CENTRIFUGAL SUPERCHARGER

A final year project submitted in partial fulfilment

of the requirements for the degree of
B-Tech.Mech.Eng.

2005

Anton Gouws
9808 6986
Mnr.J.A.Botes

ii
DECLARATION

I hereby declare that the Centrifugal Supercharger has been submitted as partial
fulfilment of the requirements for the B-Tech.Mech.Eng. at Tshwane University of
Technology. I further declare that the report is my own work and that all the sources
that were used or quoted, was given recognition by means of a complete reference.

_______________________________
Initial(s) Surname (Printed)

_______________________________
Signature

_______________________________
Date

iii
SUMMARY

The contents of this report describe the process of design of a centrifugal

supercharger for internal combustion engines, and more specific, the drive
system thereof.

A supercharger is a device that increases the pressure of the inlet charge of

an internal combustion engine, hence, increasing the density and mass of air
flowing through the engine, and combined with additional fuel, increases
power output. This process is generally known as forced induction. There are
various methods of achieving this result, all with its own advantages and
disadvantages. They are however divided into two distinct groups: Turbo
charging and supercharging.

Turbo charging implies the use of a turbine driven by otherwise wasted energy
in the form of hot exhaust gasses, driving a centrifugal air compressor.
Because of the extremely high rotational speeds obtainable, turbo charger
compressors have a very high adiabatic efficiency, often in excess of 80%.
Superchargers on the other hand may take on the form centrifugal, roots, vane
type or screw compressors. Because a supercharger is driven from the engine
crankshaft, rotational speed is the limiting factor, and because of inherent
properties of superchargers they tend to suffer from very low adiabatic
efficiency, often reaching a maximum of only 60%.

Apart from the fairly large difference in compressor efficiency, there is a much
larger variety of turbo charger compressors available on the market than there
are superchargers.

The content of this report suggests that a drive system can be developed to
utilize turbocharger compressor components driven from the engine
crankshaft resulting in a high efficiency supercharger. It also details the design
procedure for a suggested type of drive system, a planetary traction drive,
which would be capable of driving the compressor at the required speed to
obtain maximum adiabatic efficiency. Compared to conventional methods of
power transfer, the planetary traction drive would result in cheaper
manufacturing costs, lower noise, and no need for carrier bearings on the
compressor impellor shaft capable of handling speeds in excess of
100 000 rpm.

Calculations for the detail design were based on the coefficient of friction
between the rollers, and also the contact stresses between the roller faces for
determining reliability considering surface fatigue. The results suggest that this
is in fact a feasible solution to the suggested scenario.

iv
ACKNOWLEDGEMENT

Because there is no predetermined procedure for deriving calculations for the

design of a planetary traction drive as described in this report, the process was
one of trial and error. Even though this might not be particularly clear from the
report itself, the reader is urged to acknowledge the Excel spread sheet in
appendix C. This spread sheet contains the final values for the calculations
after numerous iteration processes.

v
 Table of Contents
Page

List of Figures...............................................................................................vi
List of notations, symbols and acronyms......................................................vii

1. Literature Review................................................................................1
1.1 History...................................................................................................1
1.2 Centrifugal Supercharger Drive Systems...............................................2
1.2.1 Conventional drive systems
1.2.2 Alternative drive systems
1.2.3 Bearing systems
1.3 Practical considerations in drive system design.....................................3

2. The Test Engine..................................................................................5

2.1 Modifications......................................................................................5
2.1.1 Compression ratio
2.1.2 Fuel system
2.1.3 Other modifications
2.2 Compressor Requirements................................................................8
2.3 Compressor Selection......................................................................11

3. Concept Design.................................................................................13
3.1.1 Electrical
3.1.2 Cog-belts and sprockets
3.1.3 Spur gears
3.1.4 Planetary Traction Drive
3.2 Concept Selection............................................................................17

4. Detail Design:....................................................................................18
4.1 Detail Calculations:..........................................................................18
4.2 Design Layout..................................................................................23
4.3 Conclusion.......................................................................................26

List of References........................................................................................27
Appendix A: Product Design Specification...................................................29
Appendix B: Gantt Chart..............................................................................32
Appendix C: Excel Spread Sheet.................................................................33
Appendix D: Kiss Soft Report.......................................................................34
Appendix E: K600 Properties.......................................................................35

vi
LIST OF FIGURES
Page
Figure 1 Copper Gaskets...........................................................................6
Figure 2 Carburettor Enclosure..................................................................7
Figure 3 Aluminium Cam Gear...................................................................8
Figure 4 Compressor Calculations Spreadsheet (Screen shot).................10
Figure 5 Compressor Map.......................................................................11
Figure 6 Impellor Wheel and Compressor Housing..................................12
Figure 7 Compressor Calculations Spreadsheet.......................................12
Figure 8 Cog-belt and sprockets..............................................................14
Figure 9 Planetary Gear Set.....................................................................15
Figure 10 Planetary Traction Drive.............................................................16
Figure 11 Annulus Deflection.....................................................................21
Figure 12 Planetary Rollers Locating Shaft................................................23
Figure 13 Carrier and Seal Plates..............................................................23
Figure 14 Annulus Fitted............................................................................24
Figure 15 Annulus and Drive Plate Interface..............................................24
Figure 16 Housing and Cover Plate...........................................................25
Figure 17 Impellor Fitted............................................................................25
Figure 18 Compressor Housing Fitted........................................................25

vii
LIST OF NOTATIONS, SYMBOLS AND ACRONYMS

α Rotational acceleration; Coefficient of thermal expansion

δ Deflection
σ Stress
γ Specific heat ratio
b Width of contact area
Cf Coefficient of friction
CM Corrected mass flow
Cp Specific heat capacity at constant pressure
CV Corrected volume flow
d Sun roller diameter
D Planetary roller diameter
Da Annulus diameter
Df Maximum corrected annulus diameter
E Modules of elasticity
Ec Compressor efficiency
Ev Engine volumetric efficiency
Ft Tangential force
HB Brinell hardness
I Second moment of area; Moment of Inertia
K Buckingham's load stress factor
L Length of beam (developed annulus section)
N Sun roller speed
n Safety factor
Na Annulus speed
Nc Compressor impellor speed
Ne Engine speed
Np Planetary roller speed
Pa Atmospheric pressure
Pb Boost pressure
PB Boost power
Pc Contact pressure
PR Pressure Ratio
Pt Total Power
R Specific gas constant
r Radius
Tα Inertial torque
Ta Atmospheric temperature
TB Boost Torque
Tp Planetary roller torque
Tt Total Torque
∆T Temperature difference
V Unblown volume flow
Ve Engine volumetric efficiency
Vs Circumferencial speed
w Face width of rollers

viii
1. Literature Review

1.1 History

The concept of “forced induction” or “supercharging” of internal combustion engines

has been researched for well over 120 years. Gottlieb Daimler, founding father of
Mercedes Benz and Daimler Chrysler as they are known today, patented a forced
induction engine in 1885 [1]. The principle was that of a normal four-stroke engine, with
the addition of charge air being induced, through a non-return valve, into the
crankcase, by the upward travel of the piston. This air charge is then forced into the
cylinder through a valve in the piston that only opens at the bottom end of each piston
stroke. The first V-engine, the Daimler Twin, designed by Wilhelm Maybach, was also
of this type. Needless to say, problems arose with the valve in the piston and
problems with keeping oil inside a pressurized crankcase can also be imagined.

One of Gottlieb’s methods of supercharging utilized a twin rotor “air-mover” invented

and patented in 1860, by Mr. Francis Roots from Indiana [2]. This machine would later
develop into the Roots Blower, as we know it today. Krigar, a German engineer,
invented an air pump utilizing two rotating screws intermeshing with each other, and
transfer air in an axial direction. This type of machine formed the basis for the “twin-
screw” or “Lysholm compressor”, widely used today as a better alternative to the
Roots Blower.

Swiss engineer Alfred Büchi did extensive research on exhaust gas turbo charging,
with some patents dating back to 1905, utilizing multistage axial flow turbines and
compressors [1]. Later it would be determined that compressors of the centrifugal type
were better suited to this method of forced induction, basically because of the
simplicity of operation and manufacture compared to multistage axial compressors.
The combination of the centrifugal compressor of a turbocharger, and the step up
drive system of a supercharger, resulted in the birth of the centrifugal supercharger.

Soon after the turn of the previous century, superchargers started to make their
appearance on race cars, and military aircraft engineers found it an attractive way to
compensate for reduced air density at higher altitudes. The success of the centrifugal
supercharger in motor racing starts as soon as 1908, with Lee Chadwick in the
Vanderbilt cup in Long Island, New York[2].

In the mid 1930’s Robert Paxton McCulloch’s engineering company started

manufacturing centrifugal superchargers commercially, and by 1954 they had
supercharger kits available for almost all production 6 and 8 cylinder engines [2].
This is still the case today, with Paxton being one of the leading centrifugal
supercharger manufacturers including Powerdyne, Vortech, and ATI ProCharger.

1.2 Centrifugal Supercharger Drive Systems

1
1.2.1 Conventional drive systems

Because of the high spindle speeds required by centrifugal compressors (often in

excess of 100 000 rpm.) to perform at optimal efficiency, the supercharger head unit,
more often than not, has an integral step-up drive system in addition to the accessory
belt driving it from the engine crankshaft [3]. In order to obtain the necessary step-up
ratio, various manufacturers uses various methods of power transmission, each with
its advantages and disadvantages.

Vortech, Paxton and ATI superchargers all utilize gear systems to achieve the desired
ratio [3]. The major disadvantage of gears is that they generate a whining noise while
in operation. Noise in such a system can be reduced by the use of helical gears, and
Japanese researchers claim to have optimized the thrust angle of the gears to the
axial thrust imposed on the impellor by air pressure [4]. This resulted in higher
durability in gears driven to speeds of over 100 000 rpm. Another disadvantage is that
lubrication needs to be plumbed in from the engine oil system. The oil provides
lubrication to the bearings and gears as well as cooling of heat generated due to
friction. A drive system like this should last for the life of the engine since oil change
intervals of the engine provides the necessary service procedure for the
supercharger. ATI has however developed a self-contained gear driven supercharger
[3]
. This unit has its own internal oil reservoir containing a proprietary oil that does not
require changing for the life of the system.

Powerdyne on the other hand uses a US patented system based on a cog-belt and
sprockets [5]. This system is completely self-contained and does not require the engine
oil system to be tapped. It has less friction than a gear driven system and hence less
heat and noise is generated. It does however require replacement of the belt 3 around
every 80 000km.

1.2.2 Alternative drive systems

Visteon [6] has developed a Torque Enhancement System (TES), which is basically a
centrifugal supercharger that is driven by an electric motor. Visteon’s success is in the
electronic control system, since the TES requires about 2kW of electricity from the
12/14-volt electrical system, which translates into about 170 amps! For conventional
engine control systems without high output smart-alternators and electronic control
units, a conversion like this, though not entirely impossible, but extremely difficult and
impractical, could end catastrophically. The TES is also limited to engine sizes of up
to 2litre petrol and 3litre diesel engines, and maximum boost pressure is claimed to be
0.4bar on a 1,2litre engine.

A type of drive system that has only recently come under development is the
planetary traction drive. HKS Rotrex[9] and Turbotechnics[10] are currently producing
centrifugal superchargers driven by this type of system, which is basically a planetary
2
gear set utilizing toothless gears. Torque is transmitted between the toothless gears
via friction caused by oil under pressure as a result of a contact force between the
rollers. HKS utilizes a special traction fluid inside the system which is circulated and
cooled, while Turbotechnics uses engine oil with great success. This type of system
eliminates the use of high speed bearings as the impellor shaft is carried by the
planetary rollers, and output speeds of up to
185 000 rpm has been achieved.

1.2.3 Bearing systems

For impellor shaft speeds in excess of 100 000rpm, a very reliable bearing system is
required. Modern centrifugal superchargers rely on angular contact ball bearings with
ceramic rolling elements. Because of the ceramic rolling elements reduced weight and
improved resistance against friction, wear, temperature and impact, these bearings
can stand loads at speeds of over 150 000rpm with little or no external lubrication and
has excellent acceleration capabilities [8]. There are also US patents referring to the
use of compound bearings, which is basically defined as a bearing within a bearing [7].
This will give the compound bearing a much higher speed rating than a single bearing
of the same dimensions.

1.3 Practical considerations in drive system design

From a design and manufacturing point of view, the ideal would be to use of-the-shelf
parts. This would be nearly impossible to do when designing a gear driven
supercharger since gears has to be manufactured according to application. There are
some gears in standard sizes available, but these typically lack the properties that
would be required in a transmission that provides a step up ratio in the region of 16:1.
The solution is to manufacture the gears for the specific transmission. The gears
would have to be of very high quality material and finish in order to perform reliably
under continuous acceleration and deceleration. Because of the strength required in
the teeth, unless exotic materials are used, the use of high quality steel would result in
heavy gears, which in turn raises inertial forces. The use of exotic materials in turn
would raise the cost of material and manufacturing above feasibility. The possibility to
change gear ratios would also result in redesign of almost the whole transmission
including manufacturing of new gears. Apart from the physical properties of the gears,
the transmission housing would have to be designed to incorporate a lubrication
circuit or an internal oil reservoir and some method of distributing oil.

3
Cog-belts and sprockets also provide a positive drive similar to gears. The advantage
of this system is that belts and sprockets can be bought of-the-shelf in numerous
shapes and sizes. This provides for fairly simple adjustment of ratios and simplifies
the design of the transmission housing since the need for lubrication is eliminated.
Should the design require the use of sprockets that are not commercially available,
the manufacturing process is not as sophisticated as with gears. Materials such as
aluminium can be utilized in order to reduce weight. Belt drive systems inherently
generate less noise and heat than gears, and the elasticity of the belt could in some
cases serve as a harmonic damper to reduce vibrations to the impellor, which can
prolong fatigue life. In short, compared to gears, a belt drive system would be cheaper
and easier to design and manufacture, and would run quieter and cooler, with more
possibilities of adjustment, and lower repair cost should failure occur.

2. The test engine

4
For the purpose of this project, a Ford 3.0l V6 engine has been made available for
modification and testing. The goal has been set to pressurize the inlet charge to a
maximum of 1bar at 6000rpm, and not to change the fuel requirements of the engine,
i.e. it should still be capable of operating on normal 93 octane petrol.

2.1 Modifications

2.1.1 Compression ratio

The peak pressure inside a petrol engine is governed mainly by the octane rating of
available fuel. The higher the pressure, the higher the octane rating required. This is
due to the fact that octane in petrol suppresses the tendency of the inlet charge to
detonate spontaneously under conditions of high pressure and high temperature
causing knock, or pinging.

By reducing the compression ratio, or increasing the volume of the combustion

chamber relative to the swept volume of the cylinder, the peak pressure is reduced.
Forced induction can then be utilized to increase the peak pressure to a level were the
combustion of available fuel can still be effectively controlled. Since the volume of the
combustion chamber is now increased and the peak pressure is not more than before,
there is a larger mass of charge available for combustion, and hence the brake mean
effective pressure is increased, resulting in a higher power output.

Since there is no specific guideline for altering compression ratio relative to the
amount of boost, it is accepted as a general rule of thumb, that for a boost pressure of
1bar, compression ratio should be lowered to 8:1 for use with normal 93 octane petrol
available at the reef. This would result in a combination that is tuneable enough to
obtain good results without the risk of detonation.

To reduce the compression ratio of the test engine, it was opted to increase the
thickness of the cylinder head gaskets. The thickness of the stock gaskets is 1.6mm
giving a compression ratio of 9.5:1. Doing simple compression ratio calculations
resulted in a required gasket thickness of 3mm to give an 8:1 compression ratio.
Since 3mm gaskets are not available to by as standard, special gaskets had to be
made. The material used was 3mm thick copper plate. The use of copper plate to
make thicker gaskets for forced induction engines is general practice in the
automotive aftermarket (figure1). The stock gaskets were scanned and modelled in
Solid Works, and a CNC program generated for machining.

5
Figure 1: Copper Gaskets

2.1.2 Fuel system

On engines utilizing carburettors for fuelling, it is recommended that they be enclosed

in a sealed box, for boost exceeding 0.5bar. The reason for this is to obtain a pressure
balance between the float chamber and the “atmosphere” outside. If the pressure
inside the float chamber becomes higher than atmospheric, problems arise with
keeping the fuel inside the carburettor and stop leaks. Problems also may occur with
the carburettor housing distorting since it is not designed for handling pressure on the
inside.

It was opted to build an enclosure for the carburettors using 3mm aluminium plate. A
mock-up of the box was built using 3mm thick cardboard. All the panels were
modelled in Solid Works for the purpose of laser cutting and doing a Cosmos Express
stress analyses to determine whether it would be capable of withstanding the
pressure without exploding. These results were satisfactory.

All the panels were laser cut by KARE Sheet metal products and welded together by
ProQuick welding. The welding was done with the box completely enclosed to avoid
distortion, and afterwards the top and sides were cut out to provide for access to the
carburettors. Rivnuts or expansion nuts were inserted at intervals around the
perimeter of each cut-out for fastening of sealing lids. A 51mm tube was also welded
in position for the fitment of the compressed air tube. The pictures in figure 2 display
the basic procedure up to the finished product with the lids removed.

6
Figure 2: Carburettor Enclosure

The fuel pump also had to be upgraded since the fuel pressure required would be
increased by the amount of boost. This was done by fitting a low pressure electric fuel
pump feeding fuel into a surge tank. The surge tank provides fuel to a high pressure
pump delivering fuel to the carburettors via a pressure regulator.

2.1.3 Other modifications

Apart from the engine being completely rebuilt, some additional modifications were
made to ensure that it would be able to handle the extra power and remain reliable.

The two centre main journals on the crankshaft were grooved for additional oil flow to
all the bearings. The crankshaft was also polished and nitrated for extra strength. All
the bearings are lead copper instead of normal white metal for handling extra
pressure and the connecting rod bolts are from ARP and can handle four times more
stress than the standard bolts. An aluminium cam gear (figure 3) was fitted instead of
7
the standard composite gear which tends to break teeth and destroy the cylinder
heads.

Figure 3: Aluminium Camshaft Gear

The camshaft was reprofiled to a LAMCO R70T, which should result in smooth power
delivery and extra torque from idle through to 6000rpm. All rotating and reciprocating
parts were balanced to reduce vibration. The intake manifold was sent to Ajax
Manufacturing for modification to accept two 38DGAS Weber carburettors instead of
only one as per standard and can be seen in figure 2.

2.2 Compressor requirements:

Knowing what engine would be used, and what the goal is, it is possible to calculate
the power required by the supercharger under the desired conditions. The
calculations were based on example 13.3, p.463 of Applied Thermodynamics by
Eastop and McConkey, 5th edition. The results from these calculations would be of
assistance when selecting a compressor suited to the engine, as well as in the
selection process of a feasible concept.
The formulae used to calculate these requirements are as follows:

Pa + Pb
Pressure Ratio: PR =
Pa

Where: Pa = Atmospheric pressure

Pb = Boost pressure

Ne * Ve * Ev
Unblown Volume Flow: V =
2 * 60

Where: Ne = Engine speed

Ve = Engine volume
Ev = Engine volumetric efficiency

8
  (γ−1)


Ta * PR γ −Ta 
  
Compressor Delivery Temperature: Tb =Ta + 

 Ec 
 

 


Where: Ta = Atmospheric temperature

Ec = Compressor efficiency
γ = Specific heat ratio

V * PR * Ta
Corrected Volume Flow: CV =
Tb

PR * Pa * V
Corrected Mass Flow: CM =
R * Tb

Where: R = Specific gas constant

Boost Power: PB = CM * Cp * ( Tb − Ta )

Where: Cp = Specific heat capacity at constant pressure

PB * 60
Boost Torque: TB =
2 * π * Nc

Where: Nc = Impellor speed

Inertial Torque: Tα = I * α

Where: I = Moment of inertia

α = Rotational acceleration

Total Torque: Tt = TB + Ta

2 * π * Tt * Nc
Total Power: Pt =
60

As a matter of convenience, these formulae were entered into an Excel spread sheet
(figure 4). Values in the pink cells are given information and can be changed. Values
in the blue and green cells are calculated and the values from the blue cells are used
to plot on compressor maps. Some values have also been converted to imperial
units, since the majority of available compressor maps are configured as such.

9
 Figu
re 4: Compressor Calculations Spreadsheet (Screen Shot)

10
2.3 Compressor selection:

The properties of a compressor housing and impellor combination are plotted on a

compressor flow map. These are generated by the manufacturer and indicate the
adiabatic efficiency and impellor speed of the compressor plotted on a graph of
pressure ratio against air mass flow rate. Thus, the air mass flow rate and pressure
ratio needs to be calculated for the engine at the required boost pressure and speed,
and plotted on various flow maps to find a combination with the highest possible
efficiency. For the purpose of this project, during the selection of a compressor, it
should be considered that the impellor speed be as low as possible. This is
anticipated to aid in the simplification of the design, as the required speed ratio would
then be more easily acquired.

For initial selection a random value for compressor efficiency is entered in the
spreadsheet. As the values for calculated mass flow and pressure ratio are plotted,
the value for compressor efficiency in the region of the point of coincidence is entered
in the spreadsheet and new results determined. This process is repeated until all
values on the spreadsheet and the flow map coincide. At this point the impellor speed
can also be read of and entered in the spreadsheet.

Figure 5: Compressor Map

The selected compressor combination is designated as a T04E compressor housing,

with a 60 trim compressor wheel. From the compressor map, the impellor speed was
11
determined to be approximately 93 000rpm and the adiabatic efficiency 80%. These
parts were bought from Louis Lima at Nitrous and Turbo Performance.

Figure 6: Impellor wheel and Compressor Housing

For the calculation of the impellor inertial torque, an impellor was modelled in Solid
Works and a density of 2750kg/m3 (Aluminium) was entered in the material properties
section. The resulting inertia was used in conjunction with an angular acceleration
from maximum impellor speed to stop over one second, as a worst-case scenario.
This value was added to the torque required to deliver maximum performance, as a
compensation for shock loads. The result, combined with impellor speed, gives the
values for power and torque used for specifying the drive system requirements.

Metric: Imperial:
Engine Capasity (cc): 3000.00 183.00 in^3
Engine Speed (rpm): 6000.00
Volumetric Efficiency (%): 100.00
Atmospheric prssure (kPa): 89.00 12.91 psi
Ambient Temperature (C): 30.00
Boost Pressure (bar): 1.00 14.50 psi
Compressor Efficiency (%): 80.00

Pressure Ratio: 2.12

Unblown Volume Flow (m^3/s): 0.15 317.84 ft^3/min
Corrected Volume Flow (m^3/s): 0.25 519.16 ft^3/min
Corrected Mass Flow (kg/s): 0.25 33.18 lb/min
Blower Delivery Temperature (C): 120.93
Impellor Shaft Power (kW): 22.92 30.73 hp

Impellor Shaft speed (rpm): 93000.00

Impellor Shaft Torque (Nm): 2.35

Impellor Moment of Inertia (g*mm^2): 53605.00

Maximum Inertial Torque (Nm): 0.52 Calculated from maximum impellor
shaft speed to stop over 1 second
Total Shaft power (kW): 28.00
Total Shaft Torque (Nm): 2.88

Figure 7: Compressor Calculations Spreadsheet

3 Concept design

Four concepts for drive systems were considered:

12
 1) Electrical
2) Cog-belts and sprockets
3) Spur gears
4) Planetary traction drive

3.1.1 Electrical

This concept simply implies the use of an electric motor to drive the compressor from
the existing vehicle electrical system.
From the relationship between power, voltage and current:

P = VI

Where: P = Power (Watt)

V = Voltage (Volt)
I = Current (Amps)

The resulting current on a vehicle with a conventional 12V electrical system, for a
power of 28kW, is 2333A! Even though this level of power might only be required at
occasional impulse, the power required at constant maximum boost would be 1900A.
These values do not even take into account the efficiencies of the charging system,
the electric motor, and a possible step up drive system that might be needed to reach
the desired speed.

From these simple calculations, the option of driving the supercharger by electrical
power was dismissed as a possibility.

13
3.1.2 Cog-belts and sprockets

The second concept utilizes synchronous belts. The driver pulley A, is connected to
the smaller driven pulley B, by means of a toothed belt C, similar to a timing belt
system on an engine camshaft.

Figure 8: Cog-belt and sprockets

Because of the high output speed required by the drive system, the limiting factor for
this concept is surface, or circumferential speed. During the research process, the belt
with the highest speed rating that could be found was from ContiTech. The CONTI
SYNCHROBELT® offers a maximum surface speed of 60m/s at the pcd of the
pulleys.

Selecting an absolute maximum impellor speed of 100 000rpm, the available surface
speed provides for a pcd of 11.45mm. From the ContiTech catalogue, a 17 teeth
pulley is available with a pcd of 11mm for belts with a MXL pitch (2.032mm). For the
required step up ratio of 15.5:1, a driver pulley with 263.5 teeth is needed; therefore, it
is opted for a two-stage step up system. The speed of the system input shaft can
easily be doubled from the engine speed by means of a belt drive from the engine
crankshaft. This would leave an internal drive requirement of 7.75:1, which requires a
driver pulley with 131.75 teeth. A pulley with 132 teeth and a pcd of 85.38mm is
available, and would result in a total ratio of 15.53:1 and an impellor speed of 93
176rpm.

From the power rating tables for a belt of MXL pitch and standard width of 6.35mm, a
force capability was calculated to be 23N. Calculating the force in the belt from the
total shaft torque of 2.88Nm, and the pitch circle radius of 5.5mm, the result is 523N. It
is clear that these belts will not be able to transmit the amount of power needed to
drive the compressor, and are consequently dismissed as a possibility.

14
3.1.3 Spur gears

D
C

A
B

Figure 9: Planetary Gear Set

For the concept design of a planetary gear set, Bellingham Engineering Development
was consulted. A seven-page report is included in Appendix D, which was generated
using Kiss Soft software. They suggested using a planetary gear set of module 0.75,
with a driven sun gear A of 20 teeth, three equally spaced planet gears B with 70
teeth each on a stationary carrier C, and a driving ring gear D with 160 teeth to result
in a step up ratio of 8:1. The external belt driving the input of the gear system would
have to add an additional step up ratio of 2:1. The calculations were based on EN24
steel with no surface treatment as the material and a tooth width of 7.5mm. It was
suggested that, even though the design is capable of 2 000 000 hours of service,
special attention should be given to lubrication since the circumferential speed is very
high, and that case hardening of the gears would increase the capability of the
system.

This concept eliminates the possibility of designing a self-contained unit, since

constant lubrication would be required. The lubrication system can however be
extended to incorporate the bearings, which could result in better service life. Ceramic
bearings with very high-speed ratings would still have to be used, and this would be
an expensive acquisition. The manufacture of the gears would also have to be
subcontracted since it would not be feasible to manufacture in-house.

15
3.1.4 Planetary Traction Drive

The principal of operation of this type of drive system is similar to that of a planetary
gear set, with the exception of the gears being replaced by toothless rollers.

B
A

Figure 10: Planetary Traction Drive

The sun roller or shaft C would be driven by the annulus A, through the planetary
rollers B, which are carried on the stationary carrier D.
The rollers are in contact under a certain load in order to prevent them from slipping.
Because of the simplicity of the principal of operation of this system, all parts can be
manufactured in house. Another advantage is that since the sun roller is supported
under load from three directions by the planetary rollers, and there is no backlash or
clearance, there is no need for expensive high-speed carrier bearings on the impellor
shaft. Lubrication would have to be supplied from the engine for lubrication of
bearings in the planetary rollers and input shaft, and would also act as a viscous
connection between rolling surfaces, because of very high surface speeds and
contact pressure.

In daily practice, there are not a lot of power transmission requirements for high
speed, high power, and low torque applications; hence, there is not a lot of literature
available on this type of drive system. This does not, however, mean that it is not a
feasible option. Calculations can be based on friction coefficients, radial forces, and
allowable surface pressures, in order to do a detail design.

16
3.2 Concept selection

The only two drive systems capable of transmitting the amount of power and speed
required by the selected compressor, is the planetary gear system and the planetary
traction drive. The following table compares advantages and disadvantages of these
two types of systems:

Planetary gearbox Planetary traction drive

Power Density High High
Self containing No No
Available theory Plenty Very, very little
Manufacturing time High Low
Manufacturing cost High Low
High speed bearings Yes No
Complexity High Low

It is clear from the table above that the conservative choice would be the planetary
gear system. Because of the proven capability of gears as a means of power
transmission, the design process would be aided largely by the abundance of
information and theory on this subject. By choosing this type of drive system the
outcome can almost be guaranteed to be a success. The major disadvantages of this
drive system is the manufacturing process, which would take more time and hence
result in higher costs, as well as the fact that high speed ceramic bearings should be
used for supporting the impellor shaft. These bearings can however be replaced by
plain hydrodynamic bearings since the system has to be supplied with oil from the
engine anyway.

The major disadvantage of the planetary traction drive is the lack of theory on the
subject. Since there is not a huge requirement for power transmission system capable
of transmitting high power at high speed and low torque, not a lot of research has
been done considering materials, surface treatments, pressures, etc. This type of
system has however been used in similar supercharger systems with success, and it
is anticipated that a lot of the theory considered in bearings would be applicable to the
design process. The big advantage is the sheer simplicity of the concept. It would not
require any support bearings on the impellor shaft, since the planetary rollers support
it, and the manufacturing process would be relatively quick and simple.

Even though the planetary gear system would be the conservative choice, it is opted
to further develop the traction drive system because of its simplicity and possibility for
further research.

17
4. Detail Design

Since the overall step-up ratio required to increase the rotational speed from 6000rpm
to 93 000rpm is 15.5:1, it was to decided to design the internal part of the drive
system with a ratio of 8:1, while the external drive from the engine crankshaft to the
compressor input shaft could be utilized to supply an initial step up ratio of between
1.9:1 to 2:1. This would result in some adjustability to the system for fine tuning.

Because there is no set guideline available for the design of a planetary traction drive
as there is for i.e. spur gear or belt drive systems, it was deemed necessary to first do
a basic design, and then analyze the various components and the circumstances
under which they operate within the system to determine whether or not they would
suffice. In order to simplify the whole trial and error process of calculating different
options, all calculations were again done in Excel as an extension of the initial spread
sheet.

4.1 Detail Calculations

Taking the main variables to be the ratio, sun roller diameter, coefficient of friction and
the width of the rollers, the following formulae were used to calculate initial physical
dimensions for all relevant parts and also the forces that would be acting upon them
under operating conditions:
π *d *N
Circumferential speed (m/s): Vs =
60

Where: d = Sun roller/shaft diameter (m)

N = Sun roller/shaft speed (rpm)

Tt * 2
Tangential Force (N): Ft =
d *3

Where: Tt = Total roller/shaft torque (Nm)

Note: The Tangential force is divided by 3 since there will be
three planetary rollers each carrying a third of the force.

Ft
Minimum Radial Force (N): Fr =
Cf

Where: Cf = Coefficient of friction

Note: The coefficient of friction was taken to 0.05 between
steel and steel with lubrication.

The diameter for the annulus is simply obtained by multiplying the sun roller or shaft
diameter with the ratio i.e. 12 x 8 = 96mm, and the diameter for the planetary rollers
were obtained using the following equation:
18
 Da − d
D=
2

Where: Da = Annulus diameter (mm)

d = Sun roller/shaft diameter (mm)

The speeds for the various parts were calculated from a ratio equation:

N Np Na
= =
d D Da

Where: Np = Planetary roller speed (rpm)

Na = Annulus speed (rpm)

The torque required by the annulus is simply obtained by multiplying the sun roller or
shaft torque with the ratio i.e. 2.88 x 8 = 23.04Nm.

The torque transmitted by the planetary rollers was obtained from the following
equation:

Ft * D
Tp =
2

At this stage in the calculation process, having physical dimensions and force
requirements for the various components, it has become necessary to determine what
type of materials would be suitable to provide an acceptable life for the system. The
calculations to follow are based on Herzian contact stresses between the various
components and fatigue life of the surfaces in contact with each other. These were
found chapter 7-18 Surface Strength, Mechanical Engineering Design, by Joseph
Edward Shigley.

Fr (1 E1 ) + (1 E 2 )
Width of the contact area (m): b2 =
w (1 r1 ) + (1 r2 )

Where: w = face width of rollers (m)

E1 = Modulus of elasticity of first roller
E2 = Modulus of elasticity of second roller
r1 = Radius of first roller (m)
r2 = Radius of second roller (m)
Note: For the contact area between the planetary roller and
the annulus, the annulus radius value has a negative
sign indicating a concave radius.
19
 2 * Fr
Contact Pressure (Pa): Pc =
π *b* w

 1 1 
Buckingham’s Load Stress Factor: K 1 = 2.857 * Pc 2 *  + 
 E1 E 2 

K1
n=
Safety Factor for 10^8 cycles: Fr  1 1 
 + 
w  r1 r2 

Pc + 70
Minimum Material Hardness (HB): HB =
2.76

Manipulating the above equation to accept the Brinell hardness as the variable and
the capable contact pressure as a result, a new Buckingham load stress factor was
calculated as an intermediate value to obtain an acceptable factor of safety.

For the annulus to supply the calculated minimum radial force to the system, it has to
be manufactured to a physically smaller diameter than calculated using the ratio. The
reason for this is that the annulus will deform elastically around the three planetary
rollers and also the thermal expansion of the annulus has to be accounted for.

The calculations for determining the deflection of the annulus from its mean diameter,
was based on the following assumption: if the calculated radial force were applied to
three equally spaced positions on the inside of the annulus, it would deflect to a
triangular shape. It is assumed that the triangular shape would coincide with the mean
circular shape in six positions, and these six positions are equally spaced. This
assumption is indicated in Figure 11. The blue circle indicates the mean diameter and
the green line the deflection from this diameter, caused by the forces indicated by
black arrows. The red dots mark the six equally spaced points.

20
 Figure 11: Annulus Deflection

These six positions would indicate points of inflection of stresses, and one segment
between two of these points was developed as a simply supported beam, with a
length of one-sixth the circumference of the annulus, and the calculated radial force
acting in the middle between these two points. The deflection was calculated using
the following equations, found in chapter14 of Strength of Materials for Technicians by
J.G.Drotsky, 2nd edition:

π * Da
Length of the Section (mm): L=
6

W *t3
Second Moment of Area (m^4): I =
12

Where: W = Width of the cross section (m)

t = Thickness of the cross section (m)
Note: The second moment of area was calculated for a
rectangular cross section, and the width and
thickness stated as variables.

Fr * L3
Deflection (m): δ=
48 * E * I
21
The deflection was subtracted from the mean radius to find the corrected diameter,
and the calculations reiterated backwards to find the effective diameter after
correction with the forces applied. This value was virtually indifferent to the required
mean diameter and was left at that.

To determine the correction for thermal expansion, the corrected diameter of the
annulus was assumed to be at 100oC. The following equation was used to determine
the initial diameter at 20oC:

Maximum Corrected Diameter (mm): Df = Di −α * ∆T * Di

Where: Di = Corrected annulus diameter (mm)

α = Coefficient of thermal expansion
ΔT = Difference in temperature

The stress at the annulus surface was also calculated using the simply supported
beam approach.

Fr * L * t
Stress at annulus surface: σ=
8* I

Note: This value would be negative at the annulus outer

diameter, indicating tensile stress, and positive at the
annulus inside diameter indicating compressive
stress.

Since the maximum compressive stress in the annulus would be acting at the same
point as the radial force, the value for the compressive stress due to the contact if the
roller was added to the stress due to bending. This was done to obtain a maximum
compressive stress in the annulus for comparison with material properties.

A printout of the completed spreadsheet is attached in Appendix C.

4.2 Design layout

The sun roller, or impellor shaft, as well as the planetary rollers is to be manufactured
form EN36 steel, hardened to 64HRC, and ground to a 1,6μm surface finish and
concentricity of within 10μm. The planetary rollers are each to be carried on two SKF
22
6000 bearings with no seals. These bearings has a speed capability of 36 000 rpm
when lubricated with oil. The planetary rollers with bearings are shown locating the
shaft in position figure 12.

Figure 12: Planetary Rollers Locating Shaft

Figure 13 shows the planetary rollers and shaft fixed in position between the seal
plate and the carrier plate. The seal plate is designed to fit directly to the compressor
housing and contains a groove for supplying oil to the roller surfaces and bearings.
The carrier plate houses on of the SKF 61804 bearings that would carry the drive
plate.

Figure 13: Carrier and Seal Plates

The annulus is to be manufactured of Bohler K600. This material is classified as a tool

steel and does not become brittle or loose elasticity with hardening. It is to be
hardened to 51 HRC and ground to a 1,6μm surface finish and a concentricity of
within 10μm. The relevant property sheets are available in appendix E. Because the
annulus will be under constant elastic deformation, the interface with the drive plate is
23
done by a dog-teeth principle to allow for some clearance. This is displayed in figure
14.

Figure 14: Annulus Fitted

The drive plate is to be manufactured from EN19 steel with no heat treatment, and will
be carried on two SKF 61804 bearings of which the outer one has a seal on the
outside for oil retention purposes and is located in the cover plate, and the other is
located within the carrier plate (figure 13). Figure 15 shows the interface between the
drive plate and the annulus.

Figure 15: Annulus and Drive Plate Interface

Figure 16 shows the housing and cover plated bolted to the seal plate, and the supply
hole for the oil can be seen on top of the housing. What is not seen in this picture is
the hole at the bottom side of the housing through which the oil will flow back to the
engine oil pan under its own gravity.

24
 Figure 16: Housing and Cover Plate

The Impellor is displayed fitted to the output shaft of the drive system in figure 17, and
figure 18 shows the compressor housing fitting over the impellor and directly to the
seal plate. The clearance between the impellor and housing is between 0.25 and
0.3mm.

Figure 17: Impellor Fitted Figure 18: Compressor Housing Fitted

4.2 Conclusion

The results obtained from the theory and calculations in this report, indicates that the
planetary traction drive, and more importantly the design contained within this report,
would indeed be a feasible option as a high speed drive system for centrifugal
superchargers.
25
As for the progress of the project, at the time of writing this report the project was not
complete enough for the purpose of performing tests. The test engine is however
completed and material for the manufacturing of the drive system has been supplied
and will commence shortly.

The project will continue to progress after the evaluation of this report and it is
believed that this subject could provide for further research beyond the scope of B-
Tech level. It is anticipated that research could be done on the surface fatigue life and
contact stresses between rotating parts of different materials for the purpose of
transmitting torsion through shear, and also the high pressure viscosity change of
various lubricants under the contact pressure between rollers.

26
LIST OF REFERENCES

1] ZINNER, K. 1978. “Supercharging of Internal Combustion Engines”. Berlin,

Heidelberg, New York: Springer-Verlag.

2] Et al. 2003. “Supercharger Tech: Supercharger A-Z (Part 1-3)”.

http://www.superchargersonline.com/content.asp?ID=31

3] Et al. S.a. “Superchargers 101-Introduction to Superchargers”.

http://www.superchargersonline.com/content.asp?ID=15

4] Abstract 9539356: “Development of Centrifugal Mechanical Supercharger –

Optimum Gear specification for Improvement of Durability”. JSAE Review,
Vol.17. 1996: p.89.

5] WHEELER, JR., et al. 2004. United States Patent: 6 763 812. “Centrifugal air
compressor”.

6] Visteon. S.a. “Torque Enhancement System”.

http://www.visteon.com/products/automotive/media/vtes_spec.pdf

7] JONES. 2003. United States Patent: 6 612 747. “Velocity variance bearing
arrangement for impellor shaft of centrifugal supercharger”.

8] GMN. S.a. “High Precision Ball Bearings” Catalog: PM 001/04 E.

9] HKS. S.a. “Superchargers”.

http://www.hksusa.com/categories/more.asp?id=2191

10] Turbo Technics. S.a. “Forced induction – a new approach”.

http://www.turbotechnics.com/docs/supercharger/expo.htm

Additional References

The following references were also studied and repeat and confirm information in
cited references.

11] Et al. S.a. “The Centrifugal supercharger explained”.

http://www.superchargersonline.com/content.asp?ID=21

12] FERRARI, B. 2004. “Compressor Design and Choices: Centrifugal

Compressors”.
http://www.automotivearticles.com/Superchargers_Compressor_Design_and_
Choices_.shtml

27
13] NORMAN, E. S.a. “Supercharge!”
http://www.fefcholden.org.au/techinfo/supercharge/

14] SCHMIDT, F. 1965. “The Internal Combustion Engine”. London: Chapman

and Hall.

15] JUDGE, A.W. 1972. “Motor Manuals Eighth Edition – 1 Automobile Engines”.
London: Chapman and Hall.

16] NEWTON, K. STEED, W. GARRET, T.T. 1996. “The Motor Vehicle”. Oxford:
Butterworth-Heineman.

28
 Appendix A

Product Design Specification

PRODUCT DESIGN SPECIFICATION

29
 for
DATE: 06-02- SENTRIFUGAL AUTOMOTIVE SUPERCHARGER
2005
RELATED SPECIFICATIONS:
Design a supercharger using a turbocharger compressor
ISSUING AUTHORITY:
Gouwspeed Engineering
FOREWORD:
Over the past few years the turbocharger aftermarket in South Africa has grown
extensively. Turbocharger systems, though very effective in converting otherwise
wasted heat energy into engine power, has three major disadvantages: extensive
and expensive modifications to accommodate fitment; heat soak to the inlet charge
from the exhaust turbine; and turbo lag due to the time required for the
turbocharger to spool up. Heat soak can be compensated for by fitment of an inlet
charge intercooler, which in turn adds to modifications and cost. Fitment of a
supercharger driven from a different power source, for example the engine
crankshaft, eliminates the necessity for exhaust modification, and also turbo lag,
since the compressor impellor will accelerate linearly as the engine accelerates.
Heat soak will also be minimized due to the absence of the hot exhaust turbine and
fitment of an intercooler is seldom necessary in a supercharger system.
Superchargers are however very expensive (from ± $2000-00 US up), and a
system can be manufactured locally for cheaper.
INTRODUCTION:
The objectives are:
To design a supercharger system using a turbocharger compressor.
To design the system for a universal type of fitment to most vehicles.
To design the system for easy maintenance and adjustability/tuneability.
SCOPE:
Provide a tuneable universal supercharger system that can be fitted to most
vehicles, be it fuel injection or carburettor, either by trained and/or experienced
technicians, or the d.i.y. enthusiast, with less modification than a turbocharger
system.
DEFINITIONS:
Supercharger – a system driven from the engine crankshaft that raises the
pressure of the inlet charge of an internal combustion engine.
Turbocharger – a system that is driven with a turbine that is driven from the engine
exhaust gasses and raises the inlet of the combustion engine.
Heat soak – heat conducted from one section to another.
Turbo lag – time needed for exhaust gasses to accelerate the turbocharger to
sufficient speed in order to supply required boost.
Spool up – acceleration of turbocharger.
Intercooler – during compression the temperature of the inlet charge rises. An
intercooler is a heat exchanger that transfers heat from the hot inlet charge to
cooler ambient air.
PERFORMANCE REQUIREMENTS:
System must supply inlet air at a maximum pressure of 1bar and a mass flow
30
calculated according to engine size and speed.
System must be driven from engine crankshaft.
Materials such as polished or anodized aluminium and coloured silicon hosed must
be used in order to achieve acceptable aesthetics to satisfy automotive
customisers.
System must last for at least 100 000km of combined city, highway and fast road
driving.
System will operate at under bonnet temperatures of up to 70ºC.
Ex works cost will be dependant on components used.
System must be easily installed, tuned, and serviced.
System must be of a standard acceptable in the automotive performance after-
market.
Noise levels must be kept as low as possible to eliminate any driving
inconvenience.
System must be isolated from heat of the surrounding parts, and also for heat to
the surrounding parts.
Modifications in order to lower the compression ratio of the vehicle engine
sufficiently are up to the customiser.
System must be compatible with both fuel injection and carburetted engines. The
selection of the fuel system is the customiser’s choice. A carburetted setup will be
of the pressurised type and not a suck-through-system.
Turbo compressor selection will be based on calculations of engine size and
speed, so that the operating range of the turbo shaft does not exceed 100 000 rpm,
and an efficiency of lower than 60%.
MANUFACTURE REQUIREMENTS:
All machining must be done in house.
Lightweight aesthetically pleasing materials must be used.
System must be assembled as far as possible from standard of the shelf
components.
ACCEPTANCE STANDARDS:
System must be supplied with a guarantee against defects in material and
manufacturing, but no guaranty or warranty on the life of the system.
Prototype must complete a full life cycle of 100 000km with no failures.
DISPOSAL:
Used systems can be supplied back to the manufacturer for exchange,
reconditioning or disposal as scrap metal, depending on the condition of the
system.

Appendix B
31
Gantt Chart

32
33
 Appendix C

Excel Spread Sheet

34
 Metric: Imperial:
Engine Capacity (cc): 3000.00 183.00 in^3
Engine Speed (rpm): 6000.00
Volumetric Efficiency (%): 100.00
Atmospheric pressure (kPa): 89.00 12.91 psi
Ambient Temperature (C): 30.00
Boost Pressure (bar): 1.00 14.50 psi
Compressor Efficiency (%): 80.00

Pressure Ratio: 2.12

Unblown Volume Flow (m^3/s): 0.15 317.84 ft^3/min
Corrected Volume Flow (m^3/s): 0.25 519.16 ft^3/min
Corrected Mass Flow (kg/s): 0.25 33.18 lb/min
Blower Delivery Temperature (C): 120.93
Impellor Shaft Power (kW): 22.92 30.73 hp

Impellor Shaft speed (rpm): 93000.00

Impellor Shaft Torque (Nm): 2.35

Impellor Moment of Inertia (g*mm^2): 53605.00

Maximum Inertial Torque (Nm): 0.52 Calculated from maximum impellor
shaft speed to stop over 1 second
Total Shaft power (kW): 28.00
Total Shaft Torque (Nm): 2.88

Ratio: 8.00
Sun Roller Dia. (mm): 12.00
Coefficient of friction: 0.05
Face Width of rollers (mm): 20.00

Circumfirential speed (m/s): 58.43

Tangential Force (N): 159.72
Minimum Radial Force (N): 3194.45

Sun Roller: Planet roller: Annulus:

Diameter (mm): 12.00 42.00 96.00
Speed (rpm): 93000.00 26571.43 11625.00
Torque (Nm): 2.88 3.35 23.00

Poisson's Ratio: 0.3 0.3 0.3

Modules of Elasticity (Gpa): 200 200 200

Width of Contct Area (mm): 0.0001859 0.0005257

Contact Pressure (GPa): 1.0941679 0.3868468
Buckingham's load-stress factor: 34204102 4275512.8
Safety factor for 10^8 cycles: 0.9993531 2.5572097
Minimum material hardness (HB): 421.79997 165.52419

Material Hardness (HB): 750

Buckingham's load stress factor: 114280000
Safety factor for 10^8 cycles: 3.3389584

Thikness: Width:
Annulus Cross Section (mm): 5.00 30.00
Length of Section (mm): 50.265
Second Moment of Area I (m^4): 3.13E-10
Deflection (mm): 0.135
Corrected Diameter (mm): 95.730

Deflection with Corrected Diameter (mm): 0.134

Effective Diameter (mm): 95.998

Coefficient of thermal expansion: 1.20E-05

Maximum Corrected diameter (mm): 95.638

35
 Appendix D

Kiss Soft Report by BED

36
37
38
39
40
41
42
 Appendix E

Bohler K600 Material Properties

43
44
45
46