Design Characterization and Machining Force Model of A Novel Orbital Micromachining Technology Based On Single Point Tool Tip Geometry

DESIGN, CHARACTERIZATION AND MACHINING FORCE MODEL
OF A NOVEL ORBITAL MICROMACHINING TECHNOLOGY

BASED ON SINGLE-POINT TOOL TIP GEOMETRY
by
Sumet Heamawatanachai
A dissertation submitted to the faculty of

The University of Utah
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
Department of Mechanical Engineering

The University of Utah
August 2010
Copyright Sumet Heamawatanachai 2010

All Rights Reserved
The University of Utah Graduate School
STATEMENT OF DISSERTATION APPROVAL
The dissertation of
Sumet Heamawatanachai
has been approved by the following supervisory committee members:
, Chair
4/28/2010
Bruce Gale
, Member
4/28/2010
Rebecca Brannon
, Member
Stephen Mascaro
, Member
Florian Solzbacher
, Member
Eberhard Bamberg
and by
T
A me el
"'
"'
...: =
:::.""
"'nn
"'
___________
the Department of
bate Approved
__________
Mechanical Engineering
and by Charles A. Wight, Dean of The Graduate School.
Date Approved
5/412010
Date Approved
5/2/2010
Date Approved
5/5/2010
Date Approved
'
Chair of
ABSTRACT
This research presents a new mechanical micromachining technique that can be

used to machine 2D or 3D microfeatures into many types of materials such as metals,
semiconductor materials and glass. The microscopic cutting motion of the tool is based
on orbital motion of a single-point tool tip that is made of a single crystal diamond. To
generate the microscopic cutting motion, the tool tip is attached to the free end of a piezo
tube while the other end is held stationary. Sinusoidal voltage signals are applied to the
four electrodes (x+, x-, y+, y-) of the piezo tube to create cyclic bending that results in
orbital motion of the attached tool tip. This allows various types of trajectories of the tool
tip to be created that include circular, elliptical and tapered motion. Microfeatures can be
created by combining the piezo tool, which produces the microscopic cutting motion,
with a macroscopic motion module (any 3-axis stage such as a CNC machine tool),
which is used to generate the feed motion. This research details the design and control of
the tool and also reports on its dynamic behavior.
A second aspect of this research is the development of a cutting force model
capable of predicting the forces that occur between the tool tip and the work piece during
orbital cutting. The force model is based on the determination of the contact area of a
conical tool tip with a tip radius that is typically much larger than the uncut chip
thickness. Experiments with aluminum alloy (AL 2024) were performed to obtain the
specific cutting forces as well as the friction coefficients at these very small length scales.
The model was validated by varying critical machining parameters such as depth of cut,
orbital frequency as well as feed per orbit.
Machining of brittle materials was performed on silicon, which is a very
commonly used material in optical instruments as well as semiconductor-based products.
Due to its high brittleness and low fracture toughness, it is rather difficult to machine
with mechanical techniques. Owing to its tool radius, which is much larger than the uncut
chip thickness, orbital micromachining produces relatively high hydrostatic pressure
during cutting. This allows brittle materials to be cut in the ductile regime whereby
material is removed without fracture, thereby producing very smooth surfaces. In reality,
there are many machining parameters (orbital trajectory, tool geometry, etc.) that affect
the machining result on silicon. This research presents experimental studies on single
crystal silicon work pieces that demonstrated the feasibility of ductile regime machining
for depths of cut less than 1 micron.
iv
TABLE OF CONTENTS
ABSTRACT...iii
LIST OF FIGURES.viii
LIST OF TABLES...xiii
NOMENCLATURE.xiv
ACKNOWLEDGEMENTS.xvii
Chapter
1. INTRODUCTION .......1
1.1 Introduction.1
1.2 Significance of the research ....3
1.3 Existing micromachining techniques ......4
1.3.1 Electrical discharge machining (EDM)..5
1.3.2 Electrochemical machining (ECM)6
1.3.3 Photolithography ....8
1.3.4 Micromilling ..9
1.3.5 Single point diamond machining..10
1.3.6 Vibration assisted machining .......11
1.4 Overview of orbital motion for micromachining...12
1.5 Design requirements..13
1.6 Hypotheses.14
2. DESIGN, MANUFACTURING AND CONTROL..16
2.1 Introduction ......16
2.2 Cutting module......18
2.3 System components and control of the orbital micromachining tool24
2.3.1 Microscopic motion control..26
2.3.1.1 The first version control system...27
2.3.1.2 The second version control system..31
2.3.2 Macroscopic motion control.36
2.4 Conclusions36
3. TOOL CHARACTERIZATION............................38
3.1 Piezo tube modeling. ....38
3.1.1 Static deflections...39
3.1.2 Orbital paths..40
3.1.3 Calculation of resonance frequency..45
3.2 Experimental results...53
3.2.1 Measured resonance frequency 53
3.2.2 Static performance of the piezo tube....53
3.2.3 Hysteresis..55
3.2.4 Long term stability57
3.2.5 Frequency response...58
3.3 Conclusions59
4. CUTTING FORCE MODEL..................61
4.1 Introduction...61
4.2 Cutting force models..61
4.3 Orbital single-point cutting force model65
4.3.1 Principle of orbital cutting motion66
4.3.2 Calculation of the tool tip geometry.67
4.3.3 Calculation of the machining contact area68
4.3.4 Calculation of cutting forces.72
4.4 Model verification..75
4.4.1 Experimental setup75
4.4.2 Measurement data processing...77
4.4.3 Calculation result..78
4.5 Conclusions83
5. ORBITAL MICROMACHINING OF DUCTILE MATERIALS..............86
5.1 Machining examples on ductile materials..86
5.2 Machining of microchannels into ductile material....89
5.3 Investigation of tool wear..97
5.4 Conclusions..100
6. ORBITAL MICROMACHINING OF BRITTLE MATERIALS.102
6.1 Introduction......103
6.2 Influence of hydrostatic pressure on the ductile behavior of brittle materials.105
6.3 Examples of machining 2D/3D microfeatures on brittle materials......107
6.4 The effect of orbital motion.110
6.5 The effect of lubrication...111
6.6 The effect of depth of cut.112
vi
6.7 Tool trajectories during machining..114

6.8 Improving surface finish with a finishing cut..117
6.9 Machining forces, surface quality and tool wear.117
6.9.1 Measuring of the cutting force121
6.9.2 Machined surface quality124
6.9.3 Tool wear126
6.10 Conclusions127
7. CONCLUSIONS AND FUTURE WORK.129
7.1 Conclusions .........129
7.2 Future work......132
REFERENCES134
vii
LIST OF FIGURES
Figure
Page
1.1: Concept of EDM...6

1.2: Electrochemical removal process ........7
1.3: Basic photolithography and pattern transfer.8
1.4: Principle of milling process .....9
1.5: Microgrooves..10
1.6: 2D Vibration assisted machining .......11
1.7: Orbital micromachining principle...14
2.1: The ceramic piezo tube scanner with four electrodes on the outside surface and the
center ground on the inside surface...19
2.2: The piezo tube is actuated by analog voltages that vary from 0 to -1000 V .20
2.3: The cyclic bending of the piezo tube generates a planar, orbital motion of the tool tip
....20
2.4: The cutting mechanism in 3D (a), in more detail (b), as a side view (c),
and as a top view (d)..21
2.5: The cutting module prototype consists of the piezo tube, a housing to protect the
tube, a clamp to hold the tool shank, and noncontact probes to monitor the cutting
motion....21
2.6: A clamp holding the tool shank is mounted to the free end of the piezo tube22
2.7: The cutting tool is a stylus with a 60 diamond cone with a tip radius of 10 m..22
2.8: Assembly of the tool. Electrical isolation of the piezo tube is achieved by bonding it
to ceramic insulators with high stiffness epoxy ....23
2.9: The piezo tube is attached to the housing by bonding with epoxy to the base
insulator..24
2.10: System setup #1, the microscopic motion module was mounted to the macroscopic
motion module (Micromachine tool, Optimation 24P)......25
2.11: System setup #2, the microscopic motion module was mounted to the CNC
machine (Haas TM-1).25
2.12: System setup #3, the microscopic motion module was mounted to the CNC
machine (Haas VF-E)..26
2.13: Close view of the cutting module during machining27
2.14: First version control diagram of the cutting module28
2.15: LabVIEW VI showing the output voltage from the signal generator and the
resulting motion of the tool clamp.29
2.16: Example of tool trajectories..30
2.17: Second version control diagram of the cutting module31
2.18: Signal generator control program.33
2.19: Setting parameters to calibrate and match x+ axis to the y- axis.34
2.20: Actual tool trajectories from parameters of Fig. 2.19..35
2.21: Actual tool trajectories at 100 Hz after calibrating, (a) 1V amplitude and 90o phase,
(b) 2V amplitude and 90o phase, (c) 3V amplitude and 90o phase, (d) 2V amplitude and 0o
phase, (e) 2V amplitude and 60o phase, (f) 2V amplitude and 120o phase35
3.1: Circular orbital tool path.....41
3.2: Elliptical tool path....42
3.3: Elliptical tool path with 30o phase shift in x axis ..42
3.4: Circular orbital motion with height compensation ....43
3.5: Tilted circular orbital motion .....44
3.6: Saddle-like orbital tool path. ..45
3.7: The piezo tube is modeled with a lumped mass attached to the free end...46
ix
3.8: The measured resonance frequency for excitation of the x-electrodes (a) and yelectrodes (b)..54
3.9: Response of a piezo tube to a 300 V step (3V control signal) to X-axis....55
3.10: Hysteresis of the piezo tube tool from applying voltage to the X-axis....56
3.11: Hysteresis of the piezo tube tool from applying voltage to the Y-axis.....56
3.12: Long term stability: The orbit radii for various excitation voltages (a) and the drift
of the orbit center (b).57
3.13: Frequency response of x (a) and y electrodes (b).58
4.1: Overview of cutting models: (a) orthogonal cutting model (adapted from Merchant
[43]), (b) orthogonal cutting model for large tool radii and/or small chip thickness
(adapted from Son et al. [45]), (c) grinding force model (adapted from Shaw [49]) and
(d) single-point, orbital cutting model (Heamawatanachai and Bamberg, in
press)..................................................................................65
4.2: Orbital machining of slot with (a) spherical contact area, (b) spherical and conical
contact area, (c) zoom view of contact area with tool tip, and (d) force components at the
contact area66
4.3: Cutting motion of (a) conventional, rotating tools and (b) orbiting tools..67
4.4: Front view of tool tip geometry..68
4.5: Cross section i of tool tip at height zi during orbital machining.69
4.6: Directional area components of the contact area74
4.7: Calculated cutting force components based on KN and Kf equations.75
4.8: The micro-orbital machining setup.76
4.9: Measurement data for orbital cut (100 Hz, 5.5 m depth, 10 m orbit radius, 10 m/s
feed rate) showing measured forces in (a) x , (b) y, (c) z direction, (d) tool position in x
and (e) in y. Also shown are the filtered data for the forces and tool position (f) through
(j)77
4.10: Range of measured forces vs. depth of cut (100 Hz, 10 m orbit radius, 0.1 m feed
per orbit).78
4.11: Experimental data and curve fits for the specific cutting forces KN and Kf.80
4.12: Friction coefficient in orbital machining..81

4.13: Comparison between the predicted and measured cutting forces82
4.14: Cutting force predictions based on (a) depth of cut, (b) tip radius, (c) orbit radius
and (d) feed per orbit.84
5.1: Pyramid (100 x 100 m) in copper.87
5.2: Pyramid (100 x 100 m) in titanium alloy (Ti-6Al-4V).87
5.3: Pyramid (100 x 100 m) in stainless steel (316)88
5.4: Various microfeatures machined in 316 stainless steel..88
5.5: Microslot machined in Al2024 without lubrication or flushing. The picture was taken
after machining without any cleaning91
5.6: Microslot on Al2024...93
5.7: Micro-taper-groove on Al 2024 with a depth of 70 m.....94
5.8: Micro-taper-groove with smooth side walls (left and right) on Al 202495
5.9: Microchannels combine with 6 features and 3 different depths (30, 50, and 70 m) in
Al 2024..96
5.10: Microfeature in Ti-6Al-4V...97
5.11: Comparison of tool tip before and after machining of 35e6 m3 of Al 2024..99
5.12: Comparison of tool tip before and after machining of 59.7e3 m3 on Ti-6Al-4V.100
6.1: Relationship of yielding and fracture boundary for brittle material 107
6.2: 3D view of a microfluidic feature machined into glass taken with an optical surface
profiler (Zygo NewView). The surface roughness was measured as 0.184 m...108
6.3 Pyramid feature in single-crystal silicon ...109

6.4: Pyramid feature in single-crystal germanium ..109
6.5: The effect of orbital motion on the machining quality 110
6.6: Slot machined (a) without and (b) with lubrication (BOELUBE). ..111
xi
6.7: Slot machined into single-crystal silicon with (a) 1 m, (b) 3 m, (c) 5 m and (d) 10
m depth of cut........113
6.8: Slots with (a) 1, (b) 2, (c) 5, (d) 10 and (e) 15 m depth of cut machined into singlecrystal silicon at 300Hz with 10 m orbital radius..115
6.9: Optical profile of slot machined into single-crystal silicon with a depth of cut of 2
m at 300 Hz, an orbital radius of 10 m and a feed rate of 0.1 mm/min..116
6.10: Orbital trajectories recorded at depth of cuts of (a) 0, (b) 1, (c) 2, (d) 5,(e) 10 and (f)
15 m during (g) milling of a slot into single-crystal silicon..116
6.11: Slot machined in silicon in two steps: first pass with 2 m depth of cut followed by
a second pass with 0.23 m depth of cut118
6.12: The micro-orbital machining setup.119
6.13: Measurement data for orbital cut of the experiment #1 (100 Hz, 4.4 m depth, 10
m orbit radius, 1 m/s feed rate) showing measured forces (filtered) in (a) x , (b) y, (c) z
direction...121
6.14: The effect of tip radius on the machining forces at (a) 100 Hz and (b) 300 Hz orbital
frequency..123
6.15: The effect of feed speed to the machining forces...124
6.16: SEM images of machined surfaces with (a) 1.6 m, (b) 0.9 m and (c, d) 0.1 m
depth of cut..125
6.17: The tool tip before and after machining..126
xii
LIST OF TABLES
Table
Page
3.1: The properties of the piezo tube with attached mass .52
3.2: Resonance frequencies of the piezo tube with attached mass....52
5.1: Data for elements with known diamond turning properties ..98
5.2: Components (% by weight) of aluminum alloy (Al 2024-T4)...98
5.3: Components (% by weight) of titanium alloy (Ti-6Al-4V)99
6.1: Machining parameters for experiments120
6.2: Range of forces and standard deviation of the measured machining forces in x, y and
z directions during machining..122
NOMENCLATURE
Chapter 3:
d 31
The piezoelectric strain coefficient for displacements normal to the

polarization direction [ ]
The length of the tube [m]
The change in length of the tube [m]
The voltage applied to all electrodes [V]
The wall thickness [m]
Lateral displacement in x direction [m]
Lateral displacement in y direction [m]
ID
The inner diameter of the tube [m]
Ux
The symmetric electrode voltages in x axis [V]
Uy
The symmetric electrode voltages in y axis [V]
Youngs modulus of the beam [N/m2]
The area moment of inertia of the cross section of the beam [m4]
Shear coefficient of the cross-section [ ]
The cross-sectional area of the beam [m2]
Shear modulus of elasticity [N/m2]
Slope of the beam due to bending [ ]
The deflection of the center line of the beam [m]
The density of the beam [kg/m3]
Poissons ratio [ ]
ri
The inner radius of the tube [m]
ro
The outer radius of the tube [m]
The moment of inertia of the attached mass around its center of

mass [kg.m2]
The attached mass [kg]
Non-dimensional length along the beam [ ]
Frequency [rad/s]
Chapter 4:
FN
Normal force [N]
Ff
Friction force [N]
KN
Specific cutting force in the normal direction [N/m2]
Kf
Specific cutting force in the tangential direction [N/m2]
[ X T , YT , Z T ]
Tool tip coordinate frame [m,m,m]
[ X C , YC , Z C ]
Center coordinate frame [m,m,m]
[ X T , YT , Z T ]
Tool tip coordinate frame [m,m,m]
Orbit radius [m]
Angular speed [rad/s]
Vf.
Feed speed [m/s]
rt
Tool tip radius [m]

xv
ri
Cross section radius of the ith disk [m]
Cone angle [degrees]
Orbit angle [degrees]
t0
Depth of cut [m]
Period [s]
The tool engagement angle [rad]
The tool engagement length [m]
v
Ni
The normal vector to the contact surface of the ith disk
v
FNi
The normal force vector of the ith disk
v
F fi
The friction force vector of the ith disk
ANx
The directional area component for the calculation of the normal

force in x direction [m2]
ANy

force in y direction [m2]
ANz

force in z direction [m2]
Afx
The directional area component for the calculation of the frictional

force in x direction [m2]
Afy
The directional area component for the calculation of the frictional

force in y direction [m2]
Fx
Cutting force in x direction [N]
Fy
Cutting force in y direction [N]
Fz
Cutting force in z direction [N]
xvi
ACKNOWLEDGMENTS
Special thanks to Dr. Eberhard Bamberg (my advisor) for all the help and support
that he gave me during this research work. I would also like to thank my committee
members: Dr. Bruce Gale, Dr. Rebecca Brannon, Dr. Stephen Mascaro and Dr. Florian
Solzbacher, for their support. Thanks to Dr. Dinesh Rakwal (my friend) who helped me in
taking SEM pictures. I would like to thank the National Institute of Metrology
(Thailand), the Thai people and the Thai government who gave me a scholarship to
perform this research. I would also like to thank all my teachers who taught me so that I
have good background knowledge in engineering. Thanks to Wannee Heamawatanachai
(my mother), Pichai Heamawatanachai (my father), Sakara Tunsophon (my wife), Primar
Heamawatanachai (my daughter) and Parrish Heamawatanachai (my son) who are the
sources of inspiration in my life.
CHAPTER 1
INTRODUCTION
1.1 Introduction
The use of microproducts is increasing rapidly in various industries such as
electronics, medical, optics, automotive and avionics [1-2]. The three major groups of
microfabrication technologies to create microfeatures are: lithography-based techniques,
electro-physical and chemical systems, and mechanical systems. The lithography-based
fabrication technologies are capable of producing microfeatures and are well suited for
mass production. However, there are some limitations such as the restricted choice of
work materials, the inability to produce complex 3D geometries, and the need for a clean
room environment. As a layered process that requires masks for the photolithography, the
one-time cost is very high, making this process expensive for prototyping.
The electro-physical and chemical technologies such as EDM and ECM are
capable of machining microfeatures (2D and 3D) on electrically conductive materials.
These technologies need no direct physical contact between the tool and work piece,
which helps to avoid problems such as material deformation, vibration and breakage.
There are some drawbacks such as the difficulty of controlling the machining position
depth because of tool wear in EDM, and the subsurface property change due to the
2
thermal and chemical reactions. The mechanical machining techniques such as turning,
milling and drilling need direct contact between the tool and work piece. Hence, a good
correlation between the tool geometry and the work piece surface can be obtained. They
usually provide high material removal rates, good surface finish and are able to machine
2D or 3D features in a wide range of materials. However, the increasing chance of
breakage of the tool for smaller tool sizes is a fundamental problem of these techniques.
Micromilling is a scaled-down version of conventional milling. As the size
becomes smaller, it is very difficult to manufacture the milling tool. This is because of
the complex geometry of the milling tool requiring two or more helical flutes. In addition,
the flutes also reduce the cross-sectional area of the tool, which further weakens the tool.
Therefore, for small tool sizes, the chance of tool breakage, especially for tool diameters
of less than 0.2 mm, rapidly increases [3]. The micromilling tools also face the problem
of chip clogging, which leads to the breakage of the tool [4]. Friedrich and Vasile [5]
reduced the complexity by making micro-end mills with a square shape using a focused
ion beam machining process (FIB), which allowed the effective tool diameter to be
reduced to 22 microns.
In this study, a novel micromachining tool was developed using the orbital motion
machining technique [6]. The principle of the tool is achieved by orbiting the single point
tool tip in a recirculating trajectory at high frequency. The tool tip is made of a single
crystal diamond that is known as the hardest of all commercially available materials.
Therefore, a wide range of materials can be machined with this tool. In this orbital
machining technique, any point on the circumference of the tool tip can be used as the
cutting edge, unlike the milling technique where the cutting edges are on the tool flutes.
3
The cross-section geometry of the orbital machining tool tip can be made circular, which
is a simple and strong geometry. Thus, the effective size of the tools cross-section can be
made as small as 1 m, which leads to the smaller possible features that can be made
compared to the milling technique. Moreover, in terms of machined surface quality,
single point diamond turning is known as a technique to create mirror surface finish on
brittle materials such as silicon [7, 8, 28].
1.2 Significance of the research

A new and exciting thrust of research that particularly depends on
micromachining is the area of microfluidics. This area has been called one of the
greatest engineering challenges of the century by the Stanford Microfluidics Laboratory
and is vital to drug discovery efforts and fundamental genetic research [9]. Currently,
microfluidic devices are fabricated from plastics (PMMA, PDMS) through molding [10]
or a combination of molding and bending [11-12] or even micromilling [13]. Other
methods include lift-off [14], hot embossing [15] or laser ablation [16]. While all these
techniques have proven feasible, the low melting temperature of the materials (polymers)
result in significant limitations for the fabricated devices.
Another major fabrication technique is based on photo-lithography, which
requires a sequence of patterning and etching to create the geometry desired [17-19]. This
technology requires materials that can be etched (silicon, metals) and, while economical
for mass production, is very costly to prototype because of the need for masks required
for patterning. A third method is based on glass that is machined with lasers [20-21].
4
Prototyping microfluidic devices based on high temperature materials such as
silicon is typically based on photolithography to fabricate two-dimensional (2D) layers;
these layers are aligned, stacked, and sealed to produce the three-dimensional (3D)
microfluidic network. Although this method can be applied broadly, misalignment
between adjacent layers often creates discontinuities at the interface between
connections, which in turn cause disturbances in the flow.
This study intended to develop a new mechanical micromachining technique that
is better suited to prototype 3D microfeatures compared to existing techniques. The
cutting tool tip is made of a hard material such as single crystal diamond that is known as
the hardest commercially available material and able to be used to machine a wide range
of materials (such as polymers, metals, semiconductor materials and glass). A piezo tube
scanner is used as an actuator to create the cutting motion of the tool tip. The novel tool
was designed [6] such that it can easily be integrated to any 3-axis machine such as a
CNC machine to create microfeatures. The tool is also able to be operated in a typical
machine shop environment, which means no clean room is required. Therefore, this
technique provides low investment and operating cost. Moreover, the motion of the tool
tip is generated by a piezo tube, which is a solid-state device with no conventional
moving parts that results in a very long operating time.
1.3 Existing micromachining techniques

To create microfeatures, there are two fundamental methods: top-down and
bottom-up. Top-down manufacturing methods are based on material removal and begin
with a work piece whose overall dimensions exceed that of the finished part. Then, its
5
size is reduced by selectively removing material through cutting, milling, etching and
slicing. Bottom-up methods are based on selectively adding material to form a bigger
structure.
The
manufacture
of
microfeatures
is
principally
based
on
top-down
manufacturing methods, which remove the cohesion among particles to create smaller
structures. Examples of removing methods include electrical discharge machining
(EDM), electrochemical machining (ECM), lithography-based microfabrication, laser
machining, milling and vibration assisted machining (VAM).
1.3.1 Electrical discharge machining (EDM)

Electrical discharge machining (EDM) uses electrical discharges to remove
conductive materials such as metals. EDM is a thermal machining process that can be
used to drill microholes in hard-to-machine materials such as titanium and polycrystalline
diamond (PCD). The concept of the EDM process is shown in Fig. 1.1 [22]. The pulsed
arc discharges occur in the gap between the electrode and work piece, which is filled with
an insulating medium (dielectric liquid). The insulating medium is important in avoiding
an electrolysis effect during the EDM process. The shape of the work piece is a copy of
the shape of the electrode with an offset of the gap between them. The gap width depends
on the type of electric fluid and also EDM parameters such as voltage supply and size of
the capacitor for the relaxation type pulse generator. The discharge is ignited by a high
voltage, which is high enough to overcome the dielectric breakdown strength of the small
gap. In each discharge column with high temperature, some small amount of material is
Feed direction
Dielectric fluid
Bubbles
Tool electrode (-)

Workpiece (+)
Discharge current
Discharge column
Fig. 1.1 Concept of EDM (adapted from Kunieda et al. [22])
melted and ejected into the dielectric fluid. Therefore, a small crater is developed on
both electrode and work piece surfaces. The removed material particles are cooled
and solidified by the dielectric fluid to form debris particles, which are flushed out from
the gap by the dielectric flow.
1.3.2 Electrochemical machining (ECM)

The principle of ECM is shown in Fig. 1.2. Like EDM, it is a noncontact method
that shapes the work piece according to the shape of the cathode. Hence, the materials
have to be electrically conductive. The medium between the work piece and the electrode
is a highly conductive electrolyte, which is forced through the gap at high speed (10 to 60
m/s) to prevent metal ions from plating the cathode tool and also for heat removal
purposes. The metal removal is achieved by passing a large amount of DC current up to
D. C.
power
supply
Electrolyte in
Electrode (-)
Insulator
Electrolyte out
Workpiece (+)
Fig. 1.2 Electrochemical removal process
100 A/cm2 through the electrolyte solution, which typically is a high concentrated
solution of inorganic salt such as sodium chloride, potassium chloride and sodium nitrate.
The outside of the electrode is insulated to prevent machining actions on the sides. The
supply voltage for ECM commonly ranges from 5 to 20 V. Higher voltage is used for
rough machining and lower voltage is used for finishing cuts. The DC current and the
electrolyte solution cause the electro-chemical reaction to convert the removed material
into hydroxide, which is flushed away by the solution flow. The advantages of ECM are
low running cost, very good surface quality and essentially zero wear. The disadvantages
are the difficulties in making the electrode, which needs to be insulated, and the initial
cost of the equipment.
8
1.3.3 Photolithography
Photolithography is the most widely used technique in microfabrication. It is the
process that creates a pattern on the work piece by shining light through a mask onto the
work piece. The mask is based on a pattern of opaque material that is deposited on a glass
plate. Some of the light is absorbed by the opaque pattern, while the remaining light
passes through the glass to illuminate the work piece, which is covered by a thin layer of
photo resist. This technique is always used with other techniques such as etching and
deposition to create the microstructures. Fig. 1.3 [23] shows the pattern transfer on a
silicon wafer.
Fig. 1.3 Basic photolithography and pattern transfer (adapted from Madou [23])
9
1.3.4 Micromilling
Micromilling is a mechanical machining technique that is capable of machining
microfeatures in a wide range of materials. The principle of micromilling is similar to
macromilling, except that the size of the tool is miniaturized. As seen in Fig. 1.4, the tool
rotates around its axis at high frequency in order to cut the material. The major
limitations in micromilling are the manufacturing of the smaller tool, which is difficult
due to the complex tool geometry as well as the increasing chance of tool breakage for
very small tool diameters of 0.2 mm or less [3].
Feed direction
Rotation
Milling tool
Work piece
Fig. 1.4 Principle of milling process
10
1.3.5 Single point diamond machining
Single point diamond machining is a technique that uses the single point diamond
(natural or synthetic) as a material for the tool tip. Because diamond is the hardest of all
materials, it can be used to machine a wide range of materials. It can be applied to many
applications depending on the motion of the tool and work piece. Single point diamond
turning, where the system setup is similar to a lathe, can be used to create a mirror
surface finish. The various geometries that can be created depend on the geometry (with
sharp cutting edge) and the motion of the tool (rotating and feeding) relative to the work
piece, as seen in Fig. 1.5 [27].
Tool tip
Tool tip
(a) Rectangular cross section (b) Curvy cross section
Tool tip
(c) Triangular cross section
Fig. 1.5 Microgrooves (adapted from Dornfeld et al. [27])
11
1.3.6 Vibration assisted machining
Vibration assisted machining (VAM) is a combination of precision machining that
is augmented by a tool vibration to improve the machining performance [27]. It can be
applied to many applications from turning to drilling to grinding [24]. There are two
types: 1D and 2D VAM. In 1D VAM (or linear tool path), the tool is vibrated with small
amplitude in the same direction as the feed. The 2D VAM (or circular/elliptical tool
path), as shown in Fig. 1.6, is based on a tool that is vibrated in 2 axes in the form of a
circle or ellipse to improve the performance of the orthogonal cutting process. The
literature shows that VAM helps in the reduction of machining forces and chip thickness.
Moreover, it also improves the surface finish and extends the tool life.
Feed direction
Elliptical motion
Chip
Tool
Work piece
Fig. 1.6 2D Vibration assisted machining.
12
1.4 Overview of orbital motion for micromachining
Orbital motion is applied to many applications to improve the efficiency and
capability of current systems. In drilling with EDM, the electrode (tool) is typically
rotated around its own axis while moving in the axial direction to drill the work piece
(conductive material). The size of the drilled hole is slightly larger than the size of the
electrode with a small gap, which is dependent on the machining parameters (voltage and
capacitor). Therefore, to drill various sizes of holes, it is necessary to use various sizes of
electrodes to match each size of the desired hole, which is time consuming and costly.
From the work of Bamberg and Heamawatanachai [26], a flexural x-y stage (travel length
of + 100 m) was designed to integrate orbital motion with the electrode rotating motion
of EDM. As a result, the size of the hole was decoupled from the size of the electrode,
which allowed a single size of electrode to drill various sizes of holes, thereby reducing
the time and cost for making electrodes. Moreover, the orbital motion improved the
flushing of the debris from EDM process, which increased the efficiency in drilling of the
deep holes.
In mechanical machining, generally orbital motion (such as in vibration assisted
machining) was used to improve efficiency, surface quality and tool life of the system.
Shamoto et al. [29] developed an ultrasonic vibration tool for cutting of sculptured
surfaces on hardened die steel. The cutting principle was achieved by vibrating the tool
tip at high frequency of 34.4 kHz. Thus, a mirror surface finish was obtained with this
technique.
13
1.5 Design requirements
The following are the design requirements of the novel machining technology.
Capable of machining 3D microfeatures in a wide range of materials

(polymers, metals, semiconductors, glasses): This was demonstrated by
experiments that showed that the tool can machine 2D or 3D features into many
types of materials such as metals, semiconductor materials and glass.
Smooth surface finish (Ra < 0.5 micron): The actual experimental results on
aluminum alloy (Al 2024) showed a surface roughness Ra = 0.05 micron.
Long term durability: The experiments showed that the tool orbit radius
remained constant and the center of the trajectory drifted less than 0.5 micron
over a measurement time of more than 60 minutes.
Can be operated in a typical machine shop environment (no clean room

required): The tool was attached to CNC machines to perform experiments
without a problem.
Based on the requirements listed above, a novel micromachining tool was
designed that can machine materials in any horizontal direction, similar to the material
removal with an end mill. Moreover, this tool is capable of machining in the vertical
direction to create 3D microfeatures. Unlike milling, where the tool rotates around its
axis, the cutting motion is achieved by orbiting the tool tip at high frequency, as seen in
Fig 1.7. By combining this microscopic motion tool with the macroscopic motion of a 3axis machine tool, this technique can create 3D microfeatures in a wide range of
materials [6, 25].
14
Feed direction
2
3
Center axis
Orbital motion
Cutting boundary
Tool axis
Work piece
Fig. 1.7 Orbital micromachining principle
1.6 Hypotheses
In this study, a novel orbital mechanical micromachining tool was developed
capable of creating microfeatures in a wide range of materials. The following are the
hypotheses in this study.
The single point tool tip with orbital motion only (no rotating motion), as
seen in Fig 1.6, can be used as a machining tool for micromachining
applications: This was proven by the achievement of micromachining with this
tool on many materials such as aluminum alloy (Al2024), single crystal silicon
and glass.
In this technique, the tool does not require a sharp edge, and the size of the
tool is large compared to the size of the uncut chip thickness to be removed:
15
This was proven by experiments on aluminum alloy (Al2024) that showed that the
tool with 1 micron tip radius can machine the work piece even with uncut chip
thickness of 0.1 micron.
The tool tip can be made of a hard material such as diamond or cubic boron
nitride (CBN) with a simple geometry (circular cross section) without sharp
edge (such as flutes edge of milling tool): The diamond tip that was used in this
study has circular cross-section.
This technique can be used to machine both ductile and brittle material: This
was proven by experiments that showed that the tool can create 3D microfeatures
on ductile materials such as aluminum alloy (Al2024) and brittle materials such as
single crystal silicon.
CHAPTER 2
DESIGN, MANUFACTURING AND CONTROL
2.1 Introduction
Micromilling is a versatile micromachining process used to produce parts whose
feature dimensions typically range from hundreds of microns up to several millimeters.
At present, the technology is in essence a scaled-down version of conventional milling
with end mills as small as 0.05 to 0.1 mm diameter. As a result of the small cross section
of the tools, which is further reduced by the helical flutes, Tansel et al. [3] found that tool
breakage is a major concern. A solution to this problem is to reduce the complexity of the
end mill geometry. Friedrich and Vasile [5] eliminated the helical flutes by giving the end
mill a square shape using a focused ion beam machining process (FIB). As a result, the
effective tool diameter could be reduced to as little as 22 m.
To machine even smaller features, and to reduce the cost of producing extremely
small microtools even further, a new approach to micromachining is developed in this
study. By reducing the tool geometry to a single point, and making this point from
diamond, which is the hardest and most wear resistant material known, the capability of
micromilling can be expanded not only towards smaller features but also to difficult-tomachine materials such as stainless steel, titanium or even glass. This would expand
micromilling to the machining of semiconductor materials as well, similar to the work by
17
Rusnaldy et al. [30], who used diamond-coated micro-end mills to machine single-crystal
silicon. The study found that ductile regime cutting with end mills is almost impossible,
in part due to the effective rake angles of the tool, which Patten and Gao [8] identified as
an important factor in ductile regime cutting.
The novel aspect of this technique is the generation of the cutting speed, which
does not involve a tool that rotates about its own axis. Instead, the cutting motion is
achieved by actuating the tool tip on orbital paths while the feed rate is generated by
either advancing the work piece or the tool with a separate motion. This completely
decouples the cutting speed from the feed rate, which is an advantage over the nanocutting experiments reported by Gao et al. [31-32], where material is removed by
actuating a work piece with a piezo tube along a single direction and gradually advancing
the work piece laterally after each cut.
To generate the cutting speed and to decouple it from the feed speed, the tool tip
is actuated on a small circular trajectory whose frequency and radius can be adjusted onthe-fly. This ability is achieved with a piezo tube scanner that is energized with
sinusoidal voltages that create oscillatory bending as demonstrated by Bamberg and
Heamawatanachai [6, 25]. The single-point diamond tool used in this technology results
in reduced stresses compared to standard cylindrical tools and has been shown by Pizani
et al. [7] and Chao et al. [33] in diamond turning to enable ductile material removal of
otherwise brittle materials such as silicon. This potentially increases the surface quality
and may result in the mirror-type surface finish that diamond turning is known for.
18
2.2 Cutting module
This novel micromachining technology is based on single-point diamond tip
milling, which can be used to machine microfeatures into a wide range of materials
ranging from polymers, through metals to glass. The unique characteristic of this
technology is that the cutting velocity is not created by rotating the tool tip about its axis
of rotation. Instead, the cutting motion is generated by actuating the tool tip on a recirculating, microscopic trajectory at high frequencies. This allows a standard 3-axis
stage (e.g. from a CNC machine) to generate the macroscopic motion required for the
tool path while at the microscopic level, the cutting motion and the tool tip path are being
generated by the tool itself. As a result, truly three-dimensional part geometries at the
micron level with a surface roughness of less than 0.5 micron (Ra) can be created.
The microscopic motion of the tool tip is achieved by a 76.2 mm long piezo tube
scanner, shown in Fig. 2.1, that is made from EC-63 ceramics with an outer diameter of
15.4 mm and a wall thickness of 0.79 mm with four external electrodes. The capacitance
of each electrode was measured as 8.4 nF. The microscopic cutting motion of the tool is
achieved by energizing the four electrodes with time-varying voltages that range between
0 and -1000 V, which creates periodic bending of the free end of the tube. If the voltages
are sinusoidal with a phase lag of 90o between adjacent electrodes, the resulting periodic
bending creates a planar, orbiting motion with a maximum radius of 35 m that is
perpendicular to the piezo tubes center axis.
19
Fig. 2.1 The ceramic piezo tube scanner with four electrodes on the outside surface and
the center ground on the inside surface (Heamawatanachai and Bamberg [25]).
The concept of the voltage generation is shown in Fig. 2.2 and consists of a
National Instruments DAQ card (PXI-6040E) and four high-voltage piezo amplifiers (PI
E-507.00), that amplify the 0 - 10 V analog waveforms by a factor of minus 100 to
produce the 0 to -1000 V used to drive the piezo tube.
Since the orbiting motion is planar, the machined surface is flat, making this
cutting mode the equivalent of a miniaturized fly cutter, whose tool tip offset from the
axis of rotation can be varied during cutting simply by varying the amplitude of the
driving voltages (Fig. 2.3).
The cutting mechanism is shown in more detail in Fig. 2.4 and illustrates the
interaction of the rounded tool tip with the material. As shown in Fig. 2.4(b), the material
to be removed is in the form of a very thin cylindrical shell whose curvature matches that
of the tool tip.
20
Fig. 2.2 The piezo tube is actuated by analog voltages that vary from 0 to -1000 V
(Heamawatanachai and Bamberg [25]).
Non-rotating tool shank
(with reference flat)
PZT orbital motion

(radius 0 - 35 m, 0 - 2.5 kHz)
Tool tip
orbital motion
Fe
(w ed
ork
pie
Depth of cut
ce
Tool tip
(diamond cone
with tip radius 1-10 m)
bi
Or
tal
m
dia
ete
Fig. 2.3 The cyclic bending of the piezo tube generates a planar, orbital motion of the
tool tip (Heamawatanachai and Bamberg [25]).
Fig. 2.5 shows the cutting module prototype that consists of the piezo tube
scanner, a housing to protect the tube, a clamp to hold the tool shank and two noncontact
inductive probes to monitor the cutting motion. Fig. 2.6 shows the close-up view of the
clamp holding the tool shank that is mounted to the free end of the piezo tube. Fig. 2.7
shows the single crystal diamond tool tip with 60o cone angle, which is held by the tool
holder.
21
Rtip
Top view
Fa
vc not yet machined
(a)
Rtip
Detail
Ft
Fr
(b)
already machined
Sid
ie
ev
(d)
Rtip
(c)
Rtip
Side view (with tool)
Top view
Fig. 2.4 The cutting mechanism in 3D (a), in more detail (b), as a side view (c),
and as a top view (d) (Heamawatanachai and Bamberg [25]).
Fig. 2.5 The cutting module prototype consists of the piezo tube, a housing to protect the
tube, a clamp to hold the tool shank, and noncontact probes to monitor the cutting motion
22
Fig. 2.6 A clamp holding the tool shank is mounted to the free end of the piezo tube.
Tool clamp
Diamond cone (60)

Tip radius ~10m
Stylus (steel 3mm)
Fig. 2.7 The cutting tool is a stylus with a 60 diamond cone with a tip radius of 10 m
23
As shown in Fig. 2.8, in order to provide electrical insulation, the piezo tube is
bonded to the base and top insulators with a high strength, nonconductive epoxy (Epotek
353ND). This epoxy has a shear strength at 23 C of almost 14 MPa. For a joint overlap
of 2 mm, the maximum force in the vertical direction that this bond area can hold is
1203 N. Fig. 2.9 shows the actual picture of mounting of the piezo tube to the insulator
base.
Fig. 2.8 Assembly of the tool. Electrical isolation of the piezo tube is achieved by
bonding it to ceramic insulators with high stiffness epoxy (Heamawatanachai and
Bamberg [25]).
24
Fig. 2.9 The piezo tube is attached to the housing by bonding with epoxy to the base
insulator.
2.3 System components and control of the orbital micromachining tool

In this study, several versions of system setups and control programs were
developed. Fig. 2.10 shows the first system setup of this study that combines the
macroscopic and microscopic motion modules. The macroscopic module is achieved by
using a micro-EDM machine (Optimation Profile 24P). The microscopic module, which
was the main objective of this study, consists of the cutting module, amplifier, signal
generator, position sensor and control program. One of the objectives in this study was to
design the micromachining tool that could be integrated into a general 3-axis machine
such as a CNC milling machine. Fig. 2.11 and 2.12 show system setups that use CNC
machines (Haas-TM-1 and Haas-VF-E, respectively) as the macroscopic motion
modules. The CNC machine was used in later experiments because it has higher stiffness
axes than the Optimation Profile 24P, which was designed for EDM application.
25
Fig. 2.10 System setup #1, the microscopic motion module was mounted to the
macroscopic motion module (Micromachine tool, Optimation 24P).
Fig. 2.11 System setup #2, the microscopic motion module was mounted to the CNC
machine (Haas TM-1).
26
CNC machine
Haas-VF-E
CNC control screen
CNC Z-axis
Orbital control screen
Cutting module
Microscope
Piezo amplifier
PI-E-507.00
Data acquisition system
NI-PXI-6040E
Fig. 2.12 System setup #3, the microscopic motion module was mounted to the CNC
machine (Haas VF-E).
2.3.1 Microscopic motion control

The cutting module, as seen in Fig. 2.13, consists of the diamond tip on a tool
shank that is held with the tool holder. The tool holder is attached to the free end of the
piezo tube scanner while the other end of the piezo tube is held stationary. In order to
actuate the motion of the tool tip, sinusoidal voltage signals (range -5 to 5 V) are
generated from voltage generation modules (NI-PXI-6040E or NI-PCI-7833R) and
amplified and adjusted by the piezo amplifiers (PI E-507.00) to achieve the operational
voltage range of 0 to -1000 V, which are applied to the electrodes of the piezo tube (x+,
x-, y+ and y-).
27
Fig. 2.13 Close view of the cutting module during machining
2.3.1.1 The first version control system

Fig 2.14 shows the control diagram of the first version control system. The analog
voltages that drive the piezo tube are generated with a National Instruments card (NIPXI-6040E) that is controlled by LabVIEW. The card has only two analog channels,
which produce two sinusoidal waveforms with a 90 phase. These voltages drive the x+
and y+ electrodes. To generate the voltages required to power the x- and y- electrodes, a
separate circuit with op-amps was used to invert the voltages, thereby creating the
required 180 phase between the positive and negative electrodes such that the x-axis
electrodes are driven by the same voltage but with a 180 phase lag.
28
PZT microscopic motion control diagram (first version)
2 sinusoidal signals (x-, y-)
Inverter
Signal generator
(NI-PXI-6040E)
Computer
Control Program
-Labview
Center axis
Fixed end
Signal amplifier
(PI E-507.00)
4 sinusoidal signals
(x+, x-, y+, y-) High voltage signals
2 sinusoidal signals
Housing
(x+, y+)
Piezo tube scanner (PZT)
with 4 electrodes (x+,x-,y+,y-)
Inductive probes (x&y)

x&y positions
Data Acquisition
(NI-PXI-6040E)
Position sensor
(Lion Precision UC3)
Diamond tip
Free end
Orbital motion
Tool axis
Tilt angle
Fig. 2.14 First version control diagram of the cutting module
The developed LabVIEW control program is shown in Fig. 2.15. It also illustrates
the two sinusoidal waveforms that generate the driving voltage for the two positive piezo
tubes electrodes, x+ and y+ and the circular actual motion of the tool clamp as measured
by the noncontact inductive sensors.
From Fig 2.15, it is interesting to note that the polar plots of the x+/y+ waveform
and the resulting tool motion do not match. While the tool motion is clearly circular, the
driving voltage to generate this motion is shaped like an oval with a phase of 77 between
the two waveforms at 100 Hz frequency. This is the result of manufacturing and
assembly tolerances, the variations in material properties of the piezo tube and also the
signal amplification system. The following topic will detail the experiments to identify
the major sources of phase delay behavior.
29
Fig. 2.15 LabVIEW VI showing the output voltage from the signal generator and the
resulting motion of the tool clamp.
The position measurement program is developed in LabVIEW. This program
controls the acquisition card (NI-PXI-6040E) that samples the measured tool tip position
from the inductive probes. Fig. 2.16 shows example trajectories of the tool tip position as
measured by the inductive probes for various signal parameters. It also indicates that the
developed tool is capable of creating various types of motion such as circular, ellipse,
parabolic, etc.
30
a) Ux=2V, Uy=2V, y/x=1, 90 deg Phase
b) Ux=3V, Uy=3V, y/x=1, 60 deg Phase
c) Ux=3V, Uy=3V, y/x=2, 90 deg Phase
d) Ux=3V, Uy=3V, y/x=3, 90 deg Phase
Fig. 2.16 Example of tool trajectories
31
2.3.1.2 The second version control system
As shown in Fig. 2.17, the control card for signal generation was replaced with an
FPGA National Instrument card (NI-PCI-7833R), which is controlled by LabVIEW
FPGA. This card can generate up to 8 analog signals. Four channels of the card are used
to drive the piezo tube electrodes x+, x-, y+ and y- directly, which eliminates the need to
use an inverting circuit. The new control system is also capable of controlling the z- axis
or the length of the piezo tube by applying the additional equal amount of offset voltage
to all electrodes. Fig. 2.18 shows the control screen of the signal generation program.
PZT microscopic motion module (second version)

Fixed end
Signal generator
(NI-PCI-7833R)
Center axis
X
Signal amplifier
(PI E-507.00)
4 sinusoidal signals High voltage signals

(x+,x-,y+,y-)
Housing
Computer
Control Program
-Labview
-Labview FPGA
Piezo tube scanner (PZT)

with 4 electrodes (x+,x-,y+,y-)
Inductive probes (x&y)

x&y positions
Data Acquisition
(NI-PXI-6040E)
Position sensor
(Lion Precision UC3)
Diamond tip
Free end
Orbital motion
Tool axis
Tilt angle
Fig. 2.17 Second version control diagram of the cutting module
32
Fig. 2.18 Signal generator control program
33
As seen in Fig. 2.18, the program is used to control all 4 electrodes of the piezo
tube (x+, x-, y+, y-). The signal generator for each axis requires different command
parameters (phase and voltage gain), because the frequency responses of the amplifiers to
the actual output voltage and phase are different. The effect of different phase delays can
be clearly seen in the Fig. 2.15, where the actual tool trajectories do not match the signal
generation trajectories. The voltage gain and its effect on tool amplitude is dependent on
the frequency (discussed in more detail in Chapter 3).
In order to create the actual circular trajectories as desired, calibration of all axes
for each operating frequency is necessary. In this program, the voltage signal of the yelectrode is used as a reference. All other axes can be calibrated such that the actual
trajectory matches the trajectory observed by the y-axis. For example, during calibration
of the x+ axis at 100 Hz, the command voltages applied to the electrodes (x- and y+)
were initially set to zero. The global phase difference of the x and y axes were also set to
zero. After that, the calibration parameters (phase delay and voltage gain) of the x+ axis
are adjusted to match the tool position of the y- axis (Fig. 2.19). The goal is to obtain the
actual trajectories as a straight line at a 45o angle, as seen in Fig. 2.20. This procedure is
repeated for the remaining axes. After the calibration of all axes, the actual circular
trajectories can be created by setting the global phase difference between the x and y axes
to 90o, as seen in Fig. 2.21(a), (b) and (c). Fig. 2.21 shows some examples of the actual
tool trajectories for various control parameters after calibration at 100 Hz.
34
Fig. 2.19 Setting parameters to calibrate and match x+ axis to the y- axis.
35
Fig. 2.20 Actual tool trajectories from parameters of Fig. 2.19
Fig. 2.21 Actual tool trajectories at 100 Hz after calibrating, (a) 1V amplitude and
90o phase, (b) 2V amplitude and 90o phase, (c) 3V amplitude and 90o phase, (d) 2V
amplitude and 0o phase, (e) 2V amplitude and 60o phase, (f) 2V amplitude and 120o
phase.
36
2.3.2 Macroscopic motion control
For the system setup #1, as seen in Fig 2.10, a 3-axis stage of a sinker EDM
machine (Optimation Profile 24) was used for the macroscopic motion. The control
software (Optimation rev2) is programmed with Visual Basic and was originally
developed to control the EDM machine for microdrilling applications. As such, the
program was limited to control the vertical axis only. The program was then modified to
provide full 3-axis control to support 3D milling and 3D EDM applications [26]. The new
version includes a G-Code interpreter that generates the motion instructions for the
motion controller (Galil PCI-1840).
G-Code is an internationally standardized language used to program computer
numerically controlled (CNC) machine tools. The G-Code command can be used for
controlling the CNC machines (Haas-TM-1 and Haas-VF-E), which were used in the
system setup #2 and #3, as seen in Fig. 2.11 and 2.12, respectively. This code is
generated with commercially available computer-aided manufacturing (CAD-CAM)
software (e.g. FeatureCam, MasterCam, Pro/MANUFACTURE, etc.) directly from the
solid models of the work pieces.
2.4 Conclusions
This micromachining technique is based on single-point diamond tips that are
actuated at the microlevel by a piezo tube scanner. The resulting tool tip paths, which are
used to generate the cutting motion, have trajectories that can be varied easily by
adjusting the sinusoidal excitation voltages. By actuating the tool tip on its own, recirculating tool paths, the cutting speed is decoupled from the feed speed, providing a
37
cutting mechanism with an offset from the center axis that is very similar to a miniature
fly cutter. The difference is that the tool is orbited instead of rotated about its center axis.
The FPGA National Instrument card (NI-PCI-7833R) was used to generate four
sinusoidal signals, which are amplified to drive the piezo tube electrodes x+, x-, y+ and
y- directly. This allows the system to control the z- axis or the length of the piezo tube by
applying the additional equal amount of offset voltage to all electrodes.
The dynamic motion of the tool from sinusoidal signals that is used in orbital
machining minimized the errors from hysteresis and drift, which are the major errors in
control of piezo tube in static mode. The open loop control that is used in this study
simplifies the control efforts and also lowers the cost. The error from manufacturing of
the tool such as misalignment and the imperfect output from the amplifiers can be
minimized through a calibration procedure for each operating frequency.
The orbital micromachining tool (microscopic motion module) in this study can
easily be integrated into any 3-axis machine such as a CNC machine (macroscopic
motion module) to create microfeatures. The G-code that is used to program tool paths
for the CNC machine can be generated easily from any computer aided manufacturing
software such as Pro/MANUFACTURE or FeatureCam.
CHAPTER 3
TOOL CHARACTERIZATION
For orbital micromachining with single point tool tip geometry, the actuator that
is used to drive the diamond tip needs to have sufficient stiffness to minimize unwanted
deflections from the cutting force. The frequency of operation is also an important issue,
because high orbiting frequencies can increase the material removal rate. The piezo tube
scanner is an interesting choice because it is a solid state actuator that has a very fast
response to the signal voltage and it has high strength. Moreover, as a solid state device,
it has no conventional moving parts (e.g. bearings) and therefore has a very long life.
3.1 Piezo tube modeling

Piezo ceramic tubes are used in a wide range of applications, ranging from microdispensing to scanning microscopy [34]. The material has monolithic components that
contract in the lateral and longitudinal directions when a voltage is applied between the
inner and outer electrodes. Utilizing these two basic contractions in combination with
four separate electrodes can provide three-dimensional motions in the x-, y- and zdirection as a function of the applied voltages.
39
3.1.1 Static deflections
A piezo tube actuator is an interesting choice for a micromilling operation and
must generate the cutting motion with sufficient power to allow the diamond tip to
penetrate the material to be removed and to overcome its specific cutting energy. The
material has a fast response to voltages applied to its electrodes and its geometry in
combination with its high elastic modulus should result in a reasonably large static and
dynamic stiffness. Moreover, as a solid state device, it has no moving parts (e.g.
bearings), minimizing assembly efforts and also providing a long life. For a piezo tube
driven microtool, the most important parameters are the lateral displacement of the free
end and the change in tube length as a voltage is applied.
The axial contraction of a piezo tube is given by Sun and Wolkow [34, 35]:
L = d 31
LU
t
( 3.1)
where L is the change in length of the tube, d31 is the piezoelectric strain coefficient for
displacements normal to the polarization direction, L is the length of the tube, U is the
voltage applied to all electrodes and t is the wall thickness.
The lateral displacement ( ) at the free end of the tube can be determined as [35]:
2 2d 31 L2U y
2 2d 31 L2U x
, y =
x =
I D t
I D t
(3.2)
40
where ID is the inner diameter of the tube and Ux and Uy are the symmetric electrode
voltages about the center voltage (Uc) that are applied to the x and y electrodes,
respectively. The voltage for the x+ electrode is found as Ux+ = Uc + Ux and the voltage
for the x- electrode is given as Ux- = Uc - Ux. The voltages for the y+ and y- electrodes are
found similarly. The piezo tube used in this study is made from EC-63 and the
piezoelectric strain coefficient d31 is given by the EDO Corporation [36] as 1.20e-10
m/V. This yields the tip displacements x and y as 0.0575 m/V.
Given the range of 0 to -1000 V of the amplifiers, the center voltage of all
electrodes was set to -500 V, which limits the maximum amplitude to the electrodes to
500 V. Based on Eq. (3.2), the maximum lateral displacement is predicted as 28.7 m
and according to Eq. (3.1), the tube is shortened by 5.79 m.
3.1.2 Orbital paths

The cyclic bending of the piezo tube is the principle of this device and the radial
displacement in the x-y plane can be predicted using Eq. (3.2). However, due to bending,
the z position of the tip will be affected by the amount of radial displacement according
to:
z = L L2 r 2 = L L2 ( x2 + y2 )
(3.3)
where L is the tube length, r is the radial displacement of the free end of the tube, and x
and y are deflections in x and y directions. The orbital tool path can be generated by
applying 90 phase difference sinusoidal signals to all electrodes according to:
41
U x + = U x sin(t )
U y + = U y sin(t + / 2)
(3.4)
U x = U x sin(t + )
U y = U y sin(t + 3 / 2)
By using the same amplitude signal to all axes (Ux=Uy=U1) in combination with
Eq. (3.2) and Eq. (3.3), the circular orbital tool path can be generated as illustrated in Fig.
3.1.
Elliptical orbital motion can be created by applying different amplitude signals to
the x and y axes. In this case, the axes of the ellipse are aligned with the axes of the piezo
tube. This is illustrated in Fig. 3.2,
By changing the phase of the applied voltages, the principal axes of the elliptical
motion can be rotated as seen in Fig. 3.3.
Circular orbital motion
Ux+ = U1sin(t)
Uy+ = U1sin(t+/2)
Ux- = U1sin(t+)
Uy- = U1sin(t+3/2)
0.01
zaxis (m)
0.008
0.006
0.004
0.002
0
U1 = 100 V
20
0
20
xaxis (m)
30
20
10
10
yaxis (m)
20
30
U1 = 200 V
U1 = 300 V
U1 = 400 V
U1 = 500 V
Fig. 3.1 Circular orbital tool path (Heamawatanachai and Bamberg [25]).
42
Elliptical orbital motion
Ux+ = 0.5U1sin(t)
Uy+ = U1sin(t+/2)
Ux- = 0.5U1sin(t+)
Uy- = U1sin(t+3/2)
0.01
zaxis (m)
0.008
0.006
0.004
0.002
0
U1 = 100 V
20
0
20
30
20
10
10
20
30
U1 = 300 V
U1 = 400 V
yaxis (m)
xaxis (m)
U1 = 200 V
U1 = 500 V
Fig. 3.2 Elliptical tool path.
Elliptical orbital motion (x-phase angle = 30)

Ux+ = U1sin(t+/6)
Uy+ = U1sin(t+/2)
Ux- = U1sin(t+7/6)
Uy- = U1sin(t+3/2)
0.01
zaxis (m)
0.008
0.006
0.004
0.002
0
U1 = 100 V
20
0
20
xaxis (m)
30
20
10
10
yaxis (m)
20
30
U1 = 200 V
U1 = 300 V
U1 = 400 V
U1 = 500 V
Fig. 3.3 Elliptical tool path with 30o phase shift in x axis (Heamawatanachai and
Bamberg [25]).
43
According to Eq. (3.1), by applying the same amount of offset voltage to all axes,
the length of the tube or z-axis can be manipulated.
U x + = U x sin(t ) U z
U y + = U y sin(t + / 2) U z
(3.5)
U x = U x sin(t + ) U z
U y = U y sin(t + 3 / 2) U z
To compensate for the change in height as a function of tip displacement in the xand y- direction (Fig. 3.1) and to obtain a truly planar orbital motion (Fig. 3.4), the offset
voltage can be calculated by:
U2 =
t
d 31 L
(L
L2 x2 + y2
))
(3.6)
Circular orbital motion with height compensation

Ux+ = U1sin(t) - U2
Uy+ = U1sin(t+/2) - U2
Ux- = U1sin(t+) - U2
Uy- = U1sin(t+3/2) - U2
0.01
zaxis (m)
0.008
0.006
U2 =
4 2d31LU1
t
1 1
d31
I D t
U1 = 100 V
0.004
U1 = 200 V
0.002
U1 = 300 V
U1 = 400 V
U1 = 500 V
20
0
20
xaxis (m)
30
20
10
10
20
30
yaxis (m)
Fig. 3.4 Circular orbital motion with height compensation.
44
Tilted motion can be achieved by varying the offset voltage U2 periodically, as
shown in Fig. 3.5. The resulting tilted motion allows tapered surfaces to be machined
without the need for tilting the work piece (or the spindle) or the use of a tapered end
mill. The tilt angle is adjustable by varying the magnitude of the secondary voltage while
the orientation of the tilt with respect to the z-axis is adjusted by the phase lag in the
offset voltages U2.
It is also possible to vary the offset voltage at twice the rate compared to the
voltage applied to x and y axes. This will produce a saddle-like orbit path, as shown in
zaxis (m)
Fig. 3.6.
6
5
4
3
2
1
0
1
2
3
4
5
6
Tilted circular orbital motion

Ux+ = U1sin(t) - U2
Uy+ = U1sin(t+/2) - U2
U2 = (500 V - U1)sin(t+)
Uy- = U1sin(t+3/2) - U2
U1 = 100 V
U1 = 200 V
U1 = 300 V
U1 = 400 V
U1 = 500 V
20
0
20
xaxis (m)
30
20
10
10
20
30
yaxis (m)
Fig. 3.5 Tilted circular orbital motion (Heamawatanachai and Bamberg [25]).
45
zaxis (m)
Curved circular orbital motion

Ux+ = U1sin(t) - U2
Uy+ = U1sin(t+/2) - U2
Uy- = U1sin(t+3/2) - U2
6
5
4
3
2
1
0
1
2
3
4
5
6
U1 = 100 V
U1 = 200 V
U1 = 300 V
U2 = (500 V - U1)sin(2t)
20
0
20
30
20
xaxis (m)
10
10
20
U1 = 400 V
U1 = 500 V
30
yaxis (m)
Fig. 3.6 Saddle-like orbital tool path (Heamawatanachai and Bamberg [25]).
3.1.3 Calculation of resonance frequency

For the transverse motion of a piezo tube, which usually has a length-to-diameter
ratio of less than 10, the thin beam equation based on Euler-Bernoulli is not suitable.
Instead, it is necessary to use Timoshenkos beam theory, which also considers the
rotational inertia of the cross section as well as shear deformation.
The approach closely follows the work by Huang [37], who modeled a beam
using Timoshenkos approach but without a lumped mass. Bruch and Mitchell [38] also
modeled a Timoshenko beam, but with an additional lumped mass attached to the free
end of the beam, which is assumed to be located exactly at the tip of the beam (x=L).
This assumption is not valid for the micromilling tool. Instead, an offset D is introduced,
as shown in Fig. 3.7 that allows the center of the attached mass to be located away from
the end of the tube.
46
Fig. 3.7 The piezo tube is modeled with a mass attached to the free end
The Timoshenkos beam equations of the free transverse vibration of a uniformcross-section cantilever beam with an attached mass at the free end, as shown in Fig. 3.7,
are represented in Eq. (3.7) and (3.8) [38].
2
y
2
EI 2 + KAG( ) I 2 = 0
t
x
x
2 y
2 y
)=0
KAG
t 2
x 2 x
(3.7)
(3.8)
where A is the cross-sectional area of the beam, E is Youngs modulus of the beam, G is
the shear modulus of elasticity, I is the area moment of inertia of the cross section of the
beam, K is a shear coefficient of the cross-section, is the density of the material of the
47
beam, y is the deflection of the center line of the beam and is slope of the beam due to
bending.
The shear coefficient K of the tubes cross-section can be calculated from
Hutchinson [39] as:
6(ri 2 + ro2 ) 2 (1 + ) 2
K= 4
7 ri + 34ri 2 ro2 + 7 ro4 + (12ri 4 + 48ri 2 ro2 + 12ro4 ) + 2 (4ri 4 + 16ri 2 ro2 + 4ro4 )
(3.9)
where is Poissons ratio, ri is the inner radius and ro is the outer radius of the tube. The
boundary conditions for the beam, as shown in Fig. 3.7, are taken from Abramovich [40]:
y = 0, = 0 at
x=0
y
2 y
2
) = m( 2 + D 2 ) at x = L
x
t
t
2
2 y
EI
= ( J + mD 2 ) 2 + mD 2 at x = L
x
t
t
KAG (
(3.10)
where L is the length of the beam, D is the distance from the end of the beam to the
center of the attached mass, J is the moment of inertia of the attached mass around its
center of mass and m is the attached mass. Note that the terms on the right of Eq. (3.10) at
x=L are the force and torque at the end of the beam due to the dynamics of the attached
mass. The nondimensional variables are defined as:
48
x
I
EI
m
1
, b 2 = AL4 2 , r 2 =
, s2 =
, M =
2
2
L
AL
AL
KAGL
EI
1 =
1 2(J + mD 2 )
, 2 =
L
m
2D
L
( 3.11)
where is the nondimensional length along the beam and is the frequency (rad/s). The
solution to the Timoshenkos equation with the nondimensional length can be represented
as [38]:
y = Y ( )e it
and = ( )e it
2 y
= 2Y ( )e it
2
t
and
( 3.12)
2
= 2 ( )e it
2
t
Note that:
y 1 y
=
,
x L
2 y 1 2 y
=
,
x 2 L 2
1
=
,
x L
2 1 2
=
x 2 L 2
By using the nondimensional variables in Eq. (3.12), Eq. (3.7) and Eq. (3.8) can be
rewritten as [38]:
2
1 Y
s
=0
(1 b 2 r 2 s 2 ) +
2
(3.13)
2Y
+ b 2 s 2Y L
= 0 for 0 < < 1
2
(3.14)
49
Eliminating and Y from Eq. (3.13) and (3.14) yields [38]:
2
4Y
2
2
2 Y
+
b
(
r
+
s
)
b 2 (1 b 2 r 2 s 2 )Y = 0
4
2
(3.15)
2
4
2
2
2
+ b (r + s ) 2 b 2 (1 b 2 r 2 s 2 ) = 0
4
(3.16)
The solutions to Eq. (3.15) and (3.16) are found from the roots of the differential
equations as follows [38]:
Y = C1 cosh b + C 2 sinh b + C3 cos b + C 4 sin b
(3.17)
= C1 sinh b + C 2 cosh b + C3 sin b + C 4 cos b
(3.18)
where [37]
(r 2 + s 2 ) + (r 2 s 2 ) 2 +
(r 2 + s 2 ) + (r 2 s 2 ) 2 +
4
b2
4
b2
(3.19)
Eq. (3.18) is applicable when r2s2b2 < 1. Substituting Eq. (3.17) and (3.18) into Eq. (3.14)
results in [38]:
50
b ( 2 + s 2 )
b ( 2 + s 2 )
C1 , C 2 =
C2
L
L
b ( 2 s 2 )
b ( 2 s 2 )
C3 =
C3 , C 4 =
C4
L
L
C1 =
(3.20)
Now, the boundary conditions can be written as:
= 0 at = 0
KAG Y
KAG
+ m 2Y + mD 2 = 0 at = 1
L
EI
2 ( J + mD 2 ) 2 mDY = 0 at = 1
L
Y = 0,
(3.21)
Substituting the nondimensional variables into Eq. (3.21) yields:
Y = 0,
= 0 at = 0
Mb 2 s 2 D
1 Y Mb 2 s 2
Y+
+
= 0 at = 1
L
L
L
(3.22)
b 2 M 22
b 2 M 12
Y = 0 at = 1
2
2D
Substituting Eq. (3.17), (3.18) and (3.20) into Eq. (3.22) results in [38]:
1
0 ( 2
R1
R1
0
+ s2 ) /
1
0
R2
R2
R3
R3
0
C1 0
2
( s ) / C 2 0
=
C 3 0
R4

R4
C 4 0
2
(3.23)
51
where
b
b ( 2 + s 2 )
2
sinh b + Mb 2 cosh b
R1 = + Mb D
L
2
2
b
b ( + s )
cosh b + Mb 2 sinh b
R2 = + Mb 2 D
L
b
b ( 2 s 2 )
sin b + Mb 2 cos b
R3 = Mb 2 D
2
2
b
b ( s )
cos b + Mb 2 sin b
R4 = + Mb 2 D
L
R1 =
M 22 b 2
M 12 b 2
b ( 2 + s 2 )
b cosh b
sinh b
cosh b
2
2D
L
R2 =
M 22 b 2
M 12 b 2
b ( 2 + s 2 )
b sinh b
cosh b
sinh b
2
2
L
D
R3 =
M 22 b 2
M 12 b 2
b ( 2 s 2 )
b cos b
sin b
cos b
L
2
2D
M 22 b 2
M 12 b 2
b ( 2 s 2 )
b sin b
R4 =
cos b
sin b
L
2
2D
(3.24)
(3.25)
The determinant of the coefficient matrix in Eq. (3.23) can be expressed as [38]:
2 + s2
( R3 R4 R3 R4 + R4 R1 R1 R4 )
F ( ) =

2 s2
( R2 R3 R2 R3 + R1 R2 R2 R1 ) = 0
+
(3.26)
52
The values of that satisfy Eq. (3.26) are the resonance frequencies. Using the
properties for the micromilling tool given in Table 3.1 and a numerical iteration method,
the first mode resonance frequency was found at 685 Hz and the higher modes are given
in Table 3.2. Note that the uncertainties of the numerical results in Table 3.2 are
calculated from the worst case possibility, which is obtained from the effects of the
known uncertainties of the parameters in Table 3.1.
Table 3.1 The properties of the piezo tube with attached mass.
Density
= 7500 750 kg/m3
Length
L = 0.0762 0.0003 m
Inside diameter
ID = 0.0139 0.0003 m
Outside diameter
OD = 0.0155 0.0003 m
Wall thickness
t = 7.784e-4 m
Moment of inertia
I = 9.8251e-10 0.6e-10m4
Youngs modulus E = 8.9e10 8.9e9 N/m2
Cross sectional area A = 3.6332e-5 8e-7 m2
Poissons ratio
= 0.33 0.04
Shear coefficient
Attached mass
m = 0.01858 0.002 kg
Mass moment of inertia

:
J=1.4713e-6 2e-7 kgm2
Distance of mass
D = 0.00788 0.0001 m
K = 0.5734 0.0077
Table 3.2 Resonance frequencies of the piezo tube with attached mass.
Resonance frequency mode #
Frequency (Hz)1
685+101
4864+606
11724+1382
19163+2331
27952+3610
Frequency is calculated by including the uncertainties of all relevant parameters listed in Table 3.1
53
3.2 Experimental results
For the experimental validation of the micromilling tool, the prototype was
mounted to a Profile 24P micromachine tool that is equipped with an x-y stage that is
driven by piezoelectric linear motors and encoders with a resolution of 100 nm. The
analog excitation voltages are generated with a 4-channel waveform generator (NI PCI7883R), which is controlled by LabVIEW. The voltages are amplified with four piezo
amplifiers (PI E-507.00) and the motion of the tool holder is measured with noncontact,
inductive probes whose output is amplified by two amplifiers (Lion Precision ECL-100)
and sampled with a data acquisition system (NI-PXI-6040E).
3.2.1 Measured resonance frequency
To validate the analytical resonance frequency, the excitation voltages for the xand y-electrodes were swept individually from 0 to 2500 Hz while recording the resulting
motion of the tool tip with the inductive probes in both directions. Fig. 3.8a shows the
response measured by the x- and y-axis probes for a swept excitation in the x-axis. The
highest peak at 704 Hz corresponds to the first mode in the x-direction. Fig. 3.8b shows
the response to a y-axis excitation and identifies the first mode in the y-axis as 680 Hz,
which is very close to the predicted resonance frequency of 685 Hz.
3.2.2 Static performance of the piezo tube
Piezo ceramic materials are known to experience hysteresis whereby an excitation

that moves the tube from point J to point K will require a very different excitation to
move the tube back to point J. Fig. 3.9 shows the position of the piezo tubes x-axis
before and after applying a 3 V step command input (300 V to the piezo tube). Following
the step voltage of 300 V at time A, the tubes response is highly nonlinear, and, even at
54
Fig. 3.8 The measured resonance frequency for excitation of the x-electrodes (a) and yelectrodes (b) (Heamawatanachai and Bamberg [25]).
time B, which is more than one hour later, the tube continues to creep and the
displacements continue to increase. At time B, the voltage is removed and the tube
responds in a similar, nonlinear fashion. Sixty minutes after the voltage removal, the
displacement of the tube is still 1.19 m, which is a clear indication of hysteresis. The
zoom views A and B also show the transient response of the piezo tube after changing the
applied voltage.
55
Test PZT Static Response X-axis
Position X-axis (microns)
20
15
10
0
0
50
100
Time (mins)
150
200
Fig. 3.9 Response of a piezo tube to a 300 V step (3 V control signal) to X-axis
3.2.3 Hysteresis
To investigate the hysteresis effect, a set of step command voltages was applied to
the x-axis, as seen in Fig. 3.10. The command voltage started at 0 V and was increased to
5 V in 1 V increments, and then reduced step by step to -5 V, and finally increased step
by step to 0 V. The time spent for each step was 20 seconds. Fig. 3.10 also shows the
actual positions of the x and y axes for each step command voltage. It is clearly seen that
there is a small cross talk between x and y axis, providing proof that the piezo tubes axes
and the position probes are not perfectly aligned. Fig. 3.11 shows the hysteresis effect of
the piezo tube by applying command voltage to the y axis.
56
Position of the PZT tool axes under static command voltage applied on x-axis
(with 20 seconds interval time)
40.0
30.0
Position x&y (microns)
20.0
-6
de
cre
as
o
ev
lta
10.0
Inc
0.0
-5
de
-4
a
cre
-3
se
In
vo
e
tag
a
cre
se
-2
vo
-1
rea
ge
o
ev
lta
ge
-10.0
g
lta
-20.0
e2
x-axis
-30.0
y-axis
-40.0
Applied command voltage to the x-axis (Volts)
Fig. 3.10 Hysteresis of the piezo tube tool from applying voltage to the X-axis
Position of the PZT tool axes under static command voltage applied on y-axis
(with 20 seconds interval time)
40.0
30.0
Position x&y (microns)
20.0
-6
dec
rea
ol
ev
tag
10.0
Inc
0.0
-5
-4
c
de
rea
se
lt
vo
Inc
-3 2
e
-2
-1
ag
rea
rea
e1
ol
ev
tag
e1
-10.0
o
ev
lta
ge
-20.0
x-axis
-30.0
y-axis
-40.0
Applied command voltage to the y-axis (Volts)
Fig. 3.11 Hysteresis of the piezo tube tool from applying voltage to the Y-axis
57
3.2.4 Long term stability
For dynamic applications, where the excitation voltages vary rapidly, hysteresis is
not an issue. This can be seen in Fig. 3.12, where the orbit radii and center drifts are
plotted over more than 60 minutes of running time. The orbit radii are essentially
constant over the entire time, and the drift of the orbital center is limited to less than 0.3
m. This measurement was performed with the inductive probes that are mounted to the
protective housing of the tool. As such, any potential change in shape of the housing due
to temperature fluctuations as well as the long-term stability of the probes itself are
included in these results.
Fig. 3.12 Long term stability: The orbit radii for various excitation voltages (a) and the
drift of the orbit center (b).
58
3.2.5 Frequency response
To establish the piezo tubes behavior to dynamic actuation, the response of the
piezo tube at various excitation voltages was recorded for frequencies ranging from 0 to
600 Hz. Fig. 3.13a shows the response to a symmetric voltage actuation of the x+ and xelectrodes and Fig. 3.13b demonstrates the response of the y+ and y- electrodes. For
frequencies up to 200 Hz, the response is close to the predicted tip displacements. At
higher frequencies, owing to inertia effects, the amplitude increases to much larger
values, especially as the excitation frequency approaches the first resonance frequency.
Fig. 3.13 Frequency response of x (a) and y electrodes (b).
59
3.3 Conclusions
The cutting motion of the tool can be generated by applying the sinusoidal signals
to the electrodes of the piezo tube. The circular trajectories can be generated by using
equal amplitudes with 90o phase difference between the signals to the x and y axes,
respectively. The size of the parasitic error motion due to bending of the piezo tube is
proportional to the orbit radius. Based on the calculations at an orbital radius of 30 m,
the z position of the tip moves up 6 nm compared to the z position without orbital motion.
Although this parasitic motion is small in micromachining and can be neglected, it can be
eliminated by applying an additional offset voltage to all electrodes.
By manipulating the amplitude and phase of the sinusoidal signals that are applied
to the electrodes of the piezo tube, more complex trajectories (x-y plane) such as
elliptical and linear tool paths can also be generated. The direction of the principle axis of
both these tool paths can be created at any direction along the x-y plane. This allows the
tool to be able to change its direction to match with the desired feed direction in actual
machining.
Tilted motion can be achieved by varying the offset voltage U2 periodically. The
resulting tilted motion allows tapered surfaces to be machined without the need for tilting
the work piece (or the spindle) or the use of a tapered end mill. The tilt angle is adjustable
by varying the magnitude of the secondary voltage while the orientation of the tilt with
respect to the z-axis is adjusted by the phase lag in the offset voltages U2
Timoshenkos beam theory, which also considers the rotational inertia of the cross
section as well as shear deformation, was used in this study to predict the resonance
60
frequency of the tool. The calculation results show that the first resonance frequency of
the tool is at 685 Hz, which is very close to the measured resonance frequency at 680 Hz.
The dynamic motion of the tool from sinusoidal signals that is used in orbital
machining minimized the errors from hysteresis and drift, which are the major errors in
control of piezo tube in static mode. Experiment results show that the tool has excellent
long term stability whereby the center position drifted less than 0.5 m within 70
minutes.
The amplitude and phase responses of the system vary with frequency and result
in errors of the tool trajectories. Misalignments, due to the manufacturing of the piezo
tube and assembly process, also affect tool trajectory errors. To minimize the effect of the
errors from frequency response and the misalignment from manufacturing, calibration is
required for every operating frequency.
CHAPTER 4
CUTTING FORCE MODEL
4.1 Introduction
The understanding of the cutting mechanism and machining forces is very
important in development of the machining tool. If the machining forces can be
estimated, a system with an appropriate amount of stiffness can be designed to achieve
the desired machining quality. The machining parameters such as feed rate, can also be
optimized to maximize the machining efficiency.
The orbital micromachining technique is a new technique where the cutting edge
is large compared to the uncut chip thickness, so the existing force models from other
machining techniques are not applicable. This study focused on the development of a
cutting force model for the orbital micromachining technique. The proposed model is
then validated through experiments using a 6-axis load cell as a force measuring device.
4.2 Cutting force models

The orthogonal cutting model shown in Fig. 4.1a is a very well established,
fundamental model of 2D ductile cutting that can be applied to many types of machine
operations such as turning, drilling and milling [41, 42]. The assumptions of this model
include that the tool has a perfectly sharp edge and that a continuous chip is produced
62
with no built-up edge [42]. The resultant of the cutting forces between the tool and work
piece can be expressed by either the forces applied to the tool face (NC and FC), the forces
applied on the shear plane (NS and FS) or the horizontal and vertical forces (FP and FQ).
The rake angle is the angle between the line perpendicular to the cutting direction and
the tool face and is shown as a positive angle in Fig. 4.1a. Merchant [43] presented an
orthogonal cutting force model that is derived from the resultant force between the tool
face and the chip, which is equal to the resultant force between the work piece and the
chip along the shear plane. This model is very useful in conventional machining to
predict the cutting forces related to the cutting parameters (rake angle, shear angle,
material, etc.). In general, the cutting force is estimated from the cutting area times the
specific cutting force K, which is obtained experimentally. From the literature, the
specific cutting force increases exponentially with smaller uncut chip thickness [42, 44].
In micromachining, the uncut chip thickness is small, making the effect of the
edge geometry quite important. In this case, the rake angle of the cutting condition varies
drastically with the uncut chip thickness and the edge radius. To account for the edge
radius, Son et al. [45] modified the orthogonal force model, as seen in Fig. 1b. Therefore,
the specific cutting force, which is related to the cutting geometry, is obtained from
experiments for each condition (depth of cut, tool geometry, etc.). Lee et al. [44]
presented the cutting force model for the micro-end mill by introducing the concept of the
partial effective rake angle in combination with experiments to obtain the specific cutting
energy.
From the literature on microcutting, there exists a minimum chip thickness below
which no chip will form [44-46]. In this case, the material below the minimum chip
63
thickness will deform elastically and will fully recover back to the original position after
the tool has passed. The material above the minimum chip thickness will be removed
through plastic deformation and the orthogonal cutting model can be applied. Son et al.
[45] found that the minimum uncut chip thickness is a function of the tool edge radius
and the friction coefficient between the tool and material.
Another interesting model that could be applied to orbital cutting is a grinding
force model. Typically, grinding has a very small uncut chip thickness relative to the
grain diameter combined with large negative angles, similar to orbital cutting. The
typically range of uncut chip thickness in grinding is about 0.5 to 1 m [47], which is
very small compared to conventional, precision milling where the range of uncut chip
thickness is around 6 to 15 m. The specific shear stress in grinding is closer to the
theoretical shear stress than in precision milling. In the past, the orthogonal cutting theory
was applied to predict cutting forces in grinding. However, whereas the model predicts an
energy transfer of 90% of the cutting energy into the chip, measurements show that about
80% of the energy is actually transferred into the work piece [48].
Shaw [49] proposed his grinding model that is derived from the hardness
indentation principle, as illustrated in Fig. 4.1c. The chip is formed by the upward flow of
the incline indentation of the tool on the work piece. The specific cutting energy also
increases exponentially with smaller uncut chip thickness.
Orbital micromachining is a new technique that is capable of creating 3D features
in a wide range of materials. Due to its unique 3D contact surface between the tool and
work piece, none of the existing cutting models are applicable. The uncut chip thickness
is very small compared to the tool diameter as a result of having a very large negative
64
rake angle. The geometry of the tool also affects the ratio between the uncut chip
thickness and the cross section radius. The tip geometry, which is used in this study,
consists of a conical and spherical section, resulting in variable tool radii along the height
of the tip. The uncut chip thickness varies from 0 to some maximum value and back to 0
at the tool position angle around 0, 90 and 180o, respectively. Because the radius changes
for each cross section, the minimum uncut chip thickness is different for each cross
section. This indicates that the material removal mechanism is located somewhere
between grinding and orthogonal precision machining. In order to derive a cutting model
for orbital cutting, the resultant machining force is assumed to consist of the normal force
FN and friction force Ff, as shown in Fig. 4.1d. The normal force FN is defined based on
the assumption that the contact pressure and hence the specific cutting force in the normal
direction KN are uniform across the contact area. The definition of the friction force Ff
also assumes that the specific friction cutting force in the tangential direction Kf is
constant across the contact area. For each cross-section of the tool tip, the orbital cutting
model can be presented as shown in Fig. 4.1d. The uncut chip thickness is small
compared to the cross-section radius, yielding a large negative rake angle. The normal
force is applied at the center of the contact length with the direction normal to the contact
area. The friction force is applied at the center of the contact length in the x-y plane with
the direction tangent to the contact curve.
65
FQ
NS
FP
FR
Chip
FS
Tool
Feed
direction
Feed
direction
Tool
Tool
to
te
Minimum uncut chip
thickness line
FC
Elastic deformation
Work piece
NC
(a) Orthogonal cutting model
to
Positive rake angle
Plastic deformation Chip
Negative rake angle
Work piece
piece
Work
(b) Orthogonal cutting model for large tool radii and/or small chip thickness
Feed
direction
Chip
Grain
FR
FQ
FP
to
Chip
FR
Elastic-plastic
boundary
Work piece
(c) Grinding force model
Feed
direction
Tool tip
F
N
Contact area (A)

FN = KNA Ff
Ff = KfA where Kf = KN
Work piece
= tan()
(d) Single-point, orbital cutting force model
Fig. 4.1 Overview of cutting models: (a) orthogonal cutting model (adapted from
Merchant [43]), (b) orthogonal cutting model for large tool radii and/or small
chip thickness (adapted from Son et al. [45]), (c) grinding force model
(adapted from Shaw [49]) and (d) single-point, orbital cutting
model (Heamawatanachai and Bamberg, in press).
4.3 Orbital single-point cutting force model
Fig. 4.2a and 4.2b represent two distinct examples of slot machining with the
same tool but at different depths of cut. In Fig. 4.2a, the contact area between the work
piece and the tool is created from the spherical section of the tool only, while Fig. 4.2b
shows a contact area that consists of a spherical and conical section if the depth of cut
exceeds the radius of the tool tip. The trace on the slots surface is the effect from the
previous loop of orbital motion of the tool. Fig. 4.2c shows the zoom view of the contact
area of Fig. 4.2a and also the tool geometry at the contact position. Fig. 4.2d depicts the
force components that are applied on the contact area during machining.
66
(a)
Depth of cut 1.5 m
Tip radius 3 m
Orbit radius 10 m
Feed per orbit 1 m
Cone angle 60
(c)
contact area from sphere section

not yet machined
Y
X
already machined
machined trace
(b)
Depth of cut 5 m
Tip radius 3 m
Orbit radius 10 m
Feed per orbit 1 m
Cone angle 60
contact area from sphere and cone section
(d)
Fy
not yet machined

Fx
Ff
FN
Y
X
X
machined trace
Fz
already machined
Fig. 4.2 Orbital machining of slot with (a) spherical contact area, (b) spherical and
conical contact area, (c) zoom view of contact area with tool tip, and (d) force
components at the contact area (Heamawatanachai and Bamberg, in press).
4.3.1 Principle of orbital cutting motion

The cutting force model of an orbital single-point tool tip that is actuated on
orbital tool paths exhibits significant differences to conventional, rotating tools. This
difference is illustrated in Fig. 4.3, which provides a top view of a conventional, rotating
tool such as a fly cutter (Fig. 4.3a) compared with an orbital tool motion of a single-point
tool tip (Fig. 4.3b). For the orbital tool motion shown in Fig. 4.3b, and neglecting the
effect of the tilt angle that is generated during the cyclic bending of the piezo tube, the
center of the tool tip is moving around the center axis on a circular trajectory. The tip
coordinate frame [XT, YT, ZT] does not change its direction during this motion. Instead, the
origin of the tip coordinate frame orbits around the origin of the center frame [XC, YC, ZC]
with radius R and angular speed .
67
YT
YC
Trajectory of tool tip center

(angular speed )
YC
XT
YT
1
0 4
YT
YT
4
1
0 3
2
XT
XT
2
0 1
2
0 1
XT
XC
2
0 1
XT
R
YT
2
0 1
XT XC
YT
Outer boundary
(a) Rotating cutting motion
(b) Orbital cutting motion
Fig. 4.3 Cutting motion of (a) conventional, rotating tools and (b) orbiting tools
(Heamawatanachai and Bamberg, in press).
The equations of motion of any point on the tool tip with respect to the center
coordinate frame are:
X i = R cos(t )+ T X i ,
C &
X i = R sin(t ),
C &&
X = 2 R cos(t ),
C
Yi = R sin(t )+ T Yi
C &
Yi = R cos(t )
C &&
Y = 2 R sin(t )
C
(4.1)
4.3.2 Calculation of the tool tip geometry
Fig. 4.4 shows the tool tip geometry that consists of a cone with a spherical tip.
For the cutting force model, the tip is divided into small disks of height z. The radius of
the cross section at height z can be calculated as:
68
ZT
c
tip center
rt
cone
sphere
z
ZC
XC
O
c
2
ri
ithdisk
zi
XT
P
Fig. 4.4 Front view of tool tip geometry (Heamawatanachai and Bamberg, in press).
2
2
rt (rt z )
r( z) =
rt cos c + z rt sin c tan c
2
2
2
for 0 z rt 1 sin c
2
for z > rt 1 sin c

2
(spherical tip)
(conical section)
(4.2)
where c is the cone angle and rt is the tip radius. The z position of the ith disk can be
calculated from zi = (2i-1)z/2.
4.3.3 Calculation of the machining contact area
From Fig. 4.5, the tool tip is orbiting around the center coordinate frame [XC, YC,
ZC] with angular velocity . The center coordinate frame moves in the YW direction with
feed speed Vf. The unmachined boundaries from two successive loops are indicated as
dashed circles while the machined boundaries are shown as solid circles. The
instantaneous length of cut around the tool tip L(z,) is defined by points P1 and P2.
69
Unmachined
YC,0
Feed speed Vf
YT
L
L
P2
P1
XT
r
P
Orbital direction
Angular velocity
Tool
Machined
XC,0
O
Current loop
Previous loop
YW
XW
Fig. 4.5 Cross section i of tool tip at height zi during orbital machining
For a given angle , the position of the tip origin P() with respect to the center
coordinate frame of the previous loop (C,0) is found as:
C ,0
R cos
P( ) = R sin + V f T
(4.3)
where Vf is feed speed, R is the orbit radius and the period T is found as T = 2/. Point
P1 is the contact point of the tool with the current machined boundary and its position is
found as:
70
r cos
P1 = r sin
z
(4.4)
The position angle of the tip origin with respect to the work piece coordinate frame is:
R sin + V f T
R cos
= tan 1
(4.5)
The position of point P2 with respect to the center frame coordinate system of the
previous loop is:
R( R + 2r ) + OP 2
( R + r ) cos( + )
C ,0
P2 = ( R + r ) sin( + ) where = cos 1
2( R + r ) OP
z
(4.6)
This point can be transformed into the tool coordinate system [XT, YT, ZT] using a
transformation matrix:
RT C ,0
1
0
=
0
0
1
0
0
R cos
0
0 R sin V f T
1
0
0
1
Point P2 can then be expressed as:
(4.7)
71
( R + r ) cos( + ) R cos
P2
P2 ( R + r ) sin( + ) R sin V f T
=
R

=
T C ,0
z
1
1
C ,0
(4.8)
The tool engagement angle L is found as:
P
P
1 2
1
L = cos 1
where P1P2 = ( P1, x P2, x ) 2 + ( P1, y P2, y ) 2
2
2r
(4.9)
and the tool engagement length is L = rL. To determine the force that is applied to the ith
disk at angle , it is necessary to determine the area of each disk as Ai = Lis, where the
arc length s of the disk is:
s = z
s =
for 0 z i rt 1 sin c
2
rt
ri
z
cos
for
z > rt 1 sin c
2
(4.10)
v
The normal vector to the contact surface N i in the work piece coordinate system [XW, YW,
ZW] is:
72
v
Ni
v
Ni
ri cos + 2
= ri sin + L for 0 z i rt 1 sin c

2
2
z i rt
ri cos + 2
= ri sin + L for z i > rt 1 sin c
2
2
ri tan c
2
(4.11)
4.3.4 Calculation of cutting forces
The cutting forces applied to the contact area consist of the normal and the
friction force. The normal force is calculated based on the assumption of constant contact
v
pressure over the contact area. Therefore, the normal force vector FNi on the ith disk is
v
parallel to the normal vector N i and:
v
v
Ni
FNi = v Ai K N
Ni
(4.12)
The friction force is calculated based on the assumption of constant friction coefficient
over the contact area and the direction of force is in the x-y plane and perpendicular to
v
the normal vector. Hence, the friction force F fi on the ith disk is:
v
v
v
v
v
( N i,x + N i, y ) N i
fi
F fi = v
v
v Ai K N = v Ai K f
( N i,x + N i, y ) N i
fi
(4.13)
73
v
v
where f i is a vector parallel to the friction force vector Ffi . The forces in x, y and z
direction are obtained from the combination of components of friction force and the
normal force in each direction as:
v
v
v
N i,x
f i,x
Fi , x = v Ai K N + v Ai K f
Ni
fi
v
v
v
N i, y
f i, y
Fi , y = v Ai K N + v Ai K f
Ni
fi
v
v
N i,z
Fi , z = v Ai K N
Ni
(4.14)
Ac is the total contact area. ANx, ANy and ANz are the directional area components for the
calculation of the normal force in the x, y and z directions, respectively. Afx and Afy are the
directional area components for the calculation of the frictional force in x and y
directions, respectively. The area decomposition illustrated in Fig. 4.6 can be written as:
ANx =
i =1
n
A fx =
i =1
Ai N i , x
v ,
Ni
Ai f i , x
v ,
fi
ANy =
i =1
A fy =
i =1
Ai N i , y
v ,
Ni
Ai f i , y
v ,
fi
ANz =
i =1
Ai N i , z
v
Ni
(4.15)
Ac = Ai
i =1
Define n as the total number of the disks contacted with the work piece as n = t0/z. The
total cutting force on the tip is the summation of force from all disks (Fig. 4.7).
74
Directional Area Components

8
Ac
ANx
ANy
ANz
Afx
Afy
ANi
Afi
(m2)
Depth of cut: 5.5 m

Orbit radius: 10 m
Tip radius: 1 m
Cone angle: 50 deg.
Feed per orbit: 0.1 m
0
2
4
0
30
60
90
120
Orbit angle (deg.)
150
180
Fig. 4.6 Directional area components of the contact area (Heamawatanachai and
Bamberg, in press).
n A f
n Ai N i , x
i i,x
v
v
f i K N
i =1
Fx i =1 N i
= n AN
n A f

i
i, y
i i, y K f
Fy
v
v
f i
i =1
i =1 N i
n AN
{Fz } = i v i , z {K N }
i =1 N i
(4.16)
75
Calculated Force from KN and Kf Equations
0.04
Fx
Fy
Fz
0.03
Force (N)
0.02
Depth of cut: 5.5 m

Orbit radius: 10 m
Tip radius: 1 m
Cone angle: 50 deg.
0.01
0
0.01
0.02
0.03
30
60
90
120
Orbit angle (deg.)
150
180
Fig. 4.7 Calculated cutting force components based on KN and Kf equations

4.4 Model verification

4.4.1 Experimental setup
The micro-orbital tool tip is actuated with a piezo tube scanner that allows
circular trajectory motion to be generated with a frequency of up to 2 kHz and orbital
radii that can be varied between 0 and 30 m [25]. The piezo tube tool was mounted to
the z-axis of a CNC machining center (Haas TM-1) in order to generate the feeds for
machining. The machining forces were measured with a miniature 6-axis load cell (ATI
Nano 17) that was mounted onto a x-y alignment stage to level the top surface of the
work piece (Fig. 4.8). The position of the tool tip relative to the tool housing was
measured with two inductive probes (Lion Precision UC3). All three coordinate systems
(CNC, tool tip and load cell) were aligned to each other. To synchronize the tool tip
position with the machining force measurement, the inductive probes ( x and y ) and the
76
ZCNC
YCNC
XCNC
Piezo tube housing
Inductive probe Y
ZTool
Inductive probe X
Tool tip holder
YTool
XTool
Diamond tip
Alignment screw
Work piece
Load cell (ATI Nano 17)
ZLc
YLc
XLc
x-y alignment stage
Machine vise
Fig. 4.8 The micro-orbital machining setup (Heamawatanachai and Bamberg, in press).
load cell reading signals were recorded with a single data acquisition card (NI-PXI6040E). The sampling frequency was set at 40 kHz, which yielded 400 data points per
cycle at an orbital frequency of 100 Hz.
The tool tip was a single crystal diamond with a cone angle of 50 and tip radius
of 1 m (Gilmore Diamond Tools). Highly polished aluminum alloy (AL2024) was
selected as the work piece material because it is largely free of unpaired d-shell electrons,
making it a diamond machinable material [50]. The tool was orbited counter clockwise
and moved in the y-direction to create a slot. The actual positions of the tool tip and the
force signals from the load cell were recorded simultaneously in order to study the
machining forces for each position of the tool during machining.
77
4.4.2 Measurement data processing
The micro-orbital control program was developed using LabVIEW and

LabVIEW-FPGA. The measurement of the tool position and the forces from the load cell
were recorded at 40 kHz. Fig. 4.9(a-e) shows an example of the raw measurement data
for a slot machined with 100 Hz orbital frequency, 5.5 m depth of cut, 10 m orbital
radius and 10 m/s feed rate. The data filtered with a low pass filter at 500 Hz are shown
in Fig. 4.9(f-j).
(a) Force X
(f) Force X (filtered)
(b) Force Y
(g) Force Y (filtered)
(c) Force Z
(h) Force Z (filtered)
(d) Tip X
(i) Tip X (filtered)
(e) Tip Y
(j) Tip Y (filtered)
Fig. 4.9 Measurement data for orbital cut (100 Hz, 5.5 m depth, 10 m orbit radius, 10
m/s feed rate) showing measured forces in (a) x , (b) y, (c) z direction, (d) tool position
in x and (e) in y. Also shown are the filtered data for the forces and tool position (f)
through (j) (Heamawatanachai and Bamberg, in press).
78
The filtered force data shown in Fig. 4.8(f-h) were compared with the tool tip xposition (Fig. 4.8i) to identify the orbit angle of the tool position. Each cycle of the force
data was used to calculate the specific cutting forces KN and Kf by modifying Eq. (4.16)
as:
n Ai N i , x
v
K N 1 N i
= n AN
i
i, y
K f
v
1 N i
Af
1 i fv i, x F
i
x
n A f
Fy
i i, y
1 fv
i
n
(4.17)
4.4.3 Calculation result
Fig. 4.10 shows the force range as a function of depth of cut based on 100
measurement cycles for each depth of cut for an Al2024 work piece machined with a 1
m diamond tip and a 50 cone angle actuated at 100 Hz with 10 m orbit radius and 0.1
m feed per orbit. The range of force is determined from the difference between
maximum and minimum force for each cycle.
Force Range vs. Depth of Cut
Force (N)
0.2
Rx = Fx,max - Fx,min
Material: AL2024
Ry = Fy,max - Fy,min
Tip radius: 1 m
Rz = Fz,max - Fz,min
0.15
Cone angle: 50 deg.
Orbit radius: 10 m
0.1 Orbit freq.: 100 Hz
Ry
Rz
0.05
0
Rx
5
10
Depth of cut (m)
15
Fig. 4.10 Range of measured forces vs. depth of cut (100 Hz, 10 m orbit radius, 0.1 m
feed per orbit) (Heamawatanachai and Bamberg, in press).
79
From the literature, the specific cutting force is inversely proportional to the uncut
chip thickness to the power of n (K~1/tn), where n~0.2 for conventional metal cutting and
n~0.5-1 for grinding. In this study, the specific cutting forces KN and Kf are assumed to be
inversely proportional to the contact area power of n (K~1/An). Therefore, the relationship
of the specific cutting force and the contact area can be written as:
ln( K ) = C n ln( A)
(4.18)
where C and n are constants. Using Eq. (4.18) and based on 2 million data points from
500 measurement cycles for depths of cut varying between 3.5 to 12.5 m, the specific
cutting forces KN and Kf were calculated and are illustrated in Fig. 4.11. Also shown are
the curve fits based on the least square technique. The n values for the KN and Kf
equations are 0.34 and 0.41, respectively, which are between the typical values of 0.2 and
0.5 for orthogonal cutting and grinding. The maximum error of the fitted KN values
compared to the average experimental data ranges from -11 to +70%, while the maximum
fit error for Kf, was found to range from -15 to +82%.
The resultant force that is applied to any section of the contact area consists of the
normal force and the frictional force. To simplify the model, the friction force is defined
as a force in the x-y plane that is perpendicular to the normal force. The coefficient of
friction can be then calculated from = Kf/KN or from curve fitting of the raw data. As
seen in Fig. 4.12, the friction coefficient from both methods yields a value of about 0.45.
ln(KN), ln(Kf) (ln(Pa))
80
Specific Cutting Forces ln(KN) and ln(Kf) vs. Contact Area

26
Material: AL2024
ln(KN)
Tip radius: 1 m
ln(Kf)
25 Cone angle: 50 deg.
Orbit radius: 10 m
Orbit freq.: 100 Hz
ln(KN) = 22.899 - 0.338ln(A)
24
23
22
21
ln(Kf) = 22.184 - 0.409ln(A)
20
6
8
Area (m2)
10
12
14
Fig. 4.11 Experimental data and curve fits for the specific cutting forces KN and Kf
Based on the specific cutting forces KN and Kf, the normal and frictional forces
were calculated and decomposed into their x-, y- and z-components using Eq. (4.16).
Fig. 4.13 presents a comparison of the predicted and measured machining forces based on
a depth of cut of 5.5 m as a function of the orbit angle. The predicted forces generally
match the measured data, although at a slightly lower level. The model is based on the
assumption that contact between the tool and work piece only occurs for orbit angles
81
Friction Coefficient vs. Contact Area

1
0.9
Friction coefficient
from raw Kf/KN

from fitted Kf/KN
fitted from raw Kf/KN
Material: AL2024
Tip radius: 1 m
Cone angle: 50 deg.
Orbit radius: 10 m
Orbit freq.: 100 Hz
0.8
0.7
0.6
0.5
0.4
0.3
10
12
14
Area (m2)
Fig. 4.12 Friction coefficient in orbital machining (Heamawatanachai and Bamberg,
in press).
between 0 and 180 degrees. The experiments, however, indicated that the tool is engaged
for more than 180 degrees, most likely due to chip buildup. This buildup also increases
the effective depth of cut, which is a likely reason for the fact that the measured forces
are roughly 50% higher than predicted.
As shown in Fig. 4.14a, based on the specific cutting forces KN and Kf equations,
the effect of the depth of cut on the range of forces is calculated. Based on a very small
tip radius of 1 m, depths of cut deeper than 1 m are largely performed with the conical
82
Predicted and Measured Cutting Forces (5.5 m depth of cut)
Fx (N)
0.05
0.025
0
0.025
0.05
90
180
270
360
450
540
630
Orbit angle (deg.)
720
810
900
990
1080
90
180
270
360
450
540
630
Orbit angle (deg.)
720
810
900
990
1080
90
180
270
360
450
540
630
Orbit angle (deg.)
720
810
900
990
1080
Fy (N)
0.05
0.025
0
0.025
0.05
Fz (N)
0.05
0.025
0
0.025
0.05
Material: Al2024 Cone angle: 50 deg. Orbit freq.: 100 Hz

Tip radius: 1 m Orbit radius: 10 m Feed per orbit: 0.1 m
Predicted force
Measured force
Fig. 4.13 Comparison between the predicted and measured cutting forces
section of the tool tip, which explains the linear relationship between depth of cut and
cutting forces.
Fig. 4.14b represents the effect of the tip radius on the range of the cutting forces.
The contact area increases with larger tool tip radii, which results in larger machining
forces. For the cutting conditions used in the model, the cutting force in the z-direction
will become dominant for tip radii greater than 20 m.
Fig. 4.14c shows the effect of the orbit radius on the range of cutting forces. As
can be seen, the ranges of forces increase for very small orbit radii and converge for
increasing values. This supports the idea that the orbital motion helps in material removal
by reducing the machining forces. In the case that the orbital radius is zero, which would
83
constitute a material removal mechanism similar to scratching, the contact forces will be
very high, thereby increasing the risk of tool breakage.
The effect of the feed per orbit on the cutting forces is illustrated in Fig. 4.14d.
The slope of the forces are high for small feed rates and then gradually taper off for larger
feed rates. Due to the increasing contact area with increasing feed rates, the range of
forces also increase.
4.5 Conclusions
The cutting force model developed in this study is applicable to single-point tool
tips that are orbited about their center axis. The model takes into consideration the
complex 3D shape of the contact area between the work piece and the tool tip, which
consists of a spherical tip followed by a conical section. The cross section of the tool can
be made circular, which is stronger compared with other shapes at comparable size.
Therefore, the size of the effective tool can be made as small as a few microns.
The machining force that is applied to the contact surface between the tool and the
work piece consists of a normal and a friction force. The force model is developed based
on a uniform contact pressure over the contact area. Hence, the specific normal cutting
force KN and the specific friction cutting force Kf are uniform along the contact area. The
results from the force model calculation are a close match with the results from the
experiments based on a ductile material (AL2024), although the magnitudes of the
calculation results are smaller than the measurement results. This is due to material builtup during machining, which increases the contact area.
84
Force (N)
Force (N)
Force (N)
Force (N)
Predicted Range of Force vs. Depth of Cut

0.1
0.08
0.06
0.04
0.02
0
0.12
0.1
0.08
0.06
0.04
0.02
0
0.06
0.05
0.04
0.03
0.02
0.01
0
0.06
0.05
0.04
0.03
0.02
0.01
0
(a) Tip radius: 1 m

Orbit radius: 10 m
5 6 7 8 9 10 11 12 13
Depth of cut (m)
Predicted Range of Force vs. Tip Radius
(b) Depth of cut: 5.5 m

Orbit radius: 10 m
10
20
30
40
Tip radius (m)
Predicted Range of Force vs. Orbit Radius
50
(c) Depth of cut: 5.5 m

Tip radius: 1 m
10
20
30
40
Orbit radius (m)
Predicted Range of Force vs. Feed per Orbit
50
(d) Depth of cut: 5.5 m

Tip radius: 1 m
Orbi radius: 10 m
0.05
0.1
0.15
0.2
Feed per orbit (m)
Material: Al2024
Cone angle: 50 deg.
Orbit freq.: 100 Hz
0.25
0.3
Rx = Fx,max - Fx,min
Ry = Fy,max - Fy,min
Rz = Fz,max - Fz,min
Fig. 4.14 Cutting force predictions based on (a) depth of cut, (b) tip radius, (c) orbit
radius and (d) feed per orbit (Heamawatanachai and Bamberg, in press). Note that
Rx, Ry and Rz are the ranges of forces in x, y and z directions, respectively.
85
Machining parameters such as material, depth of cut, tool geometry, orbit radius,
orbit frequency, feed rate, orbital trajectory, lubrication, etc. are important to the
machining force results. Examples of the effects of some parameters such as depth of cut,
orbit radius, tool radius and feed per orbit are shown in the calculation results. The tool
tip in this study is made from a single crystal diamond, which is the hardest commercially
material available. In reality, however, diffusion of carbon needs to be taken into
consideration. Therefore, the diamond tip can be used to machine materials such as
aluminum, brass, silicon, etc. Materials with low carbon content or with significant
numbers of unpaired d-shell electrons (e.g. steels, titanium, etc.) require a tool tip from
another material such as cubic boron nitride (CBN).
CHAPTER 5
ORBITAL MICROMACHINING OF DUCTILE MATERIALS
In this chapter, experiments were performed to investigate the ability of the

orbital micromachining technique of ductile materials. Firstly, 2D and 3D features were
machined on various materials (copper, titanium alloy and stainless steel) to verify that
the diamond tool tip with orbital motion can create microfeatures on the ductile materials.
Secondly, a more detailed investigation on the effect of selected machining parameters on
the machining results was studied. Aluminum alloy (Al2024) and titanium alloy (Ti-6Al4V) were used because of their wide use in engineering applications that require strength
and durability. The machined surface roughness was measured with an optical surface
profiler (Zygo NewView). Tool wear from machining was also observed by examining
the tool tip before and after machining.
5.1 Machining examples on ductile materials

By using G-Code, the tool paths for complex geometry microfeatures can be
created. Fig. 5.1 to Fig. 5.4 show the machined 2D and 3D features in copper, titanium
and stainless steel, indicating that the tool is capable of machining complex structures in
various ductile materials. In these experiments, a micro-EDM machine (Optimation
Profile 24P) was used for the macroscopic motion. Fig. 5.1 shows a pyramid feature on
87
Fig. 5.1 Pyramid (100 x 100 m) in copper.
Fig. 5.2 Pyramid (100 x 100 m) in titanium alloy (Ti-6Al-4V)
88
Fig. 5.3 Pyramid (100 x 100 m) in stainless steel (316)
Fig. 5.4 Various microfeatures machined in 316 stainless steel.
89
copper with a surface roughness (Ra) of 0.12 m. In Fig. 5.2 and 5.3, the pyramid feature
was created on a titanium alloy and stainless steel work piece, respectively. The
machined surfaces roughness (Ra) was 0.48 m and 0.08 m, respectively. Fig. 5.4
shows microfeatures in 316 stainless steel. The rather lengthy and complex GCode
required for this feature was created in Pro/MANUFACTURE based on a 3D model
created in Pro/ENGINEER.
From Fig. 5.1 through 5.3, the shapes of the top surface of the pyramid were
rectangular rather than square. This indicates that the tool tip was not symmetric and the
widths of the tip at the same depth along the x and y directions were not equal. This is
caused from a nonsymmetric tool tip due to manufacturing errors or damage of the tip
during machining. In Fig. 5.4, it is clearly seen that the shape of the top surface of the
pyramid is square, indicating that the tool tip used in this experiment had equal cross
sectional radii in the x and y directions. To remove the chip during machining,
pressurized air was used to blow the chip out from the cutting zone (no lubrication was
used in these experiments). The work pieces were cleaned by brushing the features with a
cotton tip soaked with acetone. The machining results of copper (Fig. 5.1) indicate that
materials of extreme ductility may produce features with significant burr buildup.
5.2 Machining of microchannels into ductile material

In the previous topic, the machining results show that a tool tip made of single
crystal diamond can be used to machine 3D features in ductile materials. However, the
achieved surface qualities, particularly the buildup of burrs along machined edges,
needed to be investigated further. There are many factors that affect the machined surface
90
quality, including tip geometry, tool cutting motion, type of materials, lubrication, etc.
This topic focuses on the ability of orbital machining to create microchannels in ductile
materials such as Al2024 and Ti-6Al-4V.
An example of the burr formation of ductile materials is shown in Fig. 5.5. A
single pass slot was machined in aluminum alloy (Al2024) with 12.5 m depth of cut, 10
m orbit radius, 100 Hz orbit frequency and 10 m/s feed rate. A CNC machine (HaasTM-1) was used instead of a micro-EDM machine to generate the macroscopic feed
motion because it has a much greater stiffness. No lubrication or pressurized air was used
in this experiment. The picture was taken right after machining without any cleaning.
From observation of Fig. 5.5, the chip formation can be separated into 3 groups: 1) the
small and weak-bond burr on the top side wall, 2) the large and strong-bond burr on the
bottom side wall, and 3) the weak bond of the removed particles that still stick on the
bottom surface. Analyzing the image, the following hypotheses can be made. Because the
orbital direction of the tool tip is counter clockwise (see Fig. 5.5), the cutting motion of
the first group (top side wall) is very similar to up-milling motion, which gives better
surface finish. The chip of the first group can be removed easily by brushing with a
cotton tip after machining. The cutting motion of the second group (bottom side wall) is
similar to climb-milling. The chip in this group forms a stronger bond with the side wall
and is difficult to be removed. The second group can be avoided by always using the upmilling principle for the finish cut. The third group is hard to be removed after machining
because it is inside a very small groove, and therefore very hard to reach. The third group
can be avoided by using a flushing system during machining such as using pressurized air
and/or lubrication.
91
Orbital micro machining of slot feature on Al2024

- 12.5 m depth of cut
- 50o cone angle (tool tip)
- 1 m tip radius
- 10 m orbit radius
- 100 Hz orbit frequency Group #1
Feed direction
(10 m/s)
Orbit direction
(100 Hz)
Group #3
Group #2
Fig. 5.5 Microslot machined in Al2024 without lubrication or flushing. The picture was
taken after machining without any cleaning.
More experiments with orbital micromachining in combination with a CNC

machine were performed to obtain a better understanding of the effects of machining
parameters on machining results. From observation using a microscope, the CNC
machine produces a significant jump (10 20 m) when reversing direction of its axes.
Compared to the typical dimensions of microfeatures, this error is quite large and can
cause significant problems during machining.
92
To minimize the effect of this error, tool paths were created that avoid reversal of
the direction while the tool is in contact with the work piece material. Fig. 5.6 shows a
portion of a microslot in Al 2024 with 15 m depth, 137 m width and 1 mm length that
was created with orbital micromachining. The tool feed direction indicates the
macroscopic motion from the CNC to machine the work piece material. At the upper end
of the feature, the tool was moved upwards to create a gap distance of 20 m between the
top surface of the work piece and the tool tip. Next, the tool tip was moved back to the
lower end of the feature and was also stepped over by 5 m before starting the next cut.
The microscopic motion of the diamond tool tip (3.5 m tip radius and 50o cone angle)
was set at the 100 Hz orbit frequency, 10 m orbit radius and counter clockwise orbit
direction, as seen in Fig. 5.6. To reduce chip buildup along the edges, a small amount of
lubrication (BOELUBE) was applied on the top surface of the work piece before
machining. The machined surface quality between the left and right side walls are
different due to the orbital direction of the tool that sweeps the chip material from right to
left. The surface roughness of the feature along the feed direction was measured as 0.05
m with an optical surface profiler (Zygo NewView).
Fig. 5.7 shows a portion of a tapered groove machined with the orbital
micromachining technique. The tool tip geometry was 3.5 m tip radius and 50o cone
angle. The tool path was created to machine 7 layers with 10 m depth of cut for each
layer, and the width of the tool path was reduced by 10 m for every subsequent layer. In
this fashion, a tapered groove with 30o taper angle and 70 m depth was created. A small
amount of lubrication was applied before machining each layer. From Fig. 5.7, the side
wall surface on the right is very smooth, indicating that this technique is capable of
93
Orbital micro machining of slot feature on Al2024

- 1 mm length, 100 m width --- tool path
- 15 m depth of cut
- Total 21 cutting paths
- 3.5 m tip radius
- 10 m orbit radius
15 m depth
- 100 Hz orbit frequency
- lubrication oil (BOELUBE)
- Ra=0.05 m
(along feed direction)
137.28 m
Feed direction
(10 m/s)
Step over direction
(5 m each step)
(Total of 21 paths)
Orbit direction
(100 Hz)
Orbital machining AL2024
Fig. 5.6 Micro-slot on Al2024
creating a deep microfeature with a smooth surface finish. There are still lots of chip
particles remaining on the left side wall because of the orbital direction that sweeps the
chip from right to left. The bottom surface roughness along the feed direction is about
0.05 m. It is interesting to note that such a deep feature can be created with this
technique, although the side wall taper angle needs to be larger than half of the tool tip
cone angle.
94
Orbital micro machining of taper groove on Al2024-T4
- 10 m depth of cut for each cutting layer totally of 7 layers
- 100 m width & 1 mm length --- tool path of first layer (top layer)
- 40 m width & 0.97 mm length --- tool path of the 7th layer (bottom layer)
- 3.5 m tip radius
- 10 m orbit radius
- 10 m/s feed speed
64.93 m
143.36 m
70 m depth
(approx.)
Step over direction
(5 m each step)
Feed direction
(10 m/s)
Depth of each layer
(10 m)
Orbit direction
(100 Hz)
Fig. 5.7 Micro-taper-groove on Al 2024 with a depth of 70 m.
From the experiment shown in Fig. 5.5, the burr on the side wall can be avoided if
the up-milling principle is used for the finish cut. To show that the quality of the side wall
depends on the orbital direction of the tool tip, another micro-taper-groove was created.
The right side wall of the feature was machined with counter clockwise orbit direction
and the left side wall was machined with clockwise orbit direction, resulting in up-milling
of both side walls. As seen in Fig 5.8, at the half width of each layer, the orbital direction
changed from counter clockwise to clockwise. In this fashion, microfeatures with smooth
side walls can be manufactured with the orbital micromachining technique.
95
Orbital micro machining of taper groove on Al2024-T4

- 10 m depth of cut for each cutting layer totally of 7 layers
- 100 m width & 1 mm length --- tool path of first layer (top layer)
- 40 m width & 0.97 mm length --- tool path of the 7th layer (bottom layer)
- 3.5 m tip radius
- 10 m orbit radius
69.43 m
140.58 m
70 m depth
(approx.)
Feed direction
(10 m/s)
Feed direction
(10 m/s)
Depth of each layer

(10 m)
Orbit direction
(CW 100 Hz)
Orbit direction
(CCW 100 Hz)
Step over direction

(5 m each step)
Step over direction

(5 m each step)
Fig. 5.8 Micro-taper-groove with smooth side walls (left and right) on Al 2024
Fig. 5.9 shows the ability of orbital micromachining to create complex microchannels. Microchannels with different depths (30 m, 50 m and 70 m depth) were
created with connections between each other. The right features with 70 m also indicate
that this technique can be used to create deep features (such as a deep hole) if the side
wall taper angle is larger than half the cone angle of the tool tip.
Fig. 5.10 shows a microfeature in Ti-6Al-4V with 5 m depth and a width
around 137 m. The feed speed was set at 5 m/s, the orbit radius was 10 m at 100 Hz
96
70 m
- 10 m depth of cut, 5 m step over

- 3.5 m tip radius
- 10 m orbit radius
- 10 m/s feed speed
38 m
Complex micro channel
30 m depth
(approx.)
143.36 m
38.82 m
70 m depth
(approx.)
71.75 m
128.63 m
67.38 m
63.44 m
89.69 m
50 m depth
(approx.)
132.14 m
Fig. 5.9 Microchannels combined with 6 features and 3 different depths

(30, 50, and 70 m) in Al 2024.
and the orbit direction changed from CCW to CW at the half way point of the width, with
lubrication applied prior to machining. The machined surface result (bottom left of
feature in Fig. 5.10) indicates that the tool tip was damaged during machining. This
shows that a single crystal diamond tip can machine titanium but tends to suffer from a
significant amount of tool wear. For titanium, in order to avoid excessive cost from tool
wear, the tool tip should be made from another, more appropriate material instead of
diamond.
97
Orbital micro machining on Titanium
- 100 m length, 100 m width --- tool path
- 50o cone angle (tool tip), 3.5 m tip radius
- 100 Hz orbit frequency, 10 m orbit radius
137.28 m
Feed direction
(5 m/s)
131.32 m
Feed direction
(5 m/s)
Orbit direction
(CW 100 Hz)
Step over direction

(5 m each step)
Orbit direction
(CCW 100 Hz)
Step over direction

(5 m each step)
(Total of 21 paths)
Orbital machining Titanium
Fig. 5.10 Microfeature in Ti-6Al-4V
5.3 Investigation of tool wear

From the report of Paul et al. [50], the major wear of the single point diamond
turning is from diffusion due to chemical aspects between work piece material and the
carbon atoms of diamond. Their hypothesis is that the material with unpaired d-shell
electrons will cause tool wear in diamond turning. In their report, for example, aluminum
has no unpaired d-shell electron, and therefore can be machined with diamond without
excessive tool wear. Titanium, on the other hand, has two unpaired d-shell electrons that
cause increased tool wear when it is machined with the diamond tool. Table 5.1 shows
98
Table 5.1 Data for elements with known diamond turning properties
(adapted from Paul et al., [50])
Element
Zn (Zinc)
Mg (Magnesium)
Al (Aluminum)
Ge (Germanium)
Ag (Silver)
Au (Gold)
Cu (Copper)
Mn (Manganese)
Si (Silicon)
Ni (Nickel)
Fe (Iron)
Ti (Titanium)
Cr (Chromium)
V (Vanadium)
Mo (Molybdenum)
W (Tungsten)
Melting
point, oC
420
649
660
937
962
1064
1083
1244
1410
1453
1535
1660
1857
1890
2617
3410
Microhardness,
kg/mm2
51
48
25
721
96
96
76
384
1211
189
142
250
248
192
348
No. of unpaired
d-shell electrons
0
0
0
0
0
0
0
5
0
2
4
2
5
3
5
4
Diamond turnable?,
Yes/No
Y
Y
Y
Y
Y
Y
Y
N
Y
N
N
N
N
N
N
N
data of the elements with known diamond turning properties from the work of Paul et al.
[50].
Tables 5.2 and 5.3 show the components of the aluminum alloy (Al 2024) and the
titanium alloy (Ti-6Al-4V), respectively. The major component in Al 2024 is aluminum,
which is known as a diamond turnable material. The major component in Ti-6Al-4V is
titanium, which is not a diamond turnable material.
Table 5.2 Components (% by weight) of aluminum alloy (Al 2024-T4) [63]

Al90.794.7
CrMax0.1
Cu3.84.9
FeMax0.5
Mg1.21.8
Mn0.30.9
SiMax0.5
TiMax0.15
ZnMax0.25
Other,eachMax0.05
Other,totalMax0.15
99
Table 5.3 Components (% by weight) of titanium alloy (Ti-6Al-4V) [63]
Ti90
Al6
V4
FeMax0.25
OMax0.2
Fig. 5.11 shows the comparison pictures of the diamond tool tip before and after
machining on the aluminum alloy (Al 2024). The total volume of machining is 25.7e6
m3. Comparing Fig. 5.11a and 5.11b, no tool wear of the tool tip can be detected,
indicating that diamond is a suitable material to machine the aluminum alloy. These
results, of course, agree with the findings of Paul et al. Although, the composition in Al
2024 contains some elements that are not diamond turnable, the majority of the elements,
namely aluminum (90.7-94.7%), are diamond turnable. Hence, the effect of unpaired dshell electrons from the nondiamond turnable elements is minimized.
a).Before machining on Al2024
b).After machining on Al2024

- 25.7x106 m3 machining volume
R=3.45 m
R=3.45 m
200 m
200 m
Fig. 5.11 Comparison of tool tip before and after machining of 35e6 m3 of Al 2024
100
Fig. 5.12 shows the comparison pictures of the diamond tool tip before and after
machining the titanium alloy (Ti-6Al-4V). The total volume of machining is 59.7e3 m3,
which is only 0.2% of the total machining volume on Al 2024. The wear of the tool tip is
clearly seen, indicating that the diamond tool tip is not suitable for machining of titanium
alloy. These results agree with Table 5.1 and 5.3.
5.4 Conclusions
From the experiments, it is clearly seen that the orbital micromachining technique
can be used to create 2D/3D microfeatures in ductile materials such as copper, stainless
steel, titanium alloy and aluminum alloy. The major problem of machining of ductile
material is the burr surface because the chip particles still stick to the machined feature.
(a)Before machining on Ti-6Al-4V
R=3.45 m
100 m
(b)After machining on Ti-6Al-4V

- 59.7x103 m3 machining volumn
tip wear
100 m
Fig. 5.12 Comparison of tool tip before and after machining of 59.7e3 m3 on Ti-6Al-4V
101
The burr surface can be avoided by using a proper flushing system (such as
pressurized air, flushing water and lubrication) and using the up-milling principle for the
finish cut. This is confirmed by using the piezo tube tool with lubrication (BOELUBE) to
machine slots and tapered groove features in aluminum alloy (Al 2024), which yielded a
very smooth surface finish (Ra) of 0.05 m. The groove with a very smooth side wall
was achieved with the up-milling principle and orbital motion. Deeper microfeatures can
be created from the orbital micromachining technique as well. However, due to the
geometry of the tool tip, the taper angle of the side wall needs to be larger than half the
cone angle of the tool tip.
The tool tip of single crystal diamond is the hardest commercially available
material, which can be used to machine any weaker material. However, the rates of tool
wear depend on the chemical composition of the work piece. The diamond tip can
machine aluminum alloy (Al 2024) with no measurable wear. On titanium alloy (Ti-6Al4V), however, the diamond tool tip exhibits a noticeable wear of 5 - 10 microns due to
diffusion of carbon atoms from the diamond into the titanium alloy.
CHAPTER 6
ORBITAL MICROMACHINING OF BRITTLE MATERIALS
In this chapter, experiments were performed to investigate the ability of the

orbital micromachining technique on brittle materials. First, 2D and 3D features were
machined on various materials (glass, silicon and germanium) to verify that the diamond
tool tip with orbital motion only (no rotating motion) can create microfeatures on the
brittle material.
Second, a more detailed investigation on the effects of selected machining
parameters (such as orbital trajectory, tool geometry, depth of cut and lubrication) on the
machining results was studied. Single crystal silicon was used because it is widely used
in microfabrication and the microfluidic industry. The machined surface roughness was
measured with an optical surface profiler (Zygo NewView). The tool position during
machining was measured by two inductive probes. The machining forces were measured
with a 6-axis load cell. Tool wear from machining was also observed by examining the
tool tip before and after machining.
103
6.1 Introduction
Products made from semiconductor materials such as silicon that require optical
quality surfaces are typically achieved by processes such as grinding, lapping and
polishing (Uddin et al. [52]). Other mechanical techniques such as single point diamond
turning and milling with proper machining parameters are also capable of producing
surfaces with a mirror finish. In the literature, this type of machining is generally referred
to as ductile regime machining that requires the thickness of the uncut chip to be very
small. Patten et al. [53] found that the depth where the ductile-to-brittle transition
typically occurs is between 10 - 100 nm, which is considerably less compared to Chao et
al. [59], who reported critical depth of cut values between 0.2 and 1.0 m for diamond
turning depending on the cutting direction and crystal orientation of silicon. The
difference between the two findings is most likely tied to the difference in hydrostatic
pressure during cutting. This effect was studied by Yoshino et al. [54], who performed
machining of highly brittle materials (soda glass, quartz glass, silicon and quartz) using a
single point tool while applying external hydrostatic pressure to the material. The
machining samples were placed inside a pressure vessel and subjected to a pressure
between 0 and 400 MPa. The results showed that the critical depth of cut (ductile-brittle
transition) increases with larger external hydrostatic pressures.
Another way to increase the hydrostatic pressure on the cutting region is to
perform cutting with negative rake angles. With a large negative rake angle, the critical
depth of cut where the work piece material deforms plastically is increased. Liu and Li
[56] performed experiments on ductile cutting of tungsten carbide using a CBN tool. The
tool with 5.8 m edge radius and negative rake angle of -32 degrees was used to make
104
tapered scratches. The scratches showed ductile mode cutting at the beginning (small
depths of cut) and changed to brittle mode cutting for larger depths of cut. The critical
depth of cut was reported as 2.485 m. Sreejith [57] studied the ductile machining of
silicon nitride using a poly-crystalline diamond (PCD) tool with rake angles of 0, -5, -10,
-15 and -20 degrees. The paper reports that the high negative rake angle provided
hydrostatic pressure that resulted in plastic deformation of the work piece. Fang and
Venkatesh [58] performed experiments by making tapered cuts on a single crystal silicon
sample with a diamond tool and rake angles that ranged from 0 to -25. They found that
as long as the depth of cut is less than 236 nm, ductile mode cutting is achieved.
Yan et al. [55] investigated the effect of tool edge radius on ductile machining of
silicon by using finite element analysis. The Johnson-Cook material model, which
normally is used in metals, was implemented in their study in order to model the ductile
behavior of silicon. From their results, the volume of material under high pressure
increased with larger cutting edge radii. Venkatachalam et al. [60] used single point
diamond turning to machine silicon in order to study the ductile-brittle transition depth of
cut. In this study, silicon was modeled with the Johnson-Cook model. They proposed a
model to predict ductile-brittle mode of machining, which was based on comparing the
stress intensity factor of machining with the fracture toughness of silicon.
The orbital single-point micromachining used in this study is a new mechanical
machining technique that is based on micro-orbital motion of a single point tool tip made
from diamond, which Paul et al. [50] identified as a suitable tool material for silicon due
to chemical aspects. Based on the cutting force model of this tool at small feed rates per
orbit (Heamawatanachai and Bamberg, in press), the uncut chip thickness for each cross-
105
section is very small compared to the tools radius. Therefore, a large negative rake angle
of cutting is obtained that increases the hydrostatic pressure on the cutting area, thereby
allowing brittle materials such as silicon to be machined in the ductile regime.
6.2 Influence of hydrostatic pressure on the ductile

behavior of brittle materials
Generally, engineering materials can be classified based on the stress-strain
relationship as either ductile or brittle. Under applied stress, ductile materials such as
metals will deform elastically while the stress is increased up until the yield strength.
Increasing the stress further will cause plastic deformation until the material breaks apart
upon reaching the fracture strength. Brittle materials such as ceramics and semiconductor materials will only deform elastically and not show any plastic deformation
because the fracture strength is smaller than the yield strength, allowing stresses to reach
the fracture strength before the yield strength, and as a result, break apart. In design
engineering, the yield stress is used as criteria for structural failure of ductile materials,
and the fracture strength is used for design of brittle materials.
It is well known that brittle materials are stronger in compression than in tension,
and that the fracture strength in the compressive mode is larger than in the tensile mode.
This is supported by Dowling [62] who reports that brittle materials can exhibit ductility
when tested under high hydrostatic pressure. The fracture strength increases with
increasing hydrostatic pressure, whereas the yield strength is constant based on the
octahedral shear stress yield criterion. Fig. 6.1 shows the yield and fracture strength
boundaries of a material along with the hydrostatic pressure line of brittle and ductile
106
materials, which is adapted from the report of Dowling [62]. The hydrostatic pressure
line is the line that has equal stress for all three principal stresses in the material. The
rectangular region represents the yield strength boundary and the parabolic region shows
the fracture strength boundary.
From Fig 6.1a, the brittle material displays brittle behavior if the hydrostatic
pressure is in the tensile or low compressive mode, as illustrated by points a and b. At
point a, the octahedral shear stress is zero and the hydrostatic pressure is low
compressive. While maintaining the hydrostatic pressure, increasing the octahedral shear
stress will make the point move upwards to reach the fracture boundary at point b.
Therefore, the brittle material with tensile hydrostatic pressure or low compressive
pressure will reach the fracture boundary before reaching the yield boundary, thereby
displaying brittle behavior. Conversely, a brittle material can also display ductile
behavior if the hydrostatic pressure is compressively high, as shown by points c, d and e.
The octahedral shear stress at point c is zero and the hydrostatic pressure is in the brittle
region. If the octahedral shear stress is increased, the point will move upwards from point
c and reach the yield boundary at point d. Therefore, the material will deform plastically
until the octahedral shear stress reaches the fracture boundary at point e where the
material brakes apart. Fig 6.1b shows the yield-fracture boundary of a ductile material.
The ductile region of the ductile material covers widely from all compressive hydrostatic
pressure to slightly high tensile hydrostatic pressure, which make these materials yield
before fracture for almost all of the general conditions.
107
Fracture
Hydrostatic line
Yield
Fracture
H (compressive)
Yield boundary
H (tensile)
Ductile region
Brittle region
Fracture boundary
(a) Yield-fracture boundary (brittle material)

Hydrostatic line
Fracture
Yield boundary
Fracture
H (compressive)
Yield
H (tensile)
Ductile region
Brittle region
Yield
Fracture boundary
(b) Yield-fracture boundary (ductile material)
Fig. 6.1 Relationship of yielding and fracture boundary for brittle material
(adapted from Dowling [62]).
6.3 Examples of machining 2D/3D microfeatures on brittle materials
To demonstrate the machining ability of the tool on brittle materials, the cutting
module was mounted on a 3-axis micro-EDM machine (Optimation Profile 24P) that is
equipped with an x-y stage driven by piezoelectric linear motors and encoders with a
resolution of 100 nm. Three different materials were machined. These were glass, single
crystal silicon and single crystal germanium. The results show that the orbital tool can
machine microfeatures on all these materials.
Fig. 6.2 shows an example of a 2.5-D feature with outer dimensions of 450 x 450
m and 15 m depth on glass. This feature was machined at a feed rate of 5 m/s, a depth
of cut of 5 m and a 10 m step over between passes. The surface roughness (Ra) of
0.184 m was measured by an optical surface profiler (Zygo NewView).
108
Fig. 6.2 3D view of a microfluidic feature machined into glass taken with an optical
surface profiler (Zygo NewView). The surface roughness was measured as 0.184 m.
Fig. 6.3 and Fig. 6.4 show machined 3D pyramid features in single crystal silicon
and single crystal germanium, respectively. The images prove that the tool is capable of
machining complex structures into brittle materials. The machined surface roughness
(Ra) on single crystal silicon is 1.42 m and on single crystal germanium is 0.68 m.
From Fig. 6.3 and 6.4, the shapes of the top surface of the pyramid were
rectangular instead of square. This indicates that the tool tip was not symmetric and the
widths of the tip at the same depth along the x and y directions were not equal. This was
caused either during grinding of the tool tip or perhaps it could be the result from tip
damage during machining. To remove the chip during machining, high pressurized air
was used to blow the chips out from the cutting zone (no lubrication was used in these
experiments). The work pieces were cleaned by brushing the features with a cotton tip
soaked with acetone. Fig 6.3 and 6.4 also show that there are a large number of voids on
the machined surface. This indicates that brittle mode cutting is a major problem of
micromachining of brittle materials.
109
Fig. 6.3 Pyramid feature in single-crystal silicon.
Fig. 6.4 Pyramid feature in single-crystal germanium
110
6.4 The effect of orbital motion
To study the effect of orbital motion, the orbital cutting module was integrated
into a CNC machine (HAAS-VF-E). The orbital cutting module was used to generate
circular microscopic motion of the diamond tip and the CNC was used to generate the
macroscopic feed motion. As seen in Fig. 6.5, two slot features were machined into a
single crystal silicon work piece with and without orbital motion. Feature 1 was
machined with orbital motion of 10 m orbit radius, 300 Hz orbital frequency, 0.1
mm/min feed speed and 2 m depth of cut. Feature 2 was machined with 0.1 mm/min
Fig. 6.5 The effect of orbital motion on the machining quality
111
feed speed and 2 mm depth of cut without orbital motion. A small amount of lubrication
(BOELUBE) was applied on the work piece before machining. From Fig. 6.5, without
orbital motion, the tool performs cutting similar to scratching and the machined feature
exhibited a large number of defects at the surface. Based on the surface characteristic, it
is obvious that the material was removed primarily in the brittle mode. Feature 1, on the
other hand, shows that most of the area of this feature was machined in the ductile
regime. This experiment proves that the orbital tool tip motion has a significant effect on
the surface quality of machined silicon.
6.5 The effect of lubrication

The effect of lubrication on the machined quality was studied by comparing the
machined single-pass slot features as seen in Fig. 6.6. The slot in Fig. 6.6a was machined
with 0.6 mm/min feed, 300 Hz orbit frequency, 10 m orbit radius and 2-2.5 m depth of
Fig. 6.6 Slot machined (a) without and (b) with lubrication (BOELUBE).
112
cut without applying any lubrication. The slot in Fig. 6.6b was machined using similar
settings (2.5-3 m depth instead of 2-2.5 m) but with a small amount of lubrication
(BOELUBE) applied before machining. A single crystal diamond tip with 60o cone angle
was used in this study. In case of machining without lubrication, the feature is machined
mostly in brittle mode, although there are some spots that were machined in ductile
mode. With lubrication, the machine surface quality improved significantly, as seen in
Fig. 6.6b. The majority of the machined surface is smooth and free of voids, indicating
that the material was primarily removed in the ductile mode. The side wall surface is very
smooth, indicating that it was machined in the ductile mode. Fig. 6.6b shows that there
are three different surface depths on the feature, which indicates that the tip geometry
was not a perfect sphere but had multiple peaks at the end of the tip. Fig. 6.6b also
indicates that the diamond tip was not damaged during the machining because the side
walls of the feature are unchanged along the cutting distance.
6.6 The effect of depth of cut

To investigate the effect of the depth of cut on the machining quality, slots were
machined with a diamond tip (60o cone angle) by varying the depth of cut ranging from 1
to 10 m, as seen in Fig. 6.7. A feed speed of 0.1 mm/min, orbit radius of 10 m and
orbit frequency of 300 Hz was used for all slots and a small amount of lubrication
(BOELUBE) was applied on the top surface of the work piece before machining. From
Fig. 6.7, most of the material was removed in the ductile mode, only some minor areas
show brittle mode machining.
113
Fig. 6.7 Slot machined into single-crystal silicon with (a) 1 m, (b) 3 m, (c) 5 m and
(d) 10 m depth of cut.
The tool tip used for these experiments was not a perfect sphere but consisted of
multiple peaks, probably due to tip breakage or perhaps poor tip manufacturing outcome.
Evidence of these multiple peaks can be found by comparing the bottom surface of Fig.
6.7b through Fig. 6.7d, which essentially look identical. The bottom surface of the feature
with 10 m depth of cut has a similar surface quality as all other slots. The side wall at
the top surface, however, exhibits numerous defects. This is most likely caused by
114
insufficient rigidity of the piezo tube tool, which allows deviations from the ideal tip
trajectory as a result of increased machining forces.
From these initial experiments, it can be concluded that for depths of cut up to 5
m, relatively smooth surface finishes that are largely free of voids can be achieved.
Increasing the depth of cut to 10 m, however, produces very poor side walls. Some of
these effects can probably be attributed to the fact that the tool tip used for the
experiments consisted of multiple peaks, which by itself resulted in machining at various
depths of cut even within a single pass.
6.7 Tool trajectories during machining

Fig. 6.8 shows SEM images at 180x of five slots that were machined with an orbital
frequency of 300 Hz and orbital radius of 10 m, which results in a cutting speed of
18.85 mm/s (=1.13 m/min). The slots were machined with a single pass and depths of
cuts of 1, 2, 5, 10 and 15 m, respectively. Lubrication was provided by applying small
amounts of BOELUBE before the cut.
As can be seen in Fig. 6.8, depths of cuts of up to 2 m result in well defined
feature outlines that are free from breakage. Increasing the depth of cut to 5 m produces
slightly rougher edges. Increasing the depth to 10 m and more, on the other hand, results
in considerable damage to the edges. Fig. 6.8a through Fig. 6.8d show that the majority of
the bottom surfaces are free of defects, indicating that the material was removed in the
ductile regime. There are, however, a few spots visible where breakouts indicate brittlemode cutting as well. For the 15 m deep cut shown in Fig. 6.8e, the majority of the
115
Fig. 6.8 Slots with (a) 1, (b) 2, (c) 5, (d) 10 and (e) 15 m depth of cut machined into
single-crystal silicon at 300Hz with 10 m orbital radius (Heamawatanachai and
Bamberg [25]).
material was removed in the brittle mode. The roughness of the bottom surfaces was
measured with an optical profiler (Zygo NewView), as shown in Fig. 6.9, and varied
between 0.03 and 0.076 m (Ra) for all slots with a depth of cut of 10 m and less. The
15 m depth of cut resulted in a noticeable rougher surface of 0.15 0.23 m (Ra).
During machining of the slots, the orbital tool trajectories were measured with the
noncontact probes and the data recorded throughout the cut. The trajectory of 300 orbits
sampled at 100 kHz is displayed in Fig. 6.10 together with the feed and orbital direction.
The shape of the trajectory for depths up to 5 m is identical, indicating that the piezo
tube tool is capable of maintaining the desired tool path without position feedback. At 10
and 15 m depth of cut, the trajectories exhibit an increased variability in the x-direction,
116
Fig. 6.9 Optical profile of slot machined into single-crystal silicon with a depth of cut of
2 m at 300 Hz, an orbital radius of 10 m and a feed rate of 0.1 mm/min
0 m depth
b.)
5
Feed
0
5
Orbit direction
5
Feed
0
5
10
Yaxis (m)
10
Orbit direction
e.)
Feed
0
5
Orbit direction
10
5
0
5
Xaxis (m)
10 m depth
5
0
5
Xaxis (m)
f.)
5
Feed
0
5
Material: single-crystal silicon

Orbital frequency: 300 Hz
Orbital radius: 10 mm
Feed rate: 0.1 mm/min
Sample rate: 100 kHz
Sample time: 1s
5
0
5
Xaxis (m)
10
15 m depth
10
10
10
10
Orbit direction
10
10
Orbit direction
10
10
Yaxis (m)
d.)
5
0
5
Xaxis (m)
5 m depth
Feed
10
10
10
5
Yaxis (m)
10
c.)
10
Yaxis (m)
10
Yaxis( m)
Yaxis( m)
a.)
2 m depth
1 m depth
10
5
Feed
0
5
Orbit direction
10
10
5
0
5
Xaxis (m)
10
10
5
0
5
Xaxis (m)
10
y
g.) Slot
Feed
Work piece
Orbit direction
Fig. 6.10 Orbital trajectories recorded at depth of cuts of (a) 0, (b) 1, (c) 2, (d) 5,(e) 10
and (f) 15 m during (g) milling of a slot into single-crystal silicon (Heamawatanachai
and Bamberg [25]).
117
which is also the direction of the feed. The average orbit radius, however, is maintained,
which is also true for the y-direction of the tool. From Fig. 6.10 it can be concluded that
the piezo tube tool easily can machine features into single-crystal with a depth of cut of
up to 5 m at a feed rate of 0.1 mm/min without sacrificing part accuracy.
6.8 Improving surface finish with a finishing cut

To improve the surface finish and to avoid any brittle mode cutting, a slot that
was machined with an initial depth of cut of 2 m was machined with a second pass
(finishing cut) without changing the z-position of the machine. Instead, the length of the
piezo tube was increased by 0.23 m by reducing the center voltage from -500 to -480 V.
This effectively lowered the tool tip by 0.23 m for the second cut. The SEM image at
6200x reveals a very smooth surface finish without any breakouts (Fig. 6.11). The surface
roughness was measured as 0.03 0.049 m.
6.9 Machining forces, surface quality and tool wear

In order to gain a better understanding of the effects of the orbital machining
parameters on the machining forces and machined quality, another set of experiments
was performed. The piezo tube tool was mounted to the z-axis of a CNC machine (Haas
TM-1) in order to generate the feed motion. A single crystal silicon work piece was
mounted on a miniature 6-axis load cell (ATI Nano 17) to measure the machining forces.
The base of load cell was mounted onto a x-y alignment stage to level the top surface of
the work piece. All coordinate systems of the tool tip, load cell and the CNC were aligned
to the same direction, as seen in Fig. 6.12. To synchronize the tool tip position with the
118
Machined surfaces (2-2.5 m depth)
Polished surface
20 m
Fig. 6.11 Slot machined in silicon in two steps: first pass with 2 m depth of cut followed
by a second pass with 0.23 m depth of cut (Heamawatanachai and Bamberg [25]).
machining forces, the inductive probes (x and y) and the load cell reading signals were
measured by one data acquisition card (NI-PXI-6040E) at the sampling rates of 40 kHz.
The tool tip in this study was made of single crystal diamond that consists of a
50 conical section and a spherical tip section with radii ranging from 4 to 15 m.
Experiments were performed to measure the machining forces in x, y and z axes. The
machining quality was analyzed using SEM and the surface roughness was measured
with an optical surface profiler (Zygo NewView).
119
Orbital micro machining setup

CNC Coordinate system
Y
X
CNC Mounting plate
PZT Housing
PZT Coordinate system

Inductive Probe X
Tool tip holder
Z
Y
X
Inductive Probe Y
Diamond tip
Loadcell Coordinate
Work piece
Loadcell
Alignment Stage
Y
X
Alignment screw
Fig. 6.12 The micro-orbital machining setup
A set of slot cutting experiments were performed in order to investigate the

machining performance of this tool on silicon work piece which is a very brittle material
and very hard to achieve ductile machining with mechanical machining technique. The
machining parameters such as depth of cut, tool tip radius, orbital frequency and feed
speed were varied for each experiment as shown in Table. 6.1. A small amount of
lubrication (BOELUBE) was applied before each machining experiment.
120
Table. 6.1 Machining parameters for experiments
Experiment #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Tip # and
radius (m)
#1, 4.7
#1, 4.7
#1, 4.7
#1, 4.7
#1, 4.7
#1, 4.7
#2, 7.4
#2, 7.4
#2, 7.4
#2, 7.4
#2, 7.4
#2, 7.4
#3, 14.6
#3, 14.6
#3, 14.6
#3, 14.6
#3, 14.6
#3, 14.6
#4, 4.1
#4, 4.1
#4, 4.1
#4, 4.1
#4, 4.1
#4, 4.1
#4, 4.1
#4, 4.1
#4, 4.1
Depth of cut
(m)*
4.4
4.5
5.1
2.4
3.2
4.3
2.5
3.4
4.2
1.7
2.8
3.5
2.8
3.9
5.0
2.5
3.5
4.4
1.6
2.6
3.5
0.9
1.9
2.6
0.1
1.3
2.2
Orbital
frequency
(Hz)
100
100
100
300
300
300
100
100
100
300
300
300
100
100
100
300
300
300
100
100
100
100
100
100
100
100
100
Feed speed
(m/s)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
3
3
3
Orbit radius
(m)
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
* Depth of cut was measured after experiment using optical surface profiler (Zygo
NewView)
121
6.9.1 Measuring of the cutting force
The forces in x, y and z axes were measured during machining using a load cell
(ATI Nano 17) via data acquisition (NI-PXI-6040E) with a sampling rate of 40 kHz. Fig.
6.13 shows a set of orbital machining forces from experiment #1 filtered with a 500 Hz
low pass filter that consists of 10 machining cycles. The machining parameters used in
this experiment are 100 Hz orbit frequency, 4.4 m depth of cut, 10 m orbit radius and 1
m/s feed rate. The range of force that is calculated from the difference between the
maximum and minimum values is a good way to compare forces of all axes (x, y and z).
For each experiment, based on 10 sets of measurement data (100 cycles for 100 Hz and
300 cycles for 300 Hz) of machining forces, the mean and standard deviations of the
range of forces (in x, y and z directions) are calculated as shown in Table. 6.2.
(a) Force X (filtered)
(b) Force Y (filtered)
(c) Force Z (filtered)
Fig. 6.13 Measurement data for orbital cut of the experiment #1 (100 Hz, 4.4 m depth,
10 m orbit radius, 1 m/s feed rate) showing measured forces (filtered) in (a) x , (b) y,
(c) z direction.
122
Table. 6.2 Range of forces and standard deviation of the measured machining forces in x,
y and z directions during machining
Experiment #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Range Fx
0.0298
0.0361
0.0356
0.0082
0.0133
0.0181
0.0262
0.0373
0.0509
0.0086
0.0153
0.0255
0.0601
0.1217
0.1092
0.0612
0.1159
0.0815
0.0072
0.0111
0.0202
0.0053
0.0103
0.0163
0.0040
0.0088
0.0123
Measured range of forces (N)

Sd Fx
Range Fy
Sd Fy
Range Fz
0.0085
0.0346
0.0108
0.0488
0.0112
0.0346
0.0127
0.0486
0.0128
0.0376
0.0124
0.0475
0.0040
0.0107
0.0049
0.0246
0.0056
0.0131
0.0055
0.0254
0.0060
0.0164
0.0069
0.0302
0.0104
0.0190
0.0103
0.0541
0.0141
0.0306
0.0109
0.0689
0.0148
0.0348
0.0130
0.0769
0.0045
0.0100
0.0047
0.0307
0.0057
0.0163
0.0077
0.0389
0.0086
0.0218
0.0057
0.0490
0.0166
0.0361
0.0097
0.2137
0.0253
0.0603
0.0172
0.2702
0.0176
0.0629
0.0214
0.2327
0.0223
0.0330
0.0080
0.1832
0.0192
0.0538
0.0117
0.2578
0.0168
0.0567
0.0121
0.2290
0.0040
0.0074
0.0044
0.0319
0.0054
0.0124
0.0072
0.0319
0.0065
0.0177
0.0067
0.0351
0.0026
0.0060
0.0033
0.0349
0.0052
0.0100
0.0052
0.0350
0.0065
0.0159
0.0071
0.0433
0.0018
0.0035
0.0011
0.0429
0.0043
0.0091
0.0050
0.0431
0.0047
0.0131
0.0062
0.0385
-Note: Range Fx = Rx=Fx,max-Fx,min, Range Fy =Ry=Fy,max-Fy,min,

Range Fz =Rz= Fz,max-Fz,min
Sd Fz
0.0182
0.0196
0.0163
0.0117
0.0125
0.0127
0.0214
0.0233
0.0247
0.0159
0.0161
0.0142
0.0551
0.0617
0.0678
0.0539
0.0510
0.0483
0.0177
0.0168
0.0159
0.0192
0.0167
0.0195
0.0172
0.0200
0.0190
123
Fig. 6.14 shows the range of force from selected experiments with 3 different tip
radii (4.7 m, 7.4 m and 14.6 m). For an orbital frequency of 100 Hz (Fig. 6.14a), the
plot indicates that larger depths of cut result in an increase of the range of forces for all
axes and tip radii. The plot also shows that larger tip radii result in a larger range of force,
because the contact area during machining from a larger tip radius is greater than the
contact area from a small tool tip radius.
Fig. 6.14b shows the effect of the tip radius on the machining forces for the
selected data with 300 Hz orbit frequency. Similar to the results at 100 Hz, the plot
indicates that the range of forces increases with the tip radius. It is interesting to note that
the range of forces at 300 Hz orbit frequency is slightly lower than the results of 100 Hz
orbit frequency. However, even though increasing the orbital frequency by a factor of 3
effectively reduces the feed per orbit by a factor of 3 as well, the reduction in machining
forces does not scale by the same amount. This is the result of the contact area, which,
due to its complex shape, does not scale linearly with the feed per orbit.
Effect of tool tip radius and orbital frequency

0.35
4.7 m
7.4 m
0.3
14.6 m
Range (N)
0.25
Range Ry
0.35
R = 10 m
Vf = 1 m/s
0.3
0.35
0.3
0.3
0.15
0.15
0.15
0.1
0.1
0.1
0.1
0.05
0.05
0.05
0.05
0.15
0.15
0.15
0.1
0.1
0.05
0.05
2
0.3
0.2
0.2
R = 10 m
Vf = 1 m/s
0.2
0.2
4
6
Depth (m)
14.6 m
0.3
0.2
0.2
7.4 m
Range Rz
0.35
0.25
0.25
4
6
Depth (m)
(a) 100 Hz orbital frequency
0.25
4.7 m
Range Ry
0.35
0.25
0.25
Range Rx
Range Rz
0.35
Range (N)
Range Rx
0
2
4
6
Depth (m)
2
4
Depth ( m)
2
4
Depth ( m)
2
4
Depth ( m)
(b) 300 Hz orbital frequency
Fig. 6.14 The effect of tip radius on the machining forces at (a) 100 Hz and (b) 300 Hz
orbital frequency.
124
Fig. 6.15 shows the effect of feed speed on the machining forces. The results in
this plot were obtained from the experiments with a 4.1 m tip radius. The cutting forces
increase slightly with the feed speed. Similar to the discussion of Fig. 6.14, as the feed
speed is increased by a factor of 3, the machining contact area will increase to a smaller
extend, resulting in an increase of machining forces that does not scale linearly with the
feed speed.
6.9.2 Machined surface quality

Fig. 6.16 shows SEM images of the machined surfaces from the experiments with
a 4.1 m tip radius. Fig. 6.16a and 6.16b show that a few of the areas of the features were
Effect of feed speed

Range Rx
0.07
1 m/s
2 m/s
0.06
3 m/s
Range (N)
0.05
Range Ry
0.07
Range Rz
0.07
f = 100 Hz
R = 10 m
rtip= 4.1 m
0.06
0.05
0.06
0.05
0.04
0.04
0.04
0.03
0.03
0.03
0.02
0.02
0.02
0.01
0.01
0.01
2
4
Depth (m)
2
4
Depth (m)
2
4
Depth (m)
Fig. 6.15 The effect of feed speed to the machining forces
125
Fig. 6.16 SEM images of machined surfaces with (a) 1.6 m, (b) 0.9 m
and (c, d) 0.1 m depth of cut.
machined in the ductile mode while the majority of the features were machined in brittle
mode. The tool tip used for these experiments consisted of a single peak, as seen by the
lack of steps that were clearly visible in earlier experiments (Fig. 6.7). Reducing the
depth of cut to 0.1 m finally yielded machined surfaces that were completely free of
cracks and voids (Fig. 6.16c and 6.26d). The measured surface roughness of 0.007 m
(Ra) is a very good indicator that the material was entirely removed in ductile mode.
126
6.9.3 Tool wear
Tool wear was investigated by examining images of the tool tip taken with an
optical microscope at 20X magnification before and after machining. Fig. 6.17 shows
pictures of the tool tip #1, #2 and #3 before and after machining. Each tip was used to
machine slot features using the cutting parameters as listed in Table 6.1 with a total
length of cut of 2 mm. Fig. 6.17 shows that none of the tips were damaged during
machining. This indicates that a diamond tip and the orbital micromachining technique
can be used to machine single crystal silicon without wear from diffusion. This result
agrees with the report from Paul et al. [50], which identifies diamond as a suitable
material to machine silicon due to chemical aspects that avoid diffusion of carbon atoms
from the diamond into the work piece material.
Fig. 6.17 The tool tip before and after machining
127
6.10 Conclusions
The machining experiments performed with materials that included glass,
germanium and silicon prove that complex 2D/3D features at the microlevel can be
machined with the orbital micromachining technique. In the orbital micromachining
technique, the uncut chip thickness is small compared to the tip radius and results in large
negative rake angles. The large negative rake angle is a major factor in generating a
compressive hydrostatic pressure around the contact area between the tool and the work
piece. This pressure facilitates the ductile mode machining of brittle materials such as
silicon.
Comparing features that were machined with and without orbital motion clearly
showed that the tool motion greatly improves the quality of the machined surfaces.
Another important factor identified experimentally is the depth of cut. While early
experiments with an irregular shaped diamond tip produced machining results that
suggested possible depths of cut of more than 5 m, later results with properly shaped
diamond tips concluded that true ductile machining requires the depth of cut to be around
0.1 m. Also important is lubrication that, when combined with the orbital tip motion,
creates a centrifugal force that facilitates the removal of chips from the cutting zone,
thereby greatly improving surface quality.
For the piezo tube tool used in this study, cuts with a depth of cut up to 5 m and
orbit radius of 10 m produce features with good quality and few voids in the machined
surfaces. The surface finish quality can be improved by using one or several finish passes
with a depth of cut around 0.1 m, which was shown to result in true ductile machining.
For the case where the macroscopic feed motions do not allow changing the position in
128
such a small increment, the overall length of the piezo tube can be increased through a
reduction in the offset voltage of all electrodes by -20 V.
In terms of machining forces, larger tip radii cause larger machining forces. The
orbit frequency and feed speed, however, have a minor effect on the machining forces.
From a comparison of the tool tips before and after machining, it can be concluded that
the shape of the tool tips did not change during machining. This indicates that diamond
tips can be used to machine single crystal silicon without excessive tool wear from
diffusion.
CHAPTER 7
CONCLUSIONS AND FUTURE WORK
7.1 Conclusions
Orbital micromachining is a new technique that is based on orbital motion of the
single point tool tip. The tool tip (made of single crystal diamond) is held with a tool
holder that is attached to the free end of a piezo tube while the other end of the piezo tube
is fixed. The trajectory of the tool tip is achieved by applying sinusoidal excitation
voltages to the piezo tube, which has four external electrodes (x+, x-, y+, y-). A variety of
tool trajectories such as circular, elliptical, etc. can be created easily by adjusting the
amplitude and phase of the sinusoidal voltages that are applied to the piezo tube.
Moreover, the z-position of the tool can be controlled by applying an equal offset voltage
to all four electrodes.
Timoshenkos beam theory, which also considers rotational inertia as well as the
shear deformation, is a suitable method to estimate the transverse motion resonance
frequency of the piezo tube tool. The first resonance frequency of this tool, based on
calculations and experiments, is around 700 Hz. This frequency should be avoided in
order to prevent damage to the piezo tube. By actuating the tool tip on its own, recirculating tool paths, the cutting speed is decoupled from the feed speed, providing a
130
cutting mechanism that is very similar to a miniature fly cutter. The difference is that the
tool tip does not rotate around its center axis but orbits around the center axis instead.
In this study, open loop control was used to control the motion of the piezo tube
due to its simplicity and lower cost. The dynamic motion from the sinusoidal excitation
voltages minimize errors from hysteresis and drift, which are the major sources of errors
in controlling of the piezo tube in static applications. Other sources of errors such as
misalignment from manufacturing and output amplitude and phase from the amplifier can
be minimized by following the calibration procedure.
The tool tips in this study are made from single crystal diamond, which is the hardest
commercially material available. Ideally, the tool is able to machine any softer material.
However, the major cause of tool wear in diamond machining is the diffusion of carbon
atom from the tool into the work piece material. The rates of tool wear from diffusion
depend on the chemical composition of the work piece. From the experiments, the
diamond tip can machine aluminum alloy (Al 2024) with almost zero wear. Titanium
alloy (Ti-6Al-4V), on the other hand, causes a tremendous amount of wear.
A force model that takes into consideration the complex, 3D interaction between the
spherical and conical sections of the tool tip and the work piece was developed. The
machining force, which is applied on the contact surface between the tool and the work
piece, consists of a normal and a friction force. The force model is developed based on
the uniform contact pressure over the contact area, hence the specific normal cutting
force (KN) and the specific friction cutting force (Kf) are uniform along the contact area.
The results from the force model calculations are well matched with the results from the
experiments on ductile material (Al 2024).
131
The piezo tube tool can be used to machine 2D or 3D features into various types
of materials (ductile and brittle materials). From the experiments on aluminum alloy (Al
2024), a very smooth surface finish with Ra=0.05 m is obtained. The quality of the side
wall of the feature depends on the orbital trajectory as well as the feed motion of the tool.
By using the up-milling principle for orbital machining of ductile materials, microfeatures free of burrs on the side wall can be created.
Due to the thickness of the uncut chip, which is much smaller than the tool tip radius,
orbital micromachining creates high negative rake angles during machining, which
creates high hydrostatic pressure around the cutting zone. This pressure enables brittle
materials to be cut in the ductile regime and create surfaces with excellent surface
roughness that are free of cracks and voids. Ductile regime cutting on single-crystal
silicon was achieved with depths of cut around 0.1 m and resulting surface roughness
(Ra) of 0.007 m.
As a nontraditional micromachining technique, this tool is intended for the
prototyping of small parts in a wide range of materials. As the hardest material available,
diamond is an excellent choice for machining selected nonferrous materials such as
aluminum and copper. To machine ferrous materials (i.e. steel), however, the diamond tip
should be replaced with cubic boron nitride (CBN) to avoid excessive wear due to
diffusion of carbon from the diamond tool.
132
7.2 Future work
A sinker EDM machine (Optimation Profile 24) was used for the system setup #1
as a macroscopic motion control machine with the resolution of 0.1 m. This machine
used linear piezo motors to generate the x and y linear motion. The forces from this
motor and the stiffness of the axes proved to be insufficient for some of the micromilling
applications. The CNC machines (Haas-TM-1 and Haas-VF-E) that were also used in this
study are designed for conventional milling applications with a resolution of 1 m. In
micromachining, a resolution of 1 m is too large, particularly for brittle materials where
the depth of cut needs to be significantly less than 1 m. Therefore, it is suggested to
develop a small machine tool with 0.1 m resolution or better and with sufficient
stiffness tailored to the requirements of orbital micromachining. Moreover, the control
system that generates the sinusoidal excitation voltages for the piezo tube tool should be
integrated into the control system responsible for the macroscopic feed motions. This
would allow even more complex 2D and 3D geometries to be machined with this
technique.
Mechanical machining such as milling, turning and drilling is widely used in
macromachining. Most literature presents force models that are based on orthogonal
cutting models at the macroscale. The microscale cutting mechanism of micromachining,
on the other hand, is not fully understood. This research proposed a model to better
understand the force cutting mechanism of this novel micromachining technique. This
model was verified with experiments on Al 2024-T4 by using 6-axis load cell to measure
forces during machining. However, the model should be tested against other materials as
well.
133
The piezo tube actuator used for the orbital micromachining tool proved to be versatile
and relatively straightforward to design, build and control. As a relatively long, slender
object, it does however have a relatively low resonance frequency of 700 Hz. This limits
the orbital frequencies to less than 700 Hz, which in turn produces relatively slow cutting
speeds. Redesigning the orbital tool and replacing the single piezo tube with piezo stack
actuators orientated at 90o to each other would result in a tool of much greater resonance
frequency and stiffness, and is an option that should be pursued in the future.
Furthermore, this study was limited to the microscale. It is strongly suggested to expand
this tool to smaller length scales and to investigate its capabilities in nanomachining as
well.
REFERENCES
[1] Rajurkar KP, Levy G, Malshe A, Sundaram MM, McGeough J, Hu X, Resnick R,

DeSilva A. Micro and nano machining by electro-physical and chemical processes. CIRP
Annals - Manufacturing Technology 2006;55(2):643-66.
[2] Alting L. Micro Engineering. Annals of the CIRP 2003;52(2):635-57.
[3] Tansel I, Rodriguez O, Trjillo M, Paz E, Li W. Micro-end-milling - I. Wear and
Breakage. Int. Journal of Machine Tool and Manufacture 1998;38:1419-36.
[4] Koenig W, Kutzner K, Schehl U. Tool monitoring of small drills with acoustic
emission. Int. Journal of Machine Tool and Manufacture 1992;32:487-93.
[5] Friedrich CR, Vasile MJ. Development of the micromilling process for high-aspectratio microstructures. J MEMS 1996;5(1):33-8.
[6] Bamberg E, Heamawatanachai S. A novel micromilling technology based on single
point tool tip geometries. In: Proc. 2007 ASPE Conf. 2007. p.28-31.
[7] Pizani PS, Jasinevecius R, Duduch JG, Porto AJV. Ductile and brittle modes in
single- point-diamond-turning of silicon probed by Raman scattering. Journal of
Materials Science Letters 1999;18:1185-87.
[8] Patten JA, Gao W. Extreme negative angle technique for single point diamond nanocutting of silicon. Precision Engineering 2001;25:165-7.
[9] Stanford Microfluidics Laboratory (http://microfluidics.stanford.edu).
[10] Muck AJ, Wang J, Jacobs M, Chen G, Chatrathi MP, Jurka V, Vyborny Z, Spillman
SD, Sridharan G, Schoning MJ. Fabrication of poly(methyl methacrylate) microfluidic
chips by atmospheric molding. Analytical Chemistry 2004;76(8):2290-7.
[11] Wu H, Odom TW, Chiu DT, Whitesides GM. Fabrication of complex threedimensional microchannel systems in PDMS. Journal of the American Chemical Society
2003;125(2):554-9.
[12] Blattert C, Jurischka R, Schoth A, Kerth P, Menz W. Fabrication and testing of
novel blood separation devices based on microchannel bend structures. Progress in
Biomedical Optics and Imaging - Proceedings of SPIE 2005;5651:196-203.
135
[13] Singh V, Desta Y, Datta P, Guy J, Clarke M, Feedback DL, Weimert J, Goettert J. A
hybrid approach for fabrication of polymeric BIOMEMS devices. Microsystems
Technology 2007;13:369-77.
[14] Bohl B, Steger R, Zengerle R, Koltay P. Multi-layer SU-8 lift-off technology for
microfluidic devices. Journal of Micromechanics and Microengineering 2005;15(6):
1125-30.
[15] Becker H, Heim U, Roetting O. Fabrication of polymer high aspect ratio structures
with hot embossing for microfluidic applications. Proceedings of SPIE 1999;3877:74-9.
[16] Hayes JP, Harvey EC, Ghantasala MK, Dempster BC. A laser-ablation and lift-off
technique for fabricating simple microfluidic devices. Proceedings of SPIE
2001;4236:115-25.
[17] Hirst LS, Parker ER, Abu-Samah Z, Li Y, Pynn R, MacDonald NC, Safinya CR.
Microchannel systems in titanium and silicon for structural and mechanical studies of
aligned protein self-assemblies. Langmuir 2005;21(9):3910-4.
[18] Yang LJ, Kang SW. Micro fluidic system of micro channels with on-site sensors by
silicon bulk micromachining. Proceedings of SPIE 1999;3877: 267-72.
[19] Weston DF, Smekal T, Rhine DB, Blackwell J. Fabrication of microfluidic devices
in silicon and plastic using plasma etching. Journal of Vacuum Science and Technology
B: Microelectronics and Nanometer Structures 2001;19(6):2846-51.
[20] Karnakis DM, Knowles MRH, Alty KT, Schlaf M, Snelling HV. Comparison of
glass processing using high repetition femtosecond (800nm) and UV (255nm)
nanosecond pulsed lasers. Proceedings of SPIE 2005;5718:216-27.
[21] Lin CH, Lee GB, Lin YH, Chang GL. A fast prototyping process for fabrication of
microfluidic systems on soda-lime glass. Journal of Micromechanics and
Microengineering 2001;11(6):726-32.
[22] Kunieda M, Lauwers B, Rajurkar KP, Schumacher BM. Advancing EDM through
fundamental insight into the process. CIRP Annals-Manufacturing Technology
2005;54(2):599-622.
[23] Madou MJ. Fundamentals of microfabrication: the science of miniaturization 2nd
edition. CRC. 2002.
[24] Brehl DE, Dow TA. Review of vibration-assisted machining. Precision Engineering
2008;32:153-72.
136
[25] Heamawatanachai S, Bamberg E. Design and characterization of a PZT driven
micromachining tool based on single-point tool tip geometry. Prec. Eng. 2009;33(4):38794.
[26] Bamberg E, Heamawatanachai S. Orbital electrode actuation to improve efficiency
of drilling micro-holes by micro-EDM. Journal of Materials Processing Technology
2009;209:1826-34.
[27] Dornfeld D, Min S, Takeuchi Y. Recent advances in mechanical micro-machining.
Annals of the CIRP 2006;52(2):745-68.
[28] Gogotsi Y, Zhou G, Ku SS, Cetinkunt S. Raman microspectroscopy analysis of
pressure-induced metallization in scratching of silicon. Semiconductor Science and
Technology 2001;16(5):345-52.
[29] Shamoto E, Suzuki N, Tsuchiya E, Hori Y, Inagaki H, Yoshino K. Development of 3
DOF ultrasonic vibration tool for elliptical vibration cutting of sculptured surfaces. CIRP
Ann 2005;54(3):214.
[30] Rusnaldy, Ko TJ, Kim HS. Micro-end-milling of single-crystal silicon. Int J Mach
Tools & Manuf 2007;47:2111-9.
[31] Gao W, Hocken RJ, Patten JA, Lovingood J. Experiments using a nano-machining
instrument for nano-cutting brittle materials. CIRP Annals Manuf Techn
2000;49(1):439-42.
[32] Gao W, Hocken RJ, Patten JA, Lovingood J, Lucca DA. Construction and testing of
a nanomachining instrument. Prec Eng 2000;24:320-8.
[33] Chao CL, Ma KJ, Liu DS, Bai CY, Shy TL. Ducile behavior in single-point
diamond-turning of single-crystal silicon. J Mat Process Tech 2002;127:187-90.
[34] Physik Instrumente (PI) GmbH & Co. KG. (http://www.physikinstrumente.com).
[35] Sun Q, Wolkow RA. Three-dimensional displacement analysis of a piezoelectric
tube scanner through finite element simulations of a tube assembly. Review of Scientific
Instruments 2006;77:113701(9pp).
[36] EDO Corporation (http://www.edocorp.com/EDOElectroCeramics.htm).
137
[37] Huang TC. The effect of rotary inertia and of shear deformation on the frequency
and normal mode equations of uniform beams with simple end conditions. ASME Journal
of Applied Mechanics 1961;26:579-84.
[38] Bruch JC, Mitchell TP. Vibrations of a mass-loaded clamped-free Timoshenko
beam. Journal of Sound and Vibration 1987;114(2):341-45.
[39] Hutchinson JR. Shear coefficients for Timoshenko beam theory. ASME Journal of
Applied Mechanics 2001;68(1):87-92.
[40] Abramovich H, Hamburger O. Vibration of a cantilever Timoshenko beam with a tip
Mass. Journal of Sound and Vibration 1991;148(1):162-70.
[41] Tlusty G. Manufacturing processes and equipment. Prentice Hall, 2000.
[42] Shaw MC. Metal cutting principles. Oxford University Press, 2005.
[43] Merchant ME. Mechanics of the metal cutting process. J. Appl. Phys. 1945;16:26775.
[44] Lee HU, Cho DW, Ehmann KF. A mechanistic model of cutting forces in microendmilling with cutting-condition-independent cutting force coefficients. J. Manuf. Sci.
& Eng. 2008;130(3):0311021-9.
[45] Son SM, Lim HS, Ahn JH. Effects of the friction coefficient on the minimum cutting
thickness in micro cutting. Int. J. Mach. Tools & Manuf. 2005;45(15):529-35.
[46] Li C, Lai X, Li H, Ni J. Modeling of three-dimensional cutting forces in microendmilling. J. Micromech. Microeng. 2007;17(4):671-8.
[47] Taniguchi N. Nanotechnology-integrated processing systems for ultra-precision and
ultrafine products. Oxford University Press. 1996.
[48] Sato K. Grinding temperature. Bull. Jap. Soc. Grind. Eng. 1961;1:31-3.
[49] Shaw MC. A new theory in grinding. Mech. & Chem. Eng. Trans. 1972;MC8(1):738.
[50] Paul E, Evans CJ, Mangamelli A, McGlauflin ML, Polvani RS. Chemical aspects of
tool wear in single point diamond turning. Prec. Eng. 1996;18(1):4-19.
[51] Heamawatanachai S, and Bamberg E (in press). Cutting force model of orbital
single-point micromachining tool. International Journal of Machine Tools &
Manufacture. DOI: 10.1016/j.ijmachtools.2010.05.002
138
[52] Uddin MS, Seah KHW, Rahman M, Li XP, Liu K. Performance of single crystal
diamond tools in ductile mode cutting of silicon. J. Mater. Process. Technol. 2007;185(13):24-30.
[53] Patten J, Gao W, Yasuto K. Ductile regime nanomachining of single-crystal silicon
carbide. J. Mater. Sci. Eng. 2005;127(3):522-32.
[54] Yoshino M, Ogawa Y, Aravindan S. Machining of hard-brittle materials by a single
point tool under external hydrostatic pressure. J. Manuf. Sci. Eng. 2005;127(4):837-45.
[55] Yan J, Zhao H, Kuriyagawa T. Effects of tool edge radius on ductile machining of
silicon: an investigation by FEM. Semicond. Sci. Technol. 2009;24(7):075018 (11pp).
[56] Liu K, Li XP. Ductile cutting of tungsten carbide. J. Mater. Process. Technol. 2001;
113(1-3):348-54.
[57] Sreejith PC. Machining force studies on ductile machining of silicon nitride. J.
Mater. Process. Technol. 2005;169(3):414-7.
[58] Fang FZ, Venkatesh VC. Diamond cutting of silicon with nanometric finish. CIRP
Annals Manuf. Technol. 1998;47(1):45-9.
[59] Chao CL, Ma KJ, Liu DS, Bai CY, Shy TL. Ductile behavior in single-point
diamond-turning of single-crystal silicon. J. Mater. Process. Technol. 2002;127(2):18790.
[60] Venkatachalam S, Li X, Liang SY. Predictive modeling of transition undeformed
chip thickness in ductile-regime micro-machining of single crystal brittle materials. J.
Mater. Process. Technol. 2009;209(7):3306-19.
[61] Rusnaldy, Ko TJ, Kim HS. Micro-end-milling of single-crystal silicon. Int J Mach
Tools & Manuf 2007;47:2111-9.
[62] Dowling NE. Mechanical behavior of materials: engineering methods for
deformation, fracture, and fatigue 2nd edition. Prentice Hall. 1998.
[63] ASM Aerospace Specification Metals Inc. ( http://asm.matweb.com).
[64] Heamawatanachai S, Bamberg E, (in preparation). Investigation of micro-machining
of single crystal silicon using an orbital single-point micromachining tool. Journal of
Materials Processing Technology.

Design Characterization and Machining Force Model of A Novel Orbital Micromachining Technology Based On Single Point Tool Tip Geometry

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Design Characterization and Machining Force Model of A Novel Orbital Micromachining Technology Based On Single Point Tool Tip Geometry

Hochgeladen von

Copyright:

Verfügbare Formate

DESIGN, CHARACTERIZATION AND MACHINING FORCE MODEL

OF A NOVEL ORBITAL MICROMACHINING TECHNOLOGY

A dissertation submitted to the faculty of

Department of Mechanical Engineering

Copyright Sumet Heamawatanachai 2010

The University of Utah Graduate School

STATEMENT OF DISSERTATION APPROVAL

has been approved by the following supervisory committee members:

and by Charles A. Wight, Dean of The Graduate School.

This research presents a new mechanical micromachining technique that can be

6.7 Tool trajectories during machining..114

1.1: Concept of EDM...6

4.12: Friction coefficient in orbital machining..81

6.3 Pyramid feature in single-crystal silicon ...109

The piezoelectric strain coefficient for displacements normal to the

The length of the tube [m]

The change in length of the tube [m]

The voltage applied to all electrodes [V]

The wall thickness [m]

Lateral displacement in x direction [m]

Lateral displacement in y direction [m]

The inner diameter of the tube [m]

The symmetric electrode voltages in x axis [V]

The symmetric electrode voltages in y axis [V]

Youngs modulus of the beam [N/m2]

Shear coefficient of the cross-section [ ]

The cross-sectional area of the beam [m2]

Shear modulus of elasticity [N/m2]

Slope of the beam due to bending [ ]

The deflection of the center line of the beam [m]

The density of the beam [kg/m3]

The inner radius of the tube [m]

The outer radius of the tube [m]

The moment of inertia of the attached mass around its center of

The attached mass [kg]

Non-dimensional length along the beam [ ]

Normal force [N]

Friction force [N]

Specific cutting force in the normal direction [N/m2]

Specific cutting force in the tangential direction [N/m2]

Tool tip coordinate frame [m,m,m]

Center coordinate frame [m,m,m]

Tool tip coordinate frame [m,m,m]

Orbit radius [m]

Angular speed [rad/s]

Feed speed [m/s]

Tool tip radius [m]

Cross section radius of the ith disk [m]

Cone angle [degrees]

Orbit angle [degrees]

Depth of cut [m]

The tool engagement angle [rad]

The tool engagement length [m]

The normal vector to the contact surface of the ith disk

The normal force vector of the ith disk

The friction force vector of the ith disk

The directional area component for the calculation of the normal

The directional area component for the calculation of the normal

The directional area component for the calculation of the normal

The directional area component for the calculation of the frictional

The directional area component for the calculation of the frictional

Cutting force in x direction [N]

Cutting force in y direction [N]