Full Text

INTEGRATION OF MACHINING INSPECTION SENSORS AND
SOFTWARE
INTEGRATION OF MACHINING INSPECTION
SENSORS AND SOFTWARE
By
D. ALAN SAWULA, B. ENG
A Thesis
Submitted to the School of Graduate Studies
in Partial Fulfilment of the Requirements
for the Degree
Master of Applied Science
McMaster University
© Copyright by D. Alan Sawula, April 2013

M. A. Sc. Thesis – D. A. Sawula; McMaster University - Mechanical Engineering.
MASTER OF APPLIED SCIENCE (2013)
McMaster University
(Mechanical Engineering)
Hamilton, Ontario
TITLE: Integration of Machining Inspection Sensors and
Software.
AUTHOR: D. Alan Sawula, B. Eng (McMaster University)
SUPERVISOR: A. D. Spence
NUMBER OF PAGES: xii, 108
ii
Abstract
Ideally, the nominal design of a part or assembly, created with 3D
Computer-Aided Design and Manufacturing (CAD/CAM) software, can be
consistently fixtured and machined. In reality, process conditions vary, and
feedback and correction methods such as integrated on-machine inspection,
analysis, and process adjustment, are required.
On-machine inspection based on touch trigger probes is well
established, but limited motion control computing capability restricts analysis
to simple arithmetic. This prevents on-line use of known whole part
mathematical analysis software that implements the part salvaging
intentions of modern Geometric Dimensioning and Tolerancing (GD&T)
standards. Additionally, no CNC integrated method exists for geometrically
adjusting nominal tool paths so that an in-tolerance final part is produced.
Machine tool support for high data rate sensors such as laser scanners is also
lacking.
This thesis reports progress towards bidirectional integration of
machine tool mounted inspection sensors with GD&T analysis software, and
subsequent toolpath adjustment. The concepts are demonstrated using a
fixture consisting of three datum spheres and a workpiece. The fixture is
clamped in the CNC machine, datum spheres are measured, and after
mathematical data fitting and registration, an in-tolerance final part is
iii
produced. To facilitate multiple tests, a tool path is split into four and
machined in four poses with measurement and tool path adjustment for each
pose. Preliminary integration of a laser scanner with axis scales and
computer software was also accomplished.
iv
Acknowledgements
I would like to thank my supervisor Dr. Spence, for his patience,
generosity, and support over our years working together. I have had the
opportunity to meet and work with many interesting and intelligent people,
and I thank Dr. Spence for giving me that opportunity. When faced with
setbacks, Dr. Spence made sure we found a solution, and that both textbook
and life lessons were learned. I am grateful for his dedication.
None of this work would be possible without the following people. Terry
Wagg, for CNC training, Joe Verhaeghe for electronics assistance, and Jim
McLaren, Mark MacKenzie and Ron Lodewyks for help in the machine shop.
Thank you to Yu Pin Lin and Hazem Mazhar for their assistance with
laser scanning equipment and software.
Funding support from NSERC Canadian Network for Research and
Innovation in Machining Technology (CANRIMT) is sincerely appreciated.
To my family and friends for all their unconditional support I am very
grateful.
To Lisa, your patience and encouragement, are more than I could ever
ask for.
v
Table of Contents
Chapter 1 Introduction ........................................................................... 1
Chapter 2 Literature Review .................................................................. 6
2.1 Measurement in the Manufacturing Process .............................. 6
2.1.1 Static Sensor Specifications .................................................... 9
2.1.2 Touch Trigger Probes ............................................................ 11
2.1.2.1 Kinematics of Touch Trigger Probes .................................. 13
2.1.2.2 Relevant CNC G-Codes for Touch Trigger Probes ............ 15
2.1.2.3 G01 Linear Motion Command ............................................ 15
2.1.2.4 G31 Skip Command ............................................................ 16
2.1.3 Laser Digitizers...................................................................... 16
2.2 Data Fitting ................................................................................. 21
2.2.1 Rigid Coordinate System Registration ................................. 21
2.2.2 Orthogonal Least Squares Geometric Fitting ...................... 23
2.2.3 Orthogonal Least Squares Fitting on Data Sets .................. 26
2.2.4 Commercial Point Cloud Software ........................................ 28
2.3 Summary ..................................................................................... 29
Chapter 3 Hardware Description ......................................................... 30
3.1 CNC Machine Tool ...................................................................... 30
3.2 Touch Trigger Probe Integration ................................................ 30
vi
3.3 Hardware Overview .................................................................... 32
3.4 Ax9150 UMI ................................................................................ 33
3.4.1 Installation ............................................................................. 34
3.4.2 Passing coordinate data to PC .............................................. 34
3.5 Calibration of Touch Trigger Probe ............................................ 35
3.5.1 Calibration Tool Path ............................................................ 36
3.5.2 Results of Calibration Data ................................................... 38
3.6 Three Sphere Probing Routine ................................................... 39
Chapter 4 Registration Experiment ..................................................... 42
4.1 Experiment Tool Path ................................................................. 42
4.2 Tool Path Adjustment ................................................................. 43
4.2.1 Part Registration in Manufacturing Processes .................... 43
4.2.2 Tooling Sphere Registration Method .................................... 43
4.3 Test Cases .................................................................................... 45
Chapter 5 Registration Experiment Results ....................................... 49
5.1 Laser Scan of Finished Workpiece ............................................. 49
5.2 Analysis of Expected Errors ....................................................... 59
Chapter 6 Laser Scanner Integration .................................................. 61
6.1 LMI Scan Head Integration ........................................................ 63
6.2 Mount Design .............................................................................. 64
6.3 Desktop Control Software ........................................................... 66
vii
6.4 Expected Errors ........................................................................... 71
6.5 Test Data ..................................................................................... 72
Chapter 7 Conclusions and Future Work ............................................ 78
7.1 Future Work ................................................................................ 79
References 109
viii
Figures and Tables.
Figure 2.1 - Accuracy, Repeatability and Resolution in 2 example
sensors, possible accuracy improvement shown for Sensor 1 ...................... 10
Figure 2.2 - Touch Probes [23] : (a) TPS1(S) Touch Trigger Probe; (b)
Touch Trigger Probe Principle of Operation; (c) SP600 Analog Touch
Probe ............................................................................................................... 12
Figure 2.3 – Touch Trigger Probe Kinematics - Low Force Direction [24] . 14
Figure 2.4 – Touch Trigger Probe Kinematics - High Force Direction [24] 14
Figure 2.5 - Pre-Travel Variation in Resistive Touch Trigger Probes ........ 15
Figure 2.6 - Triangulation using Two Cameras [26], (b) Structured Light
[27] .................................................................................................................. 17
Figure 2.7 - Specular and Diffuse Reflection in laser scanning systems .... 18
Figure 2.8 - Limitations of Laser Scanners .................................................. 19
Figure 2.9 - Laser Line with Camera Digitizer [30] ..................................... 20
Figure 2.10 - 3 Point Registration MCS→CAD ............................................ 22
Figure 2.11 - Orthogonal Least Squares Planar Circle Parameters ........... 23
Figure 3.1 - Proposed Factory Floor Architecture ........................................ 32
Figure 3.2 - Memex Ax9150 Universal Machine Interface [38] .................. 34
Figure 3.3 - Ring Gauge Probing Path. Arrows indicate 8 of 36
trajectories. ..................................................................................................... 37
ix
Figure 3.4 - 2 Hit Probing Method ................................................................ 38
Figure 3.5 - Graph of Pre-travel variation using 36 trajectories ................. 39
Figure 3.6 - 5 Point Probing of a Sphere ....................................................... 40
Figure 4.1 - Surface Finishing Operation Toolpath ..................................... 42
Figure 4.2 - Fixture Base with Tooling Spheres and Workpiece ................. 44
Figure 4.3 - Toolpath Case Divisions ............................................................ 45
Figure 4.4 - Machining Case 0 ....................................................................... 47
Figure 5.1 - Inspection Report of Experiment Case 0, Roland Scan Data .. 50
Figure 5.2 - Case 0, Roland Scan Data – Colour Map adjusted to show
detail. .............................................................................................................. 51
Figure 5.3 - Inspection Report of Experiment Cases 1-4, Roland Scan
Data................................................................................................................. 52
Figure 5.4 – Cases 1-4, Roland Scan Data – Colour Map adjusted to
show detail. ..................................................................................................... 53
Figure 5.5 - Photograph of Machined Experiment Workpiece..................... 58
Figure 6.1 - Gocator 2330 Specifications Diagram [44] ............................... 62
Figure 6.2 - CMM-Laser Scanner Hardware Architecture Diagram .......... 64
Figure 6.3 - 3D Scanner Mount with bolt hole pattern for locking pin ...... 65
Figure 6.4 – LMI 3D Gocator 2300 Scanner on an IOTA P CMM ............... 66
Figure 6.5 – Flow Chart of PC Application Scanning Function .................. 68
x
Figure 6.6 - Geometric Fit of Captured Point Cloud with 0.0684 mm
error in diameter. ........................................................................................... 69
Figure 6.7 – CMM Laser Scanner Application Graphical User Interface .. 70
Figure 6.8 - LMI Scan Data: Case 0 Surface Deviation Analysis ................ 73
Figure 6.9 - Case 0, LMI Scan Data – Colour Map adjusted to show
detail. .............................................................................................................. 74
Figure 6.10 - LMI Scan Data: Cases 1-4 Surface Deviation Analysis ......... 75
Figure 6.11 – Cases 1-4, LMI Scan Data – Colour Map adjusted to show
detail. .............................................................................................................. 76
Table 4.1 - OLS Sphere Fit ............................................................................ 47
Table 4.2 - Fitted Sphere Centre Registration Error and Registration
Components .................................................................................................... 48
Table 5.1 - Experiment Inspection Data (Roland) Statistics ....................... 54
Table 5.2 - Sphere Fit Statistics .................................................................... 55
Table 5.3 - Sphere Centre Distances ............................................................. 56
Table 5.4 - Plane Fitting Data ....................................................................... 57
Table 5.5 - Summary of Expected Errors ...................................................... 60
Table 6.1 - Gocator 2330 Laser Scanner Specifications ............................... 62
Table 6.2 - CMM Laser Cumulative Error.................................................... 72
Table 6.3 - LMI Scan Data Surface Fit Statistics ........................................ 77
xi
List of Abbreviations and Symbols

NC – Numerical Control
CNC – Computer Numerical Control
3D – Three Dimensional
CAD – Computer Aided Design
CAM – Computer Aided Manufacturing
CMM – Coordinate Measuring Machine
GD&T – Geometric Dimensioning and Tolerancing
HTM – Homogeneous Transformation Matrix
MCS – Machine Coordinate System
OLS – Orthogonal Least Squares
SVD – Singular Value Decomposition
ICP – Iterative Closest Point
CME – Checkmate Engine
DLL – Dynamic Link Library
PCL – Point Cloud Library
UMI – Universal Machine Interface
FOV – Field of View
xii
Chapter 1
Introduction
Numerical control (NC) machining is a method of automatically
operating a manufacturing machine based on instruction codes. NC
machining offers advantages over manual production such as better control of
cutting parameters, improved part quality and repeatability, reduced
manufacturing time and reduced scrap [1].
Computer numerical control (CNC) brings the local storage and
connectivity capabilities of a computer system to NC machining, replacing the
punched tape and dedicated hardware of early NC systems.
The essence of conventional computer-aided manufacturing automation
is repeatability. In CNC machining workflow, a part design created using
three dimensional (3D) Computer-Aided Design (CAD) software can be
interpreted by Computer-Aided Manufacturing (CAM) software to generate
cutting tool paths. The part is then clamped into place, the tool paths
executed, and the desired final shape is produced.
In practice, many factors add variability to the machining process, such
as:
• In Process Variability
1
The cutting tool may be worn or break unexpectedly. If this can be
detected before the part is damaged, a replacement tool can be used
to continue the operation [2].
• Geometric/Setup Variability
Examples include casting variations, stamping die / injection mould
heat treat distortion, and fixturing setup variability. For
refurbishment of turbine engine compressor blades, in service creep
and weld bead shape differs in each case. Measurement and nominal
tool path adjustment for each blade is therefore required [3].
Presently, the steps of measurement, analysis, toolpath adjustment,
and machining are accomplished with separate tools lacking direct
integration and automation.
Part measurement in manufacturing is commonly performed as, but
not limited to, a post-process inspection. Parts are measured on dedicated
gaging systems or Coordinate Measuring Machines (CMM) to verify
dimensions against nominal geometric dimensioning and tolerancing (GD&T)
specifications. This thesis focuses on processes using CMM systems to
perform part measurement. CNC and CMM design, calibration and error
compensation have long been studied, changing by a small amount in
comparison to the advancements in machine control computational ability,
software, and the measurement devices used on these machines [4]. Resistive
2
technology based touch trigger probes, while popular in use, have been
surpassed by piezoelectric and scanning touch trigger probes that offer
significant improvements. Non-contact measurement sensors have entered as
a fast and accurate alternative to touch based measurement for some
applications.
Inspection data analysis software packages are relied on for
verification of part dimensions. The output of the inspection, such as a
surface deviation report, is used for approval of parts, and also as feedback for
the manufacturing process. Unfortunately, the inspection process is usually
performed after machining, and thus after any machining errors have
occurred. Recent analysis software such as OriginSoftFit [5] offers a proactive
approach with the ability to inspect a part prior to machining and make
modifications to the part program based on inspection data. Such
modifications may include global rigid transformations, or even per-feature
toolpath adjustments such as removing refurbishment weld material [6].
After use of this software, typically employed with a dedicated CMM or laser
digitizer, the part and toolpaths must be transferred a machine tool to
produce the final part.
Common machine tool sensors include tool setters and touch trigger
probes. Tool setters are used to determine the tool length, and touch trigger
probes are used to locate the workpiece within the machine workspace.
3
Virtually all turning centres and machining centres come equipped with some
form of touch trigger probe option [7]. Although the CNC control computer is
quite capable of performing the mathematics required for interpreting
numerical control (NC) code, it cannot run least squares or other data fitting
techniques with the same efficiency as a modern computer used for processing
CMM inspection data.
The lack of integrated and automated methods for using these tools
creates opportunity for operator error in recording, entering and analysis of
data [4], discouraging adoption of the tools. Because of this, scrap production
may result from a failure to adjust nominal tool paths to actual part
geometry, even though, by using available GD&T software, a salvaging
opportunity exists. This is the primary motivation of the thesis.
The objective of this thesis is to describe and develop methods allowing
analysis and tool path adjustment software to be integrated with machine
tool based measurement, so that solutions to the problem of geometric
variability in machining can be provided. Additionally, the thesis aims to
integrate a laser scanner with exisiting CMM equipment to investigate the
feasibility of integrating a laser scanner on a CNC machining centre.
A literature review is presented in Chapter 2, and the hardware
architecture is discussed in Chapter 3. Chapter 4 presents an experiment
where additional CNC hardware, and geometric analysis of measured datums
4
were used to update nominal tool paths to match the measured workpiece
location, and an in-tolerance part was produced. Experimental results are
discussed in Chapter 5. Integration of a laser scan head with a CMM is
presented in Chapter 6, and conclusions and future work are discussed in
Chapter 7.
5
Chapter 2
Literature Review
2.1 Measurement in the Manufacturing Process
There are many tools available for measurement in manufacturing
processes, the choice of which to employ depends on many factors including
production quantity, tolerance level, part complexity, and part variety. When
dealing with high production quantities and low part variety, dedicated
automatic gauges can be faster and more cost effective than CMM machines.
When flexibility is desirable, such as with low to medium volumes and a
variety of part geometries, or initial gauge qualification, the investment in a
CMM can be justified [8]. This thesis focuses on inspecting and machining
parts in volumes as low as a single part.
A scheme has been developed for classifying measurements based on
their existence in the manufacturing timeline [9]. The four categories are
preprocess measurement, in-process measurement, process-intermittent
measurement and postprocess measurement. The remainder of this section
discusses the involvement of automated inspection methods in each of the
four categories.
Preprocess measurements are made on the manufacturing system
itself. Calibration and correction of machine tools and measuring machines
6
fall in this category [4]. The ISO 10360 standard [10] describes “Acceptance
and reverification tests for coordinate measuring machines”, likewise the ISO
230-2 standard [11] evaluates the accuracy and repeatability of a machine
tool, and the ISO 230-10 standard [12] evaluates the measuring performance
of probing systems of numerically controlled machine tools. These standards
test the machine for environmental sensitivity, displacement accuracy,
spindle errors, thermal errors, kinematics, and cutting performance where
applicable [4]. When properly characterized, controlled and corrected,
machine tools can approach the accuracy of a measuring machine [13], [14],
[15] as cited in [4].
In-process measurements are measurements taken while the feature is
being produced. These measurements may be directly measured, or indirectly
measured through “deterministic metrology”, where process variables other
than the attribute of interest (dimension) are monitored in order to give
information about the attribute of interest. Direct measurement faces many
impending hazards on the factory floor such as chips, coolant and the tool
itself [16]. The most successful indirect in-process measurements pre-
calibrate the machine and monitor on-line variables, correcting the process
according to machining error-process variable relationship models [17].
Laboratory implementations have demonstrated successful monitoring of
7
machining errors by monitoring machine temperature, cutting forces, and
vibrations, however this technique has yet to reach the factory floor [18].
Process-Intermittent measurements are performed during a
momentary break from machining where measurements are taken and used
to modify process parameters. This may be performed on or off the machine
using a touch-trigger probe, or by other gauging equipment such as hand-held
device. Higher end CMM systems employ analog scanning probes capable of
tracing surface contours, but point measurement based touch trigger probes
are much more common among machining centres. The use of a touch probe
on a machining centre has many sources for error that complicate measuring.
The steady state temperature conditions that are imposed on CMM machines
cannot be imposed on CNC machining centres. As a result, if the temperature
of critical structural components is changes between machining and probing,
the spindle position from machining to probing can be very large [19].
Additionally, there is the systematic aspect of measuring a part on the same
equipment that machined it. If a systematic error exists with the machine at
the time of machining, the same error will exist during inspection [20]. If
these sources of error are understood and diligently considered, and machine
tool measurement evaluation standards are followed, is has been shown that
on-machine probing has a valid place in the manufacturing process.
8
Postprocess measurements are the measurements performed after
machining. Depending on the process these measurements could result in
pass/fail decisions, further machining, and/or correction of process
parameters.
2.1.1 Static Sensor Specifications
Manufacturing measurement sensors and systems are typically
described based on their static performance specifications since part
dimensions are expected to be constant over the duration of measurement.
Accuracy, repeatability and resolution are common specifications used to
describe sensor capabilities. Repeatability, also referred to as precision, is
the:
‘closeness of agreement between indications or measured quantity values
obtained by replicate measurements on the same or similar objects under
specified conditions’ [21].
The standard deviation range that the repeatability value represents
must also be specified. A 1 σ repeatability is achieved 68.27% of the time,
while 2 σ and 3 σ are achieved 95.45% and 99.73% of the time. By the same
source, Accuracy is defined as the:
‘closeness of agreement between a measured quantity value and a true
quantity value of a measurand’.
9
A high degree of accuracy in multiple sensor measurements cannot be
achieved without a high degree of repeatability. Calibration can improve
accuracy in sensors with a high degree of repeatability, but a sensor with a
low degree of repeatability will have inherently lower accuracy even after
calibration. Resolution is defined as the:
‘smallest change in a quantity being measured that causes a perceptible
change in the corresponding indication’.
Resolution is important, but a higher resolution cannot compensate for
deficiencies in accuracy. The concepts of accuracy, repeatability and
resolution are illustrated in Figure 2.1.
Figure 2.1 - Accuracy, Repeatability and Resolution in 2 example
sensors, possible accuracy improvement shown for Sensor 1
10
2.1.2 Touch Trigger Probes
The touch trigger probe (Figure 2.2(a)) was invented to measure fuel
lines on the Concorde supersonic aircraft, where existing hard contact
methods caused the fuel lines to deflect during measurement [22]. The touch
trigger probe relies on three equally spaced contact points, serving as series
wired normally closed switches (Figure 2.2(b)). The probe is positioned in the
X, Y, and Z directions using CMM or CNC machine tool motorized axes, and
upon making contact with the part being measured, one of three switch
contacts will open. The open switch signals the control computer to stop
motion, and save the current axes positions in memory. The positions can
then be used for mathematical calculations. Before another point can be
measured, the probe must be moved away from the point of contact to allow
the probe switch to close, and to avoid damage to the probe due to excess
deflection. Dynamic effects limit the overall data collection rate to
approximately one point per second. Newer analog touch probes (Figure
2.2(c)), available for CMMs, measure the amount of stylus deflection and can
obtain continuous measurements by maintaining contact with the surface.
11
(a) (b) (c)
Figure 2.2 - Touch Probes [23] : (a) TPS1(S) Touch Trigger Probe; (b)
Touch Trigger Probe Principle of Operation; (c) SP600 Analog Touch
Probe
Touch trigger probes are also available for CNC machine tools. The
probe can be loaded into the spindle using an automatic tool changer, and
replaced with a cutting tool for machining. A wireless communication link
(infrared or radio frequency) is used to transmit the probe switch open signal.
As with CMM touch trigger probes, the data collection rate is slow. The
machine tool control has limited mathematical calculation ability, and lack of
direct integration prohibits use of analog probes with standard CNC controls.
12
2.1.2.1 Kinematics of Touch Trigger Probes
The proper use of a touch probe requires the understanding of pre-
travel and probe lobing concepts. Pre-travel refers to the amount the stylus
deflects as it generates the torque required to unseat the probe. The stylus
deflection d is dependent on applied force F according to the bending stiffness
of a cantilever beam modelled with cross-sectional moment of inertia I, elastic
modulus E, and length P.
P3 F
d= (2.1)
3EI
Since the stylus length P dominates the deflection, one should use the
shortest stylus possible while maintaining the ability to visit all the required
probe locations, such as those deep inside a hole.
The concept of probe lobing refers to the direction dependent pre-travel
variation shown in Figure 2.3 and Figure 2.4. To unseat the probe, the stylus
torque 𝑻𝒔 created by the force 𝑭𝒔 at the stylus tip must overcome the torque
𝑻𝒌 created by the spring force 𝑭𝒌 and distance 𝑹. The torque will be lowest
when the pivot point is the line drawn between two of the stylus supports, as
the perpendicular distance to the stylus axis is minimal. This effect will
contribute an error as if the tip were not spherical, as shown by the dotted
line in Figure 2.5. Since the effects of pre-travel are repeatable, calibration
can be applied to increase the accuracy of touch trigger probes.
13
Figure 2.3 – Touch Trigger Probe Kinematics - Low Force Direction
[24]
Figure 2.4 – Touch Trigger Probe Kinematics - High Force Direction
[24]
14
Figure 2.5 - Pre-Travel Variation in Resistive Touch Trigger Probes
2.1.2.2 Relevant CNC G-Codes for Touch Trigger Probes
Operation of a touch trigger probe on a CNC machine tool requires a
sequence of programmed motion commands. The Fanuc Series 15i-MA
control used in this thesis used the RS274D G-code standard. The essential
G-code commands and syntax are given in sections 2.1.2.3 and 2.1.2.4 .
2.1.2.3 G01 Linear Motion Command
The G01Xxx.xxxYyy.yyyZzz.zzzFff.fff linear interpolation command is
used to move along a straight line from the current position to
(xx.xxx,yy.yyy,zz.zzz) at feed rate ff.fff. If the G21 code specifying
programming in mm is active, positions are specified in mm, and feedrates in
15
mm/min. The G01 code is used to position the probe before measurement,
and to back away from the part surface after measurement.
2.1.2.4 G31 Skip Command
The touch trigger probe is electrically connected to the CNC motion
control so that a special skip signal is generated when the probe switch opens
upon contacting the workpiece. To interface with the skip signal, the
G31Xxx.xxxYyy.yyyZzz.zzzFff.fff command is used. This command is similar
to G01, in that linear motion begins at the current position and proceeds to
(xx.xxx,yy.yyy,zz.zzz) at feed rate ff.fff. If the skip signal is received, motion
halts and the CNC control records the machine coordinate system (MCS)
position to local variables. The Fanuc 15i-MA control used in this work
assigns the X,Y and Z MCS positions to local variables #5061, #5062 and
#5063 [25]. Once MCS positions are stored, the next line in the NC code will
be executed. Since the skip signal remains high while the probe is in contact,
must be backed away from the surface using a command other than G31.
2.1.3 Laser Digitizers
When data collection rates of thousands of points per second are
needed, non-contact laser digitizers have emerged as the preferred solution.
Triangulation is a technique used by 3D scanning systems (Figure 2.6 (a)) to
16
determine a 3D point location based on its location in two or more 2D images
with known camera locations and orientations [26].
(a) (b)
Figure 2.6 - Triangulation using Two Cameras [26], (b) Structured
Light [27]
The identification of pixels 𝒙! and 𝒙! in Figure 2.6(a) that correspond to
the same real-world point 𝒑 is known as the correspondence problem, and is a
very difficult computer vision problem [28]. One strategy to assist in pixel
identification is the use of structured light techniques (Figure 2.6(b)). Using a
projector or laser diode, a line can be projected from viewpoint 𝑹! . The point
𝒑, identified in 𝑹! as pixel 𝒙! , must exist at the intersection of 𝒗! and the
plane formed by the projected line.
Figure 2.9 shows a laser scanner employing these techniques, and the
laser line as seen by the camera. Synchronization and calibration with the
machine axes is required to obtain the locations in a fixed reference frame.
17
Figure 2.7 - Specular and Diffuse Reflection in laser scanning
systems
Figure 2.3 illustrates the issue of specular reflection in laser scanning
systems, where no light is reflected from the surface toward the camera [29].
18
Figure 2.8 - Limitations of Laser Scanners
Figure 2.8 illustrates two limitations of laser scanners. On the left, the
shadow caused by another part feature (corner 1) obstructs the path of the
laser. To overcome this, multiple view angles are required, which can be
obtained by reorienting the scanner and completing multiple scanning passes.
Also illustrated in Figure 2.8 is the false data scenario, where
additional reflections are visible to the camera. Since some light is absorbed
at each reflection, corrective actions include reducing the incoming beam
strength to only allow once reflected beams to trigger the sensor.
Additionally, if the light sensor stores the intensity value of the collected
19
point, intensity values that are significantly lower compared to points taken
in the same area could be identified as false data.
Other limitations include high data rate transmission requirements,
and the need for sophisticated mathematical processing. Nonetheless, this
type of sensor is emerging as an important solution.
Figure 2.9 - Laser Line with Camera Digitizer [30]
20
2.2 Data Fitting
2.2.1 Rigid Coordinate System Registration
With any inspection sensor / machine, there is a need to align various
coordinate systems. Data collected on a machine will be expressed using
different coordinates than were used in the CAD representation. Conversion
can be accomplished using Homogeneous Transformation Matrices (HTMs).
A simple procedure is to measure N = 3 points (for example sphere centres)
on the machine, and then map them to the corresponding CAD points.
Denoting the machine coordinate system with subscript MCS, and the CAD
coordinates with subscript CAD, the points are expressed as
Pi ,MCS ( xi ,MCS , yi ,MCS , zi ,MCS ) , and P1,CAD ( x1,CAD , y1,CAD , z1,CAD ) = ( 0,0,0) ,
P2,CAD ( x2,CAD , y2,CAD , z2,CAD ) = (1,0,0) , P3,CAD ( x3,CAD , y3,CAD , z3,CAD ) = ( 0,1,0)
.
21
P3,MCS
P2,MCS
P1,MCS
P3,CAD
x
P1,CAD P2,CAD
Figure 2.10 - 3 Point Registration MCS→CAD
Note that the misalignments have been greatly exaggerated for clarity.
It is assumed that a rigid transformation can be used with unit lengths
and right angles. That is P2,MCS − P1,MCS = P2,CAD − P1,CAD = 1,
P3,MCS − P1,MCS = P3,CAD − P1,CAD = 1 , and ∠P3,MCSP1,MCSP2,MCS = ∠P3,CAD P1,CAD P2,CAD = 90 .

o
With these assumptions, the HTM to convert from CAD to MCS can be
written as
𝑋 = 𝑃!,!"# − 𝑃!,!"# (2.2)
𝒀 = (𝑷𝟑,𝑪𝑨𝑫 − 𝑷𝟐,𝑪𝑨𝑫 ) (2.3)
𝒁 = 𝑷𝟐,𝑪𝑨𝑫 − 𝑷𝟏,𝑪𝑨𝑫 × (𝑷𝟑,𝑪𝑨𝑫 − 𝑷𝟐,𝑪𝑨𝑫 ) (2.4)
𝑿 𝒀 𝒁 𝑷
𝑯𝑻𝑴𝑴𝑪𝑺←𝑪𝑨𝑫 = 𝟏,𝑪𝑨𝑫 (2.5)
𝟎 𝟎 𝟎 𝟏
The MCS to CAD HTM is the inverse, 𝑯𝑻𝑴!"#←!"# = 𝑯𝑻𝑴!𝟏

!"#←!"#
22
2.2.2 Orthogonal Least Squares Geometric Fitting
After collecting the data points, such as from the surface of a sphere, a
sphere must be mathematically fit to the data. Orthogonal Least Squares
(OLS) methods were used. The fit parameters for an OLS circle fit are shown
in Figure 2.11.
+
𝑺𝟏(𝑺𝟏,𝒙 ,𝑺𝟏,𝒚)
𝒅!
𝒓
𝑺(𝒙,𝒚) 𝒅! +
𝑺𝟐(𝑺𝟐,𝒙,𝑺𝟐,𝒚 )
𝒅!
+
𝑺𝟑(𝑺𝟑,𝒙,𝑺𝟑,𝒚 )
Figure 2.11 - Orthogonal Least Squares Planar Circle Parameters
For a sphere with sample points S j , centre point S , and radius r , the
orthogonal distance error is
23
2 2 2
dj = (s j,x ) ( ) (
− sx + s j, y − s y + s j,z − sz ) −r (2.6)
M
and the OLS sum to minimize is Σ2 = ∑ d 2j where M is the number of sphere
j=1
sample points. Assuming that the sphere is very nearly perfectly round
(small form errors), a clever initial approximation is obtained using the linear
approximation [31].
θj = ⎡( s − s )2 + ( s − s )2 + ( s − s )2 ⎤ − r 2
⎣⎢ j , x x j, y y j,z z
⎦⎥
⎧ 2 2 2 1/2 ⎫ ⎧ 2 2 2 1/2 ⎫
= ⎨ ⎡⎢( s j , x − sx ) + ( s j , y − s y ) + ( s j , z − sz ) ⎤⎥ − r ⎬ ⋅ ⎨ ⎡⎢( s j , x − sx ) + ( s j , y − s y ) + ( s j , z − sz ) ⎤⎥ + r ⎬
⎩ ⎣ ⎦ ⎭ ⎩ ⎣ ⎦ ⎭
≈ d j ⋅ 2r
and then minimizing the sum
M M 2
{
Θ2 = ∑θ j2 = ∑ g j − h − 2 ⋅ ⎡⎣ s j , x sx + s j , x sx + s j , x sx ⎤⎦
j =1 j =1
}
where g j = s 2j , x + s 2j , y + s 2j , z and h = r 2 − ( sx2 + sy2 + sz2 ) . Simultaneously solving the
partial derivative equations
∂Θ2 ∂Θ2 ∂Θ2 ∂Θ2

=0 =0 =0 =0
∂sx ∂s y ∂sz ∂h
leads to the linear system
24
⎡ M M M M
⎤ ⎡ M ⎤
⎢ ∑1 ∑s j,x ∑s j, y ∑s j,z ⎥ ⎢ ∑ g j ⎥
⎢ j =1 j =1 j =1 j =1
⎥ ⎢ j =1 ⎥
⎢ M M
2
M M ⎥ ⎡ h ⎤ ⎢ M ⎥
⎢ ∑ s j , x ∑s j,x ∑s s
j,x j, y ∑ s j , x s j , z ⎥ ⎢ ⎥ ⎢ ∑ s j , x g j ⎥
⎢ j =1 j =1 j =1 j =1 ⎥ ⋅ ⎢ 2sx ⎥ = ⎢ j =1 ⎥
⎢ M M M
2
M ⎥ ⎢ 2s y ⎥ ⎢ M ⎥
⎢ ∑ s j , y ∑ s j,x s j, y ∑s j, y ∑ s s
j, y j,z ⎥ ⎢ ⎥ ∑ j , y j ⎥
⎢ s g
⎢ j =1 j =1 j =1 j =1 ⎥ ⎣ 2sz ⎦ ⎢ j =1 ⎥
⎢ M M M M ⎥ ⎢ M ⎥
2
⎢ ∑ s j , z ∑s s
j,x j,z ∑s s
j, y j,z ∑ s j,z
⎥ ⎢ s g
∑ j , z j ⎥⎥
⎣⎢ j =1 j =1 j =1 j =1 ⎦⎥ ⎣⎢ j =1 ⎦
that can be solved for h , sx , s y , and sz . Newton-Raphson or equivalent
iterative improvement is then used to minimize Σ2 in the exact problem by
solving for δ sx , δ s y , and δ sz in the matrix equation
⎡ ∂Σ 2 ⎤ ⎡ ∂ 2 Σ 2 ∂ 2Σ2 ∂ 2 Σ 2 ⎤
⎢ ⎥ ⎢ 2 ⎥
⎢ ∂sx ⎥ ⎢ ∂sx ∂sx s y ∂sx sz ⎥
⎢ ∂Σ 2 ⎥ ⎢ ∂ 2 Σ 2 ⎡δ sx ⎤
∂ 2Σ2 ∂ 2 Σ 2 ⎥⎥ ⎢ ⎥
⎢ ⎥ = ⎢ ⋅ δ sy
⎢ ∂s y ⎥ ⎢ ∂sx s y ∂s y2 ∂s y sz ⎥ ⎢ ⎥
⎢ 2 ⎥ ⎢ 2 2 ⎥ ⎢⎣δ sz ⎥⎦
⎢ ∂Σ ⎥ ⎢ ∂ Σ ∂ 2Σ2 ∂ 2 Σ 2 ⎥
⎢⎣ ∂sz ⎥⎦ ⎢ ∂sx sz ∂s y sz ∂sz2 ⎥⎦
⎣
and then updating the sphere centre estimate using sx = sx − δ sx , s y = s y − δ s y ,
and sz = sz − δ sz . The improved radius estimate is
M
1 2 2 2
r= ∑ (s j ,x
− sx ) + (s j,y
− sy ) + (s j ,z
− sz ) . Iterations continue until the
M j=1
estimates for sx , s y , sz , and r converge. Note that convergence is to a
relative extreme, and this might be a maximum rather than a minimum.
25
Using the linearized initial approximation, with small form errors, reduces
this likelihood. An example using experimental data is shown in Table 4.1.
2.2.3 Orthogonal Least Squares Fitting on Data Sets
If there are position repeatability errors ηi in the actual sphere
locations, or there are more than four points, Orthogonal Least Squares
(OLS) methods that distribute the error over the set are used [32]. For
example, separating into rotation R MCS←PCS and translation TMCS←PCS
components, the transformation with noise is
Pi ,MCS = R MCS←PCS Pi ,CAD + TMCS←PCS + ηi (2.7)
ˆ
The OLS goal is to find R ˆ
and T that minimize the sum
MCS←PCS MCS←PCS
N 2
Σ 2 = ∑ Pi ,MCS − ( R MCS←PCS Pi ,CAD + TMCS←PCS ) (2.8)
i =1
The solution is found using a Singular Value Decomposition (SVD)
N N
1 1
approach as follows. First calculate PCAD = ∑ Pi,CAD and PMCS = ∑P i ,MCS
.
N i =1 N i =1
Next define Qi ,CAD = Pi ,CAD − PCAD and Qi ,MCS = Pi ,MCS − PMCS . Calculate
N
A = ∑ Qi ,CADQiT,MCS , the SVD A = UΛVT , and X = VUT . Usually the
i=1
ˆ = X , and Tˆ = P − R
determinant det ( X ) = +1, and in this case R ˆ .
MCS MCS←CAD PCAD
26
If the det ( X ) = −1, and an eigenvalue of Λ is zero, the reflection solution has
been found. It can be corrected to a rotation by replacing V with
⎡ v1,1 v1,2 −v1,3 ⎤

⎢ ⎥
Vʹ′ = ⎢v2,1 v2,2 −v2,3 ⎥ and X with Xʹ′ = Vʹ′UT .
⎢⎣ v3,1 v3,2 −v3,3 ⎥⎦
Implementation on actual experimental data is presented in Section
4.3 . Other cases are considered in [32]. Commercial implementations for
larger point sets usually employ quaternion methods [33].
27
2.2.4 Commercial Point Cloud Software
A number of point cloud processing software packages are
commercially marketed, including GeoMagic [34], PolyWorks [35], etc. Basic
software functions include point cloud viewing, point deletion, triangular
facet surface creation, and CAD file import. To align the MCS point cloud to
CAD coordinates, registration algorithms are required. The most basic
registration methods are derived from the Iterative Closest Point (ICP) work
of McKay and Besl [33]. For manufacturing operations, the use of Geometric
GD&T datums and alignments is industry standard. This requires feature
fitting of lines, planes, etc. and the ability to lock any combination of the six
(three translational and three rotational) rigid body degrees of freedom. The
above cited software has ability to accomplish this, but, being intended for
interactive use, the full automation required to eliminate operator invention
is lacking. For automation of the work reported herein, use of the Origin
International CheckMate Engine (CME) is intended [36]. This software
provides a Microsoft Windows Dynamic Link Library (DLL) set of callable
functions that implement the feature fitting and registration functions for
GD&T analysis.
Point Cloud Library (PCL) is a large scale, collaborative, open source
software project for 2D and 3D point cloud processing. The PCL framework
contains numerous state-of-the art algorithms including filtering, feature
28
estimation, surface reconstruction, registration, model fitting and
segmentation [37]. Creation of PCL was driven by the availability of low cost,
high data rate sensors such as Microsoft Kinect, which have significant
potential for robotics navigation and object recognition. While it does not
currently have functions relating to GD&T analysis of point clouds, given the
large number of collaborators, ease of adding changes, and free for
commercial and research use license, PCL is likely to experience many
feature additions which could include GD&T analysis capability.
2.3 Summary
Touch trigger probes have evolved to now include higher data rate
analog probes, and implementation on CNC machine tools. Very high data
rate laser digitizers are now available for CMMs, but are not integrated with
commercial CNC machine tool motion controllers. Commercial point cloud
software deals only with standard cases, and does not interface with CNC
machine based inspection methods, which is required for an in-process
manufacturing implementation in a variable geometry environment.
Progress towards overcoming these shortcomings is contributed in the
remaining chapters of the thesis.
29
Chapter 3
Hardware Description
3.1 CNC Machine Tool
The work in this thesis is applicable to a range of CNC machine tools,
including CNC turning centres, 3 axis milling centres, 5 axis milling centres
as well as machining centres featuring both turning and milling.
The work in this thesis used a Matsuura FX-5, 3 axis, vertical
machining centre controlled by a FANUC Inc. Series 15i-MA control
computer.
The FX-5 was equipped with a Renishaw MP12 touch trigger probe
with a repeatability of 1.0 µμm (2𝜎) when used with the 50mm stylus.
3.2 Touch Trigger Probe Integration
Typically touch trigger probes are used in conjunction with probing
cycle programs provided by the probe manufacturer. While these cycles are
configurable and can be combined to create sophisticated probing routines,
the proposed alternative of processing data external to the CNC control
allows use of data fitting techniques such as the OLS methods described in
section 2.2 The NC DPRNT command is used to print the probe hit
coordinates to a Universal Machine Interface (UMI) board over the serial
30
connection. The UMI will then pass the data over Ethernet to a computer for
processing. To suit industries where additional computer involvement is
undesirable for security reasons, the UMI can be programmed to process the
data directly, using the same least-squares fitting techniques.
This chapter describes the hardware and necessary implementation
details that enable this tool path alignment workflow.
31
3.3 Hardware Overview
Figure 3.1 - Proposed Factory Floor Architecture
The major hardware elements are:
1. Touch Trigger Probe
2. Machine Control Computer
3. Universal Machine Interface (UMI)
4. Data processing server
32
A probing tool path is executed by the machine control, to obtain a
measurement using the touch trigger probe. This measurement is passed to
the processing computer via the UMI. Once all data points have been
received, the an updated tool path is computed and sent back to the machine
via the UMI. This proposed bidirectional communication is a critical step
towards bringing the capabilities of GD&T software to the CNC machine tool.
The following sections describe each component in greater detail.
3.4 Ax9150 UMI
The Ax9150 Universal Machine Interface developed by Memex
Automation Inc. is a multi-function device originally designed to report
machine status to improve overall equipment efficiency of machine tools on a
factory floor. It features 0-30VDC volt digital inputs and 30VDC digital
outputs for connecting to factory equipment signals, as well as RS-232 Serial
and Ethernet connectivity for data reporting and transfer of part programs.
Compatibility with the Fanuc 15i machine tool control, availability of
manufacturer supplied software, access to local technical support, and lack of
other comparable hardware were all factors in the decision to use the Ax9150
UMI.
33
Figure 3.2 - Memex Ax9150 Universal Machine Interface [38]
3.4.1 Installation
The Ax9150 UMI was installed inside the Matsuura FX-5 Machine
Tool cabinet and connected to Serial port 1 of the Fanuc Series 15i-MA
control. To support the UMI Ethernet connection, a local area network router
was also installed in the cabinet. The machine control system parameters
#20 and #21 were changed from “1” to “14” on the CNC Controller to set the
Serial port active for UMI communications.
3.4.2 Passing coordinate data to PC
In order to transfer the G31 position information from the CNC
Controller, the ‘Macro B’ feature set of the Fanuc control was utilized. The
34
Macro B feature set enables a set of codes outside the RS274D G-code
standard to be used in order to enable advanced programming features such
as data storage, logical operations, looping and more [39]. The ‘DPRNT’
command was used to send plain text output to a RS-232C device, in this case
the Ax9150 UMI, which then relays the data over Ethernet to the processing
computer.
The following code is used to measure a point, and then print the probe
coordinates stored in local variables #5061, #5062 and #5063:
G31X-137.Y-137.Z72.F25.
POPEN
DPRNT[POINT: #5061 [44] #5062 [44] #5063 [44]]
PCLOS
Resulting serial terminal output:
POINT: -135.5105 -136.3415 72.0620
3.5 Calibration of Touch Trigger Probe
Before taking a measurement the touch trigger probe must be
calibrated as mounted on the CNC machine tool. This was accomplished
using a ring gauge of known diameter 71.133 mm (2.8005 inches), measured
using the probe in the CNC. From the resulting data, the directionally
35
dependent pre-travel amount is revealed. An average pre-travel of 0.062 mm
was calculated and used for compensation in this work. The probing method
is described in 3.5.1 and the results are discussed in 3.5.2 .
3.5.1 Calibration Tool Path
The calibration procedure involves taking probe measurements in
different directions to assess the probe’s directionally dependent pre-travel
amount. In this procedure 36 probe measurements were made, travelling
radially from the centre of the ring to the inside surface of the ring as shown
in Figure 3.3.
36
Figure 3.3 - Ring Gauge Probing Path. Arrows indicate 8 of 36
trajectories.
To reduce the error associated with the probing speed and the
unpredictable 0-4ms delay between the rising edge of the skip signal and the
reaction of the machine [40], a double hit method illustrated in Figure 3.4
was used. Upon making initial contact with the ring gauge surface (1), the
probe backs away from the contact point a small amount (2). The probe
37
makes a second measurement of the location with reduced feedrate to limit
delay effects (3). The probe then moves to another inspection site (4).
Figure 3.4 - 2 Hit Probing Method
3.5.2 Results of Calibration Data
From the measured ring gauge radius data, the average pre-travel was
calculated to be 0.062 mm. The correction was applied independent of
direction, however, all probe hits in later experiments were in line with the
major axes (0,90,180,270) where pre-travel error is very close to 0.062 mm.
Pretravel in the axial direction of the probe stylus (Z- direction) is governed
by the deflection due to compressive forces, the which is much less than the
deflection due to bending. Because of this, no pre-travel compensation was
applied to axial direction probe hits.
38
Figure 3.5 shows the pre-travel amounts plotted on a circular grid.
The triangular shape is characteristic of the probe’s kinematic configuration –
three mechanical supports with a retaining spring.
Deviation from Expected Radius (mm)

10
3600.1 20
350 30
340 0.09 40
330 0.08 50
320 0.07 60
0.06
310 0.05 70
0.04
300 80
0.03
290 0.02 90 Trial 3
0.01
280 0 100 Trial 2
270 110 Trial 1
260 120
250 130
240 140
230 150
220 160
210 170
200 180
190
Figure 3.5 - Graph of Pre-travel variation using 36 trajectories
3.6 Three Sphere Probing Routine
A probing routine was created to obtain sphere centre data for the
fixture plate illustrated in Figure 4.2. The operator positions the probe above
the first sphere so that contact will be made with the top hemisphere as the
39
probe moves downwards. After contact is made in the downwards direction, a
coarse approximation of the X-Y centre of the sphere is found by probing at 4
equally spaced locations on the circular cross section located one radius
distance down from the point of first contact, as shown in Figure 3.6.
Figure 3.6 - 5 Point Probing of a Sphere
Once the centre is found, the process is repeated using this centre as a
starting point. This method ensures all the final probe hits are made normal
to the sphere surface to minimize pre-travel errors.
Because the sphere locations on the plate are known, the probing
routine can find the top hemispheres of spheres 2 and 3 by moving in X or Y
by the sphere spacing if rotation misalignment is limited. For misalignments

40
expected to be greater than 5 degrees, a robust seeking strategy must be
applied or the datum features must be probed separately. Development of
such a strategy would be specific to the workpiece and datum feature
locations, and is not in the scope of this thesis.
41
Chapter 4
Registration Experiment
4.1 Experiment Tool Path
A partial arc of a cylindrical surface is used as the test workpiece. The
machining operation is representative of a finishing operation, where the
cylindrical arc surface has been roughly cut by a band saw or other non-CNC
means, and CNC machining is used to create the desired surface finish.
Figure 4.1 shows the toolpath trajectories in CAM software [41].
Figure 4.1 - Surface Finishing Operation Toolpath
42
4.2 Tool Path Adjustment
4.2.1 Part Registration in Manufacturing Processes
In high volume production, the reliability of dedicated fixture and
clamping assemblies justifies the initial fixture design investment. These
fixtures are created with a single workpiece in mind, and are not suited for
significant workpiece variances or custom machining operations.
Herein a method is proposed whereby the clamped workpiece is located
using automated sensor measurement methods, and tool path trajectories are
updated at run-time to match the actual workpiece pose. In addition to
reducing set-up times, scrap avoidance can be achieved by using early part
inspection to detect machinability issues, such as geometric distortion in cast
or heat treated parts. In these cases, analyzing the available workpiece
material distribution with GD&T software gives the ability to adjust the tool
path to best fit the available material, so that all features can be machined to
meet the GD&T specifications for the part.
4.2.2 Tooling Sphere Registration Method
As a simple testing platform, a base or platform was constructed as
illustrated in Figure 4.2.
43
Figure 4.2 - Fixture Base with Tooling Spheres and Workpiece
The base consists of three tooling spheres fixed to a plane, on which
another workpiece can be mounted. Ceramic tooling spheres were chosen to
provide simple geometry for touch trigger probe inspection. The ceramic
tooling spheres are also suitable for laser scanning technology due to the
diffuse reflection offered by the ceramic surface.
Once the base assembly is mounted in the CNC workspace, a NC
probing routine will obtain point data to solve for the centre of each of the
tooling spheres. Once the point locations are known in the MCS, the HTM to
44
convert from CAD to MCS can be calculated as illustrated in section 2.2.3 ,
and the correction transformation can be applied to the nominal tool path.
Samples of original and adjusted toolpath programs can be found in
Appendices E and F. A laser scan of the finished workpiece will be used to
provide numerical data to evaluate the success of the proposed workflow.
4.3 Test Cases
Figure 4.3 - Toolpath Case Divisions
45
To test the consistency of the registration, the conventional machining
toolpath shown in Figure 4.1 is split into 5 sections shown in Figure 4.3. Each
section will be machined after moving and re-probing the fixture, fitting probe
data, and adjusting the tool paths.
In Case 0, the fixture is clamped in the nominal position with the
workpiece axes aligned with the machine axes. In this position the spheres
are probed and stored for reference in later registration operations.
In Case 1 the plate is rotated by 1 degree positively about the Z axis. In
Case 2 the plate is rotated by -1 degree about Z, and +16 mm in X. In Case 3,
the plate is rotated by 1 degree about Z, and moved -48 mm in X. And finally
in Case 4, the plate is rotated -2 degrees about Z and moved -35 mm in X. All
specified translations and rotations are in reference to the nominal position.
The rotation in Case 4 is the maximum rotation that the plate’s oversized bolt
holes would permit. Machining of Case 0 on the Matsuura F-X5 3 axis CNC
machine is pictured in Figure 4.4.
46
Figure 4.4 - Machining Case 0
Table 4.1 - OLS Sphere Fit
Case 0 – Sphere 3
Point X (mm) Y (mm) Z (mm)
1 -136.7110 136.4120 148.8330
2 -124.0060 136.4140 136.1330
3 -149.4150 136.4140 136.1330
4 -136.7120 149.1310 136.1330
5 -136.7120 123.6930 136.1330
Fitted X (mm) Y (mm) Z (mm)

Sphere -136.7105 136.4120 136.1212
Centre
Radius (mm) 12.7118
Standard Deviation of Residuals 0.0145
(mm)
Uncertainty of Centre (mm) 0.0073
47
Table 4.2 - Fitted Sphere Centre Registration Error and Registration
Components
Fitted Sphere Centre

Registration Error Registration Components
Case 0 X Y Z R T
Sphere 1 0.00E+00 1.71E-13 0 1.0000 0.0000 0.0000 0.00E+00
Sphere 2 5.68E-14 8.53E-14 0 0.0000 1.0000 0.0000 -1.42E-14
Sphere 3 -5.68E-14 -8.53E-14 0 0.0000 0.0000 1.0000 1.42E-14
Case 1
Sphere 1 -0.0022 0.0148 0.0002 0.9999 -0.0173 -0.0002 -0.8700
Sphere 2 0.0070 -0.0049 -0.0006 0.0173 0.9999 -0.0000 -0.5866
Sphere 3 -0.0049 -0.0098 0.0005 0.0002 0.0000 1.0000 0.0060
Case 2
Sphere 1 0.0011 0.0053 -0.0001 0.9999 0.0157 -0.0001 16.6543
Sphere 2 0.0085 -0.0098 -0.0008 -0.0157 0.9999 0.0000 -0.6807
Sphere 3 -0.0096 0.0045 0.0009 0.0001 0.0000 1.0000 0.0117
Case 3
Sphere 1 -0.0114 0.0029 0.0011 1.0000 -0.0062 -0.0002 -48.4372
Sphere 2 0.0182 -0.0064 -0.0017 0.0062 1.0000 0.0000 -1.9574
Sphere 3 -0.0067 0.0036 0.0006 0.0002 0.0000 1.0000 0.0219
Case 4
Sphere 1 -0.0113 0.0046 0.0011 0.9993 0.0377 0.0000 -35.1846
Sphere 2 0.0087 0.0026 -0.0008 -0.0377 0.9993 0.0002 -0.2768
Sphere 3 0.0026 -0.0072 -0.0003 -0.0000 -0.0002 1.0000 -0.0090
Table 4.2 shows the 3 point registration sphere centre errors, for Cases
0 through 4 using the OLS fit method described in [32].
48
Chapter 5
Registration Experiment Results
5.1 Laser Scan of Finished Workpiece
After machining, the part was removed from the fixture plate and
scanned using a Roland PICZA LPX-600 laser scanner with accuracy ±0.050
mm given by the manufacturer [42]. The 3 minute scan captured 160,000
points, spaced at the scanner minimum of 1 mm.
The scan data of Case 0 was fit with the nominal cylinder arc geometry
and a deviation report was generated, comparing scan data from Cases 1
through 4 with the nominal cylinder arc geometry. Deviation from nominal is
illustrated by the colour of the surface, as shown in Figure 5.3.
49
Figure 5.1 - Inspection Report of Experiment Case 0, Roland Scan
Data
50
Figure 5.2 - Case 0, Roland Scan Data – Colour Map adjusted to show
detail.
51
Figure 5.3 - Inspection Report of Experiment Cases 1-4, Roland Scan
Data
52
Figure 5.4 – Cases 1-4, Roland Scan Data – Colour Map adjusted to
show detail.
53
Table 5.1 - Experiment Inspection Data (Roland) Statistics
Average +0.011mm -0.021mm
Standard
0.021 mm
Deviation
These results are encouraging. During their work on surface flank
milling of jet engine and compressor airfoils, Wu observed that deviations in
the range of 0.075 mm to 1.7 mm to the optimal airfoil shape as generated by
Computational Fluid Dynamics software, did not cause noticeable losses in
performance tests [43]. From the standard deviation of 0.021 mm, the system
has a 98.26% ability to achieve the DIN ISO 2768 f (fine) class tolerance of
±0.05 mm for the standard’s smallest nominal linear dimension category, 0.5-
3 mm.
54
Table 5.2 - Sphere Fit Statistics
Standard
Fitted Uncertainty
Deviation of
Sphere Radius of Centre
Residuals
(mm) (mm)
(mm)
Case 0 Sphere 1 12.7130 0.0150 0.0075
Sphere 2 12.7130 0.0150 0.0075
Sphere 3 12.7118 0.0145 0.0073
Case 1 Sphere 1 12.7153 0.0065 0.0033
Sphere 2 12.7140 0.0130 0.0065
Sphere 3 12.7133 0.0145 0.0073
Case 2 Sphere 1 12.7125 0.0130 0.0065
Sphere 2 12.7145 0.0150 0.0075
Sphere 3 12.7128 0.0155 0.0078
Case 3 Sphere 1 12.7160 0.0100 0.0050
Sphere 2 12.7143 0.0155 0.0078
Sphere 3 12.7113 0.0155 0.0078
Case 4 Sphere 1 12.7133 0.0125 0.0063
Sphere 2 12.7160 0.0130 0.0065
Sphere 3 12.7130 0.0130 0.0065
Average 12.7136 0.0134 0.0067
Additional analysis of experimental data was performed in order to
trace the errors introduced by each step. First, the OLS sphere fit statistics
are analyzed, showing an average residual of 0.0134 mm, and an average
uncertainty of the centre location of 0.0067 mm. The complete dataset is
given in Table 5.2.
55
Table 5.3 - Sphere Centre Distances
Difference from Case 0

Distance (mm) (mm)
Sphere 1 Sphere 2 Sphere 1 Sphere 2
Sphere 2 273.0136 0.0000
Case 0
Sphere 3 273.0408 385.9398 0.0000 0.0000
Sphere 2 273.0094 -0.0042
Case 1
Sphere 3 273.0653 385.9373 0.0245 -0.0025
Sphere 2 273.0097 -0.0039
Case 2
Sphere 3 273.0416 385.9195 0.0009 -0.0203
Sphere 2 272.9890 -0.0246
Case 3
Sphere 3 273.0402 385.9184 -0.0006 -0.0214
Sphere 2 272.9935 -0.0201
Case 4
Sphere 3 273.0522 385.9466 0.0115 0.0068
A simple metric to check for warpage in the plate is the distance
between the sphere centres. Table 5.3 gives the centre-to-centre distances
calculated between each sphere for each case. To compare the distances for
each case, the difference from Case 0 was also computed. The data does not
suggest that any significant amount of was present, but Table 5.4 gives a
better analysis using an OLS plane fit of probe data obtained from the plate
surface.
56
Table 5.4 - Plane Fitting Data
Case 0 Case 1 Case 2 Case 3 Case 4
X 0.5890 3.2840 18.6880 -46.8100 -33.5140

Centroid
Y 1.5120 1.9350 -6.5650 -6.8000 -7.3950
(mm)
Z 63.9985 63.9987 63.9989 64.0005 64.0017
X 0.0000 -0.0001 -0.0001 -0.0001 -0.0001
Direction
Y 0.0000 0.0000 0.0000 0.0001 0.0001
Cosines
Z 1.0000 1.0000 1.000 1.0000 1.0000
Residual
Standard
0.0160 0.0167 0.0175 0.0189 0.0182
Deviation
(mm)
The difference between the minimum (Case 0) and maximum (Case 4)
centroid Z height is 3.2 µμm. For each case, direction cosines describe the plane
to be effectively normal to the Z axis. The standard deviation of the data fit
residuals indicates the plate surface flatness was consistent between trials. A
photograph of the machined workpiece is presented in Figure 5.5.
57
Figure 5.5 - Photograph of Machined Experiment Workpiece
58
5.2 Analysis of Expected Errors
The following components of the machining experiment can be
expected to contribute error:
1. Positional Accuracy of the CNC machine during measurement
2. Probe Accuracy during measurement
3. Positional Accuracy of the CNC machine during machining
Distortion of the workpiece induced by clamping or machining forces
can be avoided through the use of a fixture designed to provide support at
critical areas of the part. If a fixture is not available, special care must be
taken to ensure clamping forces do not induce distortion. In the experiment
detailed in this thesis, 2x4x6 blocks were used to support the entire bottom of
the fixture plate. Additionally, the mating surfaces of the machine and the
workpiece were cleaned to ensure the surfaces were free of chips or other
debris that could cause workpiece distortion or damage to either surface.
The positional accuracy of the CNC used during measurement is
expected to be repeatable after homing. The machine used in this experiment
had a digital read out with resolution of 1.0 µμm. Occasional malfunction
issues were experienced when homing the Z axis. In one instance the machine
homed to a location 0.180mm away from the previous day’s home location,
prompting a restart of the experiment.
59
The probe accuracy is 1.0 µμm when used with the 50 mm stylus.
Calibration using 36 equally spaced probing vectors, including the probing
vector directions used in the experiment, was done using a ring gauge and
pre-travel was found to be 67 µμm for the +/- X and Y probing directions.
Because a single calibration value is used the pre-travel variation may
contribute an error of up to 10.0 µμm.
The results of the OLS sphere centre calculation are used in an OLS
data set fit to compute the HTM used for Toolpath adjustment. The HTM is
calculated to 4 decimal places, making the error contribution due to rounding
negligible.
The accuracy of the machine tool is a factor again during machining.
The magnitude and effect of each of the expected errors is summarized in
Table 5.5.
Table 5.5 - Summary of Expected Errors
Error Magnitude (2𝜎)

CNC Position During 1.0 µμm
Inspection
Probe Repeatability 1.0 µμm
Pre-Travel Compensated 10.0 µμm
Probe Accuracy
CNC Position During 1.0 µμm
Machining
Total Expected 13.0𝛍m
60
Chapter 6
Laser Scanner Integration
For soft or fragile objects, the contact pressure from touch-based
digitizers can be enough distort or destroy the object. Also, the speed of touch-
based digitizers is limited by kinematic and dynamic constraints. Laser
scanning technology uses the known properties of light to establish
information about a surface. Advancements in image sensor and image
processing technology have enabled the design of digitizers that can measure
with very high speed and accuracy.
The Gocator 2300 scan head used in this chapter was purchased from
LMI Technologies Inc. The scanner uses the triangulation principle shown in
Figure 2.9 implemented using a laser stripe and 2D image sensor. The
measurement specifications given by the manufacturer are given in Table 6.1.
The definition of clearance distance, measurement range, near field of view
and far field of view are shown in Figure 6.1.
61
Table 6.1 - Gocator 2330 Laser Scanner Specifications
Resolution Z (mm) 0.006-0.014

Resolution X (mm) 0.044-0.075
Clearance Distance (mm) 90
Measurement Range (mm) 80
Field of View (FOV) (mm) 47-85
Figure 6.1 - Gocator 2330 Specifications Diagram [44]
A webserver contained in the scan head hosts a configuration web
page, and enables data transmission over Ethernet.
62
6.1 LMI Scan Head Integration
Since measurements are only made at the intersection of the laser line
and the part surface, the sensor must be attached to a movable platform in
order to obtain measurements of the entire part. The platform must be able to
report the exact location of the scan head so that the data obtained by the
scan head can be registered to a single, world coordinate system. The
following subchapters describe how the sensor was integrated with the IOTA-
P Coordinate Measuring Machine (CMM) located at McMaster University.
Three Renishaw model RGH22-X10D00 optical linear encoders installed on
the CMM were used to provide location resolution to 1.0µm. The linear
encoder output is captured by a PC based Measurement Computing
Corporation PCI-QUAD04 Quadrature Counter card. During operation, a
custom written PC application matches the encoder location information with
the collected scan head point data.
63
Figure 6.2 - CMM-Laser Scanner Hardware Architecture Diagram
6.2 Mount Design
A sensor mounting bracket designed and machine to be compatible
with a mounting system for previous equipment (Appendix B). The mount
allows the sensor attach to the z-axis and to rotate about the y-axis. A bolt-
hole pattern and locking pin allows the sensor to be angled at 15 degree
increments. To extend the mounting system to a CNC machining centre, a
tool holder is required to attach the sensor to the spindle, and the spindle
orientation would be locked using the M19 code, as is customary when using a
touch-trigger probe on a CNC milling machine.
64
Figure 6.3 - 3D Scanner Mount with bolt hole pattern for locking pin
65
Figure 6.4 – LMI 3D Gocator 2300 Scanner on an IOTA P CMM
6.3 Desktop Control Software
The developed control software performs three functions. First, the
motion of the scan head is commanded by connecting to the CMM motion
control computer over Ethernet. This allows homing, and linear move
66
commands to be issued to the motion controller remotely. In a CNC
machining setup, the motion commands must exist in a NC program.
The second function performed by the desktop application is the
reading of quadrature counter card memory values, containing the axes
position information.
The third and final function performed by the desktop application is
commanding of the laser scan head functions. The scanner settings were
configured to wait for a trigger command from the computer application
before capturing a 3D profile.
67
Read Current Location from Counter

Card Memory Register
Send Trigger Command to 3D Scannner
Send “receive_data”
command to 3D Scanner
Scan data
Wait
available?
Receive Data
Apply Scanner Location Offset based on Current Location
Store Data in Memory
Figure 6.5 – Flow Chart of PC Application Scanning Function
68
Figure 6.6 - Geometric Fit of Captured Point Cloud with 0.0684 mm
error in diameter.
After homing the CMM and the initialization of the scan head, the scan
can be started by clicking a button on the user interface. The application
begins by reading the current location from the counter memory registers,
sending the trigger command to the scan head, and then the “receive_data”
command. Data is received into allocated memory where the scanner location
offset (or rigid transformation) is applied.
Capturing a single frame of data consists of 1280 points, and is
performed at 200Hz. Faster speeds are possible by reducing sample density
69
and through optimal lighting conditions in order to minimize the camera
exposure time.
The captured data is stored uncompressed, in comma separated
variable format for importing by any point cloud viewing software.
Figure 6.7 – CMM Laser Scanner Application Graphical User
Interface
70
6.4 Expected Errors
The errors in the laser scanning system can be attributed to the
following factors:
1. Positional error of the CMM due to thermal distortion
2. Positional error of the CMM linear scales
3. Delay between capture of positional data and capture of scan frame
4. Accuracy of laser scanner
The positional error of the CMM due to thermal distortion was not
considered in these experiments. The small scanning distance and short time
span reduce the effects of these errors.
The positional accuracy of the CMM linear scales given by the
manufacturer is 1.0 µμm.
Time delay between the capture of the position and the scan data
creates a distance error equal to the scanning head velocity multiplied by the
delay. The scanning velocity was approximately 2.5mm/second, and the
scanning period of 0.005s limits the maximum distance error to 12.5 µμm in
the Y direction. This error can be accounted for and corrected if the time
delay and velocity are known at the time of scanning.
The resolution of the laser scanner is given as 0.006-0.014mm in Z, and
0.044-0.075mm in X. Since the scan line X direction is along the part cylinder
axis, only the Z scanner error will be a factor in the surface deviation report.
71
The magnitude of the expected final error is presented in Table 6.2.
Table 6.2 - CMM Laser Cumulative Error
Error Magnitude Surface Scan Error (Z)

CMM Linear Scale 1.0 µμm 1.0 µμm
Time Delay 12.5 µμm in X Negligible
Laser Scanner 0.006-0.014 mm in Z, 0.014mm
Total Error Up to 15 𝛍m
6.5 Test Data
The resulting workpiece from the machining experiment in Chapter 4
was scanned using the assembled laser scanner and CMM system. The 30
second scan captured 200 frames per second, each generating 1280 points for
a total point count of 7.68 million points. The point spacing in the X direction
of the scan head was approximately 62.5 µm, and the point spacing in the Y
direction was approximately 0.127 mm.
72
Figure 6.8 - LMI Scan Data: Case 0 Surface Deviation Analysis
73
Figure 6.9 - Case 0, LMI Scan Data – Colour Map adjusted to show
detail.
74
Figure 6.10 - LMI Scan Data: Cases 1-4 Surface Deviation Analysis
75
Figure 6.11 – Cases 1-4, LMI Scan Data – Colour Map adjusted to show
detail.
76
Table 6.3 - LMI Scan Data Surface Fit Statistics
Average +0.010mm -0.022mm
Standard
0.019 mm
Deviation
The reference surface fit in Figure 6.8 showed a standard deviation of
13 µm. The manufacturer claimed Z resolution of the LMI scanner (0.006-
0.014 mm) shows a much higher level of detail than the Roland scanner
which claims an accuracy of 0.05mm. The higher Z resolution allows the final
surface waviness characteristic of the ball nose end mill to be seen from the
surface deviation report.
The scan density in this experiment is higher than required for
successful fitting, but demonstrates the ability of laser scanning sensors to
capture a point cloud much faster than is possible with a touch trigger probe
measuring single points.
77
Chapter 7
Conclusions and Future Work
CNC machining has evolved from manually entered and tape punched
instructions, to modern digitally stored and distributed programs. These
systems currently operate on one-way data flow, and assume geometric
consistency. In reality, factors such as casting variation, heat treatment
distortion, and assembly creep wear contribute to workpiece geometry
variations that require measurement required prior to machining. Scrap
avoidance can be a reality by using CNC mounted inspection sensors, such as
a touch trigger probe, together with existing inspection software employed
with CMM inspection data. A method has been developed and demonstrated
to allows a workpiece loaded in a CNC machine to be measured with a touch
trigger probe, analyzed using GD&T software, and machined using a tool
path updated based on the measurement and analysis.
In Chapter 4 the demonstrated pre-machining inspection procedure
determined workpiece orientation in order to detect workpiece misalignment.
A computer connected to the machining centre via a Universal Machine
interface generated an updated toolpath for correct machining of the
misaligned part. This process was tested using a fixturing plate with three
spherical datums, fixtured in 4 different instances, and laser scan surface
78
analysis has shown an average deviation of +0.010/-0.022mm from nominal
geometry, and a standard deviation of 0.019mm.
In Chapter 6 the feasibility of integrating a laser scanner with CNC
was investigated using a CMM with optical linear encoders and a laser
scanner with an Ethernet data interface. A computer application was created
to link the captured 3D point data to the CMM location data, and a
calibration sphere, and the workpiece from Chapter 4 were scanned with
accuracy comparable to commercial scanning systems.
7.1 Future Work
The registration method in this work uses spherical datums, but there
may be situations where this is not possible. Future work will include trials
using part features as datums, and implementing the system on a 5-axis CNC
machining centre.
The use of a laser scanner in a CNC machine tool offers decreased
inspection time due to higher data capture rates compared to contact-based
digitizers. The position information required by the scan head may be
obtained by connecting to the existing position feedback hardware, or by
installing additional encoders that can operate in parallel. There are a
number of challenges specific to optical inspection, such as reflectivity, the
presence of coolant or chips, as well as liquid droplets on the lens of the
79
sensor. These may be overcome by methods used by other lasers in
contaminating environments such as CNC grinding machines, where a
compartment houses the scanner with a small opening allowing an un-
restricted line of sight while air pressure keeps air-borne contaminants out.
The most promising future work is the potential unlocked by having
measurement and communication ability at the CNC control. The presence of
advanced manufacturing software at the CNC control enables the design
intentions of manufacturing engineers to be present as each part is machined.
80
Appendix
A. CMM Spindle Mount
81
B. 24 Position Rotary Adapter for CMM Mounted Laser Scanner
82
C. Ring Gauge Probing NC Code (First 50 Lines Only)
%
O1040
(A.SAWULA)
(RING GAUGE CALIBRATION)
G21 G40 G90 G54
M19
#31 = 100.0
POPEN
DPRNT[Index,X,Y]
PCLOS
G1 X0.0 Y0.0
(PROBE HIT1)
G31 X21.7889 Y249.0487
#1 = #5061
#2 = #5062
G1 X0.0 Y0.0
POPEN
DPRNT[1,#1 [44], #2 [44]
PCLOS
(PROBE HIT2)
G31 X43.4120 Y246.2019
#1 = #5061
#2 = #5062
G1 X0.0 Y0.0
POPEN
DPRNT[2,#1 [44], #2 [44]
PCLOS
(PROBE HIT3)
G31 X64.7048 Y241.4815
#1 = #5061
#2 = #5062
G1 X0.0 Y0.0
POPEN
DPRNT[3,#1 [44], #2 [44]
PCLOS
(PROBE HIT4)
G31 X85.5050 Y234.9232
#1 = #5061
#2 = #5062
G1 X0.0 Y0.0
POPEN
DPRNT[4,#1 [44], #2 [44]
PCLOS
(PROBE HIT5)
G31 X105.6546 Y226.5769
#1 = #5061
#2 = #5062
G1 X0.0 Y0.0
POPEN
DPRNT[5,#1 [44], #2 [44]
PCLOS
83
D. Sphere Probing Routine NC Code
%
O1050
(PROBES 1)
(MEASURE 1INCH DIA SPHERE)
(START AT XY NEAR CENTRE, Z ABOVE CENTRE)
(PROGRAM SETS X AND Y PROBE TRAJECTORIES BASED ON FIRST HIT IN Z-)
(CENTRE CALCULATED FROM THE AVERAGE OF PAIRS OF X AND Y HITS)
#20 = 5. (PROBE TIP DIA.)
#21 = 0.062 (PROBE FLEX COMPENS.0.062)
#22 = [#20/2]-#21
#23 = 273. (SPHERE SPACING)
#25 = 15.0 (SPHERE CLEARANCE)
#30 = 25.400 (SPHERE DIA.)
#31 = 100 (FAST PROBE SPEED)
#32 = 20 (SLOW PROBE SPEED)
#33 = 4000 (SAFE RAPID MOVE SPEED)
G21 G40 G90 G54
M19
G31 Z0.0 F#31 (PROBE Z)
#1 = #5061
#2 = #5062
#3 = [#5063-#20/2]
G1 Z[#3+#25] F#33 (UP)
G31 X[#1+[#30]+#20] F#33 (OVER+)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 X[#1] F#31 (PROBE X1 FAST-)
G1 X[#5061+#22]F#33(RETRACT+)
G31 X[#1] F#32 (PROBE X1 SLOW-)
#4 = #5061-#22
#5 = #5062
#6 = #5063
G1 X[#4+#20] F#33 (RETRACT+)
G31 Z[#3+#25] F#33 (UP)
G31 X[#1-[#30]-#20] F#33 (OVER-)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 X[#1] F#31 (PROBE X2 FAST+)
G1 X[#5061-#22](RETRACT+)
G31 X[#1] F#32 (PROBE X2 SLOW+)
#7 = #5061+#22
#8 = #5062
#9 = #5063
G1 X[#1-[#30]-#20] F#33 (RETRACT)
G31 Z[#3+#25] F#33 (UP)
G31 X[#1] Y[#2+[#30]+#20] F#33 (OVER+)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 Y[#2] F#31 (PROBE Y1 FAST-)
G1 Y[#5062+#22](RETRACT+)
G31 Y[#2] F#32 (PROBE Y1 SLOW-)
#10 = #5061
#11 = #5062-#22
#12 = #5063
G1 Y[#2+[#30]+#20] F#33 (RETRACT+)
84
G31 Z[#3+#25] F#33 (UP)

G31 Y[#2-[#30]-#20] F#33 (OVER-)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 Y[#2] F#31 (PROBE Y2 FAST+)
G1 Y[#5062-#22](RETRACT-)
G31 Y[#2] F#32 (PROBE Y1 SLOW+)
#13 = #5061
#14 = #5062+#22
#15 = #5063
G1 Y[#2-[#30]-#20] F#33 (RETRACT-)
G31 Z[#3+#25] F#33 (UP)
#16=[#4+#7]/2 (X)
#17=[#11+#14]/2 (Y)
#18= #4-#7
(TO CENTRE)
G31 X#16 Y#17 F#33
G31 Z[#3-150.0] F#31
#1= #5061
#2= #5062
#3 = #5063-#20/2
#19 = #3-[#30/2]
G1 Z[#5063+#20] F#33
G31 Z[#3-150.0] F#31 (PROBE Z)
#1 = #5061
#2 = #5062
#3 = [#5063-#20/2]
G1 Z[#3+#25] F#33 (UP)
G31 X[#1+[#30]+#20] F#33 (OVER+)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 X[#1] F#31 (PROBE X1 FAST-)
G1 X[#5061+#22](RETRACT+)
G31 X[#1] F#32 (PROBE X1 SLOW-)
#4 = #5061-#22
#5 = #5062
#6 = #5063
G1 X[#4+#20] F#33 (RETRACT+)
G31 Z[#3+#25] F#33 (UP)
G31 X[#1-[#30]-#20] F#33 (OVER-)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 X[#1] F#31 (PROBE X2 FAST+)
G1 X[#5061-#22](RETRACT+)
G31 X[#1] F#32 (PROBE X2 SLOW+)
#7 = #5061+#22
#8 = #5062
#9 = #5063
G1 X[#1-[#30]-#20] F#33 (RETRACT)
G31 Z[#3+#25] F#33 (UP)
G31 X[#1] Y[#2+[#30]+#20] F#33 (OVER+)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 Y[#2] F#31 (PROBE Y1 FAST-)
G1 Y[#5062+#22](RETRACT+)
G31 Y[#2] F#32 (PROBE Y1 SLOW-)
#10 = #5061
#11 = #5062-#22
#12 = #5063
85
G1 Y[#2+[#30]+#20] F#33 (RETRACT+)

G31 Z[#3+#25] F#33 (UP)
G31 Y[#2-[#30]-#20] F#33 (OVER-)
G31 Z[#3-[#30/2]] F#33 (DOWN)
G31 Y[#2] F#31 (PROBE Y2 FAST+)
G1 Y[#5062-#22](RETRACT-)
G31 Y[#2] F#32 (PROBE Y2 SLOW+)
#13 = #5061
#14 = #5062+#22
#15 = #5063
G1 Y[#2-[#30]-#20] F#33 (RETRACT-)
G31 Z[#3+#25] F#33 (UP)
#16=[#4+#7]/2 (X)
#17=[#11+#14]/2 (Y)
#18= #4-#7
(TO CENTRE)
G31 X#16 Y#17 F#33
G31 Z[#3-150.0] F#31
#1= #5061
#2= #5062
#3 = #5063-[#20/2]
#19 = #3-[#30/2]
G1 Z[#5063+#20] F#33
POPEN
DPRNT[SPHERE1]
DPRNT[***]
DPRNT[5]
DPRNT[*#1 [44]* #2 [44]* #3 [44]]
DPRNT[*#4 [44]* #5 [44]* #6 [44]]
DPRNT[*#7 [44]* #8 [44]* #9 [44]]
DPRNT[*#10 [44]* #11 [44]* #12 [44]]
DPRNT[*#13 [44]* #14 [44]* #15 [44]]
DPRNT[***]
PCLOS
M30
%
86
E. Nominal Tool Path from CAM Software (first 50 Lines only)
%
:0001
(BladeBaseMtl2012ZUp)
N10 G21 G90 G40
N20 G10 P1 Z85.0 R4.0 T00
N30 G10 L2 P1 X0.0 Y0.0 Z0.0 (Top)
N40 (DEFINE OPERATION : PARALLEL LACE OPERATION)
N50 G0 X0.0 Y0.0 Z100.0
N60 M5
N70 G28 G91 Z0 H0
N80 G28 X0 Y0
O90 G21 G90 G40 G94
N100 T00 M06 (8.0 MM DIA BALL NOSE MILL)
N110 G54P1
N120 T00 M1
N130 S5000 M3 M42
N140 G0 X0.0 Y0.0
N150 G43 Z50.0 H00 M7
N160 G0 X-25.399 Y35.004 Z50.0
N170 G0 X-25.399 Y35.004 Z23.486
N180 G1 X-25.399 Y35.004 Z18.486 F1000.0
N190 G1 X25.391 Y35.004 Z18.486
N200 G0 X25.391 Y35.004 Z50.0
N210 G0 X-25.399 Y34.48 Z50.0
N220 G0 X-25.399 Y34.48 Z23.698
N230 G1 X-25.399 Y34.48 Z18.698
N240 G1 X25.391 Y34.48 Z18.698
N250 G0 X25.391 Y34.48 Z50.0
N260 G0 X-25.399 Y33.955 Z50.0
N270 G0 X-25.399 Y33.955 Z23.909
N280 G1 X-25.399 Y33.955 Z18.909
N290 G1 X25.391 Y33.955 Z18.909
N300 G0 X25.391 Y33.955 Z50.0
N310 G0 X-25.399 Y33.431 Z50.0
N320 G0 X-25.399 Y33.431 Z24.12
N330 G1 X-25.399 Y33.431 Z19.12
N340 G1 X25.391 Y33.431 Z19.12
N350 G0 X25.391 Y33.431 Z50.0
N360 G0 X-25.399 Y32.901 Z50.0
N370 G0 X-25.399 Y32.901 Z24.316
N380 G1 X-25.399 Y32.901 Z19.316
N390 G1 X25.391 Y32.901 Z19.316
N400 G0 X25.391 Y32.901 Z50.0
N410 G0 X-25.399 Y32.37 Z50.0
N420 G0 X-25.399 Y32.37 Z24.511
N430 G1 X-25.399 Y32.37 Z19.511
N440 G1 X25.391 Y32.37 Z19.511
N450 G0 X25.391 Y32.37 Z50.0
N460 G0 X-25.399 Y31.84 Z50.0
N470 G0 X-25.399 Y31.84 Z24.707
87
F. Transformed Tool Path (First 50 Lines only)
%
O1064
(ALAN-S-HUB4)
G21 G90 G40
(DEFINE OPERATION : PARALLEL LACE OPERATION)
G0 X24.886 Y0.215 Z35.016
G21 G90 G40 G94 (8.0 MM DIA BALL NOSE MILL)
G54
S5000 M3
G0 X24.886 Y0.215 Z35.016
G0 X10.972 Y25.275 Z-14.973
G0 X10.966 Y25.281 Z-35.588
G1 X10.964 Y25.283 Z-40.588 F250.0
G1 X12.282 Y-25.493 Z-40.603
G0 X12.290 Y-25.500 Z-14.988
G0 X10.412 Y25.261 Z-14.973
G0 X10.406 Y25.267 Z-35.671
G1 X10.404 Y25.268 Z-40.671
G1 X11.722 Y-25.508 Z-40.685
G0 X11.730 Y-25.515 Z-14.987
G0 X9.853 Y25.246 Z-14.972
G0 X9.847 Y25.252 Z-35.753
G1 X9.846 Y25.254 Z-40.753
G1 X11.163 Y-25.522 Z-40.768
G0 X11.171 Y-25.529 Z-14.987
G0 X9.294 Y25.232 Z-14.972
G0 X9.288 Y25.238 Z-35.836
G1 X9.287 Y25.239 Z-40.836
G1 X10.605 Y-25.537 Z-40.851
G0 X10.612 Y-25.544 Z-14.987
G0 X8.738 Y25.217 Z-14.972
G0 X8.731 Y25.223 Z-35.933
G1 X8.730 Y25.225 Z-40.933
G1 X10.048 Y-25.551 Z-40.948
G0 X10.056 Y-25.558 Z-14.987
G0 X8.181 Y25.203 Z-14.972
G0 X8.175 Y25.209 Z-36.033
G1 X8.173 Y25.210 Z-41.033
G1 X9.491 Y-25.565 Z-41.048
G0 X9.499 Y-25.573 Z-14.987
G0 X7.625 Y25.189 Z-14.972
G0 X7.619 Y25.195 Z-36.132
G1 X7.617 Y25.196 Z-41.132
G1 X8.935 Y-25.580 Z-41.147
G0 X8.943 Y-25.587 Z-14.987
G0 X7.068 Y25.174 Z-14.972
G0 X7.062 Y25.180 Z-36.231
G1 X7.060 Y25.182 Z-41.231
G1 X8.378 Y-25.594 Z-41.246
G0 X8.386 Y-25.602 Z-14.987
G0 X6.512 Y25.160 Z-14.97
88
G. CMM Laser Scanner Application Source Code
1 #pragma once
2
3
4 #include "stdafx.h"
5 #include "Form1.h"
6 #include <iostream>
7 #include <fstream>
8 using namespace std;
9
10 //cmm motion communication
11 #include <winsock2.h>
12 #include <ws2tcpip.h>
13 #include <stdio.h>
14
15 //encoder quadrature capture card
16 #include <conio.h>
17 #include "..\cbw.h"
18
19 //scanner
20 #include <Go2.h>
21 #include <stdlib.h>
22 #include <memory.h>
23 #include <windows.h>
24 #include <time.h>
25
26 #pragma comment(lib, "Ws2_32.lib")
27 #pragma comment(lib, "comdlg32.lib")
28 #pragma comment(lib, "Go2.lib")
29 #pragma comment(lib, "user32.lib")
30
31 #define RECEIVE_TIMEOUT 20000
32 #define INVALID_RANGE_16BIT (0x8000) // gocator transmits
range data as 16-bit signed integers. 0x8000 signifies invalid range
data.
33 #define DOUBLE_MAX (1.79769e+308) // 64-bit double - largest
positive value.
34 #define INVALID_RANGE_DOUBLE (-DOUBLE_MAX) // floating point
value to represent invalid range data.
35
36 SOCKET ConnectSocket = INVALID_SOCKET;
37 int iResult;
38 int testCount = 0;
39 /* Encoder Variable Declarations */
40 //int Row, Col;
41 int ULStat = 0;
42 int BoardNum = 0;
43 int CounterNum, Quadrature, CountingMode, DataEncoding, IndexMode;
44 int InvertIndex, FlagPins, GateEnable;
45 long LoadValue;
46 unsigned long Count, lastCount, StatusBits;
47 unsigned long xCount, xlastCount, xStatusBits;
89
48 unsigned long yCount, ylastCount, yStatusBits;

49 unsigned long zCount, zlastCount, zStatusBits;
50 int RegName;
51 float RevLevel = (float)CURRENTREVNUM;
52 char *DirectionStr;
53 //Scanner Variable Declarations
54 typedef struct
55 {
56 short x; // x-coordinate as 16-bit signed integer - position
along laser line
57 short z; // z-coordinate as 16-bit signed integer - height (at
the given x position)
58 }ProfilePoint16s;
59
60 typedef struct
61 {
62 double x; // x-coordinate in engineering units (mm) - position
along laser line
63 double y; // y-coordinate in engineering units (mm) - derived
from scanner location
64 double z; // z-coordinate in engineering units (mm) - height (at
the given x position)
65 }ProfilePoint;
66 int status;
67 Go2System scannerSystem = GO2_NULL;
68 Go2Data data = GO2_NULL;
69 Go2IPAddress address;
70 ProfilePoint ** memory = NULL;
71 int j;
72 int i;
73 unsigned int arrayIndex;
74 unsigned int timeout = 20000;
75 double XScannerOffset = 0;
76 double YScannerOffset = 0;
77 double ZScannerOffset = 0;
78 double YMax = 400;
79 double ZMax = 230;
80
81 int scanCount = 200* 120;//200 hz * scanDurationSeconds
82 int scanIndex = 0; //stores the location where to store the next scan
data
83 clock_t start, end;
84 double cpu_time;
85
86
87
88 namespace SimpleForm {
89
90 using namespace System;
91 using namespace System::ComponentModel;
92 using namespace System::Collections;
93 using namespace System::Windows::Forms;
94 using namespace System::Data;
95 using namespace System::Drawing;
96 using namespace System::Runtime::InteropServices;
90
97
98 /// <summary>
99 /// Summary for Form1
100 /// </summary>
101 public ref class Form1 : public System::Windows::Forms::Form
102 {
103 public:
104 Form1(void)
105 {
106 InitializeComponent();
107
108 }
109
110 protected:
111 ~Form1()
112 {
113 // shutdown the connection for sending since no more data will be
sent
114 // the client can still use the ConnectSocket for receiving data
115 iResult = shutdown(ConnectSocket, SD_SEND);
116 if (iResult == SOCKET_ERROR) {
117 printf("shutdown failed: %d\n", WSAGetLastError());
118 closesocket(ConnectSocket);
119 WSACleanup();
120 // return 1;
121 }
122
123 //Scanner Cleanup
124 // stop Gocator system
125 if ((status = Go2System_Stop(scannerSystem)) != GO2_OK)
126 {
127 printf ("Error: Go2System_Stop:%d\n", status);
128 return;
129 }
130
131 // destroy system handle
132 if ((status = Go2System_Destroy(scannerSystem)) != GO2_OK)
133 {
134 printf ("Error: Go2System_Destroy:%d\n", status);
135 return;
136 }
137
138 // terminate Go2 API
139 if ((status = Go2Api_Terminate()) != GO2_OK)
140 {
141 printf ("Error: Go2Api_Terminate:%d\n", status);
142 return;
143 }
144 printf("Press any key to Free Memory...\n", status);
145 getchar();
146 //free memory array
147 for(i = 0; i < scanCount; i++)
148 free(memory[i]);
149 free(memory);
150 //End Scanner Cleanup
91
151
152 if (components)
153 {
154 delete components;
155 }
156 }
157 private: System::Windows::Forms::Button^ button1;
158 protected:
159
160
161
162 public: System::Windows::Forms::TextBox^ textBox1;
163 public: System::Windows::Forms::Label^ label2;
164 public: System::Windows::Forms::RichTextBox^ richTextBox1;
165 public: System::Windows::Forms::Label^ staticStatusLabel;
166 public: System::Windows::Forms::Label^ statusLabel;
167 public: System::Windows::Forms::Button^ button2;
168
172
173 public: System::Windows::Forms::Label^ xlabel;
174 public: System::Windows::Forms::Label^ ylabel;
175 public: System::Windows::Forms::Label^ zlabel;
176 private: System::Windows::Forms::GroupBox^ groupBox1;
177 public: System::Windows::Forms::Timer^ timer1;
178 public: System::Windows::Forms::CheckBox^ checkBox1;
179 public: System::Windows::Forms::RichTextBox^ richTextBox2;
180 public: System::Windows::Forms::Button^ buttonStartScan;
181 public: System::Windows::Forms::Button^ buttonStopScan;
182 private: System::Windows::Forms::Button^ buttonSaveData;
183 private: System::Windows::Forms::Timer^ timerScan;
184 public: System::Windows::Forms::Label^ labelScanSize;
185 private: System::Windows::Forms::Label^ label1;
186 public: System::Windows::Forms::ProgressBar^ progressBar1;
187 private: System::Windows::Forms::TrackBar^ trackBar1;
188 public: System::Windows::Forms::Label^ labelScanRateHz;
189 private: System::Windows::Forms::Button^ buttonRestartScan;
190 protected:
191 private: System::ComponentModel::IContainer^ components;
192
193
194 #pragma region Windows Form Designer generated code
195 /// <summary>
196 /// Required method for Designer support - do not modify
197 /// the contents of this method with the code editor.
198 /// </summary>
199 void InitializeComponent(void)
200 {
201 this->components = (gcnew System::ComponentModel::Container());
202 this->button1 = (gcnew System::Windows::Forms::Button());
203 this->textBox1 = (gcnew System::Windows::Forms::TextBox());
204 this->label2 = (gcnew System::Windows::Forms::Label());
92
205 this->richTextBox1 = (gcnew

System::Windows::Forms::RichTextBox());
206 this->staticStatusLabel = (gcnew System::Windows::Forms::Label());
207 this->statusLabel = (gcnew System::Windows::Forms::Label());
208 this->button2 = (gcnew System::Windows::Forms::Button());
212 this->xlabel = (gcnew System::Windows::Forms::Label());
213 this->ylabel = (gcnew System::Windows::Forms::Label());
214 this->zlabel = (gcnew System::Windows::Forms::Label());
215 this->groupBox1 = (gcnew System::Windows::Forms::GroupBox());
216 this->checkBox1 = (gcnew System::Windows::Forms::CheckBox());
217 this->timer1 = (gcnew System::Windows::Forms::Timer(this-
>components));
218 this->richTextBox2 = (gcnew
System::Windows::Forms::RichTextBox());
219 this->buttonStartScan = (gcnew System::Windows::Forms::Button());
220 this->buttonStopScan = (gcnew System::Windows::Forms::Button());
221 this->buttonSaveData = (gcnew System::Windows::Forms::Button());
222 this->timerScan = (gcnew System::Windows::Forms::Timer(this-
>components));
223 this->labelScanSize = (gcnew System::Windows::Forms::Label());
225 this->progressBar1 = (gcnew
System::Windows::Forms::ProgressBar());
226 this->trackBar1 = (gcnew System::Windows::Forms::TrackBar());
227 this->labelScanRateHz = (gcnew System::Windows::Forms::Label());
228 this->buttonRestartScan = (gcnew
System::Windows::Forms::Button());
229 this->groupBox1->SuspendLayout();
230 (cli::safe_cast<System::ComponentModel::ISupportInitialize^
>(this->trackBar1))->BeginInit();
231 this->SuspendLayout();
232 //
233 // button1
234 //
235 this->button1->Location = System::Drawing::Point(288, 268);
236 this->button1->Name = L"button1";
237 this->button1->Size = System::Drawing::Size(75, 23);
238 this->button1->TabIndex = 0;
239 this->button1->Text = L"Send";
240 this->button1->UseVisualStyleBackColor = true;
241 this->button1->Click += gcnew System::EventHandler(this,
&Form1::button1_Click);
242 //
243 // textBox1
244 //
245 this->textBox1->AcceptsReturn = true;
246 this->textBox1->HideSelection = false;
247 this->textBox1->Location = System::Drawing::Point(69, 270);
248 this->textBox1->Name = L"textBox1";
249 this->textBox1->Size = System::Drawing::Size(213, 20);
250 this->textBox1->TabIndex = 2;
251 this->textBox1->Text = L"HECK2";
93
252 this->textBox1->TextChanged += gcnew System::EventHandler(this,

&Form1::textBox1_TextChanged);
253 //
254 // label2
255 //
256 this->label2->AutoSize = true;
257 this->label2->Location = System::Drawing::Point(9, 273);
258 this->label2->Name = L"label2";
259 this->label2->Size = System::Drawing::Size(57, 13);
260 this->label2->TabIndex = 3;
261 this->label2->Text = L"Command:";
262 //
263 // richTextBox1
264 //
265 this->richTextBox1->Location = System::Drawing::Point(9, 34);
266 this->richTextBox1->Name = L"richTextBox1";
267 this->richTextBox1->Size = System::Drawing::Size(271, 227);
268 this->richTextBox1->TabIndex = 5;
269 this->richTextBox1->Text = L"";
270 //
271 // staticStatusLabel
272 //
273 this->staticStatusLabel->AutoSize = true;
274 this->staticStatusLabel->Location = System::Drawing::Point(8, 9);
275 this->staticStatusLabel->Name = L"staticStatusLabel";
276 this->staticStatusLabel->Size = System::Drawing::Size(40, 13);
277 this->staticStatusLabel->TabIndex = 6;
278 this->staticStatusLabel->Text = L"Status:";
279 this->staticStatusLabel->Click += gcnew System::EventHandler(this,
&Form1::label3_Click);
280 //
281 // statusLabel
282 //
283 this->statusLabel->AutoSize = true;
284 this->statusLabel->Location = System::Drawing::Point(54, 9);
285 this->statusLabel->Name = L"statusLabel";
286 this->statusLabel->Size = System::Drawing::Size(228, 13);
287 this->statusLabel->TabIndex = 7;
288 this->statusLabel->Text = L"Unable to connect - Please Restart
Application";
289 //
290 // button2
291 //
292 this->button2->AutoSize = true;
293 this->button2->Location = System::Drawing::Point(20, 57);
294 this->button2->Name = L"button2";
295 this->button2->Size = System::Drawing::Size(118, 28);
296 this->button2->TabIndex = 8;
297 this->button2->Text = L"Set Position as Home";
298 this->button2->UseVisualStyleBackColor = true;
299 this->button2->Click += gcnew System::EventHandler(this,
&Form1::button2_Click);
300 //
301 // label3
302 //
94

304 this->label3->Font = (gcnew System::Drawing::Font(L"Calibri",
27.75F, System::Drawing::FontStyle
::Bold));
309 this->label3->Text = L"X:";
310 //
311 // label4
312 //
::Bold));
319 this->label4->Text = L"Y:";
320 //
321 // label5
322 //
::Bold, System::Drawing::GraphicsUnit::Point,
325 static_cast<System::Byte>(0)));
330 this->label5->Text = L"Z:";
331 //
332 // xlabel
333 //
334 this->xlabel->AutoSize = true;
335 this->xlabel->Font = (gcnew System::Drawing::Font(L"Calibri",
::Bold));
336 this->xlabel->Location = System::Drawing::Point(74, 94);
337 this->xlabel->Name = L"xlabel";
338 this->xlabel->Size = System::Drawing::Size(134, 45);
339 this->xlabel->TabIndex = 12;
340 this->xlabel->Text = L"123456";
341 //
342 // ylabel
343 //
344 this->ylabel->AutoSize = true;
345 this->ylabel->Font = (gcnew System::Drawing::Font(L"Calibri",
::Bold));
346 this->ylabel->Location = System::Drawing::Point(74, 151);
347 this->ylabel->Name = L"ylabel";
95
348 this->ylabel->Size = System::Drawing::Size(134, 45);

349 this->ylabel->TabIndex = 13;
350 this->ylabel->Text = L"123456";
351 //
352 // zlabel
353 //
354 this->zlabel->AutoSize = true;
355 this->zlabel->Font = (gcnew System::Drawing::Font(L"Calibri",
::Bold));
356 this->zlabel->Location = System::Drawing::Point(74, 207);
357 this->zlabel->Name = L"zlabel";
358 this->zlabel->Size = System::Drawing::Size(134, 45);
359 this->zlabel->TabIndex = 14;
360 this->zlabel->Text = L"123456";
361 //
362 // groupBox1
363 //
364 this->groupBox1->Controls->Add(this->xlabel);
365 this->groupBox1->Controls->Add(this->checkBox1);
366 this->groupBox1->Controls->Add(this->zlabel);
367 this->groupBox1->Controls->Add(this->button2);
368 this->groupBox1->Controls->Add(this->label3);
369 this->groupBox1->Controls->Add(this->ylabel);
372 this->groupBox1->Location = System::Drawing::Point(413, 34);
373 this->groupBox1->Name = L"groupBox1";
374 this->groupBox1->Size = System::Drawing::Size(239, 441);
375 this->groupBox1->TabIndex = 15;
376 this->groupBox1->TabStop = false;
377 this->groupBox1->Text = L"CMM Coordinates";
378 //
379 // checkBox1
380 //
381 this->checkBox1->AutoSize = true;
382 this->checkBox1->Location = System::Drawing::Point(20, 34);
383 this->checkBox1->Name = L"checkBox1";
384 this->checkBox1->Size = System::Drawing::Size(101, 17);
385 this->checkBox1->TabIndex = 16;
386 this->checkBox1->Text = L"Monitor Position";
387 this->checkBox1->UseVisualStyleBackColor = true;
388 this->checkBox1->CheckedChanged += gcnew
System::EventHandler(this, &Form1::
checkBox1_CheckedChanged);
389 //
390 // timer1
391 //
392 this->timer1->Interval = 5;
393 this->timer1->Tick += gcnew System::EventHandler(this,
&Form1::timer1_Tick);
394 //
395 // richTextBox2
396 //
397 this->richTextBox2->Location = System::Drawing::Point(286, 35);
96
398 this->richTextBox2->Name = L"richTextBox2";

399 this->richTextBox2->Size = System::Drawing::Size(121, 227);
400 this->richTextBox2->TabIndex = 17;
401 this->richTextBox2->Text = L"";
402 //
403 // buttonStartScan
404 //
405 this->buttonStartScan->Location = System::Drawing::Point(42, 332);
406 this->buttonStartScan->Name = L"buttonStartScan";
407 this->buttonStartScan->Size = System::Drawing::Size(123, 37);
408 this->buttonStartScan->TabIndex = 18;
409 this->buttonStartScan->Text = L"Start Scan";
410 this->buttonStartScan->UseVisualStyleBackColor = true;
411 this->buttonStartScan->Click += gcnew System::EventHandler(this,
&Form1::buttonStartScan_Click);
412 //
413 // buttonStopScan
414 //
415 this->buttonStopScan->Enabled = false;
416 this->buttonStopScan->Location = System::Drawing::Point(42, 375);
417 this->buttonStopScan->Name = L"buttonStopScan";
418 this->buttonStopScan->Size = System::Drawing::Size(123, 37);
419 this->buttonStopScan->TabIndex = 19;
420 this->buttonStopScan->Text = L"Stop Scan";
421 this->buttonStopScan->UseVisualStyleBackColor = true;
422 this->buttonStopScan->Click += gcnew System::EventHandler(this,
&Form1::buttonStopScan_Click);
423 //
424 // buttonSaveData
425 //
426 this->buttonSaveData->Location = System::Drawing::Point(226, 437);
427 this->buttonSaveData->Name = L"buttonSaveData";
428 this->buttonSaveData->Size = System::Drawing::Size(123, 38);
429 this->buttonSaveData->TabIndex = 20;
430 this->buttonSaveData->Text = L"Save Data";
431 this->buttonSaveData->UseVisualStyleBackColor = true;
432 this->buttonSaveData->Click += gcnew System::EventHandler(this,
&Form1::buttonSaveData_Click);
433 //
434 // timerScan
435 //
436 this->timerScan->Interval = 5;
437 this->timerScan->Tick += gcnew System::EventHandler(this,
&Form1::timerScan_Tick);
438 //
439 // labelScanSize
440 //
441 this->labelScanSize->AutoSize = true;
442 this->labelScanSize->Location = System::Drawing::Point(323, 375);
443 this->labelScanSize->Name = L"labelScanSize";
444 this->labelScanSize->Size = System::Drawing::Size(59, 13);
445 this->labelScanSize->TabIndex = 21;
446 this->labelScanSize->Text = L"ScanSize#";
447 //
448 // label1
97
449 //
455 this->label1->Text = L"Scan Profile Count:";
456 //
457 // progressBar1
458 //
459 this->progressBar1->Location = System::Drawing::Point(226, 408);
460 this->progressBar1->MarqueeAnimationSpeed = 1000;
461 this->progressBar1->Name = L"progressBar1";
462 this->progressBar1->Size = System::Drawing::Size(123, 23);
463 this->progressBar1->Step = 1;
464 this->progressBar1->Style =
System::Windows::Forms::ProgressBarStyle::Continuous;
465 this->progressBar1->TabIndex = 23;
466 //
467 // trackBar1
468 //
469 this->trackBar1->Location = System::Drawing::Point(202, 327);
470 this->trackBar1->Maximum = 200;
471 this->trackBar1->Name = L"trackBar1";
472 this->trackBar1->Size = System::Drawing::Size(205, 45);
473 this->trackBar1->TabIndex = 24;
474 this->trackBar1->Value = 200;
475 this->trackBar1->Scroll += gcnew System::EventHandler(this,
&Form1::trackBar1_Scroll);
476 //
477 // labelScanRateHz
478 //
479 this->labelScanRateHz->AutoSize = true;
480 this->labelScanRateHz->Location = System::Drawing::Point(210,
311);
481 this->labelScanRateHz->Name = L"labelScanRateHz";
482 this->labelScanRateHz->Size = System::Drawing::Size(35, 13);
483 this->labelScanRateHz->TabIndex = 25;
484 this->labelScanRateHz->Text = L"label6";
485 //
486 // buttonRestartScan
487 //
488 this->buttonRestartScan->Location = System::Drawing::Point(42,
442);
489 this->buttonRestartScan->Name = L"buttonRestartScan";
490 this->buttonRestartScan->Size = System::Drawing::Size(123, 23);
491 this->buttonRestartScan->TabIndex = 26;
492 this->buttonRestartScan->Text = L"Restart Scan";
493 this->buttonRestartScan->UseVisualStyleBackColor = true;
494 this->buttonRestartScan->Click += gcnew System::EventHandler(this,
&Form1::
buttonRestartScan_Click);
495 //
496 // Form1
497 //
98
498 this->AcceptButton = this->button1;

499 this->AutoScaleDimensions = System::Drawing::SizeF(6, 13);
500 this->AutoScaleMode = System::Windows::Forms::AutoScaleMode::Font;
501 this->AutoSize = true;
502 this->ClientSize = System::Drawing::Size(674, 508);
503 this->Controls->Add(this->buttonRestartScan);
504 this->Controls->Add(this->labelScanRateHz);
505 this->Controls->Add(this->trackBar1);
506 this->Controls->Add(this->progressBar1);
507 this->Controls->Add(this->label1);
508 this->Controls->Add(this->labelScanSize);
509 this->Controls->Add(this->buttonSaveData);
510 this->Controls->Add(this->buttonStopScan);
511 this->Controls->Add(this->buttonStartScan);
512 this->Controls->Add(this->richTextBox2);
513 this->Controls->Add(this->groupBox1);
514 this->Controls->Add(this->statusLabel);
515 this->Controls->Add(this->staticStatusLabel);
516 this->Controls->Add(this->richTextBox1);
517 this->Controls->Add(this->label2);
518 this->Controls->Add(this->textBox1);
519 this->Controls->Add(this->button1);
520 this->Name = L"Form1";
521 this->Text = L"DMSL CMM Laser Scanner Application";
522 this->Load += gcnew System::EventHandler(this,
&Form1::Form1_Load);
523 this->groupBox1->ResumeLayout(false);
524 this->groupBox1->PerformLayout();
525 (cli::safe_cast<System::ComponentModel::ISupportInitialize^
>(this->trackBar1))->EndInit();
526 this->ResumeLayout(false);
527 this->PerformLayout();
528
529 }
530
531 #pragma endregion
532
533 private: System::Void Form1_Load(System::Object^ sender,
System::EventArgs^ e) {
534
535 WSADATA wsaData;
536
537 progressBar1->Maximum = scanCount;
538
539 // Initialize Winsock
540 iResult = WSAStartup(MAKEWORD(2,2), &wsaData);
541 if (iResult != 0) {
542 printf("WSAStartup failed: %d\n", iResult);
543
544 statusLabel->Text = "WSAStartup failed";
545 //return 1;
546 }
547
548 struct addrinfo *result = NULL,
549 *ptr = NULL,
99
550 hints;
551
552 ZeroMemory( &hints, sizeof(hints) );
553 hints.ai_family = AF_INET;
554 hints.ai_socktype = SOCK_STREAM;
555 hints.ai_protocol = IPPROTO_TCP;
556
557
558
559 // Resolve the server address and port
560 iResult = getaddrinfo("192.168.0.35", "5010", &hints, &result);
561 if (iResult != 0) {
562 printf("getaddrinfo failed: %d\n", iResult);
563 statusLabel->Text = "getaddrinfo failed:" + iResult;
564 WSACleanup();
565 // return 1;
566 }
567
568 // Attempt to connect to the first address returned by
569 // the call to getaddrinfo
570 ptr=result;
571
572 // Create a SOCKET for connecting to server
573 ConnectSocket = socket(ptr->ai_family, ptr->ai_socktype,
574 ptr->ai_protocol);
575
576 //check socket is valid
577 if (ConnectSocket == INVALID_SOCKET) {
578 printf("Error at socket(): %ld\n", WSAGetLastError());
579 freeaddrinfo(result);
580 WSACleanup();
581
582 }
583
584 // Connect to server.
585 iResult = connect( ConnectSocket, ptr->ai_addr, (int)ptr-
>ai_addrlen);
588 ConnectSocket = INVALID_SOCKET;
589 }
590
591
592 freeaddrinfo(result);
593
594 if (ConnectSocket == INVALID_SOCKET) {
595 printf("Unable to connect to server!\n");
596 WSACleanup();
597
598 }else{
599 statusLabel->Text = "Connected to CMM Motion Control";
600 }
601
602
603 //begin counter capture card setup
100
604 /* Declare UL Revision Level */

605 ULStat = cbDeclareRevision(&RevLevel);
606
607 /* Initiate error handling
608 Parameters:
609 PRINTALL :all warnings and errors encountered will be printed
610 DONTSTOP :program will continue even if error occurs.
611 Note that STOPALL and STOPFATAL are only effective in
612 Windows applications, not Console applications.
613 */
614 cbErrHandling (PRINTALL, DONTSTOP);
615
616 /* set up the display screen */
617
618 /* set the configurable operations of the counter
619 Parameters:
620 BoardNum :the number used by CB.CFG to describe this board
621 CounterNum :the counter to be configured (0-5)
622 Quadrature :Select type of counter input
623 CountingMode :Slects how counter will operate
624 IndexMode :Selects what index signal will control
625 InvertIndex :Set to ENABLED id index signal is inverted
626 FlagPins :Select which signals will drive Flag pins
627 GateEnable :Set to ENABLED to use external gating signal */
628 CounterNum = 1;
629 //1 = Y, 2 = X, 3=Z
630 Quadrature = X4_QUAD;
631 CountingMode = NORMAL_MODE;
632 DataEncoding = BINARY_ENCODING;
633 IndexMode = INDEX_DISABLED;
634 InvertIndex = DISABLED;
635 FlagPins = CARRY_BORROW;
636 GateEnable = DISABLED;
637
638 //initialize
639
640 ULStat = cbC7266Config (BoardNum, CounterNum+1 , Quadrature,
CountingMode,
641 DataEncoding, IndexMode, InvertIndex, FlagPins,
642 GateEnable);
643 ULStat = cbC7266Config (BoardNum, CounterNum , Quadrature,
CountingMode,
645 GateEnable);
646 ULStat = cbC7266Config (BoardNum, CounterNum+2 , Quadrature,
CountingMode,
648 GateEnable);
649
650 /* send a starting value to the counter with cbCLoad()
651 Parameters:
652 BoardNum :the number used by CB.CFG to describe this
board
653 RegName :the counter to be loading with the starting
value
101
654 LoadValue :the starting value to place in the counter */

655 ////////////////////////////
656 int xLoadValue = 700000;
657 int yLoadValue = 0;
658 int zLoadValue = 0;
659
660 xlastCount = LoadValue;
661 RegName = COUNT1 + CounterNum - 1;
662 printf ("Loading counter #%u with an initial count of %u
using cbCLoad()\n", CounterNum,
LoadValue);
663 ULStat = cbCLoad32 (BoardNum, RegName, xLoadValue);
664 ULStat = cbCLoad32 (BoardNum, RegName+1, yLoadValue);
665 ULStat = cbCLoad32 (BoardNum, RegName+2, zLoadValue);
666
667 LoadValue = 4000000000;
668 RegName = PRESET1 + CounterNum - 1;
669 printf ("Setting counter #%u maximum count to %u by loading
PRESET register", CounterNum,
LoadValue);
670 ULStat = cbCLoad32 (BoardNum, RegName, LoadValue);
671 ULStat = cbCLoad32 (BoardNum, RegName+1, LoadValue);
672 ULStat = cbCLoad32 (BoardNum, RegName+2, LoadValue);
673 //end capture card setup
674
675 //Setup Connection to Scanner//
676
677 // initialize Go2 API
678 if ((status = Go2Api_Initialize()) != GO2_OK)
679 {
680 printf("Error: Go2Api_Initialize:%d\n", status);
681 return;
682 }
683 // construct Go2 system object
684 if ((status = Go2System_Construct(&scannerSystem)) != GO2_OK)
685 {
686 printf("Error: Go2System_Construct:%d\n", status);
687 return;
688 }
689 // convert ip address string to Go2IPAddress object
690 if ((status = Go2IPAddress_Parse((const Go2Char *)"192.168.0.45",
&address)) != GO2_OK)
691 {
692 printf("Error: Go2IPAddress_Parse:%d\n", status);
693 return;
694 }
695 // connect to Gocator system at default address as administrator
with empty password
696 if ((status = Go2System_Connect(scannerSystem,address,
GO2_USER_ADMIN, (const Go2Char *) "")) != GO2_OK)
697 {
698 printf("Error: Go2System_Connect:%d\n", status);
699 return;
700 }
701 // establish a Gocator data connection
102
702 if ((status = Go2System_ConnectData(scannerSystem, GO2_NULL,

GO2_NULL)) != GO2_OK)
703 {
704 printf("Error: Go2System_ConnectData:%d\n", status);
705 return;
706 }
707 // stop Gocator system
708 if ((status = Go2System_Stop(scannerSystem)) != GO2_OK)
709 {
710 printf ("Error: Go2System_Stop:%d\n", status);
711 }
712 // start Gocator system
713 if ((status == Go2System_Start(scannerSystem)) != GO2_OK)
714 {
715 printf("Error: Go2System_Start:%d\n", status);
716 return;
717 }
718 Sleep(1000); //added to ensure Scanner is finished System_Start
before calling
Go2System_ReceiveData,
719 //G02System_ReceiveData fails with the sleep removed.
720 //Allocate Memory
721 memory = (ProfilePoint **) malloc(scanCount * sizeof(ProfilePoint
*));
722 if(memory== NULL)
723 {
724 fprintf(stderr, "out of memory\n");
725 return;
726 }
727 for(i = 0; i < scanCount; i++)
728 {
729 memory[i] = (ProfilePoint *) malloc(1280 * sizeof(ProfilePoint));
730 if(memory[i] == NULL)
731 {
732 fprintf(stderr, "out of memory\n");
733 return;
734 }
735 }
736 //end Scanner setup
737
738 //Gui Initialization
739 labelScanRateHz->Text = System::Convert::ToString( trackBar1-
>Value) + " Hz";
740
741 }
742
743 public: System::Void button1_Click(System::Object^ sender,
744 //printf("Unable to connect to server!\n");
745
746 checkBox1->CheckState =
System::Windows::Forms::CheckState::Unchecked;
747 //add textboxText to
748 richTextBox2->Text = textBox1->Text + "\n" + richTextBox2->Text;
749
103
750 #define DEFAULT_BUFLEN 512

751
752 int recvbuflen = DEFAULT_BUFLEN;
753
754 char *sendbuf = (char*)(void*)
Marshal::StringToHGlobalAnsi(textBox1->Text + "\n");
755 char recvbuf[DEFAULT_BUFLEN];
756
757 // Send an initial buffer
758 iResult = send(ConnectSocket, sendbuf, (int) strlen(sendbuf), 0);
759
761 printf("send failed: %d\n", WSAGetLastError());
763 WSACleanup();
764 // return 1;
765 }
766
767 // Receive data
768 iResult = recv(ConnectSocket, recvbuf, recvbuflen, 0);
769 String^ clistr = gcnew String(recvbuf);
770 richTextBox1->Text = clistr + richTextBox1->Text + "\n";
771 if (iResult > 0){
772 printf("Bytes received: %d\n", iResult);
773 //label1->Text = L"Bytes received: " + iResult;
774 }
775 else if (iResult == 0)
776 printf("Connection closed\n");
777 else
778 printf("recv failed: %d\n", WSAGetLastError());
System::Windows::Forms::CheckState::Checked;
780 }
781
782 private: System::Void label3_Click(System::Object^ sender,
783 }
784
785 private: System::Void GrabScanProfile(){
786
787
788
789 if (Go2System_ReceiveData(scannerSystem, RECEIVE_TIMEOUT, &data)
== GO2_OK)
790 {
791
792 //each result can have multiple data items
793 for (j = 0; j < (signed int)Go2Data_ItemCount(data); ++j)
794 {
795 Go2Data dataItem = Go2Data_ItemAt(data, j); //dataItem seems
to be the single capture frame object
796 if (Go2Object_Type(dataItem) == GO2_TYPE_RAW_PROFILE_DATA)
797 {
798 ProfilePoint16s* profileData =
(ProfilePoint16s*)Go2RawProfileData_Ranges
104
(dataItem);
799 unsigned int profilePointCount =
Go2RawProfileData_Width(dataItem);
800 double XResolution = Go2RawProfileData_XResolution(dataItem);
801 double ZResolution = Go2RawProfileData_ZResolution(dataItem);
802 double XOffset = Go2RawProfileData_XOffset(dataItem);
803 double ZOffset = Go2RawProfileData_ZOffset(dataItem);
804
805 //translate 16-bit range data to engineering units and copy
profiles to memory array
806 for (arrayIndex = 0; arrayIndex < profilePointCount; ++arrayIndex)
807 {
808 if (profileData[arrayIndex].z != -INVALID_RANGE_16BIT)
809 {
810 //This section should be implemented using HTM of scanner
orientation
811 memory[scanIndex][arrayIndex].x = XScannerOffset;
812 memory[scanIndex][arrayIndex].y = YScannerOffset - (XOffset
+ XResolution * profileData[arrayIndex].x); //based on the scan stripe
aligned with the Y axis
813 memory[scanIndex][arrayIndex].z = ZScannerOffset + ZOffset +
ZResolution * profileData[arrayIndex].z;
814 }
815 else
816 {
817 memory[scanIndex][arrayIndex].x = INVALID_RANGE_DOUBLE;
818 memory[scanIndex][arrayIndex].y = INVALID_RANGE_DOUBLE;
819 memory[scanIndex][arrayIndex].z = INVALID_RANGE_DOUBLE;
820 }
821 }
822 }
823 }
824
825
826 if ((status = Go2Data_Destroy(data)) != GO2_OK)
827 {
828 printf("Error: Go2Data_Destroy:%d\n", status);
829 return;
830 }
831 } else {
832 printf("Error: Go2System_ReceiveData failed\n");
833 }
834
835
836 scanIndex++;
837
838 //endofSampleScan
839
840 }
841
842 private: System::Void GrabCount(){
843
844 //START COUNTER CAPTURE
845
846 ULStat = cbCIn32 (BoardNum, CounterNum, &xCount);
105
847 ULStat = cbCIn32 (BoardNum, CounterNum+1, &yCount);

848 ULStat = cbCIn32 (BoardNum, CounterNum+2, &zCount);
849
850 /* Parameters:
851 BoardNum :the number used by CB.CFG to describe this board
852 CounterNum :the counter to be setup
853 StatusBits :counter's status returned here */
854 //unused
855 ULStat = cbCStatus (BoardNum, CounterNum, &StatusBits);
856 if (StatusBits & C_UP_DOWN)
857 DirectionStr = "UP ";
858 else
859 DirectionStr = "DOWN";
860
861 //Write X, Y, Z Labels
862 if (xCount != xlastCount){
863 }
864 xlastCount = xCount;
865
866 if (yCount != ylastCount){
867 }
868 ylastCount = yCount;
869
870 if (zCount != zlastCount){
871 }
872 zlastCount = zCount;
873 //End of Counter capture
874
875 }
876
877
878 private: System::Void timer1_Tick(System::Object^ sender,
879 if (timerScan->Enabled == true){
System::Windows::Forms::CheckState::Unchecked;
881 }else{
882 GrabCount();
883 }
884 XScannerOffset = (double)xCount / 1000;
885 YScannerOffset = YMax - ((double)(yCount) / 1000) ;
886 ZScannerOffset = ZMax - ((double)(zCount) / 1000) ;
887 xlabel->Text = System::Convert::ToString((double)XScannerOffset);
888 ylabel->Text = System::Convert::ToString(YScannerOffset);
889 zlabel->Text = System::Convert::ToString(ZScannerOffset);
890 }
891 private: System::Void checkBox1_CheckedChanged(System::Object^
sender, System::EventArgs^ e) {
892 if (checkBox1->CheckState ==
System::Windows::Forms::CheckState::Checked){
893 //do something
894 timer1->Enabled = true;
895 }else{
896 //do seomthing else
897 timer1->Enabled = false;
106
898 }
899 }
900 private: System::Void button2_Click(System::Object^ sender,
901 // timer1->Enabled = true;
902 LoadValue = 100000;
903 //load value in micrometers
904 int xLoadValue = 700000;
905 //y and z load values declared globally
906 int yLoadValue = 0;
907 int zLoadValue = 0;
908 RegName = COUNT1 + CounterNum - 1;
909 printf ("Loading counter #%u with an initial count of %u
using cbCLoad()\n", CounterNum, LoadValue);
910 ULStat = cbCLoad32 (BoardNum, RegName, xLoadValue);
911 ULStat = cbCLoad32 (BoardNum, RegName+1, yLoadValue);
912 ULStat = cbCLoad32 (BoardNum, RegName+2, zLoadValue);
913 // testCount = 0;
914 // button2->Text = System::Convert::ToString(testCount);
915 }
916
917 private: System::Void textBox1_TextChanged(System::Object^ sender,
918
919 }
920
921 public: System::Void buttonStartScan_Click(System::Object^ sender,
922 buttonStartScan->Enabled = false;
923 buttonStopScan->Enabled = true;
924 timerScan->Enabled = true;
925
926 }
927 public: System::Void buttonStopScan_Click(System::Object^ sender,
928 buttonStopScan->Enabled = false;
929 buttonStartScan->Enabled = true;
930 timerScan->Enabled = false;
931 }
932
933 private: System::Void timerScan_Tick(System::Object^ sender,
934
935 if (scanIndex < scanCount ){
936
937 progressBar1->Value = scanIndex;
938 GrabCount();
939 //update offset
940 XScannerOffset = (double)xCount / 1000;
941 YScannerOffset = YMax - ((double)(yCount) / 1000) ;
942 ZScannerOffset = ZMax - ((double)(zCount) / 1000) ;
943 //
944 Go2System_Trigger(scannerSystem);
945 //
946 GrabScanProfile();
107
947 } else{
948 //memory buffer full
949
950 }
951 labelScanSize->Text = System::Convert::ToString((unsigned
int)scanIndex);
952 if (trackBar1->Value == 0){
953 buttonStopScan->Enabled = false;
954 buttonStartScan->Enabled = true;
955 timerScan->Enabled = false;
956
957 };
958
959 }
960 private: System::Void buttonSaveData_Click(System::Object^ sender,
961
962 //save data to file datafile.txt
963 ofstream myfile;
964
965 myfile.open ("ScanData.txt", ios::out | ios::app);
966
967 for (i=0;i < scanIndex ;i++){
968 for (j = 0; j < 1280;j++){
969 if (memory[i][j].x != INVALID_RANGE_DOUBLE)
970 myfile << memory[i][j].x << "," << memory[i][j].y << "," <<
memory[i][j].z << "\n";
971 }
972 }
973 myfile.close();
974
975 }
976
977 private: System::Void trackBar1_Scroll(System::Object^ sender,
978 labelScanRateHz->Text = System::Convert::ToString( trackBar1-
>Value) + " Hz";
979 if (trackBar1->Value != 0){
980 timerScan->Interval = int(1000/trackBar1->Value);
981 }else{
982 timerScan->Interval = int(5);
983 }
984
985 }
986 private: System::Void buttonRestartScan_Click(System::Object^
sender, System::EventArgs^ e) {
987 scanIndex = 0;
988 }
989 };
990 }
991
992
108
References
[1] James V. Valentino and Joseph Goldenberg, Introduction to Computer
Numerical Control (CNC), 5th ed.: Pearson Education Inc., 2013.
[2] Y. Altintas, I. Yellowley, and J. Tlusty, "The Detection of Tool Breakage
in Milling Operations," Trans. ASME, J. Eng. Ind., vol. 110, no. 3, pp.
271-277, August 1988.
[3] R. V. Fleisig and A. D. Spence, "Integrated Digitizing, Path Planning and
Five-Axis Machining for the Refurbishment of Compressor and Turbine
Blades," ASME Winter Annual Meeting, Manufacturing Science and
Technology, vol. 6, no. 1, pp. 31-38, 1997.
[4] Robert J. Hocken and Paulo H. Pereira, Coordinate Measuring Machines
and Systems, 2nd ed. Boca Raton, FL, USA: Taylor & Francis, 2012.
[5] Origin International Inc. (2012) CheckMate, SoftFit Solver: The What,
Why & How. [Online].
http://info.originintl.com/Portals/37807/docs/SoftFitSolver_d2.pdf
[6] Yu Pin Lin, "Geometrically Adaptive Milling of Fan Blade Assembly Weld
Fillets," Department of Mechanical Engineering, McMaster University,
Hamilton, Thesis Paper 7410, 2012.
[7] E. J. Jr Maksym, "Automated Inspection for Flexible Machining
Systems," Manufacturing Technology Information Analysis Center,
109
Chicago, SOAR MTIAC SOAR-88-01, 1988.
[8] G. L Franck, "Flexible inspection systems for automated manufacturing,"
in Test, Meas. & Inspect for Quality Control Conference/Exhibition,
1987.
[9] K. L. Blaedel, "Error Reduction," University of CA, Livermore, CA, Report
of the Machine Tool Task Force, Volume 5: Machine Tool Accuracy
UCRL-52960-S, 1980.
[10] ISO, "Geometrical Product Specifications (GPS) -- Acceptance and
reverification tests for coordinate measuring machines (CMM) -- Part 2:
CMMs used for measuring size," ISO 10360-2:2001 , 2001.
[11] ISO, "Test code for machine tools -- Part 2: Determination of accuracy and
repeatability of positioning numerically controlled axes," ISO 230-
2:2006, 2006.
[12] ISO, "Test code for machine tools -- Part 10: Determination of the
measuring performance of probing systems of numerically controlled
machine tools," ISO 230-10:2011, 2011.
[13] J. Bryan, "Design and Construction of an 84 inch diameter diamon
turning machine," Precision Engineering, vol. 1, no. 13, pp. 13-17, 1979.
[14] W. T. Estler and E. B. Magrab, "Validation metrology of the large optics
diamond turning machine," National Bureau of Standards Interagency
110
Report NBSIR-85-3182R, 1985.
[15] S. R. 1986 Patterson, "Development of Precision Turning Capabilities at
Lawrence Livermore National Laboratory," in 3rd Biennial
International Machine Tool Technical Conference, Chicago, IL, 1986,
pp. 147-59.
[16] K. Kim, K. Eman, and S. M. Wu, "In-process control of cylindricity in
boring operations," ASME Trans. Journal of Engineering for Industry,
vol. 109, no. 4, pp. 385-391, 1987.
[17] M. A. Donmez, "A real-time control system for CNC machine tool based on
deterministic metrology.," in Statistical Process Control in Automated
Manufacturing, J. B. Keats and N. F. Hubels, Eds. New York: Marcel
Dekker, Inc., 1989.
[18] J. Bryan, "The deterministic approach in metrology and manufacturing,"
in Proceedings of the 1993 International Forum on Dimensional
Tolerancing and Metrology, New York, 1993, pp. 85-95.
[19] G et al. Belforte, "Coordinate measuring machines and machine tools self-
calibration and error correction," Annals of CIRP, vol. 36, no. 1, pp.
359-364, 1987.
[20] R. Hocken, J. Raja, and U. Babu, "Sampling Issues in Coordinate
Metrology," Manufacturing Review, vol. 6, p. 282, 1993.
111
[21] "International vocabulary of metrology – Basic and general concepts and
associated terms (VIM), 3rd Edition," Joint Committee for Guides in
Metrology, JCGM 200:2012, 2012.
[22] J.A. Bosch, Coordinate Measuring Machines and Systems, G., Dieter, G.
Boothroyd, Ed.: Marcel Dekker Inc., 1995.
[23] Renishaw Plc. (2002, December) SP600 Scanning Probe System. [Online].
http://www.renishaw.com/en/sp600--6687
[24] Renishaw Plc. (2012, August) Touch-trigger Probes. [Online].
http://www.renishaw.com/en/touch-trigger-probes--6652
[25] Fanuc Ltd., "Fanuc Series 15i-MA Operator's Manual, B-63324EN/01,"
Manual 1998.
[26] R. Szelisiki,.: Springer, 2010, ch. 7, pp. 345-347.
[27] Wikipedia. (2008, May) Structured Light 3D Scanner. [Online].
http://en.wikipedia.org/wiki/Structured-light_3D_scanner
[28] Altschuler et al., "Robot Vision by Encoded Light Beams," in Three-
Dimensional Machine Vision.: Kluwer Academic Publishers, 1987, ch.
1, p. 99.
[29] ShapeGrabber. (2012, November) Shapegrabber company Web site.
[Online]. http://blog.shapegrabber.com/2012/05/optical-3d-inspection-of-
transparent-or-reflective-parts.html
112
[30] LMI 3D. (2012, August) Line Triangulation. [Online].
http://www.lmi3d.com/education/line-triangulation
[31] G. T. Anthony and M. G. Cox, "Reliable Algorithms for Roundness
Assessment to BS3730," in Software for Coordinate Measuring
Machines, M. G. Cox and G. N. Peggs, Eds. Teddington, UK: National
Physical Laboratory, 1986, pp. 30-37.
[32] K.S. Arun, T.S. Huang, and S.D. Blostein, ""Least-Squares Fitting of Two
3-D Point Sets," IEEE Trans. PAMI, vol. 9, no. 5, pp. 698-700, 1987.
[33] P.J. Besl and H.D. McKay, "A Method for Registration of 3-D Shapes,"
IEEE TRANS. PAMI, vol. 14, no. 2, pp. 239-256, 1992.
[34] Geomagic, Inc. (2013, February) Geomagic Studio. [Online].
http://www.geomagic.com/en/products/studio/overview/
[35] InnovMetric Software Inc. (2013, March) PolyWorks Solutions - Point
Cloud Engineering. [Online].
http://www.innovmetric.com/polyworks/3D-
scanners/so_pointcloud3.aspx?lang=en
[36] Origin International Inc. (2012, Nov.) Origin International, Inc. [Online].
http://info.originintl.com/Portals/37807/docs/ISO.pdf
[37] Radu Bogdan Rusu and Steve Cousins, "3D is here: Point Cloud Library
(PCL)," in IEEE International Conference on Robotics and Automation
113
(ICRA), Shanghai, China, 2011.
[38] Memex Automation Inc. a unit of Astrix Networks Inc. (2012, August)
Ax9150 Universal Machine Interface (UMI). [Online].
http://memex.ca/index.php?option=com_content&view=article&id=46&I
temid=200145
[39] P. Smid, Fanuc CNC custom macros: Programming resources for Fanuc
Custom Macro B users, Illustrated ed.: Industrial Press Inc., 2004.
[40] Renishaw PLC. (2011) (TE412) One-Touch vs two-touch probing
strategies. [Online].
http://resources.renishaw.com/download.aspx?data=32423〈=&showFor
m=true
[41] Vero Software Limited. (2013) EdgeCam. [Online]. www.edgecam.com
[42] Roland DGA Corporation. (2012, September) LPX-60DS/600DS/1200DS
3D Laser Scanners Specifications. [Online].
http://www.rolanddga.com/products/scanners/lpx600/#specifications
[43] C. Y. Wu, "Arbitrary Surface Flank Milling and Flank SAM in the Design
and Manufactoruing of Jet Engine and Compressor Coils," in ASME
Turbo Expo 2012, Copenhagen, 2012, p. 2.
[44] LMI Technologies. (2012, May) LMI Technologies. [Online].
www.lmi3D.com
114
[45] Renishaw Plc. (2012, August ) Inspection Plus - software for machining
centres. [Online]. http://www.renishaw.com/en/inspection-plus-
software-for-machining-centres--6094
[46] Rolls-Royce. (2003, April) Rolls-Royce Ships First Blisk for Joint Strike
Fighter SDD Programme. [Online]. http://www.f-16.net/f-
16_forum_viewtopic-t-9880-start-30.html
[47] D. Alan Sawula and Allan D. Spence, "Integrated CNC Measurement,
Analysis and Tool Path Adjustment," in Virtual Machining Process
Technology, vol. 1, Montreal, 2012.
[48] Allan D. Spence, D. Alan Sawula, James R. Stone, and Yu Pin Lin, "In-
situ Measurement and Distributed Computing for Adjustable CNC
Machining," in Computer-Aided Design and Applications, Bergamo,
Italy, 2013.
115

Full Text

Uploaded by

Full Text

Uploaded by

INTEGRATION OF MACHINING INSPECTION SENSORS AND

SENSORS AND SOFTWARE

D. ALAN SAWULA, B. ENG

© Copyright by D. Alan Sawula, April 2013

MASTER OF APPLIED SCIENCE (2013)

TITLE: Integration of Machining Inspection Sensors and

AUTHOR: D. Alan Sawula, B. Eng (McMaster University)

NUMBER OF PAGES: xii, 108

Computer-Aided Design and Manufacturing (CAD/CAM) software, can be

consistently fixtured and machined. In reality, process conditions vary, and

feedback and correction methods such as integrated on-machine inspection,

analysis, and process adjustment, are required.

On-machine inspection based on touch trigger probes is well

established, but limited motion control computing capability restricts analysis

to simple arithmetic. This prevents on-line use of known whole part

mathematical analysis software that implements the part salvaging

intentions of modern Geometric Dimensioning and Tolerancing (GD&T)

standards. Additionally, no CNC integrated method exists for geometrically

adjusting nominal tool paths so that an in-tolerance final part is produced.

This thesis reports progress towards bidirectional integration of

subsequent toolpath adjustment. The concepts are demonstrated using a

fixture consisting of three datum spheres and a workpiece. The fixture is

mathematical data fitting and registration, an in-tolerance final part is

pose. Preliminary integration of a laser scanner with axis scales and

computer software was also accomplished.

I would like to thank my supervisor Dr. Spence, for his patience,

and life lessons were learned. I am grateful for his dedication.

laser scanning equipment and software.

Funding support from NSERC Canadian Network for Research and

Innovation in Machining Technology (CANRIMT) is sincerely appreciated.

To my family and friends for all their unconditional support I am very

Chapter 1 Introduction ........................................................................... 1

Chapter 2 Literature Review .................................................................. 6

2.1 Measurement in the Manufacturing Process .............................. 6

2.1.1 Static Sensor Specifications .................................................... 9

2.1.2 Touch Trigger Probes ............................................................ 11

2.1.2.1 Kinematics of Touch Trigger Probes .................................. 13

2.1.2.2 Relevant CNC G-Codes for Touch Trigger Probes ............ 15

2.1.2.3 G01 Linear Motion Command ............................................ 15

2.1.2.4 G31 Skip Command ............................................................ 16

2.1.3 Laser Digitizers...................................................................... 16

2.2 Data Fitting ................................................................................. 21

2.2.1 Rigid Coordinate System Registration ................................. 21

2.2.2 Orthogonal Least Squares Geometric Fitting ...................... 23

2.2.3 Orthogonal Least Squares Fitting on Data Sets .................. 26

2.2.4 Commercial Point Cloud Software ........................................ 28

2.3 Summary ..................................................................................... 29

Chapter 3 Hardware Description ......................................................... 30

3.1 CNC Machine Tool ...................................................................... 30

3.2 Touch Trigger Probe Integration ................................................ 30

3.3 Hardware Overview .................................................................... 32

3.4 Ax9150 UMI ................................................................................ 33

3.4.1 Installation ............................................................................. 34

3.4.2 Passing coordinate data to PC .............................................. 34

3.5 Calibration of Touch Trigger Probe ............................................ 35

3.5.1 Calibration Tool Path ............................................................ 36

3.5.2 Results of Calibration Data ................................................... 38

3.6 Three Sphere Probing Routine ................................................... 39

Chapter 4 Registration Experiment ..................................................... 42

4.1 Experiment Tool Path ................................................................. 42

4.2 Tool Path Adjustment ................................................................. 43

4.2.1 Part Registration in Manufacturing Processes .................... 43

4.2.2 Tooling Sphere Registration Method .................................... 43

4.3 Test Cases .................................................................................... 45

Chapter 5 Registration Experiment Results ....................................... 49

5.1 Laser Scan of Finished Workpiece ............................................. 49

5.2 Analysis of Expected Errors ....................................................... 59

Chapter 6 Laser Scanner Integration .................................................. 61

6.1 LMI Scan Head Integration ........................................................ 63

6.2 Mount Design .............................................................................. 64

6.3 Desktop Control Software ........................................................... 66

6.4 Expected Errors ........................................................................... 71

6.5 Test Data ..................................................................................... 72

Chapter 7 Conclusions and Future Work ............................................ 78

7.1 Future Work ................................................................................ 79

Figures and Tables.