. (6.113), where m0 and a0 are the operating module and working
pressure angle.
The first thing to do is to choose the rack-type cutter from those of the standard
range. For this purpose, we choose initially the cutting tool with module, m0 ,
immediately below one of the following two modules, respectively calculated using
Eq. (6.59) or Eq. (6.113), depending on whether the gear pair under consideration
is an external gear or an internal gear:
Once this choice has been made, for which m0 and a0 are defined, we will
calculate the nominal center distance, a0 , by means of Eq. (6.11) or Eq. (6.19), and
we will verify if it meets the operating center distance, a0 , imposed as input data,
that is a0 ¼ a0 . Generally, however, we will find a0 6¼ a0 for which, depending on
whether a0 ?a0 , we will have to make a profile-shifted toothing with increase or
decrease in the center distance.
For an external gear pair, sequentially we will calculate the following quantities:
a0 , by means of Eq. (6.59) or Eq. (6.113), both written in the form
a
cos a0 ¼
0
cos a0 ð6:117Þ
a0
where a0 is given by Eq. (6.11) or Eq. (6.19), depending on whether the gear is
external or internal; Y, by means of Eq. (6.81); X, by means of Eq. (6.78), and
therefore the sum of the profile shift coefficients, ðx1 þ x2 Þ ¼ Xðz1 þ z2 Þ=2; K, by
means of Eq. (6.84); h, by means of Eq. (6.85); d a1 , and d a2 by means of
Eqs. (6.86) and (6.87).
Instead, for an internal gear pair, sequentially we will calculate the following
quantities: a0 , by means of Eqs. (6.117) and (6.118), with a0 ¼ m0 ðz2 z1 Þ=2; Y,
by means of Eq. (6.58); X, by means of Eq. (6.83), and therefore the difference of
the profile shift coefficients, ðx2 x1 Þ ¼ Xðz2 z1 Þ=2; K, by means of Eq. (6.84),
in which we replace ðx1 þ x2 Þ and ðz1 þ z2 Þ with ðx2 x1 Þ and ðz2 z1 Þ respec-
tively; d a1 , by means of Eq. (6.86); d a2 ¼ m0 ½z0 2ð1 x2 KÞ, which is
obtained from Eq. (6.87) written for an internal gear.
264 6 Profile Shift of Spur Gear Involute Toothing …
Example 3 External gear pair, and increase of the center distance. Input data:
z1 ¼ 20; z2 ¼ 40; a0 ¼ 303 mm; standard range of rack-type cutters. We will have:
m0 ¼ 2a0 =ðz1 þ z2 Þ ¼ 10:10 mm; m0 ¼ 10 mm; a0 ¼ 20 (this is the initial choice
of design); a0 ¼ m0 ðz1 þ z2 Þ=2 ¼ 300 mm. Since a0 [ a0 , we have to make a
profile shift with increase of the center distance. Continuing the calculation, we
have: a0 ¼ 21 300 ; Y ¼ 0:0100; X ¼ 0:0106; ðx1 þ x2 Þ ¼ 0:32; K ¼ 0:018 (in this
case, the corresponding decrease of the tooth depth, Km0 ¼ 0:18 mm, can be
neglected); h ¼ 22:5 mm; d a1 ¼ 10½20 þ 2ð1 þ x1 Þ and d a2 ¼ ½40 þ 2ð1 þ x2 Þ to
be calculated after splitting the sum ðx1 þ x2 Þ between the pinion and wheel.
Example 4 External gear pair, and decrease of the center distance. Input data:
z1 ¼ 20; z2 ¼ 40; a0 ¼ 298 mm; standard range of rack-type cutters. We will have:
m0 ¼ 2a0 =ðz1 þ z2 Þ ¼ 9:93 mm; m0 ¼ 10 mm; a0 ¼ 20 (for this initial design
choice, since m0 has a value close to 10 mm, it is not appropriate to choose the
lower standard module m0 ¼ 9 mm, because we would obtain either a pressure
angle very large, or a profile shift too high); a0 ¼ m0 ðz1 þ z2 Þ=2 ¼ 300 mm. Since
a0 \a0 , we have to make a profile shift with decrease of the center distance.
Continuing the calculation, we have: a0 ¼ 18 550 ; Y ¼ 0:0067; X ¼ 0:0065;
ðx1 þ x2 Þ ¼ 0:195; K ¼ 0:006: From now on, the observations and calculations to
be made are those of the previous examples.
Example 5 Internal gear pair. Input data:z1 ¼ 10; z2 ¼ 50; a0 ¼ 20 ; m0 ¼ 5 mm.
Profile shift coefficients x2 ¼ 0, i.e. wheel with zero-profile-shifted toothing, and
x1 ¼ 0:50 (thus, ðx2 x1 Þ ¼ 0:50), also known as previously chosen. We will
have: a0 ¼ 100 mm; X ¼ 0:025; a0 ¼ 14 490 ; Y ¼ 0:028; a0 ¼ 97:20 mm; K ¼
0:06 (in this case, the corresponding decrease of the tooth depth, Km0 ¼ 0:3 mm, is
neglected, to simplify the exercise; in fact, however, a careful analysis should be
carried out on the effects of this simplifying assumption on clearance); h ¼ 11:25
mm; d a1 ¼ 65 mm; d a2 ¼ 240 mm. From now on, the observations and calculations
to be made are those of the previous examples.
Example 6 Internal gear pair, and increase of the center distance. Input data:
z1 ¼ 10; z2 ¼ 50; a0 ¼ 102 mm; standard range of pinion-type cutters. We will
have: m0 ¼ 2a0 =ðz2 z1 Þ ¼ 5:10 mm; m0 ¼ 5 mm; a0 ¼ 20 (this is the initial
choice of design); a0 ¼ m0 ðz2 z1 Þ=2 ¼ 100 mm. Since a0 [ a0 , we have to make
a profile shift with increase of the center distance. Continuing the calculation, we
have: a0 ¼ 22 530 ; Y ¼ 0:020; X ¼ 0:021; ðx2 x1 Þ ¼ 0:42; K ¼ 0:02. From now
on, the observations and calculations to be made are those of the previous
examples.
However, regarding the cut of the annulus of this example, we must make the
following important consideration. We assume that the difference ðx2 x1 Þ ¼ 0:42
has been obtained by the following design choice: x2 ¼ 0:82 and x1 ¼ 0:40. To
manufacture this annulus in the design condition ðx2 ¼ 0:82Þ, we suppose to use a
pinion-type cutter with z0 ¼ 20 teeth and standard sizing ðx0 ¼ 0Þ. Therefore, we
have: a0 ¼ 75 mm; ðx2 x0 Þ ¼ 0:82; X ¼ 0:055; Y ¼ 0:048; a00 ¼ 78:56 mm
6.7 The Sum (or Difference) of the Profile … 265
(actual center distance at the end of the cutting process); r a0 ¼ 56:25 mm (radius of
the tip circle of the pinion-type cutter, whose addendum is ha0 ¼ 1:25m0 ¼ 6:25
mm); r f 2 ¼ a0 þ r a0 ¼ 134:81 mm (actual radius of the root circle of the annulus);
r a1 ¼ ½ðz0 þ 2x1 Þm0 =2 ¼ 32 mm (radius of the tip circle of the pinion that meshes
with the annulus); c ¼ r f 2 ða0 þ r a1 Þ ¼ 0:81 mm (clearance between the root
circle of the annulus and tip circle of the pinion).
Therefore, we infer that the actual clearance is decreased compared to the
standard clearance; in fact, it passes from c ¼ 1:25 mm to c ¼ 0:81 mm. This
decrease would have been even greater if the number of teeth of the pinion-type
cutter used had been smaller. This result is entirely consistent with what we have
said previously, namely that, in an internal gear pair, the clearance tends to increase
if we carry out a profile shift with variation of center distance. Therefore, if we
consider the pinion-type cutter instead of the mating pinion, we find that the cutting
tool does not cut deep enough. In certain cases, the truncation of the addendum of
the pinion teeth is therefore desirable.
It is time to address the problem of determining the profile shift coefficients, x1 and
x2 , or the problem of the distribution between pinion and wheel of their sum
ðx1 þ x2 Þ or difference ðx2 x1 Þ, depending on whether the gear is external or
internal. We have already anticipated that this problem, very complex, can be
approached from different points of view, to each of which a well-defined criterion
corresponds. Each of these criteria has advantages and disadvantages, and none of
them claims to be the best. The results that are obtained using any of these criteria
are therefore optimal from the point of view at the base of the criterion chosen, but
are not as optimal from the point of view of another criterion. The designer, aware
of the pros and cons associated with each criterion, must be able to extricate himself
from them, and from time to time choose the one best suited to the goals to be
achieved.
In Sect. 6.6.1, we have made a brief reference to the two conditions proposed by
Schiebel, which together make explicit the criterion that bears his name. Instead, in
this section, we describe in detail other criteria, some of which are now frequently
used in the gear design.
This criterion uses the Eq. (6.24) as fundamental relationship. In this equation (see
Sect. 6.4), z0 is the number of teeth that can be cut without interference, using a
266 6 Profile Shift of Spur Gear Involute Toothing …
In the first case we choose, for the pinion, a positive profile shift coefficient, x1 ,
determined by means of Eq. (6.24), and, for the wheel, a negative profile shift
coefficient, x2 ¼ x1 , that is equal in absolute value, but of opposite sign with
respect to that of the pinion. By means of equations set out in Sect. 6.4, we can
easily show that, if ðz1 þ z2 Þ [ 2z0 , the pinion is in the limit condition of cutting
interference (instead, the wheel is not yet in the limit condition), while, if
ðz1 þ z2 Þ ¼ 2z0 , both pinion and wheel are in the limit condition of cutting inter-
ference. In any case, since x2 ¼ x1 , we have a profile-shifted toothing without
variation of center distance.
In the second case we choose, for the pinion, a positive profile shift coefficient,
x1 , determined by means of Eq. (6.24), and, for the wheel, a positive or negative
profile shift coefficient, x2 , calculated using the same Eq. (6.24). In algebraic value,
however, we will always have x2 [ x1 , and thus ðx1 þ x2 Þ [ 0: Therefore, we
will have profile-shifted toothing with increase of center distance.
The following three examples are related to the use of a rack-type cutter with
pressure angle a0 ¼ 20 : In this case, z0 ¼ 17, and thus 2z0 ¼ 34.
Example 1 Input data, z1 ¼ 12 and z2 ¼ 28. Therefore, we have:
ðz1 þ z2 Þ ¼ 40 [ 34; x1 ¼ 5=17; x2 ¼ x1 ¼ 5=17. In this case, only the pinion
is in the limit condition of cutting interference.
Example 2 Input data, z1 ¼ 10 and z2 ¼ 24. Therefore, we have: ðz1 þ z2 Þ ¼ 34;
x1 ¼ 7=17; x2 ¼ x1 ¼ 7=17. In this case, both the pinion and wheel are in the
limit condition of cutting interference.
Example 3 Input data, z1 ¼ 10 and z2 ¼ 22. Therefore, we have:ðz1 þ z2 Þ ¼
32\34; x1 ¼ 7=17; x2 ¼ 5=17; ðx1 þ x2 Þ ¼ 2=17 [ 0.
It is to be noted that this criterion was the basis of the old method of the German
standard DIN, which prescribed choosing the profile shift coefficients with values
such as to avoid the cutting interference of the pinion, in the case where
ðz1 þ z2 Þ [ 2z0 , or of the pinion and wheel, in the case where ðz1 þ z2 Þ 2z0 (see
Giovannozzi [12], Niemann and Winter [22]). It is evident that, with the values of
the profile shift coefficients thus calculated, we would come to have the maximum
possible length of the path of contact (it would coincide with the maximum
6.8 Determination of the Profile Shift … 267
theoretical value, equal to the line of action, i.e. the length of the segment
T 1 T 2 ¼ a sin a, in the case where it was ðz1 þ z2 Þ ¼ 2z0 ).
It is equally evident the corresponding disadvantage of having a contact between
the mating profiles that would extend until the origin of involute curve, where the
radius of curvature is zero, and the contact takes place in very unfavourable con-
ditions, as we already said. For this reason, for the two cases of a0 ¼ 15 and
a0 ¼ 20 ; the same standard DIN subsequently prescribed to calculate the profile
shift coefficients using the practical relationships, rather than the theoretical ones.
All these relationships are collected in Table 6.3.
The use of the aforementioned old method DIN, in both formulations, theoretical
and practical, is no longer recommended because the gear designer has to control
not only the interference and undercut problems, which are undoubtedly very
important, but also many other problems certainly no less important, among which
those associated with the specific sliding and surface pressure occupy an extremely
significant role. In fact, the gear strength to the surface fatigue, wear and scuffing
depends on them.
Henriot [13] noted that, with the values of the profile shift coefficients calculated
with the old method DIN, distribution curves of the surface pressure, rH , specific
slidings, f1 and f2 , sliding velocities, vg1 and vg2 , and Almen factors, rH vg1 and
rH vg2 , very unbalanced along the path of contact were obtained. Starting from this
observation, he proposed a criterion and related method much more articulated, but
easy to apply, which not only allows avoiding the interference, but also allows
equalizing the maximum values of the specific sliding and Almen factors of the
pinion and wheel [1].
The balancing condition, i.e. the condition to balance the maximum values of the
specific sliding on the pinion and wheel leads to write the equality:
where f1A and f2E are respectively the maximum negative values of the specific
sliding of the pinion and wheel, respectively at marked points A and E (see
Fig. 2.10). The marked points A and E are the end points of the path of contact,
where the contact begins and ends respectively when the pinion is driving. In these
points, the radii of curvature of the mating profiles are: q1A ¼ T 1 A; q2A ¼ T 2 A;
q1E ¼ T 1 E; q2E ¼ T 2 E. Taking into account Eqs. (3.73), (3.59), and (3.60), the
equality (6.121) becomes:
268 6 Profile Shift of Spur Gear Involute Toothing …
As we shall see in due course, when we talk about the scuffing load carrying
capacity of the gears, we define, as Almen factor, the product of the contact stress,
rH (also called surface pressure or Hertzian pressure) and sliding velocity, vg (see
Vol. 2, Chap. 7). The balancing condition of the Almen factors of the pinion and
wheel, calculated respectively at marked points A and E, leads to write the equality:
where rHA and rHE are the contact stresses between the mating profiles at marked
points A and E, while vg1A and vg2E are the sliding velocities of the pinion and
wheel, given by Eqs. (3.65), and evaluated at the same marked points A and E. Note
that the third subscript indicates these two marked points.
We will demonstrate at the time that the Hertzian pressure rH at a given point of
contact, between the end points A and E of the path of contact, is proportional to the
square root of the sum of the curvatures of the profiles at that point. At marked
points A and E, the sum of these curvatures is given by:
Specializing the Eqs. (6.65) at these marked points, and introducing them in the
equality (6.124), together with the square root of Eqs. (6.125) and (6.126), we get:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
a sin a a sin a
ðx1 q1A x2 q2A Þ ¼ ðx2 q2E x1 q1E Þ : ð6:127Þ
q1A q2A q1E q2E
Evidently, this relationship coincides with Eq. (6.123). Therefore, we see that
the balancing of the specific slidings at marked points A and E also ensure the
balancing of the Almen factors at the same points.
6.8 Determination of the Profile Shift … 269
Starting from the relationships above [13] proposed a calculation method of the
profile shift coefficients, x1 and x2 , which ensures the equality of the maximum values
of the specific slidings and Almen factors of the two members of the gear pair, which
occur at marked points A and E. To make calculations easy and fast, he worked out
two diagrams, one valid for a ¼ 20 ; and the other valid for a ¼ 15 : Figure 6.18
shows the diagram related to a ¼ 20 (for the other diagram, we refer to [13]).
The diagram is divided substantially into two regions: the first region, to the left,
almost triangular, surrounded by the ordinate axis, and the two curves AB and A0 B,
and the second region, to the right, and external to the first. The inner part of the
region almost triangular in shape refers to ðz1 þ z2 Þ\60: The second region, the
outer to the first, refers to ðz1 þ z2 Þ [ 60: The two curves AB and A0 B, the border
Fig. 6.18 Calculation diagram of x1 and x2 , which ensure the equality of the maximum values of
specific slidings and Almen factors for pinion and wheel, for a ¼ 20 (from Henriot)
270 6 Profile Shift of Spur Gear Involute Toothing …
lines between the two regions, refer to ðz1 þ z2 Þ ¼ 60: Each of the curves appearing
in the two regions of the diagram corresponds to a given value of the gear ratio
u ¼ z2 =z1 . The diagram is to be used in the following way:
(a) In the case for which ðz1 þ z2 Þ 60, we have profile-shifted toothing without
variation of center distance, and the profile shift coefficients, x1 and x2 , with
x2 ¼ x1 , are to be calculated as a function of the number of teeth of the
pinion, z1 , and gear ratio u ¼ z2 =z1 . We enter the diagram with the number of
teeth z1 of the pinion, shown on the abscissa. Therefore, with the vertical line
having abscissa z1 , we intersect the curves u ¼ const, which are symmetric with
respect to the abscissa axis. The ordinates of the two points of intersection
above are respectively x1 and x2 ¼ x1 .
(b) In the case for which ðz1 þ z2 Þ\60, we have profile-shifted toothing with
variation of center distance. The profile shift coefficients, x1 and x2 , with
ðx1 þ x2 Þ 6¼ 0, are to be still determined as a function of the number of teeth of
the pinion, z1 , and the gear ratio u ¼ z2 =z1 , but in a different way from that
described above for the case ðz1 þ z2 Þ 60: We still enter the diagram with the
number of teeth of the pinion, z1 . Therefore, with the vertical straight line having
abscissa z1 , we intersect the two curves AB and u ¼ const, which in this case is
within the region almost triangular in shape. The ordinates of the two points of
intersection with the curve AB and the curve u ¼ const are respectively x1 and x2 .
One of the most important quantities in any gear design is the number of teeth of the
pinion, and profile-shifted toothing has its greatest relevance when the pinion has a
small number of teeth. In this regard, we have shown that the first use of profile
shift was to avoid cutting interference and undercut in pinions with numbers of
6.8 Determination of the Profile Shift … 271
teeth less than those given by the inequality (6.20). It was later realized that
profile-shifted toothing improves most aspects of the gear operating conditions, and
today it is used for many reasons including mainly:
• avoiding cutting interference and undercut;
• avoiding narrow top lands and pointed teeth, and therefore case/core separation
in case-hardened gears;
• balancing specific slidings to maximize wear strength and Hertzian fatigue
strength;
• balancing flash temperature to maximize scuffing strength;
• balancing bending fatigue life to maximize bending fatigue strength;
• reducing the frictional losses to lower the contact temperature, and increase the
strength to wear, macro-pitting, micro-pitting and scuffing;
• increasing or decreasing the center distance and changing tooth thickness and
backlash; etc.
It is evident that not all these tasks/criteria be simultaneously fulfil, also because
they are often not compatible with each other, resulting in conflicting demands.
Depending on the goals to be achieved, the gear designer must make choices,
favouring the balancing of the quantities that most influence the problem that he is
facing, at the expense of other quantities that have more limited influence, but not
nothing. In doing so, however, the problem of distribution of the sum of profile shift
coefficients between the pinion and wheel would see greatly broaden his horizons,
and gear design, in itself very demanding, would become even more complex.
To simplify the task of the designer, the various international and national
standards have dealt with the problem, and have developed easy to apply methods,
which lead to satisfactory results, although based on a reasonable compromise that
ensures the balancing of a given quantity, but tries not to penalize too much other
quantities, albeit a bit less influential. Here we present briefly three standardized
methods for the distribution of the sum ðx1 þ x2 Þ of the profile shift coefficients
between the pinion and wheel. This sum, already determined in Sect. 6.7, follows
from the requirement to ensure an imposed operating center distance, and is framed
within the limits recommended by standards.
Before describing these standardized methods, it is however necessary to clarify
a fundamental concept, that of the operating pressure angle. In fact, a common
perception is that profile shift changes the pressure angle of a gear, and this could be
a concept wrong, if it was not seen in the right light. The pressure angle we specify
on the gear drawing is the pressure angle of the cutting tool. The actual transverse
pressure angle at a point changes along the profile, from the tooth root at tooth tip,
and the operating pressure angle, i.e. the angle of the line of action, of a gear pair
depends on the gear center distance and radii of the base circles only. If the sum
ðx1 þ x2 Þ of the profile shift coefficients is zero, the gear pair operates at the ref-
erence center distance. If this sum is negative, the gear pair has smaller center
distance, and thus it has a smaller operating pressure angle, while a positive sum
requires a larger center distance, and thus an increased operating pressure angle.
272 6 Profile Shift of Spur Gear Involute Toothing …
Rebus sic stantibus, we can immediately calculate the operating pressure angle,
a0 , using the Eq. (6.59), since the other three quantities, a0 , a0 , and a0 , that appear in
this equation are all known. Once we have calculated so the value of a0 , by means
of Eq. (6.58) we can calculate the sum ðx1 þ x2 Þ to be divided between the pinion
and wheel.
This method, standardized by DIN 3992 [6], adopts a flexible criterion that leads to
toothings having load carrying capacity and sliding speed well balanced. It can also
be used for predetermined center distance. The method is applicable for gears with
number of teeth z 150 (for z [ 150, a somewhat different method is adopted, but
in these cases the profile shift does not have a significant influence on the load
carrying capacity of the gears). The method gives two pairs of diagrams permitting
the determination of the profile shift coefficient, x1 , of the pinion by the procedure
explained below. The first pair of diagrams concerns the speed reducing gears
(Fig. 6.19), while the second pair regards the speed increasing gears (Fig. 6.20).
The first diagram of each pair is of general validity, and common to both types of
drives, that is it refers both to speed reducing gears (Fig. 6.19), and speed
increasing gears (Fig. 6.20).
The first diagram of each pair does not refer, however, to the division of the sum,
Rx ¼ ðx1 þ x2 Þ, of the profile shift coefficients; instead it concerns the determina-
tion of this sum in case it had not been determined in advance with the method
described in Sect. 6.6. For this preliminary step, we enter the first diagram of each
pair with the actual sum Rz ¼ ðz1 þ z2 Þ or equivalent (or virtual) sum Rzv ¼
ðzv1 þ zv2 Þ of the numbers of teeth (see Chap. 8), depending on whether the gears
under consideration are parallel cylindrical spur or helical gears. We then identify a
point on one of the nine characteristic lines from P1 to P9 (we could also identify a
point between these lines), and we go out with a horizontal straight line up to the
ordinate axis, where we read the value of the sum, Rx ¼ ðx1 þ x2 Þ.
The characteristic lines P1 to P9 in Figs. (6.19) and (6.20) identify the main gear
peculiarities that the designer wants to achieve, namely: region between P1 and P3,
for high values of contact ratio ea ; region between P3 and P6, for usual applica-
tions, with toothing well balanced; region between P6 and P9, for high load car-
rying capacity to bending and surface fatigue. This first diagram of each pair also
highlights two special regions, to be avoided in the usual gear design: the lower
region, which refers to the cases of small operating pressure angle, a0 , and large
contact ratio, ea , and the upper region, which refers to the special cases of large
operating pressure angle, a0 , and small contact ratio, ea . On the right, the diagram
also shows that the radii of curvature of the profiles increase from the bottom
towards the top, while the contact ratios increase from the top to the bottom.
To split out the sum Rx of the profile shift coefficients between the pinion and
wheel of a speed reducing gear, we use the second diagram of Fig. 6.19. To this
end, we enter the diagram with half the sum Rz=2 or Rzv =2, on the abscissa, and
6.8 Determination of the Profile Shift … 273
(a)
(b)
Fig. 6.19 Method DIN 3992 for external speed reducing gears: a diagram to choose the sum Rx;
b diagram for partition of the sum Rx between pinion and wheel
with half the sum, Rx=2, on the ordinate axis. So, we identify a point on this
diagram, having coordinates (Rz=2, Rx=2), which will be between the seventeen
characteristic lines (lines L1 to L17, from bottom to top) that appear in the same
diagram. Therefore, a graphical interpolation line is traced between two neigh-
boring lines, and passing through the point thus identified. The ordinate of the point
having abscissa, z1 , on this interpolation line gives the profile shift coefficient, x1 , of
the pinion. Being determined the profile shift coefficient of the pinion, that of the
wheel will be given by x2 ¼ Rx x1 .
274 6 Profile Shift of Spur Gear Involute Toothing …
(a)
(b)
Fig. 6.20 Method DIN 3992 for external speed increasing gears: a diagram to choose the sum Rx;
b diagram for partition of the sum Rx between pinion and wheel
It should be noted that the characteristic lines shown in the second diagram of
Fig. 6.19 have been chosen so that the following design goals are simultaneously
achieved: the bending tooth root strengths of the pinion and wheel are balanced; the
sliding velocity at the tooth tip of the driving gear (usually, the pinion) is, as far as
possible, a little greater than that at the tooth tip of the driven gear (this design
choice is imposed by the need to reduce noise emissions); the extreme values of
specific sliding are avoided. The diagram shows two regions shaded in grey, one at
the top and the other at the bottom, which must be avoided. In addition, on the right,
it also shows that the fillet radius increases from the top to the bottom, while the
6.8 Determination of the Profile Shift … 275
tooth thickness measured on the base circle increases from the bottom to the top. It
is finally to be noted that the diagram is delimited, at the bottom left, from the limit
curve of undercut and, in the upper left side, by three limit curves, corresponding
respectively to: sa [ 0:2mn ; ean [ 1:1; sa [ 0:4mn .
To split out the sum Rx of the profile shift coefficients between the pinion and
wheel of a speed increasing gear, we use the second diagram of Fig. 6.20. The use
of this diagram, characterized by thirteen characteristic lines (S1 to S13, from
bottom to top), a region shaded in gray at the top, and a region for special cases at
the bottom, is entirely similar to the second diagram of Fig. 6.19. Finally, it is to be
noted that, to facilitate the task of the designer, the abscissa of the second diagram
of both Figs. 6.19 and 6.20, equal to Rz=2 or Rzv =2, are coordinated with those of
the first diagram of the same figures, equal to Rz or Rzv .
Example 9 Speed reducing spur gear pair. Input data: z1 ¼ 20; z2 ¼ 100: Design
goal: gears well balanced. Choice of characteristic line: between P4 and P5. From
diagram shown in Fig. 6.19a, we obtain: Rx ¼ x1 þ x2 ¼ 0:24, corresponding to
Rz ¼ z1 þ z2 ¼ 120: On the diagram shown in Fig. 6.19b, we identify the point A,
having coordinates ðRz=2 ¼ 60; Rx=2 ¼ 0:12Þ. On the same diagram, we draw the
interpolation line between L11 and L12, and passing through point A. On this
interpolation line, the ordinate of point having as abscissa z1 ¼ 20 is x1 ¼ 0:30:
Finally, x2 can be obtained as the difference
x2 ¼ Rx x1 ¼ ð0:24 0:30Þ ¼ 0:06, or as the ordinate of the point having as
abscissa z2 ¼ 100 on the same interpolation line.
Example 10 Speed increasing spur gear pair. Input data: z1 ¼ 16; z2 ¼ 40: Design
goal: high load carrying capacity to bending and surface fatigue. Choice of char-
acteristic line: P7. From diagram shown in Fig. 6.20a, we obtain:
Rx ¼ x1 þ x2 ¼ 0:84, corresponding to Rz ¼ z1 þ z2 ¼ 56. On the diagram shown
in Fig. 6.20b, we identify the point B, having coordinates
ðRz=2 ¼ 28; Rx=2 ¼ 0:42Þ. On the same diagram, we draw the interpolation line
between S9 and S10, and passing through the point B. On this interpolation line, the
ordinate of point having as abscissa z1 ¼ 16 is x1 ¼ 0:36. Finally, x2 can be
obtained as the difference x2 ¼ Rx x1 ¼ ð0:84 0:36Þ ¼ 0:48, or as the ordinate
of the point having as abscissa z2 ¼ 40 on the same interpolation line.
This method, standardized by BSI—BS PD 6457 [2], uses the following general
formula for the determination of the profile shift coefficient, x1 , of the pinion, in the
case of an external speed reducing gear pair:
276 6 Profile Shift of Spur Gear Involute Toothing …
1 Rx
x1 ¼ C 1 þ ; ð6:129Þ
u 1þu
where u ¼ z2 =z1 is the gear ratio, Rx ¼ ðx1 þ x2 Þ is the sum of the profile shift
coefficients, and C is a factor dependent on the criteria that are used.
For general applications, where slightly more than half of the tooth action occurs
during the recess stage of the meshing cycle, the factor C is equal to 1/3, for which
the Eq. (6.129) becomes:
1 1 Rx
x1 ¼ 1 þ : ð6:130Þ
3 u 1þu
where
z1
zv1 ¼ ; ð6:133Þ
cos3 b
is the equivalent or virtual number of teeth of the pinion, and b is the helix angle.
Note that, for obtaining the equivalent numbers of teeth zv1 and zv2 , the helix angle
of the reference helix of each member of the helical gear is considered (see Chap. 8,
regarding helical gears).
This method, standardized by ISO/TR 4467 [16], employs the following formula
for the determination of the profile shift coefficient of the pinion, x1 :
6.8 Determination of the Profile Shift … 277
z2 z1 z1 u1 Rx
x1 ¼ k þ Rx ¼k þ ; ð6:134Þ
z2 þ z1 z2 þ z1 uþ1 uþ1
where k is a factor to be chosen in the range ð0:50 k 0:75Þ for speed reducing
gears. If u ¼ z2 =z1 [ 5, the limit calculation value is u ¼ 5.
For the choice of the profile shift coefficient, x1 , we must take account of the
operating conditions of the external gear pair, because things change according to
whether it is a speed reducing gear or a speed increasing gear. In fact, in the first
case, the smallest gear, i.e. the pinion, which is most stressed, is driving, and thus it
requires a positive shift coefficient, such as to strengthen the tooth to its root, to
balance the specific sliding between pinion and wheel, and to limit the contact
pressure. Thus, for speed reducing gears, it is advisable to choose a factor k in the
range ð0:50 k 0:75Þ, preferably approaching to the value k ¼ 0:75 when the
sum of the number of teeth decreases.
Example 11 (see Example 3). Input data: z1 ¼ 20; z2 ¼ 40; a0 ¼ 20 ; Rx ¼ 0:32
(this value is suitable for general mechanical applications).
Results with k ¼ 0:75: x1 ¼ 0:36; x2 ¼ Rx x1 ¼ 0:04.
Results with k ¼ 0:50: x1 ¼ 0:28; x2 ¼ Rx x1 ¼ 0:04.
Therefore, we realize that, with the choice of the k-values coincident with the
extreme of its variability range, we obtain values of x1 and x2 close to those
obtained with the method used for the Example 3 (x1 ¼ 0:32, and x2 ¼ 0).
Example 12 Input data: z1 ¼ 40; z2 ¼ 120; a0 ¼ 20 ; Rx ¼ 0:90:
Results with k ¼ 0:75: x1 ¼ 0:60; x2 ¼ Rx x1 ¼ 0:30:
Results with k ¼ 0:50: x1 ¼ 0:48; x2 ¼ Rx x1 ¼ 0:42.
In the second case, that of speed increasing gears, the problem of determination
of the profile shift coefficients, x1 and x2 , is in general less demanding and requires
no special corrective actions. However, other problems can sometimes arise, and in
these cases, the design must be based on assumptions completely different. In effect,
it is necessary to take into account the fact that, due to friction (so far, we have
neglected the friction forces), the total force applied by the driving gear (the wheel)
to the driven gear (the pinion) is not directed along the line of action. In fact, it is
inclined, with respect to this, by the friction angle, u (see Fig. 3.9), which is related
to the coefficient of friction, l; according to the relationship l ¼ tan u:
Now it is to remember that the total resultant force, F, is the vector sum of the
useful force, acting along the line of action, and friction force. The friction force is
perpendicular to the useful force and has direction such as to oppose the motion.
Considering even the friction force, it happens that, generally, the total resultant
force, F, tends to approach the axis of the driven member (the pinion), when contact
occurs in the approach path of contact AC (Fig. 6.21). In this case, it is therefore
useful to have a recess path of contact wider than the approach path of contact, i.e.
CE [ AC. These favourable conditions are automatically realized with a speed
reducing gear, in which the addendum of the pinion is increased, while the
278 6 Profile Shift of Spur Gear Involute Toothing …
addendum of the wheel is decreased. For a speed increasing gear, however, this
kind of choice of the profile shift coefficients becomes disadvantageous. In fact, if
the gear ratio is large, and the number of teeth of the pinion is small, it may happen
that the line of application of the total resultant force, F, approaches dangerously to
the pinion axis. In this case, we could incur the risk of jamming and pour efficiency.
In these cases, the relationship (6.134) can still be used, with the caveat to
choose k ¼ 0: For example, with z1 ¼ 30; z2 ¼ 210; a0 ¼ 20 ; Rx ¼ 0, we obtain
x1 ¼ 0, and x2 ¼ Rx x1 ¼ 0, i.e. a standard toothing. In the case of speed
increasing gear, a good solution is to use a standard toothing, associated with the
choice of a number of teeth of the pinion equal to or greater than 36.
It should also be remembered that recess action gears could be required not only
to solve the above-mentioned issues related to speed increasing gears, but also to
use approximately their own interesting property. Indeed, in practice, it has been
found that the performance of a gear pair is smoother during the recess path of
contact rather than the approach path of contact of the meshing cycle. In Sect. 3.6
we showed that the sliding velocity changes direction and sign when the point of
contact passes through the pitch point. Even the friction force changes its direction,
while the total force, F, moves from side to side with respect to the line of action, as
Fig. 3.9 shows. This different orientation of the forces coming into play in the
recess path of contact is responsible for the smoother operation of the gear pair
during the recess stage of the meshing cycle.
For a gear pair whose members do not have to change their role (in the meaning
that one gear wheel is always the driving member, and the other is always the
driven member), it is possible to carry out its design in order to draw advantage of
the smoother recess action. The gear pair is designed in such a way that most or the
entire path of contact is the recess path of contact. To do this, the driving member
addendum must be adequately increased, and the driven member addendum must
be proportionality decreased. When the recess path of contact is much broader than
the approach path of contact, we say the gear pair has a partial recess action.
Instead, when the recess path of contact constitutes the entire path of contact (that
is, the length of the approach path of contact is reduced to zero), we say that the
6.8 Determination of the Profile Shift … 279
gear pair is an only recess action gear pair. These design goals can be achieved by
choosing suitable profile shift coefficients, x1 and x2 , using the methods described
above.
In Sect. 3.2 we said that, for an internal gear pair, unlike what happens for an
external gear pair, it is not always possible to freely share the profile shift coeffi-
cients, and simultaneously obtain optimal values of the specific sliding and contact
ratio. In fact, the tendency of teeth of an internal gear pair to interfere with each
other imposes significant limitations, so it is not easy to define a practical method to
predetermine the values of profile shift coefficients. In most cases, it is necessary to
proceed by trial and error, also because for internal gear pairs, contrary to external
gear pairs, the minimum number of teeth of the pinion to avoid the interference is
the larger, the lower is the gear ratio, u ¼ z2 =z1 .
For the determination of the profile shift coefficients, the basic relationships from
which we must start are those shown in Sect. 6.2. Bearing in mind that, very often,
the operating center distance, a0 , is an input of the problem, to be observed strictly,
from Eq. (6.113) we obtain the operating pressure angle, a0 , as follows:
a
a0 ¼ arccos
0
cos a0 : ð6:135Þ
a0
The range of the favourable values of the profile shift coefficients for this type
of profile-shifted toothing is in any case between ð0:50 x 0:65Þ. With this
kind of profile shift, choosing the profile shift coefficients to balance the
specific sliding at the end points of the path of contact, we can obtain teeth with
high load carrying capacity and favourable driving characteristics. However, in
cases of predetermined center distance, this type of profile shift shows all its
limitations, because it is not possible to satisfy strictly the design input data
with this profile shift.
(b) If the number of teeth, z1 , of the pinion is less than the minimum number of
teeth, z1min (see Fig. 3.1), it is convenient to use a gear pair with profile-shifted
toothing and variation of center distance, with profile shift coefficients to
choose from time to time in relation to the goals to be achieved. Compared to
the profile-shifted toothing without variation of center distance, those with
variation of center distance do not have a significantly higher load carrying
capacity, but allow the designer to have a greater freedom of choice for the
achievement of predetermined design requirements. Tip thicknesses, in terms of
crest-widths, space widths, and interference in the actual meshing conditions
are always checked and verified. In order to ensure a predetermined center
distance (this is impossible for profile-shifted toothing without variation of
center distance), it is preferable to choose the profile shift coefficients in such a
way that it is ðx1 þ x2 Þ\0, with x2 conventionally considered as negative.
Since the risk of interference exists, the choice above is generally necessary,
even for ðz1 þ z2 Þ [ 7 (here also the number of teeth, z2 , of the annulus is to
be considered, by convention, as negative).
From the considerations above, we infer that the profile-shifted toothing without
variation of center distance do not have appreciable disadvantages, while the
profile-shifted toothing with variation of center distance offers more, than the first,
only the ability to freely choose the center distance. For this reason, not separated
from that of considerably greater simplicity of the calculations to be performed, the
designer often opts for the profile-shifted toothing without variation of center
distance.
In the internal spur gears with profile-shifted toothing without variation of center
distance ðx2 ¼ x1 Þ, there is a correlation between ðr a2 r b2 Þ=m (i.e. the quotient of
the radial distance between the tip circle and base circle, divided by the module)
and the profile shift coefficient, x2 , which depends on the number of teeth, z2 .
Figure 6.22 shows this correlation. The choice of a positive shift coefficient, x2 , for
the annulus determines a tooth profile modification, which strengthens the tooth at
its root, and facilitates its assembly with the pinion. However, it is to be noted that,
to ensure a good condition of engagement between the pinion and annulus, a limit
exists for the maximum value of the sum of the profile shift coefficients,
ðx1 þ x2 Þmax . As Fig. 6.23 shows, this limit sum is related to the pressure angle, a,
and difference Dz ¼ jz2 j z1 ; this figure refers to an annulus to be generated with a
pinion-type cutter having nominal pressure angle, a0 ¼ 20 , and nominal adden-
dum h0 ¼ m0 .
6.10 Backlash 281
Fig. 6.22 Correlation between ðr a2 r b2 Þ=m and x2 , as a function of the number of teeth z2
6.10 Backlash
In certain applications, such as gear trains where positioning is the design key
condition, but the power to transmit is low, or lead-screws where positioning and
power are both important, backlash (sometimes also called lash or play) is an
undesirable characteristic, and should be at least minimized or even reduced to zero.
In fact, the ideal theoretical condition would be to have gears with zero backlash or
anti-backlash gears (see also 7).
However, in actual design practice of gears, some backlash must be necessarily
allowed to prevent jamming. Reasons for the presence of backlash in a gear power
transmission include allowing for manufacturing errors, thermal expansion,
deflections under load, and lubrication. Here we leave aside the problems of
anti-backlash design of gears, and focus our attention on those where the backlash
must necessarily be allowed.
Profile shift introduced by us in the previous sections, and usually used by gear
designers, is related a backlash-free contact, or zero-backlash gear set. To provide
the necessary backlash, some designers reduce the profile shift, while others
increase center distance. Factors affecting the amount of backlash required in a gear
pair, or in a gear train, include errors in profile shape, pitch, tooth thickness, center
distance, helix angle (see Chap. 8), and run-out. The greater the accuracy of
manufacture and assembly, the smaller the errors mentioned above, and therefore
the backlash needed.
282 6 Profile Shift of Spur Gear Involute Toothing …
The most commonly used method to obtain the optimum value of the backlash
consists in cutting the teeth deeper inside the gear blanks than the theoretical depth,
corresponding to the condition of backlash-free contact. This method is the most
general, and is the only one that allows satisfying a predetermined center distance
required as input for the design. It is updated by changing the profile shift, that is
moving the cutting tool radially to change the tooth thickness. Evidently, from this
point of view, we can say that all gears have some profile shift, since the backlash
must be guaranteed even in the case of zero-profile-shifted toothing.
Now, according to the sign convention for the profile shift, a positive profile shift
corresponds to thicker teeth. Thus, a shift of the rack-type cutter (Fig. 6.1) towards
the axis of an external gear wheel corresponds to a negative profile shift, since it
makes the teeth thinner. Therefore, subtracting from the zero-backlash profile shift
an additional small profile shift, corresponding to the backlash allowance, we obtain
the so-called rack shift. When gear designer specifies only the zero-backlash profile
shift (gear drawings are usually performed with respect to the zero-backlash profile
shift), the gear manufacturers will thin the teeth to ensure the necessary operating
backlash according to standards. Many of these standards require that the tooth
thinning for backlash, called upper allowance of size, is a function of the module.
6.10 Backlash 283
The values are measured on the reference cylinder, in a transverse section for spur
gears, and in a normal section for helical gears (see Chap. 8). The transverse
circular allowance, related to the normal allowance in the base tangent plane, is a
function of allowance, center distance, and tooth accuracy.
A common convention among gear manufacturers is to reduce the normal tooth
thickness of each members of the gear pair by the same amount, which may be a
predetermined value, in lm, or a function of module, such as ð0:03 0:05Þmn . This
maintains the same cutting depth for both gear members, and maximizes contact
ratio. The direction, normal or transverse, as well as the reference circle or base
tangent plane, according to which the tooth thickness reduction is to be measured,
must be specified, since there is no recognized convention. Standard practice is to
make allowance for half the backlash in the tooth thickness of each gear member.
However, if the pinion is significantly smaller than the mating wheel, it is common
practice to account for the entire backlash in the larger wheel, because this ensures
the maximum strength as possible of the pinion teeth. The amount of additional
material removed when manufacturing the gears depends on the pressure angle, a0 .
For a0 ¼ 14:5 , the cutting tool is moved towards the axis of the wheel to be cut by
an amount equal to the backlash desired; this extra distance corresponds to the
additional shift profile. For a0 ¼ 20 , this extra distance equals 0.73 times the
amount of backlash desired.
The other method consists of the increase or decrease of the center distance,
depending on whether the gear is external or internal. This method is less general
than the first, as it does not guarantee exactly a predetermined center distance.
Furthermore, the first method comprises, as a special case, the second method, since
the backlash obtained by a small change of the profile shift corresponds, in fact, to a
virtual variation (increasing or decreasing respectively for external and internal
gears) of the center distance.
Let us consider the example of an external gear pair, and determine the corre-
lation between the increase of center distance (virtual increase for the first method,
and actual increase for the second method) and backlash measured on the normal to
the profile. To this end, it is sufficient to consider that a removal Dr10 of the axis of
the pitch circle of the pinion along the center distance line O1 O2 (Fig. 2.10) causes
a decrease in the tooth thickness equal to 2Dr10 sin a0 , when it is measured along the
line of action. This is because, on the pitch circle, the normal to the profile forms an
angle, a0 , with the common tangent to the pitch circles. Similarly, a removal Dr20 of
the axis of the pitch circle of the wheel along the center distance line O1 O2
determines a decrease of the tooth thickness equal to 2Dr20 sin a0 .
Therefore, we will have a normal backlash, jn (the backlash measured along the
normal to the profile or, what is the same, along the line of action), equal to:
jn ¼ 2 Dr10 þ Dr20 sin a0 ¼ 2Da0 sin a0 ; ð6:136Þ
284 6 Profile Shift of Spur Gear Involute Toothing …
where Da0 ¼ Dr10 þ Dr20 is the virtual or actual increase of the center distance. The
corresponding transverse backlash, jt , i.e. the backlash measured on the pitch circle,
is given by:
From Eq. (6.136), once the value of the normal backlash, jn , has been fixed case
by case, and assuming a0 ¼ a0 , we obtain Da0 . In reality, for a given value of jn ,
varying the center distance, also a0 varies, although not by much. For this reason,
we should set an iterative calculation until convergence of the results is obtained.
The iterative calculations to do when the center distance, a0 , is predetermined are
even more elaborate and complex. As in the previous case, once the value of jn has
been fixed, we calculate Da0 by means of Eq. (6.136), in which we set a0 ¼ a0
(actually, we should introduce a value a0 [ a0 when ðx1 þ x2 Þ [ 0, and a0 \a0
when ðx1 þ x2 Þ\0). According to Eq. (6.59), the center distance in the condition of
backlash-free contact is given by:
z1 þ z2 cos a0
a0 Da0 ¼ m0 : ð6:138Þ
2 cos a0
References
1. Almen JO, Straub JC (1948) Aircraft gearing, analysis of test and data service. Research
Laboratories Division, General Motors Corporation, Detroit, Michigan, AGMA
2. BSI-BS PD G457: Guide to the Application of Addendum Modification to Involute Spur and
Helical Gears
3. Buckingham E (1949) Analytical mechanics of gears. McGraw-Hill Book Company Inc, New
York
4. Chauhan V (2016) A review on effect of some important parameters on the bending strength
and surface durability of gears. Int Journ Sci Res Publ 6(3):289–297
5. Colbourne JR (1987) The geometry of involute gears. Springer-Verlag, New York Inc, New
York Berlin Heidelberg
6. DIN 3992: Profilverschiebung bei Stinrädern nit Aubenverzahnung (Addendum modification
of external spur and helical gears)
7. Dooner DB, Seireg AA (1995) The kinematic geometry of gears—a concurrent engineering
approach. Wiley, New York
References 285
8. Dudley DW (1962) Gear handbook. The design, manufacture, and application of gears.
McGraw-Hill Book Company, New York
9. Dudley DW (1969) The evolution of the gear art. American Gear Manufacturers Association,
Washington, D.C
10. Dudley DW, Sprengers J, Schröder D, Yamashina H (1995) Gear motor handbook. In:
Bonfiglioli Riduttori SPA (Eds) Springer-Verlag, Berlin Heidelberg
11. Ferrari C, Romiti A (1966) Meccanica applicata alle macchine. Unione Tipografico-Editrice
Torinese, UTET, Torino
12. Giovannozzi R (1965) Costruzione di macchine, vol II, 4th ed. Casa Editrice Prof. Riccardo
Pàtron, Bologna
13. Henriot G (1979) Traité théorique et pratique des engrenages, vol 1, 6th edn. Bordas, Paris
14. ISO 1122-1 (1998) Vocabulary of gear terms
15. ISO 6336-1 Part 1: Basic principles, introduction and general influence factors
16. ISO/TR 4467 Addendum modification of the teeth of cylindrical gears for speed-addendum
modification reducing and speed-increasing gear pairs
17. Lasche O (1899) Die Ausführung mit einer Profilverschiebung nach obenstchenden Augaben
wird der grundlegenden, Z.d. VDI, S. 1488 auch als AEG-Verzahnung bezeichnet
18. Li S (2008) Effect of addendum on contact strength, bending strength and basic performance
parameters of a pair of spur gears. Mech Mach Theory 43(12):1557–1584
19. Litvin FL, Fuentes A (2004) Gear geometry and applied theory, 2nd edn. Cambridge
University Press, Cambridge
20. Maitra GM (1994) Handbook of gear design, 2nd edn. Tata McGraw-Hill Publishing
Company Ltd, New Delhi
21. Merritt HE (1954) Gears, 3rd edn. Sir Isaac Pitman & Sons Ltd, London
22. Niemann G, Winter H (1983) Maschinen-elemente, band II: getriebe allgemein,
zahnradgetriebe-grundlagen, stirnradgetriebe. Springer-Verlag, Berlin Heidelberg
23. Oda S, Tsubukura K, Namba C (1982) Effect of addendum modification on bending fatigue
strength of spur gear with high pressure angle. Bull. Jap. Soc. Mech. Eng. 259(209):1813–1826
24. Polder JW (1991) Interference of internal gears. In: Towsend P (ed) Dudley’s gear handbook.
McGraw-Hill, New York
25. Pollone G (1976) Il Veicolo. Libreria Editrice Universitaria Levrotto & Bella, Torino
26. Schiebel A, Lindner W (1954) Zahnräder, Band. I: Stirn-und Kegelräder mit geraden Zähnen.
Springer-Verlag, Berlin Heidelberg
27. Schiebel A, Lindner W (1957) Zahnräder, Band 2: Stirn-und Kegelräder mit schrägen Zähnen
Schraubgetriebe. Springer-Verlag, Berlin Heidelberg
28. Schreier G (1961) Stirnrad-Verzahnung. VEB Verlag Technik, Berlin
29. Scotto Lavina G (1996) Riassunto delle Lezioni di Meccanica Applicata alle Macchine:
Cinematica Applicata; Dinamica Applicata—Resistenze Passive—Coppie Inferiori; Coppie
Superiori (Ingranaggi)—Flessibili—Freni. Edizioni Scientifiche Siderea, Roma
30. Simon V (1989) Optimal tooth modifications for spur and helical gears. J Mach Des 3(4):
611–615
31. Woodbury RS (1958) History of gear-cutting machine: a historical study in geometry and
machines. MIT Press, Cambridge, Massachusetts, USA
Chapter 7
Profile Shift of External Spur Gear
Involute Toothing Generated
by Pinion-Type Cutter
Abstract In this chapter, the fundamentals of the profile shift of spur gears
involute toothing generated by pinion-type cutter are first described, highlighting
the interference and undercut problems related to this cutting process of the teeth.
The limitations regarding the possibility of making external spur gears with
profile-shifted toothing by means of this generation process are described, and the
greatest complication of the calculations to be carried out is highlighted, with a
comparison with the manufacturing process of profile-shifted toothing by rack-type
cutter.
The cutting tool is shaped exactly like a pinion, and is therefore called pinion-type
cutter or Fellows gear shaper (see Sect. 3.11.2). However, unlike a pinion gear, the
teeth of the pinion-type cutter are characterized by appropriate front rake and face
angles (see Fig. 3.15), and are sharpened at their race angle (see: Micheletti [7],
Galassini [2], Rossi [10]). As in a usual pinion, the teeth of the pinion-type cutter
are shaped as involute curves of a base circle having diameter
where z0 is the number of teeth of this Fellows gear shaper, while m0 and a0 are
respectively the nominal module and nominal pressure angle of its toothing; p0 ¼
pm0 is its nominal pitch.
With reference to the standard (or nominal) pitch circle, having radius
m0 z0
r0 ¼ ; ð7:2Þ
2
the addendum and dedendum are both equal to ð5=4Þm0 , and the teeth have rounded
tips and rounded root fillets having depth equal to ð1=4Þm0 , for which, in respect to
the cutting interference, this pinion-type cutter behaves as if its addendum is only
m0 . The tooth thickness s0 and space width e0 , measured on the nominal pitch
circle, are both equal to pm0 =2, i.e. s0 ¼ e0 ¼ pm0 =2 (see: Buckingham [1], Merritt
[6], Henriot [3], Niemann and Winter [8], Maitra [5]).
Using a pinion-type cutter, we can generate gear wheels having identical charac-
teristics compared to those generated by a rack-type cutter. Consider a gear wheel
having z1 teeth, nominal module m1 ¼ m0 and nominal pressure angle a1 ¼ a0 ,
generated by a rack-type cutter with a profile shift x1 m0 (then with a profile shift
coefficient x1 ), and suppose that it engages with another gear wheel, having z0 teeth
and zero profile-shifted toothing, i.e. x0 ¼ 0 [9].
The operating pressure angle a0 ; in the condition of backlash-free contact
between these two gear wheels, is given by Eq. (6.58). From this equation, written
for the case considered here, we obtain:
x1
inva0 ¼ inva0 þ 2 tan a0 : ð7:3Þ
z1 þ z0
Furthermore, the operating center distance, defined by Eq. (6.59) also written for
the same case, is given by:
cos a0 z1 þ z0 cos a0
a0 ¼ a0 ¼ m0 : ð7:4Þ
cos a0 2 cos a0
If the gear wheel with z0 teeth is a pinion-type cutter, and we place it to the
above center distance, a0 , during the generation of a gear wheel having z1 teeth, this
last will have the same dimension as would have if it had been generated by the
rack-type cutter, with a profile shift equal to x1 m0 .
On a pitch circle shifted by xm0 with respect to the nominal pitch circle, and then
having radius
z
r00 ¼
0
þ x m0 ; ð7:5Þ
2
the pressure angle a0 at a point corresponding to this radius and the correlated
module m0 are given respectively by:
7.3 Characteristic Quantities of the Pinion-Type Cutter Referred … 289
rb0 z0
cos a0 ¼ 0 ¼ cos a0 ð7:6Þ
r0 z0 þ 2x
r0 z0 þ 2x
m0 ¼ m0 ¼ m0 ; ð7:7Þ
r0 z0
The radii of the tip and root circles of the pinion-type cutter are given respec-
tively by:
z0 m0 z
0
ra0 ¼ þ 1:25m0 ¼ þ 1:25 m0 ð7:10Þ
2 2
z0 m 0 z0
rf 0 ¼ 1:25m0 ¼ 1:25 m0 : ð7:11Þ
2 2
The radii of the base circle and cutting pitch circle of a gear wheel having z1 teeth,
and generated with a profile shift coefficient x1 by means of the pinion-type cutter
defined in previous section are given respectively by:
z1 z1 m0
rb1 ¼ rb0 ¼ cos a0 ð7:12Þ
z0 2
0
During the generation, the center distance a1 between the pinion-type cutter and
the gear wheel to be generated is:
where #0s1 is the tooth thickness half angle of the gear wheel on the cutting pitch
circle. From this equation, taking into account Eq. (7.5), with x ¼ x1 , and
Eq. (7.13), which lead to define the equality r00 =r10 ¼ z0 =z1 , and bearing in mind the
Eq. (7.9), we obtain:
z0 z0 p p z0
2#0s1 ¼ 2 #e0 ¼ 2 inva0 inva01 ¼ þ 2 inva01 inva0 :
z1 z1 z0 z1 z1
ð7:17Þ
On a circle having radius r100 , the toothing pressure angle at a point is given by:
rb1 z1 m0
cos a001 ¼ ¼ 00 cos a0 ; ð7:18Þ
r100 2r1
while the tooth thickness angle 2#00s1 and space width angle 2#00e1 are given
respectively by:
p z0 z0
2#00s1 ¼ 2#0s1 þ2 inva01 inva001 ¼ þ2 1þ inva01 2inva001 2 inva0
z1 z1 z1
ð7:19Þ
p z0 z0
2#00e1 ¼ 2 1þ inva01 þ 2inva001 þ 2 inva0 ; ð7:20Þ
z1 z1 z1
because
2p
2 #00s1 þ #00e1 ¼ : ð7:21Þ
z1
7.5 External Gear Pair with z1 … 291
From equations obtained in the two previous sections, we infer that the quantities of
two external gear wheels, having respectively z1 and z2 teeth, and generated by a
pinion-type cutter with profile shift coefficients x1 and x2 , are given by the equations
described in the following subsections. Their calculation procedure, based on the
equations defined earlier as well as in the previous chapters, is not reported.
• Tooth space width angle 2#00e1 on the circle having radius r100
2p p z0 z0
2#00e1 ¼ 00
2#s1 ¼ 2 1 þ inva01 þ 2inva001 þ 2 inva0 ð7:28Þ
z1 z1 z1 z1
• Angular pitch
2p
s1 ¼ ¼ 2 #00s1 þ #00e1 : ð7:29Þ
z1
• Tooth space width angle 2#00e2 on the circle having radius r200
2p p z0 z0
2#00e2 ¼ 00
2#s2 ¼ 2 1 þ inva02 þ 2inva002 þ 2 inva0 ð7:36Þ
z2 z2 z2 z2
• Angular pitch
2p
s2 ¼ ¼ 2 #00s2 þ #00e2 : ð7:37Þ
z2
Now we put the two wheels above so that they are meshing between them with a
backlash-free contact. We denote respectively with rw1 , rw2 and aw the radii of the
two operating pitch circles and operating pressure angle (in this regard, it is
noteworthy that, in accordance with the ISO 6336-1 [4], to indicate quantities
related to the working, or operating, conditions, we use the subscript w, which
replaces the usual prime symbol, when the latter can generate misunderstanding).
The radii rw1 and rw2 are correlated by the following relationship:
rw1 z1
¼ : ð7:38Þ
rw2 z2
On these working pitch circles, the space width between the teeth of the first
wheel must be equal to the tooth thickness of the second wheel, for which
and thus
rw2 00 z2
#00e1 ¼ #s2 ¼ #00s2 : ð7:40Þ
rw1 z1
Substituting in this equation the expressions of #00e1 and #00s2 obtained from
Eqs. (7.28) and (7.35), we get:
z1 þ z0 z2 þ z0 z0
invaw ¼ inva002 ¼ inva01 þ inva02 2 inva0 : ð7:41Þ
z1 þ z2 z1 þ z2 z1 þ z2
a0 ðz1 þ z2 Þ cos a0
a0 ¼ rw1 þ rw2 ¼ ¼ m0 : ð7:42Þ
cos aw 2 cos aw
Taking into account that, during generation, the cutting center distances between
the pinion-type cutter and each of the two wheels are respectively given by
Eqs. (7.25) and (7.33), the radii of the tip and root circles of the pinion and wheel
will be given by:
m0
ra1 ¼ a01 ð z 0 2Þ ð7:43Þ
2
m0
rf 1 ¼ a01 ðz0 þ 2:50Þ ð7:44Þ
2
m0
ra2 ¼ a02 ð z 0 2Þ ð7:45Þ
2
m0
rf 2 ¼ a02 ðz0 þ 2:50Þ: ð7:46Þ
2
Placing the two wheels to mesh with each other in the condition of backlash-free
contact, we obtain that the clearance will be given by:
c ¼ a0 ra1 rf 2 : ð7:47Þ
In the case in which such clearance would be less than ð1=4Þm0 , i.e.
the tip diameters of the two wheels must be decreased by the amount
First, we examine the toothing characteristics of the pinion-type cutter and wheel to
be generated, on the generation, or cutting, pitch circle. We assume that the
pinion-type cutter has z0 teeth, nominal pressure angle a0 , nominal module m0 and
standard sizing. If with this pinion-tool we want to cut a wheel having z1 teeth, with
a profile shift coefficient x1 , the common module, m0 , and the common pressure
angle, a01 , of the toothing of the pinion-type cutter and the wheel that is being
7.6 Possibility of Making Profile-Shifted … 295
generated, both measured on the cutting pitch circles, are those given by Eqs. (7.7)
and (7.14), while the addendum of the pinion-type cutter and wheel will be given
respectively by:
ha0 ¼ m0 ð1 x1 Þ ð7:50Þ
ha1 ¼ m0 ð1 þ x1 Þ: ð7:51Þ
The profile shift coefficient x1 must have a value such as to avoid both the
interference between the tooth tip of the pinion-type cutter and the tooth profiles of
the wheel to be generated, and the interference between the tooth tip of the wheel
and the pinion-type cutter. The first of these two types of interference causes a
weakening of the wheel to its root, because of the undercut, while the second
determines a removal of part of the tooth flank profile, near the tooth tip (this is the
tip relief, which sometimes is done intentionally), and consequently a reduction in
the length of the path of contact.
To avoid the first type of interference, that between the tooth tip of the
pinion-type cutter and the tooth root profiles of the wheel to be generated, it is
necessary that the profile shift coefficient x1 does not fall below a minimum value
x1;min . In this regard, the inequality (4.4) must be satisfied. Therefore, by writing this
inequality for the case here examined and taking into account the toothing char-
acteristics on the generation pitch circles described above, it must be:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ha0
2 z20 þ z21 þ 2z1 z0 sin2 a01 z0 ; ð7:52Þ
m0
from this inequality, taking into account Eqs. (7.7), (7.14) and (7.50), we obtain:
2qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3
2
z0 4 ðz0 þ 2Þ þ z1 ðz1 þ 2z0 Þ cos a0
2
x1;min 15 : ð7:53Þ
2 ðz0 þ z1 Þ
To avoid the second type of interference, that between the tooth tip of the wheel
to be generated and the pinion-type cutter, it is necessary that the profile shift
coefficient x1 does not exceed a maximum value x1;max . In this regard, the inequality
(4.5) must be satisfied. Therefore, by writing this inequality for the case here
examined and taking into account the toothing characteristics on the generation
pitch circles described above, it must be:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ha1
2 z21 þ z20 þ 2z1 z0 sin2 a01 z1 ; ð7:54Þ
m0
from this inequality, taking into account Eqs. (7.7), (7.14) and (7.51), we obtain:
296 7 Profile Shift of External Spur Gear Involute Toothing …
z0 4ðz1 þ 1Þ z0 ðz0 þ 2z1 Þ sin2 a0
x1;max : ð7:55Þ
4ð z 0 þ z 1 Þ ð z 0 2Þ
Therefore, to avoid both types of interference, the values of the profile shift
coefficient x1 that can be chosen must necessarily be comprised between the above
limit values, that is in the range:
Table 7.1 shows the values of x1;min and x1;max , calculated using the relationships
(7.53) and (7.55), both considered as equalities rather than inequalities, in the case
of a pinion-type cutter with z0 ¼ 20 teeth, nominal pressure angle a0 ¼ 15 , and
standard sizing, for wheels to be generated having number of teeth z1 in the range
14 z1 30. The values of x1;min and x1;max shown in this table demonstrate
unequivocally that, with the pinion-type cutter here considered, it is not possible to
generate wheels with z1 19 teeth, without the interference will not occur; instead,
if z1 20, it is possible to generate wheels without the interference will not occur,
but the variability range of the x1 -values we can choose is very narrow.
The above limitations regarding the possibility of making with a generation
process external spur gears with profile-shifted toothing, by means of a pinion-type
cutter, together with the greater complication of the calculations to be performed,
make reason of the fact that these gears are almost exclusively manufactured using a
rack-type cutter.
We now consider a cylindrical spur gear, with pinion and wheel with z1 and z2
teeth, both cut by means of a pinion-type cutter, with profile shift coefficients
respectively equal to x1 and x2 . We want to define the condition that must be
satisfied, in terms of correlation that links x1 and x2 , so that these two toothing
wheels are meshing with each other without variation of center distance, i.e. with
their axis positioned at the nominal center distance.
Consider the first wheel, the one with z1 teeth, generated with a profile shift
coefficient x1 , by means of a pinion-type cutter having z0 teeth, and nominal
0
pressure angle a0 . The pressure angle a1 on the generation pitch circle is given by
Eq. (7.14). For this wheel can engage, without variation of center distance and with
backlash-free contact, with the other wheel having z2 teeth, it is necessary that the
00
working pressure angle aw ¼ a2 , given by Eq. (7.41), satisfies the condition:
aw ¼ a0 : ð7:57Þ
1
inva02 ¼ ðz1 þ z2 þ 2z0 Þinva0 ðz1 þ z0 Þinva01 : ð7:58Þ
z2 þ z0
Therefore, once the profile shift coefficient x1 of the pinion has been chosen, to
have a gear pair with profile-shifted toothing without variation of center distance, cut
by a pinion-type cutter having given characteristics (m0 ; a0 ; z0 , etc.), it is necessary
that a01 ; a02 ; x1 , and x2 are related by the relationships (7.23), (7.58) and (7.59).
Here we examine the more general case of an external gear pair, consisting of a
pinion and wheel with z1 and z2 teeth, having nominal module m0 and nominal
pressure angle a0 , and generated respectively with profile shift coefficient x1 and x2 .
Suppose first that is c 0; 25m0 , so it is not necessary to make a truncation, that is a
reduction of addendum or, what is the same, a reduction in the diameters of the gear
blanks.
298 7 Profile Shift of External Spur Gear Involute Toothing …
To calculate the transverse contact ratio, ea , we use the Eq. (3.44), in which,
however, the addendum factors k1 ¼ ha1 =m0 and k2 ¼ ha2 =m0 are to be determined.
Under the working conditions corresponding to backlash-free contact, we have:
m0 z1 cos a0
ha1 ¼ ra1 rw1 ¼ a01 ð z 2 2Þ m 0 ð7:60Þ
2 2 cos aw
m0 z2 cos a0
ha2 ¼ ra2 rw2 ¼ a02 ðz0 2Þ m0 : ð7:61Þ
2 2 cos aw
Taking into account Eqs. (7.25) and (7.33), and substituting them inside these
last relationships, we obtain:
1 z1 cos a0
k1 ¼ ½z0 ðz1 þ 2 þ 2x1 Þ þ 2z1 x1 ð7:62Þ
2z0 2 cos aw
1 z2 cos a0
k2 ¼ ½z0 ðz2 þ 2 þ 2x2 Þ þ 2z2 x2 : ð7:63Þ
2z0 2 cos aw
References
1. Buckingham E (1949) Analytical mechanics of gears. Book Company Inc, McGraw-Hill New
York
2. Galassini A (1962) Elementi di Tecnologia Meccanica: Macchine Utensili – Principi
Funzionali e Costruttivi, Loro Impiego e Controllo della loro Precisione, Reprint 9th edn.
Editore Ulrico Hoepli, Milano
3. Henriot G (1979) Traité théorique and pratique des engrenages, vol 1, 6th edn. Bordas, Paris
4. ISO 6336-1 (2006) Part 1—Basic principles, introduction and general influence factors
5. Maitra GM (1994) Handbook of gear design, 2nd edn. Tata McGraw-Hill Publishing
Company Ltd, New Delhi
6. Merritt HE (1954) Gears, 3rd edn. Sir Isaac Pitman&Sons, London
7. Micheletti GE (1958) Tecnologie Generali: Lavorazioni dei Metalli ad Asportazione di
Truciolo. Libreria Editrice Universitaria Levrotto & Bella, Torino
References 299
8.1 Introduction
Cylindrical involute spur gears with parallel axes analysed in the previous chapters
have some disadvantages, the main of which consists of the fact that only a part of
each meshing cycle is carried out with two pairs of teeth in contact, while the
remaining part is carried out with only one pair of teeth in contact. This determines
sudden variations in the total length of the contact line (this varies suddenly from
2b to b, depending on whether two tooth pairs or one tooth pair are in contact), and
abruct changes in the mesh stiffness, resulting in a very noisy operation and
working irregularity. For more, a tooth pair comes into contact over its entire face
width, b, and also loses contact for its entire face width, thus accentuating the
discontinuity of working conditions [5, 15, 26, 27].
Parallel cylindrical helical gears, or simply cylindrical helical gears, were born
and developed to overcome the disadvantages of the cylindrical spur gears. They
can be considered as the evolution of the stepped gear wheel proposed for the first
time by Hooke, and therefore called the Hooke’s wheel [4, 10, 13, 28]. This stepped
gear wheel (Fig. 8.1) is however still a cylindrical spur gear in the proper meaning
of the term, as it is constituted by a discrete set of spur gear, all the same and
connected to the same shaft, but rotated by a certain angle relative to one another.
By choosing this angle with value much lower than that corresponding to half of the
angular pitch of the teeth, Hooke obtained a gear characterized by a significant
increase in the length of the path of contact; this is because, when the first wheel of
the set thus formed comes out of contact, the others are still in contact.
A cylindrical helical gear may be considered as the limit configuration of the
Hooke’s wheel. To this end, let us first consider the infinite number of gear wheels,
having differential face widths, db, obtained by cutting a conventional cylindrical
spur gear with planes perpendicular to its axis. Now, let us rotate these gear wheels,
one with respect to the other, in the tangential direction, of a differential arc of pitch
circle, proportional to the increment of coordinate dz along the axis of the same
wheel. In this way, we get a cylindrical helical gear.
A conventional cylindrical spur gear can be considered as generated by a planar
wheel, i.e. one of these infinitesimal gear wheels, when it moves, parallel to itself
(that is, remaining invariably oriented) along its axis. However, this gear wheel with
infinitely small thickness, while moving along its axis, also rotates with respect to it,
the wheel so generated is a cylindrical helical gear. If then, in this motion of
8.1 Introduction 303
generation, the ratio between the displacement speed along the axis and the rotation
speed about it remains constant, the helicoids, which constitute the tooth flanks,
have a constant axial pitch. This is the case of almost all the cylindrical helical
gears, which today are manufactured.
It is noteworthy that the Hooke’s wheel is, in effect, a cylindrical spur gear,
because it does not give rise to axial loads on the shaft, but only generates radial
and tangential loads. The cylindrical helical gear instead generates, in addition to
the radial and tangential loads, also an axial load, correlated to the helix angle, to be
supported by suitable bearings. It is to be noted that the profiles of the teeth of the
cylindrical helical gears that characterize practical applications are involute profiles.
Therefore, we focus our attention on the cylindrical involute helical gears.
It is also to be noted that cylindrical involute helical gears are used both to
transmit motion between parallel axes, and to transmit motion between non-parallel
and non-intersecting axes. In the first case, we have parallel helical gears, which
have the same function as cylindrical spur gears, but with much more interesting
performances, as we will see in the sections below. Instead, in the second case, we
have crossed helical gears. In this chapter, only parallel helical gears are discussed,
while crossed helical gears are dealt with in Chap. 10.
Let us now examine in more detail the generation of a cylindrical involute helical
wheel, obtained by means of the planar wheel that, as it moves along its axis, rotates
around it, with a constant ratio between the displacement speed along the axis, and
the rotation speed around to it. The surface of the tooth flanks of the cylindrical
involute helical wheel thus obtained, is a portion of an involute helicoid, generated
by the helical motion, with lead pz , of an involute curve of a base circle having
radius rb , contained in a transverse plane, i.e. a plane perpendicular to the axis of
the wheel.
With the above generation motion, the circles characterizing the planar gear
wheel generate respectively the cylinders that characterize the parallel involute
helical gear wheel, i.e.: the base and pitch cylinders, and the tip and root cylinders.
The intersections of the tooth flanks with these cylinders, or any other coaxial
cylinder, are cylindrical helices, each with the same lead pz . We can regard the tooth
surface as formed by a family of helices, each helix lying in a cylinder coaxial with
the gear wheel, and each with the same lead. All helices lying on a cylinder of given
radius have the same helix angle b, defined as the angle between the tangent to the
helix, at each point, and the generatrix of the cylinder passing through the same
point (or, in equivalent terms, the angle between the tangent to the helix, and the
axis of the wheel). The helices lying on several coaxial cylinders have the same
lead, but different helix angle [23, 37].
It should be noted that the term helix angle used in connection with a helical gear
means, conventionally, the helix angle at the pitch cylinder, and it is usually
304 8 Cylindrical Involute Helical Gears
denoted by b, without any subscript. The lead angle is the complementary angle to
the helix angle; therefore, it is given by:
c ¼ ðp=2Þ b: ð8:1Þ
The lead angle cc related to any other cylinder of diameter dc ¼ 2rc can be
obtained by the following general relationship:
d r
tan cc ¼ tan c ¼ tan c ; ð8:2Þ
dc rc
the lead pz represents the axial advance of the helix in one complete turn.
More simply, the involute helicoid can also be thought of as generated by the
paths of the points of a straight line, l (generation straight line), belonging to a plane
tangent to the base cylinder, and forming the base helix angle, bb , with the axis of
the wheel, when this plane rolls without sliding on the same base cylinder
(Fig. 8.3). In fact, by doing so, the generating line is wrapped, on the base cylinder,
according to a helix with base helix angle bb , while its points describe, in planes
perpendicular to the wheel axis, involute curves of circles of radius rb , having their
points of regression on this helix. This second way of generation of the involute
helicoid gives a simple explanation of its following fundamental geometry
properties:
• The involute helicoid is a ruled surface, formed by the successive positions of
the generation straight line l. For each point P of its surface passes a straight line
l. The involute helicoid is therefore a developable ruled surface, which can be
also thought of as constituted by the successive tangents to the base helix. The
tooth flanks, as developable ruled surfaces, can therefore be grinded with rel-
ative ease by grinding wheels with developable surface. An involute helicoid
is uniquely defined by the radius rb of the base cylinder, and the base helix angle
bb .
• The normal n to the helicoid surface, in one of its points P, is contained in the
plane tangent to the base cylinder through P, and is normal to the straight line
l through P.
• The plane tangent to the base cylinder passing through any point P of the
involute helicoid surface intersects this last surface along a straight line; it is the
straight line l passing through that point.
• The plane R tangent to the helicoid surface, in one of its point P, is normal to the
plane tangent to the base cylinder through P; it touches the helicoid surface
along the straight line l through P.
306 8 Cylindrical Involute Helical Gears
Two involute helicoids with parallel axes that are touching at a point P must
have the same normal. Therefore, from what we have seen above, all contacts must
occur in the plane tangent to the base cylinder, which therefore constitutes the plane
of action, because it contains the lines of action of the parallel involute helical gear
pair (Fig. 8.4). If, moreover, at a given instant of time, we have contact at a point P,
we must have contact at all points of a straight line l through P and inclined, with
respect to the axis, the base helix angle bb . This because, in all the points of this
straight line, each of the two helicoids has the same direction of the normal that we
have at point P.
The two helicoids are therefore generated by the same straight line l (Figs. 8.3
and 8.4), during the wrapping of the plane of action on the respective base cylin-
ders, and are then to have helix angle equal, but opposite. Therefore, if the helicoid
of a wheel is a right-hand helicoid, that of the mating gear will be left-hand helicoid,
and vice versa. Therefore, the helices on the operating pitch cylinders (or any of
coaxial cylinders between the root cylinders and tip cylinders) are of opposite
direction, but the magnitude of the helix angle (or the lead angle) is the same for
both helices.
In this regard, we recall that we call conventionally right-hand and left-hand
gears those gear wheels whose helicoid surfaces are generated by generating curves
Fig. 8.4 Base cylinders, generation straight line, and plane of action
8.2 Geometry of Parallel Involute Helical Gears 307
that move with right-hand and, respectively, left-hand screw motions. Figure 8.5a,
b show a parallel helical gear pair, consisting of a right-hand pinion and left-hand
wheel, and respectively a helical rack-pinion pair, consisting of a left-hand pinion
and right-hand rack. In addition, let us remember that, according to [16] ISO
1122-1, right-hand teeth and left-hand teeth are teeth whose successive transverse
profiles show clockwise displacement and respectively counterclockwise dis-
placement with increasing distance from an observer looking along the straight-line
generatrixes of the reference surface.
On the plane of action, the two helicoids are touching along a straight line,
which is inclined of the angle bb with respect to the axis, and constitutes the
instantaneous line of action. During the progression of the relative motion, for
well-known properties of the involute curves, this straight line moves parallel to
itself on the plane of action, with a linear velocity equal to the base-line velocity
(i.e. the tangential velocity at the radii of the base cylinders). Figure 8.4 shows:
• the two pitch cylinders, having radii r1 and r2 , touching along the instantaneous
axis of rotation, or pitch line CC 0 , shown with dashes line and passing through
the pitch point C;
• the two base cylinders, having radii rb1 and rb2 ;
• the plane of action, which is limited by the two end faces of the two gear
members, whose distance measured along their axes defines the face width,
b (both gear members have the same face width);
• the instantaneous line of action, in its current position on the plane of action.
The same figure shows the theoretical rectangle of contact, T1 T10 T2 T20 , which is
bounded by the two aforementioned end faces, on two opposite sides, and, on the
other two sides, by the two straight lines of tangency of the plane of action with the
Fig. 8.5 a, parallel helical gear pair (right-hand pinion, and left-hand wheel); b, helical
rack-pinion pair (left-hand pinion, and right-hand rack)
308 8 Cylindrical Involute Helical Gears
base cylinders. In an integrated view of the parallel involute gears, which includes
both spur gears and helical gears, according to the aforementioned [16] ISO 1122-1,
as plane of action we mean the plane containing the lines of action of a parallel
involute gear pair; therefore, it is identified and coincides with the above-defined
theoretical rectangle of contact. As well as for parallel cylinder spur gears, not the
entire surface of this theoretical rectangle of contact or plane of action can be used.
The points T1 and T2 , or better the lines T1 T10 and T2 T20 , represent the loci of
primary interference. To avoid this type of interference, the tip cylinder of the two
members of the parallel helical gear pair must intersect the plane of action along lines
AA0 and EE 0 ; these are on the same plane as lines T1 T10 and T2 T20 , but are moved in the
direction of the instantaneous axis of rotation CC 0 (Fig. 8.6). We identify so the
actual rectangle of contact, AA0 EE 0 . The two opposite sides of this rectangle AA0 and
EE 0 have length equal to the face width b ðAA0 ¼ EE 0 ¼ bÞ, while the other two sides
AE and A0 E 0 have length equal to the length of path of contact ga ðAE ¼ A0 E 0 ¼ ga Þ;
the latter, as shown in Fig. 8.6, is the sum of the length of path of approach, gf , and
length of path of recess, ga , i.e. ga ¼ gf þ ga . It should be noted that all transverse
sections of the actual rectangle of contact are the same, so we call as path of contact
both the whole actual rectangle of contact and any of its transverse sections.
The length of the instantaneous line of action varies during the meshing cycle. In
the initial stage, the length increases from zero to a maximum value, which depends
Fig. 8.6 Parallel helical gear pair: meshing diagram, rectangles of contact, and instantaneous lines
of action on the tooth surface and on the actual rectangle of contact
8.2 Geometry of Parallel Involute Helical Gears 309
on the face width, b, and base helix angle bb ; in the middle stage, it is maintained
constant and equal to the maximum value; in the final stage, it decreases from the
maximum value to zero. These lengths are shown, in real size, in Fig. 8.6 at the
bottom, where the actual rectangle of contact is highlighted, after rotating the plane
to which it belongs to show its contour. Figure 8.6 also shows, at the top, the lengths
of the instantaneous lines of action, drawn on the tooth flank surface, and seen from
the inside. The procedure for the geometrical construction of these length is quite
evident. Of course, depending upon the procedure of geometrical projection, the
length 000 ; 110 ; . . . on the tooth flank surface are different from the actual lengths,
represented on the actual rectangle of contact, shown below. Note that the segments
000 ; 110 ; . . . on the tooth flank surface, if prolonged, are tangent to the base circle.
We immediately notice one of the fundamental differences between parallel spur
and helical gears. In fact, the contact between two parallel spur gears in mesh takes
place always along a contact line extending along the whole face width of the
wheels, this line being always parallel to the axis of the gear. Instead, unlike spur
gears, the instantaneous line of contact of a parallel helical gear is a diagonal across
the operating flank surface of the tooth. This instantaneous line of contact (or
instantaneous line of action) is the intersection of the plane of action, fixed and
invariable, and the surface of the operating flank of the tooth, that instead changes
its position with respect to the plane of action, instant by instant, during the
meshing cycle. The inclination of these diagonal lines with respect to the axis of the
wheel depends on bb , and is much higher the greater is bb .
We saw that the initial contact between two parallel helical gears is a point (point
A 7, in Fig. 8.6), which gradually changes into a line as the meshing cycle
proceeds. Therefore, the teeth come into meshing gradually. Similarly, they come
out gradually from the meshing, because in the final stage of the meshing cycle the
contact locus gradually changes from a line to a point (point E0 , in Fig. 8.6). With
the exception of the case of wheels having face width very thin, in the parallel
helical gears several pairs of teeth are simultaneously in engagement. This fact
explains the many advantages we have compared to parallel spur gears. In fact, the
total stiffness of the teeth (remember that this is the stiffness of all the teeth in
meshing, and that its mean value is the mesh stiffness cc (it is better defined in Vol.
2, Sect. 1.6), fluctuates less, the greater the contact ratio. Furthermore, the rotation
motion is transmitted in a uniform way, and operating noise and noise emissions are
much lower.
Figure 8.7 shows one of the above instantaneous line of action on the surface of
the tooth flank, in a generic position corresponding to its maximum length, lmax ,
which occurs in the middle stage of the meshing cycle. The length, l, of these
diagonal lines varies from zero to the maximum value, lmax , and this maximum
value is different depending on whether:
ga cot bb ? b ð8:4Þ
310 8 Cylindrical Involute Helical Gears
Fig. 8.7 Instantaneous line of action on the tooth flank surface, having maximum length
b
lmax ¼ ; ð8:5Þ
cos bb
During the meshing cycle, the total length, lt , of the line of action changes, with
the position of the teeth. This aspect and its implications are discussed in detail in
Sect. 8.7.
We now intersect the helicoid with two coaxial cylinders, the reference pitch
cylinder and base cylinder. We get two helices, which have the same lead, pz , but
different helix angle, b and bb . Figure 8.9 shows these two helices, having lead pz ,
after having spread out in the flat the two cylinders with radii r and rb . Since pz is
the same for the two helices, we have:
which coincides with Eq. (8.1), in which bc ¼ bb , and rc ¼ rb . Taking into account
that
rb ¼ r cos a; ð8:8Þ
this equation correlates between them the base helix angle, bb , helix angle, b, and
pressure angle, a.
As for a parallel cylindrical spur gear, also the main specific quantities of a parallel
cylindrical helical gear are defined with reference to the helical basic rack, which is
shown in Fig. 8.10. This figure shows the profile of a tooth, both in the transverse
section (section with a plane perpendicular to the wheel axis), and in the normal
section (section with a plane perpendicular to the tooth axis). Of course, the tooth
depth, h, is the same in both of these sections. However, the tooth appears slenderer
in the normal section (its thickness, measured in the rack reference plane, is equal to
Fig. 8.10 Helical basic rack, and transverse and normal section of a tooth
half the normal pitch, pn ), and stubbier in the transverse plane (its thickness,
measured always in the rack reference plane, is equal to half the transverse pitch p).
Four kinds of pitches can be considered for parallel helical gears. The first of
them is the lead of the helix, pz , which we have already defined in the previous
section (see Eq. (8.3)). The other three pitches, which are correlate to the toothing,
are shown schematically in Fig. 8.11. They are respectively:
• the transverse pitch, pt , or simply pitch (of course, p ¼ pt ), i.e. the length of the
arc of the transverse reference circle between two consecutive corresponding
profiles;
• the normal pitch, pn , i.e. the length of the arc of a helix normal to the pitch helix,
between two corresponding profiles;
• the axial pitch, px , i.e. the distance between two consecutive corresponding
profiles, measured along a reference cylinder generatrix.
The relationship that correlate these three pitches (Fig. 8.11) are as follows:
pn ¼ p cos b ð8:10Þ
pn
px ¼ ¼ p cot b: ð8:11Þ
sin b
8.3 Main Quantities of a Parallel Helical Gear 313
By comparison of Eqs. (8.3) and (8.11), taking into account Eq. (8.9), we
obtain:
pz ¼ zpx : ð8:12Þ
Two kinds of modules are introduced in the discussion of the parallel helical
gears: the transverse module mt ¼ m ¼ p=p, and the normal module, mn ¼ pn =p.
According to Eq. (8.10), these two modules are related by the relationship:
mn ¼ m cos b: ð8:13Þ
With reference to Fig. 8.10, we can find the relationship between the transverse
pressure angle, at ¼ a, normal pressure angle, an , and helix angle, b, of the helical
basic rack. We have in fact:
yn ¼ h tan an : ð8:15Þ
yn ¼ y cos b: ð8:16Þ
Therefore, replacing in this last relationship the Eqs. (8.14) and (8.15), we get:
The diameters of the reference pitch cylinder and base cylinder of a wheel with
z1 teeth, which meshes with the aforementioned helical basic rack, in the condition
where the reference plane of the basic rack is tangent to the reference pitch cylinder
of the wheel, are respectively given by:
314 8 Cylindrical Involute Helical Gears
d1 ¼ z 1 m ð8:18Þ
From Eqs. (8.3, 8.8 and 8.9), we infer that the lead of the helix on the reference
pitch cylinder can be expressed as:
pz1 m
pz ¼ : ð8:20Þ
tan b
The teeth flank surfaces of a parallel cylindrical helical gear pair, with z1 and z2
teeth, are involute helicoids having the same base helix angle, bb , on the base
cylinders. For one of the two wheels, the helicoid is a right-hand helicoid, while for
the other wheel it is a left-hand helicoid, or vice versa. Therefore, the transverse
pitches of the two helicoids, measured on the reference pitch cylinders, will be
given respectively by:
pz1 m pz1 m cos a
p1 ¼ ¼ ð8:21Þ
tan b tan bb
pz2 m pz2 m cos a
p2 ¼ ¼ ; ð8:22Þ
tan b tan bb
where m is the module, while a and b are respectively the pressure angle and angle
of inclination of the teeth of the generation rack-type cutter, with respect to the axes
of the two wheels. Equations (8.21) and (8.22) show that the transverse pitches of
the two wheels are proportional to their number of teeth. Of course, the normal
pitch, pn , is the same for the two wheels, as it is easy to show bearing in mind
Eqs. (8.21, 8.22 and 8.10), and considering that z1 =z2 ¼ d1 =d2 .
Referring to one of the two members of the parallel cylindrical helical gear pair,
the transverse base pitch and normal base pitch on the base cylinder will be given
respectively by:
Another key advantage of the parallel cylindrical helical gears against the par-
allel cylindrical spur gears consists of the significant increase in the contact ratio. In
fact, for these gears, we have:
ec ¼ ea þ eb ð8:25Þ
where ec , ea and eb are respectively the total contact ratio, transverse contact ratio
and overlap ratio. Therefore, the total contact ratio is equal to the sum of the
transverse contact ratio and overlap ratio.
8.3 Main Quantities of a Parallel Helical Gear 315
Referring to Fig. 8.6, we can define the transverse contact ratio similarly com-
pared to what we have already done for parallel cylindrical spur gears, that is, as:
ga
ea ¼ ; ð8:26Þ
p cos a
where ga is the length of the path of contact. This quantity is given respectively by
the Eqs. (3.24, 3.33 and 3.40), depending on whether the parallel cylindrical helical
gear pair is an external gear, a rack-pinion gear pair, or an internal gear.
To determine the overlap ratio, it is convenient to introduce the face advance or
tooth advance or offset, U. This is defined as the distance, measured on the
transverse pitch circle, through which a helical tooth moves from the initial position
at which the contact begins at one end face of the wheel to the final position at the
other end face where the contact ceases. This face advance, equal to U ¼ b tan b
(Fig. 8.11), is due to the helical orientation of the tooth course along the face width.
Therefore, we define as overlap ratio, eb , the quotient of the face advance of a
helical tooth divided by the transverse circular pitch, or, what comes to the same
thing, the quotient of the face width divided by the axial pitch. Thus, we have:
U b tan b b b sin b
eb ¼ ¼ ¼ ¼ : ð8:27Þ
p p px pmn
Generally, the face width b of a parallel cylindrical helical gear is made suffi-
ciently large, so that the face advance U corresponding to a given helix angle b is
greater than the transverse circular pitch, p. In this case we have eb [ 1. Thus, from
the theoretical point of view, it would be possible to ensure the continuity and
uniformity of the motion even if ea were zero, that is, with a tooth depth h ¼ 0. In
fact, this happens in some special toothings, as the (W/N)-toothings (Wildhaber/
Novikov toothings), which are characterized by a transversal contact ratio equal to
zero ðea ¼ 0Þ.
We will not talk of these special toothings (on this topic, we refer the reader to:
Niemann and Winter [27], Litvin and Fuentes [24], Radzevich [29, 30]), and focus
our attention on the parallel cylindrical involute helical gears for the usual practical
applications, which are always characterized by both the transversal contact ratio,
and the overlap ratio. Of course, from the kinematic point of view, the transmission of
a uniform motion is ensured if ec 1. In general, however, it is required that ea and eb
are both greater than one (ea [ 1, and eb [ 1), for which ec 2, and this in order to
obtain uniform and silent meshing condition. It is noteworthy that, to ensure conti-
nuity of motion only with the overlap ratio eb , the limiting value of the face advance
must be U ¼ p. For the sake of safety, it is customary to increase the corresponding
value of the face width b by at least 15%, so that from Eq. (8.27) we obtain:
1:15p
b : ð8:28Þ
tan b
316 8 Cylindrical Involute Helical Gears
Finally, we must determine the two principal radii of curvature at any point P on
the surface of the tooth flank, because they are used in calculation of load capacity
of helical gears, and particularly in calculation of surface durability (pitting). These
two principal radii of curvature are one maximum and the other minimum. Since the
helicoid is a developable ruled surface, they occur, the first in the plane passing
through point P and tangent to the base cylinder, and the second in the normal
plane, i.e. in the plane passing through point P, perpendicular to the previous one
and containing the normal n (Fig. 8.3). Of course, the first principal radius of
curvature, the one in the plane passing through point P and tangent to the base
cylinder, is the maximum principal radius of curvature, and is equal to infinity.
To calculate the second principal radius of curvature, the minimum one in the
normal plane passing through point P, we consider the two planes passing through
P, the transverse plane and normal plane, both perpendicular to the plane of action
(Fig. 8.3). The angle between these two planes is bb . According to well-known
properties of differential geometry of surfaces (see do Carmo [8, 9], Gauss [12],
Smirnov [34], Stoker [35]), the radii of curvature in two planes belonging to the
same star of planes, one of which is a principal plane, differ from one another by an
amount equal to the cosine of the angle between them. In other words, with ref-
erence to the case of interest here, if we indicate with qt the radius of curvature in
the transverse plane, the second, which is the minimum principal radius of cur-
vature in the normal plane, qn;min ¼ q1 , will be given by:
qt
q1 ¼ qn;min ¼ : ð8:29Þ
cos bb
qt ¼ r sin a þ gp ; ð8:30Þ
r sin a þ gp
q1 ¼ qn;min ¼ : ð8:31Þ
cos bb
Usually, parallel cylindrical helical gears are generated by generation cutting pro-
cess, carried out by means of gear cutting machines (gear shapers, gear hobbers, or
Fellows gear shapers) that use cutting tools configured in accordance with the
helical basic rack (Fig. 8.10). These cutting tools can be rack-type cutters, hobs or
helical pinion-type cutters; these last, during the generation cutting, mesh with the
helical gear wheel to be cut. The helical teeth of the gear wheel to be cut have helix
8.4 Generation of Parallel Cylindrical Helical Gears … 317
angles equal to those of the teeth of the generating cutter. A little more detail of the
cutting processes of cylindrical helical gear wheeels is made in Sect. 8.11.
Using gear cutting machine that, as cutting tools, use rack-type cutters or hobs
(then, gear shapers and gear hobbers), we can vary the helix angle of the toothing
without changing the cutting tool. Instead, using gear cutting machines that, as
cutting tools, use helical pinion-type cutters (thus, Fellows gear shapers), to vary the
helix angle of the toothing it is necessary to change the generating cutter as well as
the helical guide of the Fellows gear shaper.
The generation of parallel cylindrical helical gears with the so-called method of
the form cutting, using a cutting tool shaped exactly like the tooth space between
two adjacent teeth of the wheel, is now rarely used. However, with this method, to
obtain the normal section of the tooth space between two adjacent teeth of a wheel
with z1 teeth, reference helix angle b, and normal module mn , we must use the same
milling cutter having normal module mn ; which we use to obtain a parallel cylin-
drical spur gear wheel having number of teeth zv , given by (see next section):
z1
zv ¼ : ð8:32Þ
cos3 bb
When rack-type cutters generate parallel cylindrical helical gears, the same
cutting tools used for cutting parallel cylindrical spur gears are used. The gear-tooth
sizing of parallel cylindrical helical gear wheels is therefore carried out using the
nominal values of the normal modules mn (i.e. in the normal section to the axis of
the teeth) standardized by the ISO 54 [17]. The gear-tooth sizing most widely used
for parallel cylindrical helical gears also provides a standard addendum equal to
module mn (in mm) and a standard dedendum equal to 125 times the module mn ,
that is:
ha ¼ m n
hf ¼ 1; 25mn ¼ ð5=4Þmn ð8:33Þ
h ¼ ha þ hf ¼ 2; 25mn ¼ ð9=4Þmn :
As Fig. 8.10 shows, in transverse section the teeth appear to have stub sizing,
because in this section, according to Eq. (8.13), the module is equal to
mn
m¼ ; ð8:34Þ
cos b
and varies with the helix angle b. If we denote by an the pressure angle of the
cutting tool tooth, the pressure angle a of the tooth profile in the transverse section
is greater because, due to Eq. (8.17), we have:
tan an
tan a ¼ : ð8:35Þ
cos b
318 8 Cylindrical Involute Helical Gears
d1 þ d2 z 1 þ z 2 m n
a¼ ¼ : ð8:39Þ
2 2 cos b
Thus, it can be varied by varying the helix angle b. The minimum value that the
center distance can assume is the nominal one, namely that of the cylindrical spur
gear pair, for which b ¼ 0; this minimum value is given by:
z1 þ z2
amin ¼ mn : ð8:40Þ
2
If the parallel cylindrical helical gear is an internal helical gear pair consisting of
a pinion and an internal helical gear wheel, respectively having reference diameters
d1 and d2 , given by Eqs. (8.37) and (8.38), its center distance, a, is given by:
8.4 Generation of Parallel Cylindrical Helical Gears … 319
d2 d1 z2 z1 mn
a¼ ¼ : ð8:41Þ
2 2 cos b
Its minimum value, corresponding to that of the internal spur gear ðb ¼ 0Þ, will
instead be given by:
z2 z1
amin ¼ mn : ð8:42Þ
2
z1 þ z2 amin
cos b ¼ mn ¼ : ð8:43Þ
2a a
If, for a given value of the center distance, a, the helix angle b is too large, the
axial loads on shafts and bearings will be equally large, as shown later. In order to
reduce these axial loads within acceptable values, there are two possible solutions:
with the first solution, we can increase z1 and z2 , holding constant the gear ratio
u ¼ z2 =z1 , and so respecting an input generally imposed; with the second solution,
we can impose an acceptable value of the reference helix angle b, and make use of
profile-shifted toothings (see Sect. 8.6).
For the evaluation of all the remaining quantities, we proceed in a similar way to
what we said for parallel cylindrical spur gears. Table 8.1 summarizes the main
quantities of an external parallel cylindrical helical gear.
Table 8.1 Main quantities of an external parallel cylindrical helical gear generated by rack-type
cutter
Quantity Pinion Wheel
Number of teeth z1 z2
Reference pitch circle diameter d1 ¼ z1 mn = cos b ¼ z1 m d2 ¼ z2 mn = cos b ¼ z2 m
Tip circle diameter da1 ¼ d1 þ 2mn da2 ¼ d2 þ 2mn
Root circle diameter df 1 ¼ d1 2:50mn df 2 ¼ d2 2:50mn
Base circle diameter db1 ¼ d1 cos a db2 ¼ d2 cos a
Nominal tooth thickness and space sn ¼ en ¼ pn =2 ¼ pmn =2
width in the normal section
Nominal tooth thickness and space s ¼ e ¼ p=2 ¼ pm=2
width in the transverse section
Reference center distance a ¼ ðd1 þ d2 Þ=2 ¼ ½ðz1 þ z2 Þ=2ðmn = cos bÞ
320 8 Cylindrical Involute Helical Gears
In accordance with the geometry of the parallel cylindrical involute helical gears
described in Sect. 8.2, it is evident that the involute profile of the tooth occurs only
in transverse section; in the normal section instead the tooth profile is not an
involute curve. Therefore, the meshing process between conjugate profiles takes
place only in the transverse section, for which, for the calculation of the center
distance, as well as for the calculations to be done in connection with the use of
profile-shifted toothings, the profiles to be considered are those in the transverse
section. In addition, it is to be noted that the tooth flanks roll and slid only in the
direction of the tooth depths.
Rebus sic stantibus, consider the meshing in the reference condition (in this case,
the datum plane of the rack is tangent to the reference cylinder of the wheel),
between a cylindrical helical wheel and the mating rack, and let us ask ourselves the
question about the meaning of Eq. (8.17), which correlates an , a, and b. Both for
the rack, and for the wheel, it certainly has a meaning to speak of pressure angle, a,
in the transverse section. Similarly, for the rack, it also has a well-defined meaning
to speak of pressure angle, an , in the normal section. For the wheel instead, at least
from the theoretical point of view, it has no meaning to speak of pressure angle, an ,
in the normal section, since in this section the tooth profile is not an involute curve,
and notoriously the pressure angle is related to the involute.
On the other hand, the Eqs. (8.10) and (8.13) have meaning for both members of
the gear under consideration, i.e. the wheel and mating rack. In fact, for the wheel,
the normal pitch, pn , is the length of a portion of a helix on the reference cylinder,
perpendicular to the pitch helices, between two adjacent pitch helices. This portion
of helix corresponds to the length of wrapping of the normal pitch, pn , of the rack,
when the datum plane of the latter rolls without sliding on reference cylinder of the
wheel.
We see now to define a parallel cylindrical spur gear that, especially from the
point of view of the load carrying capacity calculations to do, but also from other
points of view that will clarify in turn, is equivalent to the parallel cylindrical helical
gear that we are analyzing. The introduction of an equivalent parallel cylindrical
spur gear is very useful, as it is thus possible to use, albeit with the necessary
8.5 Equivalent Parallel Cylindrical Spur Gear and Virtual … 321
modifications, the relevant spur gear formulae, applying them to parallel cylindrical
helical gears, which are analogous to spur gears.
To this end, we return to consider the meshing, in the reference condition, between
a cylindrical helical wheel and the mating rack (Fig. 8.12). The helix angle is b.
A plane which is normal to the rack tooth (evidently, it is also normal to the wheel
tooth) intersects the datum plane of the rack along a straight line, l ¼ MN, which is
inclined with respect to the axis of the wheel by an angle equal to ½ðp=2Þ b. The
same plane intersects instead the reference pitch cylinder of the wheel according to an
ellipse, having minor semi-axis equal to the radius, r, of the reference cylinder (CO, in
Fig. 8.12), and major semi-axis equal to r= cos b (MO, in Fig. 8.12).
During the relative motion, characterized by the rolling without sliding of the
datum plane of the rack on the reference pitch cylinder of the wheel, the straight line
l is wrapped on this cylinder describing a helix, which is a skew curve, i.e. a curve
in three-dimensional space. Therefore, from a theoretical point of view, it is not
possible to speak of meshing in a normal section. Instead, we can talk of meshing in
this normal section in approximate terms, if we confuse, within the small portion of
contact that concerns us, the above helix with its osculating circle. As known from
differential geometry of curves in two- and three-dimensional space, the radius of
this osculating circle is equal to the quotient of the square of the major semi-axis of
the ellipse, divided by the minor semi-axis, and therefore equal to (see, for example:
Gray et al. [14], Kreyszig [22] ):
2rv 2r z
zv ¼ ¼ ¼ : ð8:47Þ
mn m cos b cos3 b
3
With reference to the normal section, a cylindrical helical gear wheel of z teeth
can be considered approximately equivalent to a virtual cylindrical spur gear wheel
having the axis perpendicular to the plane of the normal section, module mn ,
diameter zv mn , and pressure angle an . Therefore, the study of the meshing of a
parallel cylindrical helical gear pair can be made also with reference to the normal
section, introducing the equivalent virtual parallel cylindrical spur gear pair,
characterized by: pressure angle, an ; module, mn ; numbers of teeth zv1 ¼ z1 = cos3 b,
and zv2 ¼ z2 = cos3 b.
The approximation above is, in some respects, similar to the well-known
Tredgold’s approximation (see, for example: Budynas and Nisbett [3], Giovannozzi
[13], Niemann and Winter [27], Collins et al. [6]) , which we will describe in the
next chapter on bevel gears (see Buchanan [2]). It is related to the fact that the
generation cutting of the toothing of a cylindrical helical wheel is performed by
322 8 Cylindrical Involute Helical Gears
Fig. 8.12 Meshing between rack and cylindrical helical wheel: transverse and normal sections;
ellipse and osculating circle in normal section
cutting tools having standard sizing in the normal section. Without the fact that calls
into question the technological process of the gear generation, this approximation is
not justified. It would be much easier to consider the profiles in transverse section;
for these transverse profiles in fact all the formulae, considerations and gearing laws
of the cylindrical spur gear are valid.
However, for calculation that do not require very high precision, this approxi-
mation allows using the formulae, tables and diagrams relating to the cylindrical
spur gears without profile-shifted toothings, with the only change consisting of the
substitution of an to a, and zv1 and zv2 to z1 and z2 . In any case, this approximation
8.5 Equivalent Parallel Cylindrical Spur Gear and Virtual … 323
cannot be used for calculations that require great precision, which are those con-
cerning parallel cylindrical helical gears with profile-shifted toothing, and the
quantities related to them.
The above mathematical discussion, which led us to define the virtual number of
teeth, zv , of a cylindrical helical gear wheel, in the terms expressed by Eq. (8.47), is
intended to be sufficiently accurate for all practical applications. To obtain the exact
value of zv , it is necessary to consider the section passing along the course of the
helix, and this results in a surface in three-dimensional space as it is a helicoid, and
not in a two-dimensional section, i.e. a plane section as the normal section that we
introduced above. This exact value of zv is given by:
z
zv ¼ : ð8:48Þ
cos2 bb cos b
Since, due to technological reasons described above, the nominal sizing of the
toothing is the one we have in the normal section, the transverse profile of the tooth,
in relation to the transverse module m, is wider and less deep than it is with
reference to the nominal sizing in the normal section. In fact, the addendum factor,
k, in the transverse section will be given by:
ha ha
k¼ ¼ cos b ¼ kn cos b; ð8:49Þ
m mn
where kn (of course, kn [ k as cos b\1) is the addendum factor in the normal
section. In the same way, we shall have:
hf hf
¼ cos b ð8:50Þ
m mn
sn
s¼ : ð8:51Þ
cos b
All this is due to the fact that, as Fig. 8.12 shows, the addendum, dedendum, and
tooth depth of the cylindrical helical gear wheel under consideration do not change
in the normal and transverse sections.
Also, the minimum number of teeth that can be obtained by generation cutting,
using a rack-type cutter, without incurring the risk of primary interference and
undercut, is reduced by a factor equal to cos3 b. In fact, with reference to this type
of interference, the cylindrical helical wheel of z teeth, normal module mn , and helix
angle b, acts as a cylindrical spur wheel having normal module mn and virtual
number of teeth zv ¼ z= cos3 b.
As for the cylindrical spur gear (see Sect. 3.2), also in this case it is admitted
that, in practice, the minimum number of teeth that can be cut by generation cutting
using a rack-type cutter is equal to 5/6 of the theoretical values. Figure 8.13, which
is valid for cylindrical helical wheels obtained by generation cutting using a
rack-type cutter, shows the variation of the theoretical and practical (in this case a
324 8 Cylindrical Involute Helical Gears
Like cylindrical spur gears, parallel cylindrical helical gears can also have
profile-shifted toothings when needed. The same reasons for which we use cylin-
drical spur gears with profile-shifted toothings are also valid for parallel cylindrical
helical gears; in this regard, we refer the reader to what we have described in the
Chap. 6. Here the subject is dealt with in its most general terms, i.e. with reference
to the case of profile-shifted toothings with variation of center distance, as the case
of profile-shifted toothings without variation of center distance is nothing but a
special case of the general subject. In addition, we will focus our attention only on
special features that characterize the parallel cylindrical helical gears compared to
the cylindrical spur gears.
The study of these profile-shifted toothings could be made with reference to the
transverse profiles of the wheel and transverse section of the mating rack. However,
in doing so, the formulae, tables and diagrams that we have introduced previously
(see Chap. 6), concerning the profile-shifted toothing of cylindrical spur gears,
could not be used, since they are inherent to generation cutting performed by a
rack-type cutter sized with reference to the normal section. In order to use these
formulae, tables and diagrams, it is therefore convenient to refer to the normal
profiles of the wheel teeth, for which the mating rack and the rack-type cutter have
their sizing referred to the normal section.
The reference framework is therefore that related to the aforementioned gener-
ation cutting process, with cutting tools sized in normal section. For the study of
parallel cylindrical helical wheels, generated with reference to a cutting plane of the
rack-type cutter shifted with respect to its datum plane, we proceed as if the
to-be-cut wheel and the generation rack-type cutter were a rack-wheel spur gear
pair, whose characteristic quantities correspond to those of the equivalent virtual
gear shown in Fig. 8.12.
It should be noted here that the subject is discussed in a traditional way, i.e.
assuming that the two members of the parallel helical gear pair are free from
assembly and manufacture errors. In this theoretical case, the instantaneous line of
contact between the mating tooth flank surfaces is a straight line. However, the
practical applications of any gear pair are always characterized by errors and
deviations from the theoretical conditions. The cylindrical parallel gears (both spur
and helical gears) with involute tooth profiles are particularly sensitive to the
above-mentioned errors. The parallel helical gears are sensitive to deviations from
axis parallelism and helix errors, which cause discontinuous linear functions of
transmission errors, resulting in vibration and noise. To reduce as far as possible the
effects of these deviations and errors, other types of modifications of conventional
involute helical gears can be made, such as the crowning of the pinion in the profile
and longitudinal directions, which was proposed by Litvin et al. [25]. Here the topic
is confined to the profile shift only.
326 8 Cylindrical Involute Helical Gears
d1 z1 mn cos at0
rb1 ¼ cos at0 ¼ ð8:53Þ
2 2 cos b
d2 z2 mn cos at0
rb2 ¼ cos at0 ¼ : ð8:54Þ
2 2 cos b
x1 þ x2
inva0t ¼ invat0 þ 2 tan a0 ; ð8:55Þ
z1 þ z2
mt0 x1
x01 ¼ x1 ¼ ð8:56Þ
mn0 cos b
mt0 x2
x02 ¼ x2 ¼ : ð8:57Þ
mn0 cos b
x01 þ x02
inva0t ¼ invat0 þ 2 tan a0 : ð8:58Þ
z1 þ z2
The operating center distance, a0 , of the helical gear pair with profile-shifted
toothings and variation of center distance is given by:
Since the operating center distance is related to the reference center distance, a0 ,
given by Eq. (6.11), by means of Eq. (6.59), we can write:
cos at0
a0 ¼ a0 ð8:60Þ
cos a0t
where
a0 ¼ a0 ð 1 þ Y Þ ð8:61Þ
cos at0
Y¼ 1 : ð8:62Þ
cos a0t
To determine the addendum reduction factor, K, and truncation Kmn0 , we use the
same procedure described in Sect. 6.6.1, for which we have:
hz þ z i
1 2 z1 þ z2 0 0
a ¼ mt0 þ ðx1 þ x2 Þ ¼ mn0 þ x1 þ x2 ð8:63Þ
2 2 cos b
where
328 8 Cylindrical Involute Helical Gears
x01 þ x02
X0 ¼ 2 ð8:66Þ
z1 þ z2
cos at0
Y0 ¼ 1 : ð8:67Þ
cos a0t
The radii r1 and r2 of the cutting pitch cylinders are given by:
d1 z1 mt0 z1 mn0
r1 ¼ ¼ ¼ ð8:68Þ
2 2 2cosb
d2 z2 mt0 z2 mn0
r2 ¼ ¼ ¼ : ð8:69Þ
2 2 2 cos b
The radii r10 and r20 of the operating pitch cylinders are given by:
z1 mt0 cos at0 z1 cos at0
r10 ¼ ¼ mn0 ð8:70Þ
2 cos a0t 2 cos b cos a0t
z2 mt0 cos at0 z2 cos at0
r20 ¼ ¼ mn0 : ð8:71Þ
2 cos a0t 2 cos b cos a0t
Finally, the helix angle b0 on the operating pitch cylinders is given by:
r10 r0
tan b0 ¼ tan b ¼ tan b 2 : ð8:72Þ
r1 r2
Example 2 This example concerns an external parallel cylindrical helical gear pair,
with an increase in the center distance. Input data: z1 ¼ 20; z2 ¼ 40; a0 ¼ 105 mm;
b ¼ 30 ; standard range of rack-type cutters. We will have:
0
mn ¼ 3:031mm (this value is sufficiently accurate, but approximate, since the cosine of the
operating helix angle differs very little from that of the cutting helix angle)
m0 ¼ 10 mm, and a0 ¼ 20 (this is the initial choice of design)
a0 ¼ m0 ðz1 þ z2 Þ=2 cos b ¼ 103:92 mm
Since a0 [ a0 , we have to make a profile shift with increase in the center distance
0
at ¼ 24 90 3500
Y 0 ¼ 0:0103
X 0 ¼ 0:0120
0 0
x1 þ x2 ¼ 0:36.
330 8 Cylindrical Involute Helical Gears
Example 3 This example concerns an internal parallel cylindrical helical gear pair,
with decrease in the center distance. Input data: z1 ¼ 10; z2 ¼ 50; a0 ¼ 105 mm;
b ¼ 20 ; standard range of rack-type cutters. We will have:
0
mn ¼ 4:93 mm (this value is sufficiently accurate, but approximate, since the cosine of the
operating helix angle differs very little from that of the cutting helix angle)
m0 ¼ 5 mm, and a0 ¼ 20 (this is the initial choice of design)
a0 ¼ m0 ðz2 z1 Þ=2 cos b ¼ 106:40 mm
Since a0 \a0 , we have to make a profile shift with decrease in the center distance.
Y 0 ¼ ½ða0 =a0 Þ 1 ¼ 0:014
X 0 ¼ 0:014
0 0
x2 x1 ¼ 0:28.
0 0
We can choice, for example, x2 ¼ 0, and x1 ¼ 0:28
K¼0
h 0
i
da1 ¼ m0 þ 2 1 þ x1 ¼ 66:00 mm
z1
cos b
h 0
i
da2 ¼ m0 cos
z2
b 2 1 x2 ¼ 255:95 mm
h ¼ 2; 25m0 ¼ 11:25 mm.
In Sect. 8.2 we have seen that, during the meshing cycle, the instantaneous line of
action moves on the surface of the tooth flank, and that its length varies from zero to
a maximum, which assumes a different value according to whether ga cot b b ? b.
We have also seen that the total length, lt , of the line of action changes with the
position of the teeth. Let us now evaluate this total length, also variable during the
meshing cycle, and its implications.
In parallel cylindrical spur gears, which are typically characterized by a trans-
verse contact ratio between 1 and 2 ð1\ea \2Þ, the total length of the line of
contact, lt , undergoes an abrupt change, passing from the value b, when only one
pair of teeth is in meshing (single contact), to the value 2b, when two pairs of teeth
are in meshing (double contact). It is to be also noted that, in the cylindrical spur
gears, the flank line and the instantaneous line of contact, on the plane of action,
coincide.
Instead, in the parallel cylindrical helical gears the total length of the line of contact
fluctuates between a maximum value lt;max ; and a minimum value, lt;min . In addition,
the active face width, bv (or equivalent face width) and the transverse contact ratio,
ea , oscillate respectively between a maximum value, bv;max , and a minimum
value bv;min , and between a maximum value ea;max , and a minimum value, ea;min .
8.7 Total Length of the Line of Action 331
It is to be noted that, also in the parallel cylindrical helical gears, the flank line and
the instantaneous line of contact, on the plane of action, coincide. While, however,
in the cylindrical spur gears, these lines are parallel to the axes of the two wheels, in
parallel cylindrical helical gears they are inclined, in the plane of action, by an angle
equal to the base helix angle, and, in the various coaxial cylinders between the tip
and root cylinders, by helix angles corresponding to them.
With reference to Fig. 8.14, which shows the actual rectangle of contact, that is
the path-area of contact in which some instantaneous lines of action are represented,
we can express the length of the transverse path of contact, ga , as:
ga ¼ ea pb ¼ ðW þ X Þpbt ; ð8:74Þ
where pbt is the transverse base pitch, W is an integer, and X is a fractional number
less than 1. Thus, we have ea ¼ ðW þ X Þ. For example, in the usual case where
1\ea \2, we have W ¼ 1, and 0\X\1. Similarly, we can express the face width,
b, as:
b ¼ eb px ¼ ðY þ Z Þpx ð8:75Þ
where px is the axial pitch, Y is an integer, and Z is a fractional number less than 1.
Thus, we have eb ¼ ðY þ Z Þ. Figure 8.14 gives a clear graphic evidence of
Eqs. (8.74) and (8.75), in the case in which W ¼ 1 and Y ¼ 2.
The actual rectangle of contact does not change during the meshing cycle, but
the instantaneous lines of action, whose inclination with respect to the axes of the
two wheels remains unchanged (it is equal to the base helix angle, bb ), move
parallel to themselves. In relative terms, we can deal with the problem of deter-
mining the total length lt keeping fixed the position of these instantaneous lines of
action compared to the actual rectangle of contact, and by translating the latter on
the plane of action, parallel to itself, in the direction perpendicular to the genera-
trixes of the base cylinders.
In doing so, Karas [20] obtained the following equations that express the
quantities defined above: (see also [27] ):
• For Z ð1 X Þ
bv;min lt;min 1
ea;min ¼ ¼ ¼ ½ðW þ X ÞY þ WZ ð8:76Þ
b lt;m eb
• For Z [ ð1 X Þ
bv;min lt;min 1
ea;min ¼ ¼ ¼ ½ðW þ X ÞY þ WZ þ ðX þ Z 1Þ ð8:77Þ
b lt;m eb
• For X Z
bv;max lt;max 1
ea;max ¼ ¼ ¼ ½ðW þ X ÞY þ ðX þ Z Þ ð8:78Þ
b lt;m eb
• For X [ Z
bv;max lt;max 1
ea;max ¼ ¼ ¼ ½ðW þ X ÞY þ 2Z ð8:79Þ
b lt;m eb
and maximum values. The ratio bvir =b then allows to calculate the contact ratio
factor for pitting, Ze , defined by the relationship:
1
Z2e ¼ ; ð8:80Þ
ðbvir =bÞ
we will talk about this contact ratio factor in Vol. 2, Chap. 1, dealing with surface
durability (pitting) of spur and helical gears.
Load analysis of parallel cylindrical spur and helical gears can be done with the
same procedure, unique for both types of gears, bearing in mind that parallel
cylindrical spur gears are a special case of parallel cylindrical helical gears. This
general procedure neglects the contribution of the friction forces, but it gives results
of great engineering value. Here we will discuss the more general case of parallel
cylindrical helical gears, which is characterized by an additional force component
(the one caused by the helix angle) with respect to the parallel cylindrical spur
gears. For this more general case, we here get the related relationships, which we
will eventually specialize for parallel cylindrical spur gears.
The total tooth force, or simply tooth force, Fn , acts normal to the tooth surface,
and its direction (namely the normal to the instantaneous line of action in the plane
of action) is the same for all points of any instantaneous line of action. This tooth
force can be resolved into three components, which act at right angles to one
another. In order to establish the interrelations between these components, we
examine Fig. 8.15, showing the meshing diagram of a parallel cylindrical helical
gear pair, as well as the actual rectangle of contact after its rotation about the
common tangent to the base circles, which are not shown in the figure. In the plane
of this rectangle, the force Fn can be resolved into two components: the axial
component, Fa ¼ Fn sin bb , acting in the direction parallel to the axes of the two
wheels, and the component Fn cos bb , acting in the direction normal to the contact
generatrix between the pitch cylinders.
This last component is inclined at an angle equal to the pressure angle, a, with
respect to the common plane tangent to the two pitch cylinders, passing throught the
pitch point C. This component, once translated along its own straight line of
application until it meets the aforementioned generatrix at point C, can be in turn
resolved into a circumferential force Ft ¼ Fn cos bb cos a (i.e. the nominal trans-
verse tangential load at reference cylinder per mesh, as it is called by the ISO
6336-1 [18]), and a radial force Fr ¼ Fn cos bb sin a. Therefore, the three compo-
nents of the tooth force Fn are as follows:
334 8 Cylindrical Involute Helical Gears
2000T
Ft ¼ Fn cos bb cos a ¼ ð8:81Þ
d
Ft tan bb
Fa ¼ Fn sin bb ¼ ¼ Ft tan b ð8:82Þ
cos a
Ft tan an
Fr ¼ Fn cos bb sin a ¼ ¼ Ft tan a; ð8:83Þ
cos b
where T is the driving torque (in Nm), d is the pitch circle diameter (in mm), b and
bb are the helix angles on the reference and base cylinders respectively, and an and
at ¼ a are the pressure angles in the normal section and transverse section
respectively. All the force components are expressed in newtons (N).
It is to be observed that, at equal circumferential component Ft , the total tooth
force Fn has a constant intensity, but its straight line of application, from a general
theoretical standpoint, changes position in dependence of the length of the instan-
taneous line of action, and the load sharing along it. With reference to the plane of
actual rectangle of contact shown in Fig. 8.15, in the case where only one pair of
8.8 Load Analysis of Parallel Cylindrical Helical Gears … 335
teeth was in meshing, the following changes in the position of the line of application
of the total tooth force Fn are evident. In fact, at the start of the meshing cycle, the
force Fn would pass through point A, in the course of the meshing cycle moves
parallel itself in the plane of rectangle of contact, and at the end of the meshing cycle
would pass through point E 0 . This behaviour would cause a variation in the loads
acting on the bearings, even if the intensity of the force Fn remains constant.
In practical applications, however, the parallel cylindrical helical gear pairs are
characterized by a total contact ratio ec (sum of the transverse contact ratio, ea , and
overlap ratio, eb ) very high, for which several pairs of teeth are simultaneously in
meshing. Therefore, with reference to the problems that interest us here, including
those related to the load carrying capacity of the gear, we can assume, with good
approximation, that the tooth force Fn and its three components are applied in the
middle planes of the two wheels, i.e. the planes bisecting the face widths.
The axial component Fa of the tooth force Fn is the thrust which acts along the
axis of the two members of the helical gear pair. Determining the intensity and
direction of this thrust is of fundamental importance in the helical gear design.
Direction of the thrust is a function of several factors, such as: the direction of helix,
i.e. right-hand helix or left-hand helix; the direction of rotation of the two members
of the helical gear pair; the relative positions of these two members. Direction of
this thrust changes by changing any one of these three factors. It is to keep in mind
that, from the design point of view, the directions of the helices of the two gear
members can be fixed only when the relative position between driving and driven
members has been established, and the direction of the thrust has been determined.
In accordance with Eq. (8.82), the axial force, or thrust, increases in proportion
to tan b; then the helix angle should be chosen carefully. For the usual parallel
cylindrical helical gears, it is prudent to limit the helix angle b within ð20 25Þ , in
order to avoid excessive values of the thrust. In exceptional cases, in which the
negative effects of the thrust are kept under control, the helix angle may go up to
30°, but for normal applications it should not exceed 20°. Of course, the intensity of
the axial thrust influences the choice of the bearings that support the shafts on
which the helical gears are mounted.
Finally, in accordance with what we said at the beginning of this section, it
should be remembered that the Eqs. (8.81–8.83) can be used for calculating the
loads acting on parallel cylindrical spur gears, as they are a special case of parallel
cylindrical helical gears. To this end, just put bb ¼ b ¼ 0 in them. So we get:
2000T
Ft ¼ Fn cos a ¼ ð8:84Þ
d
Fa ¼ 0 ð8:85Þ
Double helical gears or herringbone gears are parallel cylindrical gears having part
of the face width with right-hand teeth and part with left-hand teeth, with or without
a gap between them. They are used in those cases where the axial thrust of an usual
helical gear pair creates problems for the bearings or it is not sustainable for any
other reason concerning the design choices. A double helical gear wheel is indeed a
combination of two helical gear wheels, placed side by side, and having usually the
same helix angle but of opposite hands; it should be noted, however, that the
inclination of the two hands must not necessary be the same.
In accordance with the cutting method used, these herringbone gear wheels can
be realized in one of the following ways:
• by cutting the double toothings (one with right-hand helix, and the other with
left-hand helix, or vice versa) on the same gear blank, as Fig. 8.16a shows;
• by cutting a simple toothing (i.e. not double) on two distinct gear blanks, one
with right-hand helix, and the other with left-hand helix, or vice versa, then
connected between them by means of threaded connections, as Fig. 8.16b, c
show;
• by connecting between them two distinct gear wheels, each characterized by a
simple helical toothing, one with right-hand helix, and the other with left-hand
helix, or vice versa, as Fig. 8.16d shows.
When the two double-helical gear wheels are cut using hobs, in order to relief of
the cutting tool, a suitable groove is made on the gear blank between the right-hand
and left-hand helical halves. Instead, when these double-helical gears are cut with
gear-shaping machines that use a pair of pinion-type cutters or a pair of rack-type
cutters, the toothing is continuous in the section that bisects the face width, without
a groove or gap, for which teeth are configured in the shape of an arrow.
If the direction of rotation is in one direction only, it is convenient to place the
arrow tips (or the apices of the teeth that converge toward the central groove) in the
direction of the tangential velocity of the gear wheel. In this way, the start of the
meshing and the impact of first contact related to it occur just in correspondence of
these apices, where the strength of the teeth is greater, due to the convergence of the
opposite teeth towards an area where the stiffness is greater. When the direction of
rotation is not unique, but in both directions, the teeth do not extend up to the end
faces of the gear wheel; this instead ends with two reinforcement disks at its two
end faces.
8.9 Double-Helical Gears 337
(c) (d)
The general concepts of efficiency of the parallel cylindrical spur gears, which we
have discussed in Sect. 3.9, can still be applied to the parallel cylindrical helical
gears, with the appropriate variations of the case. Here we dwell on these variations,
which are peculiar to the parallel cylindrical helical gears. The two helicoids, which
form the flank surfaces of two mating teeth, are in contact at any instant at the
points of the instantaneous line of contact. This line, as we have seen in Sect. 8.2, is
the common generation line of the two helicoids, as well as the intersection of the
same helicoids with the plane of action (Fig. 8.6).
Consider now the parallel helical gear pair shown in Fig. 8.17, at the instant in
which the instantaneous line of contact is the straight line, l. This line is highlighted
in Fig. 8.17b, which shows the actual rectangle of contact AA0 EE 0 in the front view,
0
as well as the instantaneous axis of rotation Ci Ci where the two pitch cylinders are
tangent. Let us then consider a differential segment, ds, around the generic point
P of this instantaneous line of contact. Considering the friction in addition to what
we said in Sect. 8.8, the instantaneous infinitesimal force exerted by the driving
profile on the driven profile has two components, i.e.: a normal component, dFn ,
which lies in the plane of action and is directed as Fig. 8.17b shows (see also
Fig. 8.15); a friction component, dFl , which is directed perpendicular to the plane
(a) (b)
Fig. 8.17 Infinitesimal force components in a parallel cylindrical helical gear pair: a friction
component; b normal component
8.10 Efficiency of Parallel Helical Gears 339
of action and is oriented in the opposite direction to the relative rotation vector
x ¼ ðx2 x1 Þ of the driven wheel with respect to the driving wheel (Fig. 8.17a).
This infinitesimal friction component is related to the infinitesimal normal com-
ponent by the known relationship, dFl ¼ ldFn , where l is the coefficient of sliding
friction.
Taking into account the relationships obtained in Sect. 3.9, in the case here
examined the energy lost by friction (or lost work) on the differential segment ds in
the infinitesimal time interval dt is given by:
dWl dFn
¼l dsjx2 x1 jðCPÞ; ð8:87Þ
dt ds
where CP denotes the distance of the generic point P from the instantaneous axis of
rotation, measured along the direction perpendicular to the latter (it should be noted
that, in order to avoid misunderstandings, in Fig. 8.17 we have indicated the in-
0
stantaneous axis of rotation Ci Ci , that is differently compared to Fig. 8.6 and
Fig. 8.15, in which it is indicated by CC 0 ). This distance is given by Eq. (3.92),
with unchanged meaning of the symbols that appear in it. If s is the distance of
point P from point C
where the instantaneous line of contact, l, intersects the axis
0
Ci Ci (this distance is measured along the straight line, l), from Fig. 8.17b and
taking into account the Eq. (3.92), we obtain:
cos a
s ¼ C
P ¼ CP= sin bb ¼ u1 r1 ; ð8:88Þ
sin bb
where u1 and r1 are respectively the angular displacement and radius of the pitch
circle of the gear wheel under consideration. Then, differentiating this last rela-
tionship with respect to u1 , we get:
cos a
ds ¼ r1 du : ð8:89Þ
sin bb 1
However, the values of u1 corresponding to the points Al and El (Fig. 8.17b) are
equal to the approach angle, ðe1 =r1 Þ, and respectively to the recess angle, ðe2 =r1 Þ,
which we have already introduced in Sect. 3.9. Therefore, assuming that ðdFn =dsÞ
is a constant along the instantaneous line of contact, l, the instantaneous power
dissipated by the friction acting on the mating tooth pair under consideration, for l
variable, will be given by:
340 8 Cylindrical Involute Helical Gears
" #
dWl dFn 2 cos2 a e1Z=r1 e2Z=r1
¼ jx2 x1 j r lu1 du1 þ lu1 du1 ; ð8:91Þ
dt ds 1 sin bb 0 0
On the other hand, the rate at which the work carried out by the driving wheel is
done (that is, the power driven by the driving wheel), corresponding to the dif-
ferential segment, ds, along which the engagement between the mating tooth pair
under consideration occurs, is given by:
dWm dFn
¼ r1 x1 ðds cos bb cos a þ lu1 ds cos a lds sin aÞ; ð8:93Þ
dt ds
where we must take the minus sign or the plus sign, depending on whether the
differential segment, ds, belongs to the segment C
Al or to the segment C
El
(Fig. 8.17b). Consequently, taking into account the previous equations, we get:
dWm dFn cos2 a e1 e2 r1 cos2 a e21 e22
¼ r1 x1 cos bb r1 þ þl þ 2
dt ds sin bb r1 r1 2 sin bb r12 r1
r1 cos a e1 e2
þ lsina
sin bb r1 r1
dFn e1 e2 l e21 þ e22 l tan a e2 e1
¼ x1 r12 cot bb cos2 a þ þ þ :
ds r1 r1 cos bb 2r12 cos bb r1
ð8:94Þ
The loss of instantaneous efficiency corresponding to the mating tooth pair at the
time instant considered is therefore given by the following relationship:
dWl l r1 1 ðe1 =r1 Þ2 þ ðe2 =r1 Þ2
1 gi ¼ ¼ 1þ 2 2 ; ð8:95Þ
e þe
dWm cos bb r 2 2 e1 þ
r1
e2
r1 þ coslb 12r2 2 þ tan a e2 re
1
1
b 1
1 1 e21 þ e22
1 gi ¼ pl1 þ ; ð8:97Þ
z1 z2 e1 þ e2
where l1 ¼ l= cos bb .
Since in order to obtain the formula of Poncelet we have assumed ðe1 þ e2 Þ ¼ 1,
from the comparison of the Eqs. (3.105) and (8.97) we infer that the loss of in-
stantaneous efficiency of the parallel cylindrical helical gear pair considered above
is equal to the loss of average efficiency during the meshing cycle of the parallel
cylindrical spur gear pair, multiplied by ð1= cos bb Þ. This loss of instantaneous
efficiency does not vary for instantaneous lines of contact between EA
and E
A0 ,
i.e. for a significant portion of the meshing cycle between two mating teeth. It
follows that the loss of the average efficiency of a parallel cylindrical helical gear
pair can be considered almost equal to the loss of the instantaneous efficiency.
Therefore, the Eq. (8.97) in which gi is replaced by g, also expresses the loss of
average efficiency.
The main methods employed for the manufacture of cylindrical helical gear wheels
can be also summed up in two categories: form cutting methods, and generation
cutting methods (see Galassini [11], Rossi [32], Townsend [36], Davis [7], Radzevich
[29, 31], Kawasaki et al. [21]).
The form cutting methods are the easiest cutting methods, which are done with
milling cutting machines, using form milling cutters, such as disk-type milling
cutters or end mills. Figure 8.18 shows the different arrangement of these two types
of cutting tool when they are used for cutting the teeth of a cylindrical helical gear
wheel.
As we already pointed out in Sect. 3.11.1, concerning the cutting of cylindrical
spur gear wheels with form cutting methods, even in this case the disk-type milling
cutter allows to obtain only approximate tooth flank profiles. Instead, if end mills
are used, having the same profile as the tooth space in the normal section of the
tooth, at least theoretically, the exact tooth flank profile can be obtained. In practice,
however, major difficulties occur, because these types of cutting tools are difficult to
construct and wear out quickly. Furthermore, the subsequent sharpening changes
appreciably their initial profile. For these reasons, the accuracy of the gears obtained
with this cutting process is extremely poor.
342 8 Cylindrical Involute Helical Gears
With regard to the cutting process, whatever the type of cutting tool (disk-type
milling cutter or end mill), it must have a relative helical motion, that is a screw
motion relative to the gear-blank to be cut, defined by the helix angle, b, of the
reference helix of the helical gear wheel to be manufactured. As shown in Fig. 8.18,
this screw motion is obtained by tilting the cutting tool at an angle equal to b with
respect to the gear wheel axis. It is then moved in the direction of this axis with a
velocity, c, while the gear wheel rotates with tangential velocity at the reference
circle, v. These three quantities are correlated by the following relationship:
c ¼ v cot b: ð8:98Þ
The generation cutting methods used in the production process of cylindrical helical
gear wheels are those already described for cylindrical spur gears (see Sect. 3.11.2),
with some extra peculiarities, which are typical of these gears.
The cutting of cylindrical helical gear wheels by means of gear shaping
machines using rack-type cutters as cutting tools is quite similar to that already
described for cylindrical spur gear wheels. The only difference is that the same
rack-type cutter used to cut spur gear wheels is now tilted at an angle equal to b
with respect to the gear wheel axis (here also b is the helix angle of the reference
8.11 Short Notes on Cutting Methods of Cylindrical Helical Gears 343
helix of the helical gear wheel to be cut), and its reciprocating cutting motion takes
place along this direction. The relative motions between the gear wheel being cut
and the cutting tool are those already described for this cutting process in
Sect. 3.11.2. These relative motions simulate the pure rolling without sliding of the
pitch cylinder of the gear wheel on the pitch plane of the rack-type cutter. In the
normal section of the teeth so obtained, they have the sizing of the rack-type cutter.
The cutting of the cylindrical helical gear wheels by means of gear shaping
machines using pinion-type cutters as cutting tools is a bit different from the one
already described for cylindrical spur gear wheels (see Galassini [11], Rossi [32] ). In
fact, the pinion-type cutter is a helical pinion cutter, with helix angle equal to that of
the gear wheel to be cut, both for external gear wheels and for internal gear wheels.
In addition, the reciprocating cutting motion of the pinion cutter along the direction
of the gear wheel axis does not occur parallel to this axis, but according to a helix,
having the same helix angle of the gear wheel to be cut, and axis parallel to that of the
latter. This reciprocating helical cutting motion is obtained by means of suitable
helical guides, having lead angle equal to the reference lead angle of the gear wheel.
For helix angles greater than 30°, cutting difficulties may occur due to inappropriate
relief angles. Of course, for a given module and a given helix angle, two cutting tools
are needed, one for right-hand helix, and the other for left-hand helix.
As we have already mentioned in Sect. 3.11.2, the most commonly used gen-
eration method for cutting cylindrical spur and helical gear wheels is by hobbing.
This cutting method uses gear hobbing machines equipped with hobs whose
geometry is the one shown in Fig. 3.16. However, to obtain cylindrical helical gear
wheels, the hob is arranged with its axis inclined with respect to the axis of the gear
wheel to be cut at an appropriate angle, w. This angle is determined in such a way
that the reference helices of the hob and gear wheel to be cut have the same
common tangent at the point where the two reference cylinders of hob and gear
wheel touch each other. As Fig. 8.19 shows, this angle, w, is given by:
w ¼ c c0 ; ð8:99Þ
where c and c0 are respectively the reference lead angles of gear wheel to be cut and
hob. The plus sign or the minus sign in Eq. (8.99) is to be taken, depending on
whether helical gear wheel and hob have both right-hand teeth (or both left-hand
teeth) or the helical gear wheel has right-hand teeth and the hob has left-hand teeth
and vice versa.
To obtain a helix angle equal to b ¼ ½ðp=2Þ c, the gear wheel to be cut must
have, in addition to the main rotation about its axis, an additional rotation, pro-
portional to the advancement of the hob in the direction of the gear wheel axis (see
Figs. 3.17 and 3.18). The main rotation above, which corresponds to the trans-
mission ratio with the hob, is the only rotation of the gear wheel in the case of a
cylindrical spur gear wheel. Therefore, if c is the feed rate of hob in the direction of
the gear wheel axis, and v is the tangential velocity at the reference circle of the gear
wheel, corresponding to the additional rotation above, the Eq. (8.98) should be
satisfied. Usually, in the gear hobbing machines, the two main and additional
344 8 Cylindrical Involute Helical Gears
(a) (b)
Fig. 8.19 Hob arrangement with respect to the helical gear wheel to be cut: a right-hand hob and
left-hand gear wheel; b both members with right-hand teeth
From the geometric point of view, the cutting of cylindrical double-helical gear
wheel is identical to that of cylindrical helical gear wheels. Therefore, it can be
performed with the same gear cutting machines as described in the previous section.
In this case, we can cut separately the two halves which make up the double-helical
gear wheel, and then connect them as Fig. 8.16b–d show, or we can cut their teeth
8.12 Short Notes on Cutting Methods of Double-Helical Gears 345
on a monobloc blank, leaving a central groove in order to relief of the cutting tools,
as Fig. 8.16a shows.
There are special gear cutting machines for double-helical gear wheels, which
allow obtaining continuous arrow teeth, with consequent advantages in terms of
their strength and loading carrying capacity. This tooth continuity can be achieved
by means of a form generating process or by a generating cut process.
The form cutting methods of double-helical gear wheels are similar to those
described in Sect. 8.11.1 for helical gear wheels. End mills are used as cutting tools.
These are made to move in the direction of the gear wheel axis, with speed, c,
which is still correlated to the tangential velocity, v, at the reference pitch cylinder
of the gear wheel being cut, by the Eq. (8.98). However, when the cutting tool
reaches the mid-plane of the gear wheel, this tangential velocity reverses its
direction.
With this operating way of the gear cutting machines, at the tooth arrow point a
tooth space is formed, whose chordal dimension, l, is less than the chordal tooth
thickness, L, in the transverse mid-plane of the gear wheel. It is to remember that
the chordal tooth space and the chordal tooth thickness are the distances between
symmetrical points on two opposite flanks of the same tooth space, and respectively
on opposite flanks of the same tooth, both measured along a chord tangent to the
pitch circle. The size l of this chordal tooth space should therefore be increased to
the length L. This increase can be obtained by rounding the convex edge, A, or the
concave edge, B, as shown in Fig. 8.20.
In relation to the other peculiarities of these double-helical gears, which can be
obtained with these form cutting processes (for example, the reinforcement disks at
their end faces), we refer back to what we have already said in Sect. 8.9.
For the cutting of cylindrical double-helical gear wheels, gear cutting machines are
used, which update the same generating cutting methods already described in
Sect. 8.11.2 for cutting cylindrical helical gear wheels. Of course, these machines,
having to cut simultaneously the two halves of the same double-helical gear wheel,
have slightly different features than those used for cutting helical gear wheels. Here
we just briefly describe these peculiarities.
Two types of gear shaping machines are used: one type uses two rack-type
cutters, while the other type uses two pinion-type cutters. In both types of cutting
machines, the two cutting tools are mounted on the same slide, and are disposed in
opposition to each other. One of the two cutting tools performs the forward stroke
346 8 Cylindrical Involute Helical Gears
by cutting one of the two halves of toothing (e.g., half with right-hand helix), while
the other cutting tool perform the return stroke, which is the idle stroke. The
forward stroke of the first cutting tool is stopped exactly at the transverse mid-plane
of the gear wheel being cut. At this point, the motion is reversed and the two cutting
tools exchange their functions, in the sense that the cutting tool that first performed
the idle stroke now makes the forward stroke and vice versa; so also the other half
of the helical toothing is cut. The tooth is shaped like an arrow with a sharp edge.
Usually the helix angle in the cutting conditions is assumed equal to 30°.
The gear hobbing machines also use two hobs, the way in which they operate
has many similarities to that of the gear shaping machines described above. In this
case, however, the two halves of the double-helical toothing are not symmetrical
with respect to the mid-plane of the gear wheel, but are offset by a half circular
pitch. In this way, at this mid-plane, the tooth thickness of one-half of the toothing
faces the tooth space of the other half of the toothing.
This offset by a half circular pitch results in a very smooth and silent operation.
The helix angle in the cutting conditions is usually equal to 23°. It should be noted,
however, that during the forward strokes, the axes of the two hobs reach the
mid-plane of the gear wheel being cut, so they remove a small portion of the tooth
end faces of the other half toothing of the gear wheel.
Finally, among the generating cutting methods of cylindrical herringbone gear
wheels, it is worth mentioning a special method, which allows to cut arrow teeth.
The gear wheels obtained with this cutting method are not double-helical gear
wheels in the strict meaning of the word, but gear wheels with arrow teeth.
The gear cutting machines that implement this cutting method use a special
mechanism, conceived by Böttcher [1] and essentially referring to a planetary gear
pair (see Chap. 13). It is made up of a rotating sun gear, whose reference pitch
circle has a radius ðr0 eÞ, and a fixed annulus having a reference pitch circle with
radius r0 . The sun gear is mounted on the pin of a crank, rotating about its fixed
8.12 Short Notes on Cutting Methods of Double-Helical Gears 347
axis, so the axis of the sun gear is mobile and describes a cylinder with radius,
e. Instead, a point P rigidly connected to the sun gear, and placed to the outside of
its reference pitch circle, described an extended hypocycloid. In case r0 ¼ 4e, this
extended hypocycloid becomes a closed curve, namely a closed curvilinear
quadrilateral like that shown in Fig. 8.21.
If now a trapezoidal cutter tool is placed in place of the point P, it will describe
the tooth of a virtual generation rack, which will have in the plan view the shape of
the extended closed hypocycloid defined above. If the gear wheel to be cut is
equipped with suitable translation and rotation motions, corresponding to the pure
rolling without sliding of its pitch cylinder on the pitch plane of the virtual gen-
eration rack, one tooth space between two adjacent teeth of the gear wheel blank is
generated. In order to generate, with a continuous cutting process, all teeth of the
gear wheel, the latter must have, in addition to the two afore-mentioned motions, a
further motion corresponding to a rotation equal to a circular pitch per revolution of
the sun gear, which has the toolholder wheel function.
In this way, however, the continuous curve described by the tooth of the virtual
rack cutter is slightly asymmetrical with respect to the mid-plane of the gear wheel,
as shown in Fig. 8.22. On the other hand, we have the great advantage of cutting,
with a continuous process, all the teeth of the gear wheel, with all the related
positive consequences in terms of the accuracy grade of the gear wheel thus
obtained. In practice, three cutting tools work simultaneously: one of these gashes
the tooth spaces, while the other two finish the concave and convex flanks of the
teeth (see also Schiebel [33] ).
References
Abstract In this chapter, the fundamentals of straight bevel geometry are given
and the related main quantities are defined. Geometry of spherical involute and
octoid toothing is introduced and implications on the cutting processes of these
gears are analyzed. Tredgold approximation is then used to define the equivalent
cylindrical gears and their main characteristic quantities, including the minimum
number of teeth to avoid interference. Short notes are given on the reference profile
modifications. The problem of straight bevel gears with profile-shifted toothing and
variation of shaft angle is therefore dealt with in detail in an analytical way, also
considering the effects of transverse tooth thickness modifications. The load anal-
ysis of these gears is then performed and the thrust characteristics on shafts and
bearings are defined. Subsequently, the main kinematic quantities related to the
rolling and sliding motions of the mating teeth surfaces are defined and the
instantaneous and average efficiencies are analytically determined. Finally, short
notes on the main cutting processes of these types of gears are given and sugges-
tions for their construction and assembly are provided.
9.1 Introduction
Straight, helical or skew, and spiral bevel gears are most commonly used for geared
power transmission systems between two intersecting axes. In this chapter, we
focus our attention on the straight bevel gears. Helical or skew, and spiral bevel
gears, together with hypoid gears, will discussed in the Chap. 12. Straight bevel
gears, helical and spiral bevel gears, and hypoid gears can be used for both
speed-reducing and speed-increasing gear drives. Their gear ratio can be as low as
1, but it should not exceed about 10, although in machine tool design, where
precision hypoid gears are required, most high values of the gear ratio, in the range
between 10 and 20, have been required. However, for usual speed-increasing gear
drives, it is appropriate that this gear ratio does not exceed 5 (see Buckingham [3],
Merritt [23], Henriot [15], Niemann and Winter [25], Maitra [22], Radzevich [31]).
Basically, a straight bevel gear pair (and so also helical and spiral bevel gear
pairs) is similar to a friction drive consisting of two conical rollers in which the
conical surface of one roller drives that of the other roller by friction (see Pollone
[27], Brar and Bansal [1], Klebanov and Groper [18]). Figure 9.1 shows two pairs
of friction conical rollers, the first of which form an external friction drive
(Fig. 9.1a) and the second an internal friction drive (Fig. 9.1b). In both drives, the
two friction wheels of each pair are touching each other along the common gen-
eratrix passing through the apex or cone center, O, where the axes of the two cones
intersect. Depending on whether the friction conical pair is external or internal, the
shaft angle R between the axes is equal to the sum or, respectively, the difference,
of the pitch cone angles or simply pitch angles, d1 and d2 , of the pinion and wheel,
that is:
R ¼ d1 þ d2 : ð9:1Þ
R ¼ d2 d1 : ð9:2Þ
Notoriously, when the two friction conical rollers of a pair are pressed against
each other, they may transmit, without sliding, maximum tangential forces equal to
the friction forces, which occur in the contact. However, the tangential friction
forces that can be transmitted with these pairs of friction conical rollers are not high.
Therefore, when higher tangential forces must be transmitted, the two conical
rollers of the kinematic pair are equipped with teeth, the shape of which must be
such the teeth of one member engage, during the rotation, with the teeth of the other
member, so transmitting the motion with the same kinematics of the two afore-
mentioned friction conical rollers. In this simple way, we pass from a friction
conical pair to a bevel gear pair.
The two circular cones of the two fictitious friction wheels are the axodes, i.e. the
operating pitch cones of the two members of the bevel gear pair [21]. These pitch
(a) (b)
Fig. 9.1 Friction cone wheel pairs: a external pair; b internal pair
9.1 Introduction 353
cones are in contact along a common generatrix and meet at a common point, the
cone center or cone apex, O. This is also the crossing point Oc , i.e. the point of
intersection of axes of two bevel gear members and their shafts. Therefore, crossing
point and cone apex coincide ðO Oc Þ. The axodes are the loci of instantaneous
axes of rotation in the movable coordinate systems Oi ðxi ; yi ; zi Þ, which are rigidly
connected to rotating gear wheel i, with i = (1,2). Assuming that one of the two
members of the gear pair is fixed and the other movable, during their relative
motion the movable cone (also called movable axode or polodia axode) rolls
without sliding on the fixed cone (also called fixed axode or herpolodia axode). For
further details on this subject, see Ferrari and Romiti [10].
The intersections of these pitch cones (the herpolodia axode and polodia axode)
with a spherical surface having as its center the crossing point O Oc , are the
herpolodia and polodia, i.e. the pitch circles of the two gear wheels. In other words,
herpolodia and polodia are the directrices of two herpolodia axode and polodia
axode, which are cones having a common apex at point O Oc . Since all toothing
data of a bevel gear are given with reference to the heel (or large end), the pitch
circles of the two members of the gear pair are those corresponding to the large end
of the two cones, while AB and BC are the pitch diameters of the two gear wheels
(Fig. 9.1).
In their rolling motion without sliding, the pitch cones have spherical motion, for
which every point of a bevel gear remains at a constant distance from the crossing
point during the motion. The two circles having diameters AB and BC, and then the
two spherical segments delimited by them, roll against each other without sliding,
remaining on the sphere. In order for the motion can take place in this way, the two
wheels having the shape of spherical segments must be equipped with teeth profiled
so that the maximum circle, normal to the profiles at point of contact, passes
through the point of contact between the pitch circles. In order to shape the teeth
profiles, only the spherical involutes are generally taken into consideration among
all the curves that satisfy this condition. These are the curves generated by a point
of a maximum circle when it rolls without sliding on a smaller circle, always
keeping on the sphere [14, 27, 36].
The maximum wheel with the shape of a spherical segment is a hemispherical
wheel, having as pitch circle a maximum circle. This hemispherical wheel is the
so-called crown wheel or crown gear, i.e. a bevel gear with a reference cone angle
of 90°. For the reasons to be described in the following sections, in which we will
deepen well the generation of the spherical involute curves, it is not customary to
give the spherical shape to the heel, that is the large end of a bevel gear. Instead, this
heel is usually conical, with generatrices that are tangent to the theoretical sphere at
the pitch circle diameter: this is the so-called back cone.
By doing so, the wheels having the shape of spherical segments, and face width
equal to the width over the toothed part measured along a generatrix of their
reference cones, are in fact replaced with the bevel gears tangent to the sphere in
correspondence with the pitch circles of the same wheels. These wheels, like
Fig. 9.1 shows, belong to back cones, i.e. the cones AO1 B and BO2 C, having
apexes on the wheel axes and generatrices normal to those of the pitch cones.
354 9 Straight Bevel Gears
Of course, this is an approximate way to analyze the straight bevel gears; it gives
very reliable results, and is known as Tredgold approximation, as Tredgold con-
ceived it (see Buchanan [2]).
The back cone corresponding to the crown wheel, and therefor tangent to the
pitch circle of the maximum wheel having the shape of a spherical segment (the
hemispherical wheel introduced above), degenerates into a cylinder. Since the teeth
have a limited depth, on this cylinder we can replace the profile in the shape of
spherical involute, with (see Giovannozzi [14], Pollone [27]):
• the straight profile of the standard basic rack, for straight bevel gears;
• the cylindrical helical profile, having with respect to the pitch plane the same
inclination of the maximum osculating circle to the spherical involute, for he-
lical bevel gears.
The back cones are developable surfaces. Therefore, they permit to bring the study
of the tooth profiles of the bevel gears to the one of the teeth of equivalent
cylindrical gears. Developing on a plane the cylinder in which the back cone of the
crown wheel degenerated, we get a rack having straight profile flanks and pressure
angle a. Then the crown wheel bears the same correlation to a bevel gear as a rack
does to a cylindrical spur gear. Developing instead, always on a plane, the back
cones of the two members of the bevel gear (Fig. 9.1), we get two cylindrical gear
wheels having radii O1 B and O2 B, respectively.
Figure 9.2 shows two typical external bevel gears, each consisting of two bevel
wheels which, in order to be able to operate with correct kinematic conditions, i.e.
pure rolling without sliding between the operating pitch cones, must have the apex
of the pitch cones in common. The same figure shows the most significant cones,
and the main geometric entities and quantities characterizing the wheels of the bevel
gear pairs and their toothings, namely:
• Pitch cones, i.e. the pitch surfaces of both wheels of the bevel gear pair, and
pitch angles d1 and d2 , i.e. the angles between the axes and pitch cone gener-
atrices, which intersect at the pitch cone apex, coinciding with the crossing
point, O. As in the case of cylindrical gears, here too we must be aware that,
without a specific qualification, we mean the operating (or working) pitch cones.
When necessary, to avoid misunderstanding, we add the specific qualification,
so we will have the reference pitch cones or simply reference cones, and op-
erating pitch cones. This notation also applies, unchanged, for many of the
geometric entities and quantities defined below.
• Back cones, i.e. the cones at the outer end (the heel of the face width, whose
generatrices are perpendicular to those of the reference cones, and back cone
angles, i.e. the acute angles between the axes and generatrices of the back cones.
9.2 Geometry and Characteristics of the Bevel Gears 355
(a) (b)
Fig. 9.2 External straight bevel gears: a shaft angle R [ 90 ; b shaft angle R ¼ 90
• Tip cones, i.e. the tip surfaces of both wheels of the bevel gear pair, and face
angles or tip angles, da1 and da2 , i.e. the angles between the axes and tip cone
generatrices.
• Root cones, i.e. the root surfaces of both wheels of the bevel gear pair, and root
angles df 1 and df 2 , i.e. the angles between the axes and root cone generatrices.
• Inner cones (not shown in Fig. 9.2), i.e. the cones at the inner end (the toe) of
the face width, whose generatrices are perpendicular to those of the pitch cones.
• Mean cones (not shown in Fig. 9.2), i.e. the cones at the mean point (i.e. the
mean point of the face width), whose generatrices are perpendicular to those of
the pitch cones.
• Face width, b, i.e. the width of the toothed part of a gear wheel, measured along
a generatrix of its reference cone.
• Cone distance, i.e. the distance from the cone apex to the specified cone (then
we have: mean cone distance, Rm ; outer cone distance, Re ; inner cone distance,
Ri ; etc.), measured along a reference cone generatrix; CO1 and CO2 in Fig. 9.2a
are the back-cone distances.
• Locating face, i.e. the plane face perpendicular to the axis of the gear wheel to
be cut, by which its axial position is determined.
• Mounting distance, H, i.e. the axial distance from the locating face to the pitch
cone apex.
• Crown to crossing point, tx0 , i.e. the distance along the gear axis from the crown
point to the pitch cone apex.
• Tip distance, equal to the difference ðH tx0 Þ; i.e. the distance along the gear
axis from the locating face to the plane containing the tip circle.
356 9 Straight Bevel Gears
• Pitch circle, i.e. the circle of intersection of the pitch cone with a plane per-
pendicular to the axis, on which the pitch has a specified value (usually, the
pitch circle is the circle of intersection of the pitch cone with the back cone).
• Outer pitch diameter, de ; i.e. the diameter of the pitch circle.
• Mean pitch circle, i.e. the circle of intersection of the pitch cone with the mean
cone.
• Mean pitch diameter, dm ; i.e. the diameter of the mean pitch circle.
• Tip circle, i.e. the circle of intersection of the tip cone with the back cone.
• Outside diameter, dae , i.e. the diameter of the outside circle.
• Crown point, i.e. the point of intersection of the generatrix of the tip cone and
back cone.
• Root circle, i.e. the circle of intersection of the root cone with the back cone.
• Root diameter, dfe , i.e. the diameter of the root circle.
• Shaft angle, R, i.e. the angle equal to the sum of the pitch angles d1 and d2 , for
an external bevel gear pair, or the angle equal to the difference of the pitch
angles d2 and d1 , for an internal bevel gear pair.
• Tooth depth, h, i.e. the distance between the tip circle and root circle, measured
along a back cone generatrix.
• Addendum, ha , i.e. the distance between the tip circle and pitch circle, measured
along a back cone generatrix.
• Addendum angle, #a ¼ ðda dÞ, i.e. the difference between the face angle and
pitch angle.
• Dedendum, hf , i.e. the distance between the pitch circle and root circle, mea-
sured along a back cone generatrix.
• Dedendum angle, #f ¼ d df , i.e. the difference between the pitch angle and
root angle.
Usually, the shaft angle R, which depends on the drive conditions, is equal to
90° (Fig. 9.2b), but it can assume value different from 90°. Figure 9.3 shows the
range of the possible bevel gear pairs formed by a pinion, whose operating pitch
angle is d1 , and several wheels, characterized by different operating pitch angles d2 ,
namely:
• Shaft angle Ra \90 , with bevel gear pair formed by the pinion 1 and the wheel
2a.
• Shaft angle Rb ¼ 90 , with bevel gear pair formed by the pinion 1 and the wheel
2b.
• Shaft angle Rc [ 90 , with bevel gear pair formed by the pinion 1 and the crown
wheel 2c.
• Shaft angle Rd [ 90 , with bevel gear pair formed by the pinion 1 and the
internal bevel wheel 2d.
It is first to be noted that, generally, the operating pitch angles, for the reasons that
gradually we will specify better in the following sections, coincide with the ref-
erence pitch angles. It is then to be noted that the wheel 2c has an operating pitch
angle of 90°: it is the crown wheel, i.e. a wheel with operating pitch angle of 90°,
9.2 Geometry and Characteristics of the Bevel Gears 357
Fig. 9.3 Possible bevel gear pairs formed by a given pinion, and several wheels
and with tip and root angles respectively greater and smaller than 90°. This crown
wheel can be an actual wheel, but can also be thought of as the common bevel
wheel, which meshes properly with all the other wheels. In other words, it con-
stitutes the equivalent of the standard basic rack of the parallel cylindrical spur and
helical gears; it is of fundamental importance in the study of the bevel gears and
their behavior. Therefore, this crown wheel, as the basic rack for cylindrical gears,
is assumed as the generation wheel of all other bevel wheels.
It should finally be noted that an internal bevel gear pair, consisting of a pinion
and an internal bevel wheel (2d in Fig. 9.3), although possible from the theoretical
point of view, is to be considered an exceptional case that, in fact, does not have
practical applications, given the difficulties that we have in making an internal bevel
toothing. Therefore, in the following discussion, we consider only external bevel
gear pairs. The relationships that correlate R, d1 , d2 , z1 and z2 (or u ¼ z2 =z1 ), in the
different cases of straight bevel gears shown in Fig. 9.3, are summarized in
Table 9.1.
In cases where tip and root cone generatrices converge in the pitch cone apex
(see Sect. 9.5), the teeth of a bevel wheel taper off as they approach the apex. Thus,
the tooth size decreases from the heel towards the toe, and the tooth profile also
varies accordingly. For this reason, the quantities of interest, such as the addendum,
dedendum, tooth depth, module, pitch circle diameter, tip circle diameter, etc., are
358 9 Straight Bevel Gears
Table 9.1 Reference cone angles d1 and d2 for different straight bevel gear pairs
R d2 d1
R\90 tan d2 ¼ ðz1 =z2sin R
Þ þ cos R
d1 ¼ R d2
R ¼ 90 tan d2 ¼ z2
z1 ¼u d1 ¼ p2 d2
R [ 90 tan d2 ¼ ðz1 =zsin ðpRÞ d1 ¼ R d2
2 ÞcosðpRÞ
all conventionally measured with reference to the heel of the tooth. Therefore, they
are those relating to the back cone and the corresponding transverse section of the
toothing. In some cases (see again Sect. 9.5), these quantities are given with ref-
erence to the normal section bisecting the tooth face width. With reference to
Fig. 9.2b, we have:
h ¼ ha þ hf ð9:5Þ
da ¼ d þ # a ð9:6Þ
df ¼ d # f ð9:7Þ
Toothing of a bevel gear is sized using the same sizing of the parallel cylindrical
spur gears. The nominal sizing most widely used is the following:
ha ¼ m hf ¼ 54 m h ¼ ha þ hf ¼ 94 m: ð9:11Þ
In order not to have too different profiles of the teeth, due to the taper of the tooth
from heel to toe, the ratio ðb=Re Þ is limited within the range:
Consider any external bevel gear pair characterized by a given shaft angle R, and
denote by:
• z1 and z2 ; the numbers of teeth of the pinion and wheel;
• x1 and x2 (Fig. 9.4a), the absolute values of angular velocities of the two
members of the bevel gear pair, which generally are to be considered as vectors
(note that vectors are indicated with bold letters);
• x ¼ ðx1 x2 Þ; the relative angular velocity, which is a vector having, as
straight line of application, the common generatrix along which the operating
pitch cones of the two members of the bevel gear pair touch (Fig. 9.4b);
• d1 and d2 ; the operating pitch cone angles that, as a rule, as we have already
noted, are at the same time the reference pitch cone angles or, what is the same,
the generation pitch cone angles (Fig. 9.4a);
• d1 ¼ 2r1 and d2 ¼ 2r2 ; the diameters of the two operating pitch circles, which
are obtained as the intersections of the operating pitch cones with the back cones
(note that the two back cones have a common generatrix passing through point
C, as Fig. 9.2a shows, and that the operating pitch cone and back cone of each
of the two members of a bevel gear pair have the same axis);
• Re , Ri and Rm ; the cone distances of the back, inner and mean cones from the
cone apex, measured along a reference cone generatrix (Fig. 9.2b).
(a) (b)
Fig. 9.4 Kinematic diagram of an external bevel gear pair, angular velocities and relative angular
velocity vectors
360 9 Straight Bevel Gears
With reference to Fig. 9.4, which shows the kinematic diagram of an external bevel
gear pair, as well as the relative angular velocity vector x ¼ ðx1 x2 Þ, obtained as
the difference of the vectors of the two angular velocities, x1 and x2 , and to
Fig. 9.2, we get the following relationships, which are valid for zero-profile-shifted
bevel gears, and bevel gears with profile-shifted toothing without variation of shaft
angle (see Sect. 9.8):
R ¼ d1 þ d2 ð9:14Þ
z2 d2 x1 sin d2
u¼ ¼ ¼ ¼ ð9:15Þ
z1 d1 x2 sin d1
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
x ¼ x21 þ x22 þ 2x1 x2 cos R ð9:16Þ
x1 x2 x x
¼ ¼ ¼ ð9:17Þ
sin d2 sin d1 sin½p ðd1 þ d2 Þ sin R
x2 sin R z1 sin R
sin d1 ¼ sin R ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð9:18Þ
x 1 þ u2 þ 2u cos R z1 þ z22 þ 2z1 z2 cos R
2
x1 u sin R z2 sin R
sin d2 ¼ sin R ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi ð9:19Þ
x 1 þ u þ 2u cos R
2 z1 þ z22 þ 2z1 z2 cos R
2
sin R 1
tan d1 ¼ ðfor R ¼ 90 ; tan d1 ¼ Þ ð9:20Þ
u þ cos R u
u sin R
tan d2 ¼ ðfor R ¼ 90 ; tan d2 ¼ uÞ ð9:21Þ
1 þ u cos R
1 1 pffiffiffiffiffiffiffiffiffiffiffiffiffi
¼ ¼ u2 þ 1 for R ¼ 90 ð9:22Þ
cos d2 sin d1
d1 d2
Re ¼ ¼ ð9:23Þ
2 sin d1 2 sin d2
b
Ri ¼ Re b ¼ Rm ð9:24Þ
2
b
Rm ¼ Re ð9:25Þ
2
pd1 2pr1 pd2 2pr2
p¼ ¼ ¼ ¼ ð9:26Þ
z1 z1 z2 z2
d1 2r1 d2 2r2
m¼ ¼ ¼ ¼ ð9:27Þ
z1 z1 z2 z2
9.3 Main Equations of Straight Bevel Gears 361
In the above relationships, the meaning of the symbols already defined while
remaining unchanged, dm1 and dm2 are the diameters of the reference mean pitch
circles of the pinion and the wheel, which are used for strength calculations.
We have also the following relationships:
z1 z2
¼ ð9:34Þ
sin d1 sin d2
ha1 sin d1
tan #a1 ¼ 2 ð9:35Þ
mz1
hf 1 sin d1
tan #f 1 ¼ 2 ð9:36Þ
mz1
ha2 sin d2
tan #a2 ¼ 2 ð#a2 ¼ #f 1 Þ ð9:37Þ
mz2
hf 2 sin d2
tan #f 2 ¼ 2 ð#f 2 ¼ #a1 Þ: ð9:38Þ
mz2
In the introduction (see Sect. 9.1) we have already seen that the motion of the two
pitch cones, which roll on each other without sliding, is a spherical motion, for
which, during this motion, every point of a bevel gear remains at a constant distance
from the apex. Therefore, any point of the tooth flank of a bevel gear wheel moves
362 9 Straight Bevel Gears
on a sphere, whose center is the crossing point, O ¼ Oc ; where the axes meet, and
whose radius is the distance of this point from the cone apex. Of course, this point
describes, on the sphere surface, a curve that is not a planar curve, but a
three-dimensional curve. The profile of the tooth is obtained as the intersection of
the tooth flank of the bevel gear wheel with the surface of this sphere (see
Giovannozzi [14], Ferrari and Romiti [10], Pollone [27], Radzevich [30]).
We first examine bevel gears having spherical involute toothing. In this regard,
without less of generality, we consider a sphere of unit radius and center O, and two
cones (the base cones) that do not touch each other, with base cone angles db1 and
db2 and having their apices coinciding with point O (Fig. 9.5). These two cones
intersect the surface of the sphere according two circles, indicated with c1 and c2 in
Fig. 9.5. Consider then a plane passing through the point O and tangent to both
base cones; it intersects the sphere according to the maximum circle cmax , that is
having maximum diameter.
Let us now roll this maximum circle cmax with a uniform angular velocity x
about an axis passing through the point O and perpendicular to the plane that
contains it. We also suppose that, during this motion, the circle cmax drags by
friction, without sliding, the two not-maximum circles c1 and c2 , as well as the base
cones having base cone angles db1 and db2 , with angular velocities x1 and x2 about
their axes. Under the hypothesis that this drag motion between the three circles
cmax , c1 and c2 is a pure rolling motion, without sliding, their tangential velocities
must be equal; therefore, since the sphere has a unit radius, we can write the
following equation:
x ¼ x1 sin db1 ¼ x2 sin db2 : ð9:39Þ
However, the following relationship can be written for the operating pitch cones
to be find, which in Fig. 9.5 are tangent along the straight line OC:
Consider now the plane containing the axes of the two cones. It intersects the
sphere according to another circle having a maximum diameter, which passes
through the intersection points, V1 and V2 ; of the axes of the cones with the sphere,
and intersects at point C the other circle of maximum diameter, cmax . We denote by
½ðp=2Þ a the angle at point C formed, on the sphere, between the two circles of
maximum diameters, and between the planes that contain them. Then we denote by
a the pressure angle, i.e. the angle between the circle cmax and the common tangent
t t at point C to the two pitch circles, which is perpendicular to the plane passing
through points V1 , C, and V2 . Finally, we denote by T1 and T2 the points were the
circle cmax touches respectively the circles c1 and c2 . Applying the law of sines of
spherical trigonometry to spherical triangles CT1 V1 and CT2 V2 , which are respec-
tively rectangles at points T1 and T2 , we obtain the relationships:
Let us now roll without sliding the plane containing the circle cmax on the base
cones. During this rolling motion, a point of the circle cmax between T1 and T2
describes, on the sphere, two spherical involutes, tangent to each other, and therefore
conjugate, which are the profiles of the teeth, i.e. the intersections of the tooth flanks
with the sphere. Such spherical involutes touch at a point of the circle cmax , which
therefore is, on the sphere, the curve of action, while the arc T1 T2 is the arc of action.
If we join any point of contact on the circle cmax with the center O of the sphere,
we get the instantaneous line of contact, which, like the circle cmax , rotates, about
the axis of the latter with angular velocity x. Therefore, the plane containing the
circle cmax , which is tangent to the base cones, is the locus of points of contact, i.e.
the surface of contact. During the rolling without sliding of the plane of contact on
the base cones of the two gear wheels, the instantaneous line of contact describes
and generates the flanks of the mating teeth.
364 9 Straight Bevel Gears
The intersections of these flanks with the countless spheres between the back
cone and inner cone, which are respectively tangent to the heel and toe of the tooth
(these spheres are indicated with Se and Si in Fig. 9.3), are spherical involute
curves, the geometry of which varies as a function of the radius of the sphere under
consideration. The tooth profiles obtained as intersections of the tooth flank with
each of these numberless spheres are conjugate profiles, and therefore ensure, from
a theoretical point of view, an appropriate kinematics. Their geometry, however,
varies from one to another sphere, and this has important implications, which call
into question the choice of the virtual generating bevel wheel.
In relation to the cutting process, it is first to remember that even the toothing of
the bevel gears can be broadly classified into two main groups: the one related to
catalogue gears, and the one for gears that are not catalogue gears, with toothing for
single gears (see Galassini [12], Rossi [33], Niemann and Winter [25],
Stadtfeld [39]).
In the first case, both members of the gear pair have in common the reference
mating crown wheel, i.e. the generation crown wheel, which is a virtual wheel that
simulates the motion of the cutting tool. In this regard, Fig. 9.6 shows the mate-
rialization of the profiling motion of two planer tools with straight flanks on the
corresponding crown wheel, having also straight flanks, for cutting of two straight
bevel wheels. These wheels, being generated by the same wheel, i.e. the virtual
generation wheel or simply the generation wheel, are able to mesh between them
with a proper kinematics. This applies to all the possible bevel wheels mating with
the generation crown wheel. Any two of the bevel wheels thus obtained are
Planer tool
Generating crown wheel
Fig. 9.6 Cutting of the straight bevel wheels with gear planer
9.4 Spherical Involute Toothing and Octoidal … 365
conjugate between them, and are therefore able to ensure an appropriate kinematics,
with a linear contact between the tooth flanks. In this way, we obtain bevel wheels
that have the characteristics of catalogue gears.
The generation crown wheel of a batch of bevel wheels produced with these
characteristics is uniquely defined (Fig. 9.3) by the following data:
• external cone distance, Re ¼ OC ¼ de =2 sin d;
• face width, b;
• shape of the tooth profile, geometrically defined by the reference crown wheel
tooth profile;
• geometric shape and direction of the tooth trace, that is the line of intersection of
the tooth flank with the reference surface (see Sect. 12.1);
• tip and root cone surfaces.
The various gear wheels thus obtained are characterized by different operating
pitch cone angles d, for which it is possible to derive gears for transmission drives
between axes with different shaft angles, R. The bevel gear pair is so univocally
defined once d, R and the reference crown wheel are known.
Instead, in the second case, the one of gears that are not catalogue gears, the
toothing is not determined with reference to a common generation crown wheel. The
straight bevel gear pair consists of a generated pinion and a non-generated wheel,
and the related cutting process with which the gear pair is obtained is known as
formate-cut process. The wheel of the gear pair is produced without generation
motion, with a form cutting process, by means of inserted-blade milling cutters that
are different depending on the cutting process involved, which generally use teeth
with straight profile flanks. The pinion is instead obtained with a generation motion,
by means of a profiling tool able to simulate the pure rolling of the operating pitch
cone of the same pinion. In this way, as Fig. 9.7 shows, the straight flanks of the
Fig. 9.7 Teeth profiles of a single gear pair (continuous line), and of an octoid gear pair (dashed line)
366 9 Straight Bevel Gears
teeth of the generation wheel generate the curved flanks of the pinion teeth. Thus, we
have a single gear pair, in the sense that only that pinion and only that wheel are able
to ensure an appropriate kinematics, if the assembly and actual operating conditions
strictly fulfill the cutting conditions (see Stadtfeld [39–41].
We return now to consider the catalogue bevel gears produced by a generation
process. In this process, the tooth space of the wheel to be cut is the envelope of the
successive positions assumed by the conjugate tooth of the generation wheel,
during the relative motion of rolling without sliding of the cutting pitch cone of the
generation wheel on the cutting (or operating) pitch cone of the to-be-generated
gear wheel. In this way, all the wheels derived from the same generation wheel are
conjugate between them. However, using such a technological process, it is not
possible to choose a generation wheel having any tooth profile, and any cutting
cone angle, d. This would be possible only in the case of gears obtained by a
hot-rolling process, e.g. using the Anderson gear rolling machine or similar
machines (see Galassini [12]).
It is necessary to consider that the tooth of the generation wheel tapers off as it
approaches the apex, and that its profile has variable size and shape along its entire
face width. Therefore, it is not possible to simulate this tooth, having variable
profile, replacing it with a planer tool with constant profile, equipped with cutting
motion that will make it run from the heel to toe along a guide inclined of an angle
d. Significant cutting errors and profile deviations would be. To solve this problem,
it is necessary to choose, as generation wheel, a particular bevel wheel whose teeth
do not change the shape of their profile from heel to toe. This wheel is the crown
wheel, which has a cone angle d ¼ 90 , and reference pitch cone that degenerates
into a plane (the pitch plane). However, it must have straight tooth profiles. On the
back cone of the crown wheel, as well as on the numberless cones included between
the back and inner cones of the same crown wheel, all degenerated into cylinders,
the spherical involute profiles vary from heel to toe, also changing their curvature
above and below the pitch plane. Figure 9.8a highlights these changes of curvature
as well as the inflection points at the pitch plane.
Rebus sic stantibus, the undoubted advantages of the spherical involute profiles
(straight path of contact; unique plane of contact; proper kinematics, also chancing
the relative position of the intersecting axes; possibility of using profile-shifted
toothing with variation of the shaft angle, to achieve predetermined design goals;
etc.) are, in fact, annihilated by the almost insoluble difficulties of the cutting
process. For this reason, the spherical involute toothings, although obtainable using
bevel gear planers, which implement a template machining process, but do not
guarantee a high cutting quality, have secondary importance (Niemann and
Winter [25]).
This secondary importance of the template machining process is because the
bevel gear cutting machines that implement it are not able to guarantee the perfect
alignment -of-:
9.4 Spherical Involute Toothing and Octoidal … 367
(a) (b)
(c) (d)
Fig. 9.8 Generation crown wheels with spherical involute tooth profiles (a), and with octoid tooth
profiles (b); tooth profile shapes of the two generation crown wheels (c), and the generated wheels
(d); paths of contact on the sphere (c)
• the feeler pin or tracer, whose tracing point follows the contour of the
three-dimensional master model or template, reproducing the spherical involute;
• the shaper cutting edge, following the path taken by the tracer to machine the
spherical involute shape;
• the center of the spherical motion, which coincides with point where the axes of
the bevel wheel being cut and virtual crown wheel that is simulated by the
shaping cutter intersect.
However, the bevel gears obtained with this template machining process, which
was used sporadically as long as the CAD/CAM systems become widespread for
machining complex three-dimensional shapes [26], such as spherical involutes, are
of poor accuracy grade because, due to the shaping cutter wear, the above alignment
could not be guaranteed. In practical reality, shaper wear determines significant
deviations of the profiles relative to the theoretical geometry of the spherical
involute. In addition, the inevitable pitch errors, due to the use of the indexing head
for the single cut of each tooth, are added to these deviations. It should be noted that
this template cutting process, even upgraded with CAD/CAM systems (these reduce
and even eliminate indexing errors), is always of secondary importance for the
manufacture of straight bevel gears, as it is not suitable for large series productions.
Anyway, the toothings that are actually produced are octoid toothings. They are
related to a generation crown wheel with trapezoidal teeth (Fig. 9.8b). In this case,
the tooth profile is a straight line, and corresponds to that of the planar tool
(Fig. 9.6). The latter, in its cutting motion, will be engaged to a depth decreasing
from the heel to the toe, in the case of tooth that tapers off as it approaches the apex,
and to a constant depth, in the case of tooth with constant depth (see next section).
According to Tredgold approximation, also briefly recalled in the next section, on
368 9 Straight Bevel Gears
Fig. 9.9 Reference octoid crown wheel and bevel gear pair with octoid toothing; path of contact
in the shape of eight on the sphere (the octoid)
the cylinder in which the back cone of the octoid generation crown wheel is
degenerated, we have a basic rack with straight-line tooth flanks. Then the gener-
ation of the octoid toothing corresponds, essentially, to that of the involute tooth
profiles of the cylindrical spur gear wheels. Substantially, in this generation cutting
method, a straight sided cutting tool, simulating the generation crown wheel, and
the cutting cone of the blank of the bevel wheel to be cut roll on each other
producing the desired bevel wheel.
From a conceptual point of view, the two generation cutting processes, the
hypothetical one that would use a crown wheel with spherical involute tooth pro-
files, and the one actually used, which employs a crown wheel with octoid tooth
profiles, are quite similar. The difference consists in the fact that, considering the
rolling on the sphere, in the first case the path of contact is a straight line, while in
the second case the path of contact deviates from the straightness, albeit slightly
(Fig. 9.8c). If we consider the two wheels of a bevel gear and the common gen-
eration octoid crown wheel, this path of contact develops, on the sphere, according
to a curve in the shape of eight (Fig. 9.9). The geometry of this curve is deter-
minable using the known methods of the mating profiles (see Levi-Civita and
Amaldi [20], Ferrari and Romiti [10], Scotto Lavina [36]).
Despite the deviation of the path of contact from the straight line, the kinematics
of operation of the bevel gears with octoid toothings is correct for zero
profile-shifted toothing and profile-shifted toothing without variation of shaft angle.
The cutting process of bevel gears with octoid toothing is very precise, and leads to
high quality gears, unimaginable with a template shaping process, with which we
would get wheels with spherical involute toothing.
For brief information on the cutting methods of straight bevel gears and gear
cutting machines that implement them, we refer the reader to Sect. 9.12.
9.5 Main Conical Surfaces and Equivalent Cylindrical Gear 369
Fig. 9.10 Bevel gear with zero profile-shifted toothing, and equivalent cylindrical gear at the back
cone
As Fig. 9.10 shows, various conical surfaces are identifiable in a bevel gear pair, the
main of which are:
• the operating pitch cones, among them tangent along the common generatrix
OC ¼ Re , and defined by the cone angles d1 and d2 ;
• the reference pitch cones, which coincide with the cutting pitch cones (the
operating pitch cones in cutting conditions), and are therefore the cones that, in
cutting conditions, roll without sliding on the pitch plane of the common gen-
eration crown wheel;
• the tip and root cones, defined by the tip and root angles da1;2 and df 1;2 ;
• the back, mean and inner cones, defined by the back-cone angles ½ðp=2Þ d1
and ½ðp=2Þ d2 .
It is to be noted that the back and mean cones are used respectively to define the
equivalent cylindrical gear and to set the bending strength calculations of the teeth.
It should also be noted that the operating pitch cones and reference pitch cones of a
bevel gear pair coincide in the two cases of zero profile-shifted toothing, and
profile-shifted toothing without variation of shaft angle. Of course, in these two
cases, also the corresponding cone angles coincide. The operating pitch cones and
370 9 Straight Bevel Gears
reference pitch cones differ from each other only in the case of profile-shifted
toothing with variation of shaft angle.
The apexes of the tip and root cones need not necessarily coincide with the point
of intersection with the wheel axis and, therefore, the tooth size is not always
decreasing from the heel to the toe. This happens in the case of so-called normal
toothings, which are those of many straight bevel gears, helical bevel gears, and
spiral bevel gears of Gleason (Fig. 9.11a). In these gears, there is a non-uniform
clearance between the top land of one tooth and the bottom land of the mating tooth
along the face width. This clearance becomes progressively smaller from the heel to
the toe.
Tip and root cone generatrices can be parallel to those of the reference pitch
cone, as happens for the toothings of constant depth of the spiral bevel gears of
Klingelnberg and Oerlikon (Fig. 9.11b). In this case a uniform clearance can be
maintained. In addition, these toothings with constant depth facilitate the cutting
process, because the cutter, running along the root cone generatrices, moves on
guides oriented in the direction of the reference pitch cone and performs a gener-
ation motion very accurate, and allows an easier adjustment of the gear-cutting
machine. The teeth with constant depth are however leaner at the toe, whereby the
risk of undercut can occur.
The root cone generatrices, parallel to those of the tip cone, can intersect the
generatrices of the reference pitch cone, as happens for the toothings with constant
depth of the spiral bevel gears of Klingelnberg-Palloid (Fig. 9.11c). Finally, the tip
and root cone generatrices can degenerate into the tip and root cylinder generatrices,
parallel to the axis to the wheel, which intersect the reference pitch cone genera-
trices at the two end faces of the wheel, as occurs in the cylindrical-bevel gears
(Fig. 9.11d).
Figure 9.10 shows, in addition to the back-cone distance, Re , also the inner cone
distance, Ri . It also highlights the back cone and mean cone, i.e. two of the num-
berless cones between the back cone and inner cone. Usually, the study of tooth
profiles of a bevel gear pair is led back to the study of the tooth profiles of an
equivalent cylindrical gear pair, which is obtained by developing on a plane the
back cones. This is an approximate method, proposed for the first time by Tredgold,
(c)
(a) (b) (d)
Fig. 9.11 Tip and root cone generatrices: a converging in the operating pitch cone apex;
b parallel to pitch cone generatrices; c root cone generatrices intersecting those of the pitch cone
and parallel to those of the tip cone; d tip and root cone generatrices parallel to the wheel axis
9.5 Main Conical Surfaces and Equivalent Cylindrical Gear 371
d1 d2
rv1 ¼ rv2 ¼ ; ð9:43Þ
2 cos d1 2 cos d2
therefore, the virtual (or equivalent) numbers of teeth of the two virtual cylindrical
wheels are given by:
z1 z2
zv1 ¼ zv2 ¼ : ð9:44Þ
cos d1 cos d2
The diagram in Fig. 9.13 allows to calculate immediately the virtual number of
teeth zv as a function of the actually number of teeth, z, and the reference cone
angle, d.
Therefore, the gear ratio of the equivalent cylindrical gear pair is given by the
relationship:
It is to be noted that, for the calculations of load carrying capacity of bevel gears,
the virtual parallel cylindrical gear pair to be considered, according to the Tredgold
approximation, is that corresponding to the mean cones, i.e. the cones whose
generatrices bisect the face width, and are therefore equidistant and parallel to the
generatrices of the back and inner cones. With reference to Fig. 9.14, we have the
following quantities related to the mean cone (subscripts m and e indicate quantities
respectively related to the mean and back cones):
• Reference diameters, dvm :
Fig. 9.14 Quantities for the calculation of virtual cylindrical gears at mid-face width
374 9 Straight Bevel Gears
pffiffiffiffiffiffiffiffiffiffiffiffiffi
u2 þ 1
dvm1 ¼ dm1 dvm2 ¼ u2 dvm1 : ð9:49Þ
u
1
avm ¼ ðdvm1 þ dvm2 Þ: ð9:50Þ
2
To avoid the interference between two straight bevel wheels, having z1 and z2 teeth
and cone angles d1 and d2 , it is sufficient to prevent the interference between the
two equivalent parallel cylindrical wheels. In the case of external bevel gears, which
are the ones that actually found practical applications, the minimum number of teeth
of the pinion, z1min , to prevent the interference in the operating conditions is
obtained from Eq. (4.10) or Eq. (3.14) written in terms of z1 and z2 , substituting
therein z1 and z2 respectively with zv1 and zv2 , given by (9.44) or (9.53). Therefore,
we have:
9.6 Minimum Number of Teeth to Avoid Interference 375
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
1 þ 1 þ z2 cos d1 2 þ z2 cos d1 sin at
z1 cos d2 z1 cos d2 2
where k2 is the addendum factor of the largest wheel. Thus, the minimum number of
teeth of the pinion, z1min , varies linearly with k2 , and depends on the ratio
ðz1 =z2 Þ ¼ ð1=uÞ, transverse pressure angle at and shaft angle R ¼ ðd1 þ d2 Þ.
Recalling Eqs. (9.45), (9.54) can be written also in the following more compact form:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2k2 cos d1 uv þ u2v þ ð1 þ 2uv Þ sin2 at
z1min qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼ 2k2 cos d1 :
ð1 þ 2uv Þ sin2 at
u2v þ ð1 þ 2uv Þ sin2 at 1
ð9:55Þ
It is then necessary to check that the cutting interference does not occur during
the generation process, when cutting of the bevel wheel is done by means of a gear
cutting machine that simulates the generation crown wheel. Since the cone angle d2
of the generation crown wheel is equal to p=2, from Eq. (9.54) we infer that, to
avoid the cutting interference it must be:
cos d1
z1min 2k20 ; ð9:56Þ
sin2 at
where k20 is the addendum factor of the cutting tooth. Therefore, in this condition,
z1min varies linearly with k20 and depends on d1 , and thus on R ¼ ðd1 þ d2 Þ, and
transverse pressure angle at .
Diagrams in Figs. 9.15 and 9.16, and in Fig. 9.17, respectively show the curves
related to Eq. (9.54) or the equivalent Eqs. (9.55), and (9.56). The diagrams of
Figs. 9.15 and 9.16, respectively valid for at ¼ 15 and at ¼ 20 , highlight, for
k2 ¼ 1, the variation of z1min in the operating conditions, as a function of the shaft
angle R, shown on the abscissa, and the ratio ðz1 =z2 Þ ¼ ð1=uÞ, which is the
parameter related to each curve. The diagram of Fig. 9.17 shows in a single syn-
optic, for k20 ¼ 1, the double family of curves (top, one inherent to at ¼ 15 , and
below that relating to at ¼ 20 ) that give, in the cutting conditions performed with
generation crown wheel, the variation of z1min as a function of the shaft angle
R ¼ ðd1 þ d2 Þ and ratio ð1=uÞ ¼ ðz1 =z2 Þ.
The number of teeth, z20 , of the virtual generation crown wheel in the cutting
conditions, bearing in mind the Eq. (9.15), and remembering that d20 ¼ d0 ¼ p=2,
is given by:
376 9 Straight Bevel Gears
Fig. 9.15 Distribution curves of z1min given by Eq. (9.54), as a function of R ¼ ðd1 þ d2 Þ and
ð1=uÞ ¼ ðz1 =z2 Þ, for k2 ¼ 1 and at ¼ 15
z1
z20 ¼ ¼ z0 : ð9:57Þ
sin d1
Of course, all this is true for bevel gears with zero profile-shifted toothing.
Comparing the results obtained using the relationships (9.54) and (9.56), or the
diagrams shown in Figs. 9.15, 9.16 and 9.17, we can deduce that the more
restrictive conditions are those inherent in the cutting conditions. For bevel gear
pairs with perpendicular axes ðR ¼ 90 Þ, and with k1 ¼ k2 ¼ 1, the values of z1min
9.6 Minimum Number of Teeth to Avoid Interference 377
Fig. 9.16 Distribution curves of z1min given by eq. (9.54), as a function of R ¼ ðd1 þ d2 Þ and
ð1=uÞ ¼ ðz1 =z2 Þ, for k2 ¼ 1 and at ¼ 20
that are suitable to avoid both the operating interference with the wheel of z2 teeth,
and the cutting interference with the virtual generation crown wheel, are those
collected in Table 9.2.
It is to remember that the values of z1min taken from diagrams shown in
Figs. 9.15, 9.16 and 9.17 and those collected in Table 9.2 refer to teeth with
addendum factor k2 ¼ k20 ¼ 1. The relationships (9.54) and (9.56) show, however,
that z1min is proportional to the addendum factor. Therefore, if different sizing were
adopted for the teeth, the values of z1min to be choose are those derived from
diagrams and table above, multiplied by the actual addendum factor.
As for the cylindrical spur gear (see Sect. 3.2) and cylindrical helical gears (see
Sect. 8.5), also in this case it is admitted that, in practice, the minimum number of
teeth z1min that can be cut by generation cutting using a virtual generation crown
wheel is equal to 5=6 of the theoretical values. The diagram shown in Fig. 9.18
allows to calculate the minimum number of teeth, z1min , that can be cut without
interference, as a function of the reference pitch cone angle d1 , shown on the
abscissa, and transverse pressure angle at , shown on each curve. The broken line
curves represent the theoretical values of z1min , while the solid line curves provide
the practical values z01min of z1min . With z01min a marginal amount of undercut is
allowed.
378 9 Straight Bevel Gears
Fig. 9.17 Distribution curves of z1min given by Eq. (9.56), as a function of R ¼ ðd1 þ d2 Þ and
ð1=uÞ ¼ ðz1 =z2 Þ, for k20 ¼ 1 and at ¼ 15 , and at ¼ 20
Table 9.2 Values of z1min to avoid both operating and cutting interference, for bevel gear pair
with R ¼ 90 , and k1 ¼ k2 ¼ k20 ¼ 1
z1 =z2 z1min z1 =z2 z1min
at ¼ 15 at ¼ 20 at ¼ 15 at ¼ 20
1/1,00 21 12 1/3 28 16
1/1,05 22 12 1/4 29 17
1/1,10 22 13 1/5 29 17
1/1,25 23 13 1/6 29 17
1/1,43 24 14 1/8 30 17
1/1,67 26 15 1/10 30 17
1/2,00 27 16
Figure 9.18 also shows the advantage of the straight bevel gears with respect to
the parallel cylindrical spur gears, in terms of minimum number of teeth that can be
cut without interference. For example, for at ¼ 20 and d1 ¼ 30 , z01min drops from
14 teeth to 12 teeth. Here, too, it is obvious that, when we wanted to use pinions
characterized by a number of teeth even lower, and together dispel the risk of the
9.6 Minimum Number of Teeth to Avoid Interference 379
Fig. 9.18 Theoretical ðz1min Þ and practical z01min minimum number of teeth as a function of the
reference pitch cone angle, d1 , and transverse pressure angle, at , for bevel gears cut with virtual
generation crown wheel
In accordance with the different processes used for the production of bevel gears,
different geometric shapes of teeth can be obtained. It follows that, in this specific
manufacturing field, it is not possible to identify a reference profile of the toothing
of general use. For bevel gears with constant depth teeth, in most cases, the starting
point consists of the standard basic rack tooth profile according to ISO 53: [17],
concerning cylindrical gears (see Fig. 2.14). This rack profile is the profile of the
tooth of an imaginary (or virtual) crown wheel when its cylindrical back surface, i.e.
the back cone degenerated into a cylinder, and the tooth profile upon it, are
developed on a plane. For bevel gears having straight teeth, this profile is used as
the basis of reference.
We have seen that, in the generation process commonly used, the tooth flanks of
the virtual crown wheel are plane surfaces and generate teeth of octoid shape. This
toothing can be performed with or without profile shift. In the case of bevel gears
with profile-shifted toothing, for their proper kinematics, additional conditions
compared to those of the cylindrical involute spur gears must be guaranteed.
However, the various gear cutting machines used for their manufacture, while based
on different technologies, allow to adjust and vary the position of the cutting edges
of the two tools, which cut the operating and non-operating flanks independently of
one another. This happens for the gear cutting machines where the two finishing
tools, having straight cutting edges, work separately the two aforementioned flanks
(see Sect. 9.12). Due to these characteristics, it is possible, without additional costs,
vary the standard reference profile, albeit within well-defined limits. In fact, we
have therefore the possibility to modify the tooth profiles, to improve and optimize
the operating behavior of the bevel gears, and to facilitate their manufacturing
process, with gear cutting machines that work by chip removal.
With the exception of bevel gears with profile-shifted toothing and variation of
the shaft angle, in all other possible variants, the reference pitch cones, which are
the generation pitch cones, must be used as operating pitch cones. To meet this
restrictive condition, well-defined rules of manufacturing and mounting must be
scrupulously respected. Therefore, it is obvious that, according to the Tredgold
approximation, the imaginary rack-type cutter, with which the wheel is cut, will be
the mirror image, compared to the reference pitch line, of the equally imaginary
rack-type cutter with which the pinion is cut. The imaginary rack-type cutter is
achieved by developing on a plane the back cone of the generation octoid crown
wheel, which has degenerated into a cylinder.
For the above reasons, the range of families of bevel gears are much broader than
that of cylindrical gears. Indeed, we can have the following types of bevel
gears (see Fig. 9.19):
(A) Bevel gears with zero profile-shifted toothing, that is constituted by two bevel
wheels, both without profile shift.
9.7 Reference Profile Modifications 381
(a) (b)
(c)
(d)
(e)
Fig. 9.19 Reference profile modifications of bevel gears: a positive profile shift of the pinion;
b negative profile shift of the wheel, in absolute value equal to that of the pinion; c increase of the
transverse thickness of the pinion tooth; d variation of the tooth depth of pinion and wheel;
e variation of the nominal pressure angle of the operating and non-operating flanks (typically,
aD 9 for the operating flank, and aC 31 for the non-operating flank)
(B) Bevel gears with profile-shifted toothing without variation of the shaft angle,
i.e. consisting of two bevel wheels, both with profile shift, positive for the
pinion, and negative for the wheel, but equal in absolute value, as Figs. 9.19a
and 9.19b show. As for the cylindrical gears, the profile shift coefficient xh is
chosen so as to achieve different design goals, such as:
382 9 Straight Bevel Gears
As we have said elsewhere, the use of gears with profile-shifted toothing is related
to the involute profile. In fact, two conjugate involutes of two base circles continue
to have an appropriate kinematics in whatever way we change the relative position
of the axes of the two wheels, provided that the teeth touch each other: only the
pressure angle changes. We have also seen that, in the case of bevel wheels, the
spherical involute profiles effectively do not exist, since the bevel gears that have a
real practical interest are those with toothing having octoid profile.
The octoid profile is different, albeit not by much, from that of the spherical
involute profile, whereby the wheels with octoid toothing has a kinematic behavior
that only approximately emulates that of the wheels with toothing having spherical
involute profile. Their kinematic behavior is correct only when the generating
condition is respected, that is when operating condition and cutting condition, i.e.
the generating condition, coincide. Even modest differences between these two
conditions determine an incorrect kinematic behavior.
From a strictly theoretical point of view, it is therefore improper to speak of
profile shift for bevel gears. De facto, we speak of profile-shifted toothing of the
bevel gears, believing that the octoid profiles approximate well enough the spherical
involute profiles, so much so that the octoid toothing is confused with the spherical
involute toothing. In this framework, in approximate terms and with all the limi-
tations resulting from it, we continue to speak of profile shift of the bevel gears by
extending to them, mutatis mutandis, the concepts that we have already described
for cylindrical involute gears.
As for cylindrical involute gears, to achieve specific design goals, it is necessary
to resort to the profile shift, changing the position of the pitch cone surfaces with
respect to the tooth depth, and therefore varying the operating pitch cone angle.
Also, for straight bevel gears, the design goals to be achieved with profile shift are:
384 9 Straight Bevel Gears
• the reduction of the minimum number of teeth below the one defined in
Sect. 9.6, to avoid undercut and interference in operating conditions;
• greater bending strength of the teeth;
• higher contact ratio, and therefore less noisy gears;
• more balanced specific sliding between pinion and wheel;
• higher strength to surface fatigue and wear; etc.
To further study, we refer the reader to specialized textbooks, and the instructions
and suggestions of manufacturers of bevel gear cutting machines. We here set the
problem in its essential lines, but in entirely general terms, with the purpose of use
results obtained with cylindrical involute gears, in the framework of the Tredgold
approximation. As we have said, in almost all practical applications, bevel gear
pairs with profile-shifted toothing without variation of shaft angle are used, i.e.
bevel gear pairs with profile shift coefficients for pinion and wheel equal, but
opposite; this is the profile shift type Lasche [19], which therefore satisfies the
relationship ðxh1 þ xh2 Þ ¼ 0. Bevel gear pairs with profile-shifted toothing and
variation of shaft angle are rarely used. This is because, to achieve predetermined
design goals, unlike the cylindrical gears, we can use other modifications in the
teeth profiles, such as those described in the previous section.
Therefore, the general discussion that follows has no condition as regards the
sum of profile shift coefficients of the pinion and wheel, which can assume any
value, i.e. ðxh1 þ xh2 Þ 6¼ 0. This is because, using gear-cutting machines based on
the generation cutting process, we can choose, for the pinion and wheel, the profile
shift considered more advantageous with regard to different adoptable design cri-
teria. In this framework, we suppose to use as virtual cutting tool, a generation
crown wheel having pitch circle diameter d0 ¼ 2r0 , virtual number of teeth z0 , and
therefore module given by:
d0 2r0
m0 ¼ ¼ : ð9:58Þ
z0 z0
It is to be noted that, as for the other gears, the subscript 0 indicates quantities
relating to the cutting tool. In addition, it is to be remembered that the reference
pitch circle, which is also the operating pitch circle during the cutting without
profile shift, is the one that we get as the intersection of the reference pitch plane of
the generation virtual crown wheel with the back cylinder. The pitch plane of the
generation virtual crown wheel is the plane in which the reference pitch cone is
degenerated, and the back cylinder is the cylinder in which the back cone is
degenerated.
To have bevel gears with zero profile-shifted toothing, the operating pitch cone
of the virtual crown wheel during the cutting process has cone angle d0 ¼ 90 . This
operating pitch cone is the generation pitch cone that, in this case, coincides with
the reference pitch plane of the same crown wheel. Instead, to have bevel gears with
profile-shifted toothing, the generation pitch cone angles of the virtual crown wheel
will be different from ðp=2Þ and, for the pinion and wheel, will be equal to:
9.8 Straight Bevel Gears with Profile-Shifted … 385
Fig. 9.20 Cutting condition of a straight bevel gear wheel with profile-shifted toothing, with
respect to the generation virtual crown wheel
p p
d01 ¼ þ n1 d02 ¼ þ n2 ; ð9:59Þ
2 2
where n1 and n2 are, respectively, the deviations of the generation pitch cone angles
of pinion and wheel from 90°.
However, in correspondence of the cylinder of diameter d0 , corresponding to the
back cone of the crown wheel, degenerated in a cylinder, the module is everywhere
m0 . Therefore (Fig. 9.20), the diameters of the pitch circles of the pinion and wheel
in the cutting conditions will be given by the relationships:
d1 ¼ z 1 m 0 d2 ¼ z 2 m 0 : ð9:60Þ
On the other hand, the operating pitch cone angles d1 and d2 of the pinion and
wheel, meshing with backlash-free contact, determined by the Eqs. (9.20) and
(9.21) as a function of u and R, which are design data, differ from the generation
pitch cone angles d01 and d02 of the amounts Dd1 and Dd2 , for which we shall have:
Therefore, the shaft angle R of the bevel gear pair meshing with backlash-free
contact will be given by:
profile, shown in Sect. 9.9.4, is extensible to bevel gears with octoid tooth profile.
Then we assume that a0 is the pressure angle of the generation crown wheel.
According to the second of Eq. (9.42), in which we put d2 ¼ d0 ¼ p=2 and a ¼ a0
(note that we are applying the second of Eq. (9.42) to the generation virtual crown
wheel), we infer that the base cone angle of the generation crown wheel is given by:
We denote by db1 the base cone angle of the pinion to be cut, and a10 its cutting
pressure angle (see Fig. 9.20, on the right), which corresponds to the cutting pitch
cone d10 of the same pinion, and to the base cone angle db0 of the generation virtual
crown wheel. We then use Eq. (9.42) twice, first with reference to the cutting pitch
cone of the virtual crown wheel and then with reference to the cutting pitch cone of
the pinion to be cut. Therefore, we get the following relationships:
where the first of the two subscripts refers to the wheel under consideration (0, for
the virtual crown wheel; 1, for the pinion; 2, for the wheel), while the second
subscript, 0, identifies the cutting condition. Obtaining cos a10 from the first of these
equations and replacing what was found in the second, and recalling Eq. (9.63) and
the first of Eq. (9.59), we obtain:
cos a0 cos a0
sin db1 ¼ sin d10 ¼ sin d10 : ð9:65Þ
sin d00 cos n1
The cutting pressure angles a10 and a20 , respectively used for the pinion and
wheel, are defined by the relationships:
In relation to the profile shift, being known d1 and d2 , which are determined as a
function of the design data, u and R, using Eqs. (9.20) and (9.21), it is necessary to
determine the other unknown quantities. To this end, we consider the pinion and
wheel in their relative position corresponding to the meshing condition with zero
backlash contact (Fig. 9.21). In this position, the operating pitch cones (i.e. the
cones with cone angles d1 and d2 , which are different from the cutting pitch cones
having cone angles d10 and d20 ) of the bevel gear pair are tangent to each other. The
same figure shows the back cones of the two wheels, i.e. the cones whose common
generatrices form, with the axes of the same wheels, the cone angles given
respectively by ½ðp=2Þ ðd10 þ n1 Þ and ½ðp=2Þ ðd20 þ n2 Þ.
However, the intersections between the above-mentioned back cones, and the
cylinder of radius r0 , which delimits the generation virtual crown wheel and rep-
resents its back cone degenerated in this cylinder, are the pitch circles of the two
wheels in the cutting condition; they both have radius r0 . However, it is to keep in
mind that, according to the Tredgold approximation, for the study of meshing in the
cutting conditions, as well as for the analysis of the forces exchanged between the
teeth, it is convenient to refer precisely to these back cones. This because, in their
development on a plane, the toothing of the generation virtual crown wheel is
developed in the standard rack-type cutter with pressure angle a0 .
For ease of calculation, it is then convenient to consider, as operating pitch
circles, those having the same radii r1 ¼ r10 and r2 ¼ r20 of the cutting pitch circles,
for which the module is m0 . It is also convenient to introduce the number of teeth Z0
388 9 Straight Bevel Gears
of a fictitious cutting crown wheel, having radius OC ¼ ðZ0 m0 Þ=2, which meshes
with both members of the bevel gear pair arranged in the meshing position with
zero backlash contact (Fig. 9.21). Obviously, we have:
z1 z2
Z0 ¼ ¼ : ð9:68Þ
sin d1 sin d2
From this relationship, taking into account Eqs. (9.18) and (9.19), we obtain:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
z21 þ z22 þ 2z1 z2 cos R
Z0 ¼ : ð9:69Þ
sin R
Substituting Eq. (9.68) in Eq. (9.44), and taking into account Eqs. (9.20) and
(9.21), we obtain the following relationships that give the virtual numbers of teeth
of the equivalent cylindrical wheels:
Z0 sin R
zv1 ¼ Z0 tan d1 ¼ ð9:70Þ
u þ cos R
uZ0 sin R
zv2 ¼ Z0 tan d2 ¼ : ð9:71Þ
1 þ u cos R
Following the same procedure described for cylindrical gears, and using the
above virtual numbers of teeth, as well as the relationships or diagrams and charts
available, we obtain the optimal values of the profile shift coefficients xh1 and xh2 ,
and then the pressure angle, a0 , using the relationship:
xh1 þ xh2
inva0 ¼ inva0 þ 2 tan a0 ; ð9:72Þ
zv1 þ zv2
which is similar to Eq. (6.58), deduced and used for cylindrical gears. To this
purpose, we can use, for example: the diagrams of Schiebel (see Schiebel and
Lindner [34, 35], the relationships and diagrams of Henriot [15]; the diagrams of
Niemann and Winter [24]; etc.
Let us consider Fig. (9.20) to the left, in which however we take into account, in
place of the generation virtual crown wheel that is bounded by the cylinder of
radius r0 (the back cone degenerated in this cylinder), the fictitious crown wheel
bounded by the cylinder of radius OC ¼ ðZ0 m0 Þ=2. From this modified figure, we
infer that the cone angle variations n1 and n2 of the cutting pitch cones of the
fictitious crown wheel compared with the right angle, which we would have in the
absence of profile shift, are given by the relationships:
Since d1 and d2 are known from the design data, once the pressure angle a0 in the
operating conditions is determined by the Eq. (9.72), the base cone angles and
cutting pitch cone angles of the two wheels, and the differences Dd1 and Dd2 which
appear in Eq. (9.61) remain to be calculated.
From Eq. (9.42), we get:
these relationships, taking into account Eq. (9.68), can be written also in the fol-
lowing form:
z1 z2
sin db1 ¼ cos a0 sin db2 ¼ cos a0 : ð9:75Þ
Z0 Z0
Comparing Eq. (9.75) with Eqs. (9.65) and (9.66), we get the following rela-
tionships that allow us to calculate the cutting pitch cone angles of the two wheels:
z1 cos a0 z2 cos a0
sin d10 ¼ cos n1 sin d20 ¼ cos n2 : ð9:76Þ
Z0 cos a0 Z0 cos a0
From Fig. (9.20), we then obtain the following relationship that relates the radius
r0 with the radius r1 ¼ d1 =2 of the cutting pitch circle of the pinion:
r0
r10 ¼ sin d10 ; ð9:78Þ
cos n1
therefore, we have:
z0
z1 ¼ sin d10 : ð9:79Þ
cos n1
Substituting in this last equation the ratio ðsin d10 = cos n1 Þ that we derive from
the first of Eq. (9.76), we obtain the relationship:
cos a0
z0 ¼ Z0 ; ð9:80Þ
cos a0
which correlates the number of teeth, z0 , of the generation virtual crown wheel with
the number of teeth, Z0 , of the fictitious crown wheel, which we introduced for
convenience of calculation.
It is now necessary to reconsider the hypothesis, previously put forward, that the
two members of the bevel gear pair under consideration mesh with zero backlash
contact. In fact, as we have said for cylindrical spur gears, to ensure optimal
390 9 Straight Bevel Gears
operating conditions of the bevel spur gears, it is also necessary to leave a suitable
backlash between the flanks of the teeth in mesh. This suitable backlash both
ensures good lubrication conditions, and prevents possible interlocking of the teeth
of a wheel in space widths of the mating wheel, due to different thermal expansions
between pinion and wheel. This backlash is obtained by increasing the shaft angle,
R, of a suitable amount, DR, to be determined.
To give an idea of the calculation procedure to be made, consider the case of a
straight bevel gear characterized by tapered toothing from the heel toward the toe,
and assume that the clearance cP is decreasing linearly in the same direction above,
namely that the angular clearance is constant.
According to the generation cutting process described above, between the
operating pitch cones and root cones of the pinion and wheel we have the following
dedendum angles:
where #f 10 and #f 20 are the dedendum angles in the cutting condition (see
Fig. 9.20). With a nominal sizing, i.e. with addendum and dedendum respectively
equal to 1:00m0 and 1:25m0 , we will have:
m0 2:50
tan #f 10 þ n1 ¼ tan #f 20 þ n2 ¼ 1:25 ¼ : ð9:82Þ
r0 z0
Since between the tip cone of the pinion and root cone of the mating wheel, and
vice versa, we want to have an angular clearance equal to ðm0 =4r0 Þ ¼ ð1=2z0 Þ, it is
necessary that the angular addendum between the operating pitch cones and tip
cones of the pinion and wheel are respectively equal to:
1 1
#a1 ¼ #f 20 þ Dd2 #a2 ¼ #f 10 þ Dd1 : ð9:83Þ
2z0 2z0
Instead, according to this generation cutting process, the two actual angular
addendums are respectively equal to:
where #a10 and #a20 are the addendum angles in the cutting condition (see
Fig. 9.20). With a nominal standard sizing, we will have:
m0 2:00
tanð#a10 n1 Þ ¼ tanð#a20 n2 Þ ¼ ¼ : ð9:85Þ
r0 z0
2:50
D#a1 ¼ #a10 #f 20 ðDd1 Dd2 Þ þ
z0
ð9:86Þ
2:50
D#a2 ¼ #a20 #f 10 ðDd1 Dd2 Þ þ :
z0
To obtain between the meshing teeth a suitable normal backlash (this is the
backlash measured along the direction of the normal to the contact surface), under
the hypothesis that the toothing modification is only that described above, i.e. the
profile shift, it is necessary to increase the shaft angle R by an amount DR in such a
way to have:
Z0 m0
jn ¼ 2DR sin a0 ; ð9:87Þ
2
jn jn
DR ¼ 0
ffi : ð9:88Þ
Z0 m0 sin a Z0 m0 sin a0
With the procedure described above, all the quantities regarding gear wheels
with profile-shifted toothing are completely defined. Synthetically, the various steps
of the calculation procedure of interest here are the following:
• First, we fix the nominal backlash jn as a function of the module m (this can be
the external transverse module, met , or the mean normal module, mmn ,
depending on the case under consideration), using data or diagrams found in the
scientific literature (e.g. Niemann and Winter [25]).
• Then we calculate DR by Eq. (9.88), assuming a0 ffi a0 :
• Subsequently we calculate Z0 by means of Eq. (9.69), in which we replace the
actual shaft angle R with a calculation shaft angle, equal to ðR DRÞ:
• We then determine zv1 and zv2 by means of Eqs. (9.70) and (9.71).
• Depending on the optimization method chosen, we determine the optimal values
of the profile shift coefficients, xh1 and xh2 .
• We then calculate n1 and n2 by means of Eq. (9.73).
• We determine the operating pressure angle, a0 , by means of Eq. (9.72).
• Finally, we calculate the operating pitch cone angles and tip cone angles of the
two wheels of the straight bevel gear pair, by means of Eqs. (9.80), (9.82),
(9.85) and (9.86).
For better orientation within the profile shift of straight bevel gears, in addition to
what we said at the beginning of this section, it should be noted that, for gear ratios
next to the unit, the profile shift is not necessary. For gear ratios very different from
the unit, the meshing and operating conditions of a straight bevel gear are instead
very unfavorable, whereby the profile shift becomes necessary. We have already
said that the bevel gear pairs mostly used are those with profile-shifted toothing
without variation of shaft angle. This corresponds to profile shift coefficients, which
392 9 Straight Bevel Gears
Fig. 9.22 Diagram to determine the profile shift coefficients of straight bevel gear pairs with shaft
angle R ¼ 90 , pressure angle a0 ¼ 20 , and profile-shifted toothing without variation of the shaft
angle
satisfy the relationship ðxh1 þ xh2 Þ ¼ 0, i.e. profile shift coefficients equal in abso-
lute value, but positive for the pinion and negative for the wheel.
For the choice of the profile shift coefficients of straight bevel gear pairs with shaft
angle R ¼ 90 , pressure angle a0 ¼ 20 and profile-shifted toothing without variation
of shaft angle, we can use the chart shown in Fig. 9.22. In this chart we enter with the
number of teeth of the wheel, reported on the abscissa, and with a vertical line
downward cross the curve of the number of teeth of the pinion. The horizontal line
through this crossing point identifies, on the ordinate on the left, the profile shift
coefficient to be choose. Note that the diagram is delimited, towards the bottom, by the
limiting curve of the crest thickness of the pinion tooth and, on the left, by the
root-undercut curve for cutting tools having a tooth tip radius equal to 0:2m.
We also reported that the straight bevel gear pairs with profile-shifted toothing
and variation of shaft angle are rarely used. However, in the case we want to design
bevel gears of this type, that is with ðxh1 þ xh2 Þ 6¼ 0,, the optimum values of the
profile shift coefficients can be determined by reducing the bevel gear pair to an
equivalent virtual cylindrical gear pair corresponding, in the Tredgold approxi-
mation, to the mean cones. In this case, it is necessary first to determine the
numbers of teeth zv1 and zv2 using the relationships (9.70) and (9.71), and then use
9.8 Straight Bevel Gears with Profile-Shifted … 393
the criteria and diagrams available for the cylindrical gears. To this purpose, for
example, the diagrams of Schiebel and Lindner [15, 35], Henriot [15], and
Niemann and Winter [24], mentioned above can be used.
Finally, it should be noted that the geometry of straight bevel gears can be
defined using the criteria described in Chap. 12, concerning spiral bevel and hypoid
gears. In fact, straight bevel gears are a special case of the spiral bevel gear family,
and of the more general hypoid gear family. Furthermore, computerized methods
can be used to optimize tooth profiles (see, for example, Fuentes et al. [11]).
In Sect. 6.1 we have already indicated that the profile shift is a measure of tooth
thickness, highlighting the fact that the addendum modification without tooth depth
variation is the result of a tooth thickness variation. In this section, we examine the
same subject from a different point of view than discussed in the previous section.
This different point of view is related to a unique ability offered by the bevel gear
cutting machines. This ability consists in the fact that, with the tooth depth
remaining constant, we can change not only the addendum and dedendum of the
generation crown wheel (this is the profile shift, previously discussed), but we can
also change, simultaneously and independently from the change above, the tooth
space between the teeth and their thicknesses, the pitch remaining unchanged. This
is the transverse tooth thickness modification. In other words, the toothing modi-
fications of which we are speaking are given by the combination of those shown in
Figs. 9.19a and 9.19b, however with ðxh1 þ xh2 Þ 6¼ 0, and in Fig. 9.19c. Since all
these modifications in any case determine a tooth thickness change, without tooth
depth modification, we address here the subject in terms of transverse tooth
thickness modification.
Figure 6.3 best represents the actual geometry of the generation rack-type cutter,
obtained developing on a plane the cylinder of radius r0 ; which delimits the gen-
eration crown wheel (Fig. 9.20). Also taking into account this Fig. 6.3, from
Fig. 9.19c we infer the following relationships, which provide respectively the
space width e0 and tooth thickness s0 , both measured along the generation pitch line
of the basic rack:
pm0 pm0
e0 ¼ 2xs m0 ¼ epm0 s0 ¼ þ 2xs m0 ¼ ð1 eÞpm0 ; ð9:89Þ
2 2
394 9 Straight Bevel Gears
where
1 2xs
e¼ ð9:90Þ
2 p
z1min z⋇1min
x1 ¼ ; ð9:91Þ
z1min
where z1min is the value obtained from Fig. 9.17 for the given values of
1=u ¼ z1 =z2 , at and R. Of course, the pinion generated with a profile shift coeffi-
cient, x1 , will have addendum and dedendum given respectively by ha1 ¼
ð1 þ x1 Þm0 and hf 1 ¼ ð1:25 x1 Þm0 .
Example Input data: straight bevel gear pair, with: R ¼ 80 ; 1=u ¼ z1 =z2 ¼ 1=2;
at ¼ 15 ; z1 ¼ z1min ¼ 20. As a design goal, it is necessary to avoid the cutting
interference with the generation crown wheel. Results: (i), from curves shown in
Fig. 9.17, for k0 ¼ ðha0 =mÞ ¼ 1, we obtain z1 ¼ z1min ¼ 27; (ii), from Eq. (9.91),
we infer that, in order to avoid the cutting interference, the required profile shift
coefficient is equal to x1 ¼ ½ð27 20Þ=20 ¼ 0:259.
After this necessary premise on cutting interference, we return to the topic
outlined at the beginning of this section. Thanks to the operating flexibility of the
bevel gear cutting machines, we can vary, as best we believe from the point of view
of the design optimization, the tooth thickness s0 and space width e0 on the gen-
eration pitch line of the rack-type cutter, so as to prevent the teeth of the two
members of the gear pair are too weak. This rack-type cutter is obtained by
developing the generation crown wheel on a plane.
9.9 Straight Bevel Gears with Profile-Shifted … 395
2p
2#0s;1 ¼ e1 cos d1 þ 2ðinva0 inva0 Þ ð9:94Þ
z1
2p
2#0s;2 ¼ e2 cos d2 þ 2ðinva0 inva0 Þ: ð9:95Þ
z2
Since the angular pitches, sv1 and sv2 , of the two equivalent cylindrical wheels
are given by:
the space width angles 2#0e;1 and 2#0e;2 of the same two wheels on the pitch circles
above will be given by:
396 9 Straight Bevel Gears
2p
2#0e;1 ¼ sv1 2#0s;1 ¼ 2#0s;1 ð9:97Þ
zv1
2p
2#0e;2 ¼ sv2 2#0s;2 ¼ 2#0s;2 : ð9:98Þ
zv2
To have the profiles of the teeth of the two mating wheels in backlash-free
contact, on the above-mentioned pitch circles, the space width between the teeth of
the first wheel must be equal to the tooth thickness of the second wheel, that is the
Eq. (6.71) must be valid, in which we replace zv2 =zv1 instead of z2 =z1 :
Developing the Eq. (6.71) so modified, taking into account Eqs. (9.97), (9.94)
and (9.95), we obtain the following fundamental equation, that allows to calculate
a0 as a function of quantities all known:
pðe1 þ e2 1Þ
inva0 ¼ inva0 þ : ð9:99Þ
cos d1 þ cos d2
z1 z2
Once a0 has been determined this way, the diameters of the pitch circles in the
condition of backlash-free contact are those given by the Eq. (6.46) and (6.47),
which here become:
0 z1 m0 cos a0 0 z2 m0 cos a0
dv1 ¼ dv2 ¼ : ð9:100Þ
cos d1 cos a0 cos d2 cos a0
If, for the generation of the two members of the bevel gear pair, with
profile-shifted toothing and variation of the shaft angle, we take profile shift co-
efficients respectively equal to xh1 and xh2 , the maximum nominal addendum,
compared to the nominal pitch circles whose diameters are given by Eq. (9.92), will
be given by:
1 z1 z2 cos a0
h0a1max ¼ m0 þ 1 þ ð 1 þ x 2 Þ ð9:103Þ
2 cos d1 cos d2 cos a0
9.9 Straight Bevel Gears with Profile-Shifted … 397
Table 9.3 Main data concerning a straight bevel gear pair with shaft angle R, z1 and z2 teeth,
profile shift coefficients xh1 and xh2 , and thickness factors e1 and e2 ; generated by a virtual crown
wheel having module m0 and pressure angle a0
Quantity Pinion Wheel
Number of teeth z1 z2
Profile shift coefficient xh1 xh2
Thickness factor e1 e2
Pitch cone angle tan d1 ¼ 1 þðzð1z=z1 =z
2 Þ sin R
2 Þ cos R
tan d2 ¼ 1 þðzð2z=z2 =z
1 Þ sin R
1 Þ cos R
1 z1 z2 cos a0
h0a2max ¼ m0 þ 1 þ ð 1 þ x 1 :
Þ ð9:104Þ
2 cos d1 cos d2 cos a0
The actual dedendum of the two wheels coincide with the nominal values given
by Eq. (9.102).
Table 9.3 shows the main data concerning a straight bevel gear pair with z1 and
z2 teeth, shaft angle R, profile shift coefficients xh1 and xh2 , and thickness factors e1
and e2 , generated by means of a virtual crown wheel having module m0 and
pressure angle a0 .
Of course, when the sum of the thickness factors is equal to unit, that is, when:
ðe1 þ e2 Þ ¼ 1; ð9:105Þ
we infer that (see Eqs. 9.99 and 9.100), in the operating conditions of backlash-free
contact, the values of the pressure angle and the diameters of the pitch circles are
the nominal ones.
Some manufacturers of straight bevel gears use thickness factors related to the
profile shift coefficients by the same relationships valid for parallel cylindrical spur
gears, and therefore given by:
1 2
e¼ þ x tan a: ð9:106Þ
2 p
In this case, suppose we want to make a bevel gear pair with shaft angle R, and
characterized by a ratio ð1=uÞ ¼ ðz1 =z2 Þ and nominal values of the pressure angle
and diameters of the pitch circles. From the diagrams of Fig. 9.17, we derive, for
398 9 Straight Bevel Gears
k20 ¼ 1, the minimum number of teeth, z10 , of the pinion to avoid cutting inter-
ference with the generation crown wheel, whereby if
we can adopt, for the pinion, a profile shift coefficient equal to:
z1 z10
x1 ¼ ð9:108Þ
z1
x2 ¼ x1 : ð9:109Þ
1 2 1 2
e1 ¼ þ x1 tan a0 e2 ¼ þ x2 tan a0 ; ð9:110Þ
2 p 2 p
we see that Eq. (9.105) is satisfied, for which the values of the pressure angle and
diameters of the pitch circles will be equal to the nominal ones.
Example Input data: straight bevel gear pair with R ¼ 90 , ratio
ð1=uÞ ¼ ðz1 =z2 Þ ¼ 1=3, and nominal pressure anglea0 ¼ 20 . As a design choice,
we want to use a pinion with z1 ¼ 12 teeth, for which z2 ¼ 36 teeth. Thus, we have
ðz10 þ z20 Þ 2z1 . Therefore, we will adopt, for the pinion, a profile shift coefficient
equal to x1 ¼ ð15 12Þ=15 ¼ 0:2 and, for the wheel, a profile shift coefficient
x2 ¼ x1 ¼ 0:2. The related thickness factors are those given by Eq. (9.110).
Load analysis for straight bevel gears can be made with the same procedure used for
parallel cylindrical spur and helical gears, which neglects the contribution of the
friction forces. The total tooth force, F ¼ Fn , exchanged between the meshing teeth
acts along the direction of the normal to the tooth surface. As we said in the
Sect. 9.4 on the geometry of the spherical involute, we assume that the kinematics
of bevel gears with octoid toothing is comparable, with sufficient approximation, to
that of the bevel gears with spherical involute toothing. Under this assumption, for a
given position, the teeth are touching along the straight line of the plane of contact,
which passes through the apex of the pitch cones (Fig. 9.5). Thus, the normal to the
tooth surface in any point of this line is directed along the normal to this line, on the
plane of contact.
9.10 Load Analysis for Straight Bevel … 399
Therefore, the total tooth force has, as its line of application, the normal to the
instantaneous line of contact through the middle of the face width, and thus
coincident with a mean cone generatrix, on the plane of contact. This force,
however, actually occurs somewhere between the mean and back cones, but the
error due to the above assumption is marginal.
We denote by R the distance, which does not vary and remain constant, between
the line of application of the force Fn and the apex O of the pitch cone. In the plane
of contact, this force, constant in magnitude, changes its direction during the
meshing cycle, remaining however tangent to the circle of radius R. Referring to
Fig. 9.23, we denote with g the contact generatrix of the operating pitch cones, and
with d the cone angle of one of these cones. In the same figure, we see, to the left,
the plane of contact (it is tangent to the base cones along the straight lines OT1 and
OT2 , see also Fig. 9.5) rotated with respect to line g. In this plane, the instantaneous
position of the force Fn is defined by the angle c between the normal to the force Fn
(this normal is the instantaneous line of contact, OT in the figure) and the common
generatrix of contact g.
The total tooth force Fn can be resolved into three components, which act at right
angles to one another. These components are, respectively, the tangential force or
transmitted load, Ft , the axial force, Fa , and the radial force, Fr . In order to establish
the interrelationships between these components, consider the contact corre-
sponding to the instantaneous line of contact OT (Fig. 9.23). The force Fn is applied
at point T, and in this point it is tangent to the circle of radius R, and thus, it is
normal to the instantaneous line of contact. We now translate the force Fn parallel to
itself, to apply it at point A, whose distance from the apex O is given by OA ¼
R=cos c and is located on the common contact generatrix of the operating pitch
cones.
400 9 Straight Bevel Gears
Now we resolve the force Fn applied to point A into its two components, the first,
Fn sinc, directed along the straight line OA, and the second, Fn cos c, in the direction
perpendicular to the straight line OA. This last component lies in the plane tangent
to the sphere of radius OA (the plane having trace t in Fig. 9.23) and forms the
angle a (the pressure angle) with the normal to the plane formed by the axes of the
pitch cones. Therefore, it may be resolved into a component, Fn cosccos a, normal to
the plane shown in Fig. 9.23, to the right, and a component, Fn coscsin a; lying in
the plane shown in the same figure and perpendicular to the straight line OA.
In this way, we solved the force Fn in the three components Fn sin c; Fn cosccos a
and Fn coscsin a. Only the second of these components contributes to the trans-
mission of the torque T. In fact, we have:
R
T ¼ Fn cos c cos a OA sin d ¼ Fn cos c cos a sin d ¼ Fn Rm cos a; ð9:111Þ
cos c
where
Rm ¼ R sin d ð9:112Þ
is the radius corresponding, on the operating pitch cone generatrix, to the distance
R from the apex. The tangential component or tangential force
of the total force Fn , being the only one that transmits the torque T, is also called
transmitted load.
It is first to be noted that the force Fn is constant in absolute value only if the
torque applied to the wheel is constant. The axial and radial forces, respectively
directed parallel and radially to the axis of the wheel, are instead not constant. As
we can deduce from Fig. 9.23, these forces are given by the following relationships:
and therefore, vary as a function of d during the meshing cycle, even if Fn and
therefore the torque T remain constant.
Therefore, unlike what happens in the parallel cylindrical spur gears with
involute tooth profiles, in which the force exchanged between the meshing teeth is
constant in the absolute value and in direction along the path of contact, in this case
the force Fn is constant in absolute value, being equal to
T
Fn ¼ ; ð9:116Þ
Rm cos a
9.10 Load Analysis for Straight Bevel … 401
but it continuously varies its direction, moving with uniform angular velocity in the
plane of action, and keeping in this plane constantly tangent to the circle of radius
R ¼ Rm = sin d. The increased noise of the straight bevel gears with respect to the
parallel cylindrical gears depends, at least in a partial amount, by this variable
direction of the line of application of the force Fn .
For strength calculation of straight bevel gears, the instantaneous line of contact
characterized by c ¼ 0 is usually considered (thus it coincides with the contact
generatrix of the operating pitch cones). In this position, the Eqs. (9.114) and
(9.115) are simplified and take the form:
Bearing in mind the Eq. (9.17) and introducing the nominal transverse tangential
load Fnt referred to the radius Rm , i.e. the tangential force on the reference pitch
cone at the middle of face width, given by
T
Fnt ¼ ; ð9:119Þ
Rm
the Eqs. (9.117) and (9.118) can be written in the following form, which is the one
usually used in practical applications:
The introduction of the force Fnt is justified by the fact that a more precise
determination of the load distribution along the face width, although possible from
the theoretical point of view, is not convenient because of the unreliability of the
results arising from the inevitability to take into account the cutting and assembly
errors.
The general concepts of efficiency of the parallel cylindrical spur and helical gears,
discussed respectively in Sects. 3.9 and 8.10, can still be applied to the straight
bevel gears, with the appropriate variations of the case. Here we dwell on these
variations, which are peculiar to the straight bevel gears.
To address the problem, we assume that only one tooth pair is in engagement,
and that the mating surfaces are generated according to the procedures described in
previous sections, so the geometric properties of toothing are perfectly known. In
particular, the following two geometric entities are known:
402 9 Straight Bevel Gears
• The plane of action, that is the portion of plane having the shape of a circular
ring sector, delimited by the two circles with radii Re and Ri , and the interference
lines OT1 T10 and OT2 T20 . Therefore, the plane of action defines the theoretical
circular ring sector of contact, analogously to the theoretical rectangle of contact
we introduced in Sect. 8.2 for parallel cylindrical helical gears.
• The path of contact, that is the portion of plane having also the shape of a
circular ring sector, delimited by the same two circles with radii Re and Ri , and
the straight lines OAA0 and OEE 0 . These straight lines are the intersections of the
plane of action with the tip cones of the two members of the straight bevel gear
pair under consideration. Therefore, the path of contact defines the actual cir-
cular ring sector of contact.
These two geometric entities are shown in Fig. 9.24, where they are respectively
indicated with T1 T10 T2 T20 and AA0 EE 0 . Figure 9.24 also shows the instantaneous line
of contact l ¼ Pi Pe , defined by the angle u that it forms with the instantaneous axis
of rotation OCi Ci0 , and the differential normal component, dFn , that the pinion
transmits to the wheel along a differential segment, ds, of the instantaneous line of
contact. It acts on the plane of action, and is perpendicular to the instantaneous line
of contact. However, in addition to this differential normal component, dFn , we will
have a differential passive loss component, dFl ¼ ldFn , due to friction, acting on
the same differential segment, ds. This component, also called differential friction
Fig. 9.24 Differential normal component dFn acting on a differential segment ds of the
instantaneous line of contact
9.11 Efficiency of Straight Bevel Gears 403
force, is directed according to the relative velocity vector vr of the driven wheel
with respect to the driving pinion at point P, so it is directed perpendicularly to the
plane of action, and oriented in such a way as to obstruct the relative motion.
The relative velocity vector, vr , is given by the relationship:
where x2 and x1 are the angular velocity vectors of the two gear wheels, and PC 0 is
the distance of point P from the contact generatrix of the two operating pitch cones,
that is the instantaneous axis of rotation. Similarly to what we have already done in
Sects. 3.9 and 8.10, in the case here examined the energy lost by friction, or lost
work, on the differential segment ds in the infinitesimal time internal dt is given by:
dWl dFn
¼l dsjx2 x1 jPC 0 ; ð9:123Þ
dt ds
Pe C
PC 0 ¼ s: ð9:124Þ
Re
Substituting this last relationship into the Eq. (9.123), and assuming that
ðdFn =dsÞ is a constant along the instantaneous line of contact, we get:
Pe C
¼ sin u: ð9:126Þ
Re
On the other hand, while the two members of the straight bevel gear pair rotate
about their axes at angular velocities x1 and x2 , the instantaneous line of contact
rotates about the normal to plane of action passing through point O, with angular
velocity x1 cos a sin d1 ¼ x2 cos a sin d2 . Obviously, the instantaneous line of
contact is the common straight line along which the conical surfaces of the two
mating tooth flanks intersect the plane of action. Therefore, if u1 is the rotation
angle of the pinion about its axis (it is obtained by integrating the first Eq. (3.55) for
x1 ¼ const), the corresponding angle on the plane of action described by the
contact generatrix of conical surfaces of the mating teeth when it moves from the
position OPi Pe to position OCi Ci0 (Fig. 9.24) is given by:
404 9 Straight Bevel Gears
Developing Eq. (9.126) in Maclaurin series (see Tricomi [43]; Buzano [4];
Smirnov [38], considering only the first term of this series development, and taking
into account the Eqs. (9.127), (9.125) becomes:
dWl Re þ Ri
¼ lFn u1 cos a sin d1 jx2 x1 j : ð9:128Þ
dt 2
Now let us calculate the work done by the driving wheel. The torque resulting
from differential normal force component, dFn , applied to the differential segment,
ds, of the instantaneous line of contact, calculated with respect to the crossing point,
O, is a vector perpendicular to the plane of action, having an absolute value equal to
Fn ðRe þ Ri Þ=2. The component of this vector in the direction of the pinion axis is
given by ½Fn sin d1 ðRe þ Ri Þ=2, since the angle between the normal to the plane
of action and the pinion axis is equal to ð90 db1 Þ.
On the other hand, the torque resulting from the differential friction force
components, also calculated with respect to the crossing point, O, is a vector
perpendicular to the instantaneous line of contact, OPi Pe , having an absolute value
equal to ½lFn ðRe þ Ri Þ=2. As shown in Fig. 9.25, this torque can be resolved in
two components: the first is directed according to normal to the straight line OCi Ci0 ,
and is equal to ½lFn ðRe þ Ri Þ cos u=2; the second is directed according to the
straight line OCi Ci0 , and is equal to ½lFn ðRe þ Ri Þ sin u=2. Considering the com-
ponents of these two torque components according to the pinion axis, and taking
into account the Eq. (9.127), we get the following relationship, which expresses the
resulting torque Tl1 in the direction of the pinion axis, due to differential friction
force components:
Fig. 9.25 Torque resulting from the differential friction component dFl
9.11 Efficiency of Straight Bevel Gears 405
Re þ Ri
Tl1 ¼ lFn ½l sin d1 sin a cosðu1 cos a sin d1 Þ cos d1 sinðu1 cos a sin d1 Þ;
2
ð9:129Þ
where the plus and minus signs must be taken, depending on whether the contact
occurs during the path of approach or the path of recess.
To obtain the above torque components, we have assumed that the angular
velocities x1 and x2 are constant, that is steady state operating conditions. Now,
taking the torque components of all the forces applied to the pinion in the direction
of its axis, we get the following relationship that expresses the work done by the
driving wheel, i.e. the driving torque applied to the pinion:
Re þ Ri
T1 ¼ Fn ½sin d1 cos a
l sin d1 sin a cosðu1 cos a sin d1 Þ
2 ð9:130Þ
þ l cos d1 sinðu1 cos a sin d1 Þ:
Depending on the geometry of the straight bevel gear pair defined in Sects. 9.2
and 9.3, we can write:
Re þ Ri r1 þ r1;i Re Ri r1 r1;i
sin d1 ¼ sin d1 ¼ ; ð9:131Þ
2 2 2 2
where r1 ¼ r1;e and r1;i are respectively the radii of circles obtained by intersecting
the pitch cone of the pinion with the spheres having both centers O, and radii Re and
Ri (r1 is the already defined radius of the reference pitch circle of the pinion).
Therefore, combining the two Eqs. (9.128) and (9.130), and taking into account
the first Eq. (9.131), we get:
Assuming that the driving torque T1 is a constant, we infer that, under the
conditions considered, the work dissipated by friction, in the time corresponding to
duration of contact between the two teeth, i.e. during the complete meshing cycle, is
given by the following equation:
406 9 Straight Bevel Gears
1=2 (e =r
z21 z1 1Z 1 ðl=l0 Þu1 du1
Wl ¼ l0 T1 1 þ 2 þ 2 cos R cot d
z2 z2 0 1 þ l cos a
1
sin ðu 1 cos a sin d1 Þ tan a cosðu1 cos a sin d1 Þ
)
e2Z=r1 ðl=l0 Þu1 du1
þ cot d
0 1 þ l cos a
1
sin ðu 1 cos a sin d1 Þ þ tan a cosðu1 cos a sin d1 Þ
1=2
z2 z1
¼ l0 T1 D 1 þ 12 þ 2 cos R ;
z2 z2
ð9:134Þ
where e1 and e2 are respectively the lengths of arcs of approach and recess on the
sphere with radius Re , D briefly indicates all that the bracket contains, and l0 is the
value of coefficient of friction l corresponding to a given point of path of contact,
which can be determined conventionally. This last equation is similar to Eq. (3.98),
and reduces to it for d1 ¼ 0.
The loss of efficiency of the straight bevel gear pair examined here can be
expressed as:
1=2
Wl Wl pl0 D z21 z1
1g¼ ¼ ¼ 1 þ 2 þ 2 cos R ; ð9:135Þ
Wm T1 e1 þ e2 r1 z2 z2
r1
where Wm , given by Eq. (3.99), is the work done by the driving torque during the
same meshing cycle considered for calculation of Wl , while D is still given by
Eq. (3.101).
Equation (9.135) can be written in the form:
1=2
z2 z1
1 g ¼ 2pl0 D 1 þ 12 þ 2 cos R : ð9:136Þ
z2 z2
which coincides with Eq. (3.104) obtained for parallel cylindrical spur gear pairs.
Therefore, we can conclude that, within the approximation in which Eq. (9.137)
is valid, the efficiency of a straight bevel gear pair is given by the same
relationship
found for a parallel cylindrical spur gear pair, replacing 1
z1 z12 with
1=2
1
z2 þ 1
z2 þ z
2
z
1 2
cos R .
1 2
9.11 Efficiency of Straight Bevel Gears 407
The main methods used for the manufacture of straight bevel gears can be also
summarized up in two categories: non-generation methods or form cutting methods,
and generation methods. However, it should be noted that the cutting processes of
the straight bevel gears, and more generally the ones of the curved-toothed bevel
gears, as far as they are similar to those already described for the parallel cylindrical
spur and helical gears, have peculiar features, which deserve to be briefly high-
lighted. These peculiarities concern both gear-cutting machines and cutting tools.
In Sect. 9.4 we have already introduced the concepts underlying the template
machining process for manufacturing straight bevel gears. This process is imple-
mented by means of bevel gear planers, variously named: template-shaping
machines, tracer shaping machines, profiling shaping machines, etc. In the same
Sect. 9.4 we have also described some limitations of this non-generation cutting
method, which is still used for low production of large straight bevel gears.
Two templates are used, one for each of the two opposite flanks of two suc-
cessive teeth. From a theoretical point of view, two templates would be needed for
each gear ratio. However, for obvious reasons of economy, one template pair covers
a small range of gear ratios, so we have the same disadvantages of profile accuracy,
as we have described in Sect. 3.11.1 regarding the cutting of cylindrical spur gear
wheels by means of disk-type milling cutters. The cut is done with two simple
single-point cutting tools, each driven by the related template follower resting on a
straight guide (see [7, 12, 14]).
The tooth spaces are machined one at time. Thus, once a tooth space is finished,
the next tooth space is indexed with the indexing device and machined, and so on
until the entire tooth spaces are cut, and the teeth are formed. A dual machining
cycle is needed: in fact, tooth spaces are first roughed, using slotting tools or other
types of roughing tools, and then machined by finishing. With a proper set up,
today’s template shaping machines, which are almost all equipped with CAD/CAM
systems, allow to minimize the disadvantages described in Sect. 3.4. These systems
also allow to obtain teeth with lengthwise crowning, which is achieved by a slight
408 9 Straight Bevel Gears
motion of the tool arm as the cutting tool moves along the face width. This
crowning allows optimal location of the tooth bearing contact area.
Another form cutting method of straight bevel gears is milling. However, it is
not widely used because of the remarkable limits of accuracy that characterize it. In
fact, other specific limitations of this method are added to the above-mentioned
limits of the indexing devices. The proper technological limits are mainly because
the form-milling cutter (the disk-type cutter or end mill) has a constant profile,
while the tooth profile varies from the inner cone to the back cone. Obviously, the
milling cutter cannot have greater thickness than the tooth space at the inner cone
and, given the variability in the width of the tooth space in the direction of the face
width, it can cut only on one side.
The operation of the gear-cutting machine that implements this cutting process is
time consuming, and the cut is very approximate, so this non-generation cutting
method is sometimes used to rough straight bevel gears, which are then finished by
other methods and cutting machines. Often, to limit the aforementioned drawbacks
and have a little better-quality cut, several cutting tools are used. Nevertheless, this
method, when used to obtain straight bevel gears ready for use, is to be regarded as
a workshop gadget rather than a real cutting method.
For a more in-depth knowledge of the generation cutting methods of the straight
bevel gears, we refer the reader to specialized textbooks (see Galassini [12], Rossi
[33], Dudley [9], Townsend [42], Davis [7], Radzevich [28, 29, 32]). However, to
give an idea of these methods, we think it is useful to provide here some basic
concepts on the main generation cutting processes used in this field.
One of the traditional generation processes is the one known as planning gen-
erating method or planning generator. This method is very versatile, as it allows to
cut both straight-teeth and curved-teeth bevel gears. Moreover, it allows the cutting
of hypoid gears, with the addition of special heads to standard gear planers. These
gear planers use two planer cutters with straight cutting edges (Fig. 9.6), which
move in a reciprocating motion (a back and forth motion, with each of the two
cutters working during the return stroke of the other, and cutting off the flanks of the
same tooth). The cutters are mounted on two adjustable guides that are carried on a
cradle, which materializes the generation crown wheel and rotates about its axis,
driven by a mechanism made up of a crank and a connecting rod. Even the bevel
gear wheel being cut rotates about its stationary axis, with a rotational motion
corresponding to the rolling without sliding of its pitch cone on the pitch plane of
the virtual generation crown wheel. This motion is transmitted from the cradle to
the bevel gear wheel via a gear drive system.
In order to realize the generation cutting process, the bevel gear being cut rotates
with a constant angular velocity, synchronized with the reciprocating motion of the
cutter, and thus with the motion of the cradle, so that the cutter go right to finish in
9.12 Short Notes on Cutting Method of Straight Bevel Gears 409
the next tooth space. Therefore, at any instant, all the teeth are in the same gen-
eration stage. For straight bevel gears, the two tool-holder guides are oriented in the
radial direction. Moreover, they are adjustable, and are mounted on the cradle to
allow small oscillations of the working direction of the two cutters, for which it is
possible to obtain crowned teeth, with all the advantages already described in terms
of proper location of the tooth bearing contact area.
With this generation cutting method, large gear wheels, to be produced in small
and medium series, can be obtained. Depending on the tooth depth and shape,
several steps are required to complete the two members of a bevel gear. Generally,
the gear wheel flat blank is first roughed without generation cutting process, using
suitable roughing tools. Subsequently, the rough gear wheel is finished with at least
two finishing operations on each tooth flank, both made with a generation cutting
process. Similar machining processes are performed for cutting of the pinion, but
both roughing and finishing operations are carried out with generation cutting
processes. Of course, the generation mechanism, which synchronizes the cutters
and work-piece motions according to a definite timing, is deactivated for those
operations that are done without it.
Another traditional generation process is the one known as interlocking process,
which is realized with gear milling machines using two flat interlocking disk-type
cutters (Fig. 9.26). Therefore, due to this flat shape of the cutters, the corresponding
virtual generation crown wheel has trapeze shaped teeth, such as that shown in
Fig. 9.8b. Consequently, this process can only be used to cut straight bevel gears,
so it is much less versatile than the one described above. It is also known as
completing cutting method, as it is able to generate the teeth of the two members of
a straight bevel gear from the corresponding solid blanks in one operation.
The two milling cutters have interlocked teeth, that is their shape and size are
such that the teeth of a cutter fit the tooth space of the other. As shown in Fig. 9.26,
the cutting edges of the two cutters reproduce the tooth shape of the generation
crown wheel. They are on two planes, each of which is perpendicular to the axis of
its cutter, and both cut in the same tooth space. In order to obtain the lenghtwise
crowning, and thus to have an optimal location of the tooth bearing contact area, the
cutting edges have a concave cutting surface that removes more metal at the ends of
the tooth. The two interlocking cutters are mounted on a cradle, with their rotating
axes inclined to the platform of the same cradle. The work-piece is mounted on a
work spindle rotating about its axis at a synchronized rotational speed at an
appropriate timing rate relative to the cradle speed, to simulate the relative rolling
motion without sliding of the pitch cone of the bevel gear wheel being cut on the
cutting pitch plane of the generation crown wheel.
For reasons of brevity, we cannot describe here neither the operating modes of
the gear cutting machines using this interlocking cutting process, or the ingenious
mechanisms that implement and control the relative motion between the tool-holder
cradle and work-piece spindle. In this regard, we refer the reader to specialized
textbooks, such as those mentioned above. However, it should be noted that the
latest gear milling machines using this cutting method have CAD/CAM systems,
which allow to generate tooth flank surfaces of straight bevel gears with a com-
bination of profile and lengthwise crowning. For further information on this
interesting topic, we refer the reader to specialized textbooks and scientific literature
(see, e.g.: Stadtfeld [40], Radzevich [32], Shih and Hsieh [37]).
Finally, another generation cutting method deserves to be mentioned here, as it is
used for mass production of straight bevel gears even of considerable size and good
commercial quality at the lowest cost and in the fastest way. This method is known
as revacycle generation process. In essence, it is a generating milling method,
which uses a special cutter, which we can define as circular broach. This is because
the cutter blades, which extend radially outward from the cutter head, are dis-
tributed along the cutter circumference so that each subsequent blade is more
protruding than the preceding one, as in a traditional broach. Since each cutter blade
is progressively longer than the one before it, there is no need to feed the cutter
depth-wise into the workpiece.
With this method, straight bevel gear wheels are produced with a single com-
pleting operation. The circular broach has a remarkable diameter, since along its
circumference roughing, semi-finishing and finishing blades must be arranged
subsequently. Furthermore, between the first roughing blade and the last finishing
blade, there is an indexing gap, which also has a clearance function for automatic
loading. In addition, there is a deburring gap for collecting and expelling the tool
burrs. The cutter rotates about its vertical axis continuously and with a uniform
angular velocity. During cutting, the gear blank is kept motionless, while the cutter
is moved along a straight line parallel to the root generatrix of the gear wheel being
cut, in the direction of the face width of the latter. With this cutter motion, due to a
cam mechanism, the tooth bottom coincides with the root cone, while the desired
tooth shape is obtained as a combined effect of the cutter motion and the shape of
the cutter blades, which have concave cutting edges able to produce convex tooth
profiles.
9.12 Short Notes on Cutting Method of Straight Bevel Gears 411
Most bevel gear wheels are completed in one operation. One revolution of the
cutter completes each tooth space, and the next tooth space is indexed in the gap
between the last finishing blade and the first roughing blade (Fig. 9.27). In the few
cases where the tooth depth is too large, so the teeth could not be completed in one
cut, separated roughing and finishing operations are used. Each of these operations
is similar to the one described above, which uses a completing cutter, but requires
cutters and setups of the gear cutting machine. The first operation is roughing, and
is made with a cutter provided only with roughing blades. The second operation is
finishing, and is made with a cutter provided only with semi-finishing and finishing
blades.
Of course, the various motions described above simulate the relative rolling
motion without sliding of the pitch cone of the straight bevel gear wheel to be cut
on the pitch plane of the generation crown wheel. For reasons of brevity, here also
we cannot describe the mechanisms that implement and control this relative rolling
motion. For any further details, we refer the reader to specialized textbooks and
scientific literature (see Stadtfeld [41], Radzevich [32]).
412 9 Straight Bevel Gears
Fig. 9.28 Pinion cantilever mounted: a two preloaded tapered roller bearings, and adjustment
bush; b two preloaded tapered roller bearings, without adjustment bush; c simple supported pinion
with a cylindrical roller bearing on the left, and rigid pair consisting of two tapered roller bearings
and related adjustment device on the right; d, straddle supported pinion with a cylindrical roller
bearing on the left, and rigid pair consisting of two tapered roller bearings and related adjustment
device on the right
In the design of the straight (and curved-toothed) bevel gears, we must keep in mind
that the sizing, manufacturing, control, assembly and mounting are more difficult
than is the case of parallel cylindrical gears. Usually, the pinion is cantilever
mounted, so it is necessary to adjust its position with respect to that of the mating
gear wheel. In this regard, several construction and assembly solutions can be used,
four of which are shown in Fig. 9.28. For details of these solutions and their other
peculiarities, we refer the reader to traditional textbooks on this subject (see
Giovannozzi [13], Conti [6], Deutschman et al. [8], Chévalier [5], Niemann and
Winter [25]).
From these construction and assembly solutions, further possibilities of errors
than those associated with the parallel cylindrical gears result. They are primarily
due to the error of axial position of the pinion and wheel, as well as to the deviation
of the actual value of the shaft angle from the nominal value, especially because of
the above-mentioned cantilever mounted pinion. These errors, if not adequately
contained and controlled, can determine uneven distribution of the load along the
9.13 Construction and Assembly Solutions for Bevel Gears 413
face width, with the risk of load carrying capacity only on edges, irregular running
conditions, and, in extreme case, even interlocking of the bevel gear pair, due to the
total disappearance of the backlash.
To avoid or to mitigate these risks, we must first limit the face width, by
imposing the conditions b 0; 30Re and b 10met , and possibly use crowned teeth,
with recommended crowning depth in the range from ðb=300Þ to ðb=500Þ, for
case-hardened and grinded toothings, or equal to about ðb=1000Þ, for machining
finished toothings. The lengthwise crowning, even if does not allow to completely
remedy the misalignment error of the axes, which no longer converge at the apex,
allows to bring the contact at the middle of the face width, thus avoiding contact on
the edges. It is also appropriate that face widths of the two members of the bevel
gear pair are equal, in order to prevent that sharp edges are formed during
running-in, or due to the smoothing and polishing exerted by a case-hardened or
nitride pinion ðHRC [ 50Þ on the tooth flanks of the through hardened mating
wheel.
It is then necessary to adjust the axial positions of the pinion and wheel, so as to
get as close as possible to that of correct kinematics, which from the theoretical
point of view is only one, to work together the two members of the bevel gear, and
to mount and replace them in pairs. A complete interchange ability is only possible
with mass-produced bevel gears, and with adequate experience in the production
process and choice of materials and heat treatments. When adjusting the axial
position of the two members of the bevel gear pair, it is necessary to check that
clearance remains constant. For mounting, the shaft angle must be the same as the
design shaft angle, and the axes of the two bevel gear members must intersect.
Though small deviations from the theoretical conditions cannot be avoided in
practical applications, care must be taken to reduce these deviations to as small a
value as possible, as otherwise serious running difficulties are encountered.
To minimize the detrimental effect of the bending deflections, bearings must be
properly spaced. The spacing depends on the stiffness of the shafts carrying the gear
wheels, and the mounting type, i.e. whether the wheels are straddle mounted or
overhung mounted. It is also necessary to preload the bearings, reducing their
clearances, and so have smaller lateral forces and greater stiffness of the assembly,
to which an appropriate support sizing, perhaps with suitable ribs, makes a sig-
nificant contribution. Always to have the support maximum stiffness, and to avoid
or to limit the lateral forces, it is appropriate that the distances l1 and l2 between the
middle-planes of the bearings that support the shafts of the pinion and wheel have
suitable values, and precisely: l1 ¼ ð2:0 2:5Þdm1 for u ¼ ð3 6Þ, and l1 ¼
ð1:2 2:0Þdm1 for u ¼ ð1 6Þ, for the pinion; l2 [ 0:7dm2 , for the wheel.
Usually, pinion and its shaft make up a single piece. For pinion keyed on its
shaft, to avoid dangerous strength decreases to the tooth roots, it must be ensured
that the minimum rim thickness, sR , is at least equal to 2mn . In this way, the
maximum diameter of the shaft, on which the pinion is keyed, is determined. When
possible, to avoid the afore-mentioned errors, it is good to bear the shaft of the
pinion from both sides of the toothed part. Generally, for bevel gears with con-
verging axes, this is not possible. In these cases, in addition to reducing the face
414 9 Straight Bevel Gears
width, it is necessary that one of the two bearings that support the shaft of the
pinion cantilever mounted is as near as possible to the heel of the pinion, to limit the
bending deflection and, therefore, an uneven load distribution on the teeth (see
Giovannozzi [13].
In some cases, the pinion is supported from both sides thereof itself; this is often
possible with hypoid gear pair. In these cases, it is necessary to size the shaft ends
of the pinion in such a way that, during the generation cutting of the teeth, the
cutting tool does not interfere with the keying surface of the bearings (see Niemann
and Winter [25].
Wheels of the bevel gears are usually keyed to the shaft, and the shaft is sup-
ported by bearings located on both sides of the wheel itself. Wheels of a relatively
small diameter are made in one piece. Beyond certain diameter (800 mm), the
wheels are made of two pieces, being made up of a toothed crown ring in
case-hardened steel, having rim thickness sufficient to provide full support for the
tooth roots, and a hub in through hardened steel, joined by threaded fasteners.
Mountings must be rigid, so that the relative displacements of the two members
of the bevel gear pair under operating conditions are kept within allowable limits.
Detrimental effects of misalignment must be minimized, ensuring a proper align-
ment of the two wheels. Besides, bearing housing and mountings must be accu-
rately machined, couplings accurately mounted, and keys must be properly fitted.
References
1. Brar JS, Bansal RK (2004) A textbook of theory of machines. Laxmi Publications (P), Ltd.,
New Delhi
2. Buchanan R (1823) Practical essays on mill work and other machinary, with notes and
additional articles, containing new researches on various mechanical subjects by Thomas
Tredgold. J Taylor, London
3. Buckingham E (1949) Analytical mechanics of gears. McGraw-Hill Book Company Inc, New
York
4. Buzano P (1961) Lezioni di analisi matematica, 5th edn. Libreria Editrice Universitaria
Levrotto&Bella, Torino
5. Chévalier A (1983) Manuale del Disegno Tecnico. In: Chirone E, and Vullo V (Eds) Revised
and expanded Italian. Società Editrice Internazionale, Torino
6. Conti G (1969) I Cuscinetti a Rotolamento, vol. I and vol. II, 4th. Editore Ulrico Hoepli,
Milano
7. Davis JR (Ed) (2005) Gear materials, properties, and manufacture. J.R. Davis, Davis &
Associates, ASM International, Materials Park, OH, USA
8. Deutschman AD, Michels WJ, Wilson C (1975) Machine design: theory and practice.
Macmillan Publishing Co., Inc, New York
9. Dudley DW (1984) Handbook of practical gear design. McGraw-Hill Book Company, New
York
10. Ferrari C, Romiti A (1966) Meccanica applicata alle macchine. Unione Tipografica-Editrice
Torinese (UTET), Torino
11. Fuentes A, Gonzalez I, Pasapula HK (2017) Computerized design of straight bevel gears with
optimized profiles for forging molding, or 3D printing. Thermal processing magazine, mar
References 415
40. Stadtfeld HJ (2007) Straight bevel gears on phoenics R-machines using coniflex R-tools. The
Gleason Works, Rochester, New York
41. Stadtfeld HJ (2010) Coniflex plus straight bevel gears manufacturing. Gear Solutions,
pp 44–55
42. Townsend DP (1991) Dudley’s gear handbook: the design, manufacture and application of
gears. McGraw-Hill, New York
43. Tricomi FG (1956) Lezioni di Analisi Matematica, Parte Seconda, CEDAM— Padova: Casa
Editrice Dott. Antonio Milani
Chapter 10
Crossed Helical Gears
Abstract In this chapter, the fundamentals of general rigid kinematic pairs are first
described and the hyperboloid pitch surfaces are defined. These concepts are then
applied to the generation of cylindrical crossed helical gear pairs. The main geo-
metrical quantities of these types of gears are then defined and their kinematic
quantities are determined, particularly those correlated with the lengthwise sliding
and sliding velocity. Subsequently, the load analysis of this gears is performed and
the thrust characteristic on shafts and bearings are defined. Other important kine-
matic quantities related to the rolling and sliding motions of the mating teeth
surfaces are then defined and the instantaneous and average efficiencies of these
gears are analytically determined. Finally, short notes are given regarding the
fundamentals of profile-shifted toothing for crossed helical gears.
By definition, crossed helical gears are cylindrical gear pairs consisting of mating
helical members with crossed axes. However, the tooth flank surfaces of the parallel
cylindrical involute helical gears are in line contact, while those of the crossed
involute helical gears are in point contact. In this chapter, we will discuss only
cylindrical involute helical gears with crossed axes, without considering worm
gears, which are a special case of crossed helical gears. Worm gears will be dis-
cussed in the next chapter.
The discussion is here carried out in accordance with the classical one, available
in more or less extensive form in traditional textbooks (see Buckingham [2],
Colbourne [4], Dobrovolski et al. [5], Dudley [6], Giovannozzi [9], Henriot [10],
Litvin and Fuentes [13], Merritt [14], Niemann and Winter [17], Pollone [18],
Radzevich [20], Townsend [27]). However, to better understand the kinematics of
these gear pairs, we think it is appropriate to first introduce two introductory sec-
tions to the topic that interest us. Therefore, in this perspective, this and the next
section deal with the general rigid kinematic pairs and hyperboloid gear pairs,
respectively.
(a)
(b)
Fig. 10.1 Rigid kinematic pair, and Mozzi’s axis, in two orthogonal projections
projections of m-axis respectively in the first and second projection planes. Since
the vector of the relative angular velocity xr1;2 ¼ ðx1 x2 Þ is directed along the
Mozzi’s axis, if we compose, in the first projection plane (Fig. 10.1b), the vector x1
with the vector x2 , we get the direction of m0 , and then the angles b1 and b2 .
Therefore, we have:
x1 sin b2
¼ : ð10:1Þ
x2 sin b1
allows us to determine the angles b1 and b2 , as we have already seen in the previous
chapter, regarding the straight bevel gears, which are gears with intersecting axes.
420 10 Crossed Helical Gears
To determine the distances r1 and r2 , let us consider point O through which the
m-axis passes, as well as the velocity vectors v1 and v2 of the same point O, whose
absolute values are given respectively by v1 ¼ x1 r1 and v2 ¼ x2 r2 . Then assume
that this point O belongs once to the ruled surface having axis 1, and once to the
ruled surface having axis 2. These vectors are respectively normal to the directions
of 1 and 2, while the relative velocity vector vr1;2 ¼ ðv1 v2 Þ must be parallel to m-
axis. Therefore, if in the first projection we bring, starting from point O0 , the vector
v1 ¼ O0 A0 , normal to 10 , and the vector v2 ¼ O0 B0 , normal to 20 , the vector
vr1;2 ¼ ðv1 v2 Þ ¼ B0 A0 , must be parallel to m0 . It follows that
allows us to determine r1 and r2 . It is to be noted that, as Eq. (10.5) shows, the
Mozzi’s axis divides the shortest distance, a, in parts that are directly proportional
to the trigonometric tangents of angles b1 and b2 , which the xr -vector forms
respectively with the rotational axes of the two bodies under consideration.
The relative velocity vector vr1;2 ¼ ðv1 v2 Þ ¼ B0 A0 , which is notoriously the
sliding velocity of the points lying on Mozzi’s axis, has an absolute value equal to:
Therefore, the ruled pitch surfaces, i.e. the axodes, and the characteristics of the
relative motion, xr and vr , are thus completely defined. It should be noted that there
are no points in which the relative velocity is zero. The minimum relative velocity
possible is the one calculated above, and refers to points of Mozzi’s axis. The points
located outside of this axis have a greater relative velocity, which increases with
their distance from this axis.
10.2 Hyperboloid Pitch Surfaces 421
Let us now consider a hyperboloid gear pair. If the transmission ratio i ¼ x1 =x2 is
a constant, from Eqs. (10.1) and (10.4) we infer that the values of the angles b1
and b2 , and distances r1 and r2 are also constant. Therefore, the ruled pitch
surfaces become two double ruled hyperboloids of one sheet, also called one-
sheeted double ruled hyperboloids, of which r1 and r2 are the radii of their
symmetry cross sections or throat cross sections (Fig. 10.2). The ruled pitch
surface, which constitutes the axode of gear wheel 1, is the locus of the instan-
taneous axis of screw motion, m, which forms, with the direction of the axis 1, an
angle b1 constant, and leans on a circumference of radius r1 lying in the plane
normal to the axis 1, and passing through the shortest distance, a. The same
definition applies, mutatis mutandis, for the ruled pitch surface, which constitutes
the axode of the gear wheel 2.
By analogy with the other types of gear pairs, and assuming that the gear wheel 1
is the driving one, the absolute value of the transmission ratio i ¼ x1 =x2 , taking
into account Eq. (10.1), will be given by:
x1 sin b2
i¼ ¼ : ð10:9Þ
x2 sin b1
and the second has as cathets O4 P0 ¼ l sin b2 and P0 S equal to the throat radius r2 ,
0
(a)
(b)
Fig. 10.3 Hyperboloid gear pair, and Mozzi’s axis, in two orthogonal projections
10.2 Hyperboloid Pitch Surfaces 423
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi
2 2
r1P ¼ r1 þ l sin b1 2 r2P ¼ r2 þ l sin b2 :2 ð10:10Þ
In the first projection (Fig. 10.3b), the projection of the two circles having radii
r1P and r2P are represented by the two segments H 0 K 0 ¼ 2r1P
and L0 M 0 ¼ 2r2P
,
0 0 0
respectively drawn from point P normally to the projected axes 1 and 2 . Also in
the first projection, the generatrices of the two hyperboloids are projected in the
0
hyperbolas D0 H 0 , E 0 K 0 and F L0 , G0 M 0 , which can be drawn point by point, per-
forming the intersections of the two hyperboloids with the planes, normal to the
axes, conducted through other points of the m-axis, as we did for point P. The first
hyperbola has, as asymptotes, the straight line m0 and its symmetrical line with
respect to 10 , while the second hyperbola has, as asymptotes, the straight line m0 and
its symmetrical line with respect to 20 .
If we refer the hyperbola D0 H 0 , E 0 K 0 to a Cartesian coordinate system having its
origin at the point O0 , the abscissa axis x O0 D0 , and ordinate axis y 10 , we infer
that the coordinates of one of its current point H 0 , corresponding to the section at a
distance OP ¼ l from the straight line of shortest distance, are [29]:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi
x¼ r1 þ l2 sin2 b1 y ¼ l cos b1 : ð10:11Þ
r1 1
¼ tan2 b1 ¼ 2 : ð10:14Þ
r2 i
This last equation shows that, in the case where R ¼ p=2, the transmission ratio
i ¼ x1 =x2 is equal to the square root of the ratio r2 =r1 . The same equation shows
that, with a small transmission ratio, the radius of the throat cross section of the
greater gear wheel may be much larger than that of the other gear wheel. For
example, with i ¼ 4, we would have r2 ¼ 16r1 . Moreover, always in the case
where R ¼ p=2, taking into account Eqs. (10.13) and (10.14), one of the following
relationships can express the absolute value of the relative velocity along the
Mozzi’s axis, given by Eq. (10.7):
424 10 Crossed Helical Gears
x1 r1 x2 r2
vr ¼ x1 r1 sin b1 1 þ cot2 b1 ¼ 2
¼ x2 r2 sin b2 1 þ cot b2 ¼ :
sin b1 sin b2
ð10:15Þ
In the general case where R 6¼ p=2, if appropriate parts of the two pitch
hyperboloids are provided with teeth, we get a hyperboloid gear pair. These toothed
parts can be delimited by sections normal to the axis of rotation, located on opposite
sides or the same side with respect to the throat cross section. In the first case, we
have the so-called throat hyperboloid wheels (Fig. 10.4a) which, for limited values
of the length of the toothed part, are approximately cylindrical, and therefore
resemble cylindrical helical wheels. In the second case, we have the pseudo-bevel
hyperboloid wheels (Fig. 10.4b), so named for their resemblance to the helical and
spiral bevel wheels, the characteristics of which tend to approach the more, the
greater the distance of the sections which delimit the toothed part from the throat
cross section.
The throat hyperboloid wheels are defined by the face width (the width over the
toothed part of the wheel, measured along a hyperboloid reference generatrix), and
the values of the radii r1 and r2 , and angles b1 and b2 , which are calculated using
the relationships above. These relationships cannot however be applied for the
pseudo-bevel hyperboloid wheels, since in these wheels it is not detectable any pair
of corresponding circular cross sections whose radii are given by the above
equations. However, it is possible to show that the ratio between the two radii tends
to the transmission ratio (as in the cylindrical gears, and in the bevel gears), when
the distance of the corresponding cross sections from the throat cross sections of the
two hyperboloids tends to infinity.
The determination of the surfaces delimiting the toothing of a hyperboloid wheel
is carried out using the known general rules of Applied Mechanics, concerning the
analytical definition of the mating surface of a general rigid kinematic pair.
(a) (b)
Fig. 10.4 Hyperboloid gear pair: a throat hyperboloid gear pair; b pseudo-bevel hyperboloid gear
pair
10.2 Hyperboloid Pitch Surfaces 425
The practical solution of this problem, however, gives rise to considerable tech-
nological difficulties, due mainly to the impossibility to identify a particular
hyperboloid wheel that, for the geometric simplicity of the active surfaces of its
teeth, can be assumed as generation wheel capable of generating, by a generation
cutting process, the hyperboloid mating wheel. This fact, entirely new, does not
occur for the cylindrical and bevel wheels for which, as we have seen in previous
chapters, it was possible to identify respectively the reference basic rack, and the
reference crown wheel.
It is to be noted that the hyperboloid wheels have theoretically two main
advantages:
• the contacts between the conjugate surfaces of the meshing teeth are line
contacts;
• in such contacts, when they occur on the hyperboloid pitch generatrices and, by
extension, in their immediate vicinity, the minimum sliding velocities occur.
However, because of the above-mentioned technological difficulties, which arise for
a correct cutting of the teeth, their use is very limited. In practice, it is preferred to
forego the aforementioned advantages, in favor of a simpler and more economical
cutting process.
We can realize the above-mentioned impossibility of identifying a particular
hyperboloid wheel to be used as a generation gear wheel, considering the special
case of hyperboloid gear pair in which the shaft angle is greater than
p=2 ðR [ p=2Þ, and b2 ¼ p=2. This special case is shown in Fig. 10.5, which
highlights these two angles in the side view (Fig. 10.5a). For b2 ¼ p=2, from
Eq. (10.9) we infer sin b1 ¼ x2 =x1 , so from Eqs. (10.5) and (10.6) we get r1 ¼ 0,
and r2 ¼ a. Therefore, the Mozzi’s axis, m, lies on the plane perpendicular to the
shortest distance straight line containing the 1-axis, intersects the latter at point
P (O1 P ¼ 0, and O2 P ¼ a), and is normal to the projection on this plane of the
2-axis, so m0 is normal to 20 .
Rebus sic stantibus, it follows that the ruled pitch surface of the wheel 1
degenerates into the cone with apex P and cone angle b1 ¼ ðR p=2Þ, while the
pitch cone surface of the wheel 2 degenerates into a ruled plane. This plane is
formed by the straight lines that are tangent to the circle having as its center P (these
tangents are nothing more than the successive positions of the Mozzi’s axis, m), and
radius equal to the shortest distance, a. The second wheel is the hyperboloid crown
wheel. Figure 10.5b shows what we have just said, in the projection made in the
direction of the 2-axis.
The transmission
ratio of this special hyperboloid gear pair
is i ¼ ðx1 =x2 Þ ¼ 1= sin b1 , while in accordance with Eq. (10.7), the absolute
value of the sliding velocity is given by vr ¼ x2 a:
Assume now the ruled pitch surfaces defined above are provided with toothing.
We have so a power transmission between crossed axes with a gear pair constituted
by a straight bevel wheel and a crown wheel with curved teeth, whose intersections
with the reference plane of the same crown wheel are straight lines that are tangent
to the circle defined above, as Fig. 10.5b shows. However, it should be noted that it
426 10 Crossed Helical Gears
(a)
(b)
Fig. 10.5 Hyperboloid gear pair with R [ p=2 and b2 ¼ p=2: a projection in the side view;
b projection in the direction of the 2-axis
is impossible to use this gear pair for transmitting motion between orthogonal
crossed axes. This is because the shaft angle, R, between the crossed axes 1 and 2
cannot be equal to p=2, because in that case we should have R ¼ p=2 and b1 ¼ 0. It
follows the impossibility of identifying a particular hyperboloid crown wheel to be
used as a generation gear wheel for hyperboloid gear pairs.
10.3 Generation of a Crossed Helical Gear Pair 427
R ¼ b1 b 2 ; ð10:16Þ
where the plus and minus signs are respectively for helices that have the same
direction (i.e., both right-hand helices or both left-hand helices), and for helices that
have opposite directions (i.e., one right-hand helix and the other left-hand helix or
vice versa). Four different combinations are therefore possible, and precisely:
• Right-hand driving wheel meshing with right-hand driven wheel.
• Left-hand driving wheel meshing with left-hand driven wheel.
• Right-hand driving wheel meshing with left-hand driven wheel.
• Left-hand driving wheel meshing with right-hand driven wheel.
In most cases, the helices have the same direction, and the driving wheel is the one
with greater helix angle. Only for small values of the shaft angle, the helices have
opposite directions. The direction of rotation of the driven wheel depends on the
helix direction, and the direction of rotation of the driving wheel.
In the usual practical applications, the crossed axes are orthogonal ðR ¼ p=2Þ.
Contrary to what happens in parallel cylindrical helical gears, in which the helix
angles are equal, but of opposite directions, in the crossed helical gears the helix
angles are generally different from each other, and the wheel with greater helix
angle is the driving wheel. From a theoretical point of view, the teeth of a crossed
helical gear have a point contact between the mating tooth flank surfaces.
Actually, due to the Hertzian contact deformations and wear after some time of
operation, the theoretical point of contact develops into an ellipse of contact, which
extends in length, constituting almost a line contact. However, despite this positive
extension of the area of contact, these gears are used only for transmission of small
loads. They therefore have a limited use, and are mainly used, for example, as
secondary drives for textile machines, speedometer drives, pump drives, instru-
mentation, distributor drives of internal combustion engines, and other similar
applications.
428 10 Crossed Helical Gears
When the load to be transmitted is high, due to sliding, these gears will wear out
quickly, even if the rotational speed is not high. If then also the rotational speed is
high, the scuffing becomes the main cause of deterioration of their teeth.
If the shaft angle increases, the sliding velocity along the tooth trace increases as
well, because we have here a screw sliding. Compared to a parallel cylindrical
helical gear, where sliding is only the one in the direction of the tooth depth, we
have here the worst sliding conditions, as the lengthwise or longitudinal sliding in
the direction of the tooth axis is superimposed to the profile sliding in the direction
of the tooth depth. With small values of the shaft angle ðR\25 Þ, due to the
increasing wear, the ellipse of contact becomes increasingly extensive, thereby
enhancing the load carrying capacity appreciably.
As for a general rigid kinematic pair or a hyperboloid gear pair, for the gener-
ation of a crossed helical gear pair we can consider the same two orthogonal
projections used in Sects. 10.1 and 10.2 (Fig. 10.6). Also, in this case we indicate
with: 1 and 2, the skew shaft axes; R, the shaft angle; a ¼ O1 O2 , the center distance
Fig. 10.6 Crossed external gear pair, and instantaneous axis of rotation, in two orthogonal
projections
10.3 Generation of a Crossed Helical Gear Pair 429
or shortest distance or minimum distance between the two axes; x1 and x2 the
angular velocity vectors. If we choose an arbitrary point C of the center distance
a ¼ O1 O2 (this point is generally distinct from the point O through which the
Mozzi’s axis passes), and denote with r1 ¼ CO1 and r2 ¼ CO2 the distances of the
point C from the two axes, we can write:
a ¼ r1 þ r2 : ð10:17Þ
Assume the cylinders having axes 1 and 2, and radii r1 and r2 , as pseudo-pitch
cylinders of the two wheels. Strictly, they are not pitch surfaces. In this regard, we
remember in fact that, from a theoretical point of view, the pitch surfaces, i.e. the
axodes, are the loci of the instantaneous axes of rotation, and their cross sections are
the centrodes, i.e. the loci of the instantaneous centers of rotation. For this reason,
the cylinders above are called pseudo-pitch surfaces. These pseudo-pitch surfaces
are used here as pitch surfaces of the corresponding toothing, thus fulfilling the
function that the cylindrical pitch surfaces had in parallel cylindrical helical gears.
We will clarify this concept soon. In any case, hereafter, when we talk of pitch
surfaces of the crossed helical gears, we refer to these pseudo-pitch surfaces.
Let us now consider the common P-plane (Fig. 10.7), which is tangent to the
aforementioned cylinders (and therefore passing through the point C and normal to
the center distance) and, on it, the straight line n, having the direction of the relative
velocity vector of point C, vr ¼ ðv1 v2 Þ. If v1 ¼ x1 r1 and v2 ¼ x2 r2 are the
absolute values of the velocities of point C, considered as belonging respectively to
the first and second cylinder, the vector v1 ¼ C 0 A0 (see Fig. 10.6) will be normal to
the axis 10 , and the vector v2 ¼ C 0 B0 will be normal to the axis 20 , for which the
vector of the aforementioned relative velocity will be vr ¼ ðv1 v2 Þ ¼ B0 A0 .
Fig. 10.7 Pseudo-pitch cylinders, and generation straight line of the pitch helices
430 10 Crossed Helical Gears
Therefore, the first projection n0 of n will pass through the point C 0 and shall be
parallel to B0 A0 , for which the angles b1 and b2 between n0 and 10 and, respectively,
between n0 and 20 will have to satisfy the following condition:
x1 r2 cos b2 r2 sin c2
i¼ ¼ ¼ : ð10:19Þ
x2 r1 cos b1 r1 sin c1
The one or the other of the two equations above, together with Eq. (10.16),
allow determining the position of the straight line, n, which satisfies the condition
imposed. Wrapping the P-plane containing the straight line, n (Fig. 10.7),
respectively on the first and second of the two aforementioned cylinders, we will
obtain, on the first cylinder, the helix e1 with helix angle, b1 , and, on the second
cylinder, the helix e2 with helix angle, b2 . The helices e1 and e2 thus obtained, both
corresponding to the straight line, n, are taken as tooth traces respectively of the
wheel with axis 1 and of the one with axis 2. In the case shown in Fig. 10.7, they
are both left-hand helices (of course, they could be both right-hand helices), for
which Eq. (10.16) must be considered with the plus sign.
It should be noted that the difference between the crossed helical gears and
hyperboloid gears lies in the aforementioned generation method of tooth traces.
When the point C arbitrary chosen coincides with the point O considered in
Sects. 10.1 and 10.2 (in this case we have r1 ¼ r1 and r2 ¼ r2 ), the straight line, n,
coincides with the Mozzi’s axis, m. Also, in this particular case, a substantial
difference exists between the hyperboloid wheels, whose tooth traces coincide with
the successive positions of m, and the crossed helical wheels that we are studying,
which have, as tooth traces, the helices generated by the straight line, n, during the
rotation of the plane that contains it on the two cylinders defined above.
In addition, it should be noted that, in the special case in which the axes 1 and 2
are parallel ðR ¼ 0Þ, would be b1 ¼ b2 , and therefore the two helices would have
equal helix angles, but opposite directions, in accordance with what we said about
the parallel cylindrical helical gears. Finally, it should be noted explicitly the reason
for which we speak improperly of pitch surfaces, using the synonym terms im-
proper pitch surfaces or pseudo-pitch surfaces. In fact, the cylinders in contact at
point C, on which we have based our considerations, are not pitch cylinders, since
the instantaneous relative motion between them is not a pure rolling motion, but an
instantaneous relative helical motion, i.e. a screw motion.
For generation of teeth of the two wheels, consider the common mating rack
having, as pitch plane, the P-plane (Fig. 10.7), tooth traces parallel to the straight
line, n, normal pitch pn0 ¼ pmn0 , normal module mn0 , and pressure angle an0 . The
toothing of the common mating rack will be so completely defined, once the sizing
(for example, the standard or nominal sizing) has been chosen. Consider then the
first of the two wheels of the crossed helical gear pair, i.e. the pinion 1, having the
10.3 Generation of a Crossed Helical Gear Pair 431
cylinder with axis 1 and radius r1 as pitch surface, which meshes with the mating
rack. This wheel will show, on its pitch surface, a succession of tooth traces
consisting of helices e1 having normal pitch pn1 ¼ pn0 , and sizing based on the
normal module mn0 , while its transverse pitch is given by:
pn1
pt1 ¼ ; ð10:20Þ
cos b1
It is to be noted that, when we consider the gear pair constituted by the pinion to
be cut and the rack-type cutter, the cylinder of radius r1 , in effect, is the pitch
surface, as it meets the theoretical definition by us previously callback. The same
thing we can say when we consider the gear pair constituted by the wheel to be cut
and the rack-type cutter. This second wheel of the crossed helical gear pair, i.e. the
wheel 2, has as pitch surface the cylinder with axis 2 and radius r2 . This wheel
meshes with the mating rack, and shows on its pitch surface a succession of tooth
traces consisting of helices e2 having normal pitch pn2 ¼ pn0 , sizing based on the
normal module mn0 , and transverse pitch given by:
pn2
pt2 ¼ ; ð10:22Þ
cos b2
Since the pinion and the wheel so obtained are conjugate with the common
mating rack, they will be conjugate between them. Figure 10.8 shows a crossed
helical gear pair with R\p=2, in the projection on a plane normal to the shortest
distance straight line. The ranks of tooth traces, consisting of right-hand helices e1
and e2 , which appear in this figure, are those that are found on the front halves of
the cylinders having radii r1 and r2 , for which the point C of contact between the
two cylinders and the straight line, n, that generates these helices, are hidden to the
observer.
Let us denote, as usual, with rb1 and rb2 the radii of the base cylinders, and with bb1
and bb2 the base helix angles, and show first the following properties, quite unique,
of the involute helical gears with crossed axes. One of these properties consists in
the fact that, by varying in any way the relative position between the transmission
432 10 Crossed Helical Gears
axes 1 and 2, and thus causing them to translate or rotate relative to one another, the
transmission ratio, i, remains constant and is equal to (see Giovannozzi [9],
Herrmann [11], Schiebel [21]):
Obviously, this property is valid within the limits allowed by the tooth depth, so
that the mating teeth can touch each other. In this framework, we consider any point
of contact, P, between the active tooth flank surfaces of the two wheels, and the two
planes tangent to the base cylinders passing through point P (to be exact, one of the
two possible planes tangent to each base cylinder, the one that, by wrapping on the
base cylinder, generated the tooth surface). As we saw in Chap. 8, both of these
10.4 Fundamental Kinematic Properties 433
planes contain the common normal to the tooth surfaces through point P, and
therefore this normal cannot be anything other than the intersection of the two
aforementioned planes.
The transmission ratio is defined by the equality of the components v1Pn and v2Pn
of the velocity vectors v1P and v2P at point P in the direction of the common normal.
It is to be noted that the third subscript, n, indicates the direction of this normal. In
the plane of action passing through the point P and tangent to the base cylinder
having radius rb1 (see Fig. 10.9, where this plane is revolved in the relevant view),
the component of the velocity at the point P, supposed belonging to the corre-
sponding wheel, is equal to v1P ¼ rb1 x1 ¼ v2P ¼ rb2 x2 and it is normal to the
generatrix of contact of the plane considered with the base cylinder.
Recalling that the normal, n, to the helicoid lies in the same plane and is normal
to the generation straight line, which is inclined by the helix angle, bb , with respect
to the axis of the wheel under consideration, for the first and second wheel we can
write, respectively, the following relationships:
From the condition v1Pn ¼ v2Pn follows Eq. (10.24), which is therefore
demonstrated, whatever the relative position of the axes of the two wheels, since the
reasoning applies whatever the point P, and a point of contact is certainly possible
to find, however the axes of the two wheels are arranged. Of course, in relation to
the radii of the base cylinders and tooth depth, there are limits to the displacements
of the axes, on which it is not the case here to dwell upon.
For a given position of the axes of the two wheels, the locus of the points of
contact is a straight line (the line of action), and precisely the common normal to the
tooth surfaces, in their point of contact. If the wheels have constant angular
velocities, the instantaneous point of contact moves, on this line of action, with a
constant velocity. This line of action is tangent to the base cylinders as well as to the
base helices of the two wheels. Of course, we can have two different lines of action,
each of which corresponds to the meshing of the respective sides of the tooth
surfaces. In the purely theoretical case of an involute crossed helical gear pair
without errors, and with infinitely high tooth stiffness, these two lines of action
intersect each other at a point belonging to the shortest center distance, a. The
design of an involute helical gear pair with crossed axes that satisfies these theo-
retical conditions is called canonical design, and the gear is called canonical
crossed helical gear.
This theoretical behavior is immediately evident in that, for the common mating
rack of the two wheels, the planes of contact, or planes of action, related to the two
rack-wheel pairs are the planes passing through the point of contact P and tangent
to the base cylinders. Moreover, the intersection of these planes is the common
434 10 Crossed Helical Gears
Fig. 10.9 Plane of contact revolved in the relevant view, and components of the velocity of the
point P
normal to the two helicoids at point P. If we disregard friction, for a given position
of the axes of the two wheels, the line of application of the force that the teeth
transmit remains therefore in a fixed position in three-dimensional space, and it is
the same line of action.
10.4 Fundamental Kinematic Properties 435
Given the property expressed by Eq. (10.24), and contrary to what is sometimes
affirmed, the kinematics of the crossed helical gear pairs is always theoretically
correct, even if the relative displacements between the axes occur. The mounting
tolerances, from this point of view, can be very large. A translation without rotation
of the axes of the wheels does move in three-dimensional space, parallel to itself,
the line of action. A rotation without translation of the same axes instead does
generally translate and rotate the line of action.
Since the contact is a point contact, the crossed helical gears are more insensitive
to small errors in the helix angles than the parallel cylindrical helical gears. Instead,
they are sensitive to errors in center distance as, in this case, the sum of the helix
angles on the pitch cylinders is not equal to the shaft angle. In addition, the wheels
of a crossed helical gear can be moved in the axial direction without affecting the
tooth meshing, provided the face width is sufficient. Finally, it is to be noted that the
crossed helical gears, such as the parallel cylindrical helical gears, have a very good
and silent operation, and assure a remarkable regularity of the transmission of
motion. In addition, because they allow a wide possibility of choice of the radii of
their pitch cylinders for the same transmission ratio, it is possible to realize wheels
having limited overall dimensions.
Unfortunately, these theoretical kinematic conditions are never met in practical
applications. It is therefore necessary to deal with the misalignment problems,
always present. When the rules of canonical design are not satisfied, the two
above-mentioned lines of action do not intersect each other, but they are crossed
lines, even if each of them is still a tangent to both base cylinders and base helices.
The crossing of the two possible lines of action is the result of an error DR of the
nominal value of the shaft angle, R, or the result of an error Da of center distance,
a. The combined effects of these two errors are included in the term misalignment.
Errors of alignment, i.e. the changes of shaft angle and center distance, cause the
shift of the line of action. Depending on the amount of this shift, various effects may
occur, the main of which are:
• Shift of the bearing contact area far from the mid-face width of the two members
of the crossed helical gear pair.
• Risk of edge contact, when this shift exceeds a threshold value; if sufficient face
width of gears is not provided, edge contact is inevitable.
• High levels of vibrations and noise, increasing with the increase of this shift.
Despite the studies and research on this subject, controlling the misalignment of
involute helical gear pairs with crossed axes still constitutes a concern of designers
and manufacturers. To correct the defects arising from misalignment, some modi-
fications of the gear geometry have been proposed, such as:
• Modification of the helix angle of the pinion, which, however, weighs down the
manufacturing costs, since it involves the regrinding.
• Teeth of the two members of the gear pair characterized by tip relief and end
relief, determined according to the experience of manufacturers.
• Use of nonstandard crossed helical gears, i.e. with profile-shifted toothing.
436 10 Crossed Helical Gears
More recently, Litvin and Fuentes [13] have proposed the use of gear pairs, con-
stituted of standard gear wheels to be coupled to pinions characterized by
profile-crowning tooth surfaces. Subsequently, the same authors proposed the use
of pinions having tooth surfaces characterized not only by profile crowning, but
also by lengthwise crowning.
Here we do not dwell on this subject, even if it is very interesting. In this regard,
we refer the reader to more specialized textbooks. The sections that follow therefore
concern only the canonical design of involute helical gears with crossed axes. In
any case, it should be noted the fact that the effects caused by errors DR and Da, in
terms of shift of the line of action, increase almost exponentially with decreasing
shaft angle. From this point of view, the involute cylindrical helical gears with
parallel axes are obviously those more sensitive to the above-mentioned errors.
We assume the canonical design conditions, for which the line of action intersects
the nominal center distance O1 O2 between the axes of the two wheels, as shown in
Fig. 10.10. This assumption would not be necessary, and in any case it does not
affect the equations already written and those we will write below. In practice,
however, the canonical determination of the characteristic quantities of the crossed
helical gear pairs is performed under the assumption that, on the nominal center
distance O1 O2 , a point of contact C exists. In this case, the helices belonging to the
tooth flank surfaces and passing through point C are the pitch helices, and they are
certainly tangent to each other at point C.
Let us consider two external cylindrical helical wheels with z1 and z2 teeth,
having toothing with the same normal module mn , equal nominal pressure angle an
of the normal profiles, pitch helix angles b1 and b2 , and pitch helices of the same
direction. These wheels can work properly when they mesh mounted on crossed
axes, with shaft angle R equal to the sum ðb1 þ b2 Þ of the pitch helix angles, if the
conditions described in Sects. 10.3 and 10.4 are met. The relationship that corre-
lates the pitch angles b1 and b2 , and the shaft angle R is Eq. (10.16), taken with the
plus sign.
In the normal section passing through point C, which is perpendicular to the
pitch helices, the normal pitch must be, with reference to the common mating rack,
the same for the two wheels, for which we can write:
Fig. 10.10 Point of contact on the nominal center distance, and tooth force F resolved into its
three components, Ft , Fa , and Fr
438 10 Crossed Helical Gears
the radii of the pitch cylinders of the two wheels. The transmission ratio will be
defined by the condition that the components of the velocity of point C, supposed
belonging to two wheels, are equal in the direction normal to the common tangent
to the pitch helices at point C, for which Eq. (10.18) must be valid, and therefore
Eq. (10.19). This latter shows that, in general, the transmission ratio is not equal to
the ratio of the radii of the pitch cylinders. Therefore, we can choose the values of
these radii depending on the helix angles b1 and b2 ¼ ðR b1 Þ:
Since, according to Eqs. (10.27) and (10.28), we can write
pn z 1 pn z 2
r1 cos b1 ¼ r2 cos b2 ¼ ; ð10:29Þ
2p 2p
Taking into account Eqs. (10.28), the center distance, a, is given by:
mn z1 z2 1
a ¼ r1 þ r2 ¼ þ ¼ ðz1 mt1 þ z2 mt2 Þ; ð10:31Þ
2 cos b1 cos b2 2
where mt1 ¼ mn =cosb1 and mt2 ¼ mn =cosb2 are the transverse modules of the two
wheels.
Again, with reference to the common mating rack, the transverse pressure an-
gles, at1 and at2 , are related to the normal pressure angle an by means of the
relationships:
tan an tan an
tan at1 ¼ tan at2 ¼ : ð10:32Þ
cos b1 cos b2
• Finally, once we have chosen an , from Eq. (10.32) we get at1 and at2 .
Of course, this geometric-kinematic calculation must be followed and supple-
mented by the strength calculations, and load carrying capacity verifications. These
calculations and verifications may induce the designer to vary the module, mn , and
then the center distance, a. Obviously, during the course of design calculations,
seen in their general framework together, it is possible that the designer encounters
other problems that can be solved by using appropriately the afore-mentioned
relationships.
Finally, the following correlations allow us to calculate the quantities relating to
the base cylinders, once those relating to the pitch cylinders are known:
tan bb1 ¼ tan b1 cos at1 tan bb2 ¼ tan b2 cos at2 : ð10:34Þ
Now we focus our attention on Eqs. (10.19) and (10.30), as well as on the
following relationship:
pt1 r1 z2 cos b2
¼ ¼ ; ð10:35Þ
pt2 r2 z1 cos b1
If, for transmitting motion between orthogonal crossed axes, we make use of
crossed helical gear pairs, since b2 ¼ ½ðp=2Þ b1 ; from Eq. (10.19) we obtain:
r2
i¼ tan b1 ; ð10:36Þ
r1
so it is possible to make gear wheels having pseudo-pitch cylinders with radii r1 and
r2 arbitrarily chosen (possibly the same), provided that the helix angles b1 and
b2 ¼ ½ðp=2Þ b1 of the tooth traces meet Eq. (10.36).
Fig. 10.11 Pitch helix angles b1 and b2 of crossed helical gears with shaft angle R ¼ p=2, and
r1 ¼ r2 , as a function of the gear ratio u ¼ z2 =z1 ¼ i:
10.5 Determination of Other Characteristic Quantities … 441
Figure 10.11 shows the change of the pitch helix angles of the driving and
driven wheels (respectively, b1 and b2 ), for crossed helical gear pairs with r1 ¼ r2
and shaft angle R ¼ p=2, as a function of the gear ratio u ¼ z2 =z1 ¼ i. It highlights
the fact that, in the special case in which u ¼ i ¼ 1, the helix angles are equal
between them and equal to p=4 ðb1 ¼ b2 ¼ p=4Þ: In this case, Eq. (10.36)
becomes:
r2
i¼ : ð10:37Þ
r1
This equation, which coincides with the fundamental relationship that expresses
the transmission ratio of the parallel cylindrical gears and bevel gears, shows that
only if b1 ¼ b2 ¼ p=4 the transmission ratio of the crossed helical gear pairs is
equal to the ratio between the radii of the pseudo-pitch cylinders.
Example : The input data of this example, that represents a case arising frequently
in practical applications, are the following: shaft angle R ¼ p=2; transmission ratio
i ¼ u ¼ 2; 0; r1 ¼ r2 . Results: from Eq. (10.36), we obtain cot b1 ¼ 0:5; therefore,
we have b1 ¼ 63 260 and b2 ¼ ½ðp=2Þ b1 ¼ 26 340 .
Let us examine the special case of a crossed helical gear pair with shaft angle
R ¼ 90 , assuming that toothing is not characterized by any undercut, including tip
rounding or tip and root relief. The following discussion, however, is quite general,
but here we consider the special case of R ¼ 90 , because, as we have said else-
where, it is that we meet more frequently in practical applications. It also lends itself
better to the orthogonal projection representations, which show the genesis of
generally applicable relationships that we will write.
The tip cylinders of the two wheels intersect the line of action, defined in
Sect. 10.4, in the usual marked points A and E. Thus, the length g ¼ AE of path of
contact is uniquely defined. For the determination of the quantities of interest here,
it is convenient to consider what happens in three orthogonal projections: the first
two, on planes perpendicular to the wheel axes, and the third on the common plane
tangent to the two pitch cylinders at the point of contact C. This last plane is a plane
perpendicular to the shortest distance straight line (Fig. 10.12).
The path of contact, AE, is projected, on the two planes perpendicular to the
wheel axes (these are two axial views), according to the straight lines A0 E 0 and
A00 E 00 , which are inclined by at1 and at2 with respect to the plane tangent at point
C to the pitch cylinders of radii r1 and r2 . The straight lines A0 E 0 and A00 E 00 are the
intersections of the planes of action between each of the two wheels and the
common mating rack. The same path of contact is projected, on the plane tangent at
442 10 Crossed Helical Gears
Fig. 10.12 Path of contact of a crossed helical gear with R ¼ 90 , projected on three orthogonal
planes
point C to the pitch cylinders, according to the straight line A000 E000 , which is normal
to the common tangent to the pitch helices at point C.
In the axial view of the first wheel (the wheel 1), the point E0 is identified, while
in the axial view of the second wheel (the wheel 2), the point A00 is identified. The
other two projections E00 and E 000 of the point E are obtained by intersecting the
other two projections of the line of action with straight lines parallel to the axes of
the two wheels, and passing through point E 0 , such as Fig. 10.12 shows. The other
two projections A0 and A000 of point A are obtained in a similar manner, once the
point A00 is identified. The projection A000 E000 of the path of contact on the plane
tangent to the pitch cylinders defines the minimum face widths b1min and b2min , i.e.
the usable face widths, of the two wheels (Fig. 10.12).
Generalizing the problem to the case of shaft angle R 6¼ 90 , and also consid-
ering the two planes of action (one of these planes is shown, revolved in the
relevant view, in Fig. 10.9) and the two base cylinders, still remaining the meaning
10.6 Path of Contact, Face Width, and Contact Ratio 443
of the symbols already introduced in the previous chapters (see Figs. 1.8 and 2.10),
we can write the following relationships:
where gan1 and gan2 are respectively the lengths of the paths of recess and approach
in the normal section, while gat1 ¼ C 0 E 0 and gat2 ¼ C 00 A00 are respectively the
lengths of the paths of recess and approach in the transverse sections.
For determining the minimum face widths of the two wheels, it is more con-
venient to project the path of contact in the plane of action of the first wheel, on the
generatrix of the corresponding base cylinder, and the same path of contact in the
plane of action of the second wheel, on the generatrix of the corresponding base
cylinder. In doing so, we obtain the following relationships (see Fig. 10.12 below):
It should be noted that the parts of toothing that, in the axial direction, lie outside
of the range of minimum face width thus defined do not participate to meshing, and
they are therefore unnecessary. If we then consider the reference mating rack, which
is characterized by an infinite number of teeth, we obtain:
For ðha1 þ ha2 Þ ¼ 2mn and an ¼ 20 , i.e. with teeth having nominal depth, and
zero shifted toothing profile or profile-shifted toothing without variation of center
distance, we have:
Since these values are already valid for wheels having an infinite number of
teeth, to take account of assembly tolerances it is sufficient to increase them, by
adding ð1 2Þmn . In order to have teeth more resistant, as well as to avoid the risk
of edge contact, the inequality b
6mn must however always be satisfied.
Finally, the contact ratio en in the normal section is equal to the quotient of the
length of path of contact, AE, divided by the normal base pitch, pbn , for which we
have:
444 10 Crossed Helical Gears
AE AE
en ¼ ¼ ¼ en1 þ en2 ; ð10:44Þ
pbn pmn cos an
where en1 and en2 are respectively the addendum contact ratios of the pinion and
wheel, in the normal section, given by the relationships:
CE gan1 AC gan2
en1 ¼ ¼ en2 ¼ ¼ : ð10:45Þ
pbn pbn pbn pbn
The contact between the teeth of the two mating members of a canonical crossed
helical gear, as we have already said, is a point contact, and, during the transmission
of motion, a considerable helical sliding occurs between the teeth, which overlaps
with the profile sliding along the tooth depth. This helical sliding, also called
lengthwise cross-sliding, or longitudinal or lengthwise sliding, or axial sliding, is
evident if we consider the mode of transmission of motion from the driving wheel 1
to the driven wheel 2 of the crossed helical gear shown in Fig. 10.7.
The driving wheel 1, rotating about its own axis, causes a displacement of the
mating rack (this is materialized by its pitch plane, the P-plane) in the direction
normal to the axis 1. Suppose that this displacement, referred to the unit time, is
represented by vector CA. This vector can be resolved into two component vectors:
the first, AB, is the displacement of the pitch plane of the mating rack in the
direction of the straight line, n; the second, CB, is the displacement of the same
pitch plane in the direction normal to the axis 2. This second displacement causes
the rotation of the driven wheel 2 about its own axis. It thus becomes clear that the
transmission of motion from the driving wheel 1 to the driven wheel 2 is accom-
panied by a sliding in the direction of the tooth traces, which are helical traces.
Therefore, this sliding is known as helical sliding.
This type of transmission of motion shows a phenomenon similar to the one
described in Sect. 10.2, with reference to the mating hyperboloid gear pairs.
However, we are facing a substantial difference: in the present case, in fact, the
sliding no longer takes place along the Mozzi’s axis, but along the straight line, n,
chosen according to the criterion shown in the Sect. 10.3, and in any case distinct
from the Mozzi’s axis.
With reference to the sliding velocity, let us consider the points of contact C that
fall on the center distance O1 O2 (Fig. 10.7), i.e. the points of contact between the
pseudo-pitch surfaces. Once the choice of the straight line, n, is made, the relative
velocity vector vr ¼ ðv1 v2 Þ has the direction of this straight line, while its
absolute value is given by (see Fig. 10.6):
10.7 Longitudinal Sliding and Sliding Velocity 445
The second equality of this equation follows from the fact that v1 ¼ x1 r1 ,
v2 ¼ x2 r2 , and x2 r2 ¼ x1 r1 ðcos b1 = cos b2 Þ according to Eq. (10.19). For R ¼
90 ; which is the most frequent case of practical application, this equation becomes:
x 1 r1
vr ¼ : ð10:47Þ
sin b1
Equation (10.46) gives the sliding velocity of any point C between O1 and O2 . It
is certainly greater than that of points belonging to the Mozzi’s axis, which char-
acterizes the relative motion of a wheel with respect to the other. This factual
situation is fully justified by the following considerations. The relative velocity, vr ,
is a function of r1 and r2 , and therefore also of b1 and b2 , as Eqs. (10.19) and
(10.46) show. The condition for this relative velocity to be a minimum is obtained
by equating to zero the differential of Eq. (10.46), that is by imposing:
However, the helix angles b1 and b2 have been chosen by imposing the con-
dition given by Eq. (10.18), for which the factor within the second-round brackets
in Eq. (10.50) is equal to zero. Therefore, we get:
tan b1 tan b2
¼ ð10:52Þ
r1 r2
and thus
r1 tan b1
¼ : ð10:53Þ
r2 tan b2
446 10 Crossed Helical Gears
Therefore, the relative velocity reaches its minimum value when the distances r1
and r2 are such that their ratio equals the one between the trigonometric tangents of
the angles b1 and b2 , that is, when they coincide with the distances r1 and r2 which
define the position of the Mozzi’s axis.
If we look at the same phenomenon above analyzed from another point of view,
we have the opportunity to emphasize even more the substantial difference between
the hyperboloid gears and crossed helical gears (see Pollone [18], Schiebel and
Lindner [22, 23]). To this end, let us examine Fig. 10.13. It shows
that, for a
hyperboloid gear pair with axes 1 and 2, shaft angle R ¼ b1 þ b2 , center distance
a ¼ O1 O2 , and angular velocities x1 and x2 , the Mozzi’s axis, i.e. the axis of the
instantaneous relative helical motion passes through a point O lying between the
end points O1 and O2 of the center distance O1 O2 . Since ðx1 =x2 Þ ¼ ðz2 =z1 Þ; where
z1 and z2 are the numbers
of teeth of the two wheels, according to Eq. (10.1), the
angle b1 ¼ R b2 between the instantaneous axis of rotation and the axis 1 of
the first wheel, can be expressed with the following relationship:
Fig. 10.13 Instantaneous relative helical motion of a hyperboloid gear with shaft angle R, and
characteristic velocities: relative angular velocity vector, xr , and relative velocity vector, vr
10.7 Longitudinal Sliding and Sliding Velocity 447
Resolving the system formed by the two algebraic Eqs. (10.55) and (10.6), we
obtain:
tan b1 tan R b1
OO1 ¼ a OO2 ¼ a : ð10:56Þ
tan b1 þ tan R b1 tan b1 þ tan R b1
The instantaneous relative helical motion is defined by the relative angular ve-
locity vector xr1;2 ¼ ðx1 x2 Þ and relative velocity vector, vr1;2 ¼ ðv1 v2 Þ,
along the axis of the helical motion. The absolute values of these vectors are given
respectively by the relationships (see also Fig. 10.13):
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
z1 z1 sin R
xr ¼ x21 þ x22 þ 2x1 x2 cos R ¼ x1 1þ þ 2 cos R ¼ x1
z2 z2 sin R b1
ð10:57Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2 sin R
vr ¼ x21 OO1 þ x22 OO2 2x1 x2 OO1 OO2 cos R ¼ x1 OO1
cos R b1
tan b1 sin R
¼ x1 a :
tanb1 þ tan R b1 cos R b1
ð10:58Þ
Now, again with reference to Fig. 10.13, we replace the hyperboloid gear pair
with a crossed helical gear pair, having the same number of teeth, z1 and z2 , normal
module, mn , and helix angles, b1 and b2 . The radii of the pseudo-pitch cylinders of
the two members of this gear pair, which are tangent at point C, will be given by the
following relationship:
z1 m n z2 m n
r1 ¼ CO1 ¼ r2 ¼ CO2 ¼ : ð10:59Þ
2 cos b1 2 cos b2
0
The absolute value of the relative velocity (or sliding velocity) vr can be written
as:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi
0
vr ¼ xr CO1 O1 O2 þ vr :
2 2 ð10:60Þ
From this relationship, we infer that the minimum value of the sliding velocity is
reached when
448 10 Crossed Helical Gears
z1 m n
CO1 ¼ ¼ OO1 ; ð10:61Þ
2 cos b1
that is when b1 ¼ b1 , i.e. when the radii of the pitch cylinders are respectively equal
to OO1 and OO2 . This solution, corresponding to the minimum sliding, is feasible
only when the transmission ratio is slightly different from the unit. For greater
transmission ratios, the diameter of the pinion would be too small, as well as its
helix angle, as we can see from Eqs. (10.1) and (10.54).
It is to be remembered that the lengthwise sliding overlaps to the profile sliding,
already seen in parallel cylindrical gears (see Sect. 3.6). Generally, however,
especially when the shaft angle is large, the profile sliding is relatively low com-
pared to the lengthwise sliding. The total sliding velocity vector, vgc , is the vector
sum of two component vectors: the profile sliding velocity vector, vga , in the
direction of the tooth depth, and the helical sliding velocity vector, vgb , in the
direction of the tooth traces. Therefore, we have: vgc ¼ vga þ vgb .
This total sliding velocity is calculated at the tip cylinder of the wheel 1, because
here it has a maximum value. Of course, the calculation of the total sliding velocity
of the wheel 2 is still performed with the same procedure. Since the vectors vga and
vgb are orthogonal vectors, the absolute value of the vector vgc1 of the wheel 1 will
be given by the relationship
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
vgc1 ¼ v2ga1 þ v2gb1 ; ð10:62Þ
Fig. 10.14 Variation in intensity and direction of the total sliding velocity vector during the
meshing cycle
Here the load analysis for crossed helical gears is performed considering only the
power losses due to lengthwise sliding, and neglecting, as we did for parallel
cylindrical spur and helical gears and straight bevel gears, the power losses due to
profile sliding, in the direction of the tooth depth.
In order to perform this analysis, for ease of calculation, which does not affect
the reliability of the results obtained, we assume that the total force that the two
members of the crossed helical gear pair transmits to each other passes through
point C, where the two pitch cylinders touch one another. This force (Fig. 10.10) in
the section normal to the helicoids that are touching at point C, is characterized by
two components:
• A component, Fn , normal to the two surfaces in contact at point C, and inclined,
with respect to the common plane tangent at point C to the two pitch cylinders,
of the normal pressure angle, an .
• A friction component, Fl ¼ lFn , having the direction of the tangent to the pitch
helices at point C, and oriented so as to counteract motion.
450 10 Crossed Helical Gears
These two components, Fn and Fl ¼ lFn , of the total force acting on the mating
teeth are called here, respectively, normal force and friction force.
The coefficient of sliding friction, l, is notoriously a function of the material and
sliding velocity. With a good lubrication, its value can be taken equal to 0.1. In a
crossed helical gear, the friction force component, Fl ¼ lFn , can no longer be
ignored, as we have done in other types of gears hitherto analyzed, since it has
generally a not negligible magnitude, and therefore can affect significantly the
transverse tangential force transmitted. In other terms, the lengthwise sliding has
more pronounced effect than other kind of sliding (essentially, the profile sliding),
whose magnitude is comparatively small.
In turn, the normal force component, Fn , can be resolved in two other compo-
nents (Fig. 10.10): the first, Fn cos an , in the plane tangent to the two pitch cylinders
at point C, and, in this plane, perpendicular to the common tangent to the pitch
helices; the second, Fn sin an , passing through the point C, and perpendicular to the
plane tangent to the two pitch cylinders. This last component is the radial com-
ponent, Fr , of the total force, acting on both the wheels. In the same plane tangent to
the two pitch cylinders, vectorially adding the other component Fn cos an of the
normal force, and the friction force, Fl ¼ lFn , we obtain the vector Q applied at
point C (Fig. 10.10), and inclined, with respect to the component Fn cos an , of the
angle u, which is the angle of friction related to l by the relationship l ¼ tan u.
Finally, we resolve the vector Q according two different procedures:
• the first time, in the direction of the tangent to the pitch cylinder, in the middle
cross section of the first wheel, passing through the point C, and in the direction
of its axis;
• the second time, in the direction of the tangent of the pitch cylinder, in the
middle cross section of the second wheel, passing through the point C, and in
the direction of its axis.
By doing so, we obtain the other components of the total force, i.e. the transverse
tangential forces Ft1 and Ft2 , and the axial forces Fa1 and Fa2 acting on the two
wheels. All these components of the total force are given by:
The transverse tangential force, Ft2 , acting on the driven wheel (we assume that
the wheel 1 is the driving wheel, and the wheel 2 is the driven wheel), is known. It
is equal to the ratio between the input torque, T2 , to the driven machine (note that
10.8 Load Analysis for Crossed Helical Gears and Thrust … 451
the load capacity rating is effectively based on this torque), and the radius r2 of the
corresponding pitch cylinder ðFt2 ¼ T2 =r2 Þ. For this reason, it is best to express the
aforementioned force components as a function of Ft2 . Therefore, we have:
sinan sin an
Fr1 ¼ Fr2 ¼ Ft2 ffi Ft2 ð10:68Þ
cos an cos b2 lsinb2 cosðb2 þ uÞ
It should be noted that the expression preceded by the symbol ffi in the above
formulas are obtained by taking cos an ffi 1. They are approximate relationships.
The concepts of efficiency of the parallel cylindrical spur and helical gears, which
we have discussed in Sects. 3.9 and 8.10, can be applied also to the crossed helical
gears, with the necessary addition, as these gears are an even more general case.
Here we dwell on these additions, which are peculiar of the crossed involute helical
gears. The two helicoids, which form the flank surfaces of the mating teeth, are in
contact, at a given instant of the meshing cycle, at a single point of the path of
contact. In accordance with what we did in the previous section for calculation of
the forces acting on the teeth, also to calculate the friction losses due to the contact,
we limit our analysis to the case where this point of contact is the same point C, first
considered. According to the assumption made in the previous section, this point
belongs to the shortest distance straight line, O1 O2 .
As Cartesian coordinate systems for the two wheels that make up the crossed
helical gear pair under consideration, we take the right-hand coordinate systems
O1 ðx1 ; y1 ; z1 Þ and O2 ðx2 ; y2 ; z2 Þ, shown in Fig. 10.15. The origins of these systems
are the end points O1 and O2 of the center distance, a. Their z-axes coincide with
the axes of the two wheels; their x-axes coincide with the center distance straight
line, but are oriented from origin O1 towards point C for the first wheel, and from
the origin O2 towards point C for the second wheel; their y-axes are arranged to
complete the two right-hand coordinate systems. The distances of point C from the
end points of the center distance are the radii of the two pitch cylinders, that is
CO1 ¼ r1 and CO2 ¼ r2 . As we have already shown, these radii are related to the
radii of the base cylinders, rb1 and rb2 , and the transverse pressure angles, a1 ¼ at1
452 10 Crossed Helical Gears
and a2 ¼ at2 , by the relationships (10.33). Figure 10.15 also shows the tangential
velocity vectors v1 and v2 of the pitch point C, supposed to be rigidly connected first
to the wheel 1, and then to the wheel 2, and the projections of the common tangent,
t, to the mating tooth surfaces, which are touching at point C.
Here we assume that the shaft angle, R, is less than the sum of the absolute
values of the base lead angles, cb1 ¼ ½ðp=2Þ bb1 and cb2 ¼ ½ðp=2Þ bb2 , of the
two gear members under consideration (bb1 and bb2 are the base helix angles), for
which the following inequality is valid:
Under these conditions, the cones of normals have two generatrices in common
(see [8]). However, the condition for which a point P is a point of contact involves
the fact that two mating tooth surfaces must have at point P the same normal. It
follows that the cones of normals must have a common generatrix. However, we
have two common generatrices among the cones of normals. Therefore, there will
also be two possible directions of the common normal of the mating tooth surfaces:
one of them will be the active normal associated with one of the two possible
directions of rotation of members of the gear; the other will be the active normal
when the direction of rotation is reversed.
Fig. 10.15 Coordinate systems O1 ðx1 ; y1 ; z1 Þ and O2 ðx2 ; y2 ; z2 Þ, pitch point C on the center
distance, and velocity vectors, v1 and v2
10.9 Efficiency of Crossed Helical Gears 453
Let us consider the P-plane of Fig. 10.7, which is the plane perpendicular to the
shortest distance straight line passing through point C as well as the common
tangent plane to both pitch cylinders. As we have already seen, on this P-plane, the
total force exerted by the wheel 1 on the wheel 2 has two components: the com-
ponent of the normal force, Fn cos an , and the friction force, Fl ¼ lFn , already
defined in the previous section. These two components are shown in Fig. 10.16.
The same figure shows the tangential velocity vectors, v1 and v2 , the relative
velocity vector, vr ¼ ðv1 v2 Þ of the wheel 1 with respect to the wheel 2, and the
projections of the common tangent, t, and common normal, n, and projections of the
z-axes and y-axes of the aforementioned coordinate systems O1 ðx1 ; y1 ; z1 Þ and
O2 ðx2 ; y2 ; z2 Þ, on P-plane. Vectors v1 and v2 are inclined to the helix angles, b1 and
b2 , with respect to the application line of the component Fn cos an , while vector vr
has the same application line of the friction force, that is the aforementioned tan-
gent, t, but acts in the opposite direction to counteract motion.
With reference to the coordinate systems above, the three vectors v1 ; v2 , and vr
are given respectively by the following relationships (see Ferrari and Romiti [8],
Spiegel et al. [25]):
v1 ¼ x1 ðC O1 Þ ¼ x1 r1 j1 ð10:73Þ
Fig. 10.16 Components on the P-plane of the total force exerted by the wheel 1 on the wheel 2
454 10 Crossed Helical Gears
v2 ¼ x2 ðC O2 Þ ¼ x2 r2 j2 ð10:74Þ
vr ¼ v1 v2 ¼ x1 ðC O1 Þ x2 ðC O2 Þ ¼ x1 r1 j1 x2 r2 j2 ; ð10:75Þ
dWl
¼ lFn vr ; ð10:76Þ
dt
which, taking into account Eq. (10.46), can be expressed in the form:
dWl cos b1
¼ lFn x1 r1 sin b1 þ sin b2 : ð10:77Þ
dt cos b2
On the other hand, under steady-state conditions, for which x1 and x2 are
constant, by means of an equilibrium equation of moments about the z1 -axis, we get
the following relationship:
where T1 is the torque applied to the driving wheel, 1, and c1 ¼ ½ðp=2Þ b1 is the
lead angle of pinion.
Relationships (10.77) and (10.78) are the fundamental equations that will enable
us to define the loss of instantaneous efficiency of the crossed helical gear pair
under consideration. However, in order to do this in the most general terms, it is
first necessary to define the relationships that correlate the trigonometric functions
that appear in them and others. These relationships can be obtained by solving
another important problem simultaneously: that of the analytical definition of the
line of action, which is, as we have already seen, the intersection of planes P1 and
P2 passing through any point P of this line, and tangent to the base cylinders of the
two wheels (see Tang [26], Vaisman [28]).
If the generic point P under consideration is a point of contact, the two normals
at the mating tooth surfaces, which are tangent at this point, must coincide. In the
coordinate systems shown in Fig. 10.15, the unit vector of this common normal is
given by:
where:
• i1 ; j1 ; k1 , and i2 ; j2 ; k2 are the unit vectors of axes of the two coordinate systems;
• a1 (it is measured in the direction of the positive rotations about the z1 -axis) is
the angle between the plane (x1 ; z1 ) and the plane detected by the z1 -axis and
generatrix of contact of plane P1 with the base cylinder of the wheel 1;
• a2 (it is measured in the direction of the positive rotations about the z2 -axis) is
the angle between the plane (x2 ; z2 ) and the plane detected by the z2 -axis and
generatrix of contact of the plane P2 with the base cylinder of the wheel 2;
• cb1 ¼ ½ðp=2Þ bb1 and cb2 ¼ ½ðp=2Þ bb2 .
Among the unit vectors of the two coordinate systems above, the following cor-
relations exist:
i1 ¼ i2 j1 ¼ cos Rj2 þ sin Rk2 k1 ¼ sin Rj2 cos Rk2 ; ð10:80Þ
where R is the acute angle between the z1 -axis and z2 -axis, i.e. the shaft angle.
Introducing these last relationships in Eq. (10.79), we get:
sin a1 sin cb1 i2 þ ðcos cb1 sin R cos a1 sin cb1 cos RÞj2
ðcos a1 sin cb1 sin R þ coscb1 cos RÞk2 ð10:81Þ
¼ sin a2 sin cb2 i2 cos a2 sin cb2 j2 cos cb2 k2 :
By equating the homogeneous terms in this equality, we get the following three
equations:
Once a1 is known, from the second of the Eqs. (10.82), we get a2 , so we can
write, in the coordinate system O2 ðx2 ; y2 ; z2 Þ, the following equation of the plane
P2 , which is tangent to the base cylinder of the wheel 2, to which the line of action
to be defined also belongs:
Since
Equations (10.83) and (10.86) are the parametric equations of the intersection
straight line between the two planes, P1 and P2 ; this line is the common normal
between the mating tooth flank surfaces at points of contact. We have thus uniquely
defined the line of action of the crossed helical gear under consideration. From
Eq. (10.82), we infer the following relationships which define the angles a1 and a2
as a function of angles cb1 ; cb2 ; and R:
cos cb2 cos cb1 cos R cos cb2 cos R cos cb1
cos a1 ¼ cos a2 ¼ : ð10:87Þ
sin cb1 sin R sin cb2 sin R
Now let us consider the case where the path of contact intersects the shortest
distance straight line, that is the x1 -axis, at the pitch point C. At this point, we have
y1 ¼ y2 ¼ z1 ¼ z2 ¼ 0, so from Eqs. (10.83) and (10.84), taking into account the
first of the Eqs. (10.85), we get:
rb1 rb2
x1 ¼ ¼ r1 x2 ¼ ¼ r2 ¼ a r1 ; ð10:88Þ
cos a1 cos a2
which coincide with Eqs. (10.33). So, the angles a1 ¼ at1 and a2 ¼ at2 are the
transverse pressure angles of the two members of the crossed helical gear. Of
course, in order to transmit the motion between these two members, it is necessary
that Eq. (10.31) is satisfied.
Let us return to consider the P-plane, shown in Fig. 10.16. In this figure, in
addition to the quantities already defined above, also the projections of the common
tangent to pitch helices, t, and the common normal, n, to the mating tooth surfaces,
which are touching at point C, are highlighted. Of course, t and n are perpendicular
to each other. From Eq. (10.79) in terms of coordinate system O1 ðx1 ; y1 ; z1 Þ, it turns
out that the normal, n, oriented as in the figure, has direction parameters relative to
y1 - and z1 -axes, respectively equal to cos a1 sin cb1 and cos cb1 , and therefore
direction cosines given by:
Since, on the P-plane (Fig. 10.16), b1 and b2 are the angles that the tangent, t,
forms respectively with the z1 -axis and the ðz2 Þ-axis, we will have:
ðn; y1 Þ ¼ ðp b1 Þ, ðn; y2 Þ ¼ ðp b2 Þ, ðn; z1 Þ ¼ ½ðp=2Þ b1 and
ðn; z2 Þ ¼ ½ðp=2Þ b2 . From the second and first Eqs. (10.89), we get:
Of course, we can write the two similar following relationships for the wheel 2:
cos cb2 cos cb2
sin b2 ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 sin2 a2 sin2 cb2 1 sin2 a1 sin2 cb1
ð10:91Þ
cos a2 sin cb2 cos a2 sin cb2
cos b2 ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ;
1 sin2 a2 sin2 cb2 1 sin2 a1 sin2 cb1
because, for the first of the Eqs. (10.82), we have sin2 a2 sin2 cb2 ¼ sin2 a1 sin2 cb1 .
From Eq. (10.78), taking into account the first of the Eqs. (10.90), we get:
T1
Fn ¼ " #: ð10:92Þ
cos c1
r1 cos a1 sin c1 þ l pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 sin2 a1 sin2 c1
From Eq. (10.77), taking into account the Eqs. (10.91) and (10.92), we obtain:
we can express the losses of instantaneous efficiency of the crossed helical gear pair
here examined (they are obviously equal to the quotient of the instantaneous power
dissipated, dWl =dt, divided by the instantaneous power supplied by the driving
wheel, dWm =dt) as follows:
458 10 Crossed Helical Gears
cos cb2 ¼ cos a1 sin cb1 cos cb1 ¼ cos a2 sin cb2 : ð10:98Þ
From Eq. (10.24), taking into account these two last relationships as well as
Eq. (8.8), applied to both gear members, we get:
where q ¼ ðr1 =r2 Þ. Then, from Eqs. (10.99) and (10.98), we obtain:
cos a2
cot cb2 ¼ cot cb1 ¼ iq cos a1 : ð10:100Þ
iq
Finally, introducing these two relationships into Eq. (10.97), and taking into
account Eq. (10.96), the latter becomes:
cos a2
1 l2
iq iq l1
g¼ ¼ : ð10:101Þ
1 þ l1 iq cos a1 iqð1 þ l1 iqÞ
This is the final relationship which expresses the efficiency of a crossed helical
gear pair with shaft angle R ¼ 90 . From this equation we first infer that, if
q ðl1 =iÞ, the transmission of motion under steady
state
conditions is not possible,
as the efficiency becomes negative, so that dWl =dt [ T1 x1 , that is the power
10.9 Efficiency of Crossed Helical Gears 459
i2 q2 2l1 iq 1 ¼ 0; ð10:102Þ
If l1 is very small compared to the unit, for g ¼ gmax , we have q ffi ð1=iÞ. Under
these conditions, from Eqs. (10.99) and (10.98) we obtain:
Therefore, we conclude that, for crossed helical gears with shaft angle R ¼ 90 ,
and contact occurring on the shortest distance straight line, the frictional power
losses are minimal when the ratio of the pitch radii of the two gear members,
q ¼ ðr1 =r2 Þ is about equal to ð1=iÞ ¼ ðx2 =x1 Þ, so the helix angles are about equal,
and equal to p=4.
In global terms, it is possible to express the efficiency of toothing of a crossed
helical gear pair much more simply, as the ratio between the input power to the
driven machine, or usable power, and the output power of the driving machine, or
engine power. Therefore, taking into account Eq. (10.18) and the appropriate
relationship (10.69), we can write the following equation, which is valid whatever
the shaft angle:
If we make the derivative of this equation with respect to b1 , taking into account
that b2 ¼ ðR b1 Þ, and equating to zero the derivative thus obtained, we find that,
for a given value of the shaft angle R, the efficiency reaches a maximum for
ðb1 b2 Þ ¼ u, i.e. for:
460 10 Crossed Helical Gears
1 1
b1 ¼ ðR þ uÞ b2 ¼ ðR uÞ: ð10:107Þ
2 2
Since the value of the angle of friction u is very small, we come to the con-
clusion that the efficiency has its maximum value when
R
b1 ffi b2 ffi : ð10:108Þ
2
In the special case in which the shaft angle R ¼ ðp=2Þ, the efficiency is maxi-
mum when b1 ¼ b2 ¼ p=4.
For a crossed helical gear pair with shaft angle R, z1 and z2 teeth, normal module
mn , and helix angle b1 ¼ b2 ¼ R=2, and then in the condition of maximum effi-
ciency, the center distance, a, is given by:
1 mn
a ¼ ðz1 þ z2 Þ : ð10:109Þ
2 cosðR=2Þ
Since the pitch helix angles are equal, the pitch diameters are proportional to the
numbers of teeth, i.e.:
z1 m n z2 m n
d1 ¼ 2r1 ¼ d2 ¼ 2r2 ¼ : ð10:110Þ
cosðR=2Þ cosðR=2Þ
When the center distance, a, the shaft angle, R, the normal module, mn , and the
transmission ratio, i, are given, the numbers of teeth of the two wheels are given by:
2a cosðR=2Þ 2a cosðR=2Þ
z1 ¼ z2 ¼ : ð10:111Þ
m n ð1 þ i Þ m n i ð1 þ i Þ
It is to be noted that b1 [ b2 , and that when b2 exceeds a certain limit value, the
transmission of motion from the driving shaft 1 to the driven shaft 2 cannot take
place. This limiting value can be obtained by putting g ¼ 0 in Eq. (10.106), whence
we get tan b2 ¼ ð1=lÞ ¼ ð1= tan uÞ ¼ cot u ¼ tanð90 uÞ. Therefore, the trans-
mission of motion is theoretically possible only when b2 \ð90 uÞ. The practical
limiting value of b2 is still less. When the shaft angle is R ¼ 90 , the efficiency, g,
is given by:
tanðb1 uÞ
g¼ : ð10:112Þ
tan b1
Finally, it is to be noted that the power losses due to the lengthwise sliding
previously considered do not constitute the only power losses of a crossed helical
gear pair. Beside these losses, other losses occur, such as the losses due to toothing
action (these losses depend on the transmission ratio, the numbers of teeth, and the
normal module), the bearing losses, and other loss phenomena.
10.10 Profile-Shifted Toothing for Crossed Helical Gears 461
Like parallel cylindrical helical gears, crossed cylindrical helical gears can also
have profile-shifted toothing when needed. The same reasons for which we use
parallel cylindrical helical gears with profile-shifted toothing are also valid for
crossed cylindrical helical gears (in this regard, see Chap. 8). Here the subject is
discussed in its most general terms, i.e. with reference to the case of profile-shifted
toothing with variation of center distance, leaving the reader the task of applying the
general concepts to the special case of profile-shifted toothing without variation of
center distance.
We assume to cut the two members of the crossed helical gear under consid-
eration by means of a rack-type cutter having, in its normal section, normal module
mn , and nominal standard sizing. We continue to give to the quantities and symbols
already introduced, i.e. an ; at1 ; at2 ; b1 ; b2 ; r1 ,r2 ; mn ,mt1 , and mt2 , the same meaning
given to them that we have so far, and we denote with the subscript, 0, the cor-
responding quantities referred to the cutting pitch cylinders. The relationships
(10.27), (10.28), and (10.32) are valid, but also the equations that follow apply:
z1 mt0 z1 mn0
r10 ¼ ¼ ð10:113Þ
2 2 cos b10
cot b10
r1 ¼ r10 ; ð10:114Þ
cot b1
this is because mn0 ¼ mt0 cos b10 , and pz ¼ 2pr10 cot b10 ¼ 2pr1 cot b1 . In addition,
the following equations apply:
these equations are, respectively, in accordance with Eqs. (8.17) and (8.35).
On the cutting pitch cylinder of the first wheel, the normal tooth thickness sn10
and transverse tooth thickness st10 are given by:
p
p 2x1
invc1 ¼ invat0 þ þ tan an0 : ð10:119Þ
2z1 z1
On the operating pitch cylinder, having radius r1 , the normal tooth thickness sn1
is equal to:
For the second wheel, the driven wheel, analogous relationships apply. In par-
ticular, on its operating pitch cylinder, having radius r2 , the normal tooth thickness
sn2 is equal to:
hp i
sn2 ¼ mn þ 2x2 tan an0 þ z2 ðinvat20 invat2 Þ : ð10:121Þ
2
which, given the equations written above, takes the following form:
Once the values of b1 ; b2 ; an0 and c are chosen, and values of shift coefficients x1
and x2 have been determined based on the virtual numbers of teeth zv1 ¼
ðz1 = cos3 b1 Þ and zv2 ¼ ðz2 = cos3 b2 Þ, using the procedures described in this regard
in Chaps. 6 and 8, the value of the second member of Eq. (10.123) is uniquely
defined. Taking into account the relationships (10.32), (10.115) and (10.116), the
first member of the same Eq. (10.123), once the values of b1 and b2 are chosen, is a
function only of an , which can be determined by a trial and error procedure. Of
course, this equation can be easily solved numerically.
Always with reference to the normal section, the determination of an can be
carried out also, in a more rapid but approximated way, using the following
relationship:
x1 þ x2
invan ¼ invan0 þ 2 tan an0 : ð10:124Þ
zv1 þ zv2
10.10 Profile-Shifted Toothing for Crossed Helical Gears 463
Once the value of an has been thus calculated, using the formulas above, we
determine in succession the quantities b10 ; b20 ; at10 ; at1 ; at20 and at2 ; then we
modify x1 and x2 so that Eq. (10.123) is satisfied.
Further insights on this subject are not necessary, since the calculation procedure
does not differ from that described in the already above-mentioned Chaps. 6 and 8.
References
1. Ball RS (1876) The theory of screw: a study in the dynamics of a rigid body. Hodges,
Foster&Co, Dublin
2. Buckingham E (1949) Analytical mechanics of gears. McGraw-Hill Book Company Inc, New
York
3. Ceccarelli M (2000) Screw axis defined by Giulio Mozzi in 1763 and early studies on
helicoidal motion. Mech Mach Theory 35:761–770
4. Colbourne JR (1987) The geometry of involute gears. Springer, New York Inc, New York
Berlin Heidelberg
5. Dobrovolski V, Zablonski K, Mak S, Radtchik A, Erlikh L (1971) Éléments de Machines.
Édition MIR, Moscou
6. Dudley DW (1962) Gear handbook. The design, manufacture, and application of gears,
McGraw-Hill, New York
7. Euler L (1765) Theoria Motus Corporum Solidorum seu Rigidorum ex primis nostrae
cognitionis principiis stabilita et ad omnes motus, qui in huius modi corpora cadere possunt,
accomodata, Rostochii et Gryphiswaldiae Litteris et Impensis A. F. Röse
8. Ferrari C, Romiti A (1966) Meccanica Applicata alle Macchine, Unione Tipografica –
Editrice Torinese (UTET), Torino
9. Giovannozzi R (1965) Costruzione di Macchine, vol II. 4th edn. Casa Editrice Prof. Riccardo
Pàtron, Bologna
10. Henriot G (1979) Traité théorique et pratique des engrenages, vol 1. 6th edn. Borda, Paris
11. Herrmann R (1928) Evolventen-Stirnrädgetriebe. Julius Springer, Berlin
12. Levi-Civita T, Amaldi U (1929) Lezioni di Meccanica Razionale—Vol. 1: Cinematica—
Principi e Statica, 2nd edn. Reprint 1974, Zanichelli, Bologna
13. Litvin FL, Fuentes A (2004) Gear geometry and applied theory, 2nd edn. Cambridge
University Press, Cambridge
14. Merritt HE (1954) Gears, 3rd edn. Sir Isaac Pitman & Soins Ltd, London
15. Minguzzi E (2013) A geometrical introduction to screw theory. Eur J Phys 34:613–632
16. Mozzi del Garbo GG (1763) Discorso matematico sopra il rotolamento momentaneo dei
corpi, Stamperia Donato Campo, Napoli
17. Niemann G, Winter H (1983) Maschinen-Element Band III:Schraubrad-, Kegelrad-,
Schneckn-, Ketten-, Riemen-, Reibradgetriebe, Kupplungen, Bremsen, Freiläufe, Springer,
Berlin, Heidelberg
18. Pollone G (1970) Il Veicolo. Libreria Editrice Universitaria Levrotto & Bella, Torino
19. Poritsky H, Dudley DW (1948) Conjugate action of involute helical gears with parallel or
inclined axes. Q Appl Math 6(3):193–214
20. Radzevich SP (2016) Dudley’s handbook of practical gear design and manufacture, 3rd edn.
CRC Press, Taylor & Francis Group, Boca Raton, Florida
21. Schiebel A (1913) Zahnräder, II Teil, Räder mit schrägen Zähnen. Verlag con Julius Springer,
Berlin
22. Schiebel A, Lindner W (1954) Zahnräder, Band 1: Stirn-und Kegelräder mit geraden Zähnen,
Springer, Berlin, Heidelberg
464 10 Crossed Helical Gears
23. Schiebel A, Lindner W (1957) Zahnräder, Band 2: Stirn-und Kegelräder mit schrägen Zähnen
Schraubgetriebe, Springer, Berlin, Heidelberg
24. Scotto Lavina G (1990) Riassunto delle Lezioni di Meccanica Applicata alle Macchine:
Cinematica Applicata, Dinamica Applicata - Resistenze Passive - Coppie Inferiori, Coppie
Superiori (Ingranaggi – Flessibili – Freni). Edizioni Scientifiche SIDEREA, Roma
25. Spiegel MR, Lipschutz S, Spellman D (2009) Vector Analysis, Schaum’s Outlines, 2nd edn.
McGraw-Hill, New York
26. Tang KT (2006) Mathematical methods for engineers and scientists. Springer, Berlin,
Heidelberg
27. Townsend DP (1991) Dudley’s gear handbook, 2nd edn. McGraw-Hill, New York
28. Vaisman I (1997) Analytical geometry. World Scientific Publishing Co, Singapore
29. Zwirner G (1961) Istituzioni di Matematiche, Parte Prima, 5th edn. CEDAM-Casa Editrice
Dott. Antonio Milani, Padova
Chapter 11
Worm Gears
Abstract In this chapter, the geometry of the various types of worms and corre-
sponding worm wheels is first described and short notes about their cutting pro-
cesses are provided. The main geometric quantities of worms and worm wheels are
then defined. The parametric equations of helicoid surfaces in terms of differential
geometry are subsequently determined and the main kinematic quantities of these
types of gears are obtained. The meshing between worm and worm wheel is then
analyzed using the aforementioned concepts of differential geometry, and the
instantaneous lines of contact as well as surface of contact are analytically deter-
mined. Subsequently, the two well-known graphic-analytical methods of deter-
mining the surface of contact are described, i.e. the Schiebel’s method for
Archimedean spiral worms and Ingrisch’s method for involute worms. Interference
problems related to the outside surface of the worm wheel are then discussed and
load analysis of worm gears is performed, defining the thrust characteristics on shafts
and bearings and determining their average efficiency. Further consideration on the
worm and worm wheel sizing are made and short notes on double-enveloping worm
gear pairs are provided. Finally, standard and non-standard worm drives and special
worm drives are described.
11.1 Introduction
Worm gear pairs are gears with crossed axes, constituted by a worm (or worm
screw) of cylindrical or toroidal shape that meshes with a worm wheel having tooth
flanks capable of a line contact with the flanks of the worm threads. Usually, these
gears are used for non-parallel, non-intersecting, right-angle crossed axes (more
rarely, for crossed axes with shaft angle R 6¼ p=2), where high center distances, and
high reduction ratios are required, although in many practical applications they are
also employed for low and medium reduction ratios. In fact, worm gear pairs allow
reaching large reduction ratios, and therefore a large multiplication of torque, using
only one pair of mechanical geared members. This would not be possible, or would
not be convenient, using gear pairs such as those so far studied and analyzed in the
previous chapters.
As we have seen in the previous chapter, worm gears are a special case of
crossed helical gears. However, unlike the latter gears, which are characterized by a
point contact, the worm gears are characterized by a line contact. Given their
extensive use for interesting practical applications, they are discussed in the most
classical textbooks on gears (see Buckingham [2], Merritt [39], Dudley [13],
Giovannozzi [17], Pollone [48], Henriot [19], Niemann and Winter [44], Colbourne
[7], Townsend [65], Maitra [37], Jelaska [25], Radzevich [51, 52]). Here the dis-
cussion on this important topic is carried out not only in accordance with the
classical one, available in more or less extensive form in the traditional afore-
mentioned textbooks, but also using the most advanced numerical developments
based on rigorous theoretical analysis that uses differential geometry methods (see
Litvin and Fuentes [33]). It should be noted that the theoretical method described
here is so general that it can be used for all gears covered in this textbook, including
hypoid gears.
Since the instantaneous contact between the two members of a worm gear is a
line contact, contrary to what we highlighted for crossed helical gears, it is able to
transmit loads significantly greater. The meshing contact between the worm threads
and worm wheel teeth is a combination of rolling and sliding, with the latter
prevailing at higher reduction ratios. How best we will specify in the following
sections, the relative motion of the worm with respect to worm wheel is a screw
motion about an axis lying in a plane that is perpendicular to the line of shortest
distance between the two axes.
From the point of view of kinematics, the worm gear pairs are known to be
characterized by only one correct relative position between the axes of the two
members, which is theoretically the only one that can guarantee optimal operating
conditions. The cutting process and, more generally, the production and assembly
processes of the two members of the worm gear must be extremely accurate. The
worm wheel has curvilinear teeth, which are obtained by a generation cutting
process by means of a hob having the same geometric characteristics of the worm.
To produce the two members of a worm gear, we encounter many difficulties,
which include the construction of the worm as well as the tool with which the worm
wheel is to be generated. This cutting tool must have the same characteristics of the
worm. In fact, the hob is made of tool steel (see Davis [9], Radzevich [50]) and, due
to the necessary hardening heat treatment, undergoes considerable deformations,
also due to notches to create the cutting edges. Such deformations must be elimi-
nated by a subsequent finishing operation, generally consisting of a precision
grinding. The accuracy of this finishing operation of the worm thread and hob
strongly affects the correct operating conditions of the worm gear.
Worm gear pairs are also sensitive to assembly errors, among which the varia-
tions of shaft angle and center distance, and axial displacements of the worm wheel
play an important role. These errors result in a shift of the bearing contact to the
edge, and determine a piecewise almost-linear function of transmission errors,
which have the same frequency as the meshing cycle between the worm threads and
worm wheel teeth. A proper mismatch between the worm and hob surfaces results
11.1 Introduction 467
Fig. 11.1 Single-enveloping worm gear pair with crossed orthogonal axes, consisting of a
cylindrical two-start worm and an enveloping worm wheel, and its characteristic quantities
468 11 Worm Gears
At least one of the two members of a worm gear pair is therefore an enveloping
member. Later we will give motivation for this configuration. Worm gear pairs
most commonly used are still the first two, which we can call single-enveloping
worm gear pairs.
The mutual operating position of the two members of a worm gear pair can vary,
depending on the needs of connection with the types of driving machine (electric
motor or internal combustion engine), and driven machine. In a worm-drive unit,
the worm is generally the driving member, but not always the case: there are cases
in which the worm wheel is the driving member. As a rule, when possible, it is
convenient to have the worm, with its horizontal axis, below that of the worm
wheel, also with a horizontal axis, and not vice versa. Several cases exist, however,
in which the worm has horizontal axis and worm wheel has vertical axis, or the
worm has vertical axis and the worm wheel has horizontal axis [17]. Of course, we
are talking about the usual, conventional applications, which therefore do not
concern the countless variety of worm-drive systems that have been developed,
with characteristics entirely new and innovative, for special applications. These
types of worm gear drives are beyond the purpose of this textbook. Nonetheless, the
reader will find a brief reference to this topic in the Sect. 11.14.
The main working features of the worm gear pairs are summarized as follows:
• High gear ratios, u, with comparatively smaller overall dimensions and space
requirements and, consequently, lower weight and lower cost. For a speed
reducing unit, the range of variability of u is commonly 5 u 70, mainly
15 u 50, and for low power to be transmitted up to u ¼ 1000.
• Possibility to drive the worm from both ends of its shaft, as well as to drive
continuous shafts that carry several worms in series, which in turn drive as many
driven devices.
• Silent and anti-vibration operating conditions, due to the high sliding and low
speed of impact between the thread flanks of the worm and mating tooth flanks
of the worm wheel. With equal number of turns of the worm and torque
transmitted, the sound level is on average less than 7 dBðBÞ compared to
gear-drive systems analyzed so far.
• Self-locking ability or irreversibility of drive, in relation to the design needs.
In view of these advantages, the main disadvantages are also to be observed.
They consist of low overall efficiency, high frictional losses that give rise to heat,
and comparatively low transmitted power (see Giovannozzi [17], Ferrari and
Romiti [15], Scotto Lavina [59]).
11.2 Geometry of Worm Thread Profiles … 469
The thread flanks of this type of worm are generated as envelopes of straight lines in
axial planes, which are inclined at a constant angle ½ðp=2Þ a0t to the axis. Thus,
this type of worm is the limit case in which the above-mentioned cylinder has
diameter, d, equal to zero, and the tangent to the helix and the same helix coincide
with the axis of the worm. The generation straight line of the helicoid then intersects
the axis of the worm, forms with it a constant angle ½ðp=2Þ a0t , and moves along
it with a screw motion or helical motion with lead, pz ; this motion consists of
simultaneous uniform rotation about and translation along the axis. Therefore, we
get a worm whose axial section is a common rack with trapezoidal teeth.
470 11 Worm Gears
The worm flank surface is a ruled surface, which is the locus of the successive
positions of the generation straight line during its screw motion with respect to the
worm axis. The section of the helicoid with a transverse plane is an Archimedean
spiral, whence the name of Archimedean spiral worm that we give commonly to
this type of worm. In a normal section, the thread profiles are convex curves.
As Fig. 11.2 shows, threads may be cut on a lathe, with a cutting tool having
straight edges, the cutting plane of which lies in an axial plane of the worm. Both
flanks of a thread space may be cut simultaneously by using a cutter of trapezoidal
shape. Type A worms can be also cut by using an involute shaper to produce the
desired straight rack-type profile in an axial plane of the worm; its cutting face must
lie in that axial plane. It is also necessary that the pitch circle of the shaper rolls
without sliding on the datum line of the rack profile, which coincides with a
straight-line generatrix of the worm pitch cylinder. This second cutting method of
the worm is essentially an inversion of the process of cutting a helical gear wheel
with a rack-type cutter.
It is to be noted that the type A worms may be grounded only using suitable
profiled grinding wheels. This is because, in normal sections, the flank profiles of
the threads are convex curves, that is curvilinear profiles. For more in-depth
knowledge of the cutting and grinding methods of the various types of worms and
related worm wheels, we refer the reader to specialized textbooks (see Galassini
[16], Townsend [65], Radzevich [50]).
Fig. 11.2 Type A cylindrical worm and machining methods by turning or shaping
11.2 Geometry of Worm Thread Profiles … 471
The thread flanks of this type of worm (Fig. 11.3) are involute helicoidal surfaces,
generated by the screw motion with lead, pz , of a generation straight line not
passing through the worm axis. This generation straight line forms with worm axis
a constant angle bb (the base helix angle), and is constantly tangent to the base
(a) (b)
(c)
Fig. 11.3 Type I cylindrical worm and machining methods by: a turning; b milling or grinding
with a first placement of the cutter; c milling or grinding with a second placement of the cutter
472 11 Worm Gears
cylinder of radius rb . These quantities are related between them by the following
relationship:
With reference to the general case mentioned at the beginning of this section, in
this case the generation straight line of the helicoid is tangent to the base helix.
Therefore, the constant angle between the tangents to the base cylinder, in corre-
spondence of successive points of the base helix, and tangents to this helix in these
points is equal to zero. The base helix lies on the base cylinder of the worm, which
has base diameter db ¼ 2rb .
The helicoid thus generated is a ruled and developable surface, whose section
with a transverse plane is the involute curve of the circle of radius rb , which
constitutes the transverse section of the base cylinder. Therefore, this type of worm
is called involute worm. However, the thread shape, which is an involute in
transverse section as in the case of a helical gear wheel, is convex in normal section,
and approximately hyperbolic in the axial section.
Since the generation straight line is always tangent to the base helix in a plane,
which is tangent to the base cylinder, the flank profile of the worm is a straight line
in a plane that is tangent to the base cylinder. This straight profile can be obtained
by several cutting methods. In fact, the thread flanks can be generated by turning on
a lathe, using a cutting tool with its straight edge aligned with the generation
straight line in a plane tangent to the base cylinder (Fig. 11.3a). In order to cut both
flanks of a thread simultaneously, it is necessary to use one right-hand tool in one
plane, and one left-hand tool in another plane.
The thread flanks can be also cut by milling or grinding, using the plane side face
of a disk-type milling cutter or grinding wheel. The cutting tool can be placed in
two different ways:
• With the first placement, the cutting face is aligned in such a way that either its
axis lies in a plane that is tangent to the base cylinder, and the generation straight
line of the flank lies on the cutting face (Fig. 11.3b).
• With the second placement, the cutting face is aligned with the worm reference
helix and, in a plane perpendicular to the reference helix, it is set to the normal
pressure angle of the flank, a0n (Fig. 11.3c). Note that, in Fig. 11.3, a0t and a0n
are respectively the transverse and normal pressure angles of the cutter, while cb
and c1 are respectively the base and reference lead angles of threads.
Both these placements require that the mounting of the worm on the milling or
grinding machines must be reversed between machining left and right flanks. The
use of milling cutters or grinding wheels finds its justification in the fact that, along
each of the generatrices, the helicoid admits a tangent plane, and this makes easier
the milling or grinding of the threads. The second type of alignment of the
machining tool has the advantage that the cutting face extends to near the thread
root, and this is not simple to do with the first type of alignment.
11.2 Geometry of Worm Thread Profiles … 473
The type N worm (Fig. 11.4) is similar to type A worm, with the variation that the
thread profile is of trapezoidal shape, i.e. straight sided, in the normal section, rather
than in the axial section. It represents the more general case of ruled helicoid, in the
configuration of the thread flanks, which we discussed at the beginning of this
section. The thread profile is slightly curved in the axial section. The generation
straight line lies in a plane passing through the perpendicular to the worm axis, and
forming the angle c1 (the reference lead angle, i.e. the lead angle on the reference
pitch cylinder of the worm) with the worm axis. The transverse section of the worm
thread, i.e. the transverse worm profile, is an extended involute.
As Fig. 11.4 shows, the machining methods are several. The threads can be cut
in a lathe with a tool having trapezoidal shape and cutting edges placed in the
cutting plane, which match the profile of the thread space in a plane normal to the
reference helix of the thread space (Fig. 11.4a). The threads can be machined, in a
more or less appropriate way, in a milling or grinding machine, using a biconical
milling cutter or grinding wheel of small diameter, as Fig. 11.4b shows (when the
diameter of the grinding wheel becomes large, the type N worm approaches the type
K worm), or a small conical gear milling cutter or grinding wheel (Fig. 11.4c).
These last two methods determine profiles that are approximate, because the effects
due to the change of helix with the change of thread depth. Deviations and alter-
ations with respect to the proper profile are much larger, the bigger the diameter of
the milling cutter or grinding wheel and the lead of the worm.
The thread profiles of this type of worm are concave in the normal section. This
worm is frequently used for some advantage offered by it, inherent to the cutting
process. Unlike those of types A, I, and N, the thread flanks of type K worms do not
have a generation straight line, and their surfaces are not ruled surfaces, but
envelopes to a family of cone surfaces. The thread spaces of these worms are
generated with biconical grinding wheels or disk-type milling cutters having
straight cone generatrices, and performing a screw motion about the worm axis.
As Fig. 11.5 shows, the common perpendicular to the tool spindle and worm
axes lies on the line of intersection, l, of the middle plane of the tool and a
transverse plane of the worm; the same figure shows, below, the trace point of this
intersection line. The angle between these two planes is equal to the worm lead
angle, c1 . The straight generatrix of each conical side of the tool and the middle
plane of the same tool forms an angle equal to the normal pressure angle a0n . The
worm is turned uniformly with simultaneous axial translation of threads, so that a
474 11 Worm Gears
Fig. 11.4 Type N cylindrical worm and machining methods by turning (a), or by milling or
grinding (b, c)
point on the common perpendicular, distant r1 from the worm axis (r1 is the
reference radius of the worm), describes the reference helix. The conical sides of the
tool generate the helicoidal flanks of the worm. The profile shape is affected by
the change of helix angle with change of the thread depth, and points of the tool
flanks, which contact the worm threads, lie on a curve and not on any one-cone
generatrix.
11.2 Geometry of Worm Thread Profiles … 475
Fig. 11.5 Type K cylindrical worm and machining methods by grinding or milling
This type of worm has the advantage that the two flanks of a thread space can be
machined simultaneously, but it has the disadvantage that the shape of the thread
flanks varies with tool diameter, so the reproducibility is approximate. It should be
noted that the smaller the tool diameter, the more nearly the normal profiles of the
thread spaces approach those of type N worms, and the larger the tool diameter the
476 11 Worm Gears
more nearly the shapes of the thread flanks approach those of type I worms. It is
also to be noted that type K worms can be machined with a conical milling cutter,
but the flank surfaces have facets resulting from the cutting discontinuity due to
cutting tooth pitch.
This type of worm (Fig. 11.6) has thread profiles concave in axial and normal
sections, whereas the worm wheel teeth have convex profiles, which fit snugly into
the corresponding thread spaces of the worm during the meshing action. Unlike the
four types of worms described above, which have threads with straight or convex
profiles, the geometry of the type C worm is much more advantageous and satis-
factory as regards the requirements for a good and close contact between the mating
surfaces of worm and worm wheel, and for the creation of adequate pressure of the
lubricant oil film.
Unlike those of types A, I, and N, the thread flanks of type C worms do not have
a generation straight line, and the worm surfaces are not ruled surfaces, but
envelopes to a family of generating surfaces. Like type K worms, the thread spaces
of type C worms are generated with a disk-type milling cutter or grinding wheel. In
order to produce the concave thread profiles of this type of worm, the tool has a
cutting profile consisting of convex circular arcs. Thus, the generation surface is a
surface of revolution, which performs a screw motion about the worm axis. The
center distance, a0 ¼ ½ðdm1 þ dm0 Þ=2, varies with the tool diameter, dm0 (Fig. 11.6).
The reference lead angle, c1 , is usually equal to the angle between the projections of
the worm and tool axes onto a plane perpendicular to the center distance.
The generating process of the profiles of this type of worm is the same as for
type K worm. The shape of the thread flank profiles varies a little with the change of
tool diameter. However, in contrast with thread profiles of type K worm, thread
profiles of type C worm can be adjusted to compensate change of tool diameter by
modifying the following four tool dimensions (Fig. 11.7): the radius of cutting edge
profile, q, and mean diameter of the same cutting edge profile, dm0 , and the pressure
angle, a0n , and thickness, s0 at the mean diameter, dm0 .
Heyer and Niemann [20] proposed these type C worms. These worms have the
great advantage of improving the conditions of lubrication, by virtue of the
favorable shape of lines of contact between the worm and worm wheel mating
surfaces. Consequently, they are characterized by a continuous maintenance of
lubrication oil film and high hydrodynamic lubricant pressure, which reduces
output losses and wear, as well as by a greater load carrying capacity, lesser
frictional losses, and greater impact damping properties and noiselessness.
11.2 Geometry of Worm Thread Profiles … 477
Fig. 11.6 Type C cylindrical worm and machining methods by milling or grinding
Litvin [30] proposed a variant of this type of worm, which is generated using the
same cutting tool described above, but with a different cutting position with respect
to the worm. This difference, which concerns only the setting parameters of the
cutting machine, determines certain advantages over the aforementioned worm
proposed by Heyer and Niemann, summarized as follows by the same Litvin:
478 11 Worm Gears
• the line of contact between the grinding surface of the tool and the worm surface
is a planar curve, i.e. a circular arc of a torus in the axial section;
• the shape of the line of contact does not depend on the diameter of the grinding
wheel and the center distance.
The cutting of the worm wheel teeth is carried out by a generation cutting process,
because the shape of its teeth is not defined through a simple geometry, but only as
the envelope of the worm threads. Worm gear hobs are therefore used, having the
same shape of the actual worms, but with tip diameters that allow for the clearances
between the tip surfaces of the actual worms and root surfaces of the worm wheels.
Hob and worm wheel to be cut rotate about their axes with transmission ratio equal
to that of the actual worm gear pair.
In order to cut the tooth progressively over its entire depth, a cylindrical worm
gear hob can have a radial approach motion towards the worm wheel, until the
desired depth (Fig. 11.8a). This radial approach, however, produces cutting inter-
ferences and thus undercut, which determine the removal of parts of the usable
flank profile. The use of this type of cutting process is therefore limited to values of
the helix angle, b, not exceeding ð6 8Þ . For higher values of b, but also in
general, conical worm gear hobs can be used, equipped with axial motion in
11.2 Geometry of Worm Thread Profiles … 479
(b)
(a)
Fig. 11.8 Generation cutting process of worm wheels by: a cylindrical worm gear hob; b conical
worm gear pair
addition to motion of rotation about their axes, in so that the final tooth depth is
achieved gradually, as the hob deeper threads are progressively involved in the
cutting process (Fig. 11.8b).
In the more general case, worm and worm wheel rotate about axes z1 and z2 that
form, in a plane perpendicular to the shortest distance or center distance, a, a shaft
angle R R p=2. Figure 10.8, which refers to any crossed helical gear pair, can also
be used to show the meshing position between the two members of a worm gear
pair with R\p=2. As coordinate systems we consider three Cartesian coordinate
systems O1 ðx1 ; y1 ; z1 Þ; O2 ðx2 ; y2 ; z2 Þ and O0 ðx0 ; y0 ; z0 Þ; rigidly connected respec-
tively to worm (gear 1), worm wheel (gear 2) and frame or housing (Fig. 11.9). In
some cases, for convenience of calculation, it can be useful to introduce other
auxiliary coordinate systems that will be described from time to time.
Though orientations having other shaft angles are possible, in most practical
applications the members of a worm gear pair are generally mounted on crossed
shafts with shaft angle R ¼ p=2. For this reason, we sometimes consider a worm
gear pair consisting of a cylindrical worm, having right-hand threads, and an
enveloping worm wheel, positioned under the worm, whose axes are arranged with
90° shaft angle (Fig. 11.10). As coordinate system O1 ðx1 ; y1 ; z1 Þ, we choose the
Cartesian coordinate system with z1 -axis coinciding with the worm axis and origin
at point O1 on the worm axis, x1 -axis in the middle plane of the worm wheel and in
its initial position coinciding with the shortest distance (i.e. with the common
perpendicular to the worm and worm wheel axes), and y1 -axis that completes the
direct coordinate system. Figure 11.10 shows this right-hand coordinate system for
a worm gear pair with type A worm.
480 11 Worm Gears
The study of kinematics of a worm gear pair can be carried out based on
observation that the teeth shape of the worm wheel, when the thread profile of the
worm is known, depends only on an axial advance motion of the same worm, which
is precisely the motion that obliges the worm wheel rotates. In fact, the motion of
the worm, which is a rotation about its own axis, can be considered as the difference
of two motions: a helical motion with pitch equal to the lead, and the aforemen-
tioned axial advance motion. The first of these two motion components has no
effect on the worm wheel, that is, it does not cause the rotation of the worm wheel,
since in such a motion the thread of the worm simply slides on itself, and therefore
the only consequence of such a motion is a power loss by friction.
Worm and worm wheel, by transmitting the motion, rotate about their two
crossed axes, and their relative motion is a screw motion, that is an instantaneous
helical motion. Therefore, for the worm gear pair, we cannot talk of pitch surfaces
in the strict sense (see Sect. 10.3). However, with reference to the transmission of
motion, since the worm behaves like a rack of endless length, which drags in
11.3 Coordinate Systems and Main Geometric Quantities … 481
Fig. 11.10 Coordinate systems O1 ðx1 ; y1 ; z1 Þ, and O0 ðx0 ; y0 ; z0 Þ, and geometric quantities related
to kinematics
rotation the worm wheel, we can recognize the existence of a pitch plane for the
rack, and a pitch cylinder for the worm wheel. The relative motion between worm
and worm wheel is therefore defined by a plane, parallel to the worm and worm
wheel axes (and thus perpendicular to the shortest distance), and rigidly connected
to the worm. This plane rolls without sliding over a cylinder, having as axis the axis
of the worm wheel, and rigidly connected to the latter. Conventionally, we call this
plane rigidly connected to the worm and the cylinder rigidly connected to the worm
wheel pitch plane of the worm and, respectively, pitch cylinder of the worm wheel.
The pitch plane, also called pitch surface of the worm, is given as the locus of the
successive positions of the instantaneous axis of rotation in the relative motion of
worm wheel to the worm threads, and is parallel to the worm wheel axis.
As shown in Fig. 11.10, the intersections of the above defined pitch plane and
pitch cylinder with the middle plane of the worm wheel (this plane passes through
the axis of the worm and is normal to the axis of the worm wheel) are respectively
the straight-line p1 and the circle p2 , which are tangent at the pitch point C. They are
respectively the pitch line and the pitch circle of the rack-wheel pair, which we get
sectioning the worm gear pair with the middle plane of the worm wheel. This plane
coincides with the ðx0 ; z0 Þ-plane of the Cartesian coordinate system O0 ðx0 ; y0 ; z0 Þ,
that is with the plane having coordinate y0 ¼ 0. Performing a number of sections of
482 11 Worm Gears
the worm gear pair with planes normal to the worm wheel axis, we get as many
pairs of rack-wheel mating profiles, all corresponding to the same pitch curves p1
and p2 . We call these section planes offset planes; they are perpendicular to the axis
of worm wheel and parallel to an offset from the axis of the worm, and thus have
coordinates y0 ¼ const.
To be precise, it is necessary to remember that the actual pitch surface of the
worm would be the locus of Mozzi’s axis of the relative motion between the same
worm and worm wheel. The case that we are considering here is therefore part of
that discussed in Sect. 10.2, regarding orthogonal crossed axes 1 and 2, and
hyperboloid pitch surfaces. In this case, however, the solution corresponding to
hyperboloid wheels is not advisable from the design point of view in that, wishing
to obtain a high value of the gear ratio (for example, u ¼ 50), from Eq. (10.14) we
would have i2 ¼ 1=2500. Therefore, the Mozzi’s axis would be so close to the
worm axis, so, with a worm wheel of usual size, we would have a worm with an
extremely small diameter. With the value of this transmission ratio i, from
Eq. (10.13) we would get then i ¼ tan b1 ¼ 0:02, and b1 ffi 1 100 ; in other words,
the direction of the Mozzi’s axis would form a small angle with the worm axis, and
the hyperboloid would be very close to a cylinder.
For this reason, we introduce the two conventional pitch surfaces or pseudo-
pitch surfaces defined above. In equivalent terms, for the worm we can use a
pseudo-pitch cylinder rigidly connected to the same worm in place of the previously
mentioned pitch plane. Such pitch surfaces (the pitch plane of the worm and the
pitch cylinder of the worm wheel, or the two pitch cylinders of worm and worm
wheel, with radii r1 and r2 respectively) provide the same main point of contact C,
coinciding with the point of tangency of the pitch surfaces considered in the middle
plane of the worm wheel. Of course, as for other types of gears, we differentiate
between reference pitch surfaces and operating pitch surfaces.
In the more general case in which the shaft angle is different from
p=2 ðR 6¼ p=2Þ, the characteristic quantities of the worm gear pair are those defined
in Sects. 10.4 and 10.5 for crossed helical gears. In the most common case, the shaft
angle R is equal to p=2 and, in this case, the characteristic quantities of the worm
gear pair can be deduced as follows.
During the apparent motion of the worm, which is identified with the translation
motion of the rack to which the worm is reduced along its axis, all its points have
the same value of axial velocity, v. If we denote by x1 and pz respectively the
angular velocity and lead (i.e. the axial distance between two consecutive corre-
sponding profiles of the same worm thread) of the worm, the absolute value of
vector, v, will be given by:
pz
v ¼ x1 : ð11:2Þ
2p
However, the tangential velocity on the pitch cylinder of the worm wheel must
be equal to the translation velocity of the pitch plane of the worm-rack; so, if we
11.3 Coordinate Systems and Main Geometric Quantities … 483
denote by x2 and r2 respectively the angular velocity and radius of the pitch
cylinder of the worm wheel, this last will have to be equal to:
x1 pz
r2 ¼ : ð11:3Þ
x2 2p
The cylinder of radius r2 and the plane tangent to it and perpendicular to the
shortest distance or, in equivalent terms, the same cylinder of radius r2 and the
cylinder of radius r1 , coaxial with the worm and rigidly connected to the latter,
are therefore the pitch surfaces of the worm gear pair. Equation (11.3) shows that
the worm wheel pitch radius, r2 , is determined when the transmission ratio
i ¼ ðx1 =x2 Þ and lead pz of the worm are given. This deduction is valid regardless
of the pitch diameter d1 ¼ 2r1 of the worm, as it is kinematically possible to obtain
the desired transmission ratio whatever the pitch radius r1 of the worm. The shortest
distance between the axes, that is the center distance, varies with change of this
radius, the determination of which also depends on the geometric quantities of the
thread and sliding velocity at the start of meshing, which affect the efficiency of the
worm gear pair.
Figure 11.11 shows the axial section and front view of a four-start cylindrical
worm ðz1 ¼ 4Þ, with right-hand threads. The thread helical surface, i.e. the helicoid
that constitutes the thread flanks, is comprised between two cylindrical surfaces,
coaxial with the helicoid, which are the tip and root
cylinders.
The radial distance
between these cylinders is the thread depth, h ¼ ha þ hf . The nominal size of the
worm is given by the reference or middle cylinder diameter, d1 , whose transverse
section is the middle circle or reference pitch circle. Usually, the value of d1 lies
between ð25 60Þ% of the center distance, a. The axial pitch, px , is related to the
Fig. 11.11 Main quantities of a four-start cylindrical worm, with right-hand threads
484 11 Worm Gears
axial module, mx , by the relationship px ¼ pmx . In the case of worm gear pairs with
driving worm, mx is the reference module on which all calculations and specifi-
cations are based. For this reason, it is customary to denote this module simply by
m, i.e. without any subscript.
The axial pitch px of the worm is the distance, measured along a generatrix of its
pseudo-pitch cylinder, from a point of a given thread to the corresponding point of
the adjacent thread (Fig. 11.11). It corresponds to the axial distance between two
neighboring straight lines which represent the helices of two adjacent threads of the
worm when we develop the pitch cylinder on a plane (Fig. 8.11). The lead pz of the
worm is the axial distance between two consecutive points of the same worm-thread
when the thread helix makes a complete turn about the axis (Fig. 8.9). The axial
pitch and lead of the worm are related as follows:
pz ¼ z 1 px : ð11:4Þ
Thus, the axial pitch is equal to the quotient of the lead divided by the number of
threads (or number of starts). Of course, if the number of starts is one, we have
pz ¼ px . The lead of the worm is an integer multiple of the axial pitch of the same
worm. The number of starts usually does not exceed 4 (exceptionally 6). Only
single-start worms are used when self-locking or irreversibility of the worm drive
system is desired. Note that the worm axial pitch px does not change whatever the
coaxial cylinder of the worm between the tip and root cylinders.
All the intersections of the helicoid with these coaxial cylinders are helices
having the same lead, but different lead angle, c (Fig. 8.9). This is the angle
subtended between a tangent to the helix under consideration and a transverse plane
of the worm. If r is the radius of any coaxial cylinder between tip and root
cylinders, and pz ¼ const the lead of the worm, the lead angle of the helix obtained
as the intersection of this cylinder with the helicoid is given by:
pz pz
tan c ¼ ¼ : ð11:5Þ
2pr pd
Thus, the lead angle c of the helix decreases with increasing diameter d ¼ 2r of
the cylinder under consideration. As reference lead angle c1 of the worm we
assume, conventionally, the one related to the helix on the reference cylinder, or
middle cylinder, having diameter d1 ¼ dm1 ; so, we have:
pz z1 px z1
tan c1 ¼ ¼ ¼ ; ð11:6Þ
2pr1 pd1 d1 Pax
where Pax is the diametral pitch. The reference lead angle c1 on the worm reference
pitch cylinder, and the operating lead angle cw1 on the worm operating pitch
cylinder, having diameter dw1 ¼ 2rw1 , are related by the relationship:
11.3 Coordinate Systems and Main Geometric Quantities … 485
pz
r1 tan c1 ¼ rw1 tan cw1 ¼ ¼ p; ð11:7Þ
2p
where p ¼ ðpz =2pÞ is the screw parameter (attention must be paid to the fact that
here p has a different meaning that it has so far attributed to this symbol).
From Eqs. (11.6) and (11.7) we get:
r1 z1 px z1
tan cw1 ¼ tan c1 ¼ ¼ ; ð11:8Þ
rw1 2prw1 2rw1 Pax
where rw1 is the radius of the chosen operating pitch cylinder. Note that the dif-
ference between rw1 and r1 affects the shape of the lines of contact between the
active flank surfaces of the worm threads and worm wheel teeth.
The helix angle, b, reference helix angle, b1 , and operating helix angle, bw1 , of
the worm are complementary respectively to the lead angle, c, reference lead
angle, c1 , and operating lead angle, cw1 . Thus, we have: b ¼ ½ðp=2Þ c,
b1 ¼ ½ðp=2Þ c1 , and bw1 ¼ ½ðp=2Þ cw1 . Therefore, a worm can be thought as a
helical gear wheel whose teeth make a complete revolution about the pitch cylinder.
In order to obtain a good efficiency of the worm gear pair, it is necessary to have
a lead angle of the worm sufficiently high. The values of cw1 can reach ð30 40Þ ,
for which the lead of the worm is great. If we wanted to use a single-start worm,
with equal values of the thread thickness and thread space, to ensure good conti-
nuity of the transmission of motion, we should choose threads with depth too large.
They would give rise to excessive sliding at the beginning of the meshing, and to
cutting interferences during the generation of the worm wheel teeth. To overcome
these drawbacks, we employ multi-start worms, with less thread depth; the pitch of
their generating rack will be a sub-multiple of the lead of the worm.
We now consider the quantities of the worm related to its operating pitch
cylinder. The normal module, mn , and normal pitch, pn ¼ pmn , are related
respectively to the axial module, mx ¼ m, and axial pitch, px , by the relationships:
As indicative value, we can assume ðdw1 =15 m dw1 =6Þ, with an average
value m ffi 0:1dw1 .
From Eqs. (11.8) and (11.10), we infer:
z 1 pn z1 pn
sin cw1 ¼ ¼ : ð11:11Þ
2prw1 pdw1
The ratio between the reference pitch cylinder diameter, d1 , and the axial
module, mx ¼ m, defines the diameter quotient zF of the worm, also called form
number of the worm, i.e.:
486 11 Worm Gears
d1
zF ¼ : ð11:12Þ
m
The values of the pressure angles used in worm gear pairs depend on the values
of the lead angles. They must be large enough to avoid undercut of the threads on
the flank where the contact ends. If the lead angle increases, the cutting conditions
of the worm wheel cutter become unfavorable. For manufacturing reasons, greater
11.3 Coordinate Systems and Main Geometric Quantities … 487
pressure angles are chosen for larger values of the lead angles. In the normal, axial
and transverse sections of the worm, the reference normal, axial and transverse
pressure angles an ; ax and at are notoriously related by the following relationships:
tan an tan an
tan ax ¼ tan at ¼ ; ð11:16Þ
cos c1 sin c1
therefore, we have:
This equation relates the pressure angles in normal, axial and transverse sections,
and the lead angle of the helix at the reference pitch cylinder. Mutatis mutandis,
similar equations to Eqs. (11.16) and (11.17) can be written for any other coaxial
cylinder between the tip and root cylinders of the worm: of course, profile angles in
normal, axial and transverse sections, and the lead angle of the helix at the cylinder
under consideration are involved.
In the particular case of an involute worm, we have:
From Fig. 8.9, where we put r ¼ r1 and c ¼ c1 , and taking into account
Eq. (11.18), we get:
rb1 tan c1
¼ ¼ cos at : ð11:19Þ
r1 tan cb1
tan ax
tan at ¼ : ð11:20Þ
tan c1
The radius of the base cylinder (see Fig. 8.9), can be written as follows:
pz p p cos at p
rb1 ¼ ¼ ¼ ¼ 1=2
: ð11:21Þ
2p tan cb1 tan cb1 tan c1 tan c1 ð1 þ tan2 at Þ
This equation expresses the base cylinder radius of an involute worm as a function
of the screw parameter, p, reference lead angle, c1 , ad axial pressure angle, ax .
Figure 11.12 shows the transverse and axial sections of a single enveloping
worm wheel for a shaft angle R ¼ p=2. Many of the quantities that define the
488 11 Worm Gears
geometry of a worm wheel refer to its mid-plane, i.e. the plane perpendicular to its
axis and containing the axis of the mating worm. The toothing is between the root
toroid, i.e. the toroidal surface tangent to the root surface of the teeth, and the
outside surface. This surface is formed by the throat form surface, in the region that
straddles the mid-plane (the throat is the portion of the outside surface having
toroidal geometry), and the outside cylinder, in the two outer regions (the outside
cylinder is the cylindrical part of the outside surface). The reference circle of the
worm wheel is the inner circle of intersection of the reference toroid and the
mid-plane. For more details on the worm gear geometry, we refer the reader to ISO
1122-2: 1999 (E/F) [23].
The diameter of the reference circle of the worm wheel is given by:
d2 ¼ z2 m; ð11:23Þ
while its throat diameter, i.e. the diameter of the throat circle at mid-plane, is given
by:
Of course, Eqs. (11.23) and (11.24) refer to worm wheels with zero-shifted
toothing.
The throat form radius, also called radius of worm wheel face, is the radius of
the circle that surrounds the axial section of the throat; it is given by:
da2
rth ¼ a : ð11:25Þ
2
The diameter of the outside cylinder, i.e. the outside diameter, de , of the worm
wheel, is a function of its helix angle, b2 , and depending on whether b2 15 or
b2 [ 15 , is given respectively by:
de ¼ da2 þ m ð11:26Þ
de ¼ da2 þ mn : ð11:27Þ
We define as face width angle (Fig. 11.12) the angle at the center that, in the
generation circle of the reference toroid, is between the points of intersection of this
circle with the end faces of the teeth. The face width, b2 , of the worm wheel is the
distance between two planes perpendicular to the axis, which contain the circles of
intersection of the reference toroid with the end faces of the teeth. In the most
frequent case in which the teeth are symmetrical with respect to the mid-plane, the
face width is the length of the chord (parallel to the axis) of the generating circle of
the reference toroid, between the points of intersection of this circle with the end
faces of the teeth. This quantity is defined using, as a usual guideline, the following
relationship:
pffiffiffiffiffiffiffiffiffiffiffiffi
b2 ffi 2m 0:5 þ zF þ 1 : ð11:29Þ
We define as wheel rim and rim width (Fig. 11.12) the rim that contains the
worm wheel teeth and, respectively, the maximum axial dimension of the rim. In
the case of a worm gear pair with crossed orthogonal axes ðR ¼ p=2Þ, in order to
have a proper meshing, the helix angle of the worm wheel must satisfy the fol-
lowing conditions:
• The lead angle of the worm must be equal to the helix angle of the worm wheel.
• The axial pitch of the worm must be equal to the transverse pitch of the worm
wheel.
Let us consider the general case of a worm gear pair with shaft angle R 6¼ p=2.
Figure 11.13a, b shows respectively the operating pitch cylinders and triangle of
velocity of a right-hand worm gear pair. As Fig. 11.13a shows, the worm is placed
above the worm wheel, and the operating pitch cylinders, having respectively radii
490 11 Worm Gears
(a) (b)
Fig. 11.13 Right-hand worm gear pair: a operating pitch cylinders; b triangle of velocity
rw1 and rw2 , are tangent at pitch point C. In the plane through this pitch point and
normal to the shortest distance, both helices obtained as the intersections of the
operating pitch cylinders with the worm thread surface and worm wheel tooth
surface have a common tangent, the straight-line t–t.
We want to determine first the transmission ratio i ¼ x1 =x2 of the worm gear
pair, considering rw1 , rw2 , cw1 and R as input data (R is measured clockwise from
z0 z1 to z2 ), and assuming as given the direction and magnitude of the angular
velocity vector, x1 . We must first determine the direction and magnitude of the
worm wheel angular velocity vector, x2 , considering that the line of application of
vector x2 is the z2 -axis. As Fig. 11.13b shows, in the plane through the pitch point
C and normal to the center distance, the sliding velocity vector or relative velocity
vector, vr ¼ ðv1 v2 Þ, is collinear to the straight line t–t. In addition, the compo-
nents of the velocity vectors v1 and v2 in the direction of the normal n–n to the
tangent t–t must be equal, that is they must have the same direction and magnitude,
for which we can write:
From the same Eq. (11.30) we also infer the following relationship that gives the
transmission ratio, i, as a function of rw1 , rw2 , cw1 and R, with the condition
R 6¼ cw1 :
where the upper and lower signs correspond respectively to the cases in which
R [ cw1 , and R\cw1 .
Figure 11.14a, b shows respectively the operating pitch cylinders and triangle of
velocity of a left-hand worm gear pair. In this case, by using the same procedure
described above, we obtain:
The normal and axial pitches of the worm are related by Eq. (11.10), but the
normal pitch is the same for the worm and worm wheel. From Fig. 11.13a and
(a) (b)
Fig. 11.14 Left-hand worm gear pair: a operating pitch cylinders; b triangle of velocity
492 11 Worm Gears
taking into account Eq. (11.10), we obtain the following relationship of the
transverse pitch, pt , of the worm wheel:
pn px cos cw1 px cos cw1
pt ¼ ¼ ¼
; ð11:34Þ
cos bw2 cos½ðp=2Þ
ðcw1 RÞ sinðR cw1 Þ
where bw2 is the helix angle on the operating pitch cylinder of the worm wheel. In
this equation, which is valid provided R 6¼ cw1 , the upper and lower signs refer
respectively to the cases for which R [ cw1 , and R\cw1 ; the equation gives a
positive sign for pt .
By using the same procedure and with reference to Fig. 11.14a, for a left-hand
worm gear pair we obtain:
px cos cw1
pt ¼ : ð11:35Þ
sinðR þ cw1 Þ
In the case where R ¼ 90 , from both Eqs. (11.34) and (11.35) we obtain
pt ¼ px .
Since
pt z2 ¼ 2prw2 ; ð11:36Þ
where z2 is the number of teeth of the worm wheel (attention must be made here to
the different meaning of z2 ), from Eqs. (11.34) and (11.35) we obtain:
px z2 cos cw1
rw2 ¼
; ð11:37Þ
2p sinðR cw1 Þ
px z2 cos cw1
rw2 ¼ ; ð11:38Þ
2p sinðR þ cw1 Þ
that refer respectively to right-hand and left-hand worm gear pairs. Note that
Eq. (11.37) is valid provided R 6¼ cw1 , while the upper and lower signs refer
respectively to the cases where R [ cw1 , and R\cw1 .
Both matching Eqs. (11.31), (11.37) and (11.8), and Eqs. (11.32), (11.38) and
(11.8), we get always:
z2
i¼ ¼ u: ð11:39Þ
z1
The center distance, a, between the crossed axes of worm and worm wheel is
given by:
while rw2 is expressed by Eqs. (11.37) or (11.38), depending on whether the worm
has right-hand or left-hand helix.
In the case where R ¼ 90 and the operating pitch cylinders coincide with the
reference pitch cylinders, the center distance is given by:
px z1
a¼ þ z2 : ð11:42Þ
2p tan cw1
where rðu; vÞ is the position vector, and i, j, k are the unit vectors along the x-, y-,
and z-axes, respectively, or in scalar form, as follows:
where ðx; y; zÞ are the Cartesian coordinates of points of the surface, and
xðu; vÞ; yðu; vÞ and zðu; vÞ are some definite, continuous and single-valued functions
of the variable parameters u and v. If we eliminate the parameters u and v from
494 11 Worm Gears
(a)
(b)
Fig. 11.15 a Position vector and coordinate lines of a surface; b tangent plane to a surface at a
regular point
11.5 Elements of Differential Geometry of Surfaces 495
F ðx; y; zÞ ¼ 0; ð11:45Þ
@r @r
ru ¼ rv ¼ ; ð11:47Þ
@u @v
are the tangent vectors at any point of the surface r to the u- and v-coordinate lines,
respectively. The directions of vectors ru and rv coincide, respectively, with
directions of the vectors dru and drv , because u and v are scalar quantities. These
vectors represent the chords joining two neighboring points P and M on u-coor-
dinate line ðv ¼ constÞ, having coordinates u and u þ Du, and two neighboring
points P and N on v-coordinate line ðu ¼ constÞ, having coordinates v and v þ Dv.
In the limit, as Du ! 0 and Dv ! 0, these chords approach the tangents at a point
496 11 Worm Gears
Pðu; vÞ along the coordinate lines u and v, respectively. Since u and v are assumed
to be orthogonal, then the scalar product ru rv must be equal to zero, i.e.:
ru rv ¼ 0: ð11:48Þ
Consider now a small region of a smooth surface r near a current point Pðu; vÞ.
As smooth surface we define a surface that is continuous, i.e. the function F ðx; y; zÞ
and its first derivatives are continuous, and the latter are not simultaneously equal to
zero; in addition, it contains no discontinuity of slope, i.e. singular points, as
creases or vertices. It is to be noted that, in the case of creases, two branches of the
surface have a common line, called edge of regression, and only a half-plane exists
that is limited with the tangent line drawn at any point of this edge of regression,
while in the case of vertices the tangent plane does not exist. If we draw various
curves on the surface r through a regular point P, the tangents to these curves are
placed on one plane called the tangent plane to the surface at point P. A line that is
perpendicular to the tangent plane and passes through point P is called the normal
to the surface at point P, and is denoted by N. Since the surface is smooth, the
tangent plane and, hence, the normal to the surface at point P are uniquely deter-
mined (see also Litvin and Fuentes [33]).
Points of a surface at which a tangent plane does not exist are called singular
points. The tangent plane to a surface r at a regular point P (the point is a regular
point when the tangent plane exists) is determined by a pair of vectors ru and rv that
are tangent to the u-coordinate line and v-coordinate line, respectively
(Fig. 11.15b). We assume that ru 6¼ 0 and rv 6¼ 0 and that vectors ru and rv are not
collinear. The tangent plane at point P0 ðu0 ; v0 Þ of a smooth surface (point P0 is
defined by the position vector rðu0 ; v0 Þ) is given by the scalar triple product (also
called mixed or box product):
a ðru rv Þ ¼ 0; ð11:49Þ
a ¼ b rðu0 ; v0 Þ; ð11:50Þ
position vector b is drawn from the same origin of the position vector rðu0 ; v0 Þ to
an arbitrary point P on the tangent plane (Fig. 11.15b). Equation (11.49) indicates
that vector a ¼ P0 P , applied to point P0 defined by the position vector rðu0 ; v0 Þ,
lies in the plane drawn through vectors ru and rv .
The normal N to the surface r at point P, and therefore perpendicular to the
tangent plane to the surface at same point P, is given by the cross product:
N ¼ ru rv ; ð11:51Þ
11.5 Elements of Differential Geometry of Surfaces 497
its direction depends on the order of factors in the cross product. The normal N may
be written in terms of projections on the coordinate axes, by the relationship:
i j k
yu zu zu xu xu yu
N ¼ xu yu z u ¼ iþ jþ k: ð11:52Þ
xv yv z v yv z v z v xv xv yv
N Nx Ny Nz
n¼ ¼ iþ jþ k; ð11:53Þ
jN j jN j jN j jN j
1=2
provided jN j ¼ Nx2 þ Ny2 þ Nz2 6¼ 0.
The point rðu0 ; v0 Þ of the surface is a singular point if
ru rv ¼ 0; ð11:54Þ
and this happens when at least one of the two vectors ru and rv is equal to zero, or
when the two vectors are collinear.
where ra ð#Þ is the position vector of a current point P on the generation curve, i.e.
the polar equation of the generation curve in the plane za ¼ 0, ra ð#Þ cos # and
ra ð#Þ sin # are the components of the position vector along the xa -axis and ya -axis,
# is the independent variable, which coincides with the angle between the xa -axis
and position vector, while ia and ja are the unit vectors of the xa -axis and ya -axis. It
should be noted that, as Fig. 11.16b shows, the Cartesian coordinate system
Oa ðxa ; ya ; za Þ can be regarded as an auxiliary and movable system. This auxiliary
system, with respect to the fixed coordinate system O1 ðx1 ; y1 ; z1 Þ, rigidly connected
to the worm, has the za -axis that moves uniformly along the z1 -axis width velocity
pf, while the xa -axis and ya -axis are rotated of the angles f with respect to the
498 11 Worm Gears
(a) (b)
Fig. 11.16 a Generating transverse profile of a helicoid in the Cartesian coordinate system
Oa ðxa ; ya ; za Þ; b Cartesian coordinate systems O1 ðx1 ; y1 ; z1 Þ and Oa ðxa ; ya ; za Þ
x1 -axis and y1 -axis. The movable coordinate system Oa ðxa ; ya ; za Þ coincides with
the coordinate system O1 ðx1 ; y1 ; z1 Þ at the moment when the screw motion of the
first with respect to the second begins.
In this way, the generation transverse profile of the helicoid, which has a given
position in the plane with coordinate z1 , will have a position rotated of the angle f
with respect to this position in the plane with coordinate ðz1 þ pfÞ.
In the same auxiliary coordinate system Oa ðxa ; ya ; za Þ, the parametric equations
in scalar form of the current point P on the generation curve of helicoid can be
written as follows:
where ra ð#Þ is the absolute value of the distance of point P from the origin Oa . As
Fig. 11.16a shows, the angle, #, between the xa -axis and position vector ra ð#Þ has a
variability range ð#1 # #2 Þ. The transverse pressure angle, a, at a point, i.e. the
angle between the position vector ra ð#Þ and the tangent to the generation curve at
current point P (Fig. 11.16a) is given by:
ra ð#Þ
a ¼ arctan ; ð11:57Þ
r#
z1 -axis, and the axial displacement of the za -axis along the fixed z1 -axis with
uniform axial velocity pf. Thus, this axial velocity is proportional to f, the coef-
ficient of proportionality being the screw parameter p, which represents the dis-
placement along the z1 -axis corresponding to an angle of rotation f equal to one
radian. Of course, it will be p ¼ pz =2p [see Eq. (11.7)], since the lead pz corre-
sponds to one complete revolution. The sign of p is positive or negative for a
right-hand or left-hand screw motions, respectively.
In the fixed Cartesian coordinate system O1 ðx1 ; y1 ; z1 Þ, the parametric equations
in scalar form of a current point P of the helicoid surface thus generated (this
helicoid is the thread flank surface of the worm) can be written as follows:
where ð#1 # #2 Þ and ð0 f 2pÞ. These equations are obtained taking into
account the Eqs. (11.56) and the following matrix equation, where the coordinate
transformation matrix M 1a in the transition from auxiliary coordinate system
Oa ðxa ; ya ; za Þ to the fixed coordinate system O1 ðx1 ; y1 ; z1 Þ appears:
8 9 8 9 2 38 9
>
> x1 >
> >
> xa > cos f sin f 0 0 > xa >
< = < > = 6 < >
> =
y1 ya sin f cos f 0 07 y
¼ M 1a ¼6
4 0
7 a :
5 ð11:59Þ
>
> z > > z > 0 1 pf > z >
: >1
; : >
> a
; : a>
> ;
1 1 0 0 0 1 1
The same helicoid surface can be described, in equivalent terms, by the fol-
lowing matrix equation:
Using Eqs. (11.55) and (11.60), and taking account of the coordinate transfor-
mation matrix, M 1a , we can write:
where i1 ; j1 ; k1 are the unit vectors of the x1 -axis, y1 -axis, and z1 -axis, and p ¼
ðpz =2pÞ is the screw parameter. This equation is the parametric equation of the
helicoid surface in vector form. It should be noted that the planes # ¼ const and
f ¼ const are the axial and transverse planes of the worm, respectively.
The normal to the helicoid surface in its current point P, i.e. the perpendicular to
the tangent plane to the same surface at point P, is given by the cross product:
N1 1
n1 ¼ ¼
1=2 ½p sinð# þ f þ aÞi1 p cosð# þ f þ aÞj1
jN 1 j p þ r ð#Þ cos2 a
2 2
a
þ ra ð#Þ cos ak1 :
ð11:63Þ
r1 ¼ ðrb cos f l cos a1 sin fÞi1 þ ðrb sin f þ l cos a1 cos fÞj1 þ ðpf l sin a1 Þk1 :
ð11:64Þ
11.6 Parametric Equations of a Helicoid 501
This general vector equation represents a ruled surface, where the l-lines (lines
with f ¼ const) are straight lines, and f-lines (lines with l ¼ const) are helices. It
can be used not only for type I and type A worms, but also for type N worm, whose
helicoid is a ruled surface generated by a generating planar curve lying on a plane
that passes through the perpendicular to the worm axis and forms an angle c1 with
the worm axis.
502 11 Worm Gears
The normal to the helicoid ruled surface in its current point P is given by the
following cross product:
@r1 @r1
N1 ¼
@l @f
ð11:65Þ
¼ ½ðp cos a1 þ rb sin a1 Þ cos f l cos a1 sin a1 sin fi1
þ ½ðp cos a1 þ rb sin a1 Þ sin f þ l cos a1 sin a1 cos fj1 þ l cos2 a1 k1 :
N1 N1
n1 ¼ ¼ ; ð11:66Þ
jN 1 j m
In the particular case of type I worm, the generation straight line coincides with
the tangent to the helix on the base cylinder, and thus generates a helicoid that is an
involute surface. The vector equation of this ruled involute surface can be obtained
from the general vector Eq. (11.64) by setting a1 ¼ c1 ¼ cb (Fig. 11.17);
we get:
r1 ¼ ðrb cos f l cos cb sin fÞi1 þ ðrb sin f þ l cos cb cos fÞj1 þ ðpf þ l sin cb Þk1 :
ð11:68Þ
If l cos cb 6¼ 0, all points of the helicoid involute surface are regular points, and
the equation of the unit normal at these points can be written as follows:
Thus, the direction of this unit normal does not depend on the surface parameter,
l. Therefore, these unit normals have the same direction for all points of the gen-
eration straight line PQ, and the helicoid involute surface is a ruled developable
surface.
Instead, in the particular case of type A worm (the Archimedean spiral worm),
the generation straight line intersects the worm axis, that is the axis of the screw
motion. The vector equation of ruled surface of this worm type (the equation is
valid also for worms that are cut by straight-edged cutters) can be obtained from the
general vector Eq. (11.64) by setting rb ¼ 0. So we get:
N 1 ¼ ðp cos f l sin a1 sin fÞi1 þ ðp sin f þ l sin a1 cos fÞj1 þ l cos a1 k1 : ð11:74Þ
N1
n1 ¼ ; ð11:75Þ
m
Directions of the normal N 1 and unit normal n1 , for the Archimedean spiral
worm, depend on the location of the point on the generation straight line, and thus
they depend on l. Therefore, the helicoid surface of this worm type is a ruled
surface, but not a developable surface.
In the general case of a helicoid surface given by the vector Eq. (11.64), its
profiles in transverse sections can be obtained by cutting the helicoid surface with
transverse planes having equation z1 ¼ c ¼ const. Since the transverse sections
corresponding to the transverse planes z1 ¼ 0 and z1 ¼ c give the same planar
curve, i.e. the same profile, in two different positions, to simplify transformations
we may consider only the transverse plane z1 ¼ 0. Therefore, the transverse sec-
tions may be obtained from the one corresponding to the transverse plane z1 ¼ 0,
after rotation about the z1 -axis through an angle f ¼ c=p. From Eq. (11.64), by
setting z1 ¼ 0, for which
pf
l¼ ; ð11:77Þ
sin a1
504 11 Worm Gears
we obtain the following parametric equations in scalar form of the transverse profile
of the helicoid corresponding to the transverse plane having equation z1 ¼ 0:
The polar equation of the same transverse profile can be written as follows:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffi rhffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffii
ra ð#Þ ¼ x21 þ y21 ¼ rb2 þ ðpf cot a1 Þ2 ; ð11:79Þ
with
y1 rb tan f þ pf cot a1
tan # ¼ ¼ ; ð11:80Þ
x1 rb pf tan f cot a1
where # is the angle between the position vector ra ð#Þ and the x1 -axis
(Fig. 11.16a). In this general case, the profile in the transverse section is an
extended involute.
with
where r and q are the position vectors drawn to point P P1 P2 from two
arbitrary points respectively on the lines of application of angular velocity vectors
xð1Þ and xð2Þ . Here we choose these two points respectively coinciding with the
origins of coordinate systems O1 ðx1 ; y1 ; z1 Þ O0 ðx0 ; y0 ; z0 Þ and O2 ðx2 ; y2 ; z2 Þ.
By virtue of the parallel axis theorem, also known as Huygens-Steiner theorem,
we can express vð2Þ in an equivalent form to Eq. (11.83), replacing xð2Þ with
an equal vector applied to point O1 O0 and adding to it the vector moment
506 11 Worm Gears
a xð2Þ (a is here the shortest distance vector), for which we shall have (see
Levi-Civita and Amaldi [29], Burton [3], Kane and Levinson [26]):
vð2Þ ¼ xð2Þ r þ a xð2Þ ¼ xð2Þ r xð2Þ a ¼ xð2Þ ðr aÞ:
ð11:84Þ
Combining Eqs. (11.81), (11.82) and (11.84), we obtain the following final
expression of the relative velocity vector vð12Þ :
h i
vð12Þ ¼ vð1Þ vð2Þ ¼ xð1Þ xð2Þ r a xð2Þ : ð11:85Þ
The sliding velocity vector vð12Þ in the fixed reference system O0 O0 ðx0 ; y0 ; z0 Þ
is as follows:
ð12Þ
vx0 ¼ xð1Þ xð2Þ cos R y0 xð2Þ sin R z0
ð12Þ ð11:88Þ
vy0 ¼ xð1Þ xð2Þ cos R x0 axð2Þ cos R
ð12Þ
vz0 ¼ ða þ x0 Þxð2Þ sin R:
11.7 Relative Velocity and Coordinate Transformation 507
The same sliding velocity vector vð12Þ in the reference system O1 ðx1 ; y1 ; z1 Þ
rigidly connected to the worm is as follows:
ð12Þ
vx1 ¼ xð1Þ xð2Þ cos R y1 xð2Þ sin R cos u1 z1 axð2Þ cos R sin u1
ð12Þ
vy1 ¼ xð1Þ xð2Þ cos R x1 þ xð2Þ sin R sin u1 z1 axð2Þ cos R cos u1
ð12Þ
vz1 ¼ ða þ x1 cos u1 y1 sin u1 Þxð2Þ sin R:
ð11:89Þ
r2 ¼ M 21 r1 ¼ M 2n M nm M m0 M 01 r1 : ð11:90Þ
where
2 3
ðcos u1 cos u2 þ sin u1 sin u2 cos RÞ ðsin u1 cos u2 cos u1 sin u2 cos RÞ sin u2 sin R a cos u2
6 ðcos u1 sin u2 sin u1 cos u2 cos RÞ þ ðsin u1 sin u2 þ cos u1 cos u2 cos RÞ cos u2 sin R a sin u2 7
M21 ¼6
4
7;
5
sin u1 sin R cos u1 sin R cos R 0
0 0 0 1
ð11:91Þ
while M 01 is the rotational matrix about the z0 -axis, M m0 is the translational matrix
from O0 ðx0 ; y0 ; z0 Þ to Om ðxm ; ym ; zm Þ, M nm is the rotational matrix about the xm -axis
to obtain the shaft angle R, and M 2n is the rotational matrix about zn -axis. Om
Om ðxm ; ym ; zm Þ and On On ðxn ; yn ; zn Þ are two auxiliary systems, both rigidly
connected to the frame, which we introduce to make the coordinate transformation
easier and more evident (Fig. 11.18). In the case where R ¼ p=2, the matrix M 21 is
simplified very much, since cosðp=2Þ ¼ 0 and sinðp=2Þ ¼ 1.
From Eqs. (11.90) and (11.91) we obtain:
x2 ¼ ðcos u1 cos u2 þ sin u1 sin u2 cos RÞx1 ðsin u1 cos u2 cos u1 sin u2 cos RÞy1
ðsin u2 sin RÞz1 þ a cos u2
y2 ¼ ðcos u1 sin u2 sin u1 cos u2 cos RÞx1
þ ðsin u1 sin u2 þ cos u1 cos u2 cos RÞy1 ðcos u2 sin RÞz1 a sin u2
z2 ¼ ðsin u1 sin RÞx1 þ ðcos u1 sin RÞy1 þ ðcos RÞz1 :
ð11:92Þ
508 11 Worm Gears
ð11:93Þ
From equation r1 ¼ M 12 r2 ¼ M 1
21 r2 , we obtain:
x1 ¼ ðcos u1 cos u2 þ sin u1 sin u2 cos RÞx2 ðcos u1 sin u2 sin u1 cos u2 cos RÞy2
þ ðsin u1 sin RÞz2 a cos u1
y1 ¼ ðsin u1 cos u2 cos u1 sin u2 cos RÞx2
þ ðsin u1 sin u2 þ cos u1 cos u2 cos RÞy2 þ ðcos u1 sin RÞz2 þ a sin u1
z1 ¼ ðsin u2 sin RÞx2 ðcos u2 sin RÞy2 þ ðcos RÞz2 :
ð11:94Þ
Equations (11.92) and (11.94) are respectively the Cartesian coordinate trans-
formation in transition from reference system O1 ðx1 ; y1 ; z1 Þ to reference system
O2 ðx2 ; y2 ; z2 Þ, and vice versa.
To be able to define completely the geometry of the worm gear pair, it is necessary
to determine the surface of contact, i.e. the geometrical surface defined by contact
points between the worm and worm wheel. The successive lines of contact during
gear meshing constitute this surface, whose axial length is the sum of the length of
approach and the length of recess. When the gear is working as speed reducing unit,
the length of approach is the axial distance between the first point of contact of
threads, and the instantaneous axis of rotation, while the length of recess is the axial
distance between the last point of contact of threads, on withdrawal, and the in-
stantaneous axis of rotation. Worm face width, and face width and rim width of the
worm wheel (definitions are those of the ISO 1122-2:1999 [23]; see also Figs. 11.1
and 11.12) strongly depend on the extension of the surface of contact, which
therefore needs to be evaluated as accurately as possible. Moreover, the choice of a
worm face width larger than necessary does not give rise to disadvantages; indeed,
it gives a certain freedom in adjusting its axial position relative to the worm wheel.
In this framework, it is necessary to write the equation of meshing, that relates
parameters of the helicoid surface r1 of the worm and its angle of rotation u1 about
the z1 -axis with the corresponding parameters of the worm wheel surface r2 . The
surface r1 of the worm thread flanks and surface r2 of the worm wheel tooth flanks
11.8 Worm and Worm Wheel Meshing, and Lines of Contact 509
are in contact along curved lines at every instant during the meshing cycle. The
determination of these instantaneous curves of contact is based on the condition that
the normal N to any point P of this curve and the relative velocity vector vr , whose
line of application is the common tangent at the same point, are perpendicular
vectors.
In fact, when we consider the screw motion of a helicoid, the screw parameter
p in this motion and the screw parameter of helicoid coincide. Any point of the
helicoid traces out a helix, and the relative velocity vector vr in the screw motion is
tangent to the helix. Since the helix belongs to the helicoid and the relative velocity
vector vr (it is designed v, to avoid using too many subscripts) is tangent to helicoid,
the following condition must be satisfied:
n v ¼ N v ¼ 0: ð11:95Þ
In general terms, the relative velocity vector, i.e. the velocity vector in screw
motion, is given by:
i j k
v ¼ ðx rÞ þ px ¼ 0 0 x þ pxk ¼ xðyi þ xj þ pkÞ; ð11:96Þ
x y z
while the normal N and unit normal n to the surface r are given by:
N ¼ Nx i þ Ny j þ Nz k ð11:97Þ
n ¼ nx i þ ny j þ nz k: ð11:98Þ
N i vi ¼ 0: ð11:99Þ
ð12Þ
where vi ¼ vi1 vi2 ¼ vi is the relative velocity vector of surface r1 with respect
to surface r2 , N i is the normal to the worm surface r1 , while the subscripts
ði ¼ 1; 2; 0Þ designate coordinate systems O1 ðx1 ; y1 ; z1 Þ, O2 ðx2 ; y2 ; z2 Þ, and
O0 ðx0 ; y0 ; z0 Þ, which are rigidly connected to the worm, worm wheel and
housing-frame.
Taking into account that the worm thread surface r1 is a helicoid, the equation of
meshing can be simplified in one of the following two relationships proposed by
Litvin [30, 31]:
Introducing Eq. (11.89) in the equation above, and dividing by ðx2 sin RÞ all
members of the equation thus obtained, we get:
u cos R
ðz1 cos u1 þ a cot R sin u1 ÞNx1 þ y1 Nx1 þ ðz1 sin u1 þ a cot R cos u1 ÞNy1
sin R
u cos R
x1 Ny1 ðx1 cos u1 y1 sin u1 þ aÞNz1 ¼ 0:
sin R
ð11:104Þ
Multiplying by ½ðu cos RÞ= sin R all terms of the left-hand side of
Eq. (11.100), we infer that:
ðz1 cos u1 þ a cot R sin u1 ÞNx1 þ ðz1 sin u1 þ a cot R cos u1 ÞNy1
u cos R ð11:106Þ
ðx1 cos u1 y1 sin u1 þ aÞ p Nz1 ¼ 0:
sin R
This is the equation that expresses the meshing condition in the coordinate
system O1 ðx1 ; y1 ; z1 Þ. In this equation, Nx1 ; Ny1 ; Nz1 are the projections of the
normal N 1 to surface r1 on the x1 -, y1 -and z1 -axes, R is the shaft angle, a is the
shortest distance or center distance, and u ¼ z2 =z1 is the gear ratio.
From Eqs. (11.99), (11.97) and (11.102) written in the coordinate system
O0 ðx0 ; y0 ; z0 Þ, we obtain:
11.8 Worm and Worm Wheel Meshing, and Lines of Contact 511
Introducing Eqs. (11.88) in the equation above, and dividing by ðx2 sin RÞ all
terms of the equation thus obtained, we get:
u cos R
z0 Nx0 y0 Nx0 x0 Ny0 aNy0 cot R þ ðx0 þ aÞNz0 ¼ 0: ð11:108Þ
sin R
where M 01 ¼ M 1 10 is the rotational matrix about the z0 -axis for the coordinate
transformation in transition from coordinate system O1 ðx1 ; y1 ; z1 Þ to coordinate
system O0 ðx0 ; y0 ; z0 Þ. Thus, we have:
ra
N0 ¼ ½p sinð# þ f þ a þ u1 Þi0 p cosð# þ f þ a þ u1 Þj0 þ ra cos ak0 ;
cos a
ð11:112Þ
where ra ¼ ra ð#Þ is the magnitude of the position vector of the current point of the
worm transverse profile, and u1 is the rotation angle about the z0 -axis of the
reference system O1 ðx1 ; y1 ; z1 Þ with respect to reference system O0 ðx0 ; y0 ; z0 Þ.
Introducing into Eq. (11.110) the expressions of Nx0 , Ny0 and Nz0 obtained from
Eq. (11.112), we get:
512 11 Worm Gears
u cos R
ra cos a x0 þ a p þ ap cot R cos w ¼ pz0 sin w ¼ p2 f sin w;
sin R
ð11:113Þ
Any of the three Eqs. (11.106), (11.110) and (11.113) expresses the relationship
existing between the worm surface parameters ð#; fÞ and the angle of rotation u1 of
the worm, i.e. the function:
f ð#; f; u1 Þ ¼ 0: ð11:115Þ
where r1 ¼ r1 ð#; fÞ represents the helicoid surface r1 of the worm, describe in the
coordinate system O1 ðx1 ; y1 ; z1 Þ the family of lines of contact on surface r1 .
A given family of lines of contact corresponds to a given value of the angle of
rotation u1 of the worm. Thus u1 is the parameter of the family of lines of contact.
The following pair of equations determine the lines of contact on the worm
wheel surface r2 :
(a)
(b)
Fig. 11.19 Lines of contact on: a worm surface, r1 ; b worm wheel surface, r2
surface. Near the envelope line, the condition of lubrication and heat transfer
become unfavorable. For this reason, this envelope line should be excluded from
meshing, by choosing appropriate design parameters of the worm gear pair.
On the basis of a theorem on the necessary and sufficient conditions of existence
of the envelope line on the generation surface, proposed by Litvin et al. [32], it is
possible to demonstrate that, in the case of a conventional worm gear pair with type
I worm, the lines of contact on the worm thread surface, r1 , have an envelope line.
Since each current line of contact has two branches that are recognized by the sign
of the first derivative fu1 ¼ df =du1 with respect to the variable u1 of the function
f ð#; f; u1 Þ ¼ 0 given by the second Eq. (11.116), the worm wheel tooth surface,
r2 , which is the envelope of the worm thread surface, r1 , is formed by two
branches, designed by r12 and r22 (Fig. 11.20). These two branches, on the worm
wheel tooth surface, r2 , are divided by a common separation line, which is the
analogous of the envelope line on the worm thread surface, r1 .
514 11 Worm Gears
It is finally to be noted that the instantaneous lines of contact exist only for an
ideal worm gear pair, without errors of manufacturing and misalignments. Actually,
the contact between the surfaces r1 and r2 is an instantaneous point of contact that,
due to the aforementioned errors, can suffer from hazardous shifts toward the edge
(this is the bearing edge contact), and can adversely affect the function that
expresses the transmission errors. Such transmission errors cause vibration during
the meshing cycle.
Therefore, it is necessary to minimize the errors of manufacturing and
misalignments. To this end, it is necessary to define as accurately as possible the
shape and geometry of the bearing contact between the two mating surfaces r1 and
r2 , by reasonable averaging the results achievable with the use of theoretical and
theoretical-numerical models, but not forgetting to treasure, when available, to any
previous experience accumulated on this topic. The discussion described above,
concerning the determination of the instantaneous lines of contact, is in any case
basilar.
11.9 Surface of Contact: General Concepts and Determination … 515
The surface of contact, also called zone of contact or meshing surface, is the active
part of the surface of action. This is the surface swept out by the lines of contact
during the meshing cycle; thus, it consists of the set of instantaneous lines of
contact between the active flanks of the worm threads and worm wheel teeth. The
determination of the meshing surface is a very important aspect of the worm gear
pair design, since it determines the operating worm and worm wheel face widths,
and the worm wheel rim width.
To avoid confusion and misunderstandings, we clarify first some basic concepts.
To this end, we consider the orthogonal and right-hand worm gear pair shown
schematically in Fig. 11.21. Assuming that the worm is the driving member of the
gear pair, and considering the directions of rotation shown in this figure, we call
advancing and receding sides (front and rear parts, or entering and leaving sides,
are respectively synonyms) of the worm and worm wheel the sides (or parts)
designed with a:s: and r:s: in the same figure. The contact starts between the
Fig. 11.21 Advancing (a:s:) and receding (r:s:) sides of the contact between the worm threads
and worm wheel teeth
516 11 Worm Gears
advancing side of the worm and the receding side of the worm wheel, then it also
extends to the advancing part of the worm wheel and, finally, ends between the
receding sides of the worm and worm wheel.
Figure 11.22 shows the projection of the surface of contact of an orthogonal
worm gear pair with type A worm, on the pitch plane of the same worm, or on any
plane parallel to it, and therefore perpendicular to the shortest distance line; one of
these planes is the plane ðy0 ; z0 Þ. Such a projection, which has roughly the shape of
a horse-shoe, is bounded by the following two lines:
• The line, a, which is a smooth curve, and represents the end of the contact.
• The line, b, which is often a smooth curve at times, and represents the beginning
of the contact; the geometry of the latter curve varies depending on the shape of
the outside surface of the worm wheel.
Figure 11.22 also shows that the horse-shoe shaped surface is approximately
symmetric with respect to a straight line with direction normal to the worm thread at
radius of the pitch cylinder. Therefore, with respect to the trace of the worm wheel
mid-plane, it is the more asymmetric the greater the helix angle, b The contact starts
at the beginning of the line, b, which turns its concavity towards the outside of the
surface of contact, and ends at the top of the line, a, that instead turns outwardly its
convexity.
The position and shape of the line, b, depend on the geometry of the outside
surface of the worm wheel. From this point of view, we can distinguish the three
following different geometries of the line, b:
• The type-A line, which corresponds to an outside surface of the worm wheel
made, in the central part straddles the mid-plane, from the throat surface, and in
the two side parts, made from two cylindrical surfaces (Fig. 11.23c). The two
almost linear ends of the line, b, correspond to these cylindrical side parts of the
worm wheel toothing.
• The type-B line, which corresponds to an outside surface of the worm wheel
made of a single cylindrical surface, whereby the throat surface is missing
(Fig. 11.23d).
• The type-C line, which corresponds to an outside surface of the worm wheel
made, in the central part straddles the mid-plane, from the throat surface, and in
the two side parts, made by two conical surfaces, whose generatrices are
symmetric with respect to the mid-plane, and pass through the worm axis, as
well as the end points of the throat surface (Fig. 11.23e).
The same Fig. 11.23 shows the projections of the instantaneous lines of contact
between a single thread of the worm and a single tooth of the worm wheel on a
transverse plane of the worm (Fig. 11.23a), and the projections of the surfaces of
contact on the pitch plane of the worm (Fig. 11.23b).
After these necessary clarifications, we face the problem of determination of the
surface of contact, which constitutes the subject of this section. Notoriously, this
determination can be performed using analytical-numerical methods, graphical
methods, and mixed analytical-numerical-graphical methods. Here we use an
analytical-numerical method. Among the analytical methods described in the sci-
entific literature, we accord our preference to that proposed by Litvin and Fuentes
[33]. This method has very general validity, because it can be applied to the various
types of worms described in Sect. 11.2. It assumes that the usable surface of the
worm threads is represented as a generalized helicoid, that is, as a surface generated
by a planar curve in transverse section, r ¼ r ð#Þ, which is performing a screw
motion about the worm axis.
Figure 11.24a, b shows, in schematic form, an orthogonal worm gear pair and,
respectively, its surface of contact, projected on the plane ðy0 ; z0 Þ, where it is
bounded by the curves a and b defined above. This surface, represented in the fixed
coordinate system O0 ðx0 ; y0 ; z0 Þ, which has the worm axis as its z0 -axis, does not
518 11 Worm Gears
(a)
(c)
(b) (d)
(e)
Fig. 11.23 Right-hand worm gear pair, with Archimedean spiral worm: a projections of the
instantaneous lines of contact on a transverse plane of the worm; b projections of the surfaces of
contact on the pitch plane of the worm; c type-A worm wheel; d type-B worm wheel; e type-
C worm wheel
change in any plane parallel to the plane ðy0 ; z0 Þ, including the pitch plane of the
worm. As we said above, line b consists of the point corresponding to the entry into
meshing of those points of the worm wheel tooth flanks belonging to the worm
wheel addendum cylinder; line a consists instead of the points corresponding to the
output from the meshing of those points of the thread flanks belonging to the worm
addendum cylinder.
11.9 Surface of Contact: General Concepts and Determination … 519
(b)
The three coordinates ðxP0 ; yP0 ; zP0 Þ that uniquely define the position of the
current point P of the line a in the fixed coordinate system O0 ðx0 ; y0 ; z0 Þ are as
follows:
where ra ¼ O0 P0 is the distance from the worm axis of the point P0 (Fig. 11.24a) on
the worm addendum cylinder, corresponding to current point P, #a is the angle
between the x0 -axis and straight line O0 P0 , and aa is the pressure angle at the tip
point P0 ; the latter is the angle between the straight line O0 P0 and the tangent to the
curve ra ¼ ra ð#Þ at tip point P0 . Equations (11.118)–(11.120) are written in the
order in which the coordinates yP0 ; zP0 and xP0 are usually calculated. The quantities
ra , #a , and aa are considered as known. The input for calculation of these coor-
dinates is the value of the coordinate yP0 . Varying yP0 , we can determine the
corresponding values zP0 and xP0 of curve, a. Note that Eq. (11.118) provides two
solutions for the angle ð#a þ f þ u1 Þ, but only the solution for which xP0 \0 can be
used. This is because the surface of contact is localized all below the plane ðy0 ; z0 Þ,
as Fig. 11.24a shows.
The determination of the three coordinates ðxQ0 ; yQ0 ; zQ0 Þ that uniquely define the
position of the current point Q of the curve, b, in the fixed coordinate system
O0 ðx0 ; y0 ; z0 Þ is more complex, since it depends on the shape of the worm wheel
outside surface. This determination is here limited to the case that most often
characterizes the actual practical applications, in which the shape of the worm wheel
outside surface is the one shown in Figs. 11.23c and 11.24a. The side parts AB and
CD of this worm wheel outside surface (Fig. 11.25) are constituted by two cylinders
having equal radius, re ¼ de =2, and coaxial with the worm wheel. The central part of
the same worm wheel outside surface (BC in Fig. 11.25) is an arc of toroid, the
throat, having throat form radius rth and axis coinciding with the worm axis.
The intersection of the generating arc BC of the toroidal surface of the throat
with an offset plane m-n (Fig. 11.25), i.e. a plane parallel to the worm wheel
mid-plane, and having coordinate y0 ¼ const, is a circle of radius re . Point Q of the
line, b (Fig. 11.24), can be determined as the intersection of the curve z0 ¼ z0 ðx0 Þ
and the circle of radius re (Fig. 11.26). The curve z0 ¼ z0 ðx0 Þ is obtained as the
intersection of the surface of contact and the offset plane having coordinate
y0 ¼ const.
In order to determine the current point Q of the line b corresponding to the
toroidal surface BC of the worm wheel outside surface, we use the following
equations:
n o1=2 h i1=2
½a þ xQ0 ð#; ðf þ u1 ÞÞ2 þ z2Q0 ð#; ðf þ u1 ÞÞ a þ rth
2
y2Q0 ð#; ðf þ u1 ÞÞ
¼ f4 ð#; ðf þ u1 ÞÞ ¼ 0:
ð11:124Þ
Here rth ¼ rf 1 þ c, where rf 1 is the radius of the worm root cylinder, and c is the
clearance (usually, c ¼ 0:25m). In addition, tan a ¼ ½r ð#Þ=ðdr ð#Þ=d#Þ. Coordinate
yQ0 is considered as the input data for solution procedure, while Eqs. (11.122) and
(11.123) are used for determination of zQ0 ð#; ðf þ u1 ÞÞ and xQ0 ð#; ðf þ u1 ÞÞ that
appear in Eq. (11.124). Equations (11.121) and (11.124) can be considered as a
522 11 Worm Gears
The procedure of solution of these equations does not change compared to that
previously described. This time we consider, as a system of two non-linear equa-
tions in the two unknowns # and ðf þ u1 Þ, the system consisting of the two
Eqs. (11.121) and (11.125), while the Eqs. (11.122) and (11.123) are used for
determination of coordinates zQ0 and xQ0 for current point Q.
The determination of the surface of contact allows us to define the operating
worm and worm wheel face widths, and the worm wheel rim width.
Example 1 In order to better clarify the conclusion that can be drawn based on the
analysis of the surface of contact between the worm and worm wheel, as a first
example, we consider (Fig. 11.23) a worm gear pair with a type A right-hand worm,
i.e. a right-hand Archimedean spiral worm. This worm gear pair has the following
characteristics: number of threads, z1 ¼ 4; nominal helix angle, b1 ¼ 30 ; nominal
module of the generating rack, m ¼ 9 mm; pressure angle of the generating rack,
a ¼ 30 ; lead, pz ¼ 113:10 mm; nominal pitch diameter, d1 ¼ 2r1 ¼ 62:35 mm; tip
diameter, da1 ¼ 80:35 mm; root diameter, df 1 ¼ 41:35 mm. Instead, the worm
wheel has the following characteristics: number of teeth, z2 ¼ 28; nominal module,
m ¼ 9 mm; nominal pitch diameter, d2 ¼ z2 m ¼ 252 mm; outside surface config-
ured according to the three different shapes, A, B, and C, as Fig. 11.23c–e shows.
The center distance is a ¼ 167:17 mm.
Figure 11.23b shows the projections on the worm pitch plane of the surfaces of
contact related to the three types of outside surfaces described above, and
specifically:
• The surface of contact bounded by the mixed-line curve umprqsu, that corre-
sponds to the worm wheel outside surface type-A.
• The surface of contact bounded by the mixed-line curve vmwqv, that corre-
sponds to the worm wheel outside surface type-B.
• The surface of contact bounded by the mixed-line curve zmnoqtz, that corre-
sponds to the worm wheel outside surface type-C.
Figure 11.23a shows the projections of the instantaneous lines of contact
between a single thread of the worm and a single tooth of the worm wheel on a
transverse plane of the worm.
Figure 11.23b and, albeit in schematic form, Fig. 11.22 shows that the extension
of the surface of contact varies with the shape of the outside surface of the worm
wheel. From the analysis of the three surfaces of contact shown in these figures, we
infer that the meshing surface is more extensive in the advancing side of the worm
with respect to what happens in its receding side, and that the outside surface of the
worm wheel influences the extension of the contact in the advancing side of the
worm. By increasing the number of teeth of the worm wheel, z2 , the surface of
contact lengthens. In addition, we infer that the outside surface of the type-C worm
wheel (Fig. 11.23d and mixed-line curve C-C in Fig. 11.22) is not very favorable,
since the initial contacts take place according to instantaneous lines of small length,
which therefore carry a very low contribution to the load bearing capacity at the
524 11 Worm Gears
start of the meshing. Furthermore, the sharp edge that is generated in the tooth of
this type of worm wheel generally causes scoring and poor lubrication.
For this reason, worm wheels with outside surfaces type-A and type-B are
usually used. The type-B outside surface of the worm wheel (Fig. 11.23d) involves
the minimum surface of contact (see the mixed-line curve B-B in Fig. 11.22). The
contact, just started, extends once to the worm wheel face width, so the gradualness
of the contact fails. The best solution is therefore to use a worm wheel with type-
A outside surface. In fact, it is characterized by the maximum extension of the
surface of contact, as well as by a gradual contact at the beginning of meshing. We
also note that, according to the length of the surface of contact, for worm wheel
with type-A outside surface, three teeth are simultaneously in meshing, while for
worm wheel with type-B outside surface only two teeth are simultaneously in
meshing. The load-bearing capacity is, however, almost equal (the differences are
small), because the main contribution is due to contacts in the vicinity of the
mid-plane of the gear pair, i.e. near the pitch point C.
The worm must have the advancing and receding sides of equal lengths because,
when the direction of rotation is reversed, the advancing and receding sides of the
worm and worm wheel are exchanged between them. Thus, the usable worm face
width (i.e. the length of the threaded part of the worm that is actually usable, which
is denoted by its flanks being completely formed) must be at least twice compared
to that of the advancing side of contact. It is often desirable to have worm wheels
with type-B outside surface, as the corresponding worm is shorter, and therefore the
worm is less sensitive to unavoidable small errors in the pitch. To establish
definitively what type of worm wheel outside surface to be adopted, we must also
keep in mind that the side parts of the teeth of the worm wheels with type-A and
type-C outside surfaces may be too thin, in some cases in which the pressure angle
is large.
Finally, a consideration must be made regarding the Archimedean spiral worm.
In this case, if we choose as the generatrix of the pitch cylinder a straight line that
does not divide the tooth depth of the rack according to the standard sizing, but
instead it is closest to its axis, the extension of the surface of contact in the receding
side of the worm increases. This position of the generatrix closest to the worm axis
is only possible with large pressure angles and with high numbers of teeth, to avoid
interference. In the projection on a transverse section plane of the worm, the
instantaneous lines of contact will have greater inclinations compared to the con-
centric circles in the same transverse section (Fig. 11.23a).
Example 2 Let us now analyze, as a second example, the surface of contact
between the worm and worm wheel of a worm gear pair with involute worm having
the following characteristics: type I right-hand worm; number of threads, z1 ¼ 4;
base helix angle, bb ¼ 30 ; nominal module of the generating rack, m ¼ 9 mm;
nominal pressure angle of the generating rack, a ¼ 20 580 ; lead, pz ¼ 113:10 mm;
base cylinder diameter, db ¼ 62:35 mm; tip diameter, da1 ¼ 101:35 mm; root
diameter, df 1 ¼ 62:35 mm; pitch diameter, d1 ¼ 83:35 mm; nominal helix angle,
b1 ¼ 23 210 3000 . Instead, the worm wheel has the following characteristics: number
11.9 Surface of Contact: General Concepts and Determination … 525
(b)
threads of the worm are simultaneously in meshing with the teeth of the worm
wheel. For the same diameter of the helicoid base cylinder, the nominal pressure
angle of the thread increases with increasing pitch diameter of the worm.
It should be noted finally that Figs. 11.22, 11.23 and 11.27 refer to worm gear
pairs with zero profile-shifted toothing. In the case of worm gear pairs with
profile-shifted toothing, the projections of the surfaces of contact on the pitch plane
of the worm are translated parallel to themselves, compared to the position related
to the worm gear pairs with zero profile-shifted toothing. Figure 11.28 highlights
this change of position for a worm gear pair with Archimedean spiral worm (type A
worm) having the following characteristics: z1 ¼ 2; z2 ¼ 30; r1 ¼ d1 =2 ¼ 46 mm;
mx ¼ m ¼ 8 mm; rw1 ¼ r1 þ xm; a ¼ ½r1 þ ðz2 mÞ=2 þ xm. Two values of the shift
coefficient x are considered, namely x ¼ 0 and x ¼ 1.
Figure 11.28a shows the surface of contact of the worm gear pair considered
above, in the case of zero profile-shifted toothing ðx ¼ 0Þ. It is quite centered,
against the projection of the pitch point. Note that, when the direction of rotation is
reversed, also the surface of contact changes its orientation, assuming a position
which is the anti-symmetric image of the previous one with respect to the origin O0
of the coordinate system ðy0 ; z0 Þ; it is represented by a dashed line. Obviously, to
exploit the whole surface of contact in all its extension, for both directions of
rotation, it is necessary that the actual values of the worm face width and worm
wheel face width (here, this latter coincides with the rim width, as Fig. 11.25
shows) are greater respectively of b1 and b2 , although not by much.
Figure 11.28b shows the surface of contact of the same worm gear pair above,
but with profile-shifted toothing ðx ¼ 1Þ. It is fully offset, against the projection of
the pitch point. Also, here, when the direction of rotation is reversed, also the
surface of contact changes its orientation, assuming a position which is the
anti-symmetric image of the previous one with respect to the origin O0 of the
coordinate system ðy0 ; z0 Þ. In this case, to exploit the whole surface of contact in all
its extension, for both directions of rotation, the worm face width must be greater,
and not a little, compared to that of the previous case. On the contrary, the worm
wheel face width, well here coinciding with the rim width, does not differ much
from that of the previous case.
(b)
point (see Sect. 1.4). In the case of a worm gear pair, a current point P on the worm
thread surface becomes a point of contact, at a given instant of the meshing cycle,
when the normal to the surface in this point meets the instantaneous axis of rotation.
This instantaneous axis of rotation is the straight line where the pitch plane of the
worm-rack is tangent to the pitch cylinder of the worm wheel. In fact, if we consider
the section of the worm thread with an offset plane having coordinate y1 ¼ const,
another condition must be satisfied by virtue of the aforementioned law of conju-
gate profile. In fact, it must happen that, when the contact is at point P, the normal
to the profile at P in the section under consideration must pass through the pitch
point C, which is the trace of the instantaneous axis of rotation in this section plane.
It is easy to see that such a normal is nothing more than the normal projection of
the normal to the thread surface at point P, on the offset plane under consideration.
In fact, the normal to a given elliptic point of any surface, having Gaussian cur-
vature greater than zero, is normal to the countless curves lying on the surface and
passing through that point. Therefore, in a plane containing the tangent to one of
these curves, the normal to the surface is projected in accordance with the normal at
that point of the curve obtained as the intersection of the surface with that plane (see
Novozhilov [45], Ventsel and Krauthammer [66]).
Using the above basic law of conjugate profiles, Schiebel (see Schiebel [56],
Schiebel and Lindner [57, 58]) proposed the following method, which allows to
obtain, in the various planes normal to the worm axis, the instantaneous lines of
contact, and profiles of the worm wheel teeth, thus solving the kinematic problem of
the worm gear pair sizing. The Schiebel’s elegant method also uses another
property of the Archimedean spiral worm, the threads of which, in the axial section,
configure a rack with teeth having straight flanks, shaped as an isosceles trapeze,
and with pressure angle a. According to this property, the normal to the worm
thread surface in its current point P is projected, regardless of whether this point is
or not a point of contact, in a transverse plane, according to a straight line. This
straight line joins the projection P0 of point P in this plane with point A, which is
arranged at 90 with respect to P0 , and belongs to the circle of radius e, given by:
the radius of this circle is a constant for all points of the surface (Fig. 11.29).
In fact, given any current point P on the thread surface (in Fig. 11.29, this point
is localized in the section belonging to the plane of the same figure), at a distance
r from the worm axis, the normal to the surface in this point must satisfy the
following conditions:
11.10 Surface of Contact: Determination by Graphic-Analytical Methods 529
• It must be normal to the generation straight line passing through the point P, and
thus contained in the plane normal to this straight line passing through this
point; this plane has trace PM 0 in Fig. 11.29.
• It must be normal to the helix with lead pz and helix angle c, passing through
point P, and thus having, in the plane view, projection PA inclined by an angle c
with respect to the projection of worm axis.
The point A of intersection of the normal to the helicoid surface with the plane
passing through the worm axis, and forming an angle of p=2 compared to that
passing through point P0 , is distant from the worm axis of the quantity, e, given by:
As
pz p
tan c ¼ ¼ ; ð11:128Þ
2pr r
we infer:
pz
e ¼ r cot a ¼ p cot a ¼ const: ð11:129Þ
2pr
It should be noted that the point A would be localized, in Fig. 11.29, in the
position symmetrical with respect to the point M, if the worm had left-hand helix,
rather than right-hand helix, such as that in Fig. 11.29.
530 11 Worm Gears
Fig. 11.30 Condition under which a current point P on the thread surface is a point of contact, for
an Archimedean spiral worm
We determine now the position where a current point P, whose distance from the
worm axis is r, and belonging to the worm transverse plane having coordinate
z1 ¼ const, becomes a point of contact. Figure 11.30 shows, on the left, the thread
axial section in a given axial plane of the worm (a left-hand worm is considered),
and the trace, w, of the plane ðx1 ; y1 Þ passing through the worm wheel axis and the
instantaneous axis of rotation (see Fig. 11.10). The same figure shows, on the right,
what happens in the transverse section of the worm having coordinate z1 ¼ const,
where the point P under consideration is localized. This transverse section is par-
allel to the plane ðx1 ; y1 Þ.
Let us consider the axial plane passing through point P, and therefore containing
the worm axis. This plane, which has trace O1 R in Fig. 11.30, will cut the thread
surface according to a straight side rack, as it is shown in the same figure, on the
left, in which this plane is represented revolved in the relevant view after rotating
about its intersection with the plane ðx1 ; y1 Þ. Given the above-described properties,
in this plane, the projection of the normal PM is nothing more than the normal to
the tooth straight flank. Moreover, the intersection of the normal to the surface with
the instantaneous axis of rotation is projected into the point M, in which the normal
to tooth straight flank at point P cuts the trace of the plane ðx1 ; y1 Þ, coinciding with
the projection of the instantaneous axis of rotation. It follows that, in this plane
revolved in the relevant view, when point P is a point of contact, the intersection
M of the normal to the tooth profile with the projection of the instantaneous axis of
rotation must be away from point P of the amount (Fig. 11.30):
y1 ¼ z1 tan a: ð11:130Þ
It should be noted explicitly that, for all the points for which z1 is a constant, i.e.
for all the points belonging to a given worm transverse plane, y1 is also a constant.
11.10 Surface of Contact: Determination by Graphic-Analytical Methods 531
It is now easy to find the position for which a current point P is a point of
contact. In fact, once z1 , r, and e ¼ p cot a are known, we find the point of inter-
section H of the circle having radius e and center O1 , with the normal drawn from
point O1 to the straight line O1 P. From point P, on the straight line O1 P, we bring a
segment PR ¼ y1 ¼ z1 tan a. Successively, we join point H with point P, and then
we extend the segment HP until it meets at point T the normal drawn from point
R to the straight line O1 P. Let us now rotate rigidly about point O1 the quadrilateral
O1 HTR, until the point T is located at point T 0 , on the projection of the instanta-
neous axis of rotation. For simplicity, in Fig. 11.30, point P has been drawn in the
position in which it is a point of contact, for which T coincides with T 0 , and
P coincides with P0 . It is evident that, since the circumference of center O1 and
radius O1 T cuts in two points, T 0 and T 00 , the projection of the instantaneous axis of
rotation (Fig. 11.30), we have correspondingly two possible positions, P0 and P00 ,
for which P can become a point of contact.
Maintaining z1 ¼ const, and thus y1 ¼ const, and varying r, it is easy to find,
point by point, the projection of the line of contact on the plane z1 ¼ const, i.e. the
intersection of the surface of contact, that is the locus of all points of contact, with
the plane z1 ¼ const. If we consider a sufficient number of these sections, we can
determine the entire surface of contact with the desired accuracy. To make the
calculation procedure more streamlined and convenient, it should consider a
number of planes z1 ¼ const equally spaced of the same fraction of the product
pz cot a (for example, the planes: z1 ¼ 0; z1 ¼
0:1 pz cot a; z1 ¼
0:2 pz cot a;
etc.). Correspondingly, we will have the offset planes defined by coordinates
y1 ¼ const, as well equidistant of a same amount (for the same example, the offset
planes: y1 ¼ 0; y1 ¼
0:1 pz ; y1 ¼
0:2 pz ; etc.). It is noteworthy that z1 -coordi-
nates are to be considered positive in the advancing side of the contact, and negative
in the receding side of the contact; correspondingly, y1 -coordinates are positive
towards the worm wheel, and negative towards the worm.
We do not consider here to further deepen the operating procedure of the
Schiebel’s method in its graphical part and, in this regard, we refer the reader
directly to the already mentioned works of Schiebel or other scholars (for example,
Giovannozzi [17]). However, operating with the previously described procedure,
we can trace, point by point, the curves corresponding to various values of z1 , that is
the sections of the surface of contact with the planes having coordinate z1 ¼ const.
Figures 11.31 and 11.22 summarize the results obtained using the above procedure,
for a two-start left-hand worm. In particular, Fig. 11.31a shows the curves indicated
by 0, ±1, ±2,…, corresponding respectively to the intersections of the surface
of contact with the planes having coordinates z1 ¼ 0, z1 ¼
0:1 pz cot a,
z1 ¼
0:2 pz cot a, etc.
From the curves above, it is easy to get, point by point, the intersection curves of
surface of contact with the offset planes, i.e. with the planes parallel to the worm
wheel mid-plane, having coordinates y1 ¼ const. For example, the plane having the
straight line þ II as a trace (Fig. 11.31a) cuts the curves z1 ¼ const previously
obtained in points marked with circles, the abscissas of which are the corresponding
coordinates z1 . The curve obtained as intersection of the surface of contact with the
532 11 Worm Gears
(a) (b)
(c) (d)
Fig. 11.31 Two-start left-hand worm gear pair, with Archimedean spiral worm: a projections of
the instantaneous lines of contact on a transverse plane of the worm; b projections of the
instantaneous lines of contact on an offset plane; c projection of the surface of contact on a worm
transverse plane; d usable portion of the line þ II
with the worm thread, have a shape much more pointed, as their plane is far from
the worm wheel mid-plane.
Finally, to get a clear idea on which to base the geometric-kinematic sizing of the
worm gear pair, it is useful to analyze the projections of the surface of contact both
on a transverse plane of the worm, and on the pitch plane of the worm.
Figure 11.31c shows the projection, on a transverse plane of the worm, of the
surface of contact of the worm gear pair under consideration here, when the worm
wheel has a type-A outside surface. Figure 11.22 shows the projection of the sur-
face of contact of the same worm gear pair on the pitch plane of the worm; in this
regard, we refer to what we said in the Sect. 11.9.
It is to be noted that, in order that the contact is actually possible at all points of
the surface obtained using the previously described procedure, it is necessary that
this surface is contained within the following two well-defined limits:
• The first limit is constituted by the curve s which, in different offset planes
parallel to the worm wheel mid-plane (Fig. 11.31b), represents the locus of the
extreme points at which the contact can get there, without the interference from
occurring. As for involute cylindrical gears, on the lines of contact above, these
points are those of minimum distance from the trace of the worm wheel axis O2
(Fig. 11.31d). This curve s is localized on the side of the top of the line
a (Figs. 11.22 and 11.24b) of the surface of contact having a horseshoe shape;
therefore, the danger of interference depends on the greater or lesser value of the
worm addendum.
• The second limit is constituted by the curve r, which is the locus of the extreme
points that would be obtained on the lines of contact when the tooth flanks in the
various sections had to touch, that is, when the teeth were pointed. In the various
offset planes parallel to the worm wheel mid-plane, these extreme points are the
intersections of the circles having center O2 and passing through the apex
V (Figs. 11.31b and 2.12) of the worm wheel teeth in that section, with the
corresponding lines of contact.
The determination of the curve r can be performed in an approximate way with
the following procedure, which however provides reliable results from the appli-
cation point of view. In the worm wheel mid-plane, containing the worm axis, the
worm wheel tooth profile is an involute curve corresponding to the worm-rack and
pitch curves p1 and p2 (Fig. 11.10); therefore, it can be drawn with well known
procedures. The apex V (Fig. 11.31b) in which the extensions of the profiles of the
two tooth flanks intersect, can thus be determined. The circle having the center on
the worm wheel axis and passing through point V cuts the lines of contact on the
various offset planes parallel to the worm wheel mid-plane at points that are, with a
good approximation, the limiting points that would occur if in these planes the tooth
profiles of the worm wheel had a pointed shape. Projecting these points on the pitch
plane of the worm, we find the curve r.
The line of contact in different offset planes parallel to the worm wheel
mid-plane have variable slope, namely the higher in the receding side of the worm
534 11 Worm Gears
wheel, and the lower in its advancing side. This fact is unfavorable because, as in
the parallel cylindrical spur gears, a variation of inclination of the force exchanged
between the teeth, at equal tangential load to be transmitted, involves a variation of
magnitude of this force, and therefore less smoothness of operation. The slope
differences between the various lines of contact increase with the increasing of e,
that is, according to Eq. (11.127), they increase with the decreasing of a, and the
increasing of c. Therefore, to reduce these differences for worms with strong lead
angle of the thread, c, it is convenient to adopt the highest values of the pressure
angle, a.
The simultaneous lines of contact are the instantaneous lines of contact that are
obtained intersecting the surface of contact with the worm thread surfaces, in a
given position of the worm. For the determination of these simultaneous lines of
contact, it is therefore necessary to consider a number of angular positions of the
worm. Each of these angular positions of the worm is defined by the screw motion,
and therefore by the position of the coordinate system O1 ðx1 ; y1 ; z1 Þ, rigidly con-
nected to worm, with respect to the coordinate system O0 ðx0 ; y0 ; z0 Þ, rigidly con-
nected to the frame. Of course, it is necessary to repeat the above procedure for each
of the angular positions under consideration. These simultaneous lines of contact
are projected, on the pitch plane of the worm and within the boundary of the surface
of contact having a horseshoe shape, as approximately straight lines, parallel to the
worm threads.
Recalling that, in the parallel cylindrical spur gears, it is most appropriate not to
use the lines of contact in points too close to the limit points of tangency of the path
of contact with the base circle, even now the use of these limit lines is very
doubtful. In the parts of the surface of contact more distant from the instantaneous
axis of rotation, the sliding is maximum, and the effects of errors of manufacturing
and assembly are more consistent. Therefore, to a more extended use of the surface
of contact, a higher manufacturing and assembly precision must correspond.
Before describing the peculiarities of the Ingrisch’s method for determining the
surface of contact of an involute worm gear pair, it is necessary to premise a few
fundamental relationships typical of these worm gears, in addition to those set out
in Sect. 11.3. With symbols already introduced in that section, we have the fol-
lowing relationships:
Considering Eq. (8.52), the same normal pressure angle an can be obtained using
the following equation:
The pressure angle ax in an axial section of the worm is given by the first of
Eqs. (11.17), for which we have:
tan an tan an
tan ax ¼ ¼ : ð11:137Þ
cos c1 sin b1
tan ax tan ax
tan at ¼ ¼ : ð11:138Þ
tan c1 cot b1
Fig. 11.32 Condition under which a current point P on the thread surface is a point of contact, for
an involute worm
Therefore, once r and z1 are known, the corresponding position of the point P,
when it is a point of contact, is the one for which the circle of center O1 and radius
r, and the projection of the instantaneous axis of rotation intercept a segment of
length z1 tan cb1 ¼ z1 cot bb1 on the tangent to the base circle, passing through point
P. As we saw for the Archimedean spiral worm, even now we have, generally, two
positions of contact for point P. Therefore, to get the lines of contact in different
planes z1 ¼ const, it is sufficient to trace a set of straight lines that are tangent to the
base circle at points equidistant from it. Then we bring, starting from the inter-
sections of these lines with the instantaneous axis of rotation, segments ðz1 cot bb1 Þ
in the same direction for each of the curves to be obtained. By joining the points
corresponding to a given value of ðz1 cot bb1 Þ, we find the lines of contact, i.e. the
curves of intersection of the surface of contact with the various planes z1 ¼ const.
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
However, if the distance y1 ¼ z1 tan cb1 is equal to r1 rb1 2 (Fig. 11.33), in
that plane z1 ¼ const, there is only one position of the point of contact P, on the
base circle having radius rb1 . This point coincides with the point in which the
tangent to the base circle, passing through the pitch point C and corresponding to
the direction of the thread surface, touches the same base circle. In this regard, the
direction to be considered, of course depending on the fact that the worm is
right-hand or left-hand, must be understood as the tangents PT and P0 T 0 shown in
Fig. 11.32.
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
If z1 tan cb1 [ r1 rb1 2 , in the aforementioned plane z ¼ const, two distinct
1
points of contact P0 and P00 exist on the base circle of radius rb1 . As Fig. 11.33
11.10 Surface of Contact: Determination by Graphic-Analytical Methods 537
Fig. 11.33 Condition for which only one point of contact on the base circle exists
shows, an arc of finite length separates these two points, whereby the line of contact
is not continuous, but has a gap. In other terms, the line of contact is composed of
two separated and detached portions. As well as for an Archimedean spiral worm,
also the usable part of the surface of contact of an involute worm is projected on the
worm pitch plane according to a closed curve, having a horseshoe shape.
Figure 11.34c shows this projection, for a five-start involute worm. On the same
pitch plane of the worm, the points P0 and P00 shown in Fig. 11.33, in which the two
branches of the lines of contact in different planes z1 ¼ const begin, are projected
according to a curve s0 , with respect to which the projection of the surface of
contact must remain outside. The apex of this curve s0 corresponds to point
P shown in Fig. 11.33, and approximately to the curve indicated by +3 in
Fig. 11.34a.
Using the same identical procedure employed for the Archimedean spiral worm,
we determine the intersections of the surface of contact with the different offset
planes, parallel to the worm wheel mid-plane (Fig. 11.34b), the projections of the
surface of contact and curves r and s on the worm pitch plane, the worm wheel
tooth profiles, etc. In addition, to determine the instantaneous lines of contact, we
proceed as indicated for the Archimedean spiral worm, with the only difference that,
in the case of an involute worm, the sections of the surface of contact with the
various planes z1 ¼ const are involute curves of the base circle, instead of
Archimedean spirals. Figure 11.27b shows the projections of these instantaneous
lines of contact on the worm pitch plane, for a right-hand worm gear pair with an
involute worm.
538 11 Worm Gears
(a) (b)
(c)
Fig. 11.34 Five-start worm gear pair, with involute worm: a projections of the instantaneous lines
of contact on a transverse plane of the worm; b projections of the instantaneous lines of contact on
an offset plane; c projections of the surface of contact, and curves s, s0 and r on the worm pitch
plane
From what we have said previously, we can deduce that the intersection of the
surface of contact with any plane tangent to the pitch cylinder of the worm is a
straight line. Therefore, the surface of contact is a ruled surface. In fact, in a plane
that has as a trace, for example the straight line MT (Fig. 11.32), the points P have
coordinates y1 which are proportional to the abscissas z1 in the direction of the
worm axis.
The surface of contact has, as limiting line, the straight line a-a (Fig. 11.34b) in
the plane tangent to the base cylinder, and parallel to the worm wheel mid-plane
(the plane indicated with þ III in Fig. 11.34a). Therefore, the surface of contact has
a shape much more asymmetric (thus it is worse usable), the more rb1 is large
compared to r1 . It is therefore convenient to reduce the ratio ðrb1 =r1 Þ ¼ cos at ,
compatibly with the need not excessively to increase at , which is related to ax by
the Eq. (11.138).
For a good design of the worm gear pair, it is necessary to keep in mind that the
outside surface bounding the worm threads and worm wheel teeth must be defined
in such a way that the surface of contact does not go beyond the limiting lines s, r
and s0 (Fig. 11.34c). However, at the same time, these outside surfaces must be
drawn to ensure that the largest possible part of the lines of contact is used (see, in
this regard, the following section).
11.11 Outside Surface of the Worm Wheel and Related … 539
In Sect. 11.9 we have seen that the position and shape of the boundary line b of the
projection of the surface of contact on the worm pitch plane (see Figs. 11.22 and
11.23), and consequently the extension of the same surface of contact, depend on
the geometry of the outside surface of the worm wheel. In the same section we have
examined what are, from the point of view of the worm gear design, the effects of
three different shapes of the worm wheel outside surface (type-A, type-B, and type-
C), and we calculated, in that regard, that the best outside surface of the worm
wheel is the one called type-A.
With reference to this shape of the worm wheel outside surface, which is the
best, it is to be noted that, once the radius re of the two side cylindrical parts has
been defined (Fig. 11.26), the projection of the surface of contact on a transverse
plane of the worm (Fig. 11.31c) results limited laterally by points M and
N. Therefore, these side cylindrical parts can be chamfered with conical surfaces,
usually passing through the worm axis, and arranged beyond the corresponding
points M 0 and N 0 , thus saving a useless work during the worm wheel cutting
process. The slight conical chamfer shown in Fig. 11.31a, c responds to this con-
cept. It is to be observed that, on the right side (Fig. 11.31c), the conical chamfer
could start at point N, thus obtaining a worm wheel asymmetrical with respect to its
mid-plane. Usually, we make the worm wheel symmetrical with respect to its
mid-plane.
It should be borne in mind that, if the surface of contact does go beyond the
limiting curve s, that is, if the intersection of the surface of contact with the tip
cylinder of the worm goes beyond the line s, the interference with the worm wheel
teeth occurs. Therefore, since during the cutting of the worm wheel teeth, a cutting
tool (a hob or other suitable tools) replaces the worm, a part of the worm wheel
tooth close to its root is removed, due to undercut. The result is a decrease in the
usable part of the surface of contact.
For involute worms having nominal sizing of their threads, the curve s is very far
from the surface of contact, whereby the risk of interference due to undercut does
not occur. This factual situation is very evident in Fig. 11.34c. Instead, for
Archimedean spiral worms having mean values of the lead angle c1 , the interference
occurs for z2 \36 with a ¼ 15 , and for z2 \20 with a ¼ 20 .
Actually, the cutting interference is also higher, since the tool used to cut the
worm wheel teeth is sized in such a way as to ensure the suitable clearance between
the tip cylinder of the worm and the root surface of the worm wheel. Therefore, the
cutting tool of the worm wheel cuts the surface of action in accordance with the
dashed curve shown in Fig. 11.35. This curve is external with respect to that
corresponding to the actual worm, and is obtained using a procedure similar to that
described in the two previous sections.
540 11 Worm Gears
Fig. 11.35 Portion of the surface of contact lost for the cutting interference
(b)
(a)
Fig. 11.36 Worm gear pairs with profile-shifted toothing: a Archimedean spiral worm, with
positive profile shift; b involute worm, with negative profile shift
radius equal to 0.2 m, the profile shift coefficient can be calculated by the following
relationship:
0:15 0:01
x ¼ pz1 q ; ð11:140Þ
z1 ðz1 =z2 Þ
shift coefficient, x. In this regard, the abacus shown in Fig. 11.37 allows to calculate
this profile shift coefficient, as a function of c1 and z1 , for an axial pressure angle
ax ¼ 15 . Here too, the thread of the worm remains unchanged, while the worm
wheel teeth become stockier, as the example in Fig. 11.36b shows.
Due to the aforementioned negative profile shift, the usable surface of contact
tends to move towards the line s, becoming more centered with respect to the pitch
point C (Fig. 11.34c). Therefore, excessive values of profile shift can cause inter-
ferences related to a high value of the addendum of the worm threads. For this
reason, for type A, type N, type I, and type K worms, the trend today is to choose
values of the profile shift coefficient included in the range ð0:50 x þ 0:50Þ,
and preferably x ¼ 0, or slightly positive. In doing so, we avoid the base circle
going to land in the field of active flank (see Jelaska [25]).
For worm gears with type C worm, it is convenient to choose values of the
profile shift coefficients included in the range ð0 x 1Þ, with mean values
preferably equal to x ¼ 0:50. The lower values of this range of variability are to be
preferred for high loads ðz2
40Þ, while the highest values are more suitable for
high operating speed, which guarantee a high efficiency, but require a high preci-
sion. Other authors (see Jelaska [25]), for the type C worm, indicate a variability
range of profile shift coefficients between ð0:50 x 1:50Þ.
11.12 Load Analysis of Worm Gears, Thrust Characteristics … 543
The load analysis of worm gear pairs can be done in a similar way as in the case of
crossed helical gears. Here this analysis is performed considering the most frequent
case in practical applications, which is that of a worm-drive unit with shaft angle
R ¼ p=2. Similar to what we did for crossed helical gears, we consider only the
power losses due to sliding in the direction tangent to the helix obtained intersecting
the threads with the worm pitch cylinder, while still we neglect the power losses
due to sliding in the direction of the thread depth.
For simplicity and according to what is usually made and accepted for this type
of gear-drive unit, here we determine the forces acting on the worm and worm
wheel shafts under the hypothesis that all the force that worm and worm wheel
transmit to each other passes through the pitch point C (Fig. 11.38). We still
remember that, at this pitch point, located in the worm wheel mid-plane, the worm
pitch plane and worm wheel pitch cylinder are tangent, and thus they are touching.
For the sake of simplifying the calculation, we assume also that the transmission of
this total force takes place only at the pitch point C.
The total force that worm and worm wheel transmit to each other has two
components:
• A component Fn normal to the two mating surfaces at the pitch point C.
• A friction component Fl ¼ lFn directed along the tangent at point C to the pitch
helix, having helix angle, b1 , and lead angle, c1 ¼ ½ðp=2Þ b1 .
We must here necessarily take into account the latter component, because it,
unlike what happens for parallel cylindrical spur and helical gears, generally has a
non-negligible intensity, especially in relation to the effects of the calculation of the
tangential force acting on the worm.
Considering the normal section of the worm thread through the pitch point
C (Fig. 11.38), we see that the force Fn acts perpendicular to the thread profile, and
is inclined of the normal pressure angle, an . This force can be resolved into three
mutually perpendicular components, i.e. the tangential component,
Fnt ¼ Fn cos an sin c1 , the axial component, Fna ¼ Fn cos an cos c1 , and the radial
component Fnr ¼ Fn sin an . The friction force Fl ¼ lFn , where l ¼ tan u is the
coefficient of friction, and u is the angle of friction, can be resolved into two
components: the tangential component Flt ¼ lFn cos c1 , acting in the same direc-
tion as the force Fnt , and the axial component Fla ¼ lFn sin c1 ; acting in the
opposite direction to the axial component Fna .
Therefore, for a worm driving the worm wheel, the three resultant components of
the effective total force acting on the worm at its nominal pitch radius, r1 , are given
respectively by:
It is noteworthy that, since the axes of the worm and worm wheel form a shaft
angle R ¼ p=2, the tangential and axial forces applied to the worm wheel at its
nominal pitch radius, r2 , are respectively equal to the axial and tangential forces
applied to the worm, i.e. Ft2 ¼ Fa1 ; and Fa2 ¼ Ft1 , while the radial force Fr2 is
equal to Fr1 . Of course, their absolute values are the same, but their directions are
opposite. It is also to be noted that, in the case of an involute worm, the normal
pressure angle an is obtained from Eq. (11.137), which also applies in the case of an
Archimedean spiral worm, if we mean that ax is the axial pressure angle of this type
of worm. In the latter case, in fact, the Eq. (11.137) coincides with first of
Eq. (11.16).
In a worm-drive unit with the worm driving the worm wheel, the transmitted
load, Ft2 , is given by:
11.12 Load Analysis of Worm Gears, Thrust Characteristics … 545
where T2 is the output torque available at the worm wheel shaft. The input torque,
T1 , to the worm is given by:
T1 ¼ Ft1 r1 : ð11:145Þ
sin an
Fr2 ¼ Ft2 : ð11:147Þ
cos an cos c1 l sin c1
If we consider that Ft1 ¼ Fa2 , and take into account the Eq. (11.146), the input
torque to the worm, T1 , given by Eq. (11.145), can also be expressed as follows:
cos an tan c1 þ l
T1 ¼ Ft1 r1 ¼ Ft2 r1 ffi Ft2 r1 tanðc1 þ uÞ; ð11:148Þ
cos an l tan c1
pz þ 2plr1
T1 ¼ Ft2 r1 : ð11:149Þ
2pr1 lpz
The same Eq. (11.148) shows that, for a gear-drive unit with the worm driving
worm wheel, the tangential forces Ft1 and Ft2 are related by the equation:
With the worm wheel driving the worm, the above equation becomes:
this is because ðr2 x2 =r1 x1 Þ ¼ tan c1 , as we can infer from Eq. (11.30), written for
the nominal pitch cylinders, and with R ¼ 90 (for further theoretical details, see
Ferrari and Romiti [15]).
The diagram shown in Fig. 11.39 provides the curves that allow to calculate the
efficiency given by Eq. (11.152) as a function of the lead angle, c1 , and coefficient
of friction l ¼ tan u. From the theoretical point of view, from Eq. (11.152) we infer
that, for a given value of the coefficient of friction l, the efficiency increases with
increasing of c1 , up to reach a maximum value for c1 ¼ ½ðp=4Þ ðu=2Þ. The
curves in Fig. 11.39 show that the efficiency depends on the value of the coefficient
of friction, and that, for values of this coefficient not higher than 0.1 that we have in
practical applications, the efficiency practically reaches its maximum value already
for c1 ¼ ð15 20Þ . This factual circumstance is advantageous because, with
threads too inclined, the cutting is difficult, and in any case the related surfaces of
contact have unfavorable asymmetrical shapes and are reduced in size. In practice,
we use lead angles in the range ð15 c1 30 Þ.
The efficiency of a worm gear pair is rather low compared to other types of
gear-drive systems with similar power transmission capacity. The main cause of the
low efficiency is the large amount of frictional losses due to the relative sliding
between the mating surfaces of the worm threads and worm wheel teeth. The
efficiency of the worm gear pairs is a function of many factors, such as the lead
angle, load, speed, coefficient of friction (and thus surface finishing, and type and
condition of lubrication), type of design of the gear-drive unit, and other factors.
Since the efficiency increases with increasing lead angle, a higher efficiency can be
achieved by using multi-start worms with small diameter. Good results can be also
obtained by using rigid worms, with smooth, grounded or polished thread flanks.
If lubrication is appropriate and the gear-drive unit consists of a case-hardened
and ground worm, meshing with an accurately machined worm wheel, then the
efficiency will depend mainly on the coefficient of friction, l, and lead angle, c1 ,
according to Eq. (11.152). This equation expresses a purely theoretical condition,
according to which the efficiency is not dependent on the load or speed. In reality, it
instead depends on these two quantities, since the coefficient of friction, l, is a
coefficient of friction mediated, as between the worm and worm wheel we always
have a good lubrication. In this regard, the results of researches done in this field
confirm that a greater tangential velocity and a smaller load ensure ideal conditions
of lubrication in the best way, so that the efficiency increases.
The speed increase leads, however, a greater heating of the worm gear pair, and a
more rapid wear, so that the tangential velocity should not exceed 3 m=s for worm
gears with steel worm and cast iron worm wheel, and 10 m=s for worm gears with
steel worm and worm wheel made of phosphorus bronze. However, it is necessary
that, as speed increases, the load must be correspondingly reduced. The experience
has also shown that the efficiency and good operating conditions of the gear-drive
unit are highly dependent on the accuracy of manufacturing and assembly. The
operating conditions of this worm gear pair are in fact correct from a kinematic
point of view only for one relative position of the worm and worm wheel axes.
With the worm wheel driving the worm, the efficiency is given by:
tanðc1 uÞ
g0 ¼ : ð11:153Þ
tan c1
1 l tan c1
g¼ ð11:154Þ
1 þ tanlc
1
0
1þ l
tan c1
g ¼ : ð11:155Þ
1 þ l tan c1
2 1lgtan c1 1
g0 ¼ ffi2 ; ð11:156Þ
1 þ l tan c1 g
this is because l tan c1 can be neglected compared to unit. It follows that, if the
efficiency, g, is equal or less than 0.5, it is not possible for the worm wheel to drive
the worm. This means that the gear-drive unit is irreversible or self-locking, i.e. the
548 11 Worm Gears
worm can drive the worm wheel, but the reverse drive is not possible. We use this
property when the design requires the irreversibility condition as a categorical
imperative. However, since the fulfillment of this condition depends on the coef-
ficient of friction, l, which for various reasons is very likely to change during the
service, the irreversibility is not always automatically ensured, even if the above
conditions are satisfied, and the proper coefficient of friction was initially chosen. In
this framework, prudence requires to have always the availability of a brake in such
a gear-drive unit.
Where irreversibility is required (see, for example, certain crane systems), the
following values of the lead angle, c1 , may be taken as guideline values for
roughing calculation: c1 5 and, respectively, c1 6 , when anti-friction bearings
and journal bearings are used. The practical guideline values of the coefficient of
friction, l, are as follows:
• l ¼ 0:10, for worm made of heat-treated steel, worm wheel made of cast-iron,
and operating conditions characterized by smooth finished surfaces of threads
and teeth, and grease lubrication;
• l ¼ ð0:08 0:09Þ, for worm made of heat-treated steel, worm wheel made of
phosphorus bronze, and operating conditions characterized by surfaces of
threads and teeth without smooth finishing, and grease lubrication;
• l ¼ ð0:06 0:07Þ, for worm made of heat-treated steel, worm wheel made of
phosphorus bronze, and operating conditions characterized by machined sur-
faces of threads and teeth, and oil lubrication;
• l ¼ ð0:05 0:06Þ, for worm made of case-hardened and ground steel, worm
wheel made of phosphorus bronze, and operating conditions characterized by
machined surfaces of threads and teeth, and oil lubrication.
Figure 11.39 shows that, for a given value of the coefficient of friction, the
efficiency of a worm gear pair varies markedly with the lead angle. Since the
efficiency increases with increasing lead angle, it follows that large lead angles are
desirable. However, with increasing lead angle, the nominal pitch radius r1 ¼
ðz1 m=2 tan c1 Þ of the worm decreases correspondingly, as Eq. (11.6) shows.
Consequently, even the root diameter of the worm decreases and, with it, its
resistant area, while the worm wheel face width becomes narrower. From the same
relationships it follows that, for a given value of the module, m, it is necessary to
increase the number of threads, z1 , so that the radius r1 is not too small, and the
worm too weak. For further details on calculation of efficiency of worm gear drives,
and experimental investigations, we refer the reader to specialized literature (see,
for example: Magyar and Sauer [36], Stahl et al. [61]).
Since generally, in practice, the radius r1 is assumed equal to ð2 4Þ time the
axial pitch px ¼ pm, the maximum value of c1 , corresponding to the minimum
value of r1 ¼ 2px , should not exceed that we get from the relationship:
11.12 Load Analysis of Worm Gears, Thrust Characteristics … 549
z1
tan c1 ¼ : ð11:157Þ
4p
For this reason, the values of c1 that are adopted are generally the following:
c1 ¼ 9 , for one-start worms; c1 ¼ ð17 18Þ , for two-start worms; c1 ¼ 25 , for
three-start worms. From the above considerations, it is clear that the choice of the
lead angle cannot be made based on efficiency alone, and that the designer must
make a reasonable balance between the factors governing the efficiency and
strength.
In a worm-drive unit, the worm shaft is subjected to deflections due to bending
and shear loads. In addition to strength considerations, the design of this
worm-drive system involves stiffness considerations. From the point of view of the
strength design, we have to determine the dimensions of the gear members to
withstand the given design loads. From the stiffness point of view, it is necessary to
prevent excessive deformations, such as large deflections of the worm shaft that
might interfere with its performance. In fact, it is well known that the maximum
deflection of the worm shaft must not exceed a predetermined limit value, com-
patible with the proper kinematics of the worm gear pair.
Response analysis and design analysis of the worm shaft in terms of strength and
stiffness require the preliminary determination of the stress resultants, i.e. the
bending moments, torques, axial and shear forces. The latter here are neglected,
because their influence on the deflections is known to be small compared to that due
to bending moments. For the determination of these stress resultants, we assume
that the worm is symmetrically positioned relative to the bearings that support its
shaft, and that the total force that worm and worm wheel transmit to each other
passes through the pitch point C, located in the worm transverse plane, which is
equidistant from the bearing mid-planes (Fig. 11.40). We also assume that the input
power enters from the left side, and the thrust bearing is the one on the right in
Fig. 11.40. With this arrangement, the stress resultants acting on the worm shaft are
as follows:
• A bending moment acting in the worm wheel mid-plane, due to the total axial
component, Fa1 , and distributed as shown in Fig. 11.40a.
• A bending moment acting in the same plane as above, due to the radial com-
ponent, Fr1 , and distributed as shown in Fig. 11.40b.
• A bending moment acting in the plane orthogonal to the worm wheel mid-plane,
due to the total axial component, Fa2 , and distributed as shown in Fig. 11.40c.
• A torque T1 ffi Ft2 r1 tanðc1 þ uÞ acting on the shaft part that goes from the
worm mid-plane to the driving machine, due to the total tangential component,
Ft1 , and distributed as shown in Fig. 11.40d (in case the input power entered
from the right side, this torque would be applied in the right half instead of the
left half of the worm shaft).
550 11 Worm Gears
(a)
(b)
(c)
(d)
(e)
Fig. 11.40 Diagrams of the stress resultants acting on the worm shaft
• A compressive force, equal to the total axial component, Fa1 , acting on the shaft
part between the worm mid-plane and the thrust bearing, and distributed as
shown in Fig. 11.40e (in case the thrust bearing was the one on the left, we
would have a tensile force acting on the left half instead of the right half of the
worm shaft).
The considerations described above and the relationships that follow apply in the
theoretical case of carefully manufactured toothing and precise bearing mounting.
Obviously, these theoretical conditions are not practically achievable. Therefore,
under actual service conditions, some amount of deformations, errors and
11.12 Load Analysis of Worm Gears, Thrust Characteristics … 551
In addition to what we said in Sect. 11.3, it should be noted that, generally, the
worm sizing, that is the definition of the addendum and dedendum of the worm, is
made in relation to the module, m, of the generating rack, which coincides with the
worm axial module, i.e. m ¼ mx . The nominal pressure angles of the worm axial
profiles, i.e. the pressure angles related to the pitch cylinder, are chosen according
to the worm wheel number of teeth, z2 , and, with the same number of teeth, they are
assumed greater than those adopted for the parallel cylindrical spur gears.
By doing so, we avoid the interference phenomena, which are manifested to a
greater extent at the beginning of the worm threads meshing with the parts of the
worm wheel teeth that are more distant from the worm wheel mid-plane. Therefore,
we adopt angles a up to 30 for worm wheels having a minimum of 18 teeth
ðz2 ¼ 18Þ, and we can get off to a ¼ 15 for worm wheels with more than 60 teeth
ðz2 ¼ 60Þ.
For the Archimedean spiral worms, the pressure angle of the generating axial
profile does not change when the diameter of the cylinder we choose as nominal
pitch cylinder of the worm changes. Instead, for the involute worms, such pressure
angle changes when the chosen pitch cylinder of the worm changes. In fact, from
Eq. (11.19) we obtain:
rb1 db1
tan c1 ¼ tan cb1 ¼ tan cb1 ; ð11:158Þ
r1 d1
whereby the pressure angle of the generating pitch profile on the pitch cylinder is
given by:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
tan cb1 tan c 2 r1 rb1
tan a1 ¼ 1 2 1 ¼ 1 : ð11:159Þ
tan c1 tan cb1 rb1 r1
The values of the nominal lead angles of worm threads range from c1 ¼ ð5 6Þ
when the worm gear pair must have irreversible operating characteristics, to c1 ¼
ð30 40Þ when we want to get worm gear pairs with high efficiency.
To reduce as much as possible the maximum deflection of the worm shaft, it is
appropriate to have a comparatively large diameter of the resistant cross-sectional
area of the worm shaft, and a distance between the bearings as small as possible in
relation to the other design parameters. In this regard, the form number of the worm,
given by Eq. (11.12), plays a very important role; therefore, it must be chosen with
care. The diameter of the resistant cross-sectional area of the worm shaft, which is
necessary for calculating the section modulus of the worm shaft, must obviously be
less than the root circle diameter of the worm. It is determined on the basis of
strength or stiffness considerations, or on the basis of manufacturing economy (for
example, convenience or not to do the worm integral with the shaft, or to have the
worm press-fitted or shrink-fitted on the shaft).
Once the module m has been chosen, the nominal pitch diameter d1 ¼ 2r1
remains determined. Therefore, from Eq. (11.6) we can deduce that, in order to
obtain large lead angles, and therefore high levels of efficiency, it is necessary to
have a high number of threads, that is, it is necessary that the worm is a multi-start
worm. The number of starts is also a function of the gear ratio u ¼ z2 =z1 . As a
guideline for selecting the number of starts of the worm as a function of u, the
following indications can be used: z1 ¼ 1 for u
30; z1 ¼ 2 for u ¼ ð15 29Þ;
z1 ¼ 3 for u ¼ ð10 14Þ; z1 ¼ 4 for u ¼ ð6 9Þ.
The pressure angles used in worm gear pairs depend on the lead angles. They
must be sufficiently high, to avoid undercut of the worm wheel teeth on the side
where the contact ends. With increasing lead angle, the cutting conditions of the
worm wheel become unfavorable for which, for manufacturing reasons, greater
pressure angles are chosen for larger lead angles. As a guideline for selecting the
pressure angle as a function of the lead angle, the following indications can be used:
a1 ¼ 20 for c1 up to 15°; a1 ¼ 22:5 for c1 over 15° up to 25°; a1 ¼ 25 for c1
over 25° up to 35°; a1 ¼ 30 for c1 over 35°. It is to be noted that, by choosing in
this way the appropriate lead angle, a satisfactory tooth depth is also obtained.
In the case of involute worm, the root circle diameter of the worm, df 1 , must be
greater than or at least equal to the base circle diameter, db1 ¼ 2rb1 . For lead angles
c1 greater than 45° (in this case the worm wheel is the driving member of the worm
gear pair), while maintaining the total depth of the thread unchanged, an increased
addendum is chosen, and this increase is commensurate to the lead angle (for
example, ha1 ¼ 1:4m cos c1 ).
The operating pitch diameter of the worm may be different from its nominal
pitch diameter. This happens in the case of a worm gear pair with profile-shifted
toothing. It is obvious that, in this case, for generating the toothing of the worm
wheel, we will use a hob whose operating pitch cylinder is characterized by the
same profile shift of actual worm.
11.13 Worm and Worm Wheel Sizing: Further Considerations 553
The definition of the worm wheel face width, b2 , must take into account not only
the guideline Eq. (11.29), but also the following relationship:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
b2 ¼ da12 d2 ;
1 ð11:160Þ
according to which b2 depends on the diameters of the tip and reference cylinders of
the worm.
It is however to be noted that, in the case where the worm threads have high
thickness and large lead angle, with the value of the face width determined using
the previously mentioned procedure, the interference may occur between the side
parts of the worm wheel toothing and the worm. In the cutting conditions, this type
of interference leads to the removal of more or less substantial portions of the tooth
flank by undercut. The determination of the maximum usable face width of the
worm wheel therefore represents a critical point, and it must be made with great
care using the known methods of mating profiles (see Pollone [48], Henriot [19]).
We have already pointed out that, from a theoretical point of view, only one
relative position of the two members of a worm gear exists for its proper operating
condition. Any deviation from this proper relative position, due to design and
cutting errors or incorrect assembly, results in malfunction, abnormal tooth contact,
loss of efficiency, and reduction of load bearing capacity. To overcome these
drawbacks, it is good to foresee the use of crowned worm wheels. Indeed, this
crowning not only eliminates the above drawbacks, but also promotes the formation
of an appropriate oil film, which exalts the positive effects of lubrication.
As we have already said, a worm wheel is obtained by using a hob that is the
copy of the worm. In other words, the hob has the same pitch diameter as that of the
worm, and differs only for a greater addendum, which serves to ensure the nec-
essary clearance for the actual worm gear pair. If, to cut the worm wheel, we use a
so-sized hob, we get a worm wheel without crowning. To have crowned worm
wheels, the following four methods can be used:
• Cut the worm wheel directly using a hob whose pitch diameter is slightly larger
than that of the worm. This is a relatively simple but effective method, since it
not only provides a localized contact at the mid-plane, but also a sufficient gap
for the lubricant oil film formation.
• After cutting the worm wheel with a hob placed at the standard center distance,
recut and finish it with the same hob, whose axis is shifted parallel to the worm
wheel axis by
Dz. So the same crowing effect of the previous method is
obtained, but with a higher cost.
• The crowned teeth of the worm wheel can be obtained by a different hob
orientation, compared to its standard position to cut a crownless worm wheel. In
fact, to obtain the crowning effect on the worm wheel teeth, the hob axis is
slightly rotated by a D# angle about the shortest distance straight line, clockwise
or counterclockwise.
554 11 Worm Gears
• Use a worm with a larger pressure angle than the worm wheel. This is the only
one of the four possible methods that involves modifications that do not affect
the worm wheel (like the three previous methods), but the worm. It is also a very
complex method, both theoretically and practically, because it is necessary to
change the pressure angle and the axial pitch of the worm without changing the
pitch line parallel to the axis, in accordance with the relationship
px cos ax ¼ px;i cos ax;i . The equations that allow us to calculate the final values,
px and ax , of the two variables from their initial values, px;i and ax;i , are very
complex. The subject, though very interesting, goes beyond the purpose of this
textbook, so we refer the reader to more specialized treatises.
A cylindrical worm and a worm wheel that is throated, i.e. configured in such a way
that it can wrap partially and envelope the worm, constitute the worm gear pairs
studied in the previous sections. These worm gear pairs are therefore called single-
enveloping worm gears. In some practical applications, double-enveloping worm
gears have been used, which are characterized by the fact that both the worm and
worm wheel are throated. Therefore, the worm has globoidal shape, so that it is
curved longitudinally to fit the curvature of the worm wheel (Fig. 11.41). These
worm gears are also called globoidal worm gear pairs, and, sometimes, hourglass
worms, because of their peculiar shape.
The double-enveloping gear concept is very ancient. It dates back to
Renaissance. The development of the worm gearing principle progressed along
conventional lines from the era of Archimedes, which is universally accredited as
the inventor of worm gears, until the 15th century, when Leonardo da Vinci
evolved the double-enveloping gearing concept (see Loveless [35], Dudas [12],
Chen et al. [6]). In fact, among the hundreds of sketches from its notebooks, not
only drawings of worm gears were found, but also drawings of hourglass worm
gears.
We call globoids those solid bodies delimited externally by a surface generated
by a curved line rotating about an axis. The globoid shown in Fig. 11.42 is
delimited, at its outside periphery, from the surface generated by the rotation of the
circular arc DE about the z-axis. If, during the rotation of the globoid about its axis,
a tool ABC with cutting edge trapezoidal in shape, contained in the axial plane of
the same globoid, rotates about an axis normal to that plane and passing through the
center O of the circular arc DE (a0 is the shortest distance between the globoid and
cutting tool axes, in the cutting conditions), the cutting tool generates the globoidal
worm. All this as long as the angular velocity x0 of the cutting tool is in a constant
relationship with that of the globoid, x1 . The sides of the cutting edge of trape-
zoidal shape can be straight or curved.
11.14 Double-Enveloping Worm Gear Pairs 555
Consider, as the pitch surface, that generated by the circular arc FG which
bisects the depth of the worm thread, on which the angular width of the tool is c
(Fig. 11.42). If, for each revolution of globoid, the tool rotates by an angle 2c, we
have a single-start globoidal worm; instead, if the tool rotates by an angle 2z1 c, we
will have a z1 -start globoidal worm. In the axial section of the globoid, the thread
profile is that of the generating cutting tool. The helix angle on the pitch surface
changes with continuity as we move from the transverse mid-plane to the transverse
end planes of the worm. The circular pitch (Fig. 11.41) is constant, but the axial
pitch is not constant from thread to thread. In addition, the profile of the worm
thread is continuously variable, and so does the lead angle. Globoidal worms are
also made with thread flanks consisting of helicoids having threads with curved
flanks in axial section.
When, for the generation of the globoidal worm, a tool with cutting edge
trapezoidal in shape is used, the generating lines in the process of generation keep
the direction of tangents to the circle with center O and radius r0 . This circle is
obtained by placing the tool with its mid-plane coinciding with the mid-plane of the
worm, extending the two cutting edges, which are symmetrical with respect to the
mid-plane, and finally drawing the circle tangent to the extensions thus obtained
(Fig. 11.42). The generation of the worm wheel is obtained by using a hob identical
to the generated worm.
556 11 Worm Gears
The generating process of globoidal worms has evolved over time, staring from
the one introduced, in 1765, by the famous English clockmaker Henry Hindley (see
Loveless [35]). With this process, which is considered the first process historically
introduced in this specific field, the worm is cut by means of a trapezoidal-edged
cutting blade, while the worm wheel is cut by a worm-type cutter which is identical
11.14 Double-Enveloping Worm Gear Pairs 557
to the Hindley’s worm. Today, several cutting processes are used, with different
gear cutting machines and different cutters.
All these processes, whatever the technology used, present considerable diffi-
culties, both for the worm and for the worm wheel. Depending on the generating
process used, both the worm and worm wheel differ, more or less, from the theo-
retical shape, as the undercut phenomena cannot be completely eliminated.
Figure 11.43 shows a double-enveloping worm gear pair, with a single-start
worm. It highlights the fact that the worm face width cannot extend beyond the
point A for which the tooth flank of the worm wheel is parallel to the shortest
distance between the axes of rotation. In fact, in the case where the worm face width
extends beyond that point, it would be not possible to either assemble the worm
with the worm wheel, or mill the worm wheel with the cutting tool. Compared to a
single-enveloping worm gear pair, the double-enveloping worm gear pairs require
greater care in assembly. In fact, in addition to verifying that the shaft angle and the
shortest distance are exactly those of the generating and manufacturing processes,
it is necessary to ensure that the mid-plane of the globoidal worm contains the axis
of rotation of the worm wheel.
558 11 Worm Gears
In return, the double-enveloping worm gear pairs have quite a few advantages:
• The special shape of the worm and worm wheel increases the number of teeth
that are simultaneously in meshing and improves the conditions of torque
transmission, since the tooth surfaces in contact are more extensive compared to
those of a corresponding single-enveloping worm gear pair.
• The larger contact area enables a double-enveloping worm gear pair not only to
have a greater load carrying capacity, but also a higher efficiency at usual
speeds.
• The conditions of lubrication are substantially better compared to those of a
single-enveloping worm gear pair, due to the special shape of the lines of
contact between the active surfaces of worm and worm wheel.
• Since the total load is divided between more teeth, the pressure of contact is
lower, and thus the life is longer, if the operating conditions, including a good
lubrication, are best ones.
As a downside, it should be noted that more heat is generated at high speeds, and
hence appropriate conditions of lubrication should be ensured to prevent over-
heating and scoring. Furthermore, the double-enveloping worm gear pairs are
comparatively difficult to manufacture, as special cutting tools are required.
Nevertheless, the cases of practical applications in which it is preferable to use these
gear-drive units are constantly growing.
In Sect. 11.2 we described the five types of standardized worms by ISO, which
characterize as many types of worm drives. However, types of worm drives, which
include standard and non-standard worm drives and special worm drives, are much
more numerous. From a practical application point of view, the various types of
worm drives can be divided into power worm drives and precision worm drives.
Power worm drives are mainly used in the heavy industry (e.g., mining, petro-
chemical, metallurgy, mechanical, freight, etc.), so the problems to be solved are
those related to the improvement of their load carrying capacity, which impose the
use of double-enveloping hourglass worm drives. Precision worm drives are mainly
used in fine mechanics (e.g., measuring equipment, scientific devices, fine robotics,
indexing devices, actuators for orientation of antennas and solar panels, etc.), so the
problems to be solved are those related to improving the precision of relative
movements between worm and worm wheel, which impose the use of dual-lead
cylindricalworm drives.
11.15 Standard and Non-standard Worm Drives, and Special Worm Drives 559
Power worm drives, which have been developed to meet stringent requirements
for load carrying capacity, efficiency, and transmission accuracy, include at least the
following eight types of worm drives (see Chen et al. [6]):
(1) Cylindrical worm drives, characterized by cylindrical worms with thread flanks
having straight lines as generatrixes. Therefore, the worms of these worm
drives are type A, type I, and type N worms.
(2) Hindley worm drives, characterized by double-enveloping gear pairs (see
Loveless [35], Dudas [12], Crosher [8]). Since a trapezoidal-edged cutting
blade cuts the Hindley globoidal worm, its thread flanks have straight lines as
generatrixes. Therefore, this type of worm drive represents a generalization of
the concept of type N cylindrical worm. A worm-type hob cutter instead cuts
the worm wheel. This hob cutter, with the appropriate variations already
mentioned at the end of Sect. 11.2, is identical to the worm. The surfaces of the
worm thread flank cannot be grounded, and thus the worm cannot be subjected
to surface hardening. Its surface hardness is usually within the range ð30 40Þ
HRC, so wear in operating conditions is remarkable.
(3) Wildhaber worm drives: an enveloping worm and a cylindrical worm wheel
characterize these drives, proposed by Wildhaber, chief engineer of Gleason, in
1922 (see Dudas [12], Chen et al. [6]). The hourglass or enveloping worm is
generated by a pinion-type cutter that, with the appropriate variations already
highlighted above and at the end of Sect. 11.2, has cutting edges identical to the
normal tooth profiles of the worm wheel. Thus, the thread flank surfaces of the
enveloping worm can be easily grounded, with all the advantages that can be
achieved.
(4) Niemann worm drives: a cylindrical worm, whose threads have concave axial
profiles, and a cylindrical worm wheel, whose teeth have convex axial profiles,
characterize these drives, developed by Niemann in 1935 (see Niemann and
Winter [44], Chen et al. [6]). The worm, which originated the type C worms, is
machined with a convex circular profile disk-type cutter or grinding wheel.
Depending on the kinematic conditions between the worm to be cut and cutter,
even a circular axial profile can be obtained. As usual, the worm wheel is cut
with a worm-type hob cutter that, with the appropriate variations highlighted
above, is identical to the worm. The advantages of the worm drives obtained by
combining these two members are better lubrication conditions between the
mating active surfaces, resulting in greater efficiency, and a substantial reduc-
tion in surface stresses due to the fact that the load is transmitted from a
concave surface to the convex mating surface.
(5) Spiroid worm drives, which can be characterized by circular cone worms (these
are the spiroid worms in the strict sense) or by cylindrical worms (in this case
we are talking about helicon worms). In both cases, the mating worm wheels
have related teeth distributed over their end surfaces, for which they resemble
face-gear wheels with curved teeth. A good percentage of these teeth (from
10% to 15%) is simultaneously in meshing with threads of the spiroid worm.
Therefore, these worm drives allow reducing noise and providing good load
560 11 Worm Gears
carrying capacity. They are widely used in particular areas such as drive sys-
tems of astronomical telescopes, launcher missiles, and naval shipbuilding
platforms as well as machine tool equipment and instrumentation (see Goldfard
[18]).
(6) Toroidal involute worm drives, or shortly TI worm drives, which are charac-
terized by an involute cylindrical worm wheel in meshing with an enveloping
hourglass worm, the latter generated by the first (see Duan et al. [11]), with the
usual modes described above. Theoretical analysis, TSC (Tooth Surface
Contact) analysis as well as prototype experimental tests have shown that these
worm drives have an appropriate load distribution, multi-tooth line contact, and
high reliability. However, the performance of these worm drives is still better
with lower loads, while wear increases and efficiency decreases with increasing
load (see Sun et al. [64]).
(7) Double-enveloping worm drives: an hourglass worm, which can be cut either
using a milling cutter or a grinding wheel with conical surface, and an
enveloping worm wheel, generated by a hob-type cutter, with the usual modes
described above, characterize these drives. The hourglass worm is in fact the
generalization of the type K cylindrical worm, which is generated by a
disk-type milling cutter having straight cone generatrixes or by a biconical
grinding wheel. As in all worm gears, even in these worm drives a line contact
occurs, but the meshing region is wider. Various types of these worm drives are
described in scientific and technical literature, some of which stand out for
improved efficiency, extended life, and higher transmission power compared to
the Hindley worm drives described above, which are similar to those discussed
here only in some aspects. These more remarkable features are essentially due
to double-line contacts on the thread surface, resulting in a wider length of the
contact line (see Sakai et al. [53], Zhang and Tan [70], Zhang et al. [71], Zhao
and Zhang [72]).
(8) Internal meshing worm drives, which are characterized by an internal crown
worm wheel and an enveloping barrel worm, generally generated by the same
internal crown worm wheel, with the usual modes described above. These types
of worm drives add to the benefits of hourglass worm drives (i.e., multi-tooth
line contact, good lubrication conditions, long service life, and high load car-
rying capacity), the benefits of internal meshing, such as compacter structure,
smaller volume, and lighter weight. Various types of these worm drives are
described in scientific and technical literature (see Hoyashita [21], Chen [5],
Chen et al. [4]). Theoretical and numerical analyses and experimental tests have
substantially confirmed the aforementioned advantages. Therefore, the per-
spectives for their practical application in aerospace and other heavy-load areas
with volume and weight restrictions are extremely promising.
As we have already pointed out, each of the eight types of worm drives
described above includes one or more subtypes that fall into the same category. So
the number of possible configurations of worm drives becomes much larger. It is
then further expanded when precision worm drives are considered, which are
11.15 Standard and Non-standard Worm Drives, and Special Worm Drives 561
growing applications in several modern technical areas due to their features, such as
relative motion accuracy, adjustable backlash, and ability to compensate geometric
deviations due to wear. In fact, to achieve these performances, not only the center
distance can be changed (in this case, the worm drive configuration remains
essentially unchanged, with the only variation being able to adjust the center dis-
tance), but also special solutions can be used, which configure new types of worm
drives.
These new types of precision worm drives include (see Chen et al. [6]):
(a) Dual-lead worm drives, which allow an accurate backlash adjustment or wear
compensation by axially moving the worm, which is manufactured with two
leads, so that the thickness from one of the worm threads to another is
increased.
(b) Worm drives with a split worm, composed of a half-shank worm and a
half-hallow worm, and a worm wheel with modified tooth flank surfaces, to fit
the worm split design. With these worm drives, the backlash adjustment or
wear compensation are obtained by fixing the half-shank worm and rotating the
half-hallow worm to ensure that the two sides of the worm are meshing with the
worm wheel.
(c) Non-backlash double-roller enveloping hourglass worm drives, with the worm
wheel composed of two half worm wheels and rollers distributed over the
circumference of each half worm wheel, and the hourglass worm generated by
the rollers. With the same worm drives, the backlash adjustment and wear
compensation are obtained by rotating one of the half worm wheel.
(d) Worm drives composed of a backlash-adjustable planar worm wheel with
variable tooth thickness enveloping a hourglass worm. The backlash adjustment
or wear compensation are obtained by moving the worm wheel along its axis,
thanks to the variable tooth thickness that depends on the fact that both tooth
flanks of the worm wheel are manufactured with different inclination angles.
The precision worm drives, composed of an involute beveloid worm wheel
enveloping a hourglass worm generated by the involute beveloid worm wheel
surface, can be considered a variation of these worm drives, as they share the
same operating principle with them. However, these drives belong to
single-enveloping hourglass worm drives, with multi-tooth line contact and
high load carrying capacity.
Even these new types of precision worm drives are characterized by several
variations, which contribute to further enhancing the already large family of worm
drives (see Yang et al. [69]). In addition, in some cases we need to have worm
drives with high-precision and heavy-load capacity, so the problems to be solved
become more numerous, being the sum of those concerning both the power worm
drives and the precision worm drives. These cases include, for example, industrial
robots, indexing devices for gear cutting machines, speed reducing worm drives for
high-speed hoisting machines, precision artillery systems such as gun-laying radars,
ranging radars, quick-fire guns, etc.
562 11 Worm Gears
Chen et al. [6] attempted to classify worm drives, taking into account the fol-
lowing three factors:
– Shape of generating body.
– Tooth position and meshing area.
– Shape of generating surface.
As for the first factor, it broadens the perspective compared to the assumption that
a worm-type hob cutter generates the worm wheel; this cutter, with the appropriate
variations already mentioned at the end of Sect. 11.2, is identical to the worm. With
this usual assumption, which we have done so far, the generating body and related
generating surfaces of the worm drive are those of the worm, whereby the generated
body and corresponding generated surfaces are those of the worm wheel.
However, both theoretically and practically speaking, to get a worm drive, both the
worm and the worm wheel can be used as the generating body. The spectrum of
possibilities so widens, also because the geometric shape of the worm or worm wheel,
when one of these two members of the worm drive is assumed as the generating body,
can be a cylinder, a cone, a convex toroidal rotor or a concave toroidal rotor.
However, the differences are very marked for the worms, for which we have cylin-
drical worms, spiral bevel worms or spiroid worms, barrel worms, and hourglass
worms; instead, the differences are almost entirely evanescent for the worm wheels,
so the four possible shapes actually form a single group. Therefore, from the point of
view of the shape of generating body, worm drives can be divided into five types. In
fact, we can have worm drives with cylindrical-, spiroid-, barrel- and hourglass-worm
enveloping the worm wheel, and with worm wheel enveloping the worm.
From the point of view of the tooth (or thread) position and meshing area, five
types of worm drives can be also identified. In fact, the tooth (or thread) of the
generating body can be located on the outside circumference, end-face or inside
circumference, while different generated bodies can be obtained by choosing dif-
ferent contact lines, and positioning the tooth of the generated body on the outside
circumference, end-face or inside circumference. Therefore, we may have worm
drives with a normal worm meshing a normal worm wheel, or with an end-face or
internal worm meshing a worm wheel, or with an end-face or internal worm wheel
meshing a worm.
Finally, the shape of generation surface (it can be a surface obtained as a path of
a curve, a surface obtained as an envelope of another surface, or a surface obtained
by a rotating tooth) directly affects the thread profile and meshing performance of
the worm drives. From this point of view, Chen et al. [6] identify fourteen types of
worm drives.
Considering simultaneously the three aforementioned points of view, it turns out
that the possible types of worm drives amount to 350. These numerous sets of
worm drives include not only all the types of worm drives that are currently in use,
developed to meet the most different needs of today’s practical applications, but
also types not known, yet to be developed by researchers and designers working in
this area. This shows that gears are still an ever-living and vital search area, with a
11.15 Standard and Non-standard Worm Drives, and Special Worm Drives 563
Fig. 11.44 Speed jack with double-enveloping and recirculating ball worm drive and recircu-
lating ball screw
patents issued on this very advanced subject do not highlight the significant
problems associated with the working interference or excessive backlash, which
may occur between the balls and corresponding raceways.
The worm wheel and worm reference sections are notoriously the transverse
mid-plane of the worm wheel and the transverse mid-plane of the worm, the latter
containing the worm wheel axis. From the analysis of patent literature, it is apparent
that nobody has studied the coupling and contact conditions between balls and
corresponding raceways on the two members of the recirculating ball worm drive in
sections other than the reference ones. The study of contact, which we performed in
three-dimensional geometric conditions, has pointed out that interference or
backlash are even more pronounced in sections parallel to the reference ones, the
more the sections under consideration are far from the reference ones.
In fact, the cutting of the raceways on the two hourglass members of a recir-
culating ball worm drive, carried out with well-known traditional methods, results
in a circumferential shift of the two centers of curvature of the two semi-raceways,
leading to their displacement along the corresponding circumferential directrix.
This can result in the impossibility of moving the balls, which would be blocked by
ball jamming, or excessive backlash, resulting in loss of contact.
Figure 11.45a, b shows, respectively, what is happening in the reference
transverse section of the worm wheel and in any transverse section of the same
worm wheel, which is parallel to the ones assumed as reference. It is evident that, in
the reference transverse section of the worm wheel, the centers of curvature of the
two semi-raceways coincide, while they are shifted of a certain amount in any
transverse section parallel to the reference one. The calculations show that the shift
amount increases as the transverse section under consideration moves away from
the reference one.
11.15 Standard and Non-standard Worm Drives, and Special Worm Drives 565
Fig. 11.45 a Reference transverse section of the worm wheel, where the centers of curvature of
the two semi-raceways coincide; b any transverse section of the worm wheel parallel to the
reference one, where these two centers of curvature are shifted
According to these studies and calculations, possible worm and worm wheel
cutting technologies have been identified, which remedy the drawbacks of the
traditional methods mentioned above. The peculiarities of these technological
solutions are out of the content of this textbook. They are described in Salvini et al.
[54, 55] in the two patents RM 2004A000138 (2004) and PCT/IB2005/050898
(2005), and in Minotti et al. [40], to which the author refers the reader.
References
4. Chen Y-H, Zhang GH, Chen BK, Lou WJ, Li FJ, Chen Y (2013) A novel enveloping worm
pair via employing the conjugating planar internal gear as counterpart. Mech Mach Theory
67:17–31
5. Chen Y-H (2013) Theoretical and experimental investigations on planar tooth internal gear
enveloping crown worm drive. Chongqing University, Chongqing
6. Chen Y-H, Chen Y, Luo W, Zhang G (2015) Development and classification of worm drives.
In: 14th IFToMM world congress, Taipei, Taiwan, 25–30 Oct 2015
7. Colbourne JR (1987) The geometry of involute gears. Springer, New York, Berlin.
Heidelberg
8. Crosher WP (2002) Design and application of the worm gear. ASME Press, New York
9. Davis JR (ed) (2005) Gear materials, properties, and manufacture. Davis & Associates, ASM
International, Materials Park OH, USA
10. Dennis JE, Schnabel RB (1966) Numerical methods for unconstrained optimization and
nonlinear equations. Society of Industrial and Applied Mathematics, Philadelphia, USA
11. Duan L, Sun Y, Bi Q (2004) Tooth contact analysis of Ti worm gearing considering boundary
condition. In: Yan XT, Jiang CY, Juster NP (eds) Perspectives from Europe and Asia on
engineering design and manufacture. Springer, Dordrecht, pp 713–722
12. Dudas I (2000) The theory and practice of worm gear drives. Penton Press, London
13. Dudley DW (1962) Gear handbook. the design, manufacture, and application of gears.
McGraw-Hill, New York
14. Ernst A (1902) Die Hebezeuge auf der Industrie- und Gewerbeausstellung in Düsseldorf.
Z Ver deutsch Ing S. 1551 u.f
15. Ferrari C, Romiti A (1966) Meccanica Applicata alle Macchine. Unione Tipografica –
Editrice Torinese (UTET), Torino
16. Galassini A (1962) Elementi di Tecnologia Meccanica, Macchine Utensili, Principi
Funzionali e Costruttivi, loro Impiego e Controllo della loro Precisione, reprint 9th edn.
Editore Ulrico Hoepli, Milano
17. Giovannozzi R (1965) Costruzione di Macchine, vol II, 4th edn. Casa Editrice Prof. Riccardo
Pàtron, Bologna
18. Goldfard VI (2006) What we know about spiroid gears. In: Proceedings of the international
conference on mechanical transmissions, Chongqing, China, sept, pp 19–26
19. Henriot G (1979) Traité théorique et pratique des engrenages, vol 1, 6th edn. Bordas, Paris
20. Heyer E, Niemann G (1953) Versuche an Zylinderschneckengetrieben, In: Braunschweig:
Vieweg, VDI 95, pp 141–157
21. Hoyashita S (1996) Barrel worm-shaped tool with conjugate cutting-edge profile generated
from tooth profile of internal gear. Jpn Soc Mech Eng 62(593):284–290
22. Ingrisch (1927) Untersuchung der Eingriffsverhältnisse bei der Evolventenschnecke,
Dissertation an der deutschen technischen Hochschule in Prag
23. ISO 1122-2:1999(E/F) Vocabulary of gear terms - Part 2: Definitions related to worm gear
geometry
24. ISO/TR 10828:1997(E), Worm gears—geometry of worm profiles
25. Jelaska DT (2012) Gears and gear drives. Wiley, New York
26. Kane TR, Levinson DA (2005) Dynamics, theory and applications. McGraw-Hill, New York
27. Kreyszig E (1959) Differential geometry. University of Toronto, Toronto
28. Kulkarni S, Kajale P, Patil DU (2015) Recirculating ball screw. Int J Eng Res Sci Technol 4
(2):252–257
29. Levi-Civita T, Amaldi U (1929) Lezioni di Meccanica Razionale – vol 1: Cinematica –
Principi e Statica, 2nd edn., Reprint 1974. Zanichelli, Bologna
30. Litvin FL (1968) Theory of gearing, 2nd edn. Nauka, Moscow (in Russian)
31. Litvin FL (1989) Theory of gearing. NASA RP-1212, AVS COM 99-C-C035, Washington,
DC
32. Litvin FL, Demenego A, Vecchiato D (2001) Formation of branches of envelope to
parametric families of surfaces and curves. Comput Methods Appl Mech Eng 190:4587–4608
References 567
33. Litvin FL, Fuentes A (2004) Gear geometry and applied theory, 2nd edn. Cambridge
University Press, Cambridge
34. Louis A (1967) Differential geometry. Harper and Row, New York
35. Loveless WG (1984), Cone drive double enveloping worm gearing design & manufacturing.
Gear Technol 12–16 and 45
36. Magyar B, Sauer B (2015) Calculation of the efficiency of worm gear drives. Power Trans.
Eng.
37. Maitra GM (1994) Handbook of gear design, 2nd edn. Tata McGraw-Hill Publishing
Company Ltd., New Delhi
38. Maschmeier G (1930) Untersuchung an Zylinder- und Globoidschneckentrieben. Dissertation
an der Technischen Hochschule su Berlin
39. Merritt HE (1954) Gears, 3th edn. Sir Isaac Pitman & Sons Ltd., London
40. Minotti M, Salvini P, Vivio F, Vullo V (2007) Metodo di taglio di viti e ruote in un riduttore a
viti e ruota con ricircolazione di sfere, In: Atti XXXVI Convegno Nazionale AIAS, Ischia,
04–07 Sept 2007
41. Moré JJ, Garbow BS, Hillstrom KE (1980) User guide for MINIPACK-1. Argonne national
laboratory report ANL-80-74, Argonne, Ill
42. Moré JJ, Sorensen DC, Hillstrom KE, Garbow BS, The MINIPACK (1984) Project. In:
Cowel WJ (ed) Sources and development of mathematical software. Prentice-Hall, NJ, USA,
pp 88–111
43. Niemann G, Weber C (1942) Schneckengetriebe mit flüssiger Reibung. VDI-Forshungsh 412,
Berlin
44. Niemann G, Winter H (1983) Maschinen-Element Band III:Schraubrad-, Kegelrad-,
Schneckn-, Ketten-, Riemen-, Reibradgetriebe, Kupplungen, Bremsen, Freiläufe. Springer,
Berlin, Heidelberg
45. Novozhilov VV (1970) In: Radok JRM (ed) Thin shell theory. Wolters-Noordhoff Publishing,
Groningen (NE)
46. Octrue M, Denis M (1982) Note technique CETIM no. 22, Géométrie des roues et vis
tangentes. Ed. CETIM!, Senlis, France
47. Pohl WM (1933) Graphical determination of worm gear contact. Am Mach Eur Ed. 77:130E
48. Pollone G (1970) Il Veicolo. Libreria Editrice Universitaria Levrotto & Bella, Torino
49. Predki W (1985) Berechnung von Schneckenflankengeometrien verschiedener. Schneckentypen
Antriebstechnik 24(2):54–85
50. Radzevich SP (2010) Gear cutting tools: fundamentals of design and computation. CRC
Press, Taylor & Francis Group, Boca Raton, FL
51. Radzevich SP (2013) Theory of gearing: kinematics, geometry, and synthesis. CRC Press,
Taylor & Francis Group, Boca Raton, FL
52. Radzevich SP (2016) Dudley’s hadbook of practical gear design and manufacture, 3rd edn.
CRC Press, Taylor & Francis Group, Boca Raton, FL
53. Sakai T, Maki P, Uesugi S, Horiuchi A (1978) A study on hourglass worm gearing with
developable tooth surface. J Mech Des 100(3):451–459
54. Salvini P, Serpella D, Vivio F, Vullo V (2004) Metodo di taglio di vite e ruota in un riduttore
a vite e ruota con ricircolazione di sfere e relativi utensili di taglio. Brevetto no.
RM2004A000138, Italy, Università degli Studi di Roma “Tor Vergata”
55. Salvini P, Serpella D, Vivio F, Vullo V (2005) Method for cutting worm and worm wheel in a
worm-gear reduction unit with circulation of bearing balls and related cutting tools,
International Patent no. PCT/IB 2005/050898, Università degli Studi di Roma “Tor Vergata”
56. Schiebel A (1913) Zahnräder, II Teil, Räder mit schrägen Zähnen. Springer, Berlin
57. Schiebel A, Lindner W (1954), Zahnräder, Band 1: Stirn-und Kegelräder mit geraden Zähnen.
Springer, Berlin, Heidelberg
58. Schiebel A, Lindner W (1957) Zahnräder, Band 2: Stirn-und Kegelräder mit schrägen Zähnen
Schraubgetriebe. Springer, Berlin, Heidelberg
568 11 Worm Gears
59. Scotto Lavina G (1990) Riassunto delle Lezioni di Meccanica Applicata alle Macchine:
Cinematica Applicata, Dinamica Applicata - Resistenze Passive - Coppie Inferiori, Coppie
Superiori (Ingranaggi – Flessibili – Freni). Edizioni Scientifiche SIDEREA, Roma
60. Smirnov V (1970) Cours de Mathématiques Supérieures, tome II. Éditions MIR, Moscou
61. Stahl K, Mautner E-M, Sigmund W, Stemplinger J-P (2016) Investigations on the efficiency
of worm gear drives. Gear Solutions
62. Stoker JJ (1969) Differential geometry. Wiley, New York
63. Stribeck R (1898) Versuche mit Schneckengetrieben. Z Ver deutsch Ing S. 1156
64. Sun YH, Lu HW, Yan WY, Li GY (2011) A study on manufacture and experiment of
hardened Ti worm gearing. Chin J Mech Eng 47(9):182–186
65. Townsend DP (1991) Dudley’s gear handbook. McGraw-Hill, New York
66. Ventsel E, Krauthammer T (2001) Thin plates and shells: theory, analysis, and applications.
Marcel Dekker Inc., New York
67. Visual Numerics Inc. (1998) ISML Fortran 90 MP library user’s guide, vol 4.0. Houston,
Texas, USA
68. Wolff W (1923) Über die Erzielung günstiger Eingriffsverhältnisse an Schneckengetrieben.
Dissertation Technischen Hochschule, Aachen
69. Yang Z, Shang J, Luo Z, Wang X, Yu N, (2013) Nonlinear dynamics modeling and analysis
of torsional spring-loaded antibacklash gear with time-varying meshing stiffness and friction.
Adv Mech Eng 1–17
70. Zhang GH, Tan JP (1988) Theory investigation of sphere re-enveloping hourglass worm
drive. J Chongqing Univ 10(1):42–49
71. Zhang GH, Zhang TP, Lou WJ (2007) Selecting and optimizing of the parameters on
quasi-plane double-enveloping hourglass worm drive. Mech Trans 31(2):5–10
72. Zhao YP, Zhang Z (2010) Computer aided analysis on the meshing behavior of a
height-modified dual-torus double-enveloping toroidal worm drive. Comput Aided Des
42:1232–1240
73. Zimmer M, Otto M, Stahl K (2016) Homogeneous geometry calculation of arbitrary tooth
shape: mathematical approach and practical applications, Power Trans Eng 36–45
Chapter 12
Spiral Bevel Gears and Hypoid Gears
Abstract In this chapter, the geometry of the main types of spiral bevel gears is
first defined and considerations about the spiral angle are made. The corresponding
cutting processes are then briefly described, and the generation of active flank
surfaces of the teeth is defined. The main geometrical quantities of these gears are
then determined as well as those concerning the equivalent cylindrical gears
obtained using the Tredgold approximation. The load analysis of spiral bevel gears
is then performed, also defining the thrust characteristic on shafts and bearings.
Subsequently, the concepts concerning these gears are extended to the most general
gearing case represented by the hypoid gears. The fundamentals of these gears are
then provided, based on the kinematics already described for the hyperboloid gears.
Two approaches to theoretical analysis are described, the first of more limited
validity, and the second more general, but both capable to provide reliable results in
terms of geometric and kinematic characteristics of these types of gears. The load
analysis is extended to these types of gears, and some indications on the design
choices inherent to them to improve their efficiency are provided. Finally, the
unified ISO procedure, which allow us to calculate the geometric quantities of spiral
bevel and hypoid gears is summarized.
12.1 Introduction
In Chap. 9 we studied the straight bevel gears, which are a special case of power
transmission through intersecting axes. In this chapter, we focus our attention on the
curved-toothed bevel gears. These gears find broad application in the automotive
and truck industry, helicopter industry and, more generally, in the industry of gear
power transmissions through intersecting and crossed axes (see Buckingham [5],
Merritt [37], Henriot [18], Niemann and Winter [42], Townsend [66], Stadtfeld
[61], Maitra [36], Litvin and Fuentes [32], Radzevich [48], Klingelnberg [21]).
A curved-toothed bevel gear can be obtained by an intuitive procedure similar to
that described in Sect. 8.1, which allows to obtain a cylindrical helical gear wheel,
starting from a cylindrical spur gear wheel, divided into more and more segments,
when an infinite number of segments with infinitesimal small rotations along the
face width is considered. By applying this same procedure to a conventional
straight bevel gear wheel, we can consider a finite or infinite number of component
gear wheels having correspondingly finite or infinitesimal thickness, all obtained by
sectioning the initial straight bevel gear wheel with planes perpendicular to its axis.
Then we can rotate a given angle in the tangential direction one of the gear wheels
so obtained with respect to the adjacent one. In this way, if the number of com-
ponent gear wheels is finite, and the relative rotation angle Du of one with respect
to the next is discrete and constant, we obtain stepped bevel gears (Fig. 12.1a).
Instead, if the number of component gear wheels is infinite, and the relative rotation
of one with respect to the next is infinitesimal and proportional to the coordinate
z along the axis, we obtain curved-toothed bevel gears (Fig. 12.1b). This last figure
shows, from below upwards, the next steps that allow to obtain a curved-toothed
bevel gear starting from a simple bevel gear wheel (see Vol.3, Section 5.2).
In doing so, it is evident that, unlike what happens for straight bevel gears that
have conical conjugate surfaces, the mating surfaces of the curved-toothed bevel
gears are no longer conical surface. These curved-toothed bevel gears can be
divided into two families:
• the gears with intersecting axes, called spiral bevel gears;
• the gears with crossed axes, called hypoid gears.
Like the straight bevel gears, even the spiral bevel gears are characterized by the
fact that the axes of the pitch cones of the two members of the gear pair and the axis
of their common generation crown wheel converge at the same pitch cone apex, O,
which coincides with the crossing point, Oc O (Fig. 12.2a). The common gen-
eration crown wheel meshes properly with each of these two members, considered
separately. Instead, the two members of a hypoid gear pair have crossed axes,
which intersect the pitch plane of the generation crown wheel in two different
points, O1 and O2 , which in this plane are offset by a certain hypoid offset, a, whose
amount is generally not large (Fig. 12.2b). In this case, the crossing point, Oc , is the
apparent point of intersection of the axes of the two members of the hypoid gear
pair, when they are projected on a plane parallel to both (Fig. 12.2b). This figure
shows that any point on the shortest distance, including its end points, can be taken
as crossing point, Oc (see Niemann and Winter [42], ISO 23509:2016 (E) [20]).
It is to be noted that, strictly speaking, the hypoid gears (and so a few other gears
that seam spiral bevel gears) do not fall within the family of the spiral bevel gears.
Nevertheless, we considered, like other authors, to include in this chapter both the
families of the spiral bevel gears and hypoid gears. This is because they have the
common feature of not having straight teeth, but curved teeth, and are produced
using entirely similar cutting processes based on the same technologies (see
Micheletti [38], Galassini [13], Rossi [51], Henriot [17]; Litvin [29], Lin et al. [28],
Suh et al. [65], Stadtfeld [63], Klingelnberg [21]). Substantial differences between
these two gear families are present (see Wildhaber [72], Poritsky and Dudley [45],
12.1 Introduction 571
(a)
(b)
Fig. 12.1 a stepped bevel gear; b intuitive procedure to obtain a curved-toothed bevel gear wheel
starting from a straight bevel gear wheel
572 12 Spiral Bevel Gears and Hypoid Gears
(a) (b)
Fig. 12.2 Generation of curved-toothed bevel gears: a spiral bevel gear; b hypoid gear
Spear et al. [59], Pollone [44], Tsai and Chin [67], Fan [10]). From time to time, we
will specify what these differences.
The main advantage of the curved-toothed bevel gears over the straight bevel
gears consists in the fact that, due to the curved-shaped tooth flank surfaces, more
than one pair of teeth is simultaneously in meshing and participates in the power
transmission at any time. Moreover, similarly to what happens for parallel cylin-
drical helical gears, the contact between the curved teeth of a spiral bevel gear pair
begins at one end of the tooth and gradually and smoothly extends to the other end.
It follows an increase in the total contact ratio, which will result equal to the sum of
the transverse contact ratio and the overlap ratio, the latter due to the overlap arc,
i.e. the arc of the pitch circle between the axial planes containing the ends of one
tooth trace. The result is a large transmittable torque, and a smoother meshing
action, due to minor variations in stiffness of tooth parts involved during meshing. It
follows an appreciable reduction in vibration and noise, especially at high speed
[7, 8, 26, 43, 55, 57, 58, 74].
It is noteworthy that, with the term spiral bevel gears, we mean not only the spiral
bevel gears as such, but also the skew bevel gears, also called helical bevel gears,
and Zerol bevel gears. The latter have curved teeth with zero spiral angle. The spiral
bevel gears with spiral angles smaller than 10 are sometimes called “zerol”. Zerol
bevel gears as such are mounted as the straight bevel gears, and produce the same
thrust loads on shafts and bearings, but they have a smoother operation.
The flank lines or tooth traces have different geometry, depending on the gen-
eration processes that are typical of the conventional gear cutting machines cur-
rently used (see Giovannozzi [14], Niemann and Winter [42]). It is well known that
the shape of the tooth flank of straight and curved-toothed bevel gears is defined
according to the geometric shape of the flank lines of the generation crown gear
wheel, which simulates the cutting motion of the cutter. Indeed, during rolling
without sliding of the pitch plane of this crown gear on the pitch cone of the bevel
12.1 Introduction 573
gear to be generated, such a flank line, which is a planar curve, is wound on the
pitch cone, univocally determining the shape of the tooth flank of the bevel gear that
we want to get.
In practical applications, the flank lines actually used are those that are better
suited to the generation process of the toothing, using conventional gear cutting
machines, i.e. without Computerized Numerical Control, CNC [28]. Among them,
those shown in Fig. 12.3 deserve specific mention. They are as follows:
• Straight lines on the pitch plane of the crown wheel, passing through the apex
(Fig. 12.3a), which generate straight bevel gears as those studied in Chap. 9.
These straight bevel gears may be used, for their intrinsic limitations (the
engagement along tooth face width is not gradual, and therefore they are very
noisy), for tangential speeds not very high, i.e. less than 10 m=s; if they are
machined by grinding, tangential speeds can be a little more than 10 m=s.
Fig. 12.3 Flank lines on the generation crown wheel: a straight lines passing through the apex;
b straight lines not passing through the apex, but tangent to a concentric circle of given radius;
c arcs of circle with centers on a concentric circle; d involute curves of a given concentric circle;
e arcs of an Archimedean spiral; f arcs of circle with mean spiral angles equal to zero; g arcs of
epicycloid; h arcs of sinoid; i two straight lines converging on the middle circle (double helical
bevel gears)
574 12 Spiral Bevel Gears and Hypoid Gears
• Straight lines on the pitch plane of the crown wheel, not passing through the
apex (Fig. 12.3b), and tangent to a circle of given radius, concentric with the
inside and outside circles of the same crown wheel, which generate skew bevel
gears, also known as helical bevel gears or oblique spiral bevel gears. These
helical bevel gears, whose tooth traces are non-cylindrical helices, may be used
for tangential speeds much higher (up to 50 m=s, if machined by grinding). This
is due to the gradualness of the engagement along the tooth face width,
smoother meshing action between the mating tooth pairs as well as a greater
total arc of transmission.
• Arcs of various geometrical curves on the pitch plane of the crown wheel, such
as:
– arcs of circle of a given radius, with centers on a circle concentric with the
inside and outside circles of the crown wheel (Fig. 12.3c);
– arcs of an involute curve of a given circle, concentric with the inside and
outside circles of the crown wheel (Fig. 12.3d);
– arcs of an Archimedean spiral, with origin at the apex (Fig. 12.3e);
– arcs of a logarithmic spiral that has its asymptotic point at the apex;
– arcs of an extended epicycloid described by a point outside a circle rigidly
connected to the cutter, and rolling without sliding on a circle concentric
with the inside and outside circles of the crown wheel (Fig. 12.3g);
– arcs of a sinoid (Fig. 12.3h);
– other planar curves that is not the case here to mention.
These curves generate spiral bevel gears that, despite the greater variability
of the spiral angle along the tooth face width, represent the best of the current
bevel gear technology, being able to run at tangential speed still higher than
those permitted by the skew bevel gears (up to 100 m=s).
• Arcs of a circle with a spiral angle at mid-face width equal to zero (Fig. 12.3f),
with which we obtain the so-called Zerol bevel gears, introduced by Gleason.
These gears, thanks to an overlap ratio not too high, can operate at tangential
speeds higher compared to those permitted by the straight bevel gears (ceteris
paribus, about 60% more).
• Two straight lines converging on the mid-circle of the crown wheel (Fig. 12.3i),
or planar curves of more complex geometry (for example, an extended
hypocycloid), for the generation of double helical bevel gears and, respectively,
double spiral bevel gears, also called bevel gears with arrow teeth.
It should be borne in mind that, during the generation motion, i.e. during the
rolling motion without sliding of the pitch plane of the generation crown wheel on
the pitch cone of the helical or spiral bevel wheel to be generated, the spiral angles
are preserved. In fact, these spiral angles remain unchanged, and the various gen-
eration curves characterizing the flank lines, described above, are wrapped
according to other curves, which can be determined case by case, according to the
basic laws of the mating profiles [14]. On this subject, we will return in the next
section.
12.1 Introduction 575
The spiral angle at any point of the tooth flank of a spiral bevel gear is the angle
between the cone generatrix and the tangent to the tooth trace at that point, mea-
sured on the tangent plane to the reference cone. Usually, the mean spiral angle,
bm , i.e. the spiral angle at mid-face width, is specified.
With reference to the generation method of these gears, however, it is convenient
to consider what happens on the pitch plane of the generation crown wheel. On this
pitch plane, whatever the geometry of the flank line, which is the generation curve,
576 12 Spiral Bevel Gears and Hypoid Gears
the spiral angle, b, is the angle between the tangent to this generation curve at any
of its current points P, and the radius passing through the current point under
consideration (Fig. 12.4). This angle is a function of the point under consideration,
and therefore varies from point to point, passing from the inner spiral angle, bi , at
the inner radius, Ri , to the outer spiral angle, be , at the outer radius, Re .
The spiral angle has a constant value only in the case in which the flank line on
the pitch plane of the generation crown wheel is a logarithmic spiral. To demon-
strate this, we assume, in the pitch plane of this crown wheel, a polar coordinate
system OðR; uÞ, such as that shown in Fig. 12.4. Considering a generic flank line,
the condition b ¼ const leads to write the following relationship:
Rdu 1
tanb ¼ ¼ ¼ const; ð12:1Þ
dR k
where k is a constant. From this equation, separating the variables and integrating,
we get:
Fig. 12.4 Generation of a spiral bevel gear wheel by generation crown wheel
12.2 Considerations on the Spiral Angle 577
ln R ¼ ku þ const: ð12:2Þ
Then, passing from the logarithm function to the function, and measuring the
angles starting from the straight radial line passing through the point having radial
coordinate R ¼ Ri , we obtain the following relationship:
R ¼ Ri eku ; ð12:3Þ
which is the equation of the logarithmic spiral. This spiral has as its asymptotic
point the pole O.
During the rolling without sliding of the pitch plane of the generation crown
wheel on the pitch cone of the spiral bevel gear wheel to be generated, the loga-
rithmic spiral is wound according to a conical helix, which cuts under a constant
angle the generatrixes of the pitch cone. From the point of view of the cutting
method, the condition b ¼ const cannot be realized with accuracy by mechanical
means, i.e. simulating the motion of the cutting tool with a mechanism. Of course,
this limitation does not exist for the current Computerized Numerical Control gear
cutting machines (CNC-gear cutting machines). For further details on the geometry
of planar (but also spherical) logarithmic spiral, on its kinematic properties as well
as on some aspect concerning the technological cutting processes to obtain loga-
rithmic spiral bevel gears, (see Li et al. [27], Figliolini et al. [12], Stachel et al.
[60]).
The difficulty to realize mechanically, with precision, the logarithmic spiral on
the generation crown wheel, led to its approximate simulation, by means of curves
geometrically simple and easily achievable with the conventional cutting methods.
The logarithmic spiral was therefore replaced with an arc of the osculating circle at
the point P of the same spiral located on the pitch plane of the generation crown
wheel, in correspondence of the mid-face width (Fig. 12.4). Gleason Works were
the first to introduce and implement this fundamental concept, developing a
face-milling process that uses a milling cutter, having cutter radius rc0 and center O0
(see next section); theoretically, this center O0 is obtained as the intersection of the
normal to the logarithmic spiral at point P with the normal to the radius OP at point
O (Fig. 12.4).
The logarithmic spiral can be realized, with a good approximation, by the sinoid.
As it is well known, the sinoid is a spiral curve given by the following equation,
written in a polar coordinate system OðR; uÞ:
R ¼ k þ k0 cosðkuÞ; ð12:4Þ
where k; k 0 , and k are positive constants. As shown in Fig. 12.5, the sinoid has a
geometric profile fluctuating within the circular crown having inside and outside
radii given by ðk k 0 Þ. If we assume k ¼ m=n, with m and n integers and prime
numbers between them, the spiral curve is closed for the first time after performing
n oscillations over m angular revolutions. Figure 12.5 refers to the case m ¼ 3, and
n ¼ 10:
578 12 Spiral Bevel Gears and Hypoid Gears
Fig. 12.5 Spiral curves obtained as sinoid arcs on the pitch plane of the crown wheel
Spiral bevel gears with teeth shaped like an arc of sinoid are those obtained with
the Brandenberger method [14]. For each forward and back travel of the cutting
tool in the radial direction, the wheel to be generated rotates to an integer number of
revolutions, m, which is prime number with the number of teeth of the wheel, z ¼ n.
In this way, the sinoid closes only after it has machined all the teeth. So, we have a
simultaneous machining of all the teeth, which avoids the inevitable errors of the
universal indexing heads.
If we have the foresight to choose appropriately the values of k, k0 and k, we can
get sinoids of a shape such that, in the part actually used, the angle b changes very
little, ð2 3Þ , and the tooth size varies approximately in proportion to the dis-
tance R from the center O. The method leads to the generation of gearing char-
acterized by very regular operating conditions. Technologically, the motion
expressed by Eq. (12.4) is achievable by combining a uniform rotational speed of
the wheel with a cutting tool speed in the direction of the generatrixes, variable
according to a sinusoidal function. This is obtained, with a good approximation, by
connecting the cutting tool to the connecting rod small end of a crank mechanism,
with the connecting rod long enough with respect to the crank radius.
Before we tackle the topic regarding the various types of spiral bevel gears,
which are produced today and find practical applications, we return to the gener-
ation methods by a generation crown wheel. With reference to Fig. 12.4, we denote
by X0 and x2 the angular velocities of the generation crown wheel and,
12.2 Considerations on the Spiral Angle 579
X0 ¼ x2 sind2 : ð12:5Þ
The tangential velocity at a current point P on the flank line of the generation
crown wheel, which is at a distance R from the center O, is given by:
To ensure the desired spiral angle of the tooth of the generation crown wheel at
any current point P of the flank line, it is necessary that the cutting tool and
generation crown wheel have suitable motions. In this regard, the cutting tool must
move radially, in the direction of the center O, with an instantaneous linear speed v0
at point P, while the generation crown wheel must rotate with an instantaneous
angular velocity X0 , and thus with a tangential velocity RX0 at the same point P. If
b is the spiral angle at point P, the following relationship must be satisfied:
RX0
v0P ¼ : ð12:7Þ
tanb
Since v0 is also the instantaneous linear speed of the cutting tool in the direction
of the pitch cone apex of the spiral bevel gear wheel to be generated, to obtain the
thread of the latter also the following relationship must be satisfied:
Rx2 sind2
v0 ¼ : ð12:8Þ
tanb
This relationship, taking into account Eq. (12.5), coincides with Eq. (12.7).
Since the gear ratio u ¼ z2 =Z0 is equal to the ratio X0 =x2 , and d0 ¼ 90 , from
Eq. (9.15) we infer:
z2 ¼ Z0 sind2 ; ð12:9Þ
where Z0 is the number of teeth of the fictitious generation crown wheel. The
transverse pitch of the two wheels, measured on the reference circles, i.e. the
outside circles having radii R sind2 for the spiral bevel gear wheel to be generated,
and Re for the generation crown wheel (Fig. 12.4), is given by:
Due to the spiral angle, the arc of transmission, compared to what happens for
straight bevel gears having the same profile, is increased by the face advance, s, i.e.
the arc length corresponding to the angular displacement of the tooth profiles
between the toe and heel. Therefore, the total arc of transmission is the sum of the
580 12 Spiral Bevel Gears and Hypoid Gears
transverse arc of transmission and overlap arc. The overlap angles are equal to
ðs=Re Þ and s=ðRe sind2 Þ, respectively for the spiral bevel gear wheel to be generated,
and for generation crown wheel. Obviously, the total angle of transmission is the
sum of the transverse angle of transmission and overlap angle.
Generally, however, always with reference to the flank line geometry of the
generation crown wheel, the spiral angle changes from point to point. We define as
mean spiral angle, bm , or simply spiral angle in common usage, the spiral angle
referred to the point of intersection between the flank line and the middle circle
having radius Rm . Since the spiral angle varies as a function of the distance R from
the center O, and R varies in the range Ri R Re (Fig. 12.4), b may also vary in
the range bi b be , where bi and be are respectively the inner spiral angle and
outer spiral angle, i.e. the spiral angles at the inside and outside circles of the
generation crown wheel.
Since during the generation motion these spiral angles are preserved, they also
characterize the spiral bevel gear wheel that is generated. However, using as gen-
eration curves on the pitch plane of the generation crown wheel the usual flank
lines, which are characterized by a spiral angle that is variable as a function of the
radius, we obtain spiral bevel gear wheels whose flank lines are conical helices with
variable helix angle. As it is well known, the usual flank lines are the arc of a circle,
the arc of an involute curve, and the arc of an epicycloid, respectively shown in
Fig. 12.3c, d, g.
Still more generally, in relation to the geometry of the flank lines on the pitch
plane of the generation crown wheel, we can get spiral bevel gear wheels whose
flank lines are: conical helices with constant helix angle and variable lead; conical
helices with variable helix angle and constant lead; conical helices with helix angle
and lead both variable.
In a curved-toothed bevel gear (spiral bevel gear and hypoid gear), the hands of
the spirals of the two members are always opposite, that is one right-hand and the
other left-hand, or vice versa. In the common language of gear technology, usually
we indicate a spiral bevel gear pair, or hypoid gear pair, with reference to the spiral
hand of the pinion, which is normally the driving member. Therefore, the gear pair
is a right-hand pair or a left-hand pair depending on whether the pinion is a
right-hand pinion or a left-hand pinion.
Right-hand or left-hand teeth are respectively the teeth whose successive profiles
show clockwise or counterclockwise displacement with increasing distance from an
observer looking along the straight line generatrixes of the reference cone. The
spirals are thus right-hand spirals or left-hand spirals for an observer looking along
the generatrixes of the reference cone, regardless of the fact that he is being
positioned by the side of the cone apex, or from the opposite side of the apex.
Another way of defining the hand of the spiral is that of an observer placed on
the wheel axis, which looks like a tracing point moves along the tooth spiral. The
spiral is a right-hand spiral or a left-hand spiral, if the observer sees the tracing point
move in the clockwise direction or in the counterclockwise direction, moving away
from himself. When an algebraic sign is associated with the spiral angle, this second
convention may be usefully applied to spiral gears in which contact occurs at a
12.2 Considerations on the Spiral Angle 581
point of the common normal. In fact, in these cases, the algebraic sum of the spiral
angles is equal to the shaft angle.
According to ISO 23509:2016 (E) [20], a right-hand (or left-hand) spiral bevel
wheel is one in which the outer half of a tooth is inclined in the clockwise (or
counterclockwise) direction from the axial plane through the mid-point of the tooth,
as viewed by an observer looking at the face of the gear.
It should be noted that the rule that the spiral angles of a spiral bevel gear, at any
point where the mating tooth spirals are touching each other, must be equal and
opposite, must be considered carefully when meshing with the generation crown
wheel is analyzed. In fact, the hand of the generation crown wheels of the two
members of a spiral bevel gear must be equal and opposite, respectively, compared
to the hands of the pinion and wheel. Therefore, if a pair of spiral bevel gear is
described as right-hand pair or left-hand pair, it must be understood, as we said
above, that the description applies to the pinion.
The geometry of spiral bevel and hypoid gears and, more generally, the geometry of
the curved-toothed bevel gears, depends on the cutting process used. It is however
to be noted that the cutting process of these gears constitutes a rather complex
kinematic and technological problem. In this regard, many studies have been car-
ried out, and various solutions have been devised, none of which, however, is to be
considered strictly exact, understood as the potential of ensuring tooth profiles
having the shape of theoretical spherical involutes [19]. The reasons are to be found
in the fact that, for the cutting of spiral bevel and hypoid gears, other approxi-
mations are required, in addition to those already described regarding the cutting
process of straight bevel gears.
Obviously, however, as it happens for straight bevel gears, two wheels cut using
the same generation cutting process and, therefore, the same gear cutting machine
and the same cutting tool, are able to mesh between them correctly. This provided
that the generation pitch cones constitute also the working pitch cones (see Schiebel
and Lindner [52, 53], Giovannozzi [14], Niemann and Winter [42]).
However, the detailed study of the various solutions to the problem of the
generation of spiral bevel and hypoid gears, and various geometries of teeth that can
be achieved comes from the limits of this textbook (see Handschuh and Litvin [15],
Litvin and Fuentes [32]). Here we limit ourselves to do a summary of the main
types of curved-toothed bevel gears used in most practical applications, which are
those described in the following sections.
582 12 Spiral Bevel Gears and Hypoid Gears
Gleason spiral bevel gears have as generation curves on the pitch plane of the
generation crown wheel arcs of circle. These gears are made with spiral bevel gear
cutting machines developed by Gleason, which use circular milling cutters with
inserted blades. These cutters could be used for both right-hand and left-hand spiral
teeth. The flank lines on the pitch plane of the generation crown wheel are shaped
like arcs of circle of a given radius, rc0 , with centers on a circle concentric with the
inside and outside circles of the crown wheel. These two circles constitute the
contours of the fictitious circular crown wheel which defines the face width of the
to-be-generated spiral bevel gear wheel (see Henriot [17, 18], Niemann and Winter
[42], Stadtfeld [63]).
The original Gleason process is referred as face-milling process, as the manu-
facturing is done using the single-indexing process, where each tooth space is
generated separately. The process of generation is interrupted when generation of a
given tooth space is ended. Then the workspace is indexed to the next tooth space,
and the process of generation is repeated, and so on until the cutting process of the
to-be-generated gear wheel is completed. The curved-toothed bevel gears obtained
with this process have a tapered-depth profile along the face width.
As Fig. 12.6 shows, this face-milling process uses a milling cutter having cutter
radius, rc0 , and axis parallel to that of the generation crown wheel. The head-cutter
is mounted on a cradle held at rest, and its axis is placed an offset, qP0 (this offset is
also called crown gear to cutter center), with respect to the apex, O, where the axes
of the generation crown wheel and spiral bevel wheel to be generated converge.
Therefore, the center of curvature at the mean point and cutter center as well as the
radius of curvature of the flank line at the mean point and the cutter radius coincide.
The generation surface of the toothing is a conical surface, whose generatrixes
have as directrix curve the arc of circle AB, and are inclined with respect to the
cutter axis passing through point O0 of an angle equal to the pressure angle, a.
Otherwise placing the head-cutter with respect to the axis of the generation crown
wheel, it is possible to get the desired value and direction of the spiral angle, bm .
The theoretical location of the center O0 , described in the previous section, is
therefore not respected. It follows that the active flank of the tooth is a conical
surface, with apex V on the normal to the pitch plane of the generation crown
wheel, passing through the point O0 . The surface of the other tooth flank, i.e. the
flank that becomes active when the direction of rotation is reversed, is also a conical
surface, whose generatrixes converge at apex, V , which is the symmetrical point of
V with respect to the above-mentioned pitch plane.
This cutting process is also referred as formate-cut process (spread blade process,
fixed setting process, and five-cut process are synonymous). The head-cutter, which
is mounted on the cradle held at rest, is rotated about its axis with the suitable cutting
velocity, and generates the tooth surfaces between one tooth space as the copy of the
surfaces of its blades. During this cutting process, the to-be-generated gear wheel
does not perform any rotation about its own axis and with respect to the cradle.
12.3 Geometry and Cutting Process of the Main Types of Spiral Bevel … 583
Gleason has also developed a cutting process of spiral bevel gears referred as
face-hobbing process. With this process, the to-be-generated gear wheels are
manufactured in a continuous indexing process, and have a parallel-depth profile
along the face width. In this case, the head-cutter installed on the cradle performs a
planetary motion, consisting of a rotation in transfer motion with the cradle about
the cradle axis, and a rotation in relative motion with respect to the cradle about the
cutter axis. The to-be-generated spiral bevel wheel is placed with spiral angle, bm ,
with respect to the pitch plane of the generation crown wheel, which is simulated
and materialized by the head-cutter, and rotates about its axis, dragged in rolling
without sliding by the generation crown wheel. Rotations of the cradle, head-cutter
and spiral bevel wheel to be generated are therefore related in a timed
relationship. The spiral angle, bm , of the spiral bevel wheel to be generated rep-
resents the setting angle of the cutting machine.
The generation curves on the pitch plane of the generation crown wheel are arcs
of extended epicycloid. The milling head-cutter is provided with generation sur-
faces formed by a number of trapezoidal blades with dual cutting edges, profiling
respectively the concave and convex flanks of the teeth, or by a number of alternate
blades with single cutting edge, profiling alternatively the aforementioned tooth
flanks. Each tooth space is thus generated simultaneously, with a continuous
motion. Therefore, at any instant, all the teeth are in the same generation stage.
It is noteworthy that, with the Gleason single-indexing process, which is the one
usually used, since it allows a fast and easy set-up of the cutting machine, a
common generation crown wheel for pinion and wheel to be cut cannot exist. To
convince ourselves of this, we consider any of the two members of the spiral bevel
584 12 Spiral Bevel Gears and Hypoid Gears
gear pair to be manufactured. Cradle and head-cutter that, together, materialize the
generation crown wheel, have parallel axes. Furthermore, the root cone of the spiral
bevel member to be cut is tangent to the plane passing through a point of the cradle
axis (i.e. the axis of the generation crown wheel) and perpendicular to it. In this
same point the axis and generatrixes of root and pitch cones of the to-be-generated
spiral bevel member converge. Therefore, the generation crown wheel of this
member is not a crown wheel in the meaning that we have given so far to this term,
i.e. a bevel wheel with reference cone angle of 90 . Instead,
it is actually a bevel
wheel having reference cone angle equal to 90 #f , where #f is the dedendum
angle. Considering then the other member of the spiral bevel gear pair and repeating
the same reasoning, we can deduce that the generation crown wheels for the two
members (pinion and wheel) are different and have different axes.
On the contrary, with the Gleason continuous-indexing process, it is possible a
cutting set-up whereby the pitch cone of the spiral bevel member to be cut is
tangent to the plane perpendicular to the cradle axis, and passing through a point of
this, where the axis of the spiral bevel member and generatrixes of its pitch and root
cones converge. In this case, we have a generation crown wheel in the strict sense.
However, the axis of the head-cutter is no longer parallel to the cradle axis, but
inclined with respect to this at an angle equal to #f . Considering the two members
of the spiral bevel gear pair, we can deduce that the generation crown wheels for
pinion and wheel to be generated are equal and have the same axis.
With reference to the Gleason single-indexing process, it is to be noted that the
most used set-up of the cutting machine has the drawback of not ensuring the
correct meshing conditions along the whole face width of the spiral bevel gear pair,
but only at the mean point. This follows from the fact that the two fictitious and
improper generation crown wheels related to the two members of the gear pair to be
cut are different
from
each other,
having respectively reference cone angles equal to
90 #f 1 and 90 #f 2 . Only at the mean point, where these two fictitious
reference conical surfaces intersect, and therefore the related pitch circles are tan-
gent, the meshing conditions are correct. In all other points, the generated teeth are
not strictly conjugate. Furthermore, we are faced with a relative error of pressure
angle, which changes sign, passing from one side to the other with respect to the
midpoint, and that leads to the so-called bias-in contact and bias-out contact. It is
possible to remedy, at least partially, this drawback, by means of appropriate profile
modifications (profile crowning, lengthwise crowning, and flank twist). However,
the deepening of this topic is beyond the scope of this textbook, and we refer the
reader to more specialized literature.
Depending on the set-up of the cutting machine, with the Gleason single-indexing
process we can get: Zerol bevel gears, having spiral angle equal to zero ðbm ¼ 0 Þ;
Gleason spiral bevel gears themselves, with spiral angles up to 45 (usually,
bm ¼ 35 ). The toothing geometry is characterized by transverse tooth thickness,
tooth depth, and tooth space width tapered from the heel to toe. However, there are
Gleason spiral bevel gears with constant tooth depth.
12.3 Geometry and Cutting Process of the Main Types of Spiral Bevel … 585
Modul-Kurvex spiral bevel gears have as generation curves on the pitch plane of
the generation crown wheel arcs of circle. Special milling head-cutters with inserted
blades in two halves generate these gears. With these special milling head-cutters it
is possible to generate both the two members (pinion and wheel) of the desired gear.
Depending on the set-up of the cutting machine, we can obtain spiral angles ranging
from bm ¼ 25 to bm ¼ 45 . Modul-Kurvex spiral bevel gears are similar to
Gleason spiral bevel gears, except that the tooth depth generally remains a constant
throughout. With a proper set-up of the cutting machine, it is also possible to obtain
tooth depth tapered from the heel to toe.
Oerlikon-Spiromatic spiral bevel gears have as generation curves on the pitch plane
of the generation crown wheel arcs of epicycloid or hypocycloid. These gears, also
called eloid spiral bevel gears, are generated by milling head-cutters with inserted
blades, which are different for right-hand spiral and left-hand spiral. The Oerlikon
process is referred as a face-hobbing process, since the manufacturing is done using
a continuous indexing process. This is obtained with the cradle axis, working axis
and head-cutter axis that roll together in a timed relationship, as it is shown below.
Gears obtained with this cutting process have a parallel-depth profile along the face
width (see Henriot [17, 18], Niemann and Winter [42]).
As Fig. 12.7 shows, z0 -groups of equal blades, arranged alternatively inwardly
and outwardly of a circle having radius rc0 , form the milling head-cutter with
inserted blades (five groups of blades in the case shown in Fig. 12.7). The fictitious
generation crown wheel, whose mean radius is equal to Rm , rotates uniformly about
its axis. It, however, after the head-cutter has carried out ð1=z0 Þ turns, rotates further
of an angular pitch s0 ¼ ð2p=Z0 Þ, where Z0 is its virtual number of teeth. Of course
(see Sect. 9.8), we have Z0 ¼ ð2Rm =m0 Þ, where m0 is the module.
The ratio between the angular velocity of the generation crown wheel, X0 , and
angular velocity of the head-cutter, x0 , is given by:
X0 z0
¼ : ð12:11Þ
x0 Z0
This type of motion can be reduced to the pure rolling (i.e. rolling without
sliding) of the centrode circle having radius qc and center H, rigidly connected to
the head-cutter, on the centrode circle having radius qb and center O, rigidly
connected to the generation crown wheel. This last centrode circle is then the
epicycloid base circle, and qb is its radius. Therefore, we infer that the generation
curves on the pitch plane of the crown wheel are arcs of extended epicycloid, as
586 12 Spiral Bevel Gears and Hypoid Gears
Fig. 12.7 Operating mode of the Oerlikon spiral bevel gear cutting machine
they are described by the cutting edges of the cutter blades, which are arranged at
the outside of the rolling centrode circle having radius qc . In fact, we have:
qb x0 Z0
¼ ¼ : ð12:12Þ
qc X0 z0
For the direction of rotation shown in Fig. 12.7, the flank line on the pitch plane
of the generation crown wheel is an extended epicycloid or conical cycloid, also
called eloid (see Lawrence [24]). A continuous division process, resulting from the
combination of three motions, realizes it. These motions are as follows: rotation of
the cutting tool, rotation of the spiral bevel wheel to be cut, and motion for the tooth
profile generation. Figure 12.7 shows the combination of motions of the gear wheel
to be cut and the milling head-cutter. This last is composed of a number of identical
groups of blades. Each group has at least one blade, whose inside cutting edge cuts
the convex side of the teeth, and one blade whose outside cutting edge cuts the
concave side of the teeth. Each subsequent group of blades penetrates the following
tooth space between the teeth. Blades for roughing the tooth spaces can characterize
the milling cutter. If one of the directions of rotation shown in Fig. 12.7 is changed,
the flank lines on the pitch plane of the generation crown wheel are extended
hypocycloids.
The rotational speed of the milling head-cutter is uniform. Therefore, since a
group of blades must penetrate into a tooth space of the toothing to be cut, while the
next group of blades must penetrate into the next tooth space, the so-called division
ratio must be respected. It is defined as the quotient of the angular velocity of the
spiral bevel wheel to be cut divided by the angular velocity of the milling
head-cutter, given by Z0 =z, where z is the number of teeth of the gear wheel to be
cut. The angular velocities of the milling head-cutter and spiral bevel gear to be cut
allow to materialize the fictitious generation crown wheel, which rotates about an
axis coinciding with the cradle axis.
12.3 Geometry and Cutting Process of the Main Types of Spiral Bevel … 587
To get the toothing of the spiral bevel wheel by generation motion, it is nec-
essary that the pitch plane of the fictitious generation crown wheel rolls without
sliding on the pitch cone of the spiral bevel wheel to be generated. Without this
motion, the milling cutter would cut the same tooth spaces. Therefore, it is nec-
essary that the cradle on which the milling cutter is mounted, and then the fictitious
generation crown wheel, are equipped with an additional angular velocity. It fol-
lows a corresponding additional angular velocity of the gear wheel to be generated.
It is to be add to that due to the division ratio above. This additional rotational
speed is expressed by the so-called generation ratio, defined as the quotient of the
additional angular velocity of the spiral bevel wheel to be generated divided by the
angular velocity of the cradle, given by z0 =z. This additional rotational speed of the
gear wheel to be cut is obtained by a differential device arranged in the kinematic
chain of the gear cutting machine.
Through the combination of the three motions above-mentioned, the Eq. (12.12)
is satisfied. Apparently, the spiral bevel gears cut with Oerlikon process have the
same appearance of those obtained with the Gleason process. Instead, they are
fundamentally different, at least for the following two aspects:
• The Oerlikon toothing has a constant depth, while usually a taper depth char-
acterizes the Gleason toothing.
• The axes of milling cutters used to cut the pinion and wheel are perpendicular to
the generatrixes of the pitch cones. Therefore, both milling head-cutters mate-
rialize a common generation crown wheel, in the strict meaning of the term, so
that the cutting process complies with the theoretical conditions for conjugate
toothing. Thus, the problems regarding the contact distribution between the
active flanks of the teeth, which are typical of the Gleason process, are
overcome.
Depending on the adjustment of the Oerlikon cutting machine, which may be
performed with reference to a point between the inside and outside circles of the
generation crown wheel, not necessarily coinciding with the mean point, it is
possible to obtain two types of toothing. The first type consists of N-type
Oerlikon-Spiromatic spiral bevel gears, with spiral angle ranging from bm ¼ 30 to
bm ¼ 50 , and maximum normal module at the mean point which decreases
towards the tooth sides. The second type consists of G-type Oerlikon-Spiromatic
spiral bevel gears, with spiral angle ranging from bm ¼ 0 to bm ¼ 50 :
generated in a manner different, but equivalent, to what we have seen in Sect. 2.1.
In fact an involute can be generated by a point P on a tangent line to the base circle
(tangent line and base circle rotate rigidly connected, i.e. without relative rolling),
when the point P moves on the tangent line with a linear velocity equal to the
tangential velocity on the same base circle.
To obtain a given spiral bevel gear with this generation process, we place first
the gear wheel to be generated in such a way that its cutting pitch cone, having a
pitch angle d2 , is tangent to the pitch plane of the generation crown wheel
(Fig. 12.8). Then we give to the conical hob an additional rotation x0 about its axis.
Since the conical hob is a single-start hob, and the generation crown wheel has a
fictitious number of teeth Z0 , this additional rotation causes a rotation X0 ¼ x0 =Z0
of the generation crown wheel considered in meshing with the hob, and thus a
rotation x2 ¼ X0 =sind2 ¼ x0 =ðZ0 sind2 Þ of the spiral bevel wheel to be generated.
By virtue of the three rotations considered above, the conical hob, while drags in
rotation the fictitious generation crown wheel, and with it the spiral bevel wheel to
be cut, moves sweeping all the generation crown wheel, which also rotates, and cuts
progressively the spiral bevel wheel to be generated over all its face width.
In practical applications, the teeth are shaped like shortened involutes, for rea-
sons on which it is no need to dwell here. The spiral angle, bm , is usually equal to
35 , but it may be up to 38 . Teeth of constant depth characterize the geometry of
the toothing. The pitch and tooth thickness are the nominal ones. The surface
finishing is not high, and the tooth flank surfaces are scaly, due to the conical hob.
590 12 Spiral Bevel Gears and Hypoid Gears
Skew bevel gears, also called improperly helical bevel gears, have as generation
curves on the pitch plane of the generation crown wheel straight lines not passing
through the apex, but tangent to a circle of given radius, concentric with the inside
and outside circles of the generation crown wheel (Fig. 12.3b). These gears are
obtained using the Bilgram cutting process or the same Gleason cutting process of
the straight bevel gears, described at the end of Sect. 9.4. Obviously, in this case,
the guides that lead the two cutters are no longer oriented in a radial direction, but
rather in such a way that, on the pitch plane of the generating fictitious crown
wheel, the working strokes of the cutting tools are constantly tangent to a circle of
given radius, as Fig. 12.3b shows. The active flank of the tooth obtained by this
cutting process is a plane surface.
With this cutting process, which is the most important for large curved-toothed
bevel gears (diameters up to 2500 mm, and modules up to 20 mm), we can obtain
skew (or helical) bevel gears characterized by a small value of the spiral angle (bm
equal to a few degrees). Other similar cutting processes are based on flank lines on
the pitch plane of the generation crown wheel that are only virtually straight-line
segments oriented as Fig. 12.3b shows; indeed, they are obtained as arcs of sinoid.
With an appropriate sizing of this sinoid, deviations from the straight line are of the
order of micrometer (thousandth of a millimeter). With this cutting process, the
teeth are also obtained with a continuous and simultaneous generation motion [14].
Archimedean spiral bevel gears have as generation curves on the pitch plane of the
generation crown wheel arcs of Archimedean spirals. These gears are not much
used in practical applications, since the cutting process with which they are gen-
erated is not able to ensure an accuracy grade comparable to the one typical of the
previously described processes.
Branderberger spiral bevel gears have as generation curves on the pitch plane of
the generation crown wheel arcs of sinoid (Fig. 12.5). We have already described
the cutting process of these gears, and its potential, in the previous section, to which
we refer the reader.
12.3 Geometry and Cutting Process of the Main Types of Spiral Bevel … 591
Double-helical bevel gears or double spiral bevel gears have as generation curves
on the pitch plane of the generation crown wheel two straight lines of opposite
hand, converging on the mean circle (Fig. 12.3i). These gears have a spiral tooth in
two portions of opposite hand, but are uncommon. They in fact have been devel-
oped and made with the aim of eliminating the end thrust. However, the results
have been disappointing so far, since the maldistribution of the load on the meshing
teeth constitutes a limit, which is difficult to overcome.
These gears are generated using a variation of the Reinecker-Böttcher gear
cutting machine with which the double-helicoid cylindrical gears are produced. In
reality, the denomination of double-helical bevel gears is somewhat improper, since
the generation curves on the pitch plane of the generation crown wheel are not two
straight lines of opposite hand, converging on the mean circle, but rather a con-
tinuous curve in the shape of an arc of extended hypocycloid. The more correct
name would be spiral bevel gears with arrow-shaped teeth.
In addition to the above-mentioned main families of spiral bevel gears, other
families exist that take their trade names from the cutting processes with which they
are generated. This is not the place to describe their characteristics that, essentially,
are attributable to the generation processes previously highlighted [13].
A general process of generation of the tooth active flank of spiral bevel gears is
described here. It is however be borne in mind that the active surfaces of the tooth
flanks obtained with the process made by means of the existing gear cutting machines
are not perfect spherical involutes, but present with respect to the latter small differ-
ences. Therefore, the properties of the mating surfaces theoretically obtained may be
considered sufficiently representative of those actually realized with the current gear
cutting machines. They are therefore to be considered reliable and can be used for
calculation purposes. Thus, we are faced with the same type of approximation between
the spherical involute teeth and octoid teeth we talked about the straight bevel gears
(see Poritsky and Dudley [45], Romiti [49, 50], Ferrari and Romiti [11] ).
Without loss of generality of the discussion, we here describe the generation
process considering a skew bevel wheel, i.e. a curved-toothed bevel wheel that has,
as generating flank line on the pitch plane of the generation crown wheel, a straight
line not passing through the apex (Fig. 12.3b). With reference to Fig. 9.5, which
shows the generation of spherical involutes and tooth profiles of straight bevel
gears, let’s consider the base cone ðO; c1 Þ of the curved-toothed bevel wheel 1, and
the plane p, tangent to it, as well as any straight-line segment, l1 , not passing
through the apex, O (Fig. 12.9).
592 12 Spiral Bevel Gears and Hypoid Gears
We then roll without sliding the plane, p, on the base cone. Each point, P, of the
straight line, l1 , describes, on the sphere having radius OP and center O, a spherical
involute of the base circle obtained as the intersection of the base cone with the
afore-mentioned sphere. In this rolling motion without sliding, the successive
positions of the straight line, l1 , define a surface, which we assume as the active
surface of the wheel tooth. This surface, r1 , is a ruled surface, whose generation
lines, l1 , are tangent to the base cone under consideration. The locus, k1 , of points of
tangency of straight lines, l1 , with the base cone is a geodesic on the base cone, i.e.
a curve of shortest path between two points on the cone surface. It is a developable
curve, in the sense that, with a reverse operation, that is by rolling without sliding
the base cone on the plane p, the geodesic, k1 , develops in the straight line, l1 .
In each point P of the geodesic, the tangent to it lies on the plane p tangent to
the base cone, which also includes the generation line l1 of surface r1 passing
through point P . Since then the tangent to geodesic and generation line l1 of
surface r1 have the same inclination with respect to the generatrix of the base cone
passing through the same point P , they obviously coincide. The surface r1 thus
proves to be the ruled surface, defined as the locus of tangents to the geodesic; it is
therefore a developable surface.
From this generation process of surface r1 , we can deduce the following theo-
retical properties, which concern the same surface r1 and the engagement between
two mating curved-toothed bevel wheels, whose active tooth surfaces are obtained
by the same generation process.
1. The normal n1 to surface r1 at any current point P is the straight line on the
plane p (the plane tangent at point P to the base cone), and is perpendicular to
the generation line l1 of surface r1 passing through the same point (Fig. 12.9).
In fact, this normal must be perpendicular to the tooth profile at point P. This
profile is, however, the path described by point P during the rolling motion of
plane p about the generatrix of contact of the same plane p with the base cone;
therefore, the tooth profile is perpendicular to the plane p. Consequently, the
12.4 Generation Process of the Tooth Active Flank of Spiral Bevel Gears 593
consideration belongs to the plane of action, which is the common tangent plane
p to the two base cones of the same bevel pair. The meshing surface is the
region of this plane of action having the shape of a circular ring sector between
(Fig. 12.10).
• the two circles Ke and Ki , obtained as intersections of the plane of action
with the two spheres that surround the face width of each tooth, having radii
Re and Ri respectively (Figs. 9.2 and 9.3);
• the two radii that project the points E and A from the point O, where the axes
of the two wheels intersect (points E and A are the intersections of the circle
Ke with the tip circles of the two wheels on the sphere having radius Re ).
At every instant, the active surfaces of the teeth, r1 and r2 , are touching along a
generation line segment belonging to the aforementioned region. This segment
appears to rotate about the normal to the plane of action passing through point
O with the angular velocity X1 ¼ X2 ¼ X, defined above, while wheels 1 and 2
rotate with angular velocities x1 and x2 (Fig. 12.10). The positions on the circle
Ke of points E and A as well those of points of tangency T2 and T1 of circle Ke
with circles c1 and c2 (Fig. 9.5) on the sphere having radius Re can be easily
obtained by considering the spherical triangles ðO1 T1 C Þ, ðO1 T1 E Þ, ðO2 T2 CÞ,
and ðO2 T2 AÞ, where O1 and O2 are the poles of the pitch surfaces. Figure 12.11
shows the first two spherical triangles, which refer to wheel 1. Mutatis mutandis,
the other two spherical triangles are obtained in a similar way. Considering these
spherical triangles, which are right angled at point T1 or T2 , we get respectively
the following relationships:
cosd1 ¼ cosdb1 cosdT1 cos d1 þ m
Re ¼ cosdb1 cosdE
ð12:13Þ
cosd2 ¼ cosdb2 cosdT2 cos d2 þ m
Re ¼ cosdb2 cosdA
where (Fig. 12.10): dT1 is the angle between the straight lines OC and OT1 ; dA is
the angle between the straight lines OA and OT2 ; dT2 is the angle between the
straight lines OC and OT2 ; dE is the angle between the straight lines OE and
OT1 ; m is the arc shown in Fig. 12.11. Equations (12.13) allow to obtain the
angles dT1 , dT2 , dE and dA and the relative positions of the various
corresponding points on the circle Ke .
For each position of the generation line of contact, and thus for each relative
position of the bevel pair, it is easy to determine the mutual force transmitted
between two teeth that touch each other along the same generation line, under
the assumption that only one pair of teeth are in meshing. To this end, we
consider the generation line of contact, l1 , in the position where it is divided into
two equal halves by the segment of straight line OC between the circles Ki and
Ke (Fig. 12.10). We assume also that the force, F, that the wheel 2 exerts on
wheel 1 is applied at point of intersection between the generation line, l1 , and the
straight line, OC (this is the mid-point, usually used by the technical standards),
and that friction is negligible, for which it is directed perpendicularly to the same
generation line, l1 . Finally, we denote by b the angle between the generation
line, l1 , and straight line, OC.
As Fig. 12.10 shows, the force F has a component normal to the straight line
OC, Fn ¼ Fcosb, and a component directed along the line OC, Ft ¼ Fsinb. The
force F as well as its two components Fn and Ft are vectors that lie on the
meshing surface, which is a portion of the pitch plane of the generation crown
wheel. By using the coordinate system Oðx; y; zÞ shown in Fig. 12.11, the
components of this force F are given by the following relationships:
Fx ¼ Fn cosa ¼ Fcosbcosa
Fy ¼ F ðcosbsinacosd1 þ sinbsind1 Þ ð12:14Þ
Fz ¼ Fsinbðsinasind1 cosd1 Þ:
In the case in which the pinion 1 is the driving wheel, and the driving torque T1
is applied to it, for steady state operating conditions the following relationship
must be true:
Re1 þ Ri1
Fx ¼ T1 ; ð12:15Þ
2
2T1
F¼ : ð12:16Þ
cosbcosaðRe1 þ Ri1 Þ
gear cutting machines enjoying greater success today use, as generation lines, the
circular arc (Fig. 12.3c), the arc of epicycloid (Fig. 12.3g), and the involute arc
(Fig. 12.3d).
Figure 12.12 shows a generation crown wheel on whose pitch plane generation
lines are traced, having geometry in the form of an arc of circle. They constitute a
special case of curved generation lines. In this case, as well as in the general case of
generation lines of any geometry, repeating the procedure described above, we can
derive the active surface of the tooth flank of the to-be-generated curved-toothed
wheel, whose properties are quite similar to those obtained above. In particular, at
each instant, the mating gear wheels are touching along the points of the generation
line, l1 , in the common plane p tangent to the two base cones of the same wheels.
This generation line appears rotate in this plane with an angular velocity X pre-
viously defined, while the two wheels rotate about their axes with angular velocities
x1 and x2 . Also the mutual force, F, that the teeth exchange each other, in the
assumption that only one pair of teeth is in meshing, can be obtained in the same
way.
Finally, it should be noted that the meshing between the curved-toothed bevel
gears obtained with the above described generation process remains kinematically
correct with respect to small changes of the shaft angle. Instead, it is no longer
kinematically correct when displacements are such that the axes of the two mem-
bers of the gear pair are no longer converging, so that the gear becomes a hypoid
gear.
Various types of sizing are used for spiral bevel gears: some of them provide teeth
of constant depth, while others require geometric parameters related to those we
have described in Chap. 9 about straight bevel gears. Frequently, spiral bevel gears
with profile-shifted toothing are used, most often characterized by profile shift
coefficients equal and opposite for the two members of the gear pair ðx2 ¼ x1 Þ.
These coefficients are related to the thickness factors given by Eq. (9.111). This
discussion on the sizing of the curved-toothed bevel gears as well as that of the
hypoid gears will still be deepened in Sect. 12.11.
Nominal profiles of the teeth are those related to the back cone. The tooth
profiles obtained as sections of the same teeth with countless cones between the
back and inner cones (from a theoretical point of view, they are infinite), are similar
to each other, with dimensions linearly proportional to the distances from the apices
of the pitch cones.
The Tredgold method (see Buchanan [4]), already described for straight bevel
gears in Sect. 9.5, can be used also for spiral bevel gears. Depending on the
geometric shape of the tooth trace (i.e., the line of intersection of the tooth flank
with the reference pitch cone), we will have, generally, virtual equivalent cylin-
drical wheels with straight, skew (or improperly helical) and spiral teeth, which are
used for the analysis of contact. In the case of spiral bevel gears of interest here,
often we are content to reduce the actual gear wheel to a cylindrical helical wheel,
equivalent to the spiral bevel wheel under consideration, having virtual number of
teeth zv ¼ z=cosd, and helix angle equal to the spiral angle, bm . Obviously, if the
teeth of the actual spiral bevel gear pair are crowned, also those of the virtual
equivalent cylindrical helical gear pair will be crowned.
If the teeth are cut with a cutter having nominal standard sizing in the normal
section, we can assume that the minimum number of teeth, zmin , that can be cut
without the risk of undercut, decreases according to a factor equal to cosdcos3 bm
[14]. This extends the considerations already made for cylindrical involute helical
gears (see Sect. 8.5) to spiral bevel gears.
At a given radius R in the range Ri R Re (Fig. 12.4), the normal pitch, pn , i.e.
the pitch measured in a section made with a plane perpendicular to the flank line of
the tooth on the pitch plane of the generation crown wheel, is equal to the circular
pitch, or transverse pitch, p, at that radius, multiplied by cosb. The angle b is the
spiral angle at the same radius.
The various types of toothing of spiral bevel gears, generated by the cutting
processes developed so far, are different, both because they have a spiral angle that
varies with the distance from the apex of the reference pitch cone, both for the way
in which the tooth depth and tooth transverse thickness change as a function of the
radius, R. Once again it is to remember that only the logarithmic spiral used as
generation line on the pitch plane of the generation crown wheel determines a spiral
angle that is a constant when the radius changes.
12.5 Spiral Bevel Gears: Main Quantities and Equivalent Cylindrical Gears 599
About the operating conditions, it is to point out that, by usual convention, the
concave pinion flank in meshing with the convex wheel flank is generally the
drive-side, i.e. the driving flank. Otherwise we would have a greater spiral angle on
the inner cone, bi (it corresponds to the inside radius Ri on the pitch plane of the
generation crown wheel in Fig. 12.4) and, consequently, a pointed tooth. This
applies in the recommended case in which the directions of the spiral and rotation
are the same. Nevertheless, there are cases where the pinion-driving flank is the
coast-side. In these cases, the convex pinion flank is in meshing with the concave
wheel flank.
In addition, it is to be noted that the use of the drive-side as main load trans-
mission direction is for spiral bevel gears (but also and especially for hypoid gears)
a rather binding rule. Transmission of speed and torque and the additional
lengthwise sliding forces would lead on the coast-side to a deflection of pinion
towards the gear wheel, consequently leading to a reduction of the backlash, in
extreme cases to zero backlash. This occurs already under moderate torque, com-
promising the lubrication conditions. Surface damages followed by tooth fracture
can occur.
The curvature of the flank line on the pitch plane of the generation crown wheel is
that resulting from the radius of the inserted blade milling head-cutter used for cutting
as well as the type of gear cutting machine employed. Gear cutting processes can be
those already described in Sect. 12.3 (Gleason, Modul-Kurvex, Oerlikon-Spiromatic,
Klingelnberg-Ziclo-Palloid, Klingelnberg-Palloid, Branderberger, etc.) or others.
Instead, the spiral angle, bm , depends on the set-up and adjustment of the gear cutting
machine. Prudence demands that the generation of these gears is carried out following
the instructions of the manufacturers of gear cutting machines used for this purpose. As
a broad indication, it is appropriate that the spiral angle is sufficiently high so that the
overlap ratio eb is greater than 1.5. This in order to put a remedy to the disadvantages
resulting from a surface of contact that is limited in depth and length (see Niemann and
Winter [42], Handschuh and Litvin [15], Su et al. [64]).
The spiral angle also significantly affects the axial thrust on the shafts, which
must be supported by suitable bearings. In any case, the spiral angle and direction
of rotation should be selected in such a way that the axial thrust on the shafts of the
pinion and wheel has direction such as to move both members of the spiral bevel
gear pair out of mesh when this gear pair is operating in the predominant working
conditions. Otherwise, the pinion would tend to be sucked from the wheel, resulting
in the above-mentioned significant reduction or total loss of backlash, and corre-
lated risk of scuffing. To avoid this, it is appropriate that both members of a spiral
bevel gear pair (or hypoid gear pair) are held against axial movement in both
directions. Often the hand of spiral is imposed by the mounting conditions.
The outer transverse module, met , i.e. the transverse module on the back cone,
and the mean normal module, mmn , i.e. the normal module on the mean cone, are
given respectively by the relationships:
600 12 Spiral Bevel Gears and Hypoid Gears
d1 d2 men
m ¼ met ¼ ¼ ¼ ð12:17Þ
z1 z2 cosb2
where remaining unchanged the meaning of other symbols, men and mmt are
respectively the normal module on the back cone, and the transverse module on the
mean cone (it is to be noted that, to avoid misunderstandings, we used the
double-subscript et to indicate the nominal module, m).
In addition to those obtained in Sect. 9.3, we have to consider other quantities
that are typical of the spiral bevel gears, the main of which are described below. The
mean pitch diameters of the pinion and wheel are given by:
mmn z1
dm1 ¼ d1 bsind1 ¼ ð12:19Þ
cosbm
tanan
tanat ¼ ; ð12:21Þ
cosb
where at and an are respectively the transverse and normal pressure angles. To have
teeth not too tapered, it is necessary that the following condition is met:
ðb=Re Þ ð0:25 0:35Þ.
In the case of teeth tapered from the heel towards the toe (Fig. 9.11a), but
assuming that the nominal clearance, c, is a constant (in this case the apex of the tip
cone will be located at the inside of the reference pitch cone), the reference quantity
is met , whereby we have the following relationships:
da1 ¼ d1 þ #f 2 df 1 ¼ d1 # f 1 ð12:22Þ
da2 ¼ d2 þ #f 1 df 2 ¼ d2 # f 2 ð12:23Þ
hfe1 hfe2
tan#f 1 ¼ re tan#f 2 ¼ re
ð12:24Þ
btan#a1
hae1 ¼ met ð1 þ xhe Þ ¼ ham1 þ ; hfe1 ¼ met ð1 c xhe Þ ð12:25Þ
2
btan#a2
hae2 ¼ met ð1 xhe Þ ¼ ham2 þ ; hfe2 ¼ met ð1 c þ xhe Þ ð12:26Þ
2
12.5 Spiral Bevel Gears: Main Quantities and Equivalent Cylindrical Gears 601
We must keep in mind that, with tapered teeth from heel to toe, it is possible to
have a greater contact ratio, without the danger of meshing interference is mani-
fested at the tooth toe. The nominal clearance, c, is generally assumed equal to
ð0:1 0:3Þm.
In the case of teeth having constant depth (Fig. 9.11b), the reference quantity is
mmn , whereby we have the following relationships:
da1 ¼ df 1 ¼ d1 ð12:27Þ
da2 ¼ df 2 ¼ d2 ð12:28Þ
#a ¼ #f ¼ 0 ð12:29Þ
tanan
tanavt ¼ tanamt ¼ ðwith avn ¼ an Þ: ð12:32Þ
cosbm
We will have then: mvt ¼ mmt ¼ dm1 =z1 ¼ dv1 =zv1 ¼ dv2 =zv2 , and mvn ¼ mmn
¼ mvmt cosbm , where mvt and mvn are respectively the transverse and normal
modules of the virtual gear pair. The profile shift coefficients are given by:
ham1;2 ham2;1
xhm1;2 ¼ : ð12:35Þ
2mmn
602 12 Spiral Bevel Gears and Hypoid Gears
1 h 2 1=2 2 1=2 i
gva ¼ dva1 dvb1
2
þ dva2 dvb2
2
av sinavt : ð12:36Þ
2
gva cosbm
eva ¼ : ð12:37Þ
pmmn cosavt
bsinbm beH
evb ¼ : ð12:39Þ
pmmn b
With reference to the overlap contact ratio and total contact ratio, we must keep
in mind that, when calculating the load carrying capacity of the gear pair, we
usually assume an effective face width, beH , equal to 0:85b, i.e. beH ffi 0:85b. Of
course, the transverse module and pitch of bevel gears change along the face width.
As already we mentioned, depending on the problem to be talked, we use or the
outer transverse module, met , or the mean normal module, mmn . Although it is not
necessary, as a series of cutters cover a large range of modules, these are generally
chosen with reference to the standard set of modules for parallel cylindrical gears.
12.6 Load Analysis for Spiral Bevel Gears and Thrust Characteristics … 603
Load analysis for spiral bevel gears can be made with the same procedure used for
parallel cylindrical spur and helical gears and for straight bevel gears, i.e. neglecting
the contribution of the friction forces. These forces are in fact not very significant
from the design point of view. To simplify this load analysis and evaluate the thrust
characteristics on shafts and bearings, we assume that the total force, F ¼ Fn ,
exchanged between the meshing teeth, acts along the direction of the normal to the
tooth surface. We also assume that this force is applied at the mean point, M, of the
tooth, and thus in the middle face width, in the tooth position where that point is
located in the plane containing the axes of the two members of the gear (see
Giovannozzi [14]; Brown [3]; Su et al. [64]).
With reference to Fig. 12.13, we consider the normal section of the tooth,
passing through its mean point, M, and suppose that this point lies in the plane of
the figure (Fig. 12.13a). The total tooth force, F ¼ Fn , acting on the tooth, lies in
the plane of the normal section to the tooth, and is inclined, with respect to the
perpendicular to the plane of Fig. 12.13a through point M, of the pressure angle, an .
Therefore, the force Fn can be resolved into two orthogonal components, Fn sinan
and Fn cosan (Fig. 12.13c), the first of which acts in the axial plane of the gear
wheel under consideration in the direction perpendicular to the generatrix of the
pitch cone (Fig. 12.13a), while the second in the front view of the same gear wheel
shown in Fig. 12.13b is inclined of the angle bm with respect to the perpendicular
through point M to the projection of the straight line OM. From the same figure,
which also shows the axial view of the gear wheel (Fig. 12.13b), as well as the
various components of the total tooth force, F ¼ Fn , we can infer the following
relationships:
where Ft , Fa , and Fr are respectively the tangential, axial, and radial components
(or tangential, axial, and radial forces).
If we obtain Fn from Eq. (12.43), and we replace it in Eqs. (12.44) and (12.45),
we get the following relationships, which express, with the signs on the top, the
axial and radial components, Fa and Fr , as a function of the tangential force, Ft :
tanan sind
Fa ¼ Ft tanbm cosd ð12:46Þ
cosbm
604 12 Spiral Bevel Gears and Hypoid Gears
(b)
(a)
tanan cosd
Fr ¼ Ft
tanbm sind : ð12:47Þ
cosbm
– If the gear wheel is driving like above, in Eqs. (12.46) and (12.47) the signs
below are valid in the other two cases shown in Fig. 12.14, i.e. the cases (c) and
(d).
If the gear wheel is driven, the signs opposite to those aforementioned are valid.
We must also keep in mind that the direction of the axial thrust in a spiral bevel
gear (and in a hypoid gear) depends on the direction of rotation, hand of the spiral
angle, relative position of the driving and driven members, and pitch cone angle.
Furthermore, it depends on the fact that the gear wheel is the driving or driven
member. All these factors determine the condition for which the axial component,
Fa , will make the gear wheel move away or towards the cone apex.
The following fundamental difference exists between a straight bevel gear and a
spiral bevel gear or a hypoid gear. In a straight bevel gear as well as in a Zerol bevel
gear, the axial components Fa1 and Fa2 always tend to force the pinion and the
wheel out meshing. On the contrary, in a spiral bevel gear as well as in a hypoid
gear, the direction to which the axial components act may be either towards or away
from the cone apex, depending on the algebraic sign appearing in Eq. (12.46).
Since in the case of spiral bevel gears and hypoid gears the axial thrust may act
in both directions, because of the aforementioned factors, it follows that the choice
of the direction and magnitude of the spiral angle constitutes a fundamental point on
which the designer has to focus his attention. The members of the spiral bevel and
hypoid gear pairs can move away or move towards each other. In the first case, an
unnecessary large backlash is created between them. In the second case, a tight
mesh is to be determined, and this may result in jamming or scuffing.
The proper running and operating conditions of a spiral bevel or hypoid gear need a
right kind of support, which can prevent possible axial displacements of both members
of the gear. To this end, meticulous design calculations of the axial thrusts are nec-
essary, in order to choose the appropriate types of support and the constructive and
(c) (d)
606 12 Spiral Bevel Gears and Hypoid Gears
assembly solutions related to them. Figure 12.15 shows the direction of the axial and
radial forces in a spiral bevel gear pair with shaft angle R ¼ 90 ; it highlights the
influence of the hand of spiral and direction of rotation on the direction of the axial
thrust.
It is to be noted that, in the curved-toothed bevel gears, the hand of the spiral and
the direction of rotation of the two members are opposite. Furthermore, in the case
where the shaft angle is equal to 90 ðR ¼ P=2Þ, the axial thrust acting on one of
the two members is equal, in absolute value and sign, to the radial force acting on
the other member, and vice versa. It is also to be kept in mind that the axial thrust,
Fa , is a vector parallel to the axis of the gear wheel, which is applied at a distance
from the axis equal to the mean pitch radius, rm . Therefore, this axial force, Fa ,
causes a concentrated bending moment, Mb ¼ Fa rm , which in turn determines
related reaction forces on the bearings.
Figure 12.15 also shows the radial force acting on the two members of the gear:
it has positive sign when it is directed towards the wheel axis, while it has negative
sign when it is directed away from the wheel axis, i.e. towards the mating gear
wheel.
Therefore, we can deduce that, in the case of a spiral bevel gear characterized by
a shaft angle R ¼ 90 (this is the most frequent case), if Fa1 and Fr1 are the axial
and radial forces acting on the pinion, respectively calculated using Eqs. (12.46)
and (12.47), the axial and radial forces Fa2 ¼ Fr1 and Fr2 ¼ Fa1 acting on the
mating wheel are those obtained using respectively Eqs. (12.47) and (12.46).
(a)
(c)
(b)
Fig. 12.15 Direction of the axial and radial forces in a spiral bevel gear pair with shaft angle
R ¼ 90
12.7 Hypoid Gears: Basic Concepts 607
x2 z1 1
sinb1 ¼ ¼ ¼ ð12:48Þ
x1 z2 i
while from Eqs. (10.5) and (10.6), since tanb2 ¼ tan90 ¼ 1, we get:
Therefore, in this case, the Mozzi’s axis, m (i.e. the axis of the instantaneous
relative helical motion, see Sect. 10.1), lies in the plane perpendicular to the shortest
distance O1 O2 containing the first axis 1, intersects the axis 1 at point O1 , and is
normal to the projection of the second axis 2 on this plane. Thus, m0 is normal to 20 as
Fig. 10.5 shows (see Mozzi del Garbo [39], Scotto Lavina [54]). Consequently, the
pitch hyperboloid of the first wheel, having as axis the axis 1, degenerates into a cone
with apex O1 and cone angle b1 ¼ ðR p=2Þ. Instead, the pitch hyperboloid of the
608 12 Spiral Bevel Gears and Hypoid Gears
second wheel, having as axis the axis 2, degenerates into a ruled plane according to
the tangent lines to the circle having axis 2 and radius equal to the center distance,
a (these tangent lines coincide with the successive positions of the Mozzi’s axis).
This second wheel is the so-called hyperboloid crown wheel. Figure 10.5 also shows
the successive positions of the
Mozzi’s axis for a transmission of motion between
two crossed axes with R ¼ b1 þ b2 [ 90 and b2 ¼ 90 , and hyperboloid crown
wheel.
We can achieve the same result starting from Fig. 10.4b, and considering what
happens when the distance of the axis of instantaneous screw motion, m (remember
that it generates the pair of double pseudo-bevel hyperboloids of one sheet), from
the axis of the pinion 1 is cancelled (such distance, in the general case, is equal to
the radius of the throat). In this case, the hyperboloid 1 is transformed into a cone.
In addition, if such instantaneous axis forms a right angle with the axis of the wheel
2, the hyperboloid 2 is transformed into a planar hyperboloid, which can be taken as
a hyperboloid crown wheel.
The transmission ratio i ¼ x1 =x2 is equal to 1=sinb1 , while the sliding
velocity at the points of contact line m (the Mozzi’s axis), according to Eq. (10.8),
will be equal to vr ¼ x2 a. Equipping of appropriate sets of teeth the pitch surface
thus defined, we can implement the transmission of motion between crossed axes
with a gear pair whose members are, respectively, the first a straight bevel wheel
and, the second, a crown wheel with curved teeth. It is however to be noted that
such a gear pair cannot be used to transmit the motion between orthogonal crossed
axes, since the shaft angle between these axes cannot be equal to p=2 (in this case
should be b2 ¼ p=2 and b1 ¼ 0).
However, since the hyperboloid crown wheel defined above actually coincides
with the crown wheel, the hypoid gear pairs can be generated using as generation
wheel the hyperboloid crown wheel. According to this generation method, the
pinion is generated by placing it in such a way that its axis is positioned with a
crossed axis with respect to that of the generation hyperboloid crown wheel, while
the wheel is generated by positioning it in such a way that its axis is coplanar with
that of the generation hyperboloid crown wheel.
The reference pitch surfaces of the two members of the hypoid gear pair thus
obtained are two conical surfaces, each of which has a common straight line with
the generation hyperboloid crown wheel. The two straight lines do not coincide
when the two-hypoid gear wheels are in contact meshing. In fact, the pitch cone of
the hypoid pinion touches the pitch plane of the generation hyperboloid crown
wheel along the straight line O1 A (see Fig. 12.16, where b2 ¼ b2 ¼ p=2 and
b1 ¼ b1 ), while the pitch cone of the hypoid wheel touches the pitch plane of the
generation hyperboloid crown wheel along a radius.
Figure 12.17 shows schematically, in three orthogonal projections, a hypoid gear
pair characterized by a shaft angle R ¼ 90 , and having a gear ratio u ¼ z2 =z1 ¼ 1.
The shortest or center distance a ¼ O1 O2 (Fig. 12.17a), between the two crossed
axes of wheels 1 and 2, is called hypoid offset. In all three projections, the boundary
line of the hyperboloid crown wheel in its position relative to the two conical pitch
12.7 Hypoid Gears: Basic Concepts 609
Fig. 12.16 Position of the pitch cone of the hypoid pinion with respect to the pitch plane of the
generation hyperboloid crown wheel
(a) (c)
(b)
Fig. 12.17 Hypoid gear pair with R ¼ 90 and u ¼ 1, in three orthogonal projections
610 12 Spiral Bevel Gears and Hypoid Gears
surfaces is indicated with a dashed line. This hyperboloid crown wheel has in
common with the pitch cone of the hypoid pinion, 1, the generatrix O1 P, and with
the pitch cone of the hypoid wheel, 2, the generatrix O2 P (point P is the mean
point).
Therefore, the two conical pitch surfaces have in common only the point P. For
this reason, the contact between the teeth of the two members of the hypoid gear
pair is almost a point contact, because it has a very limited length.
The discussion of the geometry and kinematics of these types of gearing is very
laborious, and beyond the limits of this textbook. For the study of this case, we refer
the reader to more specialized literature (see Fan [9, 10], Shih et al. [56], Vimercati
[68]). As of now, however, it is to be noted that the hypoid gear pair no longer has
the characteristics of a hyperboloid gear pair. The main reason of this is that the
location of the axodes is out of the meshing zone of hypoid gears, for which the
concept of axodes of hypoid gears cannot be used for practical applications; it is
used only for visualizing the relative velocity.
In this framework, the design of blanks of hypoid gears is aimed at determining
the operating pitch cones instead of the hypoid gear axodes, the latter coinciding
with the two hyperboloids of revolution of the hypoid gear pair. These operating
pitch cones must satisfy the following main requirements:
– Their axes must form the design crossing angle, R, between the axes of rotation
(usually, R ¼ 90 ).
– The hypoid offset, a, between the axes of the operating pitch cones must be
equal to the design value of the hypoid gear set.
– These pitch cones are in tangency at point P located in the meshing zone
between the active flank surfaces of two members of the hypoid gear pair.
– The sliding velocity at point P is directed along the common tangent to the two
curves obtained as intersections of pitch cones with the tooth surfaces; these
curves are called, improperly, helices.
The generation method of mating surfaces, corresponding to a given law of
motion, for power transmission between crossed axes, deserves further attention
with respect to what we said in Chap. 10. Let’s consider a transmission drive with
shaft angle R and shortest distance a ¼ O1 O2 , such as that shown in Fig. 12.18a.
We obtain firstly the pitch surfaces of the gear pair, assuming that the transmission
ratio is a constant. To this end, we assume, as coordinate systems, the systems
O1 ðx1 ; y1 ; z1 Þ and O2 ðx2 ; y2 ; z2 Þ shown in the same Fig. 12.18a. In the first system,
z1 coincides with the axis of the driving wheel 1 (the pinion) and is directed as the
corresponding angular velocity vector x1 , x1 coincides with the shortest distance
straight line and is directed from O1 to O2 , and y1 completes the right-hand
Cartesian coordinate system. In the second system, z2 coincides with the axis of the
driven wheel 2 (the wheel) and is directed as the corresponding angular velocity
vector x2 , x2 coincides with the shortest distance straight line, but is directed from
O2 to O1 , and y2 completes the right-hand Cartesian coordinate system.
12.7 Hypoid Gears: Basic Concepts 611
(a)
(b)
Fig. 12.18 a Axis of the instantaneous screw motion in the relative motion of a gear pair with
crossed axes; b parallelogram of composition of vectors x1 and x2
During the relative motion of the pinion with respect to the wheel, the coordinate
system O1 ðx1 ; y1 ; z1 Þ rotates about the z2 -axis with angular velocity x2 , while the
motion of member 1 with respect to member 2 of the gear pair is a screw motion
which is obtained by composing the two angular velocity vectors x1 and x2 . At
any instant of time t, the axis of the instantaneous screw motion is defined by the
condition that the relative velocity vector vr at any point P is directed as the same
instantaneous axis, i.e. as the instantaneous relative angular velocity vector
xr ¼ ðx1 x2 Þ. In the coordinate system O1 ðx1 ; y1 ; z1 Þ, the axes of which have as
unit vectors respectively ði1 , j1 , k1 Þ, we can write x1 ¼ x1 k1 and
x2 ¼ ðx2 sinRj1 þ x2 cosRk1 Þ. Therefore, we get:
vr ¼
x1 ðP O1
Þ
x2 ðP O1 Þ
i1 j1 k1
i1 j1 k1
0 0 x1
ð12:50Þ
x 1 y 1 z 1
ð x 1 aÞ y1 z1
The parallelism condition between vr and xr leads to write the following ratios:
that define the instantaneous axis of the screw motion with respect to the coordinate
system O1 ðx1 ; y1 ; z1 Þ. From the first of these ratios, we get:
and then
x1 ðx1 þ x2 cosRÞx1 þ x22 þ x1 x2 cosR ðx1 aÞ ¼ 0; ð12:55Þ
and finally
ða x1 Þ x1 ðx1 þ x2 cosRÞ x1 x2 þ cosR
x1
ði þ cosRÞ
¼ ¼ ¼i : ð12:56Þ
x1 x2 ðx2 þ x1 cosRÞ x2 1 þ 1 cosR
x ð1 þ icosRÞ
x2
x1 sinb2
¼ ; ð12:57Þ
x2 sinb1
where b1 and b2 are the angles between the relative angular velocity vector xr and,
respectively, the vectors x1 and x2 or, what is the same, z1 and z2 . Then,
introducing Eq. (12.57) into Eq. (12.56), and considering that R ¼ ðb1 þ b2 Þ, we
obtain:
12.7 Hypoid Gears: Basic Concepts 613
This equation can be write as follows, since we have ½sinb2 þ sinb1 cosðb1 þ b2 Þ
¼ cosb1 sinðb1 þ b2 Þ and ½sinb1 þ sinb2 cosðb1 þ b2 Þ ¼ cosb2 sinðb1 þ b2 Þ:
Thus, we infer that the instantaneous axis of the screw motion divides the center
distance a ¼ O1 O2 between the axes of the two wheels in parts that are directly
proportional to the tangents of angles between the instantaneous relative angular
velocity vector xr and the axes of rotation of the two wheels. Bearing in mind that
R ¼ ðb1 þ b2 Þ, from Eq. (12.57), we get:
(b)
The envelope of the same surface r0 , in the motion of the generation crown
wheel relative to the wheel 2, is instead another surface r2 , which touches the
generation surface r0 in points of a curved line, s20 , different from s10 , which can be
determined analytically. The contact between the active surfaces r1 and r2 is
therefore a point contact, in a point coinciding with the common point to the two
characteristic curved lines s10 and s20 .
Generally, the larger wheel between the two members of the hypoid gear pair is
the driven wheel; its cone angle d2 is thus somewhat greater than that of the other
member of the gear pair, which is the driving pinion ðd2 [ d1 Þ. From the generation
process shown in Fig. 12.19 and described above, it appears that the pinion 1 is a
usual spiral bevel wheel. The hypoid wheel 2 is generated with the same generation
crown wheel used for pinion, i.e. cutting its teeth with the same cutter used to cut
the pinion teeth, with the difference however that the apex of wheel 2 has a different
position compared to that corresponding to the apex of the wheel 1. Furthermore,
the cutter feed motion is now corresponding to the simultaneous rolling and sliding
of two round ruled hyperboloids, instead of the pure rolling of two cones of
revolution.
The contact between the two members of the hypoid gear pair is a point contact.
It is correct from a kinematic point of view only for a given position of the two
members of the gear pair, and thus for given values of the shaft angle and shortest
distance. Therefore, the hypoid gear pair requires a very accurate assembly.
However, hypoid gears have the considerable advantages that we summarized in
Sect. 12.1, which usually overshadow the limitations above. Due to these advan-
tages, in many practical applications it is preferable to use hypoid gears even when
it would possible to realize the transmission of motion with straight bevel and spiral
bevel gears. The most practical applications of these gears are still those charac-
terized by a shaft angle R ¼ 90 .
Finally, it should be noted that the generation of the hypoid gear pair described
above, related to the diagram shown in Fig. 12.19, is only an example to understand
the generation process. We can consider the dual situation with respect to the
example of Fig. 12.19, where the wheel 2 is a usual spiral bevel wheel, while the
pinion is a hypoid wheel. This is done for the generation process of the hypoid gear
pair made with the most of the Gleason gear cutting machines.
The method developed by Gleason for cutting hypoid gears pairs, which is the
most used method, employs the hyperboloid crown wheel as a virtual generation
cutter. The hypoid wheel is generated as a common spiral bevel wheel, i.e. by
arranging the to-be-generated wheel and the generation hyperboloid crown wheel
simulating the cutting tool in such a way that the apices of the two cutting pitch
cones coincide. Instead, the pinion is generated by placing it in such a way that the
axis of rotation of the generation hyperboloid crown wheel and that of the pinion to
be generated are crossed axes. The tooth flank lines on the pitch plane of the
generation hyperboloid crown wheel may be straight lines not converging on the
axis (Fig. 10.5) or curved lines. Of course, it is also possible to cut hypoid gear
pairs whose members are both hypoid gear wheels. In this case, both pinion and
616 12 Spiral Bevel Gears and Hypoid Gears
wheel are generated by placing their blanks in such a way that the axis of rotation of
the generation hyperboloid crown wheel and those of to-be-generated pinion and
wheel are crossed axes.
Let’s consider the pitch cones of the two members 1 (the pinion) and 2 (the wheel)
of a hypoid gear, that are tangent to the plane p, and assume that pinion and wheel
are located above and, respectively, below the pitch plane p, as Fig. 12.19b shows.
In this condition, O1 P ¼ l1 and O2 P ¼ l2 are the generatrixes of contact of the two
pitch cones with plane p; they intersect each other at point P. We can appreciate the
ability to transmit motion and torque, imagining that a fictitious generation crown
wheel is associated with each of the two pitch cones, and precisely the crown wheel
W01 conjugated with wheel 1, and the crown wheel W02 conjugated with wheel 2.
The pitch plane of both crown wheels coincides with the plane p, while their axes of
rotation are different. In fact, the crown wheel W01 has as its axis the straight line h1 ;
perpendicular to plane p and passing through point O1 , while the crown wheel W02
has as its axis the straight line h2 , perpendicular to plane p and passing through
point O2 (Fig. 12.19b).
We now rotate the wheel 1 about its axis, with angular velocity x1 ; the crown
wheel W01 will rotate about its axis h1 with angular velocity X01 ¼ x1 sind1 . We
then rotate the wheel 2 about its axis, with angular velocity x2 ; the crown wheel
W02 will rotate about its axis h2 with angular velocity X02 ¼ x2 sind2 . If the crown
wheel W02 is driven in its motion by the crown wheel W01 , to make possible the
transmission of motion between the two crown wheels, it is necessary and sufficient
that the relative motion of W01 compared to W02 induces, in the point of contact,
velocity vectors which must be tangent to the mating surfaces touching at point
P (see Ferrari and Romiti [11]). As point of contact, P, we consider the mean point.
Let u1 and u2 be the unit vectors of axes of the two wheels, directed in
accordance with the respective angular velocity vectors, x1 and x2 , as Fig. 12.20a
shows. Let’s consider the usual case in which u1 and u2 , and thus the axes of the
12.8 Approximate Analysis of Hypoid Gears 617
two wheels, are orthogonal (R ¼ 90 ). Then consider the Cartesian coordinate
system Pðx1 ; y1 ; z1 Þ with origin at point P, x1 -axis coinciding with the generatrix of
contact of the pitch cone of wheel 1 with plane p, y1 -axis in this plane and normal to
x1 -axis, and z1 -axis normal to plane p; these axes are directed as Fig. 12.20a shows.
From this figure, we can easily derive the following relationships, where fmp \p=2
is the angle between O1 P and O2 P (this is the pinion offset angle in pitch plane,
according to ISO standards), while ði1 , j1 , k1 Þ are the unit vectors of x1 -, y1 -, and z1 -
axes, respectively:
u1 ¼ cosd1 i1 sind1 k1
ð12:61Þ
u2 ¼ cosd2 cosfmp i1 cosd2 sinfmp j1 sina2 k1 :
Since we have assumed R ¼ 90 , i.e. the axes of the two wheels are orthogonal,
the dot product of unit vectors u1 and u2 must be equal to zero ðu1
u2 ¼ 0Þ, for
which, by imposing this condition, from Eqs. (12.61) we get:
and then
(a) (b)
Fig. 12.20 a Axes of wheels 1 and 2, and generatrices of contact of their pitch cones with plane p;
b m-plane through the axis of wheel 1 and shortest distance straight line
618 12 Spiral Bevel Gears and Hypoid Gears
that ðaÞ represents the bending arm of the unit vector u2 applied to point O2 with
respect to the line of action of the unit vector u1 passing through point O1 , so we
can write the following equation, formed by a dot product of cross product:
a ¼ ðO2 O1 Þ u2 u1 ; ð12:64Þ
where ðO2 O1 Þ is the position vector of point O2 with respect to point O1 . In the
coordinate system Pðx1 ; y1 ; z1 Þ, the position vector of points O1 and O2 with respect
to point P are respectively: ðO1 PÞ ¼ Rm1 i1 and ðO2 PÞ ¼ Rm2 cosfmp i1 þ
Rm2 sinfmp j1 , where Rm1 and Rm2 are the mean cone distances. Therefore, from
Eq. (12.64) we have:
Rm2 cosfmp Rm1 Rm2 sinfmp 0
a ¼
cosd2 cosfmp cosd2 sinfmp sind2
cosd1 0 sind1
ð12:65Þ
¼ ðRm2 cosd1 sind2 þ Rm1 sind1 cosd2 Þsinfmp
¼ ðRm1 tand1 þ Rm2 tand2 Þcosd1 cosd2 sinfmp :
Similarly, we get:
a Lcosd2
¼ : ð12:68Þ
sinfmp tand1
(a) (b)
Fig. 12.21 a Taper ratio; b calculation diagram of the relative velocity at mean point P on the
p-plane
differential elements ds1 and ds2 of these curved lines, which can be assumed as
linear, as well as belonging to the plane p. Let t be the straight line of plane p, to
which ds1 and ds2 belong.
We indicate with v2 ¼ ðV2 PÞ and v1 ¼ ðV1 PÞ the velocity vectors at point
P of the wheels 1 and 2, corresponding to their angular velocity vectors x1 and x2
about the respective axes. The condition that the relative velocity vector vr at point
of contact P has component equal to zero in the direction of the common normal of
contact between the mating surfaces involves the condition that this relative ve-
locity vector is directed along the straight line t (Fig. 12.21b). Since the absolute
values of two vectors, v2 and v1 ; are given respectively by v2 ¼ PV2 ¼ x2 Rm2 sind2 ;
and v1 ¼ PV1 ¼ x1 Rm1 sind1 , the following equality must necessarily be satisfied:
where b1 and b2 are the spiral angles of the two wheels, i.e. the angles between the
straight line t and, respectively, the straight lines O1 P and O2 P.
Suppose now that each of the wheels 1 and 2 is in meshing, at point P, with the
associated generation crown wheels, W01 and W02 , in the manner set out above. The
angles b1 and b2 coincide with the spiral angles of the fictitious teeth of these
generation crown wheels, while the tooth-axes of the same crown wheels, at the
instant under consideration, are tangent to each other at point P. We denote these
tooth-axes with k1 and k2 .
To determine the centers of curvature of k1 and k2 at point P, we first observe
that the motion of the crown wheel W02 relative to the crown wheel W01 is a planar
motion, and that the center of the instantaneous rotation C of this motion, at the
instant in which k1 and k2 are touching at point P, belongs to the extended straight
line O1 O2 (Fig. 12.22). On the other hand, point Cmust also belong to the common
620 12 Spiral Bevel Gears and Hypoid Gears
normal to k1 and k2 , which are always tangent to each other, and therefore it
coincides with the point of intersection of these two straight lines.
After an infinitesimal time dt, the point of contact between the meshing teeth
moves from point P to point Q, which can be considered belonging to the p-plane.
Thus point Q will be the new tangency point between k1 and k2 . For ease of
understanding, we materialize the common normal to k1 and k2 through point
P with a semi-infinite rod, q, whose end point on the finished side coincides at every
instant with the point of contact between k1 and k2 . The instantaneous center of
motion of this rod on p-plane must be on the normal to path of this extreme point,
that is the normal to the straight line PQ through point P. It is then to be observed
that, since the angular velocities of rotation of the crown wheels W01 and W02 are
respectively X1 ¼ x1 sind1 and X2 ¼ x2 sind2 , the point C, which is the instanta-
neous center of rotation of their relative motion, will have to satisfy the relationship:
CO1 x2 sind2
¼ : ð12:70Þ
CO2 x1 sind1
Since i ¼ ðx1 =x2 Þ ¼ const, and positions of points O1 and O2 do not vary, from
this latter relationship we infer that also the position of point C on the p-plane
remains in a fixed position during the mutual engagement of the two crown wheels.
Therefore, the common normal to k1 and k2 must constantly pass through this point,
12.8 Approximate Analysis of Hypoid Gears 621
where
are the mean pitch diameters of the two gear members, while K ¼ ðcosb1 =cosb2 Þ
is a ratio whose value Gleason assumes within the range ð1:3 1:5Þ. Therefore,
once the value of K was chosen, since the transmission ratio i is an input datum of
the gear design, Eq. (12.71) allows to derive one of the two mean pitch diameters
when the other is fixed. Correspondingly, Gleason recommended that the taper ratio
was determined using the following relationship:
3 z1 z2
L ¼ dm1 þK ; ð12:73Þ
8 z2 z1
where the subscripts 1 and 2 indicate respectively the driving and driven members
of the gear pair.
622 12 Spiral Bevel Gears and Hypoid Gears
dm1 dm2 2L
þ ¼ : ð12:74Þ
cosd1 cosd2 sind1
L sind1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð12:75Þ
a þL
2 2
sin d1 þ cos2 d1 cos2 d2 sin2 fmp
2
since then, taking into account Eq. (12.63), it can be shown that
2
sin d1 þ cos2 d1 cos2 d2 sin2 fmp ¼ cos2 d2 , Eq. (12.75) becomes:
L sind1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ : ð12:76Þ
a þL
2 2 cosd 2
L 2L dm1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ tand1 : ð12:77Þ
a þL
2 2 dm2 dm2
With this equation, we calculate d1 , since the other quantities that appear in it are
now known. By Eq. (12.76), we calculate d2 . After that, in order to determine fmp ,
Rm1 and Rm2 , just use Eqs. (12.63), (12.65) and (12.66). Finally, taking into account
that ðcosb2 =cosb1 Þ ¼ ð1=K Þ and ðb2 b1 Þ ¼ fmp , we obtain:
cosfmp ð1=K Þ
tanb1 ¼ ; ð12:78Þ
sinfmp
xi ¼ ui sindi cos#i
yi ¼ ui sindi sin#i ð12:79Þ
zi ¼ ui cosdi ;
where ui an #i are the Gaussian coordinates (see Sect. 11.5), while i ¼ ð1; 2Þ.
According to Eqs. (11.51) and (11.53), the surface normal N i and unit normal ni are
given by the equations:
@ri @ri
N i ¼ @u i
@# i
ni ¼ jN
Ni j ;
i
ð12:80Þ
To define the equations of tangency of the operating pitch cones at mean point P,
which is the pitch point, it is convenient to represent these pitch cones in a fixed
coordinate system. To this end, we consider, in addition to the two coordinate
systems, O1 ðx1 ; y1 ; z1 Þ and O2 ðx2 ; y2 ; z2 Þ, rigidly connected to the two members 1
and 2 of the hypoid gear pair, also the coordinate system O0 ðx0 ; y0 ; z0 Þ, rigidly
connected to the frame. Figure 12.24 shows the position and orientation of the two
systems O1 ðx1 ; y1 ; z1 Þ and O2 ðx2 ; y2 ; z2 Þ, one with respect to the other, as well as
that of these two systems with respect to the fixed system O0 ðx0 ; y0 ; z0 Þ. By means
of a coordinate transformation from O1 ðx1 ; y1 ; z1 Þ and O2 ðx2 ; y2 ; z2 Þ to
O0 ðx0 ; y0 ; z0 Þ, we obtain the following vector functions:
2 3
r1 cos#1
r1;0 ðu1 ; #1 Þ ¼ 4 r1 sin#1 5 ð12:82Þ
r1 cotd1 l1
624 12 Spiral Bevel Gears and Hypoid Gears
2 3
cosd1 cos#1
n1;0 ð#1 Þ ¼ 4 cosd1 sin#1 5 ð12:83Þ
sind1
2 3
r2 cos#2 þ a
r2;0 ðu2 ; #2 Þ ¼ 4 r2 cotd2 þ l2 5 ð12:84Þ
r2 sin#2
2 3
cosd2 cos#2
n2;0 ð#2 Þ ¼ 4 sind2 5; ð12:85Þ
cosd2 sin#2
which express, in the fixed coordinate system O0 ðx0 ; y0 ; z0 Þ, the pitch cones of pinion
and wheel and their unit normals. It is to be noted that the first of the two subscript
refers to the member 1 or 2 of the hypoid gear pair, while the second refers to the
fixed reference system, O0 ðx0 ; y0 ; z0 Þ. In these vector functions, the meaning of
symbols already introduced remains unchanged. Instead, the new symbols represent:
• ri ¼ ui sindi (with i ¼ 1; 2), is the radius of the cross section at point P of each of
the two pitch cones (being P the mean point, ri is the mean pitch radius, that is
ri ¼ rmi ¼ dmi =2; the use of a single subscript is here preferred for the sake of
brevity, but the indication with double subscript will be used in the final part of
this section);
• li (with i ¼ 1; 2), indicates the location of the apex of each of the two pitch
cones (Fig. 12.24);
• a, is the shortest distance.
The equations of tangency of the operating pitch cones at the pitch point P are
given by:
r1;0 ðu1 ; #1 Þ ¼ r2;0 ðu2 ; #2 Þ ¼ rP;0 ð12:86Þ
where rP;0 and nP;0 are respectively the position vector and the common normal to
the pitch cones at pitch point P. Since the mating pitch cones are located above and
below the pitch plane (see also Fig. 12.19), their surface unit normals at point
P have opposite directions. Therefore, the coincidence of the surface unit normal is
obtained with a negative sign in Eq. (12.87). Vector Eqs. (12.86) and (12.87) give
the following system of six scalar equations:
It should be noted that only two of the last three equations of the system (12.88)
are independent, because it must be jn1:0 j ¼ jn2:0 j ¼ 1.
Making some transformations and eliminating the trigonometric functions con-
taining the variable #i , from Eqs. (12.88) we obtain the following relationships,
which are the basis for design of hypoid pitch cones:
r2 acosd1 cotd1
l1 ¼ þ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð12:90Þ
sind1 cosd2 cos2 d1 sin2 d2
r2 asind2
l2 ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð12:91Þ
sind2 cosd2 cos d1 sin2 d2
2
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
r2 cos2 d1 sin2 d2
xP;0 ¼ a ð12:92Þ
cosd2
asind2
yP;0 ¼ r2 tand2 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð12:93Þ
cos d1 sin2 d2
2
r2 sind1
zP;0 ¼ ð12:94Þ
cosd2
626 12 Spiral Bevel Gears and Hypoid Gears
8 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
< nP;x0 ¼ cos2 d2 sin2 d1 ¼ cos2 d1 sin2 d2
nP;y0 ¼ sin d2 ð12:95Þ
:
nP;z0 ¼ sin d1 :
It is to be noted that, in Eqs. (12.95), two subscripts are used: the first indicates
the point at which the normal refers (in this case, point P); the second, composed of
a letter with numeric subscript, indicates the component of the unit normal in the
direction of the corresponding axis of the coordinate system O0 ðx0 ; y0 ; z0 Þ.
We indicate now with s1 and s2 the unit vectors of the generatrices of contact of
the pitch cones with the pitch plane, which intersect each other at pitch point P, and
are directed as Fig. 12.24 shows. Consider then the tooth traces k1 and k2
(Fig. 12.22) of the mating tooth flanks at point P on the pitch plane, i.e. the
intersections of the tooth flank surfaces with the pitch plane. These tooth traces or
flank lines, improperly called spirals or helices, are curves depending on the
geometry of the teeth, and therefore on the cutting process used to obtain
the toothing. Obviously, these flank lines have a common tangent at point P on the
plane, p. The so-called mean spiral angles, bm1 ¼ b1 and bm2 ¼ b2 , in the pitch
plane are the angles between the common tangent to the tooth traces k1 and k2 and
the generatrices of the respective pitch cones passing through point P (Fig. 12.25).
The pinion offset angle in pitch plane, fmp (Fig. 12.25) is given by the following
dot product:
cosfmp ¼ s1
s2 : ð12:96Þ
Considering that the unit vectors s1 and s2 can be expressed by the following
relationships:
O1 P @r1:0 =@u1
s1 ¼
¼ ¼ ½ sind1 cos#1 sind1 sin#1 cosd1 T
O1 P
j@r1:0 =@u1 j
ð12:97Þ
O2 P @r2:0 =@u2 T
s2 ¼
¼
¼ ½ sind2 cos#2 sind2 sind2 sin#2 :
O2 P j@r2:0 =@u2 j
We have thus obtained the same Eq. (12.63), previously deduced otherwise.
The relative velocity vector or sliding velocity vector of the pinion with respect
to the wheel at point P (Fig. 12.25) can be expressed as (see also Eq. 11.85):
Fig. 12.25 Pinion offset angle in pitch plane, mean spiral angles, and relative velocity
action of vector x2 with respect to the origin O0 of the same coordinate system, i.e.
O0 Q (Fig. 12.24). The angular velocity vector x1 is a vector passing through the
origin of coordinate system O0 ðx0 ; y0 ; z0 Þ. For point P, vectors v1 and v2 lie in the
pitch plane, and are perpendicular to the generatrices of contact of the pitch cones
with this plane. By means of some transformations (see Sect. 11.7), from
Eq. (12.99) we obtain: 2 3
0
v1;2 ¼ x1 r1 ðtanb1 tanb2 Þcosb1 4 sinb1 5 ð12:100Þ
cosb1
x1 r2 cosb2 z2
i¼ ¼ ¼ : ð12:101Þ
x2 r1 cosb1 z1
The relative velocity vector vr ¼ v1;2 given by Eq. (12.100) is related to the
coordinate system Oe ðe1 ; e2 ; e3 Þ, where e1 is the unit vector normal to the pitch
plane, e3 ¼ s1 is the unit vector corresponding to the generatrix of contact of the
pinion pitch cone with the pitch plane, and e2 ¼ e3 e1 . From Eqs. (12.100) and
(12.101) we get:
ir1 r2 cosfmp
tanb1 ¼ ð12:102Þ
r2 sinfmp
The quantities bi , di and li (width i ¼ 1; 2) are the basic design parameters of the
hypoid gear pair under consideration. Parameters li , which determine the location of
628 12 Spiral Bevel Gears and Hypoid Gears
the pitch cone apexes in the coordinate system O0 ðx0 ; y0 ; z0 Þ, as Fig. 12.24 shows,
are calculated with Eqs. (12.90) and (12.91), once the other parameters that appear
in these equations have been determined. Instead, the four parameters bi and di are
related by three equations, which briefly can be written in the following form:
f1 ðd1 ; d2 ; b1 Þ ¼ 0
f2 ðd1 ; d2 ; b1 ; b2 Þ ¼ 0 ð12:104Þ
f3 ðd1 ; d2 ; b1 ; b2 Þ ¼ 0:
To determine these three equations in explicit form, we consider the usual case
for which the parameter b1 is a design data (usually, b1 ¼ 45 ). The first two of
these equations do not change for the two possible types of hypoid gear pairs that
we have described in Sect. 12.3, i.e. those cut with a face-milling process, which
have tapered teeth, and those cut with a face-hobbing process, which have teeth of
uniform depth. The third equation instead changes, for which it must be specifically
derived for each of these two types of hypoid gear pairs. This is because the
face-milled tapered teeth are generated by a surface (the cone surface of the
head-cutter), while the face-hobbed teeth having uniform depth are generated by a
straight-line segment materialized by the blade edge.
For derivation of the first two Eq. (12.104) in explicit form, first we obtain from
Eqs. (12.89) and (12.101) the following relationship:
cosb1
cosb2 ¼ ; ð12:106Þ
p
2 2
cos b1
q ¼ ðsinb1 sinb2 Þ2 ¼ 1 cos2 b1 1 cos2 b2
p
cos2 b1
¼ 1 cos b1 1
2
; ð12:109Þ
p
ð1 q2 Þb2
f1 ðd1 ; d2 ; b1 Þ ¼ cos2 b1 ¼ 0: ð12:110Þ
ð1 þ p2 2pqÞ
Thus the first of Eqs. (12.104) has been obtained. In fact, q is a function of d1
and d2 (see Eq. 12.108), but also p is a function of d1 and d2 , because the quantities
z1 , z2 , a and r2 appearing in Eq. (12.107) are considered as known.
The second of Eqs. (12.104) written in explicit form has been already obtained.
In fact, it consists of the Eq. (12.98), which can be written as follows:
The derivation of the third of Eqs. (12.104) for hypoid gear pairs with
face-milled teeth is based on the limit contact normal, introduced for the first time
by [73], and applied by Gleason to design of these types of gears. With reference to
the Wildhaber’s concept of limit contact normal and associated limit pressure
angle, we refer the reader directly to the Wildhaber’s paper [73]. However, it is
worth noting that, in the Gleason approach, by a procedure on which it is not the
place to dwell, it is possible to express the limit pressure angle alim , i.e. the angle
between the limit contact normal to the hypoid wheel tooth surface at point P and
the pitch plane, by the following relationship:
r2
sinb2 sind
r1
sinb1
tanalim ¼ sind2 1
: ð12:112Þ
r2
sind2 þ r1
sind1
From this equation, negative values of the normal pressure angle can be derived
ðalim \0Þ, which have no physical meaning. It follows that the limit pressure angle,
alim , can assume the value alim ¼ 0 as the minimum value (see also: Litvin [30];
Litvin and Fuentes [32]).
Figure 12.26 shows both tooth profiles of the hypoid wheel, in the normal
section passing through the pitch point P, which are not symmetrical with respect to
the normal to the pitch plane passing through the same pitch point P. In this figure,
n is the unit vector of the limit contact normal, while n1 and n2 are the unit normals
to the concave and convex tooth sides, whose lines of action form with the line of
action of n the same angle. It follows that the normal pressure angles an1 and an2 of
the corresponding profiles must satisfy the following relationship:
630 12 Spiral Bevel Gears and Hypoid Gears
It is to be noted that an1 is different from an2 , and in accordance with the Gleason
approach, the pressure angle an1 on the concave side is larger than the pressure
angle an2 on the convex side, i.e. an1 ¼ an2 þ 2jalim j.
Another equation can be derived that correlates the limit pressure angle with the
design parameters of the pitch cones, and is independent from the previous one. The
derivation of this equation is based on the consideration that a hypoid gear pair with
face-milled teeth (and then obtained by means of a formate-cut process) consists of
a generated pinion (which does not interest here), and a non-generated wheel whose
tooth flank surfaces coincide with the head-cutter surface. In accordance with what
we said in Sect. 12.3 , the intersection of the wheel tooth surface with the pitch
plane is an arc of circle having radius rc0 , which coincides with the mean radius of
the head-cutter; it is given by the following relationship:
tanb1 tanb2
rc0 ¼ : ð12:114Þ
sind1
r1 cosb1 r2sind
cosb
2
tanb1rcosd
1
1
þ tanb2 cosd2
r 2
tana lim
2
The Eqs. (12.112) and (12.114) taken together make it possible to obtain the
explicit form of the requested third of Eqs. (12.104), valid for hypoid gear pairs
with face-milled teeth.
Instead, the derivation of the third of Eqs. (12.104) for hypoid gear pairs with
face-hobbed teeth is based on the specific location of the head-cutter for the gen-
eration of these types of gears. Figure 12.27 shows the generatrices of contact O1 P
and O2 P of the two pitch cones with the pitch plane (the plane passing through
points O1 , O2 , and P), the pitch circles that are tangent each other at point P, and the
unit vectors s1 ¼ O1 P and s2 ¼ O2 P of the pitch cone generatrices, whose lines of
action converge at point P. We find now the point C of intersection of the
12.8 Approximate Analysis of Hypoid Gears 631
(b)
head-cutter with the pitch plane, assuming that it belongs to the extended line O1 O2 ,
in accordance with the set-up of the gear cutting machine, comprised of head-cutter.
We denote then with z0 the number of finishing blade groups of this head-cutter.
Let’s consider now the imaginary generation crown wheel, simultaneously in
meshing with pinion and wheel of the hypoid gear pair, whose axode is the
aforementioned pitch plane, which can be regarded as a circular cone. Assume also
that, while the head-cutter rotates about its axis with angular velocity x0 , the
generation crown wheel rotates about its axis with the angular velocity X0 ,
remembering that both these axes are perpendicular to the pitch plane, and pass
respectively through points C and O2 . If I is the instantaneous center of rotation of
the head-cutter in its relative motion with respect to the generation crown wheel, its
position on the extended line O1 O2 will be given by the following relationship:
632 12 Spiral Bevel Gears and Hypoid Gears
O2 I x0 Z0 z2
¼ ¼ ¼ ; ð12:115Þ
IC X0 z0 z0 sind2
where Z0 ¼ z2 =sind2 is the number of teeth of the imaginary crown wheel (of
course, Z0 must be an integer number); z2 and d2 are respectively the number of
teeth and pitch cone angle of the hypoid wheel.
The finishing blade is located in the plane perpendicular to the pitch plane,
passing through the line PI. Point P of this blade generates in the pitch plane an
extended epicycloid whose normal at point P coincides with the line PI. From
Fig. 12.27b, we can infer the following proportion:
O2 A O2 B
¼ : ð12:116Þ
O1 A CB
O1 A ¼ O1 Psinðb1 b2 Þ
O2 A ¼ O2 P O1 Pcosðb1 b2 Þ
ð12:117Þ
CB ¼ r0 cosðb2 d0 Þ
O2 B ¼ O2 P r0 sinðb2 d0 Þ;
z0 r2 cosb2
sind0 ¼ : ð12:120Þ
z2 r0
The derivation of this last equation is based on the following three relationships,
as inferred from Fig. 12.27b:
12.8 Approximate Analysis of Hypoid Gears 633
bsind2
r2 ¼ rm2 ¼ re2 : ð12:122Þ
2
About the iterative procedure of solution, at each iteration we can consider the
equations represented in echelon form, so that they can be solved separately in
the case in which one of the unknowns (for example, d2 ) is regarded as given. Then
the third nonlinear equation is used for checking of the iterative procedure. The
determination of the three unknowns d1 , d2 and b2 the appear in the equation
system (12.104), written in explicit form, involves the use of computer-aided
programs, with appropriate subroutines for the numerical solution of nonlinear
equations. However, when these subroutines are used, calculations must be com-
pleted by requiring that the following requirements be met:
In the two previous sections, we have seen that the hypoid gears are gear wheel
pairs of conical or approximately conical shape, which mesh with their axes crossed
and offset. We can say that the hypoid gears are special spiral bevel gears that
transmit motion and torque between crossed axes, arranged with a given offset each
634 12 Spiral Bevel Gears and Hypoid Gears
other. Therefore, the axis of the pinion does not intersect the axis of the wheel, and
has a hypoid offset, a, with respect to it. Generally, this offset has a not high value.
With a suitable value of this offset, it is possible to have the axis of the pinion
below that of the wheel. Thus, it becomes possible to arrange the bearings strad-
dling the toothed part. Typical advantages of this relative position between the two
axes that transmit the motion are manifold.
The first great advantage of the hypoid gear pairs is that they allow us to have
crossed axes of the two gear members, and therefore to support their shafts with
bearings arranged at the two sides of the same gear wheels. Thus, the cantilever
mounting of the pinion, inevitable for gear pairs with limited space, is avoided.
The second great advantage of these gears is related to the fact that the shaft of
the pinion, which is the driving member, occupies a lower position with respect to
the shaft of the wheel. This relative position of the two members of the gear pair is
adopted by many automobile manufacturers, and used in the differential for the rear
axle drive of the cars (see Powell and Barton [46], Pollone [44], Lechner and
Naunheimer [25], Naunheimer et al. [41]). This arrangement with a lowered drive
shaft, and consequently lowered transmission shaft, allows in fact having a
lower-floored body of the vehicle. It follows a lowering of the center of gravity of
the transmission system and, therefore, of that of the whole vehicle, and thus a
greater stability of the same vehicle.
A third advantage achievable with hypoid gears, which are characterized by
sufficient offset between the two shafts, concerns the possibility of supplying power
to several machines, using a single input shaft on which several hypoid pinions are
mounted. This advantageous and useful feature is used in various industrial
applications (e.g., textile machines and plants, machine tools, etc.).
However, the kinematics of the hypoid gear pair is characterized by sliding, not
only in the direction of the tooth depth, but also along the tooth trace, and the latter
sliding increases with the increasing of the offset. Therefore, the fourth advantage of
the silent operating conditions, typical of this gear and related to the longitudinal or
lengthwise sliding, is associated with other significant disadvantages, such as lower
efficiency, more high wear, and higher operating temperatures, with consequent risk
of scuffing (see Coleman [6], Kolivand et al. [23], Stadtfeld [62]).
At first sight, the hypoid gears can be confused with the spiral bevel gears, as the
shape of their teeth is very similar to that of the latter. A substantial difference exists
however between the two types of gears: in fact, the axodes, i.e. the pitch surfaces,
are two cones in the spiral bevel gears, while they are two hyperboloids of revo-
lution in hypoid gears. Hypoid gears transmit roughly the same power of spiral
bevel gears, having equal nominal pitch diameters and equal transmission ratio.
The hypoid offset a ¼ O1 O2 between the axes of the two members of a hypoid
gear pair may reach 30% of the mean pitch diameter of the hypoid wheel, dm2 , when
the gear ratio is equal to 1ðu ¼ 1Þ, by must not exceed 20% of the same mean pitch
diameter if u [ 3. For gear ratios included in the range ð1 u 3Þ, the hypoid
offset can be made to vary linearly between the two aforementioned values. Other
limitations exist in regard to the magnitude of the hypoid offset. Indeed, it should
12.9 Main Characteristics of the Hypoid Gears, and Some Indications … 635
Fig. 12.28 Hypoid gear pairs with: a positive hypoid offset; b negative hypoid offset; c without
hypoid offset (spiral bevel gear pair)
not exceed 40% of the outer cone distance, Re2 , of the hypoid wheel and, for heavy
duty equipment, it should be nearer 20% of the same outer cone distance.
The hypoid offset can be positive or negative, depending on whether the axis of
the hypoid pinion is placed below or above the axis of the hypoid wheel.
Figure 12.28 shows three hypoid gear pairs, respectively with positive hypoid offset
(Fig. 12.28a), negative hypoid offset (Fig. 12.28b), and without hypoid offset
(Fig. 12.28c). In this latter case, the hypoid gear becomes a normal spiral bevel
gear, which can therefore be considered a special case of the more general family of
hypoid gears (see also Niemann and Winter [42]). Figures 12.28a and 12.28b also
highlight the pinion offset angle in axial plane of the same pinion, fm (see also
Fig. 12.36b).
With a positive hypoid offset, the hypoid pinion has a larger mean spiral angle,
bm1 , than that of the hypoid wheel,
for which bm2 \bm1 . With reference to the mean
cones, we have in fact bm1 ¼ bm2 þ fmp , where fmp is the pinion offset angle in
pitch plane, i.e. measured in the pitch plane of the common generation hyperboloid
crown wheel. With the same mean pitch diameter, dm2 , of the hypoid wheel and the
same transmission ratio, i, the mean pitch diameter, dm1 , and pitch angle, d1 , of the
hypoid pinion are larger than the values that define the corresponding spiral bevel
gear pair. Also, the outer transverse module, met , will be increased, as a result of the
increase of bm1 . At the same time, the overlap ratio increases, and so does the axial
force with respect to a spiral bevel gear pair, i.e. a gear pair without offset. The
greater diameter of the pinion also makes possible an increase of the diameter of its
shaft. This is the preferred solution in the automotive industry. In borderline cases,
bm1 ¼ 0 (straight bevel and Zerol bevel hypoid pinion), and bm2 ¼ 0 (straight bevel
or Zerol bevel hypoid wheel).
Conversely, with a negative hypoid offset, the hypoid pinion has a smaller mean
spiral angle than that of the hypoid wheel: thus bm1 ¼ bm2 fmp . With the same
mean pitch diameter of the hypoid wheel and the same transmission ratio, the mean
pitch diameter, pitch angle, overlap ratio, and axial force are smaller than the values
that define the corresponding spiral bevel gear pair without offset. In the extreme
case, the hypoid pinion assumes a cylindrical shape, becoming a cylindrical pinion.
Besides the advantages mentioned above, the hypoid gears are more suitable
than the ordinary spiral bevel gears in industrial applications that require a smooth
636 12 Spiral Bevel Gears and Hypoid Gears
and quiet running. In addition, the hypoid gears can ensure high-speed reductions,
with transmission ratios that, normally, can reach and even exceed values of 60:1.
Finally, a hypoid gear pair is a very compact gear drive, so that it is possible to
reduce the overall size of the same gear drive, with a lower number of teeth of the
pinion and, together, a greater strength of the driving member, i.e. the pinion.
Of course, the meshing action is subject to considerable lengthwise sliding along
the teeth, and this sliding leads to great power losses and, therefore, to local heating.
Temperatures related to this local heating can be so high as to determine hot spots
and consequent micro-welding. To inhibit the tendency to the generation of these
local micro-welding, as well as to withstand the high pressure of contact, anti-scuff
lubricants (the so-called EP, Extreme Pressure) are needed.
Finally, it is to emphasize the need to use bearings able to withstand the axial
thrusts, and an adequate lubrication circuit, which also has the task of dissipating the
heat generated due to the high power losses. In the hypoid gears, these power losses,
and the consequent efficiency, are highly dependent on many parameters, such as the
hypoid offset, mean spiral angles, load in the design operating conditions, finishing
of the tooth flank surfaces, rotational speed, characteristics of the lubricant, etc.
Only the study and most recent analyses have shown the great influence of the
load exerted on the hypoid gear efficiency. This predominant influence as well as
that exerted by the coefficient of friction have led to the proposal of approximate
empirical formulas, other than those previously developed and proposed, which
were characterized by the specific parameters of the tooth geometry. The disap-
pearance of these parameters in the more recent formulas is due to the fact that
many of these geometric parameters have an influence entirely negligible in front of
the coarse approximation that inevitably characterizes the evaluation of the coef-
ficient of friction, l.
To improve the efficiency of the hypoid gears, as well as that of the spiral bevel
gears, now the margins are not very high, since the manufacturing technologies and
assembly techniques of these gears are so advanced as not to allow further large
gains. From the design point of view, it is to be remembered that the efficiency of
the hypoid gears increases, reducing the hypoid offset, pressure angle, and the
difference between the mean spiral angles, increasing the operating load and rota-
tional speed, improving the finishing quality of the tooth flank surfaces, and using
the best lubricants.
One of the approximate empirical formulas, mainly used for the calculation of
the hypoid gear efficiency, is that developed and proposed by Gleason; we can write
this formula in the following form:
100
g¼
;
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð12:124Þ
2
1þ Tmax
T lsecaeD ðtanbm1 tanbm2 Þ þ k1 þ k2
12.9 Main Characteristics of the Hypoid Gears, and Some Indications … 637
where: Tmax is the maximum torque, to be taken equal to 2:75 times the torque TFlim
corresponding to the nominal stress number for bending rFlim Tmax ¼ 2:75Tflim ;
T is the operating torque; l is the coefficient of friction; aeD is the effective pressure
angle on drive side of the driving member of the hypoid gear pair;
secaeD ¼ ð1=cosaeD Þ; bm1 and bm2 are respectively the mean spiral angles of the
pinion and wheel; k1 and k2 are two numerical dimensionless factors that take into
account the profile sliding losses and, respectively, the bearing losses. The
expression ðtanbm1 tanbm2 Þ2 appearing in this formula takes into account the
lengthwise sliding losses.
To take into account the profile sliding losses, Gleason recommended to take
k1 ¼ 2:25 102 . For the most frequent case in which the shafts of pinion and
wheel are supported by anti-friction journal bearings, Gleason recommended to take
k2 ¼ 0:01. When other types of bearings are used, this value of k2 must be
increased, determining it preferably by means of suitable experimental measure-
ments. The coefficient of friction, l, varies mainly depending on the lubricant used
and its viscosity, rolling and sliding velocities between the tooth mating surfaces,
finishing quality of these surfaces, and transmitted load. The experimental evidence
in this regard shows that it is included in the range ð0:01\l\0:12Þ. Usually, a
value l ¼ 0:05 is considered as a reasonable average value to be taken in the
efficiency calculations.
The curves shown in Fig. 12.29 allow us to immediately see the great influence
exerted by the ratios ða=dm2 Þ and ðT=Tmax Þ on the efficiency of the spiral bevel
gears and hypoid gears. These curves were obtained with aeD ¼ 20 , l ¼ 0:05, and
with certain values of bm1 and bm2 (of course, bm1 ¼ bm2 for spiral bevel gears),
considered to be representative of mean operating conditions. The same curves
.
. . . . .
638 12 Spiral Bevel Gears and Hypoid Gears
show that, ceteris paribus, the hypoid gears have a lower efficiency than that of
spiral bevel gears, with much higher differences, the greater the hypoid offset.
The values of the efficiency of straight bevel, Zerol, and spiral bevel gears are
almost equivalent. They are very much influenced by the characteristics of the
bearings and assembly, and by the finishing quality of the tooth mating surfaces.
For ðT=Tmax Þ ¼ 1, that is for transmission of torques equal to Tmax defined above,
efficiency values included in the range ð0:98 g 0:99Þ can be achieved with these
types of gears, when they have a high quality accuracy, and are properly mounted
and supported by anti-friction journal bearings. For lower loads, the efficiency
decreases, assuming values of about 0:96, for torques that cause root bending
stresses equal to the nominal stress number for bending, rFlim .
Due to the lengthwise sliding losses, hypoid gears are characterized by much
lower efficiency values than those of the aforementioned gears. In fact, with the
exception of gear pairs characterized by high transmission ratio (i ¼ 10, and
greater), the values of efficiency of these gears are included in the range
ð0:90 g 0:98Þ, for loads for which the ratio T=Tmax is equal to 1. For torque
values which cause root bending stresses equal to the nominal stress number for
bending, rFlim , the efficiency of these gears decreases again to values in the range
ð0:86 g 0:97Þ, depending on the transmission ratio and pinion hypoid offset.
Of course, it is to consider that the above-mentioned need to reduce the hypoid
offset, difference between the mean spiral angles, and pressure angle, in order to
increase the efficiency of these gears, contrasts with the current conveniences in the
design of automotive transmissions. In fact, for these applications, it is required to
have centers of gravity at a lower level (to this end, a larger hypoid offset is
required), low noise (to this end, a larger difference between the mean spiral angles
is required), and greater tooth strength (to this end, a larger value of the pressure
angle is required). However, we know that the design of any mechanical system is
the result of a reasonable compromise between conflicting requirements.
Finally, it is not the case to investigate what the contribution of the various
dissipation phenomena that contribute to determine the total efficiency, g, of the
hypoid gears. On this subject, we refer the reader to textbooks that are more
specialized as well as to the scientific literature references (see, for example:
Coleman [6]; Kolivand et al. [23]; Stadtfeld [62]; Artoni et al. [1]).
For determining the forces acting on the hypoid gear pairs, the considerations
carried out in Sect. 12.6, regarding the analysis of loads acting on the spiral bevel
gears, are still valid. In fact, the hypoid gears are the generalization of spiral bevel
gears, which are a special case of hypoid gears. We focus our attention on the
general aspects of this problem, while for the more particular ones we refer the
reader to specialized scientific literature (see, for example: Vimercati and Piazza
[69]; Wang et al. [70]; Kolivand and Kahraman [22]).
12.10 Load Analysis for Hypoid Gears 639
The tangential force, Fmt2 (in N), at the mean pitch diameter dm2 (in mm) of the
hypoid wheel is given by:
2000T2
Fmt2 ¼ ; ð12:125Þ
dm2
where T2 (in Nm) is the torque transmitted by the hypoid wheel. The tangential
force, Fmt1 (in N), at the mean pitch diameter dm1 (in mm) of the mating hypoid
pinion is instead given by:
where T1 (in Nm) is the torque transmitted by the hypoid pinion, and bm1 and bm2
are the mean spiral angles of pinion and wheel.
The Eqs. from (12.43) to (12.45), which express respectively the tangential, axial,
and radial forces acting on the meshing teeth of spiral bevel gears, as a function of
the total tooth force, normal pressure angle, pitch angle and spiral angle on the mean
cone, and Eqs. (12.46) and (12.47) resulting from the equations above, are also valid
for hypoid gears. Of course, when they are written with reference to the hypoid
pinion or hypoid wheel, we add respectively additional subscripts 1 and 2.
The axial force, Fa , can be a repulsive force (in this case, the two members of the
hypoid gear pair tend to move away each other, and the axial force is considered
positive), or a suction force (in this case, the two members of the hypoid gear pair
tend to approach each other, and the axial force is considered negative). For signs
that appear in Eqs. (12.46) and (12.47), the same conventions as described in
Sect. 12.5 are valid.
Figure 12.30 highlights, in a synoptic view, as the direction of the axial thrust,
Fa1 , acting on the pinion of a hypoid gear pair changes with the hand of the spiral,
direction of rotation and position of the hypoid pinion with respect to the hypoid
wheel (positive or negative hypoid offset). The four possible cases are considered,
namely: pinion with right-hand spiral (RH-spiral) and clockwise rotation
(Fig. 12.30a); pinion with right-hand spiral (RH-spiral) and counterclockwise
rotation (Fig. 12.30b); pinion with left-hand (LH) spiral and clockwise rotation
(Fig. 12.30c); pinion with left-hand spiral (LH-spiral) and counterclockwise rota-
tion (Fig. 12.30d).
Figure 12.31 shows as the direction of the total tooth force, and that of its axial
and tangential components, acting on the pinion of a hypoid gear pair, change with
the hand of the spirals, direction of rotation of the two members, and position of the
hypoid pinion with respect to the hypoid wheel (positive or negative offset).
It is necessary to note that the hand of the spiral must be such that the spiral
angle of the pinion at mid-face width is greater than that of the hypoid wheel, i.e.
bm1 [ bm2 . In addition, if the axial thrust acting on the pinion is a repulsive force,
the active flank of the pinion is the concave flank, while, if the axial thrust acting on
the pinion is a suction force, the active flank of the pinion is the convex flank.
640 12 Spiral Bevel Gears and Hypoid Gears
Fig. 12.30 Direction of the axial thrust acting on the pinion as a function of the hand of the spiral
(RH and LH), direction of rotation, and positive or negative offset
while for coast side flank loading, the pinion axial force, Fax1;C , and wheel axial
force, Fax2;C , are given respectively by:
sind1
Fax1;C ¼ Fmt1 tananC tanbm1 cosd1 ð12:129Þ
cosbm1
sind2
Fax2;C ¼ Fmt2 tananC þ tanbm2 cosd2 : ð12:130Þ
cosbm2
In these equations, anD and anC are the generated pressure angles on drive and,
respectively, coast side, while the meaning of other symbols already introduced
remains unchanged. It is also to be noted that the positive sign (+) indicates
12.10 Load Analysis for Hypoid Gears 641
Fig. 12.31 Direction of the total tooth force and its axial and tangential components, acting on the
pinion as a function of the hand of the spirals (RH or LH), direction of rotation, and positive or
negative offset
Table 12.1 Loaded tooth Driver hand Rotation of driver Loaded flank
flank of spiral Driver Driven
Right Clockwise Convex Concave
Counterclockwise Concave Convex
Left Clockwise Concave Convex
Counterclockwise Convex Concave
direction of thrust is away from pitch apex, while the negative sign (−) indicates
direction of thrust is toward pitch apex.
For drive side flank loading, the pinion radial force, Frad1;D , and wheel radial
force, Frad2;D , are given respectively by:
642 12 Spiral Bevel Gears and Hypoid Gears
cosd1
Frad1;D ¼ Fmt1 tananD tanbm1 sind1 ð12:131Þ
cosbm1
cosd2
Frad2;D ¼ Fmt2 tananD þ tanbm2 sind2 ; ð12:132Þ
cosbm2
while for coast side flank loading, the pinion radial force, Frad1;C , and wheel radial
force, Frad2;C , are given respectively by:
cosd1
Frad1;C ¼ Fmt1 tananC þ tanbm1 sind1 ð12:133Þ
cosbm1
cosd2
Frad2;C ¼ Fmt2 tananC tanbm2 sind2 : ð12:134Þ
cosbm2
In these equations, the symbols are all already known. As for the signs, it should
be noted that the positive sign (+) indicates direction of force is away from the
mating member (this force is usually called the separating force), while the neg-
ative sign (−) indicates direction of force toward the mating member (this force is
usually called the attracting force).
In this section, we take a unified discussion of bevel and hypoid gear geometry, in
accordance with ISO 23509:2016 (E) [20]. Therefore, from now on, the term bevel
gear(s) is to be understood in its more general meaning, i.e. including straight, skew
(or helical) and spiral bevel, Zerol and hypoid gear(s). Thus, symbols and rela-
tionships that gradually we introduce, unless noted otherwise, are applicable to all
these types of gears. This unified treatment is also compliant with the introduction
of universal, multi-axes, CNC-gear cutting machines that, in principle, are capable
of producing nearly all types of gearing, including the different curved-toothed
bevel gears previously described as well as hypoid gears.
Hypoid gear pairs, which consist of pinion and wheel with skew and
non-intersecting axes, and with curved teeth in the lengthwise direction, can be
considered the most general type of gear pair. All other types of gear pairs are
subsets of the hypoid gear pairs. So spiral bevel gears, including skew (or helical)
bevel gears and Zerol bevel gears, are hypoid gears with zero hypoid offset, straight
bevel gears are hypoid gears with zero hypoid offset and zero tooth curvature, and
parallel cylindrical helical gears are hypoid gears with zero shaft angle and zero
tooth curvature.
12.11 Bevel and Hypoid Gear Geometry: Unified Discussion 643
Figure 12.32 shows a hypoid gear pair with shaft angle R ¼ 90 , and highlights
some of its geometrical quantities. In hypoid gears, axes do not intersect, so we call
conventionally, as a crossing point, Oc , the apparent point of intersection of these
axes, when they are projected on a plane parallel to both. Given the high flexibility
allowed by the current hypoid gear cutting machines, apices of face (or tip), pitch,
and root cones of a member of the hypoid gear pair may be beyond or before of the
crossing point, which is the trace of the centerline of mate on projection plane
above. Conventionally, the distances between these apices and the crossing point
are considered positive or negative, depending on whether the apices are beyond the
crossing point or before the crossing point. Therefore, according to this convention,
for the hypoid pinion, tzF1 (face apex beyond crossing point), tzR1 (root apex beyond
crossing point), and tz1 (pitch point beyond crossing point) are all positive, while
for hypoid wheel the corresponding quantities tzF2 , tzR2 , and tz2 are all negative
(these quantities are not shown in Fig. 12.32). Obviously, all these quantities would
be zero in the case in which the apices of both members of the hypoid gear pair
coincide with the crossing point.
(b)
(c)
12.11 Bevel and Hypoid Gear Geometry: Unified Discussion 645
depth is a constant along the face width (Figs. 9.11b and 12.33c), regardless of
cutter radius. The sum of the dedendum angles of pinion and wheel for uniform
depth, R#fU , is equal to zero. In this case, the cutter radius, rco , should be greater
than Rm2 sinbm2 , but less than 1:5Rm2 sinbm2 . In this way, the variation in normal
circular thickness along the face width is a minimum for both the pinion and
wheel. In the case in which a narrow inner topland occurs on the pinion, a small
tooth tip chamfer may be provided; the length measured along the face width,
and the angle between the chamfer generatrix and pinion axis define its
geometry (see Fig. 12.34).
– Space width taper and thickness taper, i.e. respectively the change in the space
width and tooth thickness along the face width, both measured generally in the
pitch plane.
– Slot width taper, i.e. the change in the point width identified by a V-shaped
cutting tool with nominal pressure angle, whose top is tangent to the root cone
and whose sides are tangent to the two sides of the tooth space, along the face
width. This taper is of primary consideration for production, since the width of
the slot at its narrowest point defines the point width of the cutting tool, and
limits the edge radius of the cutter blade. It depends on the dedendum angle and
lengthwise curvature, and for straight bevel gears can be changed by varying the
depth taper, i.e. by tilting the root line (Fig. 9.11c). This rotation is generally
carried out about the mid-section at the pitch line, in order to maintain the
desired working depth at the tooth mean section. For spiral bevel and hypoid
gears, the amount of the root line tilting is further dependent on a number of
geometric characteristics including the cutter radius. The wheel and pinion root
line can be rotated about the mean point, with a resulting dedendum angle
modification or tilting, normally ranging between 5 and þ 5 (Fig. 9.11c);
this is done to avoid cutting interference with a hub or shoulder.
From the point of view of the tooth depth configuration, we can have bevel gear
wheels with constant slot width, and bevel gear wheels with modified slot width
(Fig. 12.33b). The gear wheels with constant slot width have zero width slot
taper on both members of the gear pair; therefore, this taper is related to a tilt of
the root line such that the slot width is constant while maintaining the proper
space width taper. Instead, the gear wheels with modified slot width have a slot
width taper characterized by a root line tilted about the mean point of an
intermediate amount between zero and the value related to the constant slot
width. In this case, the slot width of the wheel is constant along the face width,
while the pinion does not have any slot width taper.
The formulas for calculating the sum of the dedendum angles, R#f (in degrees),
of the pinion and wheel, for the above-mentioned four possible cases of depthwise
taper that are chosen in the accordance with the cutting method, are as follows:
hfm1 hfm2
R#fS ¼ arctan þ arctan ð12:135Þ
Rm2 Rm2
R#fU ¼ 0 ð12:136Þ
90met Rm2 sinbm2
R#fC ¼ 1 ð12:137Þ
Re2 tanan cos bm rc0
which apply respectively to standard depth taper, uniform depth, constant slot width
and modified slot width. It is to be noted that subscripts S, U, C and M for quantities
appearing in Eqs. (12.135) to (12.138) refer respectively to standard depth taper,
uniform depth, constant slot width, and modified slot width depthwise taper.
Instead, the formulas that allow us to calculate the addendum and dedendum
angles of the wheel, #a2 and #f 2 (both in degrees), apportioning the sum of the
dedendum angles between pinion and wheel as a function of the desired depthwise
typer, are as follows:
hfm1
#a2 ¼ arctan ð12:139Þ
Rm2
#a2 ¼ #f 2 ¼ 0 ð12:141Þ
ham2
#a2 ¼ R#fC ð12:142Þ
hmw
ham2
#a2 ¼ R#fM ð12:144Þ
hmw
It is to be noted that: Eqs. (12.139) and (12.140) apply to standard depth taper;
Eq. (12.141) applies to uniform depth; Eqs. (12.142) and (12.143) apply to con-
stant slot depth; Eqs. (12.144) and (12.145) apply to modified slot width. It is also
to be noted that the quantities and related symbols in Eqs. (12.135) to (12.145) are
all known, but we feel it opportune to specify that the subscript, m, refers to the
mean cone quantities, and that the mean working depth, hmw , is the depth of
engagement of two members of the hypoid gear pair at mean cone distance.
For constant slot width, Eq. (12.137) shows that the sum of dedendum angles,
R#fC , is strongly influenced by the cutter radius, rc0 , as this has a significant effect on
the amount by which the root line is tilted. It is to be noted that, for a given design, a
large cutter radius and a small cutter radius increases and, respectively, decreases the
sum of dedendum angles. However, the cutter radius should be neither too large nor
too small: it must be within the limits ð1:1Rm2 sinbm2 rc0 Rm2 Þ. In fact, if the
cutter radius is too large (beyond the upper limit, i.e. rc0 [ Rm2 ), the resultant
depthwise taper could adversely influence the tooth depth at both ends, with a too
shallow tooth at inner end for a proper tooth contact, and a too depth tooth at the
outer end that may cause undercut and narrow toplands. Then, if the cutter radius is
equal to Rm2 sinbm2; R#fC becomes zero. Therefore, we would get uniform depth
teeth. If rc0 is less than Rm2 sinbm2 , we would get a reverse depthwise taper, and the
teeth would be deeper at the inner end than at the outer end. In order to not have teeth
with excessive depth at the inner end, with the risk of undercut and narrow toplands,
it is appropriate that rc0 is at least equal to the minimum value mentioned above. In
addition, it is to be noted that standard taper is the norm for gears cut with planer
tools; in this case, the cutter center is considered to be at infinity, and root lines are
not tilted.
For modified slot width, the sum of dedendum angles, R#fM , must not exceed
either 1.3 times the sum of the dedendum angles for standard depth taper, R#fS , nor
the sum of the dedendum angles for constant slot width taper, R#fC . In practice,
therefore, the smaller of the values 1.3 R#fS or R#fC is used (see Eq. 12.138).
Since the geometry of bevel and hypoid gears is a function of the cutting method
used, as well as the setting of the cutting machine that implements it, an infinite
number of pitch surfaces will exist for any hypoid gear pair. However, once the
initial data related to a given method have been defined, we will have one pitch
surface for each method. If we limit our attention to the cutting method of spiral
bevel gears, and the three cutting methods of hypoid gears (Gleason, Oerlikon, and
Klingelnberg), we can identify the following four design procedures, referred as
Method 0, Method 1, Method 2, and Method 3 by ISO 23509:2016 (E) [20]. The
fields of use of these four methods are as follows:
– Method 0 is used for spiral bevel gears, i.e. non-hypoid gears (curved-toothed
bevel gears, without hypoid offset).
– Method 1 is used for hypoid gears manufactured by the face-milling process.
The pitch surfaces of these gears are chosen in such a way that the hypoid radius
648 12 Spiral Bevel Gears and Hypoid Gears
(a) (b)
Fig. 12.35 Geometry of the main cutting processes of spiral bevel gears: a face-milling process;
b face-hobbing process
of curvature matches the radius of curvature of the cutter at the mean point for
the to-be-generated gear wheels (so the mean radius of curvature of the tooth is
equal to the cutter radius, i.e. qmb ¼ rc0 as Fig. 12.35a shows).
– Method 2 is used for hypoid gears manufactured by the face-hobbing process.
The pitch surfaces of these gears are chosen in such a way that the hypoid radius
of curvature matches the mean epicycloid curvature at the mean point, with the
condition that the wheel pitch apex, pinion pitch apex and cutter center lie on a
straight line.
– Method 3 is also used for hypoid gears manufactured by the face-hobbing
process. The pitch surfaces of these gears are chosen in such a way that the
hypoid radius of curvature matches the mean epicycloid curvature at the mean
point for gear wheel to be generated, without the condition described for Method
2 (Fig. 12.35b).
For spiral bevel gears, Method 0 has to be used, and face width factor is set to be
cbe2 ¼ 0:5; for initial data other than those shown in Table 12.2, the formulas can
be easily converted. With Method 1, which is used by Gleason, it is necessary to
determine the face width factor cbe2 , since the calculation point is not in the middle
of the wheel face width. The pitch cone parameters obtained with this method have
similar values compared to those obtained using Method 3, which is the method
used by Klingelnberg. Method 2 is the method used by Oerlikon.
The quantities shown in Fig. 12.35, and related symbols are as follows: mean
cone distance, Rm2 ; spiral angle, bm2 ; intermediate angle, u2 ; crown gear to cutter
center, qP0 ; cutter radius, rc0 ; lengthwise tooth mean radius of curvature, qmb ; first
auxiliary angle, k; second auxiliary angle, g1 ; lead angle of the cutter, m; epicycloid
base circle radius, qb .
To define in more detail the geometry of the hypoid gears, which represent the
most general case of gears, it is necessary to consider not only the three main views,
12.11 Bevel and Hypoid Gear Geometry: Unified Discussion 649
Table 12.2 Initial data for calculation of the pitch cone parameters
Symbol Description Method Method Method Method
0 1 2 3
R Shaft angle x x x x
a Hypoid offset 0 x x x
z1;2 Number of teeth x x x x
dm2 Mean pitch diameter of wheel – – x –
de2 Outer pitch diameter of wheel x x – x
b2 Wheel face width x x x x
bm1 Mean spiral angle of pinion – x – –
bm2 Mean spiral angle of wheel x – x x
rc0 Cutter radius x x x x
z0 Number of blade groups (only x x x x
face-hobbing)
but also three appropriate sections, as shown in Fig. 12.36, where: O1 and O2 are
the pinion pitch apex and respectively the wheel pitch apex; Oc is the crossing
point, i.e. the apparent point of intersection of axes, when it is projected on a plane
parallel to both axes (therefore, it is also the trace of the common normal to pinion
and wheel axes through crossing point); P is the mean point; M and N are the
intersection points of contact normal through mean point at pinion axis and
respectively at wheel axis; F and G are the crossing points at pinion axis and
respectively at wheel axis; l is the distance along wheel axis between crossing point
and intersection of contact normal.
The other quantities shown in Fig. 12.36, and related symbols are as follows:
hypoid offset, a; pinion pitch angle and wheel pitch angle, d1 and, respectively, d2 ;
pinion spiral angle and wheel spiral angle, bm1 and, respectively, bm2 ; pinion offset
angle in axial plane and wheel offset angle in axial plane, fm and, respectively, g;
pinion mean cone distance and wheel mean cone distance, Rm1 and, respectively,
Rm2 ; pinion offset angle in pitch plane, fmp ; pinion pitch apex beyond crossing point
and wheel pitch apex beyond crossing point, tz1 and, respectively, tz2 ; crossing point
to mean point along pinion axis and crossing point to mean point along wheel axis,
tzm1 and, respectively, tzm2 ; tangent to tooth trace at mean point, t.
Calculations of bevel and hypoid gears are divided into the two main steps
described in the following two subsections. The calculation development basically
follows that from ISO 23509:2016 (E) [20], which uses relationships obtainable
according to the theoretical treatment described in the previous sections. However,
we have tried to eliminate repeated equations, while retaining some repetitions so as
not overload the logic flow with references to previous calculation steps.
650 12 Spiral Bevel Gears and Hypoid Gears
Fig. 12.36 Main angles and quantities of a hypoid gear pair with R ¼ 90 : a side view looking
along the pinion axis; b side view looking along the wheel axis; c top view showing the shaft
angle; d view of the wheel section along the plane making the offset angle, fm , in the pinion axial
plane; e view of the pitch plane of the hypoid gear pair; f view of the pinion section along the plane
making the offset angle, g, in the wheel axial plane
In this first step of calculation, the pitch cone parameters are determined from the
initial data. This determination is carried out using a specific set of formulas for
each of the four methods described above. For spiral bevel gears (Method 0), it is
possible a simple determination of the pitch cone parameters. For hypoid gears, this
is not possible; in this case, a procedure of successive approximation or iteration
must be used. To start the calculation procedure related to each of the four methods,
it is necessary to have a series of initial data, which are summarized in Table 12.2.
However, it is to be noted that the pitch cone parameters of spiral bevel gears also
can be determined with different initial data as given in Table 12.2 (see Sect. 12.5).
The shaft angle, R, and in most cases also hypoid offset, a, are imposed by the
practical application involved. A positive pinion hypoid offset (Figs. 12.30 and
12.31) is recommended because of the increasing diameter of the pinion, higher
pitting load carrying capacity, and higher face contact ratio. It is however necessary
to check the scuffing load capacity due to additional lengthwise sliding. For this
reason, the pinion hypoid offset should not exceed 0:25de2 and, for heavy-duty
applications, it should be limited to half of the above value.
12.11 Bevel and Hypoid Gear Geometry: Unified Discussion 651
The calculation of the outer pitch diameter of wheel, de2 , involves the prelimi-
nary determination of the outer pitch diameter of pinion, de1 , which is correlated to
the pinion torque, T1 (in Nm), given by:
P 60P
T1 ¼ ¼ ; ð12:146Þ
x1 2pn1
where P is the power (in W), x1 is the pinion angular velocity (in rad/s), and n1 is
the pinion rotational speed (in min1 ; revolutions per minute). When the load is not
constant, the pinion torque will vary, and its operating value must be calculated
considering the power and speed values at which the expected operating cycle of
the driven machine is carried out. When peak loads are present, if their total
duration exceeds ten million cycles during the total expected life of the gear, the
(a)
(b)
Fig. 12.37 Outer pitch diameter of pinion, d e1 , as a function of pinion torque, T 1 , for: a pitting
strength; b bending strength
652 12 Spiral Bevel Gears and Hypoid Gears
determination of the gear size should be made according to peak load. If, however,
their total duration is less than ten million cycles, the evaluation of the gear size
should be started based on the greater of the two values corresponding to the
highest sustained load or half of the peak load.
For spiral bevel gears with R ¼ 90 , the charts shown in Fig. 12.37 enable us to
determine the outer pitch diameter of commercial quality spiral bevel pinions of
case-hardened steel, at 55 minimum HRC, as a function of the pinion torque,
respectively for pitting and bending strengths. For shaft angles R 6¼ 90 , the charts
give preliminary values of de1 less accurate, and this could require additional
adjustments of design choices. For straight bevel and Zerol bevel gears, the values
of de1 obtained by these charts are to be multiplied respectively by 1.2 and 1.3 (for
Zerol bevel gears, this is due to a face width limitation). For hypoid gears, the
values of de1 obtained by the same charts are to be considered as equivalent pinion
outer pitch diameters.
In the more general case of hypoid gears, to calculate the outer pitch diameter of
the hypoid wheel, de2 , it is necessary to determine first a preliminary hypoid pinion
pitch diameter, deplm1 , given by:
a
deplm1 ¼ de1 ; ð12:147Þ
u
where de1 (in mm) is the greater of the two values of the pinion outer pitch diameter
obtained from the charts shown in Fig. 12.37, a (in mm) is the hypoid offset, and
u ¼ z2 =z1 is the gear ratio. The actual outer pitch diameter of the hypoid wheel, de2 ,
is then calculated as:
where Reint2 and dint2 are respectively the intermediate wheel outer cone distance
and intermediate wheel pitch angle, whose approximate values are given by:
deplm1
Reint2 ¼ ; ð12:149Þ
2sindint1
The intermediate pinion pitch angle, appearing in equations above is given by:
sinR
dint1 ¼ arctan : ð12:151Þ
u þ cosR
from the charts shown in Fig. 12.37a, and the value of deplm1 calculated by
Eq. (12.147) are to be multiplied by 0.8. Furthermore, for materials other than
case-hardened steel at 55 minimum HRC, the values of de1 obtained from the charts
shown in Fig. 12.37, and the values of deplm1 determined by Eq. (12.147) are to be
multiplied by a material factor, kM , given in Table 12.3.
For statically loaded gears, only the bending strength is considered. In fact, the
load conditions related to bending strength are more restrictive than those related to
the pitting resistance. Depending on whether or not the gears are subjected to
vibration, the values of the outer pitch diameter of pinion taken from the charts
shown in Fig. 12.37b or calculated by Eqs. (12.147) are multiplied by 0.7 or
respectively by 0.6.
Theoretically, the choice of the numbers of teeth z1 and z2 can be made in an
arbitrary manner, as long as the assigned gear ratio u ¼ z2 =z1 is respected. In
reality, however, for general applications, it is appropriate that the number of pinion
teeth, z1 , is made using the two charts shown in Figs. 12.38a and 12.38b, respec-
tively valid for spiral bevel and hypoid gears, and for straight bevel and Zerol bevel
gears; these two charts provide approximate but reliable values of the number of
teeth of pinion, as a function of dei and u. For automotive applications, the pinion of
spiral bevel gears and hypoid gears often has a fewer number of teeth, as shown in
Table 12.4. In order to achieve an acceptable contact ratio without undercut,
straight bevel gears and Zerol bevel gears are respectively designed with 12 teeth
and higher, and with 13 teeth and higher.
Spiral bevel gears and hypoid gears can have a fewer numbers of teeth (see
Table 12.6), due to the additional overlap ratio resulting from oblique teeth, which
allows the teeth to be stubbed to avoid undercut and still maintain an acceptable
contact ratio. A careful analysis of undercut must however be done.
654 12 Spiral Bevel Gears and Hypoid Gears
The wheel face width, b2 (in mm), of spiral bevel gears with R ¼ 90 can be
determined as a function of the outer pitch diameter of pinion, de1 (in mm), and gear
ratio, u, as Fig. 12.39 shows. This figure refers to face widths corresponding to 30%
of the outer cone distance. For shaft angle less than or greater than 90°, respectively
a face width larger or smaller than that given in Fig. 12.39 can be used. Generally,
the face width should not exceed 30% of the outer cone distance or 10met2 , where
met2 is the outer transverse module of the wheel. The face width of Zerol bevel
gears should not exceed 25% of the outer cone distance, and should be determined
by multiplying by 0:83 the value read in the chart of Fig. 12.39. The face width of
hypoid gear wheels is also determined with the chart in this figure; the hypoid
pinion face width is generally greater than that of hypoid wheel (see
Sect. 12.12.2.5).
The outer transverse module of the wheel is obtained by dividing the outer pitch
diameter by the number of teeth, i.e. met2 ¼ de2 =z2 . It is not necessary that this
module be an integer, since the cutting tools for bevel gears are not standardized
according to the module. The determination of the mean pitch diameter of wheel,
dm2 , is carried out using the above defined quantities and those that follows.
The mean spiral angles of pinion and wheel, bm1 and bm2 , are chosen so as to
achieve a face contact ratio, eb , approximately equal to 2:0. For maximum
smoothness and quietness and high-speed applications, it is appropriate to have
eb [ 2, although values of eb less than 2:0 are allowed. The chart shown in
Fig. 12.40 is a guideline for choosing the spiral angles of spiral bevel gears with
face width equal to 30% of the outer cone distance, i.e. b=Re ¼ 0:3. For any value
of b=Re , the face contact ratio, eb , is given by:
Kz3 3 Re
eb ¼ Kz tanbm tan bm ; ð12:152Þ
3 pmet
12.11 Bevel and Hypoid Gear Geometry: Unified Discussion 655
(a)
(b)
Fig. 12.38 Approximate number of pinion teeth, for: a spiral bevel and hypoid gears; b straight
bevel and Zerol bevel gears
where bm is the mean spiral angle at pitch surface, quantities Re , met , and b are
given in mm, and Kz is a dimensionless factor depending on the ratio b=Re ,
given by:
656 12 Spiral Bevel Gears and Hypoid Gears
Fig. 12.39 Face width of spiral bevel gears with R ¼ 90 , as a function of the outer pitch
diameter of pinion, and gear ratio
Fig. 12.40 Face contact ratio for spiral bevel gears as a function of spiral angle and ratio ðb=met Þ
12.11 Bevel and Hypoid Gear Geometry: Unified Discussion 657
b 2 b
Re
Kz ¼ : ð12:153Þ
2Re 1 b
Re
For the use of chart shown in Fig. 12.40, where b=Re ¼ 0:3, we have
eb ¼ ½ð0:3885tanbm 0:0171tan3 bm Þðb=met Þ.
The mean spiral angle of the pinion of a hypoid gear pair is determined by the
relationship:
pffiffiffi a
bm1 ¼ 25 þ 5 u þ 90 ; ð12:154Þ
de2
with a and de2 in mm. The mean spiral angle bm2 of the wheel depends on the
hypoid geometry, and is calculated as shown in Sect. 12.12.1.
Design and manufacture of bevel and hypoid gears also depend, for face-milled
gears, on the cutter radius, rc0 , and number of blade groups, z0 , and, for face-hobbed
gears, only on the cutter radius. For both cutting processes, Table 12.5 provides a
data list of standard cutters. For the face-milling process, the cutter diameter 2rc0 is
given in inches, for 2rc0 \500 mm, and in millimeters, for 2rc0 500 mm.
By using the above-described initial data as well as the appropriate equations in
the next section, we can determine the pitch cone parameters Rm1 , Rm2 , d1 , d2 , bm1 ,
bm2 , and cbe2 , by which a schematic diagram of a spiral bevel or hypoid gear can be
drawn, like the one shown in Fig. 12.41.
The parameter cbe2 ¼ ðRe2 Rm2 Þ=b2 , i.e. the face width factor, must neces-
sarily be considered for Method 1, as in this case the calculation point does not
always coincide with the mean point P of the wheel face width. For Methods 0, 2
and 3, the calculation point coincides with the mean point, for which cbe2 ¼ 0:5.
The schematic diagram of Fig. 12.41 shows a common tangential plane, T,
between both pitch cones, with mean pitch diameters dm1 and dm2 , which are in
contact each other at the mean point, P. Besides, both pitch cones contact with the
tangential plane, T, along two straight lines, which are the mean cone distances, Rm1
and Rm2 , and include the offset angle, fmp . The normal straight line to the plane, T,
through the mean point, P, intersects the pinion axis xP at point NP , and the gear
wheel axis xG at point NG . The straight line NP NG represents the center distance,
av , of the equivalent mean virtual cylindrical gear.
In doing so, we applied to hypoid gears the concepts we have described in
Chap. 9, concerning straight bevel gears. However, contrary to what happens in the
last gears, the pinion and wheel axes of a hypoid gear pair are not in the same plane.
Therefore, to have virtual cylindrical gears with parallel axes, an approximation is
made according to which both axes are arranged in the direction that divides the
offset angle fmp into half. This does not mean that the thus-defined virtual cylindrical
gears have the same meshing conditions such as hypoid gears. This goal is achieved
afterwards, by introducing several appropriate correction factors such as the hypoid
factor, ZHyP , which takes into account the influence of the lengthwise sliding of
658 12 Spiral Bevel Gears and Hypoid Gears
hypoid gear teeth. Nevertheless, virtual cylindrical gears supply the required geo-
metrical basis to achieve a practicable rating system for all types of bevel gears.
In this second step of calculation, the gear dimensions are determined, starting from
the pitch cone parameters, and introducing the set of additional data presented in
Table 12.6. These additional data can be given in terms of European standards (data
type I) or in terms of AGMA standards (data type II). For example, European
standards describe the gear tooth proportions with an addendum factor, khap , a
dedendum factor, khfp , a profile shift coefficient, xhm , and a thickness modification
coefficient, xsmn , while AGMA standards describe the same tooth proportions with a
depth factor, kd , a clearance factor, kc , a mean addendum factor of wheel, cham , and
a thickness factor, kt or wheel mean slot width, Wm2 . European factors and AGMA
factors are related to each other (see Table 12.7), and both lead to the same result of
tooth geometry. This geometry, contrary to pitch cone parameters, is expressed by
only one set of formulas for bevel and hypoid gears, no matter which method was
chosen. All formulas for hypoid gears, with the hypoid offset, a, set to zero, also
apply to spiral bevel gears. Figure 12.42 highlights some of the above-mentioned
quantities, with reference to the basic rack tooth profile of the wheel.
As Table 12.6 shows, some additional data for calculation of gear dimensions
are common to European and AGMA standards. Other additional data are rather
different. In this regard, in this textbook, we refer to data type I, and we leave to the
reader the ability to turn them into data type II, according to the correlations
reported in Table 10.7. In any case, whatever the standards, European or AGMA,
the choice of these additional data should be done carefully, to ensure optimal
performance of the gear to be designed.
With regard to the normal pressure angles, it is to distinguish between:
• Nominal design pressure angle, ad (adD on drive side, and adC on coast side),
which is the start value for the calculation, and may be half of the sum of
pressure angles or different on drive and coast sides.
• Generated pressure angle, an (anD on drive side, and anC on coast side), which is
the pressure angle of the generation crown wheel, and characterizes the tooth
flank in the mean normal section.
• Effective pressure angle, ae (aeD on drive side, and aeC on coast side), which is a
calculated value as reported below.
Generally, the nominal design pressure angles on drive side, adD , and on coast
side, adC , are balanced, but in some optimized applications these angles are
unbalanced. For bevel gears, the most commonly used design pressure angle is
ad ¼ 20 . This pressure angle greatly influences the gear design. In fact, lower
generated pressure angles reduce the axial and separating forces, and increase the
transverse contact ratio as well as toplands and slot widths, while higher generated
pressure angles cause adverse effects. In addition, lower effective pressure angles
increase the risk of undercut.
For hypoid gears, in many practical applications, it would be appropriate to have
unequal generated pressure angles on the drive and coast sides, to obtain mesh
balanced conditions. In the case in which full balanced mesh conditions are
required, the effective pressure angle on drive side, aeD , has a different value
compared to that of the effective pressure angle on the coast side, aeC , and the
influence factor of limit pressure angle, falim , is set equal to the unit. Thus, the
generated normal pressure angles on the drive side, anD , and on the coast side, anC ,
are obtained from the corresponding design pressure angles, adD and adC , by adding
and respectively subtracting the limit pressure angle, alim , which is given by:
tand1 tand2 Rm1 sinbm1 Rm2 sinbm2
alim ¼ arctan : ð12:155Þ
cosfmp Rm1 tand1 þ Rm2 tand2
In the general case of balanced mesh conditions, the generated normal pressure
angles on drive side, anD , and on coast side, anC , are given respectively by the
following relationships:
while the effective pressure angles on drive side, aeD , and on coast side, aeC , are
given respectively by:
12.11 Bevel and Hypoid Gear Geometry: Unified Discussion 661
Reducing the generated pressure angles on drive side, several benefits can be
obtained in terms of contact ratio, contact stress, and axial and radial forces.
However, we must not go down to values lower than ð9 10Þ , due to inherent
limitations of the cutting tools, as well as to the risk of undercut. In all cases,
however, the effective pressure angles are calculated with Eqs. (12.158) and
(12.159).
For not-hypoid gears, i.e. straight, Zerol, and spiral bevel gears, the limit pres-
sure angle is always equal to zero, for which the nominal design pressure angles and
generated pressure angles have the same values. If then the effective pressure angles
have the same values, the mesh conditions on drive side and coast side are equal.
The following guidelines are to be observed for the choice of the nominal
pressure angle, in order to prevent undercut:
– For straight bevel gears, ad ¼ 20 or higher for pinions with 14 to 16 teeth, and
ad ¼ 25 for pinions with 12 or 13 teeth.
– For Zerol bevel gears with high transmission ratios or low tooth numbers, or
both, ad ¼ 22:5 for pinions with 14 to 16 teeth, and ad ¼ 25 for pinions with
13 teeth.
– For spiral bevel gears, ad ¼ 20 or higher for pinions with 12 or fewer teeth.
– For hypoid gears, ad ¼ 18 or ad ¼ 20 for light-duty drives, and ad ¼ 22:5 or
ad ¼ 25 for heavy-duty drives (it is to be noted that, to balance the mesh
conditions on drive and coast sides, it should be falim ¼ 1, but for use of standard
cutting tools, the value of falim can be different from unity).
As regard the tooth depth components, it should be noted that, in terms of
European standards, in common cases, the basic crown wheel addendum factor,
khap , and basic crown wheel dedendum factor, khfp , are chosen respectively equal to
1.00 and 1.25. To prevent then undercut, the profile shift coefficients have to be
included within specified ranges (see Sect. 12.12.3). AGMA standards, however,
provide more detailed data type II, which can be used to calculate the corresponding
data type I, with the help of equations collected in Table 12.7.
khap ¼ kd =2 kd ¼ 2khap
khfp ¼ kd kc þ 12 khfp
kc ¼ 12 khap 1
h i
xsmn ¼ k2t ¼ 12 W kt ¼ 2xsmn
mmn þ kd kc þ 2 ðtananD þ tananC Þ 2
m2 1 p
662 12 Spiral Bevel Gears and Hypoid Gears
Fig. 12.42 Basic rack tooth profile of wheel, and wheel tooth profile with profile shift and
thickness modification
Table 12.8 Suggested Type of gear Depth factor Number of pinion teeth
values of depth factor, kd
Straight bevel 2.000 12 or more
Spiral bevel 2.000 12 or more
1.995 11
1.975 10
1.940 9
1.895 8
1.835 7
1.765 6
Zerol bevel 2.000 13 or more
Hypoid 2.000 11 or more
1.950 10
1.900 9
1.850 8
1.800 7
1.750 6
For constant clearance along the whole tooth depth, the calculation is made at
mean point, and a clearance factor kc ¼ 0:125 is used, but also this value can be
changed to obtain other specified design requirements. For fine pitch gearing
(met ¼ 1:27 and finer), and for teeth to be finished in a secondary machining
operation, the corresponding values of the clearance must be increased by
0.051 mm.
The mean addendum factor, cham , apportions the mean working depth, hmw ,
between the pinion and wheel addendum. To prevent undercut, the pinion adden-
dum is usually greater than the wheel addendum, with the exception of the case in
which the numbers of teeth are equal. Table 12.9 gives a guideline to choose the
mean addendum factor, cham , for R ¼ 90 , as a function of type of gear, number of
pinion teeth, and equivalent ratio, ua , defined as:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
cosd1 tand2 cosg
ua ¼ : ð12:160Þ
cosd2
where
Table 12.9 Mean addendum factor, cham , for shaft angle R ¼ 90
Type of gear Mean addendum factor Number of pinion teeth
Straight bevel 0:210 þ 0:290=u2a 12 or more
Zerol bevel 0:210 þ 0:290=u2a 13 or more
Spiral bevel and hypoid 0:210 þ 0:290=u2a 12 or more
0:210 þ 0:280=u2a 11
0:175 þ 0:260=u2a 10
0:145 þ 0:235=u2a 9
0:130 þ 0:195=u2a 8
0:110 þ 0:160=u2a 7
0:100 þ 0:115=u2a 6
664 12 Spiral Bevel Gears and Hypoid Gears
and wheel, the curves shown in Fig. 12.43 may be used. These curves have been
obtained with the following relationship:
that the backlash allowance is proportional to the outer transverse module met (in
mm), and depends on the ISO accuracy grades, which are divided into two ranges
of values.
In the previous section, we said that the calculation of spiral bevel and hypoid gears
is divided in two main steps, which respectively relate to the determination of the
pitch cone parameters, and gear dimensions. Figure 12.44 shows the flow-chart of
the aforementioned calculation structure, which deserves to be described here in
more detail.
Both the flow chart in Fig. 12.44 and Table 12.2 show the initial data needed to
make the first step of calculation, the one concerning the determination of the pitch
cone parameters. For all four methods (Method 0, Method 1, Method 2 and Method
3), these initial data include the numbers of teeth of pinion and wheel, z1 and z2 , and
therefore the gear ratio u ¼ z2 =z1 . However, the calculation procedure is different
Fig. 12.44 Flow-chart of calculation structure of spiral bevel and hypoid gears
666 12 Spiral Bevel Gears and Hypoid Gears
for each of these four methods. It is very simple for spiral bevel gears (Method 0),
while it is an iterative procedure for the other three methods, respectively used by
Gleason (Method 1), Oerlikon (Method 2), and Klingelnberg (Method 3).
Therefore, the four methods here are differentiated from each other.
12.12.1.1 Method 0
The pinion pitch angle, d1 , and wheel pitch angle, d2 , are given by:
d1 ¼ arctan sinR
cosR þ u d2 ¼ R d1 : ð12:163Þ
The outer cone distance, Rei , and mean cone distance, Rmi , are given by:
Rei ¼ 2sind
dei
i
Rmi ¼ Rei b2i ; ð12:164Þ
with i ¼ ð1:2Þ.
Finally, the mean spiral angles of pinion and wheel are the same ðbm1 ¼ bm2 Þ,
and the face width factor of the wheel is assumed equal to 0.5 ðcbe2 ¼ 0:5Þ.
12.12.1.2 Method 1
Using the initial data, and according to the shaft angle departure from 90°, given by
DR ¼ ðR 90 Þ, and the desired pinion spiral angle, bD1 ¼ bm1 , we calculate first
the following preliminary quantities: approximate wheel pitch angle, dint2 ; wheel
0
mean pitch radius, rmpt2 ; approximate pinion offset angle in pitch plane, e1 ;
approximate hypoid dimension factor, K1 ; and approximate pinion mean radius,
rmn1 . These five quantities are given respectively by the following relationships:
ucosDR
dint2 ¼ arctan ð12:165Þ
1:2ð1 usinDRÞ
de2 b2 sindint2
rmpt2 ¼ ð12:166Þ
2
0 asindint2
e1 ¼ arcsin ð12:167Þ
rmpt2
rmpt2 K1
rmn1 ¼ : ð12:169Þ
u
12.12 Calculation of Spiral Bevel and Hypoid Gears 667
Now we start the iterative calculation procedure, and for the purpose we intro-
duce, in succession, the quantities defined below, some of which are common to
gears obtained with face-hobbing process and with face-milling process, while
others are specific of each of these two cutting processes. The quantities common to
the two cutting processes are those described below:
– Wheel offset angle in axial plane, g, with which the iterative calculation pro-
cedure begins; it is given by:
a
g ¼ arctan : ð12:170Þ
rmpt2 ðtandint2 cosDR sinDRÞ þ rmn1
a rmn1 sing
e2 ¼ arcsin ð12:171Þ
rmpt2
sing
dint1 ¼ arctan þ tanDRcosg ð12:172Þ
tane2 cosDR
0 sine2 cosDR
e2 ¼ arcsin ð12:173Þ
cosdint1
K1 cose02
bmint1 ¼ arctan : ð12:174Þ
sine02
sing
d1 ¼ arctan þ tanDRcosg : ð12:178Þ
tane1 cosDR
– Pinion offset angle in pitch plane, e01 , pinion spiral angle, bm1 , wheel spiral
angle, bm2 , and wheel pitch angle, d2 , which are given respectively by the
following relationships:
sine1 cosDR
e01
¼ arcsin ð12:179Þ
cosd1
K1 þ DK cose01
bm1 ¼ arctan : ð12:180Þ
sine01
– Mean cone distance of pinion, Rm1 , mean cone distance of wheel, Rm2 , and
pinion mean pitch radius, rmpt1 , which are given respectively by the following
relationships:
rmn1 þ Drmpt1
Rm1 ¼ ð12:183Þ
sind1
rmpt2
Rm2 ¼ ð12:184Þ
sind2
– Limit pressure angle, alim , and limit radius of curvature, qlim , which are given
respectively by:
tand1 tand2 Rm1 sinbm1 Rm2 sinbm2
alim ¼ arctan ð12:186Þ
cose01 Rm1 tand1 þ Rm2 tand2
The quantities specific of the face-hobbed gears are the following: number of
crown gear teeth, zp , lead angle of cutter, m, first auxiliary angle, k, crown gear to
cutter center distance, qP0 , second auxiliary angle, g1 , and lengthwise tooth mean
radius of curvature, qmb . These quantities are given respectively by the following
relationships:
z2
zp ¼ ð12:188Þ
sind2
Rm2 z0
m ¼ arcsin cosbm2 ð12:189Þ
rc0 zp
where rc0 and z0 are respectively the cutter radius and the number of blade groups.
The specific quantity of face-milled gears is only the lengthwise tooth mean
radius of curvature, qmb , which is given by:
The iterative procedure involves several calculation steps, the first of which
involves the determination of all the above-mentioned quantities. With the second
step, the initial value of the offset wheel angle in axial plane, g, given by
Eq. (12.170), is changed, and consequently all the quantities given by
Eqs. (12.171) to (12.187) are recalculated, and so on, until the following condition
is satisfied:
qmb
1 0:01: ð12:195Þ
qlim
At this point, the iterative procedure is stopped, and the face width factor, cbe2 , is
calculated as follows:
1 de2
cbe2 ¼ Rm2 : ð12:196Þ
b2 2sind2
670 12 Spiral Bevel Gears and Hypoid Gears
12.12.1.3 Method 2
Using the initial data, we calculate first the following preliminary quantities:
– Lead angle of cutter, m, first auxiliary angle, k, first approximate pinion pitch
angle, d1app , first approximate wheel pitch angle, d2app , and first approximate
pinion offset angle in axial plane, fmapp , which are given respectively by the
following relationships:
z0 dm2 cosbm2
m ¼ arcsin ð12:197Þ
2rc0 z2
cosbm2
Fapp ¼ ð12:202Þ
cos bm2 þ fmapp
dm1app ¼ Fapp dm2 =u ð12:203Þ
" #
ucosfmapp
u2 ¼ arctan ð12:204Þ
u
tand2app þ Fapp 1 sinR
common pitch plane (see Fig. 12.36), which are given respectively by the fol-
lowing relationships:
tanðbm2 þ g1 Þ
u3 ¼ arctan ð12:207Þ
sinu2
2 3
dm1app sinR
d001 ¼ arctan4 5 ð12:208Þ
dm2 cosfmapp þ dm1app cosR tan 2a
ðu3 þ fmapp Þ
Now we start the iterative calculation procedure, and for this purpose we
introduce, in succession, the quantities defined as follows:
– Improved wheel pitch angle, d2imp (the iterative procedure begins with this first
quantity), auxiliary angle, gp , approximate wheel offset angle, gapp , improved
pinion offset angle in axial plane, fmimp , and improved pinion offset angle in
pitch plane, fmpimp , which are given respectively by the following relationships:
d2imp ¼ arctan tand002 cosfmapp ð12:210Þ
" #
sinfmapp cosd2imp
gp ¼ arctan ð12:211Þ
cos R d2imp
2 3
6 2a 7
gapp ¼ arctan4 cosgp sinðbm2 þ g1 Þ
5 ð12:212Þ
dm2 tand2imp þ dm1app cos Rd
ð 2imp Þ
" #
2a Fapp tangapp sind2imp cosgp
fmimp ¼ arcsin ð12:213Þ
dm2 ucos R d2imp
" #
tanfmimp sinR
fmpimp ¼ arctan : ð12:214Þ
cos R d2imp
cosbm2
F¼ ð12:215Þ
cos bm2 þ fmpimp
– Improved auxiliary angle, gpimp , wheel offset angle in axial plane, g, pinion
offset angle in axial plane, fm , pinion offset angle in pitch plane, fmp , pinion
spiral angle, bm1 , pinion mean pitch diameter, dm1 , and auxiliary angle, n (this
quantity is expressed by two different equations, depending on whether R 6¼ 90
or R ¼ 90 ), which are given respectively by the following relationships: the
two Eqs. (12.228) are respectively valid for R 6¼ 90 and R ¼ 90 .
tangapp sinu5
gpimp ¼ arctan ð12:222Þ
cosðR u5 Þ
2 3
2a
g ¼ arctan4 cosg sinu5
5 ð12:223Þ
dm2 tand2 þ dm1 cosðpimp
Ru Þ 5
dm2 cosbm2
dm1 ¼ ð12:227Þ
ucosbm1
– Pinion pitch angle, d1 , mean cone distance of pinion, Rm1 , mean cone distance of
wheel, Rm2 , crown gear to cutter center distance, qP0 , intermediate angle, u6 ,
complementary angle, ucomp , and checking variable, Rmcheck , which are given
respectively by the following relationships:
d1 ¼ arctan tanncosfmp ð12:229Þ
Rm2 sinu6
Rmcheck ¼ : ð12:235Þ
sinucomp
The iterative procedure involves also here several calculation steps, the first of
which involves the determination of all the quantities above. With the second step,
the initial value of the improved wheel pitch angle, d2imp , given by Eq. (12.210), is
changed, and consequently all the quantities given by Eqs. (12.211) to (12.235) are
recalculate, and so on, until the following condition is satisfied (note that d2imp must
be increased if Rm1 \Rmcheck , and vice versa):
Rm1
1
0:01: ð12:236Þ
R
mcheck
At this point, the iterative procedure is stopped, and the face width factor cbe2 is
taken equal to 0.5 (cbe2 ¼ 0:5).
674 12 Spiral Bevel Gears and Hypoid Gears
12.12.1.4 Method 3
Using the initial data, we calculate first the wheel pitch angle, d2 , and pinion pitch
angle, d1 , which are to be considered as two preliminary quantities; they are given
respectively by the following relationships:
sinR
d2 ¼ arctan ð12:237Þ
ðF=uÞ þ cosR
d1 ¼ R d2 ; ð12:238Þ
– Hypoid dimension factor, F, pinion mean pitch diameter, dm1 , mean cone dis-
tance of pinion, Rm1 , mean cone distance of wheel, Rm2 , lead angle of cutter, m,
and auxiliary angle, #m , which are given respectively by the following
relationships:
12.12 Calculation of Spiral Bevel and Hypoid Gears 675
sinðbm2 þ #m mÞsinfmp
A7 ¼ cosbm1 cosðbm2 þ #m Þ ð12:255Þ
cosðbm2 mÞ
A3 A4
Rmint ¼ ð12:256Þ
A5 A6 þ A3 A7
dm1
d1 ¼ arcsin ð12:257Þ
2Rmint
sind1 cosfm sinR þ cosd1 cosfmp cosR
d2 ¼ arccos : ð12:258Þ
1 sin2 Rsin2 fm
Here also, the iterative procedure involves several calculation steps, the first of
which involves the determination of all quantities above. With the second step, the
initial value of the wheel mean pitch diameter, dm2 , given by Eq. (12.239), is
changed, and consequently all the quantities given by Eqs. (12.240) to (12.258) are
recalculated, and so on, until the following condition is satisfied:
676 12 Spiral Bevel Gears and Hypoid Gears
At this point, the iterative procedure is stopped, and the face width factor cbe2 is
taken equal to 0.5 ðcbe2 ¼ 0:5Þ.
Once the pitch cone parameters are known, for determination of the gear dimen-
sions, which constitutes the second main step of the calculation procedure of spiral
bevel and hypoid gears (see flow chart shown in Fig. 12.44), the set of additional
data summarized in Table 12.6 is required. Some of the gear dimensions to be
determined are common to the four calculation methods described in the two
previous sections, other instead change from method to method, and other, for a
given method, are different depending on whether the gears under consideration are
face-hobbed gears or face-milled gears. Whenever differences are found in this
regard, we will care to highlight them on time.
Using the pitch cone parameters and additional data, we calculate successively the
basic data, to be used at least partially in further calculation, as follows:
– Pinion mean pitch diameter, dm1 , wheel mean pitch diameter, dm2 and shaft
angle departure from 90°, DR, which are given respectively by the following
relationships:
DR ¼ R 90 : ð12:262Þ
– Offset angle in pinion axial plane, fm , and offset angle in pitch plane, fmp ; which
are given respectively by the following relationships:
2acosd1
fm ¼ arcsin ð12:263Þ
dm2 cosd1 þ dm1 cosd2
sinfm sin R
fmp ¼ arcsin : ð12:264Þ
cosd1
12.12 Calculation of Spiral Bevel and Hypoid Gears 677
– Offset in pitch plane, ap , mean normal module, mmn , and limit pressure angle,
alim , which are given respectively by the following relationships:
– Generated normal pressure angles on drive side, anD and on coast side, anC , and
effective pressure angles on drive side, aeD , and on coast side, aeC , which are
given respectively by the following relationships:
– Outer and inner pitch cone distances of wheel, Re2 and Ri2 , outer and inner pitch
diameters of wheel, de2 and di2 , and outer transverse module, met2 , which are
given respectively by the following relationships:
– Wheel face width from calculation point to outside, be2 , wheel face width from
calculation point to inside, bi2 , crossing point to calculation point along wheel
axis, tzm2 , crossing point to calculation point along pinion axis, tzm1 , and pitch
678 12 Spiral Bevel Gears and Hypoid Gears
apex beyond crossing point along axis, tz1:2 , which are given respectively by the
following relationships:
The quantities concerning the tooth depth are the mean working depth, hmw , mean
addendum of wheel, ham2 , mean dedendum of wheel, hfm2 , mean addendum of
pinion, ham1 , mean dedendum of pinion, hfm1 , clearance, c, and mean whole depth,
hm . These quantities are given respectively by the following relationships (see
Table 12.8 for other symbols):
The quantities concerning the root angles and face angles are, in ordered sequence,
as follows:
– Face and root angles of wheel, da2 and df 2 , auxiliary angle for calculating pinion
offset angle in root plane, uR and auxiliary angle for calculating pinion offset
angle in face plane, u0 , which are given respectively by the following
relationships:
df 2 ¼ d2 # f 2 ð12:290Þ
a tan DR cos df 2
uR ¼ arctan ð12:291Þ
Rm2 cos #f 2 tz2 cos df 2
aðtanDRcosda2 Þ
u0 ¼ arctan : ð12:292Þ
Rm2 cos#a2 tz2 cosda2
– Pinion offset angle in root plane, fR , pinion offset angle in face plane, f0 , face
angle of pinion, da1 , root angle of pinion, df 1 , addendum angle of pinion, #a1 ,
and dedendum angle of pinion, #f 1 , which are given respectively by the fol-
lowing relationships:
a cosuR sindf 2
fR ¼ arcsin uR ð12:293Þ
Rm2 cos#f 2 tz2 cosdf 2
aðcosu0 sinda2 Þ
f0 ¼ arcsin u0 ð12:294Þ
Rm2 cos#a2 tz2 cosda2
da1 ¼ arcsin sinDRsindf 2 þ cosDRcosdf 2 cosfR ð12:295Þ
# f 1 ¼ d1 df 1 : ð12:298Þ
680 12 Spiral Bevel Gears and Hypoid Gears
– Wheel face apex beyond crossing point along wheel axis, tzF2 , wheel root apex
beyond crossing point along wheel axis, tzR2 , pinion face apex beyond crossing
point along pinion axis, tzF1 , and pinion root apex beyond crossing point along
pinion axis,tzR1 , which are given respectively by the following relationships:
For determination of the pinion face width, b1 , we calculate first the pinion face
width in pitch plane, bp1 , and pinion face width from calculation point to front
crown, b1A , using the following relationships:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
bp1 ¼ R2e2 a2p R2i2 a2p ð12:303Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
b1A ¼ R2m2 a2p R2i2 a2p : ð12:304Þ
(a) (b)
Fig. 12.45 Pinion geometry: a inner and outer diameters, and face width; b tooth depth, and face
width
For Method 1, the quantities involved are the auxiliary angle, k0 , pinion face
width in pitch plane, breri1 (see Fig. 12.46), pinion face width increment along
pinion axis, Dbx1 , increment along pinion axis from calculation point to outside,
Dgxe , increment along pinion axis from calculation point to inside, Dgxi , pinion face
width from calculation point to outside, be1 , pinion face width from calculation
Fig. 12.46 Quantities in pitch plane for determination of pinion spiral angle
682 12 Spiral Bevel Gears and Hypoid Gears
point to inside, bi1 , and pinion face width along pitch cone, b1 , which are given
respectively by the following relationships:
sinfmp cosd2
k0 ¼ arctan ð12:306Þ
ucosd1 þ cosd2 cosfmp
b2 cosk0
breri1 ¼ ð12:307Þ
cos fmp k0
1
Dbx1 ¼ hmw sinfR 1 ð12:308Þ
u
cbe2 breri1
Dgxe ¼ cosda1 þ Dbx1 hfm2 c sind1 ð12:309Þ
cos#a1
ð1 cbe2 Þbreri1
Dgxi ¼ cosda1 þ Dbx1 þ hfm2 c sind1 ð12:310Þ
cos#a1
For Method 2, the quantities involved are the pinion face width along pitch cone,
b1 , pinion face width from calculation point to outside, be1 , and pinion face width
from calculation point to inside, bi1 , which are given respectively by the following
relationships:
b1 ¼ b2 1 þ tan2 fmp ð12:314Þ
For Method 3, the quantities involved are the pinion face width along pitch cone,
b1 , additional pinion face width, bx , pinion face width from calculation point to
inside, bi1 , and pinion face width from calculation point to outside, be1 , which are
given respectively by the following relationships:
b1 ¼ int bp1 þ 3mmn tan
fmp
þ 1 ð12:316Þ
bx ¼ b1 bp1 =2 ð12:317Þ
12.12 Calculation of Spiral Bevel and Hypoid Gears 683
From what has been said above, we realize that each method involves their
specific formulas, which are related to the corresponding calculation procedures of
the pitch cone parameters. For spiral bevel gears, the pinion and wheel face widths
are equal. The calculation point is located in the middle of the pinion face width for
hypoid gears calculated with Method 2; consequently, the inner and outer pinion
face widths, bi1 and be1 , at the pitch cone are equal. For hypoid gears calculated
with Method 1 and 3, the inner and outer pinion face widths, bi1 and be1 , are instead
different. With Method 3, an integer value of the pinion face width along the pitch
cone, bi1 , is chosen. Figure 12.45a shows the quantities of interest as well as how to
get bi1 and be1 , which represent the actual distances from calculation point to inside
and outside, measured along the pitch cone generatrix. In accordance with what has
been done for pinion, also the inner and outer wheel face widths, bi1 and be1 , can be
determined. Once bi1 , be1 , bi2 and be2 have been calculated, it is possible to
determine the inner and outer diameters of pinion and wheel, dai1 , dae1 , dai2 , and
dae2 , for all methods with the same equations.
For the determination of the inner and outer spiral angles, we must distinguish
between pinion and wheel, because they are calculated with different formulas for
face-hobbed and face-milled wheels, while the previous quantities have been cal-
culated using the same formulas both for face-hobbed and face-milled wheels.
The spiral angles of the pinion are determined at its front and rear crowns, i.e. at
the inner and outer cone distances, Ri1 and Re1 (Fig. 12.46). To determine the
values of these spiral angles, it is first necessary to define the spiral angles of the
wheel at boundary points of the hypoid gear pair in pitch plane (points A and B in
Fig. 12.46). At these points, the corresponding cone distances, Ri21 and Re21 , of the
wheel may be smaller/larger than the inner/outer wheel cone distances, Ri2 and Re2 ;
this is due to the overlap of the pinion face over the wheel face. The circles with
center O2 and radii Ri21 and Re21 are shown by dashed lines in Fig. 12.46, which
also highlights the pinion offset angles in pitch plane, fip21 and fep21 , at inner
boundary point, A, and at outer boundary point, B. These angles are used to
determine the inner and outer spiral angles, bi1 and be1 , of pinion. However, it is
noteworthy that, to arrive at common formulas valid for both face-hobbed and
face-milled pinions, a different set of intermediate formulas is used for face-hobbing
and face-milling.
For the determination of the inner and outer spiral angles of pinion, we first
calculate the wheel cone distance of outer pinion boundary point, Re21 (it may be
larger than Re2 ), and the wheel cone distance of inner boundary point, Ri21 (it may
be smaller than Ri2 ), using the following equations:
684 12 Spiral Bevel Gears and Hypoid Gears
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Re21 ¼ R2m2 þ b2e1 þ 2Rm2 be1 cosfmp ð12:319Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Ri21 ¼ R2m2 þ b2i1 2Rm2 bi1 cosfmp : ð12:320Þ
For face-bobbed pinions, the intermediate quantities involved are the lead angle
of cutter, m, crown gear to cutter center distance, qP0 , epicycloid base circle radius,
qb , auxiliary angle, ue21 , auxiliary angle, ui21 , wheel spiral angle at outer boundary
point, be21 , and wheel spiral angle at inner boundary point, bi21 , which are given
respectively by the following relationships:
z0 mmn
m ¼ arcsin ð12:321Þ
2rc0
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
qP0 ¼ R2m2 þ rc0 2 2R r sinðb
m2 c0 m2 mÞ ð12:322Þ
qP0
qb ¼ ð12:323Þ
½1 þ ðz0 =z2 Þsind2
2
Re21 þ q2P0 rc0
2
ue21 ¼ arccos ð12:324Þ
2Re21 qP0
2
Ri21 þ q2P0 rc0
2
ui21 ¼ arccos ð12:325Þ
2Ri21 qP0
Re21 qb cosue21
be21 ¼ arctan ð12:326Þ
qb sinue21
Ri21 qb cosui21
bi21 ¼ arctan : ð12:327Þ
qb sinui21
For face-milled pinions, the intermediate quantities involved are the wheel spiral
angle at outer boundary point, be21 , and wheel spiral angle at inner boundary point,
bi21 , which are given respectively by the following relationships:
2Rm2 rc0 sinbm2 R2m2 þ R2e21
be21 ¼ arcsin ð12:328Þ
2Re21 rc0
2Rm2 rc0 sinbm2 R2m2 þ R2i21
bi21 ¼ arcsin : ð12:329Þ
2Ri21 rc0
point, fip21 , outer pinion spiral angle, be1 , and inner pinion spiral angle, bei1 , which
are given respectively by the following relationships:
ap
fep21 ¼ arcsin ð12:330Þ
Re21
ap
fip21 ¼ arcsin ð12:331Þ
Ri21
Of course, despite the final formulas are the same, the results will be different,
since the quantities appearing in Eqs. (12.332) and (12.333) are different for
face-hobbed and face-milled pinions.
For the determination of the inner and outer spiral angles of wheel, we have to
use different formulas for face-hobbing and face-milling.
For face-hobbed wheels, the quantities involved are the auxiliary angle, ue2 ,
auxiliary angle, ui2 , outer wheel spiral angle, be2 , and inner wheel spiral angle,bi2 ,
which are given respectively by the following relationships:
2
Re2 þ q2P0 rc0
2
ue2 ¼ arccos ð12:334Þ
2Re2 qP0
2
Ri2 þ q2P0 rc0
2
ui2 ¼ arccos ð12:335Þ
2Ri2 qP0
Re2 qb cosue2
be2 ¼ arctan ð12:336Þ
qb sinue2
Ri2 qb cosui2
bi2 ¼ arctan : ð12:337Þ
qb sinui2
For face-milled wheels, the quantities involved are the outer wheel spiral angle,
be2 , and inner wheel spiral angle, bi2 , which are given respectively by the following
relationships:
2Rm2 rc0 sinbm2 R2m2 þ R2e2
be2 ¼ arcsin ð12:338Þ
2Re2 rc0
2Rm2 rc0 sinbm2 R2m2 þ R2i2
bi2 ¼ arcsin : ð12:339Þ
2Ri2 rc0
686 12 Spiral Bevel Gears and Hypoid Gears
Figure 12.45b shows the quantities concerning the inner and outer tooth depth of
pinion. It is to be noted that quantities hai1 , hfi1 , ham1 , hfm1 , hae1 , and hfe1 are
measured along the normal to the pitch cone generatrix, while quantity ht1 is
measured along the normal to the root cone generatrix. The inner and outer
addendum and dedendum can obtained easily taking into account inner and outer
face widths of the pinion.
Here, the quantities that interest us are the outer addendum, hae , outer dedendum,
hfe , outer whole depth, he , inner addendum, hai , inner dedendum, hfi , and inner
whole depth, hi , which are given respectively by the following relationships:
In order to determine the tooth thickness of pinion and wheel, the outer normal
backlash, jen , normal backlash at calculation point, jmn , transverse backlash at
calculation point, jmt2 , and outer transverse backlash, jet2 , are first to be calculated.
As Fig. 12.42 shows, the thickness modification coefficient, xsmn , which is included
among type-I data, is a theoretical value, and does not take into account the
backlash. To take into account the backlash, the thickness modification coefficients,
xsm1 and xsm2 , are to be determined. The backlash has to be set equal to zero only
when the theoretical tooth thickness is considered.
For the determination of the tooth thickness of pinion and wheel, it is necessary
first to calculate the mean normal pressure angle, an , using the following
relationship:
1
an ¼ ðanD þ anC Þ: ð12:346Þ
2
These equations are equivalent to each other, and we choose that for which the
backlash that appears in their expressions is one of the additional data summarized
in Table 12.6.
The mean normal circular tooth thickness of pinion, smn1 , is given by:
These four equations are also equivalent to each other, and we choose that for
which the backlash appearing in their expressions is one of the additional data
shown in Table 12.6.
The mean normal circular tooth thickness of wheel, smn2 , is given by:
Finally, the mean transverse circular thickness, smt , mean normal diameter, dmn ,
mean normal chordal tooth thickness, smnc , and mean chordal addendum, hamc , are
given respectively by the following expressions:
dm1;2
dmn1;2 ¼ ð12:352Þ
1 sin bm1;2 cos2 an cosbm1;2 cosd1;2
2
smnc1;2 ¼ dmn1;2 sin smn1;2 =dmn1;2 ð12:353Þ
smn1;2
hamc1;2 ¼ ham1;2 þ 0:5dmn1;2 cosd1;2 1 cos : ð12:354Þ
dmn1;2
688 12 Spiral Bevel Gears and Hypoid Gears
The remaining dimensions to be determined are the outer pitch cone distance of
pinion, Re1 , inner pitch cone distance of pinion, Ri1 , outer and inner pitch diameters
of pinion, de1 and di1 , outside diameter, dae , diameters dfe , dai and dfi , crossing point
to crown along axis, txo1;2 , crossing point to front crown along axis, txi1;2 , and pinion
whole depth perpendicular to the root cone, ht1 , which are given by:
tzF1 þ txo1
ht1 ¼ sin #a1 þ #f 1 ðtzR1 tzF1 Þsindf 1 : ð12:365Þ
cosda1
The equations in this section are used to check if undercut occurs or not. They are
valid for generated gears with non-constant and constant tooth depth, and can be
applied selecting any point on the pinion or wheel face widths. These equations are
based on the concept of the common generation crown wheel, which is able to mesh
with pinion and wheel at the same time; this common generation crown wheel has
already been described in detail in the previous sections.
To make things even clearer, we can add to what we have already said that the
action of the blades in the cutter-head simulates the one of a tooth of the generation
crown wheel. The axis of rotation of this generation crown gear coincides with the
generation cradle axis of the gear-cutting machine. The gear wheel to be generated
(e.g., the pinion) rolls without sliding with the imaginary generating and mating
12.12 Calculation of Spiral Bevel and Hypoid Gears 689
crown gear, and in this motion its tooth spaces and flanks are manufactured. To
generate the mating gear wheel to be fit properly with this pinion, a mirrored
arrangement is used, i.e. this mating gear wheel is cut on the backside of the same
generation crown wheel. This is a general principle, which applies independently of
the lengthwise tooth form (circular, epicycloid or involute arcs).
For the undercut check of pinion, we first choose the pinion cone distance of the
point to be checked, Rx1 , obviously within the range ðRi1 Rx1 Re1 Þ, and then we
define the wheel cone distance of the appropriate pinion boundary point, Rx2 (as
Fig. 12.46 shows, this pinion boundary point may be smaller than Ri2 and larger
than Re2 ), using the relationship:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Rx2 ¼ R2m2 þ ðRm1 Rx1 Þ2 2Rm2 ðRm1 Rx1 Þcosfmp : ð12:366Þ
For face-hobbed pinion, we determine in sequence the auxiliary angle, ux2 , and
wheel spiral angle at checkpoint, bx2 , with the equations:
2
Rx2 þ q2P0 rc0
2
ux2 ¼ arccos ð12:367Þ
2Rx2 qP0
Rx2 qb cosux2
bx2 ¼ arctan ; ð12:368Þ
qb sinux2
while for face-milled pinion we determine the wheel spiral angle at checkpoint, bx2 ,
with the equation:
2Rm2 rc0 sinbm2 R2m2 þ R2x2
bx2 ¼ arcsin : ð12:369Þ
2Rx2 rc0
z1 cosbx2
dEx1 ¼ dx2 ð12:375Þ
z2 cosbx1
dEx1
REx1 ¼ ð12:376Þ
2sind1
z1
znx1 ¼ ð12:377Þ
1 sin2 bx1 cos2 an cosbx1 cosd1
tand1 tand2 REx1 sinbx1 Rx2 sinbx2
alimx ¼ arctan ð12:378Þ
cosfmp REx1 tand1 þ Rx2 tand2
It is to be noted that, for further calculations, the smaller effective pressure angle
must be chosen, i.e. aeminx ¼ aeCx , if aeCx \aeDx , and aeminx ¼ aeDx , if aeCx aeDx .
Now the minimum profile shift coefficient of pinion at calculation point, xhmminx1 ,
should be determined. The quantities involved are the working tool addendum at
checkpoint, khapx , minimum profile shift coefficient of pinion at checkpoint, xhx1 ,
and minimum profile shift coefficient of pinion at calculation point, xhmminx1 , which
are given respectively by the following relationships:
face-milled wheels. We calculate the wheel pitch diameter at checkpoint, dx2 , and
normal module at checkpoint, mxn , using Eqs. (12.373) and (12.374). Then we
calculate the intermediate value, znx2 , with the equation:
z2
znx2 ¼ ; ð12:384Þ
1 sin bx2 cos2 an cosbx2 cosd2
2
and we note that, for further calculations, the smaller effective pressure angle must
be chosen, i.e. aeminx ¼ anC , if anC \anD , and aeminx ¼ anD , if anC anD .
Now the maximum profile shift coefficient of pinion at calculation point,
xhmmaxx1 , should be determined. The quantities involved are the working tool
addendum at checkpoint, khapx , and maximum profile shift coefficient of pinion at
calculation point, xhmmaxx1 , which are given respectively by the following
relationships:
References
34. Litvin FL, Petrov KM, Ganshin VA (1974) The effect of geometrical parameters of hypoid
and spiroid gears on their quality characteristics. J Eng Ind 96:330–334
35. Litvin FL, Wang AG, Handschuh RF (1998) Computerized generation and simulation of
meshing and contact of spiral bevel gears with improved geometry. Comput Methods Appl
Mech Eng 158:35–64
36. Maitra GM (1994) Handbook of gear design, 2nd edn. Tata McGraw-Hill Publishing
Company Ltd., New Delhi
37. Merritt HE (1954) Gears, 3th edn. Sir Isaac Pitman&Sons, Ltd, London
38. Micheletti GF (1958) Lavorazioni dei Metalli ad Asportazione di Truciolo. Libreria Editrice
Universitaria Levrotto&Bella, Torino
39. Mozzi del Garbo GG (1763) Discorso matematico sopra il rotolamento momentaneo dei
corpi. Stamperia di Donato Campo, Napoli
40. Müller H (2007) Face-off: face hobbing vs. face milling. Gear Solution Magazine, 1 Sept
2007
41. Naunheimer H, Bertsche B, Ryborz J, Novak W (2011) Automotive transmissions:
fundamentals, selection, design and application, 2nd edn. Springer, Berlin, Heidelberg
42. Niemann G, Winter H (1983) Maschinen-Elemente, Band III: Schraubrad-, Kegelrad-,
Schnecken-, Ketten-, Rienem-, Reibradgetriebe, Kupplungen, Bremsen, Freiläufe. Springer,
Berlin, Heidelberg
43. Pauline K, Irbe A, Torims T (2014) Spiral bevel gears with optimized tooth-end geometry.
Procedia Eng 69:383–392
44. Pollone G (1970) Il Veicolo. Libreria Editrice Universitaria Levrotto & Bella, Torino
45. Poritsky H, Dudley DW (1948) Conjugate action of involute helical gears with parallel or
inclined axes. Quaterly Appl Mathem VI(3)
46. Powell DE, Barton HR (1959) Analytical study of surface loading and sliding velocity of
automotive hypoid gears. ASME Trans II(2):173–183
47. Radzevich SP (2010) Gear cutting tools: fundamentals of design and computation. CRC
Press, Taylor&Francis Group, Boca Raton, Florida
48. Radzevich SP (2016) Dudley’s handbook of practical gear design and manufacture, 3rd edn.
CRC Press, Taylor&Francis Group, Boca Raton, Florida
49. Romiti A (1960) Sopra un tipo di ingranaggi conici per assi concorrenti e sghembi, vol XCIV.
Atti dell’Accademia delle Scienze di Torino, Torino
50. Romiti A (1960) Problemi di dimensionamento, accoppiamento e fabbricazione di ingranaggi
conici per uso universale, vol XCIV. Atti dell’Accademia delle Scienze di Torino, Torino)
51. Rossi M (1965) Macchine Utensili Moderne. Editore Ulrico Hoepli, Milano
52. Schiebel A, Lindner W (1954) Zahnräder, Band 1: Stirn-und Kegelräder mit geraden Zähnen.
Springer, Berlin, Heidelberg
53. Schiebel A, Lindner W (1957) Zahnräder, Band 2: Stirn-und Kegelräder mit schrägen Zähnen
Schraubgetriebe. Springer, Berlin, Heidelberg
54. Scotto Lavina G (1990) Riassunto delle Lezioni di Meccanica Applicata alle Macchine:
Cinematica Applicata, Dinamica Applicata - Resistenze Passive - Coppie Inferiori, Coppie
Superiori (Ingranaggi – Flessibili – Freni). Edizioni Scientifiche SIDEREA, Roma
55. Sheveleva GI, Volkov AE, Medvedev VI (2007) Algorithms for analysis of meshing and
contact of spiral bevel gears. Mech Mach Theory 42:198–215
56. Shih YP, Fong ZH, Lin GC (2007) Mathematical model for a universal face hobbing hypoid
gear generator. J Mech Des 129(1):38–47
57. Simon V (2007) Computer simulation of tooth contact analysis of mismatched spiral bevel
gears. Mech Mach Theory 42:365–381
58. Simon V (2009) Design and manufacture of spiral bevel gears with reduced transmission
errors. ASME J Mech Des 131(4):041007-1–041007-11
59. Spear GM, King CB, Baxter ML (1960) Helixform bevel and hypoid gears. ASME Trans
82-B(III):179–190
60. Stachel H, Figliolini G, Angeles J (2019) The logarithmic spiral and its spherical
counterpart. J Ind Des Eng Graph 14(1):91–98
694 12 Spiral Bevel Gears and Hypoid Gears
61. Stadtfeld HJ (1993) Handbook of bevel and hypoid gears: calculation. Rochester Institute of
Technology, Manufacturing and Optimization, Rochester
62. Stadtfeld HJ (2011) Tribology aspects in angular transmission systems, Part VII: hypoid
gears. Gear Technol 66–72
63. Stadtfeld HJ (2014) Gleason bevel gear technology. Gleason Works, Rochester
64. Su J, Fang Z, Cai X (2013) Design and analysis of spiral bevel gears with seventh-order
function of transmission error. CSAA, Chin J Aeronaut 26(5):1310–1316
65. Suh S-H, Jung D-H, Lee E-S, Lee S-W (2003) Modeling, implementation, and manufacturing
of spiral bevel gears with crown. Int J Adv Manuf Technol 21:775–786
66. Townsend DP (1991) Dudley’s gear handbook. McGraw-Hill, New York
67. Tsai YC, Chin PC (1987) Surface geometry of straight and spiral bevel gears. J Mech
Transmissions Autom Des 109(4):443–449
68. Vimercati M (2007) Mathematical model for tooth surfaces representation of face-bobbed
hypoid gears and its application to contact analysis and stress calculation. Mech Mach Theory
42(6):668–690
69. Vimercati M, Piazza A (2005) Computerized design of face hobbed hypoid gears: tooth
surface generation, contact analysis and stress calculation. AGMA Fall Technical Meeting,
Paper N. 05FTM05
70. Wang J, Lim TC, Li M (2007) Dynamics of a hypoid gear pair considering the effects of
time-varying mesh parameters and backlash nonlinearity. J Sound Vib 308:302–329
71. Wang P-Y, Fong Z-H (2005) Adjustability improvement of face-milling spiral bevel gears by
modified radial motion (MRM) method. Mech Mach Theory 40:69–89
72. Wildhaber E (1946) Basic relationship of hypoid gears. Am Machinist, XC, February, pp 14,
and 28, March, p 14, June, pp 6 and 20, July, p 18, and August, pp 1 and 15
73. Wildhaber E (1956) Surface curvature. Prod Eng 27:184–191
74. Zhang Y, Litvin FL, Handschuh RF (1995) Computerized design of low-noise face-milled
spiral bevel gears. Mech Mach Theory 30(8):1171–1178
Chapter 13
Gear Trains and Planetary Gears
Abstract In this chapter, the general concepts of ordinary gear trains are first
described, particularly those concerning their efficiency and the obtainable trans-
mission ratios. The same concepts are then extended to planetary gear trains and
various methods of calculating the transmissions ratios achievable with these types
of gear drives are defined. Some problems related to simple planetary gear trains are
then examined, and the main characteristics of some of these gear drives are
described and discussed. Particular planetary gear trains are then dealt with, such as
those concerning summarizing and differential gear trains. The analysis is then
extended to multi-stage planetary gear trains and parallel mixed power trains
consisting of ordinary and planetary gear trains, and the calculation procedure is
described of the braking torque to be applied to the members of the planetary gear
train that are held at rest. Finally, the problem of the efficiency of a planetary gear
train is addressed and discussed in its general terms, also with reference to those
particular planetary gear trains used to realize large transmission ratios.
13.1 Introduction
A gear pair is a mechanism consisting of two gears rotating about axes the relative
position of which is fixed, and one gear turns the other by action of teeth succes-
sively in meshing contact. A train of gears or gear train is any combination of gear
pairs, i.e. a combination of two or more gear pairs (or a combination of two or more
gears), mounted on rotating shafts, to transmit power or torque. In common prac-
tical applications, a gear train acts as a speed reducer, but the industrial application
where the gear train acts as a speed increaser are also frequent.
The more or less numerous gear pairs of a gear train can be arranged in an
endless variety of ways to meet the particular shaft position, direction of rotation,
and transmission ratio requirements (see: Merritt [37], Dudley [12], Ferrari and
Romiti [14], Pollone [51], Maitra [36].
From the theoretical point of view, we can get a large transmission ratio using
only two gears. In doing so, however, we would get a gear pair characterized by a
small pinion and an enormously large wheel, with obvious problems of overall
dimensions and technical feasibility. Such a design choice is not technically and
economically viable. The desired transmission ratio is then achieved by using a gear
train, which may be constituted by different types of gears (e.g. parallel cylindrical
spur and helical gears, cylindrical crossed helical gears, straight bevel, spiral bevel
and hypoid gears, worm gears, etc.).
The gear trains can be classified into two main families:
• the family of the ordinary gear trains, in which all the gears rotate about axes
which are fixed with respect to a housing, which is the reference frame;
• the family of the epicyclic gear trains, also called planetary gears trains, in
which at least one gear rotates with respect to its axis, which in turn rotates with
respect to the fixed frame.
Each of these two families of gear trains has its own characteristics. However, we
can consider the family of the ordinary gear trains as a special case of the more general
family of planetary gear trains, since the former can be obtained from the second when
the moving axes of the planetary gear trains are restrained and held at rest.
Ordinary gear trains as well as planetary gear trains are widely used in practical
geared transmissions, which are often irreplaceable in many technological areas
(industrial plants, machine tools, textile machines, power trains for cars, aircrafts
and helicopters, auxiliary devices for steam and gas turbines, mills and kilns for
cement production, ceramic and brick industry, food industry, lifting and transport
machines, wind generators, etc.) They are in fact used to mediate between the
characteristics of the power delivered by the prime mover and those of the driven
machine. For example, in automotive transmissions, ordinary and planetary gear
trains have the function of adapting the available traction power to the required
power, while also ensuring the desired performance [31, 45, 51].
(d)
Fig. 13.1 Examples of ordinary gear trains with non-coaxial driving and driven shafts
Fig. 13.2 Example of ordinary gear trains with coaxial driving and driven shafts
where x1 and x2 are the angular velocities of the driving and driven gears,
expressed in rad/s, n1 and n2 are the rotational speeds of the same gears, expressed
in min−1 (rpm, revolution per minute), and z1 and z2 are their numbers of teeth. In
this case, the transmission ratio, i, coincides with the gear ratio, u, which is defined
as the quotient of the number of teeth of the wheel divided by the number of teeth of
the pinion.
For a compound ordinary gear train constituted by two gear pairs, arranged in
series (Fig. 13.1b, c), the transmission ratio is given by:
x1 n1 z2 z4
i¼ ¼ ¼ ; ð13:2Þ
x2 n2 z1 z3
where z1 and z3 are the numbers of teeth of the two driving gears, while z2 and z4
are the numbers of teeth of the two driven gears.
Generally, for a compound ordinary gear train consisting of more than two gear
pairs, arranged in series, the transmission ratio is equal to the quotient of the
product of numbers of teeth of the driven gears divided by the product of numbers
of teeth of the driving gears. It is therefore expressed by a relationship that is the
generalization of Eq. (13.2).
If the gear axes of the compound ordinary gear train are all parallel, we can
assign a sign to the transmission ratio, making it positive if the first and last gear
wheels have the same direction of rotation and negative in the opposite case. It is
therefore apparent that, if a gear wheel, Wi , mounted on any intermediate shaft,
operates both as a driving and driven gear wheel (it then simultaneously engages
with the proceeding gear wheel ðWi 1Þ and with the following gear wheel
ðWi þ 1Þ, it does not alter the absolute value of the transmission ratio, corre-
sponding to the direct engagement of the gear wheels, ðWi 1Þ, and ðWi þ 1Þ, but
reverses the direction of rotation that the gear wheel ðWi þ 1Þ would have in the
case of its direct engagement. Thus, the effect of each idle gear is to reverse the
direction of rotation, while the overall transmission ratio is not changed.
Intermediate gear wheels can also be used in the case of gear trains including
straight bevel and spiral bevel gears.
13.2 Ordinary Gear Trains 699
u1
i¼ : ð13:3Þ
u2
Then neglecting the power losses due to friction and other causes, by the
resulting equality between the input power, P1 ¼ T1 x1 , and the output power,
P2 ¼ T2 x2 , we deduce:
T2
i¼ ¼ iT ; ð13:4Þ
T1
where T1 and T2 are respectively the input and output torque, and iT is the torque
conversion factor defined in Sect. 1.1.
If we want to take into account the inevitable power losses due to various causes,
we have to introduce the concept of efficiency, g, defined as the quotient of the
output power from the last driven gear divided by the input power in the first
driving gear of the gear train. Equation (13.4) is no longer valid because, due to the
lost power Pd ¼ P1 P2 , the efficiency is always less than one ðg\1Þ. In this
regard, we must keep in mind that it is necessary to consider at least two values of
the efficiency, depending on whether the prime mover power flows from gear 1 to
gear 2, or vice versa. In the first case, the efficiency will be g1 , while in the second
case the efficiency will be g2 [54].
In the first case, it is g1 ¼ P2 =P1 , for which the Eq. (13.4) should be replaced
with the equation:
T2
ig1 ¼ ; ð13:5Þ
T1
while, in the second case, it is g2 ¼ P1 =P2 , for which the Eq. (13.4) should be
replaced with the equation:
i T2
¼ : ð13:6Þ
g2 T1
The two ratios between the output and input torques given by Eqs. (13.5) and
(13.6) are different from each other even when g1 ¼ g2 . Under the hypothesis that
discontinuities do not exist passing from the static condition to a dynamic condi-
tion, we infer the rule according to which gears initially at rest continue to remain at
rest in case where the ratio ðT2 =T1 Þ is in the range:
while the motion takes place if this ratio comes out of aforementioned range.
When the transmission ratio is slightly different from the unit, as it occurs for
spur and bevel gears, we can obtain very high efficiencies g1 and g2 , which differ
little from one another. In practice, we consider a single value of gear efficiency,
13.2 Ordinary Gear Trains 701
given by g ¼ g1 ¼ g2 . The parallel cylindrical spur gears can reach value of effi-
ciency equal to g ¼ g1 ¼ g2 ¼ 0:98, while bevel gears are characterized by a little
lower efficiency. Frequently, due to manufacturing inaccuracies, the values of ef-
ficiency are lower.
Instead, when the transmission ratio is small or large, the efficiencies g1 and g2
generally get worse, and become very different from each other. Indeed, we find
that the best efficiency is when the gear works as a speed reducer ði [ 1Þ, i.e. when
the driving gear wheel is the fast wheel, while the worst efficiency is when the gear
works as a speed multiplier ði\1Þ, i.e. when the driving gear wheel is the slow
wheel. In the first case we speak of direct motion, while in the latter case we speak
of retrograde motion.
A possible reason for this behavior can be given assuming that the lost power,
Pd ¼ P1 P2 , in the gear is proportional to the power, P2 , applied to the slower
wheel, on which the most relevant torque acts. In direct motion, for which 1 and 2
are respectively the fast and slow gear wheels, we have:
Pd P1 P2 1
¼ ¼ 1: ð13:8Þ
P2 P2 g1
In retrograde motion, while maintaining the ratio ðPp =P2 Þ characterizing the
direct motion, being
P1 Pd
g2 ¼ ¼1 ; ð13:9Þ
P2 P2
we get:
1
g2 ¼ 2 : ð13:10Þ
g1
several mechanisms, i.e., mechanism linked so that the follower (or driven member)
of any of the component mechanisms is the driving member of the next one. It is
also an example of a machine that can be considered as absolute regime machine,
i.e. capable of working, under normal operating conditions, so that its kinetic
energy is maintained constant, so its variation in any time interval is equal to zero.
In fact, during the engagement of two gear wheels, even under steady-state
conditions, a periodicity in operating conditions exist, correlated with the time
needed to rotate each gear wheel of an angular pitch. The corresponding period is,
however, so small, and so does the kinetic energy variation within the same period,
so the effects of the related variations can generally be neglected. Now, the absolute
regime machines resulting from a series connection of several mechanisms possess
a well-known property: the one that the efficiency of the machine is the product of
the efficiencies of the single mechanism that make up the same machine [14]. Rebus
sic stantibus, it follows that the efficiency of an ordinary gear train, whatever it is,
simple or compound, is equal to the product of the efficiencies of the single gear
pairs, since they are arranged in series between them. Therefore, with product
notation P, we will have:
Y
n
g¼ gi ð13:11Þ
i¼1
where gi is the efficiency of the i-th gear pair that makes up the gear train con-
sidered. Therefore, with regard to efficiency, the idler gears (they, as we have
already seen, do not affect the transmission ratio) are playing the same role as all the
other gear wheels that make up the gear train.
In the case of more complex gear trains in which the input power first is divided
in two or more flow lines, and then it rejoins downstream of the various gears
through which the same power has flowed, the calculation of the efficiency is more
complicated. In this case in fact the efficiency is no longer equal to the product of
the efficiencies of each single gear pair. Therefore, Eq. (13.11) is no longer valid,
since the efficiency of the single gears has a different weight, depending on the
amount of power that passes through them (see Sects. 13.8 and 13.12).
[37], Dudley [12], Henriot [21], Lyndwander [35], Maitra [36], Litvin and Fuentes
[32], Jelaska [24], Radzevich [53]); many textbooks, which deal with general
machine design topics (see: Giovannozzi [18], Juvinall [25], Niemann and Winter
[46], Budynas and Nisbett [3]); textbooks on automotive design and transmissions
(see: Pollone [51], Morelli [43], Lechner and Naunheimer [31], Genta and Morello
[17], Fisher et al. [15]); specific textbooks and literature (see: Kelley [28], Henriot
[22], Pearson [49], Ruggieri [54], Litvin et al. [33], Cooley and Parker [5],
Balbayev [1]). However, this topic continues to attract the attention of prominent
scholars, which still make it an advanced research subject, especially with reference
to new and innovative applications, such as gear drive systems for helicopters and
hybrid electric vehicles. In this regard, in addition to the works referred by Litvin
and Fuents [32], we limit ourselves to citing the following papers: Coy et al. [6],
Simionescu [55], Miller [41], Chen and Angeles [4], Galvagno [16], Liu and Peng
[34], Pennestrì et al. [50], Davies et al. [7], Gupta and Remanarayanan [20], Dheaud
and Pullen [9], Gupta et al. [19], Yang et al. [62], Kahraman et al. [27], Xue et al.
[61], Essam and Isam [13].
According to ISO 1122-1:1998 [23], an epicyclic gear train (also called plan-
etary gear train, or more simply epicyclic gear or planetary gear) is a combination
of coaxial elements, of which one or more are annulus gears (or ring gears) and
one or more are planet carriers (or train arms or spiders). Planet carriers rotate
around the common axis and support one or more planet gears that mesh with the
annulus gears and one or more sun gears. Usually, the planet gears are equally
spaced, to balance the forces acting on sun gear, annulus gear and planet carrier.
The first and last gear wheels of the planetary gear train, taken in the direction of the
subsequent engagements of their teeth, are called main wheels. Generally, a plan-
etary gear train can be derived from an ordinary gear train whose housing, instead
of being held at rest, is rotated about the driving axis. In so doing, the speed of the
driven axis becomes a function of speeds of the driving axis and housing.
However, it should be borne in mind that the above-mentioned ISO definition
does not encompass all the possible configurations of planetary gear trains con-
ceived and developed to solve real practical problems. In fact, by limiting attention
to usual applications, the axes of the main wheels and the axis of the planet carrier
are parallel, while the intermediate axes may have different directions. The family
of planetary gear trains is therefore much larger, as it includes (see Ferrari and
Romiti [14]:
(a) planetary gear trains whose end axes are fixed and coincide with the axis of the
planet carrier;
(b) planetary gear trains with a mobile end wheel, whereby the last axis is distinct
from the first, but parallel to it;
(c) planetary gear trains with a mobile end wheel, whereby the last axis is distinct
from the first, but is directed according to any angle with respect to the latter,
while the axis of the planet carrier coincides with the axis of the first main
wheel;
704 13 Gear Trains and Planetary Gears
(d) planetary gear trains in which also the axis of the planet carrier is mobile.
These mechanisms are systems with two degrees of freedom. When, however, they
are considered in the framework of the machine to which they belong, the number
of degrees of freedom is reduced to one by virtue of the constraints that immobilize
the first (or the last) wheel or by virtue of the couplings that establish a determined
correlation between rotations of the main wheels and rotation of the planet carrier.
A relationship between these rotations then exist, which depends on the geometry
of the gear wheels that make up the planetary gear train. This relationship depends
only on the geometry of the gear system and can easily be obtained if the axes of the
main wheels and planet carrier are parallel, and the axis of the planet carrier is held
at rest (see next section).
The simplest of the planetary gear trains consists of two gears mounted so that
the center of one gear revolves around the center of the other gear (Fig. 13.3).
A planet carrier connects the centers of the two gears and rotates to carry the planet
gear around the sun gear. The planet and sun gears mesh so that their pitch circles
roll without sliding. When the sun gear is fixed and the planetary gear roll around
the sun gear, a point on the pitch circle of the planet gear traces an epicycloid curve.
The simple epicyclic gear train described above may be housed at the inside of
an annulus gear and assembled in such a way that the pitch circle of the planet gear
rolls without sliding on the inside of the pitch circle of the annulus gear supposed
fixed (Fig. 13.4). In this case, a point on the pitch circle of the planet gear traces a
hypocycloid curve.
Generally, we call planetary gear train a mechanism consisting of the combi-
nation of all four coaxial component members shown in Fig. 13.4, i.e. the mech-
anism composed of at least one planet gear, one sun gear, one planet carrier, and
one annulus gear. However, in most cases the terms planetary gear train and epi-
cyclic gear train are used interchangeably. In the most frequent case, the annulus
gear is usually fixed, the planet carrier is driving, and the sun gear is driven.
Albeit in a more streamlined shape, this kind of mechanism has an ancient
history: in fact, it dates back to the 5th century BC, when the Greeks invented the
idea of epicycles, i.e. of circle traveling on the circular orbits. According to the
epicyclic theory, the Antikythera Mechanism (about the ’80 BC) had gearing that
was able to approximate the moon’s elliptical path through the heavens, while
Claudius Ptolemy in the Almagest (in 148 AD) was able to predict planetary orbital
paths (Ptolemy, in translated edition, [52].
Compared to an ordinary gear train, which has only one degree of freedom
movement, a planetary gear train has two degrees of freedom of movement.
Therefore, the planetary gear trains, as they are characterized by three main com-
ponent members (the sun gear, annulus gear, and planet carrier), can be used in
various ways to obtain combinations of rotary motions [48, 57].
A planetary gear system, even in its simplest configurations, possesses unique
characteristics, which are summarized as follows: compactness, also due to the
coaxial arrangement of the driving and driven shafts; high power-to-weight ratio;
large speed reduction possibilities compared to its overall size; possibilities of a
large number of combinations of driving and driven inputs and outputs; possibilities
of large torque conversion; different possibilities of orientation of the gear system.
Planetary gear trains can be typically classified as simple or compound planetary
gear trains. A simple planetary gear train has one sun gear, one annulus gear, one
planet carrier and one planet gear set. A compound planetary gear train involves
one or more of the following three types of substructures:
• meshed-planet gears, for which the compound planetary gear train is charac-
terized by at least two more planet gear in mesh with each other;
• stepped-planet gears, for which the compound planetary gear train is charac-
terized by a shaft connection between two planet gears;
• multi-stage structures, for which the compound planetary gear train contains
two or more planet gear sets.
The already remarkable characteristics, which we highlighted above for the
simple planetary gear trains, are further enhanced in the compound planetary gear
trains. The latter in fact have the advantages of higher torque-to-weight ratio, larger
reduction ratio and more flexible configurations.
706 13 Gear Trains and Planetary Gears
A simple planetary gear train may have different configurations. Usually, the
axes of all gears are parallel. However, for special cases, such as differentials or
pencil sharpeners, these axes can be placed at an angle, introducing members
configured as bevel gears. Furthermore, the axes of the sun gear, planet carrier and
annulus gear are usually coaxial. Depending on whether the axes are parallel (or
coaxial) or not parallel, we speak of planar planetary gear trains or spherical
planetary gear trains.
A compound planetary gear train with meshed-planet gears is characterized by at
least two more planet gears in mesh with each other. One of these two planet gears
meshes with the sun gear, while the other planet gear meshes with the annulus gear.
With a compound planetary gear train so configured, the annulus gear rotates in the
same direction of the sun gear when the planet carrier is held at rest; in this way, the
mechanism provides a reversal in direction of rotation compared with a simple
usual epicyclic gear train.
Instead, a compound planet gear, composed by two differently sized planet gears
forming a single block with their common shaft, characterizes a compound plan-
etary gear train with stepped-planet gears. The smaller planet gear meshes with the
sun gear, while the larger planet gear meshes with the annulus gear. We use this
type of compound planetary gear train when the overall package size is limited, and
it is necessary to achieve smaller step changes in gear ratio. The speed hub of
internally geared bicycle hubs constitutes an example of this type of compound
epicyclic gear train. As a down side the problems posed by these types of com-
pound epicyclic gear trains, which have a relative gear mesh phase (or timing
marks), should not be underestimated. Moreover, their assembly conditions are
more restrictive than those of a simple planetary gear train. In fact, these compound
planetary gear trains with stepped-planet gears must be assembled in such a way
that their initial relative position is correct, in order to prevent their teeth from
simultaneously meshing the annulus and sun gear at opposite ends of the compound
planet, as this condition would determine a very rough running and short life.
Finally, a multi-stage structure is a compound planetary gear system where more
planet and sun gear units are placed in series into the same annulus gear housing.
The various component units (or substructures) are assembled in such a way that the
output shaft of the first stage become the input shaft of the next stage, and so on. As
an example, consider a simple planetary gear train, where no component member is
fixed, and the power comes from the sun gear and does go out from the planet
carrier. In this case, the sun gear drives the planet gears assembled with the annulus
gear to operate. The planet carrier, connected to the planet gears, revolves on its
own axis and along the annulus gear, and the output shaft, connected to the planet
carrier, rotates with a speed reduced with respect to the input speed. A higher
reduction ratio can be obtained by doubling the multiple staged gears and planet
carriers. A gear system so configured may provide a larger or smaller gear ratio,
depending on the sizing of the various component substructures. By choosing a
suitable combination of the component units, speed ratio of 104:1 can be easily
obtained. Some automatic transmission of vehicles works in this way. The number
13.3 Epicyclic or Planetary Gear Trains: Definitions and Generalities 707
The transmission ratio in a planetary gear train can be obtained using several
methods, among them equivalent, that lead to the same results. Here we describe
only three of these methods, which are commonly used. In addition, we mention
two other methods also usually used.
Consider the most frequent case of planetary gear train, which is the one corre-
sponding to the schematic drawing shown in Fig. 13.5 for a planar planetary gear
train, or a similar schematic drawing for a spherical planetary gear train.
Let xa , xs , xc , and xp respectively be the angular velocities of the annulus gear,
sun gear, planet carrier and planet gear. Of course, in place of these angular
velocities (in rad/s), the equivalent rotational speeds na , ns , nc , and np (in min−1, i.e.
rpm), can be considered. From what is known on the kinematics of mechanisms, it
is obvious that the intercorrelation existing between the three angular velocities xa ,
xs and xc does not change, when we do rotate the whole planetary system with
angular velocity xc about the common axis of rotation. In doing so, the angular
velocities of the annulus gear and sun gear are respectively equal to ðxa xc Þ and
ðxs xc Þ, while the angular velocity of the planet gear becomes equal to zero
ðxc xc ¼ 0Þ. In this condition, the planetary gear train is thus reduced to an
ordinary gear train, with a driving shaft and a driven shaft, and with the axes of
rotation of the planet gears fixed in three-dimensional space (see: Paul [48], Uicker
et al. [57]).
where xr1 ¼ ðx1 x3 Þ and xr2 ¼ ðx2 x3 Þ are, in value and sign, the relative
angular velocities of the first gear wheel and the last gear wheel with respect to the
planet carrier. Introducing these two relative angular velocities in Eq. (13.12), we
obtain:
x1 x3
i0 ¼ : ð13:13Þ
x2 x3
This is the famous and well-known Willis formula, which correlates linearly
between them the three angular velocities x1 ¼ xs , x2 ¼ xa , and x3 ¼ xc . This
710 13 Gear Trains and Planetary Gears
formula allows to calculate one of the three angular velocities, once the values of
the other two angular velocities, and the numbers of teeth of the gear wheels of the
planetary gear train are known [59].
The Willis formula can also be written in terms of angles of rotation, in
accordance with Eq. (13.3). Thus indicating respectively with u1 , u2 , and u3 the
angles of rotation of the three main members of the planetary gear train (in the
order, the sun gear, 1, annulus gear, 2, and planet carrier, 3), all three measured
starting from a given initial position, we can write:
u1 u3
i0 ¼ : ð13:14Þ
u2 u3
In the technical and scientific literature, the use of the reciprocal of the char-
acteristic ratio, that is the quantity w ¼ 1=i0 , is often preferred, for which the Willis
formula is written in the following form:
1 x2 x3
w¼ ¼ : ð13:15Þ
i0 x1 x3
Like i0 , also w can be positive or negative, and the two sign conventions do not
differ. Resolving the Eq. (13.15) with respect to x1 , x2 , and x3 , we get the fol-
lowing three relationships:
1 1
x1 ¼ ½x2 þ x3 ðw 1Þ; x2 ¼ wx1 þ ð1 wÞx3 ; x3 ¼ ðwx1 x2 Þ:
w w1
ð13:16Þ
These equations show that each of the three angular velocities x1 , x2 , and x3 is
a linear function of the other two. From the second of Eq. (13.16), we infer the
following three special cases:
(a) If w ¼ þ 1, we have x2 ¼ x1 , whatever the value of x3 . Therefore, the
angular velocity x2 is independent of x3 , and the planetary gear train behaves
like a rigid coupling.
(b) If w ¼ þ ð1=2Þ, we have:
1
x2 ¼ ðx1 þ x3 Þ: ð13:17Þ
2
1
x3 ¼ ðx1 þ x2 Þ: ð13:18Þ
2
13.4 Transmission Ratio in a Planetary Gear Train 711
The planetary gear trains for which the Eqs. (13.17) and (13.18) are valid, are
used in cases in which the speed of rotation of a shaft should be proportional to the
sum of the speeds of rotation of the other two shafts.
Equations (13.16) allow us to easily handle the two particular cases in which one
of the two main wheels, 1 or 2, is held at rest. In fact, if the first main wheel is held
at rest, i.e. x1 ¼ 0, from the second of Eqs. (13.16) we find that the transmission
ratio between the shafts of the second main wheel and the planet carrier is given by
x2 =x3 ¼ ð1 wÞ. If instead the second main wheel is held at rest, i.e. x2 ¼ 0,
from the first of Eqs. (13.16) we find that the transmission ratio between the shafts
of the first main wheel and the planet carrier is given by x1 =x3 ¼ ðw 1Þ=w. Both
of these two conditions allow us to obtain planetary gear trains capable to achieving
very small transmission ratios between two parallel shafts, that of the planet carrier
and that of the main wheel, which is rotating. In fact, from what we have said
above, we infer that, to reach this interesting design goal, it is sufficient to design
the gear wheels so that i0 ¼ 1=w is close to the unit as much as is desired; con-
sequently, the transmission ratio x1 =x3 (or x2 =x3 ) becomes close to zero as much
as is desired.
The characteristic ratio, i0 , and its reciprocal, w, may be expressed as the quo-
tient of the product of number of teeth of driving gears in a gear train divided by the
product of number of teeth of driven gears in the same gear train, i.e. in the form
1 x2 x3 z1 z3 . . .
w¼ ¼ ¼ ; ð13:19Þ
i0 x 1 x 3 z2 z4 . . .
where z1 , z3 , … are the numbers of teeth of driving gears, and z2 , z4 , … are the
numbers of teeth of the driven gears. In Eq. (13.19), we take the plus or minus sign
depending on whether, with the planet carrier held at rest, the directions of rotation
of the first gear wheel (the driving gear wheel, 1) and the last gear wheel (the driven
gear wheel, 2) are the same or opposite.
Furthermore, it is to be noted that the angular velocities x1 , x2 , and x3 (and thus
the corresponding rotational speeds n1 , n2 , and n3 ) are absolute angular velocities,
that is angular velocities in absolute motion of the three main members of the
planetary gear train with respect to its housing, which constitutes the frame.
Finally, it should be noted that the above-described algebraic method, from
which the Willis formula is obtained, is applicable to planetary gear trains of type
(a) and (b), as defined in the previous Sect. 13.3. To solve the kinematic problems
of planetary gear trains of type (c) and (d) defined in the same Sect. 13.3, other
more elaborate methods must be used. For planetary gear trains of type (c), a
method similar to the one described above may be used, with the addition of other
conditions due to their geometric configuration (see Ferrari and Romiti [14]). For
planetary gear trains of the type (d), a more complex method must be used, such as
the one proposed and described by Ferrari and Romiti [14]. This method, on which
we do not dwell here for brevity, is quite general, so it is applicable in all possible
cases, including the three previous cases, though it is less easy to apply in these
cases than the one above described.
712 13 Gear Trains and Planetary Gears
This method involves the analysis of the balance of torques acting on the single
members of the planetary gear train. For the description of the method, we consider
the Fig. 13.6, where these members are shown separately, that is in exploded view
that allows a free-body force analysis to be carried out (see: Juvinall [25], Juvinall
and Marshek [26]). With the notations shown in this figure, we find that the distance
of the axes of the planet gears from the axis of the gear system is equal to the sum of
the
radii of pitch circles of the sun gear and planet gear. This distance, equal to
rs þ rp , can be expressed as follows:
ra rs ra þ rs
rs þ rp ¼ rs þ ¼ ; ð13:20Þ
2 2
where rs , rp , and ra are respectively the radii of the pitch circles of the sun gear,
planet gear, and annulus gear.
Let us assume that the power losses are equal to zero, the sun gear is held at rest,
the driving torque or input torque, Ti , is applied to the annulus gear, and the
resistant torque or output torque, T0 , is applied to the planet carrier. To satisfy the
conditions of balance of torques acting on the three members of the planetary gear
train, through which the driving power flows (they are, in order, the annulus gear,
planet gears, and planet carrier), we introduce the necessary forces acting at various
radii. This procedure, summarized in Fig. 13.6, leads to the following final result:
xi T0 rs
¼ ¼ 1þ ð13:21Þ
x0 Ti ra
Fig. 13.6 Determination of the transmission ratio by free-body force analysis, and balance of
torques (annulus gear as input, planet carrier as output, and sun gear as fixed member)
13.4 Transmission Ratio in a Planetary Gear Train 713
where xi and x0 are respectively the input and output angular velocities.
Under the same aforementioned conditions, using the second Eq. (13.16) in
which we place x1 ¼ 0, x3 ¼ xi , and x2 ¼ x0 , and bearing in mind that in this
case w ¼ ðzs =za Þ ¼ ðrs =ra Þ, we infer that the second Eqs. (13.16) and (13.21)
give the same result.
This method involves the analysis of the tangential velocity vectors on the pitch
circles of the single members of the planetary gear train, as well as on the circle
passing through the axes of the planet gears. For a description of the method, we
denote by v an arbitrary value of the tangential velocity at the point of contact
between the pitch circle of the annulus gear and the pitch circle of any one of the
planet gears of the planetary gear train examined (Fig. 13.7).
Even here we assume that the sun gear is held at rest. In this condition, the
tangential velocity of point of contact between the pitch circles of the sun gear and
the planet gear under consideration will obviously be zero. The tangential velocity
of the planet carrier at point corresponding to the axis of the planet gear is evidently
equal to v=2. The angular velocities of the annulus gear and the planet carrier can
then be determined as the quotient, respectively, of the tangential velocities v and
v=2 divided by the corresponding radii ra and ðra þ rs Þ=2. In doing so, in another
way, we get the same Eq. (13.21).
Fig. 13.7 Determination of the transmission ratio by tangential velocity vectors (annulus gear as
input, planet carrier as output and sun gear as fixed number)
714 13 Gear Trains and Planetary Gears
In addition to the three above described methods, other methods exist for the
determination of the transmission ratio of a planetary gear train. We make here a
brief reference to the tabulation method and geometrical method, illustrating the
basic concepts on which they are based. For details, we refer the reader to more
specialized textbooks (see: Müller [44], Lyndwander [35], Maitra [36], Jelaska
[24]).
The tabulation method is a summation process in which the planetary gear train
is first considered with all its members locked to one another (thus the whole gear
system rotates as if it is a single block). Successively, the same planetary gear train
is considered as an ordinary gear train (thus all the gears are free to rotate about
their own axes, and planet carrier is held at rest). Results related to these two steps
are then added and presented in a tabulation form. Compared with the algebraic
method, which provides only the final transmission ratio of the gear system, this
method has the advantage of giving a complete picture of the angular motions of the
various members, and also to provide any intermediate transmission ratio related to
them.
Ex nomine, the geometrical method is based on geometrical considerations.
According to this method, starting from a given initial position, the pitch circle of
any of the planet gears is rolled without sliding on the pitch circle of the sun gear,
assumed as fixed. The angle between the initial and new positions of the planet
carrier is then considered and, in correlation with it, the transmission ratio is
evaluated.
If we interpose between two shafts a planetary gear train, and rigidly connect to
each of them two of the three main members of the gear system, the transmission of
motion cannot occur, as the third member rotates freely. To carry out the trans-
mission of motion, it is therefore necessary that the third member is locked, that is
held stationary; by applying to it a brake-clutch, and activating or deactivating this
last, we can carry out or not the transmission of motion, at will.
Usually, we can assign to all three main members of the planetary gear train
three different functions: input of motion, output of motion, and fixed member. By
locking one of the three main members (of course one at a time), and then con-
sidering three different possible operating conditions, we can obtain six different
transmission ratios, as follows.
13.5 Transmission Ratios Achievable with Planetary Gear Trains 715
(A) With the planet carrier locked ðx3 ¼ 0Þ, we have an ordinary gear train (all
axes are fixed), and the sun gear and annulus gear (the one or the other can be
indifferently driving or driven members) rotate in the opposite direction,
realizing a reversal of the direction of rotation. From Eqs. (13.16) or (13.15)
we get the following expression of the transmission ratio:
x1 1
i ¼ i0 ¼ ¼ : ð13:22Þ
x2 w
Therefore, depending on whether we take the sun gear or the annulus gear as
driving gear 1 and driven gear 2, or vice versa, two different transmission
ratios are possible.
(B) With the sun gear locked, the annulus gear and planet carrier (again one or the
other can be indifferently driving or driven members) rotate in the same
direction, but with different angular velocities. Since x1 ¼ 0, from Eqs. (13.16)
or (13.15) we get the following expression of the transmission ratio:
x2
i¼ ¼ 1 w: ð13:23Þ
x3
In addition, in this case, depending on whether we take the annulus gear or the
planet carrier as driving gear 2 and driven gear 3, or vice versa, two different
transmission ratios are possible.
(C) With the annulus gear locked, the sun gear and planet carrier (also here the one
or the other can be indifferently driving or driven members) rotate in the same
direction, but with different angular velocities, and with ratios between these
velocities different from those obtained above for the case where the sun gear
was locked. Since x2 ¼ 0, from Eqs. (13.16) or (13.15) we get the following
expression of the transmission ratio:
x1 1
i¼ ¼1 : ð13:24Þ
x3 w
Also, in this case, depending on whether we take the sun gear or the planet
carrier as driving gear 1 and driven gear 3, or vice versa, two different
transmission ratios are possible.
We also have a seventh chance. In fact, if the gear system is characterized by a
brake-clutch which allows to rigidly connect between them any pair formed by any
two of the three main members of the planetary gear train, the whole gear system
will rotate as a single block, coming to operate as a rigid coupling. We have
therefore the possibility to realize the direct drive, i.e. the direct engagement
between the driving and driven shafts ði ¼ 1Þ. This mode of operation is common
to the three possible interlocks between any two of the three main members of a
planetary gear train. The three related trivial states of motion, mentioned by
716 13 Gear Trains and Planetary Gears
someone, are reduced de facto to a single trivial state of motion in which the gear
system rotates as a rigid coupling.
In summary, we can say that a planetary gear train provides seven combinations
of possible states of motion (they obviously become nine when the three trivial
states of motion are considered as distinct states of motion), which are derived from
the fact that, in principle, the position of the annulus gear, planet carrier and sun
gear can be locked, for which they act as a frame. The two remaining members of
the planetary gear train can be used as input or output of the power to be trans-
mitted. The ratios of the single states of motion cannot be selected independently of
each other, but are defined by the number of teeth z1 of the sun gear and z2 of the
annulus gear. Table 13.1 summarizes, as an example, the six non-trivial states of
motion of a planetary gear train such as that shown in Fig. 13.5, and provides, for
each of these states of motion, the operating conditions and transmission ratios
obtainable.
Therefore, this type of gear system makes it possible to achieve different
transmission ratios between the input and output shafts, depending on how the
individual main members are linked together and which members are in a locked
position. The members are linked together by clutches, while they are locked to the
housing by brakes. The operating condition with the sun gear locked is the one
commonly used. Depending on the relative sizes of the various members of the
planetary gear train, the transmission ratio calculated by Eqs. (13.16) or (13.15) can
take values between 1 and 2. With the annulus gear as a driving gear and planet
carrier as output shaft, the planetary gear train is often used as a speed reducing gear
drive in propeller planes. With the planet carrier as an input shaft and annulus gear
as an output gear, the planetary gear train forms the basis of the overdrive used in
automobiles [25, 43, 45].
If several planetary gear trains of the type shown in Fig. 13.5 are connected
together, the result is a compound planetary gear train of the multi-stage structure
type (see Sect. 13.3). This type of compound gear train makes it possible to obtain
different transmission ratios between input and output, which depend on how
individual transmission members are connected together, and which members are in
a fixed position. The connection between the various members of the compound
Table 13.1 States of motion of a planetary gear train such as that shown in Fig. 13.5, operating
conditions, and transmission ratio obtainable
State of motion Operating condition Input Output Frame Transmission ratio
1 Ordinary gear train 1 2 3 i¼x
x2 ¼ z1
1 z2
2 2 1 i¼x
x1 ¼ z2
2 z1
5 2 3 1 i¼x
x3 ¼ 1 þ z1
2 1
z2
6 3 2 i¼x
x1 ¼ 1 þ
3 z1
z2
13.5 Transmission Ratios Achievable with Planetary Gear Trains 717
Let z1 , z2 and z0 , respectively, the number of teeth of the sun gear, annulus gear, and
planet gears. In a simple planetary gear train, a first problem arises in relation to
these numbers of teeth, because it is possible to have n planet gears arranged at an
angular distance ð2p=nÞ from one another, so that the centrifugal forces acting on
them have zero resultant, thus being balanced.
718 13 Gear Trains and Planetary Gears
z1 þ z2 ¼ multiple of n: ð13:26Þ
In the second case (z0 , odd number), repeating the same reasoning with reference
to the diagram shown in Fig. 13.8b, we find that the condition to be satisfied is as
follows:
(a) (b)
Fig. 13.8 Conditions between the numbers of teeth to be met for: a z0 even number; b z0 odd
number
13.6 Some Problems Related to Simple Planetary Gear Train 719
A0 B000 z2 1 z1 1
¼ þ Z þ ¼ Z; ð13:27Þ
p n 2 n 2
from which we still gain the condition given by relationship (13.26). However, this
necessary condition is not sufficient. The following condition must be added to it:
z2 z1 ¼ 2z0 ¼ Z; ð13:28Þ
where Z is an even number; this condition expresses that the numbers of teeth of the
sun gear and annulus gear must be both even, or both odd.
Sometimes it happens that the planet gears are mounted on ball bearings.
A second problem arises in this case: the one related to the determination of the
number of revolutions according to which these bearings must be calculated, based
on the indications of catalogs of the manufactures. This number of revolutions,
which corresponds to the relative motion between the bearing and the pin, is
calculated by impressing to the whole gear system an angular velocity x3 , equal
and opposite to the angular velocity of the planet carrier. We have already said that,
by doing so, the planetary gear train becomes an ordinary gear train. The angular
velocities of the sun gear and annulus gear pass from the values x1 and x2 to the
values ðx1 x3 Þ and ðx2 x3 Þ; therefore, the number of revolutions required is
given by:
1 z1 1 z2
ðx1 x3 Þ ¼ ðx2 x3 Þ : ð13:29Þ
2p z0 2p z0
The ball bearing of the planet gear is loaded by a force equal to the resultant of
the two forces (they are approximately of equal value), exchanged between the sun
gear and planet gear, and between the planet gear and annulus gear (see Fig. 13.6).
This resultant force is about double of the tangential force acting on the planet gear.
Therefore, often the need occurs to increase the size of the gear planet than would
be required according to the strength calculation of the planet gear itself, so that it is
possible to house the bearing inside the coaxial hole practiced in the center of the
same planet gear.
Some planetary gear trains are examined here as examples and, for each of them,
the main characteristics are described, with reference to the conditions that must be
met in terms of numbers of teeth of the toothed members, angular distance between
the planet gears and obtainable transmission ratios. All the examples mentioned
below are mainly derived from practical applications in vehicle transmissions.
720 13 Gear Trains and Planetary Gears
As Fig. 13.9 shows, this planetary gear train is composed of two coaxial sun gears
having z1 and z4 teeth, and a given number n of planet gears, each of which consists
of two gear wheels having z2 and z3 teeth, and between them rigidly connected. The
various toothed pairs in meshing with each other have the same face width, and
their mid-planes coincide. As usual, all teeth of the toothed members have equal
module, m.
The condition that the numbers of teeth of the gear wheels must meet, resulting
from the need that the two gear pairs have the same center distance, is as follows:
z1 þ z2 ¼ z3 þ z4 : ð13:30Þ
Since the sizing we are considering is the nominal one, the condition that two
successive planet gears do not interfere with each other is expressed by the fol-
lowing relationship:
u
ðz1 þ z2 Þ sin [ z2 þ 2; ð13:31Þ
2
where u ¼ 2p=n ¼ s is the angular pitch. This angular distance is then given by
any of the following two equations, each of which is expressed in radians or
degrees:
2p 360
u¼Z ðradÞ u¼Z ð Þ ð13:32Þ
z4 z1 zz32 z4 z1 zz32
2p 360
u¼Z ðradÞ u¼Z ð Þ; ð13:33Þ
z1 z4 zz23 z1 z4 zz23
where Z is an integer selected so that the Eq. (13.31) is satisfied. Thus, we can
deduce that u must be a multiple according to an integer of the angle expressed by
the fraction that appears in the above equations. The planet gears are arranged in
pairs diametrically opposed, unless they are put at u ¼ 120 .
The determination of the transmission ratio of this planetary gear train (as well as
the determination of the transmission ratios of the planetary gear trains described in
the following sections) is performed using the Eqs. (13.23) and (13.24), in the case
in which only two members are mobile, and with Eqs. (13.16) in the case in which
all three members are mobile.
Putting ðz2 z3 Þ ¼ x, and taking account of eq. (13.30), it must also be
ðz4 z1 Þ ¼ x, and therefore:
z1 z3 z1 ðz2 xÞ 1 ðx=z2 Þ
w¼ ¼ ¼ \1: ð13:34Þ
z2 z4 z2 ðz1 þ xÞ 1 þ ðx=z1 Þ
1 ðx=z2 Þ
x2 ¼ x1 w ¼ x1 ð13:35Þ
1 þ ðx=z1 Þ
x½ð1=z1 Þ þ ð1=z2 Þ
x2 ¼ x3 ð1 wÞ ¼ x3 ð13:36Þ
1 þ ðx=z1 Þ
1 x½ð1=z1 Þ þ ð1=z2 Þ
x1 ¼ x3 1 ¼ x3 ; ð13:37Þ
w 1 ðx=z2 Þ
which we will have: x2 ¼ x3 ½1=ðz1 þ 1Þ for x1 ¼ 0, and x1 ¼ x3 ð1=z1 Þ for
x2 ¼ 0. With the first gear pair having profile shifted toothing and z1 ¼ 19 and
z2 ¼ 17, and the second gear pair having zero profile shifted toothing and z3 ¼ 17
and z4 ¼ 20, and with a ¼ 20 , the minimum gear ratio is given by x2 ¼ x3 =20, so
that the bulk of the planetary gear train will be ðz4 þ 2z3 þ 2Þm ¼ 56m.
As shows Fig. 13.10, this planetary gear train is composed of two sun gears, having
z1 and z6 teeth, and of p planet gears, each of which is composed of two rigidly
connected pairs of gear wheels having z2 , z3 , z4 , and z5 teeth. The relative position
of the two gear wheels of each pair is identical to that determined for the planet gear
of the planetary gear train described in the previous Sect. 13.7.1. All teeth of the
toothed members have equal module, m.
It must be verified that the pins of the planet gear with z2 and z3 teeth do not
interfere with the sun gear having z6 teeth. The angular distance u must correspond
to one of the values given by the following relationships:
2p 360
u¼Z z3 z5 ðradÞ ¼ Z ð Þ ð13:38Þ
z6 þ z1 z2 z4 z6 þ z1 zz32 zz54
2p 360
u¼Z ðradÞ ¼ Z ð Þ; ð13:39Þ
z1 þ z6 zz45 zz23 z1 þ z6 zz45 zz23
The numbers of teeth of the single pairs of the gear wheels that mesh between
them should be chosen so as to avoid interference between the tooth flank profiles.
The planet gears are arranged in opposed pairs along a diameter. The angle c
between the two radial straight lines passing through the axes of each of the two
rigidly connected pairs of gear wheels of the planet gears is given by the following
relationship:
is not very interesting from the point of view of practical applications, due to its
constructive complexity. The case in which z1 ¼ z6 and z2 ¼ z3 ¼ z4 ¼ z5 other-
wise makes exception; since in this case w ¼ 1, these types of planetary gear
trains with three mobile members are used as speed adders or as differential torque
splitters into two equal parts.
As Fig. 13.11 shows, this planetary gear train is composed of a sun gear having z1
teeth, an annulus gear with z2 teeth, and p planet gears, each of which is composed
of a single gear wheel having z0 teeth. All teeth of the toothed members have equal
module, m.
Considering that the pitch diameter of the annulus gear is equal to the sum of the
pitch diameter of the sun gear plus twice the pitch diameter of the planet gears, we
obtain the following relationship that correlates the number of teeth of the various
toothed members:
z2 z1 ¼ 2z0 : ð13:42Þ
Since 2z0 is an even number, we infer that z1 and z2 must be both even or both
odd. The planet gears need to be arranged at an angular distance, u, given by:
2p 360
u¼Z ðradÞ ¼ Z ð Þ; ð13:43Þ
z1 þ z2 z1 þ z2
z1 þ z2
¼ p; ð13:44Þ
Z
Substituting z0 in place of z1 in Eqs. (5.10) and (5.11), and taking into account
Eq. (13.42), for which it is z2 ¼ ðz1 þ 2z0 Þ, we get the following two relationships
that express, as a function of z1 , the minimum number of teeth, z0 , of the planet
gears which does not cause interference with the annulus gear having z2 teeth:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4 z1 sin2 a þ 16 12 sin2 a þ 4z1 sin2 a þ z21 sin4 a
z0;min ¼ : ð13:46Þ
3sin2 a
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi
4ðh1 =mÞ z21 sin2 a þ 16 12sin2 a hm1 þ 4 hm1 z1 sin2 a þ z21 sin4 a
z0;min ¼ ;
3sin2 a
ð13:47Þ
the second Eq. (13.16), which is valid in the case where its three main members are
all mobile, becomes:
z1 z1
x2 ¼ x1 þ 1 þ x3 : ð13:49Þ
z2 z2
As Fig. 13.12 shows, this planetary gear train differs from that considered in the
previous section only because each planet gear consists of two gear wheels rigidly
connected, having z2 and z3 teeth. The sun gear and the annulus gear have respectively
z1 and z4 teeth. All teeth of the toothed members have equal module, m.
Since the radii r1 , r2 , r3 , and r4 of the four gear wheels are linked by the
relationship r4 ¼ r1 þ 2r2 ðr2 r3 Þ, we obtain:
z4 ¼ z1 þ z2 þ z3 ; ð13:53Þ
for which z1 and z4 may be either odd or even. The angular distance between the
straight lines passing through the axes of the planet gears is given by one of the two
following relationships:
2p 360
u¼Z ðradÞ ¼ Z ð Þ ð13:54Þ
z4 þ z1 ðz3 =z2 Þ z4 þ z1 ðz3 =z2 Þ
2p 360
u¼Z ðradÞ ¼ Z ð Þ; ð13:55Þ
z1 þ z4 ðz2 =z3 Þ z1 þ z4 ðz2 =z3 Þ
where Z is an integer. In order that the planet gears do not interfere with each other,
Z must be selected so as to satisfy the following inequality:
u
ðz1 þ z2 Þ sin [ z2 þ 2: ð13:56Þ
2
This type of planetary gear train is much more cumbersome than the previous
one, without realizing much larger reduction ratios. For this gear train we have:
z1 z3
w¼ : ð13:57Þ
z2 z4
The various reduction ratios achievable are obtained by substituting this value of
w in Eqs. (13.35), (13.36) and (13.37), which refer respectively to cases where
x3 ¼ 0, x1 ¼ 0, and x2 ¼ 0.
As Fig. 13.13 shows, this planetary gear train consists of two annulus gears having
z1 and z4 teeth and a single planet gear, composed of two gear wheels having z2 and
z3 teeth and rigidly connected to each other. All teeth of the toothed members have
equal module, m.
For this planetary gear train, the first condition to be satisfied is given by the
following relationship:
z4 z3 ¼ z1 z2 ; ð13:58Þ
which indicates that the two internal gear pairs have the same center distance. The
second condition to be met imposes that the choice of the numbers of teeth of the
various gear wheels in meshing is carried out so as to avoid the meshing inter-
ference. For this planetary gear train, we have:
z1 z3
w¼ ; ð13:59Þ
z2 z4
and the various reduction ratios achievable are obtained by substituting this value of
w in Eqs. (13.35), (13.36) and (13.37), which refer respectively to cases where
x3 ¼ 0, x1 ¼ 0, and x2 ¼ 0. From Eqs. (13.36) and (13.37), we can deduce that,
z4 z3 ¼ y ð13:60Þ
Once we have chosen the value of y, using the diagram shown in Fig. 5.7, we
determine z3;min , and then z4;min by Eq. (13.60). The minimum values of z3 and z4 can
be also read directly on the diagram of Fig. 5.4, since they correspond to the points that
delimit, at the bottom right, the curves relating to the various pressure angles.
If we denote by x the difference between the numbers of teeth of the two gear
wheels of the planet gear, i.e. if:
z2 z3 ¼ x; ð13:62Þ
z2 ¼ z3 þ x ð13:63Þ
z4 ¼ z3 þ y ð13:64Þ
z1 ¼ z3 þ x þ y: ð13:65Þ
Therefore, the ratio w given by Eq. (13.59) can be written in the form
z3 ðz3 þ x þ yÞ
w¼ ; ð13:66Þ
ðz3 þ xÞðz3 þ yÞ
13.7 Main Characteristics of Some Planetary Gear Trains 729
and it is obviously smaller than 1 ðw\1Þ. Since y depends on the pressure angle, to
achieve large reduction ratios, the value of x should be chosen as smaller as pos-
sible, that is x ¼ 1. With reference to the angular velocities indicated in Fig. 13.13,
the transmission ratios achievable are as follows:
x2 ¼ wx1 ð13:67Þ
x2 ¼ ð1 wÞx3 ð13:68Þ
1 1
x1 ¼ 1 x3 ¼ 1 x3 ; ð13:69Þ
w w
As Fig. 13.14 shows, this planetary gear train is composed of two sun bevel gears,
having z1 and z2 teeth, and p planet gears made up of bevel gears with z0 teeth. All
teeth of the toothed members have equal module, m.
Among the numbers of teeth of the toothed gear wheels and their geometric
quantities, the following relationships exist:
R ¼ d1 þ d0 ð13:70Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4z20 ðz2 z1 Þ2
tan d0 ¼ ð13:71Þ
z1 þ z2
z1 ðz1 þ z2 Þ
tan d1 ¼ tan d0 ð13:72Þ
2z20 þ z1 ðz2 z1 Þ
z2 ðz1 þ z2 Þ
tan d2 ¼ tan d0 ð13:73Þ
2z20 z2 ðz2 z1 Þ
The lengths of the segments AO, OB and AB ¼ ðAO þ OBÞ are given by:
mz1
AO ¼ ð13:74Þ
2 tan d1
mz2
OB ¼ ð13:75Þ
2 tan d2
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
z z 2
m z1 z2 m 2 1
AB ¼ þ ¼ z20 : ð13:76Þ
2 tan d1 tan d2 2 2
If we want the pitch cone angle of the largest sun bevel gear is less than
p=2ðd2 \p=2Þ, due to Eq. (13.73) the following inequality must be satisfied:
that is
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
z2 ðz2 z1 Þ
z0 [ : ð13:78Þ
2
We will choose, therefore, a number of teeth of the planet bevel gears z0 that
satisfies inequality (13.78). The angular distance u at which the planet bevel gears
are arranged is given by:
2p 360
u¼Z ðradÞ ¼ Z ð Þ; ð13:79Þ
z1 þ z2 z1 þ z2
The various reduction ratios achievable are obtained by substituting this values
of w in Eqs. (13.35), (13.36), and 13.37). Thus, we have:
z1
x2 ¼ wx1 ¼ x1 ð13:81Þ
z2
z1
x2 ¼ ð1 wÞx3 ¼ 1 þ x3 ð13:82Þ
z2
1 z2
x1 ¼ 1 x3 ¼ 1 þ x3 ; ð13:83Þ
w z1
If none of the members of a planetary gear train is held at rest, it can become a
summarizing gearbox, a transfer gearbox, a differential drive or a differential gear
depending on how it is used. From this last point of view, we can classify the
planetary gear trains into the following three distinct families (see also Ferrari and
Romiti [14]):
• Planetary gear trains with only one input and one output; those described in the
previous section belong to this family, provided the condition of one input and
one output is respected.
• Planetary gear trains with two inputs and one output; these are the summarizing
planetary gear trains.
• Planetary gear trains with one input and two outputs; these are the transfer
gearboxes and the differential gears.
The angular velocities x1 and x5 of the sun gear, 1, and pinion, 5, are of course
equal to those of the corresponding engines, E1 and E2. The angular velocities x2
and x5 of the annulus gear, 2, and pinion 5, are then correlated by the following
relationship:
where the minus sign indicates that these two members of the planetary gear train
rotate in opposite directions, while z5 and z2 are the numbers of teeth of the pinion,
5, and respectively of external and internal teeth of the annulus gear, 2.
If, for the planetary gear train under consideration, we assume 1 and 2 as input
power members, and the planet carrier, 3, as output power member, by the third of
the Eq. (13.16) we find the following relationship which express the angular
velocity x3 of the planet carrier as function of x1 and x5 :
z1 =z2 z5 =z2 z1 z5
x3 ¼ x1 x5 ¼ x1 x5 ; ð13:85Þ
ðz1 =z2 Þ þ 1 ðz1 =z2 Þ þ 1 z1 þ z2 z1 þ z2
this equation indicates how the planetary gear train under consideration allows to
combine the angular velocities of the two engine shafts.
Now let’s consider the mechanism described above in steady-state conditions,
and assume that it works in ideal conditions, that is, without friction losses. In these
conditions, the power balance leads to writing the following relationship:
T1 x1 þ T2 x2 þ Tr x3 ¼ 0; ð13:86Þ
where: T1 is input torque applied to the sun gear, 1; T2 is the input torque applied to
the annulus gear, 2, related to the force transmitted to it by the pinion, 5; Tr is the
resistant torque, acting on the pinion 6 rigidly connected to the planet gear 3, and
applied to it by the mating gear wheel 7.
On the other hand, the torque balance equation of the system constituted by
members 1, 2 and 3 requires the following relationship to be satisfied:
T1 þ T2 þ Tr ¼ 0: ð13:87Þ
Taking into account the equations found before, which correlate the angular
velocities, we can write the following relationship:
z1 z2
T1 x1 þ T2 x2 ¼ ðT1 þ T2 Þ x1 þ x2 ; ð13:88Þ
z1 þ z2 z1 þ z2
734 13 Gear Trains and Planetary Gears
We will have:
z5
T2 ¼ T20 ; ð13:90Þ
z2
where T20 is the drive torque provided by engine E2. Therefore, assuming z1 ¼ z5 , it
follows that T1 ¼ T20 , that is, the drive torques applied to the two input shafts are
identical in steady-state conditions.
Instead, in the actual operating conditions for which frictional power losses have
to be considered, the power balance applied to the two engines, E1 and E2, and the
planetary gear train leads to writing the following relationship:
gI T1 x1 þ gII T2 x2 þ Tr x3 ¼ 0: ð13:92Þ
jTr x3 j g T1 x1 þ gII T2 x2
g¼ ¼ I : ð13:93Þ
T1 x 1 þ T 2 x 2 T1 x1 þ T2 x2
This last relationship shows that the efficiency of the compound mechanism
ðI þ IIÞ is the weighted average of individual mechanism efficiencies.
The result thus obtained corresponds to a general property of mechanical sys-
tems resulting from the union of several component mechanisms, made in one of
the two following ways:
• by linking together all the input members of the component mechanism, so as to
have a mechanical system with just one input, and as many output as there are
mechanisms;
13.8 Summarizing and Differential Planetary Gear Trains 735
Usually, in automotive engineering, the power delivered by the engine is fed to the
driving wheels on one powered axle. In vehicles with more than one powered axle,
the power must be distributed to the various powered axles. Especially in com-
mercial vehicles, the need to provide power to auxiliary units can also arise. The
mechanism that enables us to meet these needs is the differential.
In the automotive industry, a distinction is made between [31, 45]:
• the interaxle differential or transferbox, which is a mechanism intended to split
the power to more than one powered axle, in the longitudinal direction, i.e. in
the direction of travel of the vehicle;
• the interwheel differential or differential gear unit, which is a mechanism
intended to split the power to the driving wheels of one axle, in the transverse
direction with respect to the direction of travel;
• the power take-offs, which is a mechanism intended to split the power from
actual power train to auxiliary units.
The differentials can be spur gear planetary differentials or bevel gear differentials.
We give here the general concepts that apply to one and other type of differential.
However, by way of example, we focus our attention primarily on the bevel gear
differentials, which find a general use as final drives for vehicle transmissions. The
following three types are commonly used: (i), spur gear final drives; (ii), bevel gear
final drives, with helical bevel or hypoid gears; (iii), worn gear final drive (see
Naunheimer et al. [45]). Therefore, with reference to the differential shown in
Fig. 13.14, we denote by x3 and T3 the angular velocity of the driving member (the
planet carrier) and the torque acting on it, and by x1 and x2 the angular velocities
of the two driven members, and T1 and T3 the torques acting on them.
To perform its design functions, a differential must meet the following three
requirements:
1. For a given angular velocity, x3 , of the driving member, the angular velocities
x1 , x2 and x3 of the three main members of the differential must be linked by a
unique relationship, so that the difference ðx1 x2 Þ is indeterminate.
2. The driving torque, T3 , must be splitted into the torques T1 and T2 acting on the
driven members, according to a constant ratio.
3. The two torque T1 and T2 acting on the driven members must have the same
sign, that is they must operate in the same direction of travel.
736 13 Gear Trains and Planetary Gears
The epicyclic gear train meet the first two requirements described above. In fact, if
we denote by x1 and x2 the angular velocities of the two driven members, x3 the
angular velocity of the planet carrier, which is the driving member, and w ¼ 1=i0
the reciprocal of the characteristic ratio of the ordinary gear train from which the
planetary gear train comes, we get the relationship between the above three angular
velocities. As we showed in Sect. 10.4, this relationship is given by any of the three
Eqs. (13.16). Subtracting member to member, from identity x1 ¼ x1 , the two
members of the second Eq. (13.16), we can deduce the following other equation:
x1 x2 ¼ ð1 wÞðx1 x3 Þ: ð13:94Þ
This equation shows that the difference ðx1 x2 Þ can take any value.
In a mechanical system in conditions of steady-state equilibrium, the sum of
torque must be equal to zero (see Eq. 1.5); therefore, we have:
T3 ¼ T 1 þ T2 : ð13:95Þ
Furthermore, in the ideal case in which there are no losses of efficiency ðg ¼ 1Þ,
and the system is in conditions of steady-state equilibrium, by virtue of the theorem
of virtual work, also the algebraic sum of the powers that came into play must be
equal to zero (see Eq. 1.12). Therefore, we can write the following relationship:
T3 x3 ¼ T1 x1 þ T2 x2 : ð13:96Þ
From Eqs. (13.16), (13.95) and (13.96), we obtain the following expressions:
w
T1 ¼ T3 ð13:97Þ
1w
1
T2 ¼ T3 : ð13:98Þ
1w
From these equations, we can deduce that the ratio between the two torques T1
and T2 applied to the driven shafts is a constant, and equal to:
T1
¼ w: ð13:99Þ
T2
To satisfy the third condition, according to which the two torques T1 and T2 must
have the same sign, the ratio w must be negative. So we come to the conclusion that
the differentials are epicyclic gear trains with negative ratio, w.
If w ¼ 1, from the previous equation we get
x1 þ x2
x3 ¼ ; ð13:100Þ
2
13.8 Summarizing and Differential Planetary Gear Trains 737
and, in the case where there are no losses of efficiency ðg ¼ 1Þ, we have:
T3
T1 ¼ T2 ¼ : ð13:101Þ
2
n
T1 ¼ T3 ð13:103Þ
1þn
1
T2 ¼ T3 ð13:104Þ
ð 1 þ nÞ
T1
¼ n: ð13:105Þ
T2
This differential splits the driving torque in different parts between the driven
shafts. Also, in this differential, when the driven shafts rotate at the same angular
velocities ðx1 ¼ x2 Þ, the relative motions between the members of the differential
do not occur, as Eq. (13.102) shows. This type of differential can be used in the
transmissions of a vehicle with two powered axles, to split the driving torque
between the two powered axles in proportion to the weights acting on each of them,
that is according to the maximum forces of adhesion that are available.
All the equations written above are valid under the hypothesis that the losses of
efficiency are equal to zero ðg ¼ 1Þ. To take into account the losses of efficiency
that inevitably occur in the real case, it is convenient to consider the motion of the
planetary gear train, which has two degrees of freedom, as the sum of two ele-
mentary motions, which both satisfy Eq. (13.15). For example, we can consider the
actual motion as the sum of two motions. A first motion in which the three main
members of the planetary gear train rotate rigidly connected to each other, with the
same angular velocity x3 of the planet carrier; this is the so-called dragging motion,
738 13 Gear Trains and Planetary Gears
T3 x3 ¼ T1 x3 þ T2 x3 ð13:106Þ
Eliminating x3 from Eq. (13.106), we get Eq. (13.95). Instead, taking into
account Eqs. (13.12) and (13.15), from Eq. (13.107) we obtain the Eq. (13.99).
Therefore, the transmission of power between the three main members of the
planetary gear train takes place through two different flow ways, namely, in part
through the dragging motion, and in part through the relative motion.
For the determination of the actual efficiency of the gear system, it should be
noted that the losses of efficiency occur only in the relative motion, in the meshing
between the various toothed wheels that characterize the planetary gear train.
Instead, the losses of power in the dragging motion are negligible, as similar to those
that might occur in a rotating rigid coupling. Since during the relative motion the
planet carrier is locked, in Eq. (13.107) we can consider as driving member the gear
wheel 1 (in this case, the input power is given by Pri ¼ T1 xr1 [ 0, while the output
power is given by Pr0 ¼ T2 xr2 \0), or the gear wheel 2 (in this case, the input power
is given by Pri ¼ T2 xr2 [ 0, while the output power is given by Pr0 ¼ T1 xr1 \0).
In both cases, the power dissipated for the various causes of loss, Pd , is given by:
Pd ¼ ð1 g0 ÞPri ; ð13:108Þ
Pd
g¼1 ; ð13:109Þ
Pi
13.8 Summarizing and Differential Planetary Gear Trains 739
1g Pri
¼ ; ð13:110Þ
1 g0 Pi
Pri
g ¼ 1 ð 1 g0 Þ : ð13:111Þ
Pi
T1 w
¼ ð13:113Þ
T2 g0
T1
¼ wg0 : ð13:114Þ
T0
with spur gear planetary differential (Fig. 13.16b). Both of these transfer gearboxes
make it possible to equalize the speed and forces between the powered axes. With a
bevel gear differential, as we have already seen, the torque is split equally between
the front and rear axles. Instead, with a spur gear planetary differential, the torque is
split unevenly.
Here, it is not necessary to further investigate the topic concerning these transfer
gearboxes with spur gear planetary differential, also because their discussion does
not show any significant differences compared to transfer gearboxes with bevel gear
differential described above. The reader can deepen the topic with specialized
textbooks (see Naunheimer et al. [45]).
According to the definition already given in Sect. 13.3, multi-stage planetary gear
trains are compound structures containing two or more planetary gear trains
arranged in series. These compound mechanisms are widely used in semi-automatic
and fully automatic transmissions for passenger cars and commercial vehicles. To
clarify the potential of these mechanisms, we would like to examine here the
simplest case of multi-stage planetary gear train made up of two planetary gear
trains like the one shown in Fig. 13.5, arranged in series.
With reference to Fig. 13.17, and with the same notations used in Sect. 13.4.1,
the sun gears, annulus gears and planet carriers of the two component planetary
gear trains are indicated by 1, 2, and 3. One superscript and two superscripts are
used respectively to distinguish the first from the second of the two component
planetary gear trains. The notation xi (with i ¼ 1; 2; 3) is used to indicate the
angular velocities of the three main members of the two planetary gear trains, and
notation zi (with i ¼ 0; 1; 2) for number of teeth of the planet gears, sun gears, and
annulus gears. Finally, the quantities w1 ¼ 1=i01 and w2 ¼ 1=i02 , given by
(a) (b)
Fig. 13.16 Transfer gearboxes with: a bevel gear differential; b spur gear planetary differential
13.9 Multi-stage Planetary Gear Trains 741
Eq. (13.15), indicate the reciprocal of the transmission ratios of the two planetary
gear trains, which have become ordinary gear trains (i.e., with planet carriers held
at rest).
Let us introduce the following quantities:
The quantities w1 and w2 , which are negative, are then given by:
Therefore, for the two planetary gear trains under consideration, the second of
the Eqs. (13.16) becomes respectively:
x02 ¼ w01 x01 þ 1 þ w01 x03
ð13:117Þ
x002 ¼ w02 x001 þ 1 þ w02 x003 :
If a member of the first planetary gear train is connected with a member of the
second, so that these two members rotate at the same speed, a compound planetary
gear train is obtained. The six rotational speeds of the members of this compound
gear system are however correlated only by three equations. Therefore, in order to
determine the relationship between the various rotational speeds, it is necessary to
assign the speed values of three members not rigidly connected two-to-two between
each other. In this way, any of the other three members can constitute the driven
member, and its rotational speed is a linear combination of those of the three
driving members whose speeds are assigned.
x03 ¼ x003
x02 ¼ x001 ð13:118Þ
x002 ¼ 0:
In addition, the angular velocity x01 of the driving sun gear 10 is known. From the
above equations, we get the following relationship that gives the transmission ratio
of the compound gears system under consideration:
It should be noted that, when it is not possible to make the mechanical con-
nection between two members that need to be braked, two separate brakes must be
used for their braking.
The second example of connection is that shown in Fig. 13.19, which highlights
the following characteristics: members 10 and 200 as well as members 30 and 100 are
interconnected; member 20 is the driving member; member 300 is the driven member;
with only one brake, it is possible to held at rest the mechanical subset constituted
by the interconnection of the members 10 and 200 . In this case, in addition to the
Eq. (13.117), the following equalities are valid:
x01 ¼ x002
x03 ¼ x001 ð13:120Þ
x01 ¼ 0:
In addition, the angular velocity x02 of the driving annulus gear 20 is known.
From the above equations, we get the following relationship that gives the trans-
mission ratio of the compound gear system under consideration:
x002 1 þ w01 1 þ w02
¼ : ð13:121Þ
x03 w02
Several special cases of connection between the members of the two aforemen-
tioned planetary gear trains are possible. Let us consider for example the case in
which one of the two connections is made between two of the members of one of the
two planetary gear trains. The planetary gear train where two of its main members are
interconnected comes to operate as a rigid coupling, so its three main members rotate
all at the same speed. The compound planetary gear train behaves as if this planetary
gear train did not exist, so only the other planetary gear train is active.
As another example, we consider the case of two pairs of interconnected
members between the two planetary gear trains, where each pair has a member
belonging to the first planetary gear train and the other member belonging to the
second gear train. We also assume that a pair of connected members is the driving
pair, and the other is the driven pair or is held at rest. In this case, only that of the
two planetary gear trains whose third member is respectively the driven member or
the stationary member remains active with reference to power transmission.
Therefore, when operating conditions dictate that both planetary gear trains partic-
ipate in power transmission, only one pair of connected members must be the driving
pair, or driven pair, or held at rest. For example, if one of the two pairs of connected
members is the driving pair, one of the two non-interconnected members of the plan-
etary gear trains may be stationary member, and the other the driven member.
The number of transmission ratios achievable by a compound planetary gear
train consisting of two planetary gear trains arranged in series varies depending on
whether the interconnected members belong to one and the other of the two
planetary gear trains or only to one of them. With a compound planetary gear train,
with two members of the first planetary gear train separately connected with two
members of the second, 108 different transmission ratios can be achieved, each of
which can be obtained with multiple connections. All 108 transmission ratios are
functions of the quantities w01 and w02 , given by Eq. (13.115).
With the same compound planetary gear train where one of the connections is
made between two members of the same planetary gear train, only 13 different
transmission ratios can be obtained. Six of these ratios are only a function of w02 ,
since they are obtained with the first planetary gear train rotating rigidly connected
with the driving member of the second, due to one of the connections
10 30 ; 10 20 ; 20 30 . Other six of these ratios are only a function of w01 , since
they are obtained with the second planetary gear train rotating rigidly connected
with the driven member of the first, due to one of the connections
100 300 ; 100 200 ; 200 300 . The thirteenth ratio is unitary, as two members of each
planetary gear train are interconnected, and a further connection exists between the
two planetary gear trains, which makes them rotate rigidly connected.
In total, therefore, 121 different transmission ratios are achievable with the
compound planetary gear train under consideration. Here, for brevity, it is not the
case to deepen this topic further. For details, including the values in terms of w01 and
w02 of the achievable transmission ratios, we refer the reader to specialized text-
books (e.g., Pollone [51]).
13.9 Multi-stage Planetary Gear Trains 745
However, before leaving this subject, we have to point out that a compound
planetary gear train consisting of two planetary gear trains arranged in series,
automatically performs the two-way motion transmission, so it can be used as a
summarizing planetary gear system, similar to those discussed in Sect. 13.8.1. In
fact, in a compound planetary gear train in which two connections exist between the
two component planetary gear trains, the drive power is transmitted through the two
paths formed by these connections. The two examples described below clarify the
salient aspects that are of interest.
As a first example, we here consider the compound planetary gear train shown in
Fig. 13.18, and calculate the power distribution between the two paths. Let’s denote
with T10 ; T20 ; T30 ; T100 ; T200 , and T300 the torques applied respectively to members
10 ; 20 ; 30 ; 100 ; 200 , and 300 of the compound planetary gear train under the condition
shown in Fig. 13.18, and with Ti and T0 the driving and resistant torques respec-
tively applied to members 10 and 300 . The following relationships are valid between
the above torques:
Due to the connection between members 20 and 100 , the following equality is also
valid:
Bearing in mind the values of w01 and w02 given by Eqs. (13.115), for the first
planetary gear train we will have:
T100 T20
T200 ¼ 0 ¼ 0 : ð13:125Þ
w2 w2
Since the resistant or output torque, T0 , applied to the driven shaft is given by:
Since both members 30 and 300 are rigidly connected to the driven shaft, in the
ideal case where the friction losses are zero, so the efficiency of the whole planetary
gear train is equal to 1ðg ¼ 1Þ, the drive power, Pi , is divided between the two paths
according to the two following fractions:
P0 T30 1 þ w01 w02
¼ ¼
Pi T0 1 w01 w02
ð13:128Þ
P00 T300 1 þ w02
¼ ¼ :
Pi T0 1 w01 w02
1 þ w01
T30 ¼ T10 ; ð13:129Þ
w01
xi T0 1 w01 w02
¼ ¼ : ð13:131Þ
x0 Ti w01 w02
In the ideal case where the friction losses are zero, and thus g ¼ 1, the drive
power is divided between the two paths according to the two following fractions:
P0 T100 1 þ w01 w02
¼ ¼ 0
Pi Ti w1 w02
ð13:133Þ
P00 T10 1 þ w02 w01
¼ ¼ ;
Pi Ti w01 w02
13.9 Multi-stage Planetary Gear Trains 747
where Ti00 is the portion of driving torque directly transmitted to the member 100 ,
while T10 is the part transmitted through the other path.
The transmission ratio is given by:
xi T0 1 þ w01
¼ ¼ 0 : ð13:134Þ
x0 Ti w1 w02
The planetary gear systems discussed in the previous sections allow to transmit the
motion when some members of the compound planetary gear trains are held at rest
(note that in the previous figures, these members are highlighted by hatching).
Instead, when these members are free to rotate, transmission of motion cannot take
place. It is therefore necessary to equip them with brakes capable of transmitting to
them braking torques whose value depends on that of the drive torque to be
transmitted, and on the characteristics of the compound planetary gear train under
considerations. In other words, these braking torques must be of such magnitude as
to make the braking members rigidly connected with the housing. Four examples of
calculation of braking torque to be applied to the members to be held at rest are
described below.
The first example corresponds to the schematic diagram shown in Fig. 13.21,
which highlights that the two members 10 and 100 are interconnected and held at rest,
and that the driving and driven members of the compound planetary gear train in
the operating condition under consideration are respectively member 30 and
748 13 Gear Trains and Planetary Gears
member 200 . The same figure shows that it is not possible to mechanically inter-
connect the two members 10 and 100 , so to held them both at rest, it is necessary to
use two distinct brakes simultaneously.
As usual, we indicate with: z01 ; z02 ; z00 ; z001 ; z002 ; z000 ; the number of teeth of the toothed
members of the two component planetary gear trains (here too we must pay
attention to the fact that z00 and z000 indicate the numbers of teeth of the planet gears);
Ti and T0 , the driving and driven torques applied to the corresponding shafts, which
are respectively the ones of planet carrier of the first planetary gear train, and of the
annulus gear of the second planetary gear train; w01 and w02 , the quantities given by
Eq. (13.115). The braking torque to be applied to member 10 shall be:
z01 w01
T10 ¼ Ti ¼ T : ð13:135Þ
1 þ w01
i
z01 þ z02
Therefore, the torque T300 applied to planet carrier 300 will be given by:
z02 1
T300 ¼ Ti ¼ Ti ; ð13:136Þ
z01 þ z02 1 þ w01
The second example concerns the schematic diagram shown in Fig. 13.18,
which also highlights how the power transmission is carried out through two paths.
As we have seen in the previous section, the torque T300 that the planet carrier of the
second planetary gear train transmits to the driven shaft is given by the second of
Eq. (13.127), which here for convenience is rewritten in the from:
13.10 Braking Torque on Members of a Planetary Gear System Held at Rest 749
1 þ w02
T300 ¼ T0 : ð13:138Þ
1 w01 w02
Taking into account the transmission ratio, given by Eq. (13.131), the same
torque T200 will also be given by:
1 þ w02
T300 ¼ Ti : ð13:139Þ
w01 w02
Therefore, the braking torque to be applied to the annulus gear 200 of the second
planetary gear train will have to be equal to:
z002 1 1
T200 ¼ T300 ¼ T300 0 ¼ Ti 0 0 : ð13:140Þ
z001 þ z002 1 þ w2 w1 w2
The third example corresponds to the schematic diagram shown in Fig. 13.19,
which highlights that the two members 10 and 200 are interconnected and held at rest.
In this case, the mechanical interconnection between these two members is possi-
ble, so they can be braked with a single brake, which will have to develop a braking
torque equal to the sum of the braking actions that the two members must exercise.
Since the driving torque Ti applied to the annulus gear 20 of the first planetary
gear train is known, the braking torque to be applied to the sun gear 10 of the same
planetary gear train will be given by:
z01
T10 ¼ Ti ¼ Ti w01 : ð13:141Þ
z02
while the braking torque to be applied to the annulus gear 200 of the second plan-
etary gear train shall be equal to:
z002 1 1 þ w01
T200 ¼ T0 ¼ T0 0 ¼ Ti : ð13:143Þ
z001 þ z2 00
1 þ w2 w02
The fourth and last example refers to the compound planetary gear train shown
in schematic form in Fig. 13.20, whereby two-way motion transmission is evident.
In the previous section, we found that the portion of the driving torque transmitted
by the sun gear 10 of the first planetary gear train is given by the second
Eq. (13.132), which here is rewritten for convenience in the form:
1 þ w02 w01
T10 ¼ Ti ; ð13:145Þ
w01 w02
The braking torque to be applied to the annulus gear 20 must therefore be equal
to:
z02 0 1 1 þ w02
T20 ¼ T10 ¼ T1 0 ¼ T : ð13:146Þ
w01 w02
i
z01 w1
Parallel mixed power trains are gear trains that transmit power between the driving
and driven shafts through two parallel ordinary gear trains, which drive the driven
shaft through a planetary gear train. In these mixed gear drives, torque and power
are divided between two transmission paths. These mixed gear trains differ from the
summarizing planetary gear trains, discussed in Sect. 13.8.1, both because they
incorporate ordinary gear trains in addition to a planetary gear train, and because
they are characterized by only one prime mover instead of two. Three types of these
parallel mixed power trains are shown in schematic form in Fig. 13.22.
The schematic diagrams in Fig. 13.22 show that the driving shaft, connected to
the prime mover (it is not shown in the figure), drives one of the two ends of two
parallel transmission lines, consisting of as many ordinary gear trains. The other
end of each of these two transmission lines drives one of the mobile members of a
planetary gear train, which constitute the driving members of the latter.
Let w1 and w2 denote respectively the transmission ratios of the two ordinary
transmission lines. They will be positive or negative, depending on whether the
member of the planetary gear train operated by the corresponding transmission line
and the driving shaft rotate in the same direction or in the opposite direction. We
also introduce the following quantities: Ti , the driving torque applied to the driving
shaft; T1 and T2 , the two parts of driving torque, respectively operating the two
transmission lines, I and II; T10 and T20 , the torques applied to the two driving
members of the planetary gear train; T0 , the driven torque applied to the driven shaft
of the power train under consideration; xi ¼ x, the angular velocity of the driving
13.11 Parallel Mixed Power Trains 751
(a) (b)
(c)
Fig. 13.22 Three types of parallel mixed power trains: a first example; b second example; c third
example
shaft; x1 , x2 and x3 , the angular velocities of the two shafts and planet carrier of
the planetary gear train; w, the transmission ratio of the planetary gear train, in the
condition where the planet carrier is held at rest. It should be noted that, in this
discussion, we are here using the transmission function in ¼ w defined at the end of
Sect. 1.1, in place of the strictly defined transmission ratio.
Let us now examine how driving torque and driving power are shared, in the
ideal condition of zero friction losses, and therefore efficiency equal to 1ðg ¼ 1Þ. To
this end, the three examples shown in Fig. 13.22 are discussed separately.
In the first example, as shown in Fig. 13.22a, the parallel mixed power train uses
a planetary gear train characterized by an annulus gear, as the one shown in
Fig. 13.11 (planetary gear train type 3). Its sun gear is driven by the transmission
line, I, and its angular velocities is x1 ; its planet carrier is driven by the trans-
mission line, II, and its angular velocities is x3 . Being xi ¼ x the angular velocity
of the driving shaft, and w1 and w2 the two above defined transmission ratios, we
will have:
x1 ¼ w1 x
ð13:147Þ
x3 ¼ w2 x:
752 13 Gear Trains and Planetary Gears
where z1 and z2 are the number of teeth of the sun gear and annulus gear respec-
tively. The angular velocity, x0 ¼ x2 , of driven shaft of the gear system, that is the
angular velocity of the annulus gear of the planetary gear train, by virtue of the
second of the Eq. (13.16) is given by:
where
x0 x2
q¼ ¼ ¼ ww1 þ ð1 wÞw2 ð13:150Þ
xi x
T1
T10 ¼ ; ð13:151Þ
w1
Under the assumed conditions of absence of friction losses, i.e. g ¼ 1, for which
T0 x0 ¼ Ti xi , the following relationship between the input and output torques must
be also be met:
xi x Ti
T0 ¼ Ti ¼ Ti ¼ : ð13:154Þ
x0 x2 q
For this type of planetary gear train, the following relationship is obtainable:
w
T10 ¼ T20 : ð13:155Þ
1w
13.11 Parallel Mixed Power Trains 753
ww1 ww1
T1 ¼ Ti ¼ Ti
q ww1 þ ð1 wÞw2
ð13:156Þ
ð1 wÞw2 ww1
T2 ¼ Ti ¼ Ti T1 ¼ Ti 1 ;
q q
q ww1
w2 ¼ : ð13:157Þ
1w
The two parts T1 and T2 in which the driving torque is divided between the two
transmission lines vary by varying the ratios w, w1 and w2 , and therefore varying
the total transmission ratio, q. In some cases, one of these two torques can assume
higher values than the driving torque, Ti , and the other can become negative.
Reaction torques are then born in the gear system, and the transmission lines are
much more stressed than they would if each of them transmitted all the input power.
This power is divided between the two transmission lines in the same proportions as
the two torques T1 and T2 .
In the second example, as shown in Fig. 13.22b, the parallel mixed power train
also uses a planetary gear train with an annulus gear, such as the one of the first
example. However, in this second example, while the sun gear is still driven by the
transmission line I, it is the annulus gear that is driven by the transmission line II.
Furthermore, the output shaft of the mixed power train is the one of the planet
carrier of planetary gear train that rotates with angular velocity x0 ¼ x3 and to
which the resistant torque T0 is applied. Using the same notations of the first
example, some of which are shown in Fig. 13.22b, in this second example instead
of Eq. (13.147) we will have:
x1 ¼ w1 x
ð13:158Þ
x2 ¼ w2 x;
w2 ww1
x3 ¼ x ¼ xq: ð13:159Þ
1w
w2 ww1
q¼ : ð13:160Þ
1w
754 13 Gear Trains and Planetary Gears
Equations (13.151) and (13.152) are still valid, with the warning that T20 is this
time the torque applied to the annulus gear, which rotates at the angular velocity x2 .
In the usual steady-state equilibrium conditions, the algebraic sum of torques
applied to the planetary gear train must be zero, so we can write the same
Eq. (13.153), where however T20 has the different meaning just highlighted.
Therefore, given the mode of use of the planetary gear train, the following rela-
tionship can be written, which relates T10 and T20 :
1
T20 ¼ T10 : ð13:161Þ
w1
ww1 ww1
T1 ¼ Ti ¼ Ti
ð1 wÞq w2 ww1
ð13:162Þ
w2 w2
T2 ¼ Ti ¼ Ti ¼ Ti T1 ;
ð1 wÞq w2 ww1
Here too the input power is shared between the two transmission lines in the
same proportion as the input torque.
In the third example, as shown in Fig. 13.22c, the parallel mixed power train
uses a planetary gear train characterized by only external gear wheels, whose
transmission ratio with planet carrier held at rest is given by:
z1 z3
w¼ ; ð13:164Þ
z2 z4
with the same total transmission ratio, q, we will obtain the same torque distribu-
tions, in the event that for the gear system shown in Fig. 13.22c we adopt a
transmissions ratio w1 that has an opposite sign with respect to the one used for the
gear system shown in Fig. 13.22a.
All the discussion above relates to some design solutions of speed variators in
which the transmission ratio w1 of one of the two transmission lines has a constant
and predetermined value, while the other transmission ratio, w2 , is varied using a
speed variator, to vary the speed of the driven shaft, in the belief that the speed
variator can transmit only a small part of the drive power. This conviction is not
always correct since, as we have just seen, it follows that, due to the reaction
torques that arise between the two transmission lines, when the variation range of
the total transmission ratio is wide, the speed variator may be in the condition of
having to transmit powers, positive or negative, in absolute value greater than the
drive power. It is therefore preferable to adopt a simpler type of transmission, with
the speed variator directly connected to it.
For further details, we refer the reader to more specialized textbook (e.g.,
Naunheimer et al. [45]).
of an ordinary gear train. This equation is not the only one that is available to
estimate the gear train efficiency with sufficient approximation. Other equations
have been introduced, also because this subject has been studied for over a century,
even though no simple and all-encompassing theory has so far been established.
This is due to the fact that several sources of power loss arise in a power train, such
as gear mesh losses, windage and churning losses, bearing and seal losses, and oil
pump losses, and none of the so far proposed theory is able to consider all
simultaneously.
We will focus our attention on gear mesh losses, as if they were the only losses
in an ordinary gear train. As it is well known, the losses of each gear pair in mesh,
and consequently the related efficiency, depend on the manufacturing and assembly
accuracy, the lubrication conditions, as well as the number of teeth of each gear
wheel. With reference to this last aspect, Fig. 13.23, which is based on data by
Merritt [38], shows how the losses are so much higher (and therefore the efficiency
is much lower), the more the number of teeth is low. The chart of Fig. 13.23 points
out, however, that losses for a single gear pair are very small, usually about 2%
or less.
Fig. 13.23 Chart of percentage power loss for a single gear pair, as a function of the number of
teeth of the two gear wheels
13.12 Efficiency of a Planetary Gear Train 757
In accordance with the first equality of Eq. (1.4), the basic efficiency of a single
gear pair is defined as the quotient of the output power, P0 , divided by the input
power, Pi . Instead, in accordance with the last equality of the same Eq. (1.4), the
basic efficiency is defined as ð1 LÞ, where L ¼ Pd =Pi is the percentage power
loss, which can be estimated using the chart shown in Fig. 13.23. This is an initial
estimate, which can be refined using the rules outlined below. With the notations
that are relevant here (gt ¼ g0 and L), Eq. (1.4) can be written in the following
form:
P0 Pd
g0 ¼ ¼1 ¼ 1 L: ð13:165Þ
Pi Pi
For an ordinary gear train made up of two or more gear pairs, the efficiency is
given by the Eq. (13.11). Therefore, considering the simplest example of an ordi-
nary gear train consisting only of two gear pairs, we can write the following
relationship:
g0 ¼ g1 g2 ¼ ð1 L1 Þð1 L2 Þ; ð13:166Þ
which expresses the overall efficiency of the gear train under consideration as the
product of the efficiencies of the individual component gear pairs, in terms of g and
ð1 LÞ.
Expending the last equality of this equation, we get:
g0 ¼ 1 L1 L2 þ L1 L2 : ð13:167Þ
g0 ¼ 1 L1 L2 Li Ln : ð13:168Þ
• for bevel gears, the number of teeth to be used to enter the chart in Fig. 13.23
are the virtual numbers of teeth zv , calculated according to the Tredgold
approximation.
It should be noted that, within the aforementioned approximation limits, the values
of losses L, obtained using the chart in Fig. 13.23 and the above described rules, are
the same, whichever of the two members of a gear pair is the driving wheel. Thus,
the basic efficiency of an ordinary gear train does not depend on the direction of the
power flow. In addition, with regard to a reasonably accurate bearing design, this
type of calculation provides sufficient tolerance margin, which also includes bearing
losses. It should not be forgotten that the values of L so deduced are approximate,
and that the actual values depend on speeds, loads, gear materials and heat treat-
ments, tooth flank surface finishing, shaft rigidity and alignment, lubricant prop-
erties, lubrication system, etc. This is why, at present, general theoretical models
that can provide great accuracy in this regard are not available. These calculations
are still very important, especially for comparing alternative proposed designs,
while their validity is much more limited when they are used to develop existing
gear drives; in this latter case, the efficiency tests in working conditions are
unavoilable.
Let us now see how to calculate as easily as possible reliable values of the
overall efficiency of a planetary gear train. The first step to achieve this important
goal is in any case to calculate its basic efficiency, g0 . The subject has great design
interest in this specific area, as the designer must try to design planetary gear trains
with a higher efficiency than g0 , avoiding the real risk of planetary gear drives with
an efficiency so low that they are unusable. It is worth remembering that the power
losses of a planetary gear train can be surprising high, and therefore the efficiency is
too low.
Several methods have so far been introduced and used to calculate more or less
accurate values of the overall efficiency of a planetary gear train, such as:
straightforward methods, more elaborate manual methods, and systematic com-
puterized methods. To have a brief idea of these methods, we refer the reader to Del
Castillo [8], who also proposed other methods.
We here only deal this peculiar subject in accordance with the procedure pro-
posed by Molian [42], which allows to obtain sufficiently accurate values of the
efficiency of a planetary gear drive. It is obvious that all the friction power losses in
a planetary gear train occur only under conditions of relative motion among its main
members, as they are due to the rubbing of the teeth between them. When all these
members rotate like a rigid coupling, because two of its main members are rigidly
connected to each other, the power loss is theoretically equal to zero (actually, the
power loss is not equal to zero, but so small that it can to be considered equal to
zero, and therefore neglected).
In Sect. 13.4.1 we have already introduced the relative angular velocities xr1
and xr2 of the first gear wheel and last gear wheel, 1 and 2, of a planetary gear train
with respect to its planet carrier, 3; they are respectively given by:
13.12 Efficiency of a Planetary Gear Train 759
xr1 ¼ x1 x3
ð13:169Þ
xr2 ¼ x2 x3 :
Mesh power losses due to rubbing of the teeth between them can only occur
when these relative angular velocities are both nonzero. When the entire planetary
gear train rotates as a rigid coupling, we will get x1 ¼ x2 ¼ x3 , and therefore
xr1 ¼ xr2 ¼ 0. From all this, we infer that all friction power losses are proportional
to the relative angular velocities, xr1 and xr2 , rather than to the absolute angular
velocities, x1 , x2 and x3 . So, we assume that the power flow inside a planetary
gear train and friction losses can only be affected by relative motions that cause
rubbling between the tooth flank surfaces of mating teeth, but not by the gear drive
rotation as a rigid coupling.
In order to derive the expression of efficiency, it is necessary to use the two
fundamental Eqs. (1.5) and (1.2), which represent respectively the torque balance
equation and the power balance equation. To avoid misunderstanding, considering
a planetary gear train such as the one shown in Fig. 13.5, it is convenient to write
the two equations above-mentioned in the following form:
T 1 þ T2 þ T3 ¼ 0
ð13:170Þ
T1 x1 þ T2 x2 þ T3 x3 ¼ 0:
The first of these equations, i.e. the torque balance equation, points out that the
torques applied to the three main members of the planetary gear train cannot be all
positive or all negative. The usual sign convention is then valid, according to which
the torque is positive if it acts in the same direction (and thus has the same sign) of
the corresponding angular velocity, while it is negative if it acts in the opposite
direction (and hence has opposite sign) to that of the corresponding angular
velocity. In the first case, the corresponding power involved is positive ðP [ 0Þ,
and the shaft to which the torque is applied is a driving shaft, which is connected to
the driving member. In the second case, the corresponding power involved is
negative ðP\0Þ, and the shaft to which the torque (this is the resisting torque) is
applied is a driven shaft, which is connected to the driven member. The second of
the Eq. (13.170), i.e. the power balance equation, expresses energy conservation in
the ideal condition where the total frictional losses are equal to zero, so the power
entering the planetary gear drive is the same as the one that comes out of it.
In relation to the goal to be achieved, it is convenient to express the power
balance equation (the second of Eq. 13.170) by showing in it the relative angular
velocities given by Eq. (13.169). This is because, as mentioned above, friction
power losses are only proportional to these relative angular velocities. Therefore,
we get the following relationship:
760 13 Gear Trains and Planetary Gears
However, by virtue of the first of the Eq. (13.170), the sum ðT1 þ T2 þ T3 Þ is
equal to zero, so the Eq. (13.172) reduces to the following simpler form:
whose terms express respectively the power consumed by the shafts of members 1
and 2 of the planetary gear train under consideration.
Let us now consider, instead of the ideal condition of total losses equal to zero,
and hence unitary efficiency, the actual condition of nonzero total losses. We also
assume that the gear set is operating in steady-state conditions, that is at constant
speed, or that at least the speed change so slowly that changes in internal kinetic
energy of the same gear set can be ignored. Thus, while the first of Eq. (13.170)
remains valid, instead of Eq. (13.173) we will have:
or
• the planet carrier, 3, constitutes the input member, while 1 is the output member
and 2 is the fixed member or vice versa;
• the planet carrier, 3, is the member held at rest, while 2 is the input member and
1 is the output member.
Apart from the aforementioned indications for their use, in any case to be
well-attended, it should be noted that the Eqs. (13.174) and (13.175) are quite
general. In fact, they come from the two Eq. (13.170), with the only variation to
consider friction power losses; for the rest, we have not made any assumption on
which of three main members of the planetary gear train under consideration is held
at rest or not. Therefore, these equations can be used for any planetary gear set
13.12 Efficiency of a Planetary Gear Train 761
operation mode, provide we are able to identify the power flow path in relation to
the planet carrier.
The approximate method of calculating the efficiency of a planetary gear train,
proposed by Molian [42], supported by Kudryavtsev et al. [30]\with few significant
variations, and endorsed by Michlin and Myunster [40] and Klebanov and Groper
[29], is based on the use of the first of the Eq. (13.170) and one of the two
Eqs. (13.174) or (13.175), depending on the cases highlighted above. It assumes
that the geometry of the planetary gear train, whose efficiency we want to evaluate,
has already been determined on the basis of other design considerations. Therefore,
the number of teeth z1 , z2 , and z3 of the sun gear, annulus gear and planet gears are
already known, so the basic ratio, i0 ¼ 1=w, is also known.
As we have already mentioned in Sect. 13.4.1, this basic ratio should be con-
sidered as an algebraic quantity. Therefore, it is positive ði0 [ 0Þ if the directions of
rotation of the input member and output member of the planetary gear train, which
has become an ordinary gear train, coincide (see, for example, the planetary gear
train type 1, shown in Fig. 13.9). Indeed, i0 is negative ði0 \0Þ if these directions
are opposite (see, for example, the planetary gear train shown in Fig. 13.5). The
algebraic value of i0 can be easily determined by Eq. (13.19).
We then choose and introduce a plausible value of the basic efficiency, g0 ,
determined by the above-described method for an ordinary gear train, assuming that
power flows, and thus friction power losses, are equally divided between the dif-
ferent planet gears. Furthermore, we assume that the torque acting on the driven
member of the planetary gear train under consideration is known. It is also to be
stated that the Ti -torques (with i ¼ 1; 2; 3) are to be regarded as positive if i is the
driving member, whereas they are to be considered as negative if i is the driven
member or the member held at rest. For the driving member, even the angular
velocity xi (with i ¼ 1; 2; 3) must be considered as positive, since it has the same
direction of torque Ti , so also the power applied to its shaft, Pi ¼ Ti xi is positive,
that is Ti xi [ 0. Instead, for the driven member and for the member held at rest, the
angular velocities xi must be considered as negative as they are in the opposite
directions to the Ti -torques, so that the power Pi applied to the corresponding shafts
are also negative, that is Ti xi \0.
The overall efficiency of a planetary gear train depends on the power flow path
and, in turn, this is strongly influenced by the actual operating conditions of the
planetary gear set. Therefore, for the determination of the efficiency of a planetary
gear train, once its geometry is known, and the parameters dependent on it have
been calculated, the next step is to define the power flow path in the operating
conditions of the same planetary gear set.
For positive values of the basic ratio ði0 [ 0Þ, six possible cases are identifiable.
The corresponding results are shown in Table 13.2. In the columns of this table, the
following indications or quantities are listed in order, from left to right: case under
consideration, designed with a progressive number from 1 to 6; input member; fixed
member; output member; transmission ratio, defined as the quotient of the angular
velocity of the driving member divided by the angular velocity of the driven
member, and expressed both in terms of i0 and w; torques T1 , T2 , and T3 ; efficiency
762 13 Gear Trains and Planetary Gears
Table 13.2 Efficiencies in the six possible cases of a planetary gear train with i0 [ 0
Case Input Fixed Output Transmission T1 T2 T3 Efficiency
member member member ratio, i g
T3 i 0 g0 T 3 i0 g0 1
1 1 2 3 1 i0 ¼ w1
w i0 g0 1 1i0 g0
T3
i0 1
i0 g0 1
2 2 1 3 1i0
i0 ¼w1 T3
i0 g0 1
i 0 g0 T 3
1i0 g0
T3
g0 ði0 1Þ
gi00 g0 i0 i0 1
3 3 1 2 i0
i0 1 ¼ 1w
1 T2 T2
i0 T2 i0 g0
i0 g0 g0 ði0 1Þ
4 3 2 1 1
1i0
w
¼ w1 T1 gi0 T1 g0 T1 i0 g0
0
5 1 3 2 i0 ¼ w1 i0Tg2 T2 1i0 g0
i 0 g0 T2 g0
0
i0 g0
6 2 3 1 1
i0 ¼w T1 gi0 T1 g0 T1 g0
0
T3
T1 ¼
i0 g0 1
ð13:176Þ
i 0 g0 T 3
T2 ¼ ;
1 i0 g0
13.12 Efficiency of a Planetary Gear Train 763
which express the torques T1 and T2 as a function of the torque T3 and quantities i0
and g0 .
By definition, the overall efficiency, g, is given by Eq. (1.4), which we rewrite
here in more complete terms, and neglecting the subscript, t:
P0 T0 x0 T3 x3
g¼ ¼ ¼ : ð13:177Þ
Pi Ti xi T1 x1
The angular velocity x3 of the planet carrier can be determined from the third of
the Eq. (13.16). From this equation, written in terms of i0 ¼ 1=w, we get the
following relationship:
x1 i0 x2
x3 ¼ ; ð13:178Þ
1 i0
Finally, from Eq. (13.177), taking into account this last equation as well as the
first of the two Eq. (13.176), we get:
i 0 g0 1
g¼ : ð13:180Þ
i0 1
which express the torques T1 and T3 as a function of the torque T2 and quantities i0
and g0 .
764 13 Gear Trains and Planetary Gears
From this equation, taking into account the second of the two Eqs. (13.181) and
(13.178) with x1 ¼ 0, we get:
i0 1
g¼ : ð13:183Þ
g0 i 0
T2
T1 ¼
i 0 g0
ð13:184Þ
1 i 0 g0
T3 ¼ T2 :
i 0 g0
We leave the reader the solution of the other three case in Table 13.2, here not
explicitly performed.
For negative values of the basic ratio ði0 \0Þ, six other possible cases are
identifiable. The corresponding results are shown in Table 13.3. In the columns of
this table, the same indications and quantities shown in Table 13.2 are listed, in the
same order from left to right.
Here also the results shown for the cases 1, 2 and 5 of the above table are
obtained using the first of the two Eqs. (13.170) and (13.174), while those shown
for the cases 3, 4, and 6 of the same table are obtained using the first of the two
Eqs. (13.170) and (13.175). It is not the case here to dwell on these results, con-
cerning the negative values of the basic ratio, i0 , since the same considerations
apply to them, which we have made for positive values of the basic ratio.
Instead, it is very important to draw design guidelines from the detailed
examination of the results obtained above. To do this, we leave aside cases 5 and 6
shown in both tables, as they refer to conditions in which the planetary gear train
became an ordinary gear train, as its planet carrier was held at rest. Instead, we
focus our attention on the results concerning cases 1 to 4. To give a quantitative
idea of how things go, let’s consider two different values of the basic efficiency, g0 ,
and precisely g0 ¼ 95%, and g0 ¼ 80%.
13.12 Efficiency of a Planetary Gear Train 765
Table 13.3 Efficiencies in the six possible cases of a planetary gear train with i0 \0
Case Input Fixed Output Transmission T1 T2 T3 Efficiency
member member member ratio, i g
T3 i0 g0 T3 i0 g0 1
1 1 2 3 1 i0 ¼ w1
w i0 g0 1 1i0 g0
T3
i0 1
i0 1 i0 g0
2 2 1 3 i0 ¼1w g0 T3
i0 1
i0 T3
g0 i0
T3
i0 1
3 3 1 2 i0
¼ 1w
1
i0Tg2 T2 1g0 i0 g0 ði0 1Þ
i0 1 0 i0 g0 T2 i0 g0 1
i0 g0 g0 ði0 1Þ
4 3 2 1 1
1i0
w
¼ w1 T1 gi0 T1 g0 T1 i0 g0
0
5 1 3 2 i0 ¼ w1 i0Tg2 T2 1i0 g0
i0 g0 T2 g0
0
i0 g0
6 2 3 1 1
i0 ¼w T1 gi0 T1 g0 T1 g0
0
Figures 13.24a, b show, for the four cases above, the distribution curves of
overall efficiency, respectively obtained for positive and negative values of the
basic ratio, under the assumption that the basic efficiency is equal to 95%
ðg0 ¼ 0:95Þ. The examination of these curves suggests that:
(a) Cases 1 and 4 are very similar to each other and are characterized by unacceptable
low efficiencies for positive basic ratios, and relatively low efficiencies for neg-
ative basic ratios. Moreover, efficiency decreases drastically when the basic ratio
decreases, for positive basic ratios, while it increases slightly when the absolute
value of the basic ratio decreases, for negative basic ratios, until it reaches its
maximum value when the absolute value of the basic ratio is equal to 1.
(b) Cases 2 and 3 are also very similar to each other and are characterized by a
trend of efficiency distribution curves similar to that of cases 1 and 4, but with
higher efficiency values for positive basic ratios. For negative basic ratios,
however, the trend of the distribution curves shows an increase in efficiency
when the absolute value of the basic ratio increases, starting from a minimum
value corresponding to the unitary absolute value of the basic ratio.
It is, however, necessary that too hasty conclusions are not drawn, based on the
above trends of the overall efficiency distribution curves. In fact, at first glance, and
therefore without the necessary insights, it would seem that the designer should
always accord his preference to cases 2 and 3 with a negative basic ratio, since they
have the highest efficiency. However, a more comprehensive analysis should take
into account that cases 2 and 3 do not allow large overall transmission ratios (see
the fifth column in the Table 13.2). The usual compromise choices, which we have
often talked about, can lead the designer to opt for design solutions based on cases 1
and 4, which instead provide the required high transmission ratios.
Figures 13.25a, b show, for the same four cases above, the distribution curves of
overall efficiency, respectively obtained for positive and negative values of the
basic ratio, under the assumption that the basic efficiency is equal to 80%
ðg0 ¼ 0:80Þ. The examination of these curves leads to the same considerations we
have made on the distribution curves shown in the previous figure, with the fol-
lowing two variations:
766 13 Gear Trains and Planetary Gears
(a)
(b)
Fig. 13.24 Distribution curves of overall efficiency g (in %) as a function of the absolute value of
basic ratio, for cases 1, 2, 3, and 4, with basic efficiency g0 ¼ 0:95: a positive basic ratio;
b negative basic ratio
13.12 Efficiency of a Planetary Gear Train 767
• ceteris paribus, due to the lower values of basic efficiency (80% instead of
95%), the overall efficiency values are much lower, so much that a different
scale of the ordinates is needed;
• as before, here also cases 1 and 4, and cases 2 and 3, are two-to-two very
similar, but the differences between the two curves of each pair are more pro-
nounced, and overlaps are virtually nonexistent.
Here we want to analyse the efficiency of planetary gear trains capable of achieving
large transmission ratios (see also Pollone [51]. To this end, let us consider the
planetary gear train type 5 shown in Fig. 13.13. As we already showed in
Sect. 13.7.5, this planetary gear train allows for obtaining large transmission ratios
in cases where w is as close as possible to 1, and x1 or x2 are equal to zero. In this
framework, let us assume that the operating conditions of this planetary gear set is
as follows: second annulus gear held at rest by suitable braking torque, for which
x2 ¼ 0; driving member is the planet carrier, whose angular velocity is x3 ; driven
member is the first annulus gear, which rotates with angular velocity x1 , and has T0
as resisting torque applied to its shaft. Finally, let z1 , z2 , z3 , and z4 be the numbers of
teeth of toothed gear wheels, as shown in Fig. 13.13.
As we have shown in Sect. 13.7.5, the angular velocity x1 is correlated to the
angular velocity x3 and ratio w ¼ 1=i0 by the Eq. (13.69), with w given by
Eq. (13.59). The output power applied to the driven shaft is given by P0 ¼ T0 x1 .
Since the torque, T0 , is applied to the rotating annulus gear and its relative velocity
with respect to the planet carrier is given in absolute value by ðx1 x3 Þ ¼
½x1 =ð1 wÞ, the gear pair with z1 and z2 teeth behaves as if transmitting the power
P1;2 given by:
P0
P1;2 ¼ ; ð13:185Þ
1w
(a)
(b)
Fig. 13.25 Distribution curves of overall efficiency g (in %) as a function of the absolute value of
basic ratio, for cases 1, 2, 3, and 4, with basic efficiency g0 ¼ 0:80: a positive basic ratio;
b negative basic ratio
13.13 Efficiency of Planetary Gear Trains for Large Transmission Ratios 769
so, the input power applied to the planet gear with z2 teeth is given by:
" ! # " ! #
1 1 1 1
P2 ¼ P0 1 þ 1 ¼ T0 x1 1 þ 1 : ð13:187Þ
g1;2 1w g1;2 1w
Thus, the torque applied to the annulus gear with z4 teeth is as follows:
" ! # " ! #
1 1 z2 z4 1 1 1
T4 ¼ T0 1 þ 1 ¼ T0 1 þ 1 : ð13:188Þ
g1;2 1 w z1 z3 g1;2 1w w
Since the angular velocity of the same annulus gear relative to the planet carrier
is given by ðx2 x3 Þ ¼ x3 ¼ x1 =½ð1=wÞ 1, the output rolling power will be
given by:
" ! #
ð1=wÞ 1 1
P3;4 ¼ T0 x1 1þ 1
½ð1=wÞ 1 g 1w
" ! 1;2 #
P0 1 1
¼ 1þ 1 : ð13:189Þ
1w g1;2 1w
If, for simplicity of calculation, we assume that the efficiency of the gear pair with
z3 and z4 teeth is equal to that of the gear pair with z1 and z2 teeth, that is g3;4 ¼ g1;2 ,
we get the following relationship, which expresses the lost power in this gear pair:
! " ! #
1 1 1 1
Pd3;4 ¼ P0 1 1þ 1
g1;2 1w g1;2 1w
8 ! " ! #2 9 ð13:190Þ
< 1 1 1 1 =
¼ P0 1 þ 1 :
: g1;2 1w g1;2 1w ;
so, the overall efficiency of the planetary gear train under consideration, in the
operating conditions described above, will be given by:
0 12
P0 1 B 1w C
g¼ ¼" #2 ¼ B
! @ 1
C:
A ð13:192Þ
Pi 1 1 1
1þ 1 g1;2
g1;2 1w
770 13 Gear Trains and Planetary Gears
References
1. Balbayev G (2015) New planetary gearbox: design and testing. LAP Lambert Academy
Publishing, Riga, Latvia
2. Birch TW (2012) Automatic transmissions and transaxles. Pearson Education, Upper Saddle
River, NJ
3. Budynas RG, Nisbett JK (2008) Shigley’s mechanical engineering design, 8th edn.
McGraw-Hill Book Companies Inc, New York
4. Chen C, Angeles J (2007) Virtual-power flow and mechanical gear-mesh power losses of
epicyclic gear trains. J Mech Des 129(1):107–113
5. Cooley CG, Parker RG (2014) A review of planetary and epicyclic gear dynamics and
vibrations research. ASME Appl Mech Rev 66:040804-1–040804-15
6. Coy JJ, Townsend DP, Zaretsky EV (1985) Gearing, NASA Reference Publication 1152,
AVSCOM Technical Report 84-C-15
7. Davies K, Chen C, Chen BK (2012) Complete efficiency analysis of epicyclic gear train with
two degree of freedom. J Mech Des 134(7):071006
8. De Castillo JM (2002) The analytical expression of the efficiency of planetary gear trains.
Mech Mach Theor 37:187–214
9. Dhand A, Pullen K (2014) Analysis of continuously variable transmission for flywheel energy
storage systems in vehicular application. In: Proceedings of Institution of Mechanical
Engineers, Part. C: Journal of Mechanical Engineering Science, vol 229, issue 2
10. Dooner DB (2012) Kinematic geometry of gearing, 2nd edn. Wiley&Sons, New York
References 771
11. Drachmann AG (1963) The mechanical technology of Greek and Roman antiquity, a study of
the literary sources. Munksgaard, København
12. Dudley DW (1962) Gear handbook. the design, manufacture, and applications of gears.
McGraw-Hill, Book Companies, Inc., New York
13. Essam LE, Isam EI (2018) Influence of the operating conditions of two-degree-of-freedom
planetary gear trains on tooth friction losses. J Mech Des 140(5)
14. Ferrari C, Romiti A (1966) Meccanica Applicata alle Macchine. Unione Tipografico-Editrice
Torinese (UTET), Torino
15. Fisher R, Kücükay F, Jürgens G, Najork R, Pollak B (2015) The automotive transmission
book. Springer International Publishing Switzerland, Cham, Heidelberg
16. Galvagno E (2010) Epicyclic gear train dynamics including mesh efficiency. Int J Mech
Control 11(2):41–47
17. Genta G, Morello L (2007) L’Autoveicolo: Progetto dei componenti, vol 1. ATA-
Associazione Tecnica dell’Automobile, Orbassano (TO)
18. Giovannozzi R (1965) Costruzioni di Macchine, vol II, 4th edn. Casa Editrice Prof. Riccardo
Pàtron, Bologna
19. Gupta AK, Kartik V, Ramanarayanan CP (2015) Design, development and performance
evolution of a light diesel hybrid electric pickup vehicle using a new parallel hybrid
transmission system. In: Transportation Electrification Conference (ITEC), IEEE International,
Chennai, India, 27–29 Aug 2015
20. Gupta AK, Ramanarayanan CP (2013) Analysis of circulating power within hybrid electric
vehicle transmissions. Mech Mach Theor 64:131–143
21. Henriot G (1979) Traité théorique and pratique des engrenages, vol 1, 6th edn. Bordas, Paris
22. Henriot G (1990). Gears and planetary gear trains. Reggio Emilia: Brevini (ed)
23. ISO 1122-1:1998. Vocabulary of gear terms—Part. 1: Definition related to geometry
24. Jelaska DT (2012) Gears and gear drives. Wiley, New York
25. Juvinall RC (1983) Fundamentals of machine component design. Wiley, New York
26. Juvinall RC, Marshek KM (2012) Fundamentals of machine component design, 5th edn.
Wiley, New York
27. Kahraman A, Hilty DR, Singh A (2015) An experimental investigation of spin power losses
of a planetary gear set. Mech Mach Theory 86:48–61
28. Kelley OK (1973) Design of planetary gear trains, Chapter 9 of design practices-passenger car
automatic transmissions. SAE-Society of Automotive Engineers, New York
29. Klebanov BM, Groper M (2016) Power mechanisms of rotational and cyclic motions. CRC
Press, Taylor&Francis Group, Boca Raton, Florida
30. Kudryavtsev VN, Derzhavets YA, Gluharev EM (1994) Calculation and design of gear
transmissions-handbook. Politeknika Publishing, St. Petersburg (in Russian)
31. Lechner G, Naunheimer H (1999) Automative transmissions: fundamentals, selection, design
and applications. Springer, Heidelberg
32. Litvin FL, Fuentes A (2004) Gear geometry and applied theory, 2nd edn. Cambridge
University Press, Cambridge, UK
33. Litvin FL, Fuentes A, Vecchiato D, Gonzalez-Perez I (2004) New design and improvement of
planetary gear trains, NASA Technical Report, NASA/Cr-2004-213101, ARL-CR-0540
34. Liu J, Peng H (2010) A systematic design approach for two planetary gear split hybrid
vehicles. Veh Syst Dyn 48(11):1395–1412
35. Lyndwander P (1983) Gear drive system design and application. Marcel Dekker Inc,
New York
36. Maitra GM (1994) Handbook of gear design, 2nd edn. Tata McGraw-Hill Publishing
Company Ltd, New Delhi
37. Merritt HE (1954) Gears, 3th edn. Sir Isaac Pitman&Sons, London
38. Merritt HE (1972) Gear engineering. Wiley, New York
39. Meyer PB (2011) The Wilson Preselector Gearbox, Armstrong-Siddley Type. pbm verlag,
Seevetal, Hamburg, Germany
772 13 Gear Trains and Planetary Gears
40. Michlin Y, Myunster V (2002) Determination of power losses in gear transmissions with
rolling and sliding friction incorporated. Mech Mach Theor 37:167–174
41. Miller JM (2006) Hybrid electric vehicle propulsion system architectures of the e-CVT type.
IEEE Trans Power Electron 21(3):756–767
42. Molian S (1982) Mechanism design: an introductory text. Cambridge University Press,
Cambridge
43. Morelli A (1999) Progetto dell’autoveicolo: concetti di base. Celid, Torino
44. Müller HW (1982) Epicyclic drive trains: analysis, synthesis, and applications. Wayne State
University Press, Detroit
45. Naunheimer H, Bertsche B, Ryborz J, Novak W (2011) Automatic transmission: fundamen-
tals, selection, design and application, 2nd edn. Springer, Heidelberg
46. Niemann G, Winter H (1983) Maschinen-Element, Band II: Getriebe allgemein,
Zahnradgetriebe-Grundlagen, Stinradgetriebe. Springer, Berlin
47. Pastore G (2013). Il Planetario di Archimede Ritrovato, and The Recovered Archimedes
Planetarium. Pastore Ed, Rome
48. Paul B (1979) Kinematics and dynamics of planar machinery. Prentice Hall, New Jersey
49. Pearson T (1991) Planetary gear. Roof Books, New York
50. Pennestrì E, Mariti L, Valentini PP, Mucino VH (2012) Efficiency evaluation of gearboxes for
parallel hybrid vehicles: theory and applications. Mech Mach Theory 49:157–176
51. Pollone G (1970) Il veicolo. Libreria Editrice Universitaria Levrotto&Bella, Torino
52. Ptolemy C (2014) The Almagest: introduction to the mathematics of the heavens, Translated
by Bruce M. Perry, William H. Donahue (ed). Green Lion Press
53. Radzevich SP (2016) Dudley’s handbook of practical gear design and manufacture, 3rd edn.
CRC Press, Taylor&Francis Group, Boca Raton, Florida
54. Ruggieri G (2003) Rotismi Epicicloidali. McGraw-Hill Companies, srl, Milano
55. Simionescu PA (1998) A unified approach to the assembly condition of epicyclic gears.
ASME J Mech Des 120(3):448–453
56. Smith JD (1983) Gears and their vibration: a basic approach to understanding gear noise.
Marcel Dekker Inc, New York
57. Uicker JJ, Pennock GR, Shigley JE (2003) Theory of machines and mechanisms. Oxford
University Press, New York
58. Warwick A (2014) Who invented the automatic gearbox. North West Transmissions Ltd.,
(ed). Retrieved 11 Oct
59. Willis R (1841) Principles of mechanism. Longmans-Green&C, London
60. Wright MT (2007) The Antikythera mechanism reconsidered. Interdisc Sci Rev 32(1):27–43
61. Xue HL, Liu G, Yang XH (2016) A review of graph theory approach research in gears.
In: Proceedings of the Institution of Mechanical Engineers, Part. C: Journal of Mechanical
Engineering Science, vol 230, issue 10
62. Yang F, Feng J, Zhang H (2015) Power flow and efficiency analysis of multiflow planetary
gear trains. Mech Mach Theor 92:86–99
Chapter 14
Face Gear Drives
Abstract In this chapter, the ever-increasing importance of face gear drives is first
underlined, and short notes on their cutting processes are provided. The generation
processes and geometrical elements of a face gear pair are then defined, and the
basic topics of face gear manufacturing and bearing contact are given. The basic
equations of the tooth flank surfaces of these gears are then obtained using dif-
ferential geometry methods, and the non-undercut conditions of the tooth surfaces
and those for avoiding pointed teeth are also defined. The correct meshing and
contact conditions between the two members of a face gear pair are then studied and
analyzed using the same methods, and the point contact conditions are determined.
Finally, the generation and grinding processes of face gears by worms are deepened,
always using the aforementioned analytical procedures.
According to the definition of ISO 1122-1 [15], a face gear, or contrate gear, is a
bevel gear with tip and root angles of 90° (Fig. 14.1). The same ISO defines as face
gear pair, or contrate gear pair, a gear pair consisting of a face gear and its mating
pinion, with either intersecting or crossed axes, with a shaft angle equal to 90°.
Figure 14.2 shows a conventional face gear pair formed by a spur involute pinion
(this can also be a cylindrical helical pinion) and a conjugate face gear (see Feng
et al. [8], Kawasaki [16], Litvin et al [23], Townsend [35]).
Face gear drives are among the oldest types of gears of which we have historical
memory, but also they can be considered as futuristic gears, because we do not
know well yet all their features, and above all their potential in relation to possible
practical applications, in competition with other types of gears. These types of gear
drives, albeit in a rudimentary form, have been used by the Chinese (see, for
example, the chariot compass, described in Vol. III, concerning A concise history
of gears), as well as in ancient Egypt, Mesopotamia, and in the Roman world, for
waterwheels, watermills and windmills. A face mill pair, consisting of a face gear
Fig. 14.2 Face gear pair formed by a spur involute pinion and a mating face gear, with
intersecting axes and shaft angle R ¼ 90
with pins, and a lantern pinion, in fact characterized these ancient practical
applications.
The revival of interest in these types of gears came from the middle of the past
century [3, 5, 6, 9]. This interest was intensified towards the end of the same
14.1 Introduction and General News 775
century, and the beginning of the 21th century [1, 2, 10–13, 17, 19, 20, 25, 29, 33,
40]. This interest is also witnessed by specific research programs, such as the
program ART (Advanced Rotorcraft Transmission) of the US Army, USA, and
BRITE EURAM FACET (development of face gears for use in aerospace trans-
mission) in Europe.
The reasons of this delayed interest on these types of gears are to be found in the
fact that the benefits of using face gears in the helicopter transmissions also began
to be obvious with some delay. Early studies on the subject clearly showed that the
use of face gear technology would have allowed a power density improvement and
lower cost when applied to a helicopter transmission. In fact, the ability of face gear
pairs to provide high reduction ratios as well as a self-adjusting torque splitting,
allows to replace conventional multi-stage gearboxes with gear units requiring a
lower number of stages. This leads to a better power/weight ratio, a small number of
parts of the gearbox, and a reduction of volume. In addition, the split torque design
of this type of transmission offers improved reliability and reduced costs against
existing conventional gearing designs used in large horsepower applications.
A helicopter transmission must meet, among other things, the following two
requirements: it must allow a split of torque from the engine shafts to the combining
gear; it must be able to ensure a drastic speed reduction from about 25 103 rpm
(*2.618 rad/s) of the engine shafts to about 300 rpm ( 31,4 rad/s) of the main
rotor. The combining gear is rigidly connected with the input member, the sun gear,
of a planetary gear train, whose output is the planet carrier, which drives the main
rotor shaft. Figure 14.3 highlights the concept of split of torque. Figure 14.3a
shows a more traditional design solution, which uses two spiral bevel pinions
forming a single rigid body, which drive two mating spiral bevel gear wheels that
transmit power to the combining gear. Figure 14.3b shows instead a far more
advanced design solution, in which a single spur (or helical) pinion is in mesh with
two face gears. In both figures, the transmission part made up of the planetary gear
train, which drives the main rotor, is omitted; in this regard, see Litvin et al. [29].
The same Fig. 14.3 shows that the solution with face gears has at least two main
advantages (with all that follows) over the one utilizing spiral bevel gears. On the
one hand, the torque split configuration allows to transmit reduced loads on bear-
ings that support the various shafts, and, on the other hand, the pinion is single and
consists of a simple spur (or helical) pinion, compared to the double spiral bevel
pinion, which in itself is much more complex. It results in greatly reduced weight
and cost. This is another very important design goal, also because the transmission
is called to support the weight of the entire aircraft, including that of several flight
accessories.
Now, we focus our attention on the face gear pair. The axes of its two toothed
members (pinion and face gear) generally, but not necessary, intersect under a right
angle. When this configuration is made, the face gear pair is called on-center face
gear pair; instead, if the axes do not intersect (this is a much rarer case in practical
applications, also because it is much more complex), the face gear pair is called
off-set face gear pair. We here limit the discussion only to face gear pairs with
776 14 Face Gear Drives
(b)
Combining gear
Face gear
intersecting axes, that is to on-center face gear pairs. Below, in the absence of
further details, for face gear pairs we mean those with intersecting axes.
The peculiarities of an on-center face gear pairs (as well as those of an off-set
face gear pair) are closely related to the generation process of its two toothed
members. We already know very well the generation process of the spur (or helical)
pinion. For the generation of the mating face gear, a generation shaper is usually
used. This shaper, with its relative motions compared to the face gear to be gen-
erated, must simulate the engagement of the actual pinion with the same face gear.
In this generation process, the surfaces of the teeth of shaper and face gear are in
line contact of all times, just as the tooth surfaces of the actual face gear pair (see
Fig. 14.2). However, when the shaper is exactly identical to the actual pinion (with
the obvious exception of the cutting edges), the face gear pair thus generated
becomes sensitive to misalignment, hence an undesirable shift of the bearing
contact and also separation of surfaces, i.e. absence of contact, may occur.
Therefore, to eliminate or decrease such a misalignment sensitivity, which is
proper to an on-center face gear pair, and is due to line contact, it is necessary to
make the instantaneous contact between the tooth surfaces of the pinion and mating
face gear become a point contact. In fact, with a localized contact, theoretically
limited to one point, the face gear pair is less sensitive to misalignment. This
important goal is achieved by using a shaper with a number of teeth z0 [ z1 , where
z1 is the number of teeth of the pinion. To this end, it is generally enough that the
difference ðz0 z1 Þ is equal to ð2 3Þ. For an off-set face gear pair, this attention is
not necessary, since contact is for itself a point contact; in this case, however, the
operating conditions are worse, because they are characterized by rolling and
considerable sliding.
14.1 Introduction and General News 777
Fig. 14.4 Geometry of a face gear tooth: a fillet surface, and working surface, with its contact
lines; b cross sections of a face gear tooth
778 14 Face Gear Drives
Re ¼ ð1:10 1:20Þz2 mn =2 and Ri ¼ ð0:95 1:05Þz2 mn =2, where the higher and
lower values are to be selected, respectively, for low gear ratios (about u ¼ 1:5), and
for large gear ratios (about u ¼ 8); b ¼ ðRe Ri Þ ffi 0:07z2 mn . Other scholars
express the permissible face width of the face gear as a function of the module, m, by
the relationship b ¼ ðRe Ri Þ ¼ km, where k is a coefficient that mainly depends on
the gear ratio, u. Usually, k has values within the range ð8 k 15Þ. If the pinion
was realized with transverse crowning, the corresponding face gear pair would be
insensitive to misalignment, but would have a load carrying capacity a little less.
As long as adequate cutting process were not available, face gears were not con-
sidered suitable for transmitting high power. The limited use of these types of gear
drives, which lasted almost to the end of the last century, can be attributed both to
the unavailability of technologically and economically viable manufacturing pro-
cesses, capable of meeting the demanding requirements of high-power applications,
and to the insufficient knowledge of three-dimensional geometry of their tooth
surfaces, and their meshing properties. The grinding processes of face-gear teeth
(which could have improved the accuracy grade by promoting their use for aero-
space applications) were even more inadequate than cutting processes.
Of course, no difficulty exists for grinding the spur involute pinion. Significant
difficulties existed, however, for grinding the face gear; it was therefore not possible
to harden its tooth surfaces, which could have resulted in a substantial increase in
permissible contact stresses. A first attempt to develop a face-gear manufacturing
process is certainly that made by Miller [31], who proposed a method of generating
a face gear by a hob. Due to its not a few limitations (see Litvin and Fuentes [24]),
this method did not have much luck, but had the merit of drawing the attention of
scholars and experts, and launching a research line on the subject.
These researches were intensified towards the end of the last century and the
beginning of this century, and culminated with the substantial contribution of Litvin
and his followers (see Litvin et al. [20, 22, 26–29]). This contribution is to be
considered as a substantial step in development of the grinding and cutting pro-
cesses of face gears. These processes use a worm for grinding, and a hob for
generation cutting, both conceived and developed according to a common idea,
based on the use of a worm having a special shape. This method therefore has the
advantage of being able to be used not only for grinding, but also for the generation
of face gears. This is because the related bases allow us to define the geometry of a
suitable worm by which we can configure the tools needed to perform both oper-
ations, grinding and generation cutting, to be made respectively with grinding
wheel or hob. Thus, many advantages are obtained with respect to conven-
tional cutting shaper. For grinding, a worm with only one thread is used, but the
number of turns of this thread has to be limited to avoid the appearance of
singularities on the worm surface.
14.2 About the Face Gear Cutting Process 779
In the previous section we have already said that the conventional face-gear gen-
eration process is based on the use of an involute shaper, and on the meshing simu-
lation of this shaper with the face gear to be generated. Here we will focus our attention
on the concept of generation of face gears by grinding or cutting worm. The discussion
is in line with the afore-mentioned method of generating face gears by a worm, which
was development by Litvin and his followers, and therefore called from here on
Litvin’s method. This is a general method as it can be applied both for grinding the face
gears, using a grinding worm, and for generating the face gear by a hob used as
generation wheel. However, we cannot forget another major contribution, developed
by Tan [34], which provides an improved method of dressing the grinding worm,
which offers many unique features and advantages (see Heath et al. [13]). In the
application of the Litvin’s method, we will consider this important contribution.
Figure 14.5 shows in schematic form the grinding operation of a face gear
according to the Litvin’s method, which also incorporates the contribution of Tan.
In this figure, an imaginary pinion, a face gear, and a grinding wheel are mutually
overlapped with respect to each other in their meshing positions. The dressing (or
truing) tool is also shown in the same figure. The grinding worm has a thread
geometry that, in synchronous rotation with the face gear to be generated, must
simulate the generation action of the imaginary pinion in rotation.
Fig. 14.5 Face gear grinding configuration, and dressing configuration with respect to pinion and
face gear as references
780 14 Face Gear Drives
The installation of this grinding worm on the face gear must therefore take into
account the lead angle cw of its thread, which is given by:
zw dv1
sin cw ¼ ; ð14:1Þ
zv1 dw
where zw and zv1 are respectively the number of threads of the worm and the
number of teeth of the imaginary pinion (usually zw ¼ 1), while dv1 and dw are
respectively the pitch diameter of the imaginary pinion and the diameter of the
circle passing through the pitch point on the grinding wheel (see also next section).
The generation motion is relatively easy to get. Instead, it is much harder to
obtain the correct shape of the grinding worm (and so the one of the generation
hob). Often this step is the keystone and the biggest challenge in the development
of face gears manufacturing processes. For dressing the grinding worm, Litvin uses
a tool whose geometry follows that of the tooth space of the imaginary pinion. Tan
has instead proposed a dressing tool configured to disk shape, having a plane as
generation surface. This extremely simple geometry, besides being very easy to
prepare with the desired high accuracy, has the inherent advantage of absorbing
installation and motion errors. This is because any component of error contained in
the generation plane has no effect on the generation action, and position errors of
constant amount along the plane normal, which often occur, produce deviations in
the generated surface that are of constant magnitude along the local surface normal.
However, it is well known that surface deviations having constant amount in the
direction of local surface normal have significantly alleviated damaging effects, as
the relative errors are substantially reduced.
Another important feature of the dressing method proposed by Tan is its true
conjugate action. In fact, the installation and motion of the dressing tool compared
to the grinding worm can be specified with reference to the imaginary involute
pinion. The dressing tool is positioned in such a way that it is always in tangency
with the pinion tooth surface. The positioning of this dressing tool can be repre-
sented in terms of some geometric parameters, so it is possible to define accurately
the kinematic quantities that describe the relative motion between grinding worm
and face gear to be obtained.
For further details on this important subject, which made it possible to use face
gears for high power transmissions, we refer the reader to the aforementioned paper
by Heath et al. [13].
In evaluating the meshing conditions of a face gear pair, some of its major geo-
metric elements come into play, such as reference surfaces, pitch surfaces (or
axodes), and pitch point. For a correct definition of these geometric elements,
14.3 Geometric Elements of a Face Gear Pair, and Its Generation 781
where xi and zi (with i ¼ 1; 2) are respectively the angular velocities and the
number of teeth of the pinion (i ¼ 1) and the bevel wheel (i ¼ 2). The reference
cones of these two members are the cones having cone angles d1 and d2 , given by
the Eqs. (9.20) and (9.21). These two cones touch each other along the common
tangent line, OB, which is the instantaneous axis of rotation in their relative motion
(Fig. 9.1). Therefore, they are also the pitch cones (or axodes) of the bevel gear
drive. In fact, by definition, axodes are the loci of the instantaneous axes of rotation
generated in coordinate system Si (with i ¼ 1; 2) rigidly connected with the pinion,
1, and wheel, 2. In other words, for a straight bevel gear pair with zero
profile-shifted toothing, reference and pitch surfaces coincide.
However, in our case, the face gear pair consists of a cylindrical spur pinion and
mating face gear. As shown in Fig. 14.6, the reference surfaces are, for the pinion,
the cylinder of radius r1 , and, for the face gear, the cone having cone angle R,
which coincides with the shaft angle. These reference surfaces touch each other
along the common tangency line O0 t0 , which is the reference line. In the case of
intersecting orthogonal face gear pair, for which the shaft (or crossing) angle is
R ¼ 90 , the reference surface of the face gear is a plane. However, under the
operating conditions, the two members of the face gear pair under consideration
touch each other along the instantaneous line of contact Ot, which is the instan-
taneous axis of rotation in relative motion. Therefore, the pitch surface (or axodes)
of the face gear pair are the two cones having cone angles d1 and d2 , with
ðd1 þ d2 Þ ¼ R. Thus, reference and pitch surfaces are no longer coincident, but
different.
The point of intersection, C, of the reference line, O0 t0 , with the instantaneous
axis of rotation Ot, is what we call pitch point. The location of the pitch point along
the instantaneous axis of rotation is of great importance as it affects the size of the
meshing area as well as the hazards associated with the pointed teeth. At pitch
point, the relative motion is pure rolling, while at any other point of the reference
line, O0 t0 , and instantaneous line of contact, Ot, the relative motion is a combined
rolling and sliding motion.
As we have already anticipated in Sect. 14.1, the conventional method of gen-
eration of the face gear uses an involute shaper as a cutting tool, with a number of
teeth, zs , which exceeds two or three teeth the number of teeth, z1 , of the pinion. In
this generation process, the shaper and face gear rotate about their respective
intersecting axes with angular velocities xs and x2 , which are correlated by the
relationship (see also Fig. 14.7):
xs z2
¼ ; ð14:3Þ
x2 zs
If the face gear was generated by a shaper that, apart from the cutting edges, was an
identical duplication of the pinion, its generation process would be an exact sim-
ulation of the meshing between pinion and face gear, so we would have a line
contact at every instant. However, as we have already said, such a generation
process cannot be practically used because of the sensitivity to the misalignment of
the face gear pair made through it. This sensitivity is then exacerbated by assembly
and manufacturing errors, which can cause separation of the contacting surfaces or
undesirable contact at the edge.
We have also said that, to eliminate the above sensitivity to misalignment, the
bearing contact between pinion and mating face gear must be a localized contact
and this goal can only be achieved if the generation process is able to provide a
point contact between the tooth flank surfaces of the two members that form the
face gear pair. To achieve this goal, it is sufficient that the face gear is obtained as
the envelope to the family of surfaces of an involute shaper having a number of
teeth zs [ z1 , with the difference ðzs z1 Þ equal to 2 or 3.
To analyze the contact conditions occurring during the manufacturing process,
made using an involute shaper, we introduce an imaginary internal gear pair con-
sisting of the pinion, 1, whose pitch cylinder has radius r1 , and shaper, s, whose
pitch cylinder has radius rs . This imaginary internal gear pair is shown in
Fig. 14.10, exaggerating the differences between the radii of these two members,
one with external teeth, and the other with virtual internal teeth, in order to better
highlight the concept involved here.
The pitch cylinders of radii rs and r1 , which are the axodes in the shaper and
pinion meshing conditions, are tangent along a common tangent, ts1 , which is
parallel to the respective axes of rotation. This common tangent is the pitch line,
that is, the instantaneous axis of rotation in the relative motion of the shaper with
respect to the pinion. The intersections of the pitch cylinder, pitch line, and axes of
rotation with a transverse plane are respectively the pitch circles, pitch point, C, and
centers Os and O1 , shown in Fig. 14.10. This figure also shows the two coordinate
systems, Os ðxs ; ys ; zs Þ and O1 ðx1 ; y1 ; z1 Þ, rigidly connected to shaper and pinion
respectively, whose zi -axes coincide with their axes of rotation, and form right-hand
coordinate systems with the corresponding axes xi and yi (with i ¼ s; 1).
By generalizing the problem, we now consider not only the engagement of two
members of the imaginary internal gear pair, as defined above, but also the
engagements between the shaper and face gear to be generated, and between the
pinion and generated face gear. By doing so, we come to consider three surfaces,
rs , r1 , and r2 , which are respectively the tooth flank surfaces of shaper, pinion, and
face gear. These three surfaces are in mesh simultaneously. During the process of
generation of the face gear by shaper, surfaces rs and r2 are in line contact at every
instant, along the instantaneous axis of rotation, ts2 . During the process of imagi-
nary meshing between the shaper and pinion, surfaces rs and r1 are in line contact
at every instant, along the instantaneous axis of rotation ts1 . Finally, in actual
operating conditions, surface r1 and r2 are in point contact at every instant at point
C (the pitch point), which belongs to the instantaneous axis of rotation, t12 , in the
relative motion of the pinion with respect to the face gear.
Figure 14.11 summarizes everything we have just said. It shows the location of
the three instantaneous axes of rotation during the two-to-two meshing of the three
surfaces rs , r1 , and r2 . The instantaneous axis of rotation, ts1 , in the meshing of the
imaginary internal gear pair, is parallel to the axes of rotation of shaper and pinion.
The other two instantaneous axes of rotation, ts2 and t12 , are oriented with respect to
the direction of ts1 by angles ds and d1 . Angle d1 that is formed between the pinion
axis and t12 is given by Eq. (9.20). Angle ds that is formed between the shaper axis
and ts2 is given by the following equation:
sin R sin R
tan ds ¼ ¼ ; ð14:4Þ
us2 þ cos R ðz2 =zs Þ þ cos R
ðz2 z1 Þ
a ¼ ðrs r1 Þ ¼ m: ð14:5Þ
2
It is very interesting to see how the contact lines are localized and oriented on the
shaper tooth flank surface. Figure 14.12a shows the contact lines, Ls2 , corre-
sponding to the meshing of the shaper with the face gear; they are differently
inclined with respect to the generatrixes of the shaper tooth surface. Figure 14.12b
shows the contact lines, Ls1 , corresponding to the meshing of the shaper with the
pinion; these contact lines are all parallel to the generatrixes above. Finally,
Fig. 14.12c shows the current instantaneous point of tangency, P, between surfaces
r1 and r2 , also represented on the shaper tooth surface. This point is the intersection
point of respective current contact lines Ls1 and Ls2 .
(a) (b)
(c)
Fig. 14.12 Contact lines on the shaper tooth flank surface: a contact lines, Ls2 ; b contact lines,
Ls1 ; c current instantaneous point of tangency of surfaces r1 and r2 , obtained as intersection point
of respective current contact lines Ls1 and Ls2
14.4 Basic Topics of Face Gear Manufacturing and Bearing Contact 787
The tooth flank surfaces of the face gears, which are of interest here, are those
obtained under cutting conditions. Of course, these surfaces are strongly related to
the shape of the cutting tool edge used in their manufacturing process and, more
generally, to the geometry of tooth flank surfaces of the same cutting tool. This is
because the face gear tooth surfaces are obtained as the envelope of the cutting tool
surfaces, used for this purpose.
Here we will only consider the case of face gears generated by an involute
shaper, following the theoretical development proposed by Litvin and Fuentes [24].
However, it is to be remembered that Litvin and his followers have further
developed the theory, extending it to the face gears generated by shapers conjugate
to parabolic rack cutters as well as face gears generated by involute helical shapers
(see Litvin et al. [26–28]). We limit our discussion here, circumscribing it to the
simplest case of face gears generated by an involute shaper, which is already quite
complex for itself. The discussion is, however, of a very general nature, so that,
with the proper variations, it can easely be extended to any geometry of the tooth
profiling surface of shaper.
In this framework, we must first write the equation of the involute surface of the
spur involute shaper. To this end, let us consider the shaper transverse section
shown in Fig. 14.13, and assume, as coordinate system, the system Os ðxs ; ys ; zs Þ
rigidly connected to it. This coordinate system has its origin at the crossing point
O Os O2 between axes of shaper and face gear (see Fig. 14.7). Its zs -axis is
coinciding with the shaper axis, and directed as Fig. 14.7 shows, while its xs -axis is
coinciding with the straight line that symmetrically divides space width and is
directed from the origin Os to the pitch point C (this axis is the trace of the plane of
equation ys ¼ 0, in Fig. 14.13). The ys -axis completes the right-hand coordinate
system.
In the transverse section shown in Fig. 14.13, i.e. in the Cartesian plane
Os ðxs ; ys Þ, the position of a current point P on the involute profile I is defined by the
position vector Os P ¼ Os T þ TP; the absolute value of this last vector is given by
TP ¼ TP d0 ¼ rbs #s , where rbs is the radius of the shaper base circle, and #s is the
involute roll angle (see Sect. 2.1). Considering also the current point P0 on the
involute profile II, which is symmetric of point P with respect to the xs -axis, we
obtain the following system of Cartesian equations that define the position of these
two points in the Cartesian plane Os ðxs ; ys Þ:
788 14 Face Gear Drives
xs ¼ rbs cos #s þ #0s þ #s sin #s þ #0s
ð14:6Þ
ys ¼ rbs
sin #s þ #0s #s cos #s þ #0s ;
where #0s is space width half angle on the base cylinder, which for a standard shaper
is given by
p
#0s ¼ invas ; ð14:7Þ
2zs
where zs is the number of teeth of the shaper, and as is the pressure angle. The upper
and lower signs in equation system (14.6) correspond respectively to involute
profile I and II in Fig. 14.13.
When we move from the Cartesian plane Os ðxs ; ys Þ to Cartesian space
Os ðxs ; ys ; zs Þ, according to what we have seen in Sect. 11.5, for the parametric
representation of the shaper involute tooth surface, rs , we must introduce two
independent variable parameters, u and v. In this case, parameter v coincides with
the involute roll angle, #s , while parameter u ¼ us represents the third coordinate in
the direction of the zs -axis of the current point P on the involute tooth surface in the
three-dimensional space.
By limiting the discussion to the involute profile I in Fig. 14.13, and using
homogeneous coordinates, in order to benefit from their known advantages for
matrix calculus, we can represent the involute tooth surface of the shaper, rs , by the
following vector function:
14.5 Tooth Flank Surfaces of Face Gears, and Their Equations 789
2 3
rbs cos #s þ #0s þ #s sin #s þ #0s
6 rbs sin #s þ #0 #s cos #s þ #0 7
r s ð us ; # s Þ ¼ 6
4
s s 7:
5 ð14:8Þ
us
1
According to what we have seen in Sect. 11.5, the unit normal ns to the tooth
flank surface of the shaper is given by:
2 3
Ns sin # s þ #0s
ns ¼ ¼ 4 cos #s þ #0s 5: ð14:9Þ
jN s j
0
We can now determine the face gear tooth surface, r2 , as the envelope to the
family of the shaper tooth surface, rs . For this purpose, it is first necessary to define
the relative positions of shaper and face gear to be generated under cutting con-
ditions. To do this in the best and easiest way, we use the following coordinate
systems:
• Coordinate system Os ðxs ; ys ; zs Þ rigidly connected to the shaper, already defined
above (see Figs. 14.13, 14.14 and 14.7).
• Coordinate system O2 ðx2 ; y2 ; z2 Þ rigidly connected to the face gear (see
Figs. 14.7 and 14.14a), with origin at the crossing point O Os ¼ O2 between
axes of shaper and face gear, z2 -axis coinciding with the face gear axis, and
directed as Figs. 14.7 and Fig. 14.14 show, and axes x2 and y2 in the plane
perpendicular to the z2 -axis and passing through the origin O2 , and directed as
Fig. 14.14b shows.
(b)
790 14 Face Gear Drives
where z2 and zs are the numbers of teeth of the shaper and face gear, respectively.
In the coordinate system, O2 ðx2 ; y2 ; z2 Þ, the family of shaper surfaces, rs , is
described by the following matrix equation:
is the rotational matrix which describes rotation about the zn -axis with the unit
vector,
14.5 Tooth Flank Surfaces of Face Gears, and Their Equations 791
2 3
cos R 0 sin R r2
6 0 1 0 0 7
M mn ðus Þ ¼ 6
4 sin R
7 ð14:13Þ
0 cos R r2 cot R 5
0 0 0 1
is the rotational matrix, which describes rotation about the z2 -axis with the unit
vector.
The meshing equation to be used is expressed by the dot product given by
Eq. (11.95), which we can write here in the double form:
where ns and N s are respectively the unit normal and normal to any point P of the
shaper surface rs , and vðss2Þ is the relative velocity, i.e. the velocity of a point on
surface rs with respect to a point on surface r2 . Of course, when these two points
coincide with each other and the two surfaces rs and r2 are regular surfaces, they
represent the common point of tangency of the two surfaces.
In this respect, it should be noted that the necessary condition of existence of
surface r2 provides that, if it exists, it is in tangency with surface rs . Indeed, the
sufficient condition of existence of surface r2 provides that it is not only in tan-
gency with surface rs , but also a regular surface, i.e. a surface without singular
points. It should also be emphasized that instantaneous line contacts are typical for
cases where rs is the cutting tool surface and r2 is the generated surface, and this is
our case.
Using the second of the Eq. (14.15) and remembering the expressions of the
relative velocity, given in Sect. 11.7, we get the following meshing equation
between the surfaces rs and r2 , which is expressed as a dot product of cross
product:
@r2 @r2 @r2
ns vðss2Þ ¼ fs2 ðus ; #s ; us Þ ¼ ¼ 0; ð14:16Þ
@us @#s @us
in fact, the product within the parenthesis is a cross product, while the other product
is a dot product.
792 14 Face Gear Drives
This last equation is called the meshing equation as it relates the curvilinear
coordinates ðus ; #s Þ of surface rs with the generalized parameter of motion, us . It
represents the necessary condition of existence of the envelope to the family of
surfaces defined by Eq. (14.11). If this equation is satisfied and the envelope really
exists, we can proceed as indicate below. It should be noted, however, that vector
@r2 =@us has the same direction as the already defined vector vðss2Þ . Therefore,
substituting vðss2Þ instead of @r2 =@us in the last expression of Eq. (14.16), and
taking into account that the cross product represents the normal N s , we get the first
equality of the same equation, which is a dot product. The engineering approach to
the problem discussed here uses this dot product, as it does not depend on the
chosen coordinate system. Of course, normal N s can be replaced by unit normal ns .
Surface r2 of the face gear can be determined using Eq. (14.11), which describes
the family of shaper surfaces rs in the coordinate system O2 ðx2 ; y2 ; z2 Þ, and simul-
taneously the meshing Eq. (14.16). The matrix equation of r2 ðus ; #s ; us Þ, given by
Eq. (14.11), is expressed in a three-parameter form: the first two parameters that
appear in this equation designate the surface parameter of the shaper, ðus ; #s Þ, while
the third parameter is the generalized parameter of motion, us .
From the previous discussion it follows that the shaper surface rs is in line
contact with the face gear tooth surface r2 . Such a type of line contact is due to the
fact that surface r2 is generated as the envelope of the shaper surface rs . We have
already designated by Ls2 the lines of tangency between surfaces rs and r2 in
relative motion (see Fig. 14.12).
Surface r2 can be written in a two-parameter form, using the theorem of implicit
function system existence (see Korn and Korn [18], Zalgaller [39]). To this end,
following the procedure proposed by Zalgaller and Litvin [38], we assume that the
meshing equation, fs2 ðus ; #s ; us Þ ¼ 0, is satisfied at a point P0 u0s ; #0s ; u0s . We also
assume that at this point the following condition is satisfied:
@fs2
6¼ 0: ð14:17Þ
@us
us ¼ us ð#s ; us Þ 2 C 1 ð14:18Þ
and the face gear tooth surface, r2 , can be determined locally as:
Finally, lines of contact Ls2 on the face gear tooth surface are determined by
vector function R2 ð#s ; us Þ, taking us ¼ const.
14.6 Conditions of Non-undercut of Face Gear Tooth Surface 793
The appearance of singularities on the face gear tooth surface, r2 , is the first
symptom of its imminent undercut. As it is well known, in order to avoid the
undercut of this surface, it must be a regular surface, i.e. a surface without singu-
larities, which occur at points where a tangent plane does not exist, and therefore
the local normal, N 2 , becomes equal to zero. Now, if the generated surface r2 is
expressed by Eq. (14.19), i.e. it is represented in a two-parameter form, the normal
N 2 will be given by the following cross product:
@R2 @R2
N2 ¼ : ð14:20Þ
@#s @us
Therefore, a singular point at this surface appears if at least one of the vectors in
the cross product is equal to zero or if the two vectors are collinear. However,
detection of the singular points of the face gear surface r2 through this method is
very laborious and heavy from a computational point of view. For the determination
of singularities of this surface, Litvin et al. [30] proposed a different approach which
is described below. This other approach is preferable, since it is less costly
computationally.
As it well known, the absolute velocity vector, vi (with i ¼ s; 2), of a point of
contact between surfaces rs and r2 , for each of the two members of the gear pair
under consideration, is expressed as the vector sum of two components. The first is
ðiÞ
the component vtr in transfer motion with one of the two members of the gear pair.
The second is the component vðriÞ in relative motion over the tooth surface ri (with
i ¼ s; 2). Since the point of contact belongs to both surfaces, for continuity the
resulting velocities at a given point of contact must be the same for both gear
members, so we get:
ðsÞ ð2Þ
vs ¼ vtr þ vðrsÞ ¼ v2 ¼ vtr þ vðr2Þ : ð14:21Þ
ðsÞ ð2Þ
vðr2Þ ¼ vðrsÞ þ vtr vtr ¼ vðrsÞ þ vðss2Þ ð14:22Þ
ðsÞ ð2Þ
where vðss2Þ ¼ vtr vtr is the sliding velocity at the point of tangency between the
two surfaces, while vðrsÞ is the velocity at a point of contact in its motion over surface
rs of the shaper.
Litvin et al. [30] showed that the following condition must be satisfied at a
singular point of the surface r2 :
794 14 Face Gear Drives
d
½fs2 ðus ; #s ; us Þ
¼ 0 ð14:24Þ
dt
where rs ðus ; #s Þ is given by Eq. (14.8). It is to be noted that ð@rs =@us Þ, ð@rs =@#s Þ,
and vðss2Þ are three- or two-dimensional vectors for spatial and planar gears,
respectively, and are expressed in coordinate system Os ðxs ; ys ; zs Þ.
Thus, from Eq. (14.24) we get:
Equations (14.25) and (14.26) are a system of four linear equations in the two
unknowns ðdus =dtÞ and ðd#s =dtÞ, since ðdus =dtÞ is considered as given. This
system has a well-defined solution for the unknowns only if the matrix:
" @r @rs
#
@us
s
@#s vðss2Þ
M¼ @fs2 @fs2 @fs2 dus ð14:27Þ
@us @#s @u dt
s
2 @xs @xs 3
@us @#s vðxss2Þ
6 @zs @zs
vðzss2Þ 7
D2 ¼ 4 @us @#s 5¼0 ð14:29Þ
@fs2 @fs2 @fs2 dus
@us @#s @u dt
s
14.6 Conditions of Non-undercut of Face Gear Tooth Surface 795
2 @ys @ys 3
@us @#s vðyss2Þ
6 @zs @zs
vðzss2Þ 7
D3 ¼ 4 @us @#s 5¼0 ð14:30Þ
@fs2 @fs2 @fs2 dus
@us @#s @u dt
s
2 @xs @xs
3
@us @#s vðxss2Þ
6 @ys @ys 7
D4 ¼ 4 @us @#s vðyss2Þ 5 ¼ 0 ð14:31Þ
@zs @zs
@us @#s vðzss2Þ
which is just the meshing equation. Therefore, this equation cannot be used to
determine the condition of singularity for surface r2 . To this end, only Eqs. (14.28)
to (14.30) can be used.
The simultaneous fulfillment condition of the system consisting of the three
Eqs. (14.28) to (14.30) can be expressed by the following vector equation, which
can easily be obtained by combining the same three equations:
@fs2 @rs ðs2Þ @fs2 @rs ðs2Þ @fs2 dus @rs @rs
m¼ vs vs þ ¼ 0: ð14:33Þ
@us @#s @us @us @us dt @us @#s
This vector equation is satisfied when both of the following conditions are met:
at least one Eqs. (14.28) to (14.30) is satisfied, and vector m is not perpendicular to
any of the axes of the coordinate system Os ðxs ; ys ; zs Þ.
A sufficient condition for detecting the singularities of the generated surface r2 ,
and therefore the solutions to avoid undercut, is represented by the following simple
equation:
r s ¼ r s ð us ; # s Þ
fs2 ðus ; #s ; us Þ ¼ 0 ð14:35Þ
Fs2 ðus ; #s ; us Þ ¼ 0
overcoming generates singularities of the face gear tooth surface, r2 : From the point
of view of the face gear design, it is therefore important to choose the appropriate
settings for surface rs that generates the surface r2 .
The procedure for detecting the aforementioned limit line implies, first, that the
plane ðus ; #s Þ of parameters of the shaper tooth surface is considered, and that on
this plane, the lines of tangency Ls2 between shaper surface rs and face gear surface
r2 are plotted, as well as the line M of the points ðus ; #s ; us Þ corresponding to the
singular points of the surface r2 . To do this first step, we use the second and third of
Eq. (14.35). The result of this first step is therefore a representation of the plane
ðus ; #s Þ, where lines Ls2 and line M are plotted, as shown in Fig. 14.15. The second
decisive step is to completely remove line M of the space of parameters ðus ; #s Þ.
Since line M designates the shaper parameter measured along its axis, to avoid
singularities of the surface r2 it is sufficient to eliminate the parts of the Ls2 -lines to
the left of the vertical line us K, where us corresponds to the addendum of the
shaper and K is the limiting point on the parameter plane defined by the coordinates
us ; #s . For generation of face gear by a shaper, as the one shown in Fig. 14.15,
singularities are avoided in the parameter sub-area for which #s \#s and us [ us .
Based on the above considerations, we can now determine the magnitude of Ri
that will avoid singularities of the face gear tooth surface r2 . Instead, quantity Re
determines the area free of pointed teeth of the same surface
r2 . For the calculation
of Ri , we must first consider the limiting point K us ; #s of the line of singularities,
which belongs to the addendum of the shaper. The parameter #s of this limiting
point is determined by the following equation:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 r2
ras
#s ¼ bs
ð14:36Þ
rbs
where ras and rbs are respectively the radii of addendum circle and base circle of the
shaper. We must then take into account the following fact, which is confirmed by
the investigations: to determine singularities of surface r2 , rather than solving
Eq. (14.34), just take D2 ¼ 0 or D3 ¼ 0. The equations D2 ¼ 0 or D3 ¼ 0 include
partial derivatives @zs =@us and @zs =@#s , while equation D1 ¼ 0 does not include
these partial derivatives. Using, for example, the equation D2 ¼ 0 as well as the
meshing equation fs2 ðus ; #s ; us Þ ¼ 0, we get two equations in the two unknowns us
and us . The parameter us thus obtained determines the sought-for magnitude of Ri .
Even for face gears, it is necessary to prevent the tooth thickness on the tooth top
land becoming zero, so the teeth become pointed teeth. To break the danger of the
pointed teeth, it is necessary to locate the area where it can manifest, considering
the intersection of the two opposite flank surfaces at the tooth top land. To locate
this area, special computing programs can be used. Here we will use an approxi-
mate method that guarantees sufficiently reliable results for face gears generated by
an involute shaper.
With reference to Figs. 14.14a and 14.16, we indicate with Om Q ts2 the in-
stantaneous axis of rotation in the generation process, which sees the generation
shaper and the to-be-generated face gear respectively rotating about their axes, zs
and z2 . Consider the two transverse sections of the shaper obtained as intersections
with the two transverse planes p1 and p2 , which are perpendicular to the zs -axis and
pass respectively through the pitch point, C, and a point, Q, located on the
instantaneous axis of rotation, ts2 (see Fig. 14.16). We have to determine the
location of p2 -plane where the two opposite tooth flanks of the face gear intersect.
Figures 14.17a and 14.17b show respectively the tooth profiles of the shaper and
face gear in planes p1 and p2 . The point of intersection of the two opposite
transverse profiles of the face gear in plane p2 is indicated by A (Fig. 14.17b). This
point must be located on the addendum line of the face gear, so its location on the
x0n -axis is given by ðrs mÞ. It is now necessary to determine the magnitude of Re ,
which is defined by distance Dl between the planes p1 and p2 (Fig. 14.16). This last
figure and Fig. 14.17b highlight the procedure for calculating Dl and Re .
It is first necessary to determine the pressure angle a of the pointed teeth
(Fig. 14.17b). For this purpose, we use the following vector equation:
where P is the point of tangency of profiles of the shaper and face gear in plane p2 .
ks and rbs #s the absolute values of vectors PA and TP, that is
Let’s denote then with
PA ¼ ks , and TP ¼ rbs #s . Moreover, the absolute value of the vector O0 A is
n
given by (Fig. 14.17a):
(a) (b)
Fig. 14.17 Transverse tooth profiles of shaper and face gear: a in plane p1 ; b in plane p2
O0n A ¼ rs m; ð14:38Þ
zs 2
rbs ðcos a þ #s sin aÞ ks cos a ¼ m
2 ð14:39Þ
rbs ðsin a #s cos aÞ ks sin a ¼0;
where rbs ¼ ðzs =2Þm cos as and #s ¼ a #0s , with #0s given by Eq. (14.7).
Eliminating ks , from equation system (14.39) we obtain the following nonlinear
equation:
zs 2 p
a sin a ¼ invas ; ð14:40Þ
zs cos as 2zs
from whose solution we get the sought-for pressure angle a. From Fig. 14.17b we
then get the following relationship:
rbs zs cos as
O0n Q ¼ ¼ m: ð14:41Þ
cos a 2 cos a
O0n Q zs cos as
Re ¼ ¼ m: ð14:42Þ
tan ds 2 cos a tan ds
14.7 Conditions to Avoid Face Gear Pointed Teeth, and Fillet Surface 799
ðzs þ 2:5Þ
ras ¼ rs þ 1:25 m ¼ m: ð14:43Þ
2
where M 2s ¼ M 2m M mn M ns ðus Þ.
At the end of Sect. 14.3, we emphasized the importance of an appropriate choice
of the dimensionless coefficient k ¼ ðRe Ri Þ=m ¼ b=m. The guidelines we had
given then were based on results of investigations concerning the influence that this
coefficient exerts on the structure of the face gear teeth. The above investigation
assumed that the outer radius Re was known, as previously determined by imposing
conditions that could guarantee the absence of pointing, i.e. pointed teeth.
Therefore, once Re is known, we can eliminate the portion of the tooth where the
fillet exists, by increasing the inner radius Ri (Fig. 14.16). This means that the
coefficient k will have to be decreased (see Fig. 14.9). However, choosing a suf-
ficient value of coefficient k, allows us to obtain a more uniform structure, elimi-
nating the weaker part of the face gear tooth.
The technological development and the improvement of the quality of a face gear
pair necessarily entail an in-depth knowledge of the meshing and bearing contact
between the tooth flank surfaces of the pinion and face gear, which make up a face
gear pair. This in-depth knowledge can be acquired through expensive experiment
tests on prototypes, or more often by computerized simulation techniques of
800 14 Face Gear Drives
meshing and bearing contact between the tooth surfaces, in the actual operating
conditions of the gear drive under consideration.
These simulation techniques undoubtedly have the advantage of being able to be
used in the design stage of the face gear pair in order to optimize design choices in
relation to the goals to be achieved. To do so, Tooth Contact Analysis
(TCA) programs are available, which are based on the theoretical considerations
described below. With these programs, several goals can be achieved, such as:
• definition and location of the bearing contact as a series of instantaneous contact
ellipses, in order to eliminate the risk of localized edge contacts;
• definition of the paths of contact on tooth flank surfaces;
• evaluation of transmission errors due to misalignment of the face gear members.
This type of analysis involves first the definition of the continuous tangency
conditions between the tooth surfaces r1 and r2 of the pinion and face gear. To this
end,
we introduce
three coordinate systems, O1 ðx1 ; y1 ; z1 Þ, O2 ðx2 ; y2 ; z2 Þ and
Of xf ; yf ; zf , which are rigidly connected with the pinion, face gear and frame of
the face gear drive respectively. Coordinate systems O1 ðx1 ; y1 ;z1 Þ and O 2 ð x2 ; y 2 ; z 2 Þ
are therefore mobile systems, while coordinate system Of xf ; yf ; zf is a fixed
system. It should be noted that, since in this section we are interested in determining
the meshing and contact conditions between tooth surfaces of a face gear pair in the
actual assembly conditions, the fixed coordinate system Of xf ; yf ; zf is unique.
We then introduce an additional fixed-coordinate system, Oa ðxa ; ya ; za Þ, with
which we simulate the misalignment. In addition, we introduce two more auxiliary
coordinate system, Ob ðxb ; yb ; zb Þ and Oc ðxc ; yc ; zc Þ, which are used not only to
simulate the misalignment of the face gear, but also to better define the complex
geometry of the problem being discussed. These six coordinate systems are shown
in Fig. 14.18a–d.
All misalignments are referred to the face gear. Quantities Da, b, and b cot R
determine the position Oa of the Oa ðxa ; ya ; za Þ-system with respect to
of origin
origin Of of the Of xf ; yf ; zf -system (Fig. 14.18c). Quantity Da represents the
shortest distance between the axes of pinion and face gear, when they do not
intersect, but are crossed axes (Fig. 14.18c); b and b cot R represent the position and
orientation of the coordinate
system Oa ðxa ; ya ; za Þ with respect to the coordinate
system Of xf ; yf ; zf (Fig. 14.18c). The misaligned face gear rotates about the zc -
axis (Fig. 14.18b). The position of coordinate system Oc ðxc ; yc ; zc Þ with respect to
coordinate system Ob ðxb ; yb ; zb Þ simulates the axial displacement Dc of the face
gear (Fig. 14.18d). The orientation of coordinate system Ob ðxb ; yb ; zb Þ with respect
to coordinate system Oa ðxa ; ya ; za Þ simulates the actual (or simulated) crossing
angle Ra ¼ R þ DR, where R is the shaft angle, while DR is its variation due to
misalignment (Fig. 14.18d).
In the coordinate systems O1 ðx1 ; y1 ; z1 Þ and O2 ðx2 ; y2 ; z2 Þ, the tooth flank sur-
faces r1 and r2 are represented by the two following vector functions of class C 2 :
14.8 Meshing and Contact Condition Between Tooth Surfaces … 801
(a) (b)
(c) (d)
Fig. 14.18 Coordinate systems used for simulation of meshing: a coordinate system O1 and Of ;
b coordinate systems O2 and Oc ; c coordinate systems Oa and Of ; d coordinate systems Oa , Ob ,
and Oc
r i ð ui ; # i Þ 2 C 2 ; ð14:45Þ
with (i ¼ 1; 2) and ðui ; #i Þ 2 Ai (i.e. the elements ui and #i belong to a set Ai ), under
the condition that singular points do not exist, so the following condition must be
meet:
@ri @ri
6¼ 0: ð14:46Þ
@ui @#i
This last cross vector is different from zero only when the pair of vectors
ru ¼ @ri =@ui and r# ¼ @ri =@#i , which are tangent to the u-lines and #-lines
respectively, are both different from zero, and are not collinear.
802 14 Face Gear Drives
@ri @ri
Ni ¼ ð14:47Þ
@ui @#i
@ri
Ni @ri
ni ¼ ¼ @ui @#i ; ð14:48Þ
jN i j @ri @ri
@ui @#i
with (i ¼ 1; 2). If now the pinion (i ¼ 1) and face gear (i ¼ 2), and so the tooth
flank
surfaces r1 and r2 , rotate about a fixed axis of the coordinate system
Of xf ; yf ; zf , which is rigidly connected to the frame, two families of tooth flank
surfaces are generated in this fixed coordinate system. In this system, these gen-
erated surfaces as well as their unit normals can be expressed by matrix equations
(see Litvin and Fuents [24]) or by means of the following vector functions:
ðiÞ
rf ¼ ðui ; #i ; ui Þ ð14:49Þ
ðiÞ
nf ¼ ðui ; #i ; ui Þ ð14:50Þ
with (i ¼ 1; 2). In these equations, ui and #i are the surface parameters of r1 and
r2 , while ui are the angles of rotation of pinion and face gear during their meshing.
The contacting surface must be in tangency at every instant, i.e. in continuous
tangency, and this condition is met if their position vectors and normals coincide at
any instant, i.e. when:
ð1Þ ð2Þ
rf ðu1 ; #1 ; u1 Þ ¼ rf ðu2 ; #2 ; u2 Þ ð14:51Þ
ð1Þ ð2Þ
nf ðu1 ; #1 ; u1 Þ ¼ nf ðu2 ; #2 ; u2 Þ: ð14:52Þ
Vector Eq. (14.51) gives three independent scalar equations, while vector
Eq. (14.52) yields only two independent scalar equations, since the absolute values
of the unit normals coincide, that is:
ð1Þ ð2Þ
nf ¼ nf ¼ 1: ð14:53Þ
Therefore, vector Eqs. (14.51) and (14.52) yield a system of five independent
scalar equations in six unknowns u1 , #1 , u1 , u2 , #2 , u2 . This system of equations
can be expressed as:
fi ðu1 ; #1 ; u1 ; u2 ; #2 ; u2 Þ ¼ 0 ð14:54Þ
One of the six quantities that come into play can be chosen as input parameter.
Here we choose u1 as an example. The condition that surface r1 and r2 are in point
contact leads to the following inequality:
@ ðf 1 ; f 2 ; f 3 ; f 4 ; f 5 Þ
Df ¼ 6¼ 0; ð14:55Þ
@ ðu1 ; #1 ; u2 ; #2 ; u2 Þ
whose left-hand side represents the Jacobian matrix, i.e. the matrix of all first-order
partial derivatives of the functions fi with respect to the five unknowns. In this case,
the Jacobian matrix is a square matrix, for which both the matrix and its determinant
are called simply Jacobian (see Buzano [4], Hirsch et al. [14], Tricomi [36]).
Therefore, the solution of system of Eq. (14.54) can be expressed by the
functions:
In accordance with the theorem of implicit function system existence (see Korn
and Korn [18], Miller [30]), we can state that functions (14.56) exist in the
neighborhood of a point
ð0Þ ð0Þ ð0Þ ð0Þ ð0Þ ð 0Þ
Pð0Þ u1 ; #1 ; u1 ; u2 ; #2 ; u2 ; ð14:57Þ
@f1 @f1 @f1 @f1 @f1
@u1 @#1 @u2 @#2 @u2
@ ðf1 ; f2 ; f3 ; f4 ; f5 Þ .. 6¼ 0:
Df ¼ ¼ ... ..
.
..
.
..
. . ð14:58Þ
@ ðu1 ; #1 ; u2 ; #2 ; u2 Þ @f @f5 @f5 @f5 @f5
@u5 @#1 @u2 @#2 @u
1 2
When this last requirement is met, functions fi have locally, in the neighborhood
of point Pð0Þ , their inverse functions that are differentiable if and only if the
Jacobian determinant is nonzero at Pð0Þ .
Functions (14.56) provide complete information on the meshing condition
between the pinion and face gear, which are in point contact. The function u2 ¼
u2 ðu1 Þ represents the law of motion, as it correlates the angles of rotation of the
face gear pair members about their axes. The solution of system of Eqs. (14.51) and
(14.52) by means of functions (14.56) is an iterative procedure. This procedure
requires, as a first step, the identification of the set of parameters characterizing the
point Pð0Þ , and satisfying locally the system of Eqs. (14.51) and (14.52).
804 14 Face Gear Drives
The second goal is to determine the paths of contact on surface r1 and r2 , which
are given, respectively, by the equations:
It is not the case here to go further into the theoretical discussion described
above. Instead, we should focus on some results obtained by Litvin and Fuentes
[24], who applied the aforementioned theory to a face gear generated by an involute
shaper with ideal theoretical cutting profiles. These results show an important shift
in bearing contact, due to alignment errors. This bearing contact is oriented across
the tooth surface, and is sensitive to the variation DR of the shaft angle. Therefore,
the risk of dangerous edge contact may occur. The sensitivity to the shaft angle
variation can be compensated by an axial correction Dc (Fig. 14.18d) of the face
gear during assembly. Faced with these disadvantages, the geometry of the face
gear pair obtained with the above-mentioned generation process has the consider-
able advantage of not causing transmission errors, whatever the misalignment.
The same Litvin and Fuentes [24] have also proposed and studied other face gear
pair geometries, obtained by a process that generates pinion and shaper by means of
rack-cutters with a modified profile compared to a reference rack-cutter with
straight-line profile. The face gear pair thus obtained are characterized by a
lengthwise orientation of the bearing contact, which removes or decreases the
danger of edge contact, and results in an appreciable reduction in stress. However,
they have the disadvantage of being still sensitive to misalignment, which is
moreover accompanied with transmission errors. For more information on this
topic, we refer the reader to Litvin et al. [25, 26].
In Sect. 14.2 we introduced the Litvin’s method that, as we have already pointed
out, can be used for generation or grinding of face gears, respectively by hobs (or
generation worms) or grinding worms. We have already highlighted the advantages
of this method in comparison with the conventional generation method of face
14.9 Generation and Grinding of Face Gears by Worms 805
gears, by means of a spur involute shaper and a manufacturing process that sim-
ulates the meshing of this shaper and face gear to be generated.
In this section we want to describe the theoretical bases of the Litvin’s method
[22, 23]. This method solves the following design problems of the generating and
grinding worms: (i), determination of the shaft angle between axes of the imaginary
shaper and actual worm, and definition of the worm thread surfaces; (ii), choices to
avoid singularities of these worm surfaces, and definition of their dressing process.
The surfaces involved in the problem we are discussing are those of the imag-
inary shaper, worm and face gear, which are here indicated with rs , rw and r2
respectively. Figure 14.5 shows the simultaneous meshing of these three surfaces.
We here assume that surface rs has been generated as the envelope to the family of
an ideal rack cutter having straight tooth profiles. Therefore, surface rs is the
involute surface of a spur gear, which is represented by the vector function
rs ðus ; #s Þ given by Eq. (14.8). It should be noted that if we want to take into
account both the profiles shown in Fig. 14.13, in the second row of the matrix in
Eq. (14.8) we must introduce the signs that appear in the second Eq. (14.6).
Surfaces rw and r2 are generated as the envelope to the family of shaper surfaces
rs .
We must first determine the shaft angle between axes of imaginary shaper and
worm. In Sect. 14.2 we have highlighted that the installation of the generating or
grinding worm on the face gear to be generated or grinded must take into account
the lead angle, cw , of its thread, so the worm and face gear axes are skewed or
crossed axes. Consequently, the installation of the worm with respect to the shaper
must take into account this fact, so shaper and worm axes will also have crossed
axes, with the same shaft angle of the gear pair formed by worm and face gear.
To determine this shaft angle between axes of shaper and worm, and to solve
other problems, such as those relating to tangency of surfaces rs and r2 , and
generation of the face gear by the worm, we use the coordinate systems shown in
Fig. 14.19, namely:
• mobile coordinate system Os ðxs ; ys ; zs Þ, rigidly connected to the shaper, whose
zs -axis coincides with the axis of rotation of the shaper;
• mobile coordinate system Ow ðxw ; yw ; zw Þ, rigidly connected to the worm, whose
zw -axis coincides with the axis of rotation of the worm;
• fixed coordinate system Oa ðxa ; ya ; za Þ, rigidly connected to the frame, and
having the same origin and the same z-axis of the Os -system, that is Oa Os
and za zs ;
• fixed coordinate systems Ob ðxb ; yb ; zb Þ and Oc ðxc ; yc ; zc Þ, rigidly connected to
the frame, both of which have the same origin of the Ow -system (i.e.
Ob Oc Ow ), and with zb -axis parallel to the axes za and zs , and zc -axis
coinciding with the zw -axis;
• fixed coordinate system Om ðxm ; ym ; zm Þ rigidly connected to the frame as defined
in the Sect. 14.5 and also shown in Fig. 14.14.
Axes zs and zw of shaper and worm are crossed axes, and their positive directions
form the shaft angle R ¼ ðp=2Þ
cw , where cw is the lead angle of the worm
806 14 Face Gear Drives
(a) (b)
Fig. 14.19 Coordinate systems to: a define the position of the worm with respect to the shaper;
b impose the simultaneous meshing of surface rs , rw and r2 , and generate the face gear by the worm
thread, given by Eq. (14.1). Just remember the installation of the same worm on the
face gear described in Sect. 14.2. It is to be noted that here we depart slightly from
the definition of ISO 1122-1[15], according to which the shaft angle is the smallest
angle through which one of the axes of the crossed gear pair must be swiveled so
that the axes are parallel and their directions of rotation are opposite. The upper and
lower sign in expression above corresponds to the application of a right-hand or
left-hand worm. The center distance, asw , is the shortest distance between axes zs
and zw .
rðaCÞ ¼ ½ rs 0 0 T ; ð14:62Þ
where N ðsÞ is the normal to rs -surface, while vðswÞ ¼ vðsÞ vðwÞ is the relative
velocity between the two surfaces at point C. Of course, vðsÞ and vðwÞ are the
velocity vectors of point C of the shaper and worm. By developing Eq. (14.62), we
get the following expression of the lead angle of the worm thread:
rs
cw ¼ arcsin ; ð14:64Þ
zs ðaws þ rs Þ
where zs and rs are respectively the number of teeth and radius of pitch circle of the
shaper, while aws is the shortest distance between axes of shaper and worm. The
magnitude of the center distance, asw , affects the dimensions of the generation or
grinding worm as well as the avoidance of surface singularities of the worm.
Figure 14.20 shows schematically the meshing between worm and shaper.
Figure 14.5 shows the imaginary pinion, worm and face gear in simultaneous
meshing, during which all three surfaces rs , rw and r2 are involved. For our
purpose, the rs -surface of the imaginary shaper is here considered as a given
generation surface. Instead, surfaces r2 and rw are generated as the envelopes to the
family of rs -surface of the shaper.
Now, the generation surface rs , whose geometry is known, is given by the vector
function (14.8). Even surface r2 of the face gear has already been determined,
simultaneously using Eq. (14.11) describing the family of shaper surfaces rs in the
coordinate system O2 ðx2 ; y2 ; z2 Þ, and the meshing equation given by Eq. (14.16).
Therefore, only the surface rw of the worm remains to be determined. To determine
this surface in coordinate system Ow ðxw ; yw ; zw Þ, we simultaneously use the fol-
lowing two equations [24, 25]:
where zs is the number of teeth of the shaper, and zw the number of threads of the
worm. In accordance with the usual choices in this regard, we here assume that the
worm has only one thread, so zw ¼ 1.
14.9 Generation and Grinding of Face Gears by Worms 809
The meshing Eq. (14.66) can be written in the form [21, 24]:
where N s is the normal to any point of the shaper surface rs , and vðsswÞ is the relative
velocity of a point on surface rs with respect to a point on surface rw . Of course,
when these two points coincide with each other, and the two surfaces rs and rw are
regular surfaces, they represent the common point of tangency of the two surfaces.
Vector function rw ðus ; #s ; us Þ given by Eq. (14.65), and equation of meshing
fws ðus ; #s ; us Þ ¼ 0, given by Eq. (14.66) or Eq. (14.68), represent the worm thread
surface rw expressed by the three related parameters us , #s and us .
From all this, it follows that the generation shaper surface rs is in line contact
with the worm thread surface rw . Such a type of line contact is due to the fact that
surface rw is generated as the envelope of the shaper surface rs . We designate by
Lsw the lines of tangency between surfaces rs and rw in relative motion.
The surface rw can be also written in two-parameter form, using the theorem of
implicit function system existence (see Korn and Korn [18], Zalgaller [39]). Here
too, following the procedure proposed by Zalgaller and Litvin [38], we assume that
the meshing equation, fsw ðus ; #s ; us Þ ¼ 0, is satisfied at a point P0 u0s ; #0s ; u0s . We
also assume that at this point the following condition is satisfied:
@fsw
6¼ 0: ð14:69Þ
@us
Thus, the meshing equation fsw ðus ; #s ; us Þ ¼ 0 can be solved in the neighbor-
hood of point P0 by a function of class C 1 as:
# s ¼ # s ð us ; u s Þ 2 C 1 ð14:70Þ
Lines of contact Lsw on the worm thread surface are determined by vector
function Rw ðus ; us Þ, taking us ¼ const.
Results concerning the lines of contact, obtained using the equations described
above, show that the lines Lsw and Ls2 do not coincide with each other, but intersect
at any point where the two surfaces rw and r2 are in meshing. Both of these lines of
contact are a function of the parameter of motion, us . In the plane of the parameter
ðus ; #s Þ, the qualitative results shown in Fig. 14.21 are obtained. Figure 14.21a
shows that the shaper surface rs is in line contact with the worm surface rw , and
that the corresponding contact lines referring to a given value of uðsiÞ are non-linear
curves. Figure 14.21b shows that the shaper surface rs is also in line contact with
810 14 Face Gear Drives
(a) (b)
(c)
Fig. 14.21 Lines of contact: a lines Lsw between surfaces rs and rw ; b lines Ls2 between surfaces
rs and r2 ; c, intersection of lines Lsw uðsiÞ and Ls2 uðsiÞ at meshing position where surfaces rw
and r2 are in point contact
the face gear surface r2 , and that the corresponding contact lines referring to a given
value of uðsiÞ are straight lines all parallel to the us -axis. Figure 14.21c shows that
lines of contact Lsw uðsiÞ and Ls2 uðsiÞ intersect with each other at a position of
meshing for which us ¼ uðsiÞ . This point of intersection corresponds to the point of
tangency between surfaces rw and r2 .
From what we have seen in the previous section, we know that both contact
between surfaces rs and rw , and between surface rs and r2 , are line contacts.
Instead, surfaces rw and r2 of worm and face gear are in point contact with each
other at any instant. This means that neither the generation cutting nor the finishing
grinding of the surface r2 of face gear by the worm surface rw can be carried out
using a one-parameter enveloping process. In effect, such a process would not
provide the entire required surface r2 , but only a small strip of it.
To obtain the entire required surface rs , however, it is necessary to use a
two-parameter enveloping process, characterized by two set of parameters, inde-
pendent of each other, namely:
• a set of angles of rotation uw and u2 of the cutting (or grinding) worm and face
gear to be generated (or grinded), which must be correlated by the relationship:
14.9 Generation and Grinding of Face Gears by Worms 811
where
• z2 and zw (usually zw ¼ 1, as a single start worm is used) are respectively the
number of teeth of the face gear, and the number of threads of the worm;
xw uw z2
¼ ¼ ; ð14:72Þ
x2 u2 zw
• a translation motion of the worm collinear to the axis of the imaginary shaper
(see Fig. 14.5), and characterized by the parameter lw .
To obtain the surface r2 of the face gear generated by the worm, we use the same
theoretical basis that allowed us to derive the surfaces rw and r2 as the envelopes of
surface rs of the imaginary shaper. We also use the coordinate systems described in
Sect. 14.9.1 and shown in Fig. 14.19. In this way we get the following equations, to
be solved simultaneously:
where
parameter uw ¼ const, and variable parameter lw (lines I in Fig. 14.22a), while the
second family consists of lines corresponding to variable parameter uw , and
parameter lw ¼ const (lines II in Fig. 14.22b). The instantaneous point of tangency
P between the worm thread surface rw and the face gear tooth surface r2 is the
intersection point of lines I and II (Fig. 14.22c).
The correctness of the results obtained with the above-mentioned calculation
procedure is easily controllable, because the surface r2 generated using the worm as a
generation cutting tool must coincide with that previously obtained using the shaper
as a generation cutting tool. In other words, the results obtained using Eqs. (14.73)
to (14.75) must coincide with those obtained using Eqs. (14.11) and (14.16).
The thread surfaces rw of the generation worm must be designed so that they are
regular surfaces, i.e. without singular points, as these points cause undercut of the
face gear tooth surfaces to be generated. The procedures for determining the sin-
gularities of worm thread surfaces are the same two procedures we described in
Sect. 14.6, and used to determine the singularities of face gear tooth surface.
With the first procedure, we find the singular points of the worm thread surface
rw under the condition that the local normal N w is equal to zero. Therefore, in the
case that we are here examining, instead of Eq. (14.20) we will have:
@rw @rw @fws @rw @rw @fws @rw @rw @fws
þ þ ¼ 0; ð14:77Þ
@#s @us @us @us @us @#s @us @#s @us
which are respectively quite similar to relationships (14.23), (14.26), (14.27) and
(14.32), already obtained and discussed for find singular points on face gear tooth
surface, r2 . Quantities and expressions that appear in these relationships are: vðrsÞ ,
the velocity of a point that moves over surface of the imaginary shaper; vðsswÞ , the
sliding velocity at a tangency point between the meshing surface rs and rw .
As we have already seen in Sect. 14.6, it should be noted that:
• vectors that appear in Eq. (14.78) are expressed in the coordinate system
Os ðxs ; ys ; zs Þ;
• Equation (14.78) and the differential meshing Eq. (14.66), considered simulta-
neously, allow to obtain Eq. (14.79);
• Equations (14.78) and (14.79) result in a system of four linear equations in two
unknowns, which has a certain solution if the matrix M, given by Eq. (14.80),
has the rank r ¼ 2;
• Equation (14.81) results from the solution of the above system of equations;
• the meshing Eqs. (14.66) and (14.81) allow to determine the sought-for line on
surface of the imaginary shaper rs , that generates singular points on the surface
of the worm rw .
The calculation procedure for determining and highlighting singularities on the
worm surfaces consists of the following step sequence:
• By means of the meshing Eq. (14.66), we determine the lines of contact Lsw
between shaper and worm surfaces, and represent them on the plane of
parameters of the shaper ðus ; #s Þ as functions of the generalized parameter us . In
this way we can see what happens on both the side surfaces that delineate the
shaper tooth space.
• With Eq. (14.81), we represent on the same plane ðus ; #s Þ the limiting lines of
the singular points of the worm. These lines enable us to determine the maxi-
mum angle of rotation of the shaper, which is permissible for avoidance of
worm singularities. It is thus possible to determine the maximum number of
turns of the worm thread.
• Using Eqs. (14.66) and (14.81) as well as the equation of the shaper tooth
surface rs , we determine the limiting lines of the regular points of the shaper,
which generate the worm singularities. These limiting lines are represented on
814 14 Face Gear Drives
both the side surfaces that delineate the shaper tooth space. Each of these side
surfaces is characterized by two limiting lines, located close to the extreme
transverse planes delimiting the face width.
• Finally, using coordinate transformation from the shaper surface rs to the worm
thread surface rw , we determine the limiting regular points on surface of the
shaper, and the singularity points on the worm surface that are generated by
these limiting regular points of the shaper.
For more details on this subject, we refer the reader to the paper of Litvin et al. [25].
So far, we have never dealt with technological processes on how to sharp or dress
the tools that are used to cut, shave or grind the gears. This is because this is a
typical topic of mechanical technology, which is therefore outside the scope of this
textbook or is entirely marginal. We are here to make an exception, with a short
note of the sharpening and dressing problems of cutting tools (hobs) and grinding
wheels used for manufacturing face gears. The reason for this exception follows
from the fact that, as we have already mentioned in Sects. 14.1 and 14.2, the
solution to the face-gear grinding problem was the key that opened the doors for the
growing affirmation of these types of gears.
We are faced with two typical problems: those concerning the sharpening of
cutting tools, and those concerning the dressing of grinding tools. From a theo-
retical point of view, there are no differences between the two types of problems.
Therefore, for brevity, we only talk about dressing problems, with the tacit
understanding that even the sharpening problems are included.
The worm dressing can be done using both planar disks, as proposed by Tan
[34], and conical disks, conforming to the tooth space of the pinion, as proposed by
Litvin et al. [22]. Face gears ground with any method must have tooth surfaces
perfectly conjugate to the mating pinion. To achieve this goal, the installation and
motion of the dressing tool relative to the worm (i.e. the grinding wheel) must meet
well-defined conditions, which we here briefly describe, with reference to a
dressing tool having the shape of a planar disk. As Fig. 14.5 shows, this dressing
tool is a disk with a flat surface, having mathematically a plane as generation
surface. In Sect. 14.2 we have already described the great advantages associated
with this extremely simple geometry, so it is no longer the case to come back to the
subject.
The installation and motion of the dressing planar disk relative to the grinding
worm can be specified with reference to the imaginary involute pinion. The
dressing disk must be positioned in such a way that it is always in tangency with the
pinion tooth surface. As shown in Fig. 14.5, this positioning can be uniquely
defined by the two parameters a and s, given by:
14.9 Generation and Grinding of Face Gears by Worms 815
a ¼ u1 þ # ð14:82Þ
s ¼ #rb1 ð14:83Þ
where: a is the angle between the dressing surface and the shortest distance line
through the axes of rotation of the grinding worm and imaginary pinion; s is the
shortest distance between the dressing surface and the pinion axis; u1 is the angle of
rotation of the pinion; # is the roll angle on the pinion involute profile at point
where it is in tangency with the dressing surface; rb1 is the base radius of the pinion.
The installation of the grinding worm on the face gear to be grind must also take
into account the lead angle cw , given by Eq. (14.1). The motion of the dressing disk
relative to the grinding worm, when it is specified with reference to the imaginary
involute pinion, implies that we consider:
• the meshing between the pinion and face gear, whose rotations are governed by
Eq. (14.2), which determines the transmission ratio of the gear drive;
• the grinding process, where rotation of the grinding worm is related to rotation
of the face gear by Eq. (14.72).
Using simultaneously Eqs. (14.1), (14.2), (14.72), (14.82) and (14.83), we get the
following relationship, which correlates the positioning parameters of the dressing
planar disk and grinding worm:
zw
s ¼ rb1 a
uw ; ð14:84Þ
z1
where the plus-sign is to be used for dressing of the right side of a left-hand thread
or the left side of a right-hand thread, while the minus-sign is to be used for dressing
the right side of a right-hand thread or the left side of a left-hand thread.
An infinite number of pairs of parameters ða; sÞ satisfies Eq. (14.84), for a given
value uw of the angular position of the grinding worm. Each of these pairs cor-
responds to the dressing disk in tangency condition with different points on the
thread surface of the grinding worm.
If we assume that the angular velocity of the grinding worm is constant
ðxw ¼ duw =dt ¼ constÞ, differentiating Eq. (14.84) we obtain the following
equation expressing the velocity vd of the dressing disk along the direction of the
normal to its flat surface (Fig. 14.5):
ds da zw
vd ¼ ¼ rb1
xw : ð14:85Þ
dt dt z1
Since the grinding worm is rotating, so that its angular position defined by
parameter uw is continuously variable, the instantaneous position and orientation of
the dressing disk as well as its motion are specified with two parameters ða; sÞ.
These two parameters can be varied independently, while still satisfying conditions
of meshing. All dressing disk positions corresponding to different combinations of
816 14 Face Gear Drives
parameters ða; sÞ, which even satisfy Eq. (14.84), constitute a family of generating
planes on the grinding worm’s thread surface. The final shape of the grinding worm
is generated as the envelope to a family of these planar surfaces.
In gearing theory, such a generation process is called two-parameter enveloping
process. Practical application of this process involves the introduction of specifi-
cally designed relationships between the two theoretically independent parameters
in order to make the implementation realistically feasible.
As for the dressing process discussed here, the motion of the dressing disk is
specified to be in the direction of normal to its flat surface, so a ¼ const. In this
case, the motion of the dressing disk is a pure translation in the direction of this
normal, and Eq. (14.85) becomes:
zw
vd ¼
rb1 xw ; ð14:86Þ
z1
therefore, the magnitude of the dressing disk velocity is constant for any value of
a ¼ const. The pure translation of the dressing disk makes it possible to avoid
working into sensitive areas in the grinding worm where undercut may occur. This
is one of the main advantages of the use of a dressing disk with planar surface
compared to a dressing tool with an involute profile.
Another important feature of the two-parameter enveloping process is that it
guarantees at every instant a point contact condition between the generation and the
generated surfaces. This contact condition, combined with the use of CNC tech-
nology, allows topological control or modifications of profile of the grinding worm.
These control or modifications can then be transferred on the face gear during the
grinding process. The desired meshing characteristics between the pinion and face
gear can be translated into specifications for modifications of the face gear teeth,
and eventually into motion control in the dressing process by CNC technology.
This is due to one-to-one correspondence between the position of the dressing disk,
a point on the grinding worm, a point on the face gear tooth and a point on the
pinion. Topological control of teeth of the face gear has important practical
implications, on which we cannot dwell more extensively. In this regard, we refer
the reader to the paper of Heath et al. [13]. Concluding this topic, we must highlight
an elegant analytic discussion on the subject, made by Litvin et al. [25], to which
we refer the reader.
References
1. Akahori H, Sato Y, Nishida Y, Kubo A (2001) Test of the durability of face gears. In: The
Proceedings of the JSME International Conference on MTP, Fukuoka, Japan
2. Basstein G, Sijtstra A (1993) New developments concerning design and manufacturing of face
gears. Antriebstechnik 32(11)
3. Bloomfield B (1947) Designing face gears. Mach Des 19(4):129–134
References 817
4. Buzano P (1961) Lezioni di Analisi Matematica – Parte Prima, 5th edn. Libreria Universitaria
Levrotto&Bella, Torino
5. Chakraborty J, Bhadoria BS (1971) Design of face gears. Mechanisms 6:435–445
6. Chakraborty J, Bhadoria BS (1973) Some studies on hypoid face gears. Mech Mach Theory
8:339–349
7. Dudley DW, Winter H (1961) Zahnräder: Berechnung, Enfwurf und Herstellung nach
amerikanischen Enfahrungen. Springer, Berlin, Göttingen, Heidelberg
8. Feng GS, Xie ZF, Zhou M (2019) Geometric design and analysis of face-gear drive with
involute helical pinion. Mech Mach Theory 134:169–196
9. Francis V, Silvagi J (1950) Face Gear Design Factors. Prod Eng 21:117–121
10. Goldfarb V, Barmina N (eds) (2016) Theory and Practice of Gearing and Transmissions. In
Honor of Professor Faydor L. Litvin. Springer International Publishing Switzerland, Chan
Heidelberg
11. Guingand M, de Vaujany J-P, Jacquin C-Y (2005) Quasi-static analysis of a face gear under
torque. Comput Methods Appl Mech Eng 194:4301–4318
12. Handschuh RF, Lewicki DG, Bossler RB (1994) Experimental testing of prototype face gears
for helicopter transmissions. Proc Inst Mech Eng Part G J Aerosp Eng 208(2):129–136
13. Heath GF, Filler RR, Tan J (2002) Development of face gear technology for industrial and
aerospace power transmission, NASA/CR-2002-211320, ARL-CR-0485, 1L18211-FR-01001
14. Hirsch MW, Smale S, Devaney RL (2004) Differential equations, dynamical systems, and an
introduction to chaos. Elsevier Academic Press, Amsterdam
15. ISO 1122-1:1998. Vocabulary of gear terms—part 1: Definition related to geometry
16. Kawasaki K, Tsuji I, Gunbara H (2018) Geometric design of a face gear drive with a helical
pinion. J Mech Sci Technol 32(4):1653–1659
17. Kissling U, Beermann S (2007) Face gears: geometry and strengh. Gear Technol 1(2):54–61
18. Korn GA, Korn TM (1968) Matematics hanbook for scientist and engineers, 2nd edn.
McGraw-Hill Inc, New York
19. Lewicki DG, Handschuh RF, Heath GF, Sheth V (1999) Evaluation of carbonized face gears.
In: The american helicopter society 55th annual forum, Montreal, Canada
20. Litvin FL (1992) Design and geometry of face gear drives. Trans ASME, J Mech Des 114(4):
642–647
21. Litvin FL (1998) Development of gear technology and theory of gearing, NASA reference
publication 1406, ARL-TR-1500
22. Litvin FL, Chen YS, Heath GF, Sheth VJ, Chen N (2000) Apparates and method for precision
grinding face gears. U.A. Patent No. 6.146.253
23. Litvin FL, Egelja A, Tan J, Chen DY-D, Heath G (2000) Handbook on face gear drives with a
spur involute pinion. NASA/CR-2000-209909, ARL-CR-447
24. Litvin FL, Fuentes A (2004) Gear geometry and applied theory, 2nd edn. Cambridge
University Press, Cambridge
25. Litvin FL, Fuentes A, Zanzi C, Pontiggia M (2002) Face gear drive with spur involute pinion:
geometry, generation by a worm, stress analysis. NASA/CR-2002-211362. ARL-CR-491-
2002
26. Litvin FL, Fuentes A, Zanzi C, Pontiggia M, Handschuh RF (2002) Face-gear drive with spur
involute pinion: geometry, generation by a worm, stress analysis. Comput Methods Appl
Mech Eng 191:2785–2813
27. Litvin FL, Fuentes A, Zanzi C, Pontiggia M (2002) Design, generation and stress analysis of
two versions geometry of face-gear drives. Mech Mach Theory 37(23):1179–1211
28. Litvin FL, Gonzalez-Perez I, Fuentes A, Vecchiato D, Hausen BD, Binney D (2005) Design,
generation and stress analysis of face-gear drive with helical pinion. Comput Methods Appl
Mech Eng 194:3870–3901
29. Litvin FL, Wang JC, Bossler Jr RB, Chen Y-JD, Heath G, Lewicki DG (1992) Application of
face-gear drives in helicopter transmission. In: Proceeding of the 6th international power
transmission and gearing conference, Scottsdale, Arizona, AVSCOM Technical Report
91-C-036 and NASA Technical Memorandum 105655, 13–16 Sept 1992
818 14 Face Gear Drives
30. Litvin FL, Zhang Y, Krenzer TJ, Goldrich RN (1989) Hypoid gear drive with face-milled
teeth: condition of pinion non-undercut and fillet generation, AGMA Paper 89FTM7
31. Miller EW (1942) Hob for generation of crown gears. U.S. Patent No. 2.304.588
32. Novozhilov VV (1970) Thin shell theory. 2nd augmented and revised edition.
Wolters-Noordhoff Publishing, Groningen
33. Radzevich SP (2012) Dudley’s handbook of practical gear design and manufacture, 2nd edn.
CRC Press, Taylor & Francis Group, Boca Raton
34. Tan J (1998) Apparates and method for improved precision grinding of face gears. U.S. Patent
No. 5.823.857
35. Townsend DP (ed) (1991) Dudley’s gear handbook, 2nd edn. McGraw-Hill, New York
36. Tricomi FG (1956) Lezioni di Analisi Matematica – Parte Seconda, 7th edn. CEDAM-Case
Editrice Dott. Antonio Milani, Padova
37. Ventsel E, Krauthammer T (2001) Thin plates and shell: theory, analysis, and applications.
Marcel Dekker Inc, New York
38. Zalgaller VA, Litvin FL (1977) Sufficient condition of existence of envelope contact lines and
edge of regression on the surface of the envelope to the parametric family of surfaces
represented in parametric form. Proc Univ Math 178(3):20–23 (in Russian)
39. Zalgaller VA (1975) Theory of envelopes. Publishing House Nauka, Department of Physics
and Mathematics, Moscow
40. Zanzi C, Pedrero JI (2005) Application of modified geometry of face gear drive. Comput
Methods Appl Mech Eng 194(27–29):3047–3066
Index of Standards
B C
Balbayev, G., 703 Cai, X., 599, 603
Ball, R.S., 98, 418 Capelle, J., 192
Bansal, R.K., 352 Carnevali, L., 325
Barmina, N., 775 Cattaneo, C., 99
Barton, H.R., 634 Caubet, J.J., 114
Basstein, G., 775 Ceccarelli, M., 418
Baxter, M.L., 572, 622 Chaing, W.-S., 622
Beermann, S., 775 Chakraborty, J., 774
Belfiore, N.P., 97 Champman, A., 301
Bertsche, B., 634, 717 Chang, Y.L., 577
Bhadoria, B.S., 774 Chauhan, V., 224
Bilgram, H., 590, 596 Chen, B.K., 560, 703
Binney, D., 778, 787 Chen, C., 703
Bi, Q., 560 Chen, D.Y-D., 773, 805
Birch, T.W., 717 Cheng, H.H., 44
Björke, Ö., 3, 37, 63, 126, 217 Chen, N., 778, 805, 814
Bloomfield, B., 774 Chen, Y., 554, 559–562
Boothroyd, G., 3, 63, 126, 217 Chen, Y.H., 554, 559–563
Bossler, Jr. R.B., 775, 778 Chen, Y-J.D., 775
Tuplin, W.A., 40, 181 559, 569, 570, 572, 582, 585, 588, 599,
635, 702, 777
U Wolff, W., 540, 541
Uesugi, S., 560 Woodbury, R.S., 225
Uicker, J.J., 705, 708 Woodbury, R.W., 3, 63
Wright, M.T., 702
V
Vaisman, I., 454 X
Valentini, P.P., 703 Xie, Z.F., 773
Vecchiato, D., 46, 513, 703, 778, 787 Xue, H.L., 703
Veldkamp, G.R., 98
Ventsel, E., 528, 783 Y
Vimercati, M., 610, 638 Yamashina, H., 225
Visconti, A., 303 Yang, F., 703
Visual Numeric, Inc., 522 Yang, X.H., 703
Volkov, A.E., 572 Yang, Z., 561
Vullo, V., 12, 32, 37, 565 Yan, W.Y., 560
Yu, D.D., 141
W Yu, N., 561
Walker, E., 164
Wang, A.G., 622 Z
Wang, J., 638 Zablonski, K., 417
Wang, J.C., 181, 775, 778 Zalgaller, V.A., 41, 792, 809
Wang, L.T., 577 Zanzi, C., 774, 775, 778, 782, 787, 804, 807,
Wang, P-Y., 575 814, 816
Wang, X., 561 Zaretsky, E.V., 703
Warwick, A., 717 Zhang, G.H., 559–562
Weber, C., 526 Zhang, H., 554, 559, 562, 703
Weisbach, J., 113 Zhang, T.P., 560
Wildhaber, E., 572, 629 Zhang, Y., 572, 793, 803
Willis, R., 710 Zhang, Z., 560
Wilson, C., 412 Zhao, Y.P., 560
Wilson, W.G., 717 Zhou, M., 773
Winter, H., 7, 16, 64, 75, 106, 110, 112, 139, Zhou, X., 181
223, 266, 288, 301, 315, 321, 332, 351, Zimmer, M., 469
364, 366, 388, 391, 393, 412, 414, 417, Zwirner, G., 423
Subject Index
worm(s), 465, 470, 503, 518, 523, 524, 244, 246, 250, 251, 258, 271, 275, 280,
526, 528, 530, 532, 535–537, 539–541, 287, 289, 291, 292, 303, 309, 319, 333,
544, 551 383, 534–542, 552, 585, 588, 589, 592,
Arc of action, 28, 57, 59, 61, 120, 121 648, 684, 787, 796
Arc of approach, 28, 29, 59, 89–92, 95, 96, 120 Base cone
Arc of recess, 28, 29, 59, 89–91, 95, 96, 120 angle(s), 362, 386, 389
Arc of transmission, 574, 579, 580 Base cylinder(s), 304–308, 311, 313, 314, 316,
Arhnold-Kennedy instant center theorem, 98 326, 332, 334, 431–433, 435, 439, 442,
Arm, 6, 111, 117, 408, 618, 703 443, 451, 454, 455, 472, 487, 500, 501,
Attracting force, 642 502, 524, 526, 535, 538, 551, 588, 788
Automatic transmissions, 706, 717, 740 Base diameter(s), 40, 374, 472
Automobile differential, 707 Base helix
Auxiliary angle(s), 156, 512, 648, 669–672, angle, 304–306, 309, 311, 314, 318, 326,
674, 679, 681, 684, 685, 689, 690 331, 332, 431, 452, 459, 471, 524, 601
Average efficiency, 116, 119, 123–125, 341, Base pitch
465 deviation, 33
Average rolling velocity, 104, 105 Basic efficiency, 738, 755, 757, 758, 761,
Average sliding velocity, 106, 449 764–768
Axial advance, 22, 304, 480 Basic rack, 23, 26, 27, 30, 39, 40, 69, 71–74,
Axial load or force, 303, 319, 329, 335, 337, 76, 78, 130, 144, 154, 163, 225, 235,
399, 450, 544, 604, 606, 635, 639, 640 239, 255, 311–313, 316, 354, 357, 368,
Axial module, 484–486, 540, 551 380, 382, 393, 425, 659, 662
Axial pitch, 135, 303, 312, 313, 315, 332, Basic ratio, 709, 761, 764–766, 768
483–485, 489, 491, 548, 554, 555 Basic train, 6
Axial pressure angle, 29, 487, 540, 542, 544 Bending fatigue life, 271
Axial profile, 23, 469, 476, 551, 559 Bending fatigue strength, 271
Axial runout, 34 Bending strength, 73, 76, 78, 159, 232, 276,
Axial sliding, 444 369, 382, 384, 651–653, 663, 664
Axial thrust, 335–337, 599, 604–606, 636, 639, Bernoulli-Euler-Goldbach double generation
640 theorem, 172
Axis of screw motion, 418, 421, 607 Bevel gear
Axis of symmetry, 72, 163, 165, 166 differential(s), 735, 739, 740
Axis of the helical relative motion, 418 planetary gear train, 707
Axode(s), 14, 16, 21, 352, 353, 418, 420, 421, with arrow teeth, 574
429, 607, 610, 631, 634, 780, 781, 784 Bias-in contact, 584
Bias-out contact, 584
B Bilgram cutting process, 590
Back-cone Blank(s), 21, 31, 62, 75, 99, 127, 130, 131,
angle(s), 354 133–135, 149, 150, 168, 169, 200, 201,
distance(s), 355, 370 215, 217–219, 225, 257, 282, 336, 342,
Backlash 345, 347, 368, 390, 409, 410, 427, 610,
free contact, 234, 248–250, 253, 281, 282, 616, 644
284, 288, 293, 294, 297, 326, 385, 387, Bottom clearance(s), 25, 73–75, 85, 162, 179,
396, 397 486
Ball’s reciprocity relation, 98 Bottom land, 21, 76, 230, 370
Barrelling, 23 Box product, 496
Base circle(s), 27, 39–46, 47, 49–57, 59, Braking torque, 695, 742, 743, 747–750, 767
61–63, 65, 66, 69–71, 76, 77, 88, 91, 96, Brandenberger method, 578
99, 100, 103, 110, 113, 120, 127, 132, Brandenberger spiral bevel gears, 590
133, 139, 149, 154, 155, 159–161, 170,
176, 182–185, 190, 191, 198, 202, 207, C
211, 213, 224, 226, 228, 233, 237, 243, CAD/CAM systems, 367, 407, 410
Subject Index 829
Canonical crossed helical gear, 433, 444 Cone apex (or cone center), 352–355, 357,
Canonical design, 433, 435, 436 359, 362, 370, 570, 579, 580, 605, 628,
Carrier, 6, 703–716, 719, 731–738, 740, 742, 644
748, 751–755, 758, 760–764, 767, 769, Cone distance, 355, 359, 365, 370, 371, 618,
775 635, 644, 647–649, 652, 654, 657, 666,
Case-hardened steel, 414, 652, 653 668, 673, 674, 677, 683, 688–690
Cast iron, 547, 653 Cones of normals, 452
Center distance, 7, 8, 10–12, 15, 17, 24, 27, 31, Conical cycloid, 586
35, 36, 40, 50, 52–54, 61, 67, 68, 85, Conical helix(ces), 577
114, 115, 128, 130–132, 136, 142, 144, Conjugate curve(s), 39
145, 151, 153, 162, 182, 200, 204, 206, Conjugated action, 15
207, 211–213, 215, 217, 223–225, 227, Conjugate profile(s), 15, 16, 39, 97, 99, 102,
232–240, 248, 249, 251–254, 256, 105, 108, 140, 185, 186, 320, 364, 528,
258–266, 270–272, 279–284, 288, 532
290–294, 297, 301, 318–320, 325, 327, Constant slot width, 645–647
329, 330, 374, 418, 428, 429, 433, Contact circle(s), 161
435–439, 443, 444, 446, 451, 452, 460, Contact of first order, 140
461, 465–467, 476, 478, 479, 483, 486, Contact of second order, 140
490, 492, 493, 504, 510, 523, 553, 561, Contact ratio
607, 608, 612, 613, 617, 657, 669, 673, factor, 333
684, 720, 721, 727, 806, 807 Contact stress(es), 71, 76, 268, 661, 778
Center of contact, 99 Contact temperature, 271
Centrode(s), 16, 101, 140, 429, 585, 586 Continuous Variable Transmissions (CVTs),
Chamfering 717
sliding, 112 Contrate gear
Characteristic efficiency, 738, 755 pair, 773
Characteristic ratio, 709–711, 736, 755 Conventional involute curve, 45, 47
Chordal tooth space, 345 Conventional pitch surface(s), 482
Chordal tooth thickness, 64, 345, 687 Coordinate line(s), 494–496
Circles(s) of friction, 125 Correction
Circular-arc profile(s), 16, 18 factor, 657
Circular cone worms, 559 for undercut, 224
Circular pitch, 23, 24, 29, 32, 59–62, 66, 83, Coulomb’s law
93, 116, 120, 121, 129, 135, 211, 212, of friction, 114
250, 315, 346, 347, 555, 598, 718 Counterpart rack, 163
Circumferential backlash, 26, 64, 75 Crackle ratio, 101
Clearance Crest, 21–23, 29, 30, 148, 153, 244, 256, 280,
factor, 659, 663 392
CNC-gear cutting machines, 577, 642 Cross-axial shaving, 448
Coast side Crossed gear(s), 7, 8, 11, 22, 806
flank loading, 640, 642 Crossed helical gear(s), 9, 303, 417, 418,
Coefficient of friction, 11, 50, 111, 119, 277, 427–433, 435, 436, 438–442, 444,
340, 406, 544, 546–548, 636, 637 447–451, 454, 456–461, 466, 479, 482,
Coefficient of sliding friction, 114, 339, 450 543, 607, 696
Combining gear, 775 Crossing angle, 610, 800
Complementary rack, 71 Crossing point
Completing cutting method, 409 to mean point along pinion axis, 649
Composite deviation texts, 34 to mean point along wheel axis, 649
Compound epicyclic gear train(s), 6, 706 Cross product, 46, 102, 496, 497, 499, 502,
Compound ordinary gear train(s), 697, 698 618, 791–793, 808, 811
Compound planetary gear train(s), 6, 705, 706, Crown gear
716, 717, 741, 744–747, 749, 750 pinion pair(s), 9
Compound train, 6 to cutter center, 582, 648, 669, 673, 684
830 Subject Index
Crowning, 2, 23, 34, 325, 407, 410, 413, 436, Cylindrical worm
553, 584, 778 drive(s), 558, 559
Crown point, 355, 356 wheel, 10, 467, 559, 560
Crown to crossing point, 355, 644
Crown wheel, 10, 344, 353, 354, 356–358, D
364–369, 375, 377, 379, 380, 382, Damage(s), 27, 81, 110–112, 599, 707
384–389, 393, 395, 397, 398, 408–411, Datum circle, 169, 172, 173
425, 426, 570, 573–591, 595–599, 608, Datum line, 71–73, 163, 164, 168, 226–230,
609, 613–616, 619–621, 631, 632, 635, 241, 242, 394, 470
660, 661, 688, 689 Dedendum
Cumulative factor, 104 angle modification, 645
Cumulative semi-velocity vector, 104, 105 angle(s), 356, 390, 584, 644–647, 659, 679
Cumulative velocity vector, 104 circle(s), 23, 24, 30, 66–68, 106, 112, 151,
Curved-toothed bevel gear(s), 407, 569–572, 153, 159, 162, 163, 169, 177, 179, 182,
580–582, 590, 597, 598, 606, 642, 647 185, 187, 203, 206, 207
Curve of action, 363 factor(s), 659, 661
Curvilinear coordinates, 495, 792 flank, 22, 28–30, 58, 79, 103, 109, 141,
Cutter 142, 144, 145, 147–149, 185
active addendum, 163 line, 63
datum-line, 163 surface, 21, 22
offset, 223, 224, 230 Density, 707, 775
radius, 577, 582, 644, 645, 647–649, 657, Depth factor, 659, 662
658, 669 Depth taper, 644–647
reference circle, 169 Design analysis, 549
tip circle, 169 Design helix, 34
tip line, 163 Design profile(s), 33
Cutting circular pitch, 211 Diameter
Cutting interference, 2, 69, 87, 132, 139, 148, quotient of worm, 485, 486, 800
149, 154–156, 158, 160, 163, 175, 176, Diametral pitch, 24, 39, 60, 68, 70, 225, 484
213, 217, 228, 232, 237, 241, 242, Differential
244–247, 265, 266, 270, 271, 287, 375, drive(s), 732
377–379, 394, 398, 478, 485, 539, 540, gear train(s), 695
575, 645 gear unit, 735
Cutting module, 211 planetary gear trains, 732, 735
Cutting pitch circle(s), 130, 172, 173, 201, 211, torque splitter(s), 723
212, 226, 228, 230, 232–234, 237, 239, Dimensionless wear coefficient, 111
244, 248, 250, 251, 256, 257, 259, Direct drive, 715
289–292, 295, 387, 389 Direct motion, 701
Cutting pitch cone(s), 366, 369, 386–389, 588, Direct problem, 223, 261, 262
589, 613, 615, 777 Disk-type milling cutter, 341, 342, 344, 407,
Cutting pitch line, 164, 168, 228–230, 232, 472, 473, 476, 560
241, 242, 394 Displacement function, 10
Cutting pressure angle(s), 211, 212, 219, 291, Division ratio, 586, 587
292, 386 Double-enveloping recirculating ball worm
Cycle of engagement, 36, 57 drive, 563
Cycloid profiles(s), 16, 18, 127 Double-enveloping worm drive(s), 560
Cycloid(s), 12, 164–166, 586 Double-enveloping worm gear pair(s), 10, 465,
Cylindrical gear(s), 7–10, 22, 23, 40, 60, 71, 554, 555, 557, 558
331, 336, 351, 354, 357, 369–373, Double flank composite testing, 34
380–382, 384, 388, 392, 393, 395, 401, Double-helical bevel gears, 9, 591
412, 417, 424, 441, 448, 533, 569, 591, Double-helical gear(s), 7–9, 301, 336, 337,
598, 602, 657, 659 344–346
Cylindrical helix(ces), 303–306, 314, 318, 321, Double ruled hyperboloids of one sheet, 421,
323, 324 614
Subject Index 831
Double spiral bevel gears, 574, 591 External bevel gear(s), 9, 354, 356, 357, 359,
Driven machine, 3–5, 450, 459, 468, 545, 651, 360, 374
696, 699 External helical gear(s), 8, 757
Driver, 1, 113, 114, 116, 640, 641, 697, 717 External spur gear(s), 7, 49, 139, 141, 142,
Drive-side 157, 158, 183, 184, 223, 234–236, 247,
flank loading, 640, 641 249, 287, 296, 757
Driving machine, 3, 4, 459, 468, 545, 549
Dual-lead worm drive(s), 561 F
Dynamic transmission error, 36 Face advance, 315, 579
Face angle(s)
E of blank, 644
Edge contact, 435, 443, 514, 800, 804 Face apex beyond crossing point, 643, 680
Edge of regression, 496 Face contact ratio(s), 650, 654, 656
Effective addendum line, 164 Face gear(s)
Effective driving torque, 126 pair, 773–778, 780–784, 789, 799, 800,
Effective face width, 602 803, 804
Effective pressure angle, 637, 660, 661, 677, Face-hobbing process, 583, 585, 588, 628, 648,
689–691 667
Effective tip line, 164 Face-milling process, 577, 582, 628, 647, 648,
Efficiency, 1, 4, 5, 11–14, 81, 113, 119, 120, 657, 667
123–126, 278, 338, 341, 401, 406, 448, Face width
451, 454, 458–460, 465, 468, 483, 485, angle, 489
542, 543, 545–549, 552, 553, 558–560, factor, 648, 657, 666, 669, 673, 676
569, 607, 634, 636–638, 695, 700–702, Fatigue life, 271
707, 734, 736–739, 746, 749, 751, Feed rate of cutter, 343
755–770 Fellows cutter, 154
Elliptical gear(s), 17, 19, 100 Fellows gear shaper(s), 86, 131, 287, 316, 317
Elliptic point, 528 Fellows shaping machine(s), 87, 131
Eloid Fictitious average sliding velocity, 106
spiral bevel gears, 585 Fictitious rack, 71, 163
End mill(s), 147, 341, 342, 345, 408 Fillet
End relief, 2, 23, 34, 435 circle, 30, 159, 161, 162, 168, 173–175,
Energy lost by friction or lost work, 339, 403, 203–206, 209, 210, 213, 214, 258
454 interference(s), 139, 141, 158, 159, 161,
Engine power, 459 162, 175, 177, 181, 183, 203–207
Entering side, 515 profile, 139, 159–164, 166, 167, 169, 172,
Entrainment velocity, 99, 105 177, 178, 207, 208, 213
Enveloping worm radius, 30, 66, 68, 73, 74, 76, 154, 238,
wheel, 467, 479, 487, 488, 560 274, 541
Epicyclic gear Final drive(s), 735
train(s), 6, 696, 703–706, 717, 736 Finish stock, 75
Epicycloid(s) First law of gearing, 1, 15, 98, 100
base circle, 585, 648, 684 Five-cut process, 582
Euler-Savary equation, 97, 98, 178 Fixed axode, 353
Evolute Fixed setting process, 582
curve(s), 140 Flame-hardened steel, 653
Evolvent of the circle, 39 Flank line(s), 22, 31, 33, 34, 330, 331,
Extended cycloid, 165 572–576, 579, 580, 582, 586, 588, 590,
Extended epicycloid(s), 172, 173, 177, 178, 591, 598, 599, 615, 626
574, 583, 585, 586, 632 Flank twist, 584
Extended hypocycloid(s), 190, 191, 196, 347, Flash temperature, 271
574, 586, 591 Follower, 114, 116, 407, 622, 697, 702, 778,
Extended involute curve(s), 45, 47, 48, 166 779, 787
832 Subject Index
Formate-cut process, 365, 582, 630 Gear ratio, 6, 19, 20, 62, 84, 85, 87, 145, 152,
Form cutting method(s), 126, 128, 130, 341, 186, 194, 224, 237, 238, 270, 276, 278,
345, 407, 408 279, 319, 351, 372, 391, 394, 407, 440,
Form-milling cutter, 126–128, 133, 408 441, 467, 468, 482, 489, 510, 552, 575,
Form number of the worm, 485, 552 579, 600, 608, 634, 652–654, 656, 663,
Formula of Poncelet, 120, 340, 341 665, 697, 698, 706, 722, 757, 778, 781,
Fouling, 183, 188, 190 783
Fretting, 111 Gear reduction unit(s), 3
Friction amplification factor(s), 458 Gear shaper(s), 86, 131, 147, 169, 287, 316,
Friction angle, 277 317
Friction force(s), 1, 20, 110, 111, 124, 277, Gears with zero backlash, 146, 281
278, 333, 352, 398, 403, 450, 453, 603 Gear tooth proportion(s), 659
Friction loss(es), 271, 468, 476, 546, 759 Gear tooth sizing, 64, 70, 317, 318
Front crown to crossing point, 644 Gear train(s), 3, 6, 7, 18, 20, 39, 62, 100, 146,
Front rake, 287 281, 695–717, 719–727, 729–738,
740–762, 764, 765, 767, 769, 770, 775
G Gear wheel(s), 1, 6, 7, 9, 15, 18, 21–23, 27, 30,
Gaussian coordinates, 495, 623 31, 33–36, 40, 47, 49, 50, 52, 54, 59, 62,
Gaussian curvature, 528, 782 63, 71, 74, 76, 78, 79, 81–83, 88, 94,
Gear, 14 109, 113, 114, 116, 117, 120, 121,
Gear blank(s), 62, 75, 99, 127, 130, 131, 124–132, 134–136, 140, 141, 146, 150,
133–135, 149, 169, 200, 201, 215, 153, 154, 158, 169, 175, 176, 181, 182,
217–219, 225, 257, 282, 336, 410 223, 227, 228, 230, 241, 244, 248,
Gearbox, 3, 35, 36, 121, 125, 200, 699, 707, 251–254, 259, 260, 278, 282, 288–291,
717, 732, 739, 740, 775 301–304, 306, 316–318, 321, 323, 326,
Gear hobber(s), 316, 317 336, 337, 339, 341–347, 353–355, 362,
Gear hobbing machine(s), 133, 343, 346 363, 365, 366, 368, 386, 391, 403, 407,
Gear multiplier unit(s), 3 409, 410, 412, 413, 421–423, 425, 426,
Gear pair(s), 1, 6–15, 18–22, 25, 27, 30, 31, 35, 436, 440, 470, 472, 485, 559, 569–572,
36, 39, 49–54, 56, 57, 61, 62, 64, 67, 68, 582, 583, 586, 587, 589, 597–599,
70, 71, 81–85, 87, 89–100, 106–110, 603–608, 615, 633, 634, 645, 646, 648,
114, 116, 118–120, 123, 126, 130, 131, 654, 657, 688, 689, 697–699, 701–704,
141, 142, 144–147, 150–153, 157–159, 709–711, 720–723, 726–728, 730, 732,
161, 162, 169, 172, 176, 181–187, 191, 733, 738, 739, 754–756, 758, 767, 770
194–196, 200–207, 211, 224, 225, 227, Generated normal pressure angle, 660, 677
232–236, 238–240, 247–249, 253, Generated pressure angle(s), 640, 660, 661
256–259, 261–265, 269, 271, 275, Generation crown wheel, 364, 365, 367–369,
277–281, 283, 291, 297, 306–308, 314, 375, 377, 379, 384, 386, 393–395, 398,
315, 318, 321, 325–327, 329–331, 408–411, 570, 573–591, 596–599,
333–336, 338, 341, 346, 352–356, 613–616, 619, 631, 660, 688, 689
364–366, 370–373, 382, 392, 394, 395, Generation cutting method(s), 126, 128, 130,
406, 414, 417, 418, 422, 424–428, 431, 342, 368, 407–410
433, 435, 436, 439, 444, 446, 447, Generation cutting tool, 150, 812
465–468, 478–486, 489—493, 504, Generation for envelope, 147
508, 512–518, 521, 523–528, 532–535, Generation pitch circle, 228, 295, 297
537–540, 541, 543, 545–549, 552–555, Generation pitch cone angle(s), 359, 384, 385
557–559, 570, 572, 580, 584, 597–599, Generation pitch line, 229, 393, 394
601, 602, 604, 605, 608–611, 613–615, Generation process(es), 86, 88, 130, 147, 149,
620–624, 627–639, 642, 643, 645, 647, 151, 153, 169, 225, 237, 287, 296, 329,
650, 657, 683, 695, 698, 699, 702, 709, 366, 375, 380, 408–411, 469, 572, 573,
720, 722, 727, 755–758, 767, 769, 588, 589, 591–593, 596, 597, 614, 615,
773–778, 780–785, 789, 793, 799, 800, 773, 776, 779, 782, 784, 797, 804, 811,
803–806 816
Subject Index 833
Instantaneous efficiency, 114–116, 124, 340, 103, 110, 113, 114, 125, 127, 130, 135,
341, 454, 457, 458 140, 141, 150, 154, 155, 158–160, 169,
Instantaneous helical motion, 418, 480, 607 171, 175, 184, 185, 198, 207, 209, 213,
Instantaneous line of contact (or of action), 215, 217, 225, 303, 320, 366, 383, 467,
309, 330, 331, 338, 339, 363, 399, 469, 784, 787, 788, 815, 816
401–404, 781, 782 rack, 54
Instantaneous screw motion, 418, 608, 611 roll angle, 39, 42, 43, 46, 65, 787, 788
Interaxle differential, 735 vectorial angle, 43
Interference, 2, 22, 28–30, 55, 57, 69, 83, worm(s), 40, 465, 472, 487, 524, 525,
86–88, 95, 104, 110, 131, 132, 534–539, 541, 542, 544, 551, 552, 560
139–158, 160–163, 167, 175–177, 181, Isogonality factor, 20
183–185, 188, 192, 200, 207, 213, 215,
217, 223, 228, 232, 236–239, 241–247, J
266, 267, 270, 271, 279, 280, 287, 295, Jacobian
296, 323, 374, 375, 377–379, 384, 394, determinant, 803
398, 402, 465, 478, 524, 533, 539–542, matrix, 803
551, 553, 563, 564, 575, 601, 645, Jerk ratio, 101
723–725, 727
Interference in working conditions, 139, 161 K
Interlocking process, 409 Klingelnberg-Palloid spiral bevel gears, 588
Intermediate pinion pitch angle, 652, 667 Klingelnberg-Ziclo-Palloid spiral bevel gears,
Intermediate wheel outer cone distance, 652 587
Intermediate wheel pitch angle, 652
Internal bevel gear(s), 9, 356, 357 L
Internal double helical gear(s), 8 Lash, 281
Internal gear(s), 15, 18, 20, 51, 52, 54, 57, Law of gearing, 1, 39, 51, 98
66–68, 70, 85, 89, 90, 92, 94, 96, 97, 99, Law of motion, 610, 803, 804
106, 107, 110, 114, 116, 132, 141, 171, Lead
172, 181, 183–188, 194–196, 198, angle, 133, 135, 149, 304, 306, 343, 344,
200–205, 207, 211–213, 225, 238–240, 452, 454, 473, 484–487, 489, 500, 534,
259–263, 279, 283, 315 539, 540, 544, 546–549, 552, 553, 555,
Internal helical gear(s), 8, 318, 757 648, 669, 670, 674, 684, 780, 805–807,
Internal meshing worm drives, 560 815
Internal spur gear(s), 7, 81, 181–183, 188, 203, angle of a helix, 22
204, 223, 238, 259, 280, 319, 757 of a helix, 22
Intersecting gear(s), 7–9, 11, 22 Leaving side(s), 515
Interwheel differential, 735 Left-hand teeth, 307, 336, 343, 580
Inverse problem, 223, 261, 262 Length of approach path, 28
Involute Length of contact line, 36
curve(s), 39, 41–49, 56, 88, 96, 99, 131, Length of path of contact, 2, 27, 29, 57, 59,
133, 159, 163, 185, 198, 200, 213, 215, 224, 308, 443, 602
219, 224, 226, 234, 248, 267, 287, Length of path of rotation, 28, 29
303–305, 307, 320, 353, 364, 472, 533, Length of recess path, 28
537, 573, 574, 580, 588 Lengthwise cross-sliding, 444
function, 43 Lengthwise crowning, 407, 410, 413, 436, 584
helicoid(s), 303–306, 314, 427, 471, 588 Lengthwise tooth mean radius of curvature,
interference, 55, 56, 81, 139, 141, 142, 183, 648, 669
186 Limit circle(s), 161, 162, 196, 203, 204, 206
of the circle, 39, 40 Limit contact normal, 629
polar angle, 39, 43, 65, 191 Limit point(s), 55, 203, 206, 534
profile(s), 16, 18, 27, 28, 30, 49, 50, 53, 54, Limit pressure angle, 629, 630, 660, 661, 668,
61, 62, 69, 71, 78, 81, 96, 97, 99, 100, 677, 689
Subject Index 835
Multi-start worms, 467, 469, 485, 547, 552 142–147, 149, 151–154, 176, 177, 182,
186–189, 192, 194, 202, 203, 224, 226,
N 227, 230, 237–239, 241, 244, 247, 258,
Niemann worm drives, 559 265, 270, 272, 276–281, 287, 296, 301,
Nominal center distance, 67, 68, 85, 142, 145, 314, 317, 319–321, 323, 324, 351, 372,
151, 153, 233, 262, 263, 297, 436, 437 374, 375, 377–379, 384, 387, 389, 392,
Nominal diameter(s), 63, 326 394, 397, 398, 431, 443, 447, 467, 492,
Nominal driving torque, 126 523, 525, 540, 541, 551, 558, 575, 578,
Nominal module, 63, 66, 287, 288, 294, 297, 579, 585, 586, 589, 598, 601, 632, 636,
321, 395, 523–525, 600 649, 653, 654, 663, 698, 699, 709, 711,
Nominal pitch 716, 717, 721, 724, 725, 728, 729, 731,
circle, 153, 228, 232, 288, 396 740, 748, 752, 756, 758, 761, 776,
Nominal pressure angle, 63, 70, 151, 280, 287, 780–784, 788, 807, 808, 811
288, 294, 296, 297, 381, 382, 395, 398, Number of threads, 134, 467, 484, 523, 524,
436, 524, 526, 551, 645, 661 540, 548, 552, 780, 808, 811
Nominal quantity(ies), 63, 70, 142, 153, 154
Nominal radius, 63 O
Nominal size, 483 Oblique spiral bevel gears, 574
Nominal space width, 63, 64 Octoid toothing(s), 351, 367, 368, 382, 383,
Nominal stress number 386, 398
for bending, 637, 638 Oerlikon-Spiromatic spiral bevel gears, 585,
Nominal tooth thickness, 63, 319 587, 599
Nominal torque, 113–115 Off-set
Nominal transverse tangential load, 114, 333, face gear pair, 775
401 plane(s), 482, 520, 521, 528, 531–533, 537,
Non-backlash double-roller enveloping 538
hourglass worm drives, 561 Oil-hardened steel, 653
Non-circular gears, 10, 20, 100, 101 On-center face gear pair, 775, 776
Non-standard toothing or gear, 229 One-sheeted double ruled hyperboloids, 421,
Normal 607
backlash, 27, 283, 284, 391, 462, 664, 686 Only recess action gear pair, 279
base pitch, 314, 443 Operating cylinder(s), 14
force, 112, 404, 450, 453 Operating depth, 24, 243
helix, 22 Operating lead angle, 484–486
module, 225, 313, 317, 318, 323, 326, 328, Operating (or working) center distance, 233,
329, 391, 430, 431, 436, 447, 460–462, 248, 249, 252, 253 , 256, 262, 263, 271,
485, 486, 587, 588, 599–602, 632, 677, 279, 284, 288, 293, 327, 329
689, 691 Operating (or working) pitch circle(s), 233,
pitch, 135, 312, 314, 320, 344, 430, 431, 234, 248, 249, 252, 253, 293, 359, 384,
436, 485, 491, 588, 598 387
pressure angle, 29, 313, 324, 326, 329, 344, Operating (or working) pitch cone angle(s),
438, 449, 472, 473, 535, 544, 600, 629, 359, 365, 383, 385, 391
639, 660, 677, 686 Operating (or working) pitch cone(s), 352, 354,
profile, 23, 325, 436, 473, 475 359, 363, 365, 366, 369, 370, 380,
spiral, 22, 635 382–385, 387, 390, 391, 399–401, 403,
standard toothing, 228 610, 623, 625
toothing(s), 229, 233, 370 Operating pitch surface(s), 482
Notch effect(s), 30 Operating surface(s), 14
N-type Oerlikon-Spiromatic spiral bevel gears, Ordinary cycloid, 165
587 Ordinary epicycloid(s), 172, 173, 177–179,
Number of starts, 467, 484, 552 574, 583, 585, 586, 632
Number of teeth, 1, 7, 15, 19, 23, 24, 40, Ordinary gear trains, 695–699, 701–703, 705,
59–62, 66, 68, 70, 71, 73, 76, 78, 81–88, 707–709, 714–716, 719, 736, 741, 750,
120, 123, 128, 131, 132, 134, 136, 755–758, 761, 764, 770
Subject Index 837
Ordinary hypocycloid(s), 172 offset angle in pitch plane, 617, 626, 627,
Ordinary involute curve, 44–46, 48 635, 649, 666, 668, 671, 672, 674, 684,
Origin of the involute, 41, 43, 198 689
Orthogonal curvilinear coordinate(s), 495 offset angle in root plane, 679
Outer normal backlash, 664, 686 pitch angle, 649, 666, 667, 670, 671,
Outer pitch diameter, 356, 649, 651–654, 656 673–675
Outer radius, 576, 799 pitch apex, 648, 649
Outer spiral angle, 576, 580, 683, 685 pitch apex beyond crossing point, 649, 678
Outer transverse module, 599, 602, 635, 654, rack pair, 54, 89, 91
664, 665, 677 spiral angle, 649, 666, 668, 672, 681, 685,
Output power, 2, 4, 11, 14, 459, 545, 700, 733, 689
738, 739, 757, 762, 767 Pinion-type cutter(s), 28, 86, 87, 130–133, 139,
Output torque, 4, 5, 20, 101, 545, 700, 712, 147, 148, 150, 151, 153, 154, 160,
745, 752, 754 169–178, 181, 192, 194, 200–202, 206,
Outside circle(s), 182, 356, 574, 579, 580, 582, 208–219, 224, 264, 265, 279, 280,
587, 588, 590 287–291, 294–297, 318, 324, 336, 343,
Outside cylinder, 488, 489, 532 345, 559
Outside diameter(s), 356, 489, 688 Pitch
Outside surface, 21, 465, 488, 516, 517, 520, angle(s), 262, 352, 354, 356, 436, 589, 622,
522–525, 532, 533, 538, 539 635, 639, 673
Overdrive(s), 716 apex beyond crossing point, 678
Overgear(s), 3 axis, 14
Overlap angle(s), 580 circle(s), 15–17, 23, 24, 26, 28, 29, 50,
Overlap arc, 572, 580 52–54, 57–59, 61–63, 76, 91–93, 95,
Overlap ratio, 314, 315, 335, 572, 574, 575, 96, 99, 101, 103, 106, 108, 112, 116,
599, 602, 635, 653 120, 127, 130, 149, 164, 166, 171, 172,
177–179, 182–185, 190, 195, 197, 198,
P 201, 212, 226, 228, 234, 235, 248–253,
Parallel axis theorem, 505 256, 257, 283, 284, 287, 288, 293, 294,
Parallel gear(s), 7, 8, 325, 707 302, 304, 315, 318, 319, 321, 334, 339,
Parallel helical gear(s), 303, 308, 309, 312, 345, 347, 353, 354, 356, 357, 359, 363,
313, 318, 325, 338 371, 384, 385, 387, 395–398, 405, 470,
Parallel mixed power train(s), 695, 750, 751, 481, 572, 584, 604, 630, 704, 712–714,
753, 754 718, 784, 785, 790, 806, 807
Parametric representation, 44, 493, 788 cone apex, 354, 355, 357, 570, 579, 628,
Partial recess action, 278 644
Path of approach, 89–91, 114–118, 308, 405 cone(s), 22, 352–356, 361, 380, 383,
Path of contact, 2, 27–29, 31, 53, 54, 56, 57, 397–400, 408, 410, 411, 425, 570, 572,
59, 81, 85, 89, 91–93, 102, 103, 106, 574, 577, 581, 584, 587, 598, 600,
108–110, 113, 114, 116, 117, 119–121, 603–605, 608–610, 614, 616–618,
123, 139, 152–158, 161, 168, 174, 177, 622–627, 630, 632, 644, 648–650,
184, 185, 188, 203, 205, 206, 210, 213, 657, 659, 665, 676, 682, 683, 686, 731,
214, 224, 236, 237, 258, 266–268, 777
276–278, 280, 295, 302, 308, 315, 331, curve(s), 140, 482, 532, 533
366, 368, 400, 402, 406, 441–443, 448, cylinder of worm wheel, 481, 482, 490,
451, 456, 534, 602, 787 492, 528
Path of recess, 78, 81, 89–91, 114–116, 118, cylinder(s), 14–16, 22, 131, 184, 303, 306,
308, 405 307, 311, 313, 314, 326, 328, 333, 338,
Pencil sharpener(s), 706 345, 347, 429, 430, 435, 438, 439, 441,
Permissible contact stress(es), 778 442, 448–451, 453, 461, 462, 470, 473,
Pinion 481–485, 489, 491, 493, 517, 524, 538,
offset angle in axial plane, 635, 649, 667, 540, 543, 546, 551, 552, 784
670–672, 674 deviation(s), 33, 36, 130, 131, 134
838 Subject Index
Total helix deviation, 34 Transverse profile, 8, 23, 26, 61, 161, 173, 307,
Total index deviation, 33 322, 323, 325, 497–504, 511, 797
Total length of line of contact (or of action), Transverse space-width, 26
311, 330, 332 Transverse tooth thickness
Total no-load transmission error, 35 modification, 351, 393
Total profile deviation, 33 Tredgold approximation, 354, 371
Total radial composite deviation,, 36, 37 Tredgold method, 371
Total single flank composite deviation, 36 Trimming
Total single flank composite tolerance, 36 factor, 207, 209
Total sliding velocity, 448, 449 Triple-helical gear(s), 8, 337, 338
Tracing point, 40–42, 47, 51, 55, 140, 165, Trochoid
166, 172, 175, 367, 580 interference, 159
Train arm(s), 703 Truncation
Train of gears, 62, 695, 697 coefficient, 255
Transferbox, 735 Type A worm, 526, 527
Transfer gearbox(es), 732, 739, 740 Type C worm, 476, 478, 518, 523, 542, 559
Transmission error, 1, 35, 36, 325, 466, 467, Type I worm, 527
514, 800, 804 Type K worm, 473, 476, 542
Transmission function, 6, 10, 20, 98, 100, 101, Type N worm, 473, 475, 501, 559
751, 752, 804
Transmission ratio(s), 1, 3–6, 9, 11–13, 15–20, U
35, 62, 100, 120, 126, 147, 194, 259, Undercut, 23, 30, 69, 74, 79, 139, 142, 147,
343, 421–425, 430, 432, 433, 435, 438, 149, 150, 160, 163, 175, 213–215, 217,
439, 441, 448, 460, 478, 482, 483, 486, 224, 232, 236, 241, 243, 267, 270, 271,
490, 491, 575, 607, 608, 610, 613, 621, 275, 287, 295, 323, 324, 329, 370, 377,
634–636, 638, 661, 695–702, 708, 709, 379, 382, 384, 392, 441, 478, 486, 539,
711–717, 719, 721, 725, 729, 741, 743, 552, 553, 557, 575, 598, 647, 653, 660,
744, 746, 747, 749–753, 755, 765, 767 661, 663, 688–691, 777, 793, 795, 799,
Transmitted load, 99, 122, 399, 400, 544, 637 812, 816
Transmitted power, 739 Uniform depth, 628, 644–647
Transverse angle(s) of transmission, 57, 61, Unit vector(s), 39, 45, 46, 454, 455, 493, 497,
580 499, 611, 616–618, 626, 627, 629, 630,
Transverse angle of action, 28 790, 791
Transverse arc(s) of transmission, 57, 579 Upper allowance of size, 282
Transverse arc of action, 28 Usable flank, 22, 24, 478
Transverse base pitch Usable power, 459
deviation, 33 Usable profile(s), 24, 139
Transverse circular pitch, 23, 315
Transverse contact ratio, 29, 58, 61, 76, 81, V
93–96, 113, 123, 146, 150, 157, 238, Variation of center distance, 40, 130, 132, 136,
258, 314, 315, 330, 335, 383, 572, 575, 207, 223, 235, 238–240, 247, 254, 258,
602, 660 259, 265, 266, 270, 279, 280, 297, 301,
Transverse diametral pitch, 24, 60 325, 327, 443, 461
Transverse line of action, 53 Vector form of position vector, 420, 493, 497,
Transverse module(s), 1, 23, 24, 60, 313, 318, 499, 500, 504
323, 326, 391, 438, 599, 600, 602, 635, Vector function(s), 44, 46, 623, 624, 788, 792,
654, 664, 665, 677 799, 800, 802, 805, 807–809, 811, 812
Transverse path of contact, 53, 331 Vectorial angle, 43
Transverse pitch, 24, 26, 59, 312, 314, 315, Vector loop equation, 98
344, 371, 431, 489, 492, 579, 588, Vector moment, 505
598 Velocity function, 10
Transverse pressure angle, 29, 271, 313, 326, Velocity ratio, 50, 53, 101, 104, 108, 110, 149
328, 329, 344, 371, 375, 377, 379, 438, Velocity vector(s), 102, 104, 105, 359, 360,
451, 456, 487, 498, 500, 601 403, 418, 420, 429, 433, 444, 446–449,
844 Subject Index
452, 453, 490, 504–507, 509, 607, Working depth, 24, 78, 645, 647, 662, 663, 678
610–613, 616, 619, 626, 627, 713, 793, Working interference, 158, 564
807 Worm
Vibration(s), 13, 31, 36, 147, 325, 435, 468, drives, 465, 558–563
514, 653 composed of a backlash-adjustable
Virtual (or equivalent) cylindrical gear, 371 planar worm wheel, 561
Virtual face width, 332 with a split worm, 561
Virtual number of teeth, 276, 301, 320, 321, face width, 486, 508, 524, 526, 557
323, 372, 384, 585, 598, 601 gear, 8–10, 13, 417, 465–468, 478–486,
488–493, 504, 508, 512–518, 521,
W 523–528, 532–535, 537–542, 545–549,
Wear 552–555, 557, 558, 560, 563, 696, 701
coefficient, 111 pair(s), 9, 10, 465–468, 478–486,
depth, 111, 112 489–493, 504, 508, 512–518, 521,
equation, 111 523–528, 532, 533, 537–540, 541,
strength, 271, 382 543, 545–549, 552–554, 557, 558
velocity, 111 screw, 465
Wheel threads or worm wheel(s), 466, 485, 515,
angle, 669 538, 546, 563
mean slot width, 659 wheel(s) or wormwheel(s), 10, 465–470,
offset angle in axial plane, 649, 663, 667, 476, 478–483, 485, 486, 488–492, 504,
672 505, 508, 509, 512–526, 528, 530–534,
pitch angle, 649, 652, 666, 668, 670, 671, 537–549, 551–565, 701
673–675
pitch apex, 648, 649 Z
pitch apex beyond crossing point, 649 Zero-backlash gear set, 281
rim, 489, 515, 523 Zerol bevel gears, 572, 574, 584, 605, 642,
spiral angle, 649, 668, 684, 685, 689, 690 652–655, 661
Wildhaber/Novikov toothings, 315 Zerol bevel hypoid pinion, 635
Wildhaber-Novikov profile(s), 16 Zerol bevel hypoid wheel, 635
Wildhaber worm drives, 559 Zero-profile-shifted toothing, 224, 229, 233,
Willis formula, 708–711 235–238, 240, 241, 259, 262, 264, 266,
Wilson gearbox, 717 282
Wobble, 34, 37 Zone of contact, 515
Work done by driving torque, 117, 119, 406