<%BANNER%>

Design, Assembly, and Testing of an Ultra-High-Speed Micro-Milling Spindle


PAGE 1

DESIGN, ASSEMBLY, AND TESTING OF AN ULTRA-HIGH-SPEED MICRO-MILLING SPINDLE By JAY PRAKASH PATHAK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003

PAGE 2

ACKNOWLEDGMENTS The author would like to acknowledge with thanks the contribution of the following people who helped him during this project. Dr. John Ziegert provided distinguished and expert tutelage as the chair of supervisory committee throughout the development of the spindle system right from discussing idea during the design phase to the end with actual assembly and testing. Dr. John Schueller provided support during assembly of the spindle system. Dr. Ashok Kumar provided help during the model development and finite element analysis. Dr. Sanjay Ranka was kind enough to be part of my committee. Dr. Tony Schmitz provided relevant feedback during assembly and force measurements. Mr. Paul Frederickson of Precise Corporation was kind enough to do the finite element analysis for analyzing the natural frequencies of the spindle system as well as provided help in the design of the mount for the Precise motor. Mr. Bernhard Jokiel and David Gill of Sandia National Labs instigated this project and provided funding for the development of the spindle system. Mr. Andrew Dewitt and Mr. Tim Claffey of New Way Machine Components, Inc., extended their cooperation to make custom designed air bearing relevant to our needs. Mr. Ron Brown, machinist at University of Florida, helped in making some final changes during assembly of the machine. ii

PAGE 3

The author would like to thank his friends, lab-mates and others who one way or the other helped him throughout this project. Lastly, the author would like to thank his parents for moral support and motivation. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS..................................................................................................ii LIST OF FIGURES...........................................................................................................vi ABSTRACT.....................................................................................................................viii CHAPTER 1 INTRODUCTION........................................................................................................1 1.1 Micro-Milling.........................................................................................................1 1.2 Difficulties in Scaling Milling Processes...............................................................2 1.3 Feasibility and Need For Very High Speed Micro-Milling Spindles.....................4 1.4 Ultra-High Speed Micro-Milling Spindle System..................................................5 1.5 Organization of Remainder of Thesis.....................................................................6 2 BACKGROUND AND LITERATURE REVIEW......................................................7 2.1 Development of Micro-Milling Process.................................................................7 2.2 Micro-Milling Tool Fabrication.............................................................................7 2.3 Modeling Micro-End-Milling Operations..............................................................8 2.3.1 Analytical Cutting Force Model...................................................................8 2.3.2 Influence of Tool Run-Out...........................................................................9 2.3.3 Influence of Tool-Wear..............................................................................10 2.3.4 Cutting Force Model for Micro-Milling of Heterogeneous Materials.......11 2.4 Current High-Speed Spindle Technology.............................................................11 3 MOTIVATION AND CONCEPTUAL DESIGN......................................................13 3.1 Motivation.............................................................................................................13 3.1.1 Low Chip Thickness...................................................................................13 3.1.2 High Tool Run-out.....................................................................................14 3.1.3 Low Metal Removal Rate...........................................................................14 3.1.4 High Cutting Force Coefficients................................................................14 3.1.5 Low Cutting Speed.....................................................................................15 3.1.6 Unpredictable Tool Life.............................................................................15 3.2 Conceptual Design................................................................................................15 3.2.1 Conceptual Design......................................................................................15 3.2.2 Design Description.....................................................................................17 iv

PAGE 5

4 DETAILED SPINDLE DESIGN AND ASSEMBLY...............................................19 4.1 Base Frame...........................................................................................................22 4.2 Air Bearing...........................................................................................................22 4.3 Force Sensing System...........................................................................................25 4.4 Assembly of Base Frame, Air-Bearing and Force Sensor....................................26 4.5 Motor Mount.........................................................................................................27 4.6 Assembly of Motor Mount with Base Frame.......................................................28 4.6 Assembly of Precise Spindle with Base Frame and Friction Wheel....................30 4.7 Assembly of Micro-Spindle with HSM-2.............................................................30 4.8 Micro-Spindle System Connections.....................................................................31 4.9 Computer, Electronics and Software....................................................................33 5 DESIGN VALIDATION............................................................................................37 5.1 FEA of Motor and Friction Wheel Assembly.......................................................37 5.2 FEA of the Micro-Tool, Air-Bearing and the Friction Drive...............................39 5.3 Stress/Strain Analysis...........................................................................................41 6 TESTING AND RESULTS........................................................................................43 6.1 Run-out Measurements.........................................................................................43 6.1.1 Run-out Measurement at Motor Arbor.......................................................43 6.1.2 Run-out Measurements at Friction Wheel and Gage Pin...........................45 6.2 Force Measurements.............................................................................................46 6.3 Machining of Hexagonal Test Part.......................................................................50 6.4 Air-Bearing Stiffness Measurements....................................................................51 7 CONCLUSIONS AND FUTURE WORK.................................................................53 APPENDIX A PART LIST.................................................................................................................55 B DETAIL DRAWINGS...............................................................................................59 LIST OF REFERENCES...................................................................................................66 BIOGRAPHICAL SKETCH.............................................................................................69 v

PAGE 6

LIST OF FIGURES Figure page 1-1 Solid model of ultra-high spindle micro-milling spindle system.............................5 3-1 3D model of ultra-high spindle micro-milling spindle system..............................16 4-1 The assembled micro-milling spindle....................................................................20 4-2 The assembled micro-milling spindle system mounted to HSM-2........................21 4-3 3D model of the base frame...................................................................................22 4-3 3D model of the air bearing...................................................................................23 4-4 Isometric view of the air bearing...........................................................................23 4-5 Air cleaning system and valve...............................................................................24 4-6 Three axes force sensor..........................................................................................25 4-7 Assembly of base frame, air-bearing and force sensor..........................................26 4-8 CAD model of the motor mount............................................................................27 4-9 Motor mount..........................................................................................................28 4-10 Spherical washer set...............................................................................................28 4-11 Fine adjustment set screw......................................................................................29 4-12 Assembly of motor mount and base frame............................................................29 4-13 Friction wheel........................................................................................................30 4-14 Interface plate.........................................................................................................31 4-15 Schematic of system connections..........................................................................32 4-16 Bread board circuit.................................................................................................34 4-17 Enclosure containing i/o connector and bread board.............................................34 vi

PAGE 7

4-18 Labview interface for spindle control and measurement.......................................35 4-19 Labview interface for cutting force measurement.................................................36 5-1 Motor, motor arbor and friction wheel assembly for FEA....................................38 5-2 Results of FEA of motor and friction wheel assembly..........................................38 5-3 FEA model of the tool, air-bearing and friction wheel..........................................39 5-4 First mode: 4857.3 Hz............................................................................................40 5-5 Second mode: 1.35e5 Hz.......................................................................................41 5-6 Third mode: 2.65e5 Hz..........................................................................................41 5-7 Fourth mode: 3.89e5 Hz........................................................................................41 6-1 Experimental set-up...............................................................................................44 6-2 Run-out of drive spindle arbor...............................................................................44 6-3 Run-out of friction wheel and gage pin.................................................................45 6-4 Force plot for micro-milling of Al using 127m micro-tool.................................47 6-5 Mathematical model of the tool, air-bearing and force sensor system..................48 6-6 Slots made from 508m micro-tool.......................................................................49 6-7 Slots made from 254m micro-tool.......................................................................49 6-8 Hexagonal feature with a wall thickness of 30m.................................................50 6-9 Force vs. displacement plot to measure the stiffness.............................................52 vii

PAGE 8

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science DESIGN, ASSEMBLY, AND TESTING OF AN ULTRA-HIGH-SPEED MICRO-MILLING SPINDLE By Jay Prakash Pathak December 2003 Chair: Dr. John C. Ziegert Major Department: Mechanical and Aerospace Engineering Micro-milling is an emerging fabrication technology that broadens applicable material ranges including metals and plastics. It is potentially the technology of choice to create complex three-dimensional microand meso-scale components in hard engineering materials. The micro-milling process is characterized by the use of milling tools 100 m or less in diameter. Presently, these tools are used to create miniature features in plastics and some soft metals, typically with a very low material removal rate. Spindles for micro-milling often show excessive run-out relative to the chip thickness required by these small tools. This thesis documents the design, assembly and testing of an ultra-high speed micro-milling spindle; aimed at improving the efficiency of micro-milling operations. viii

PAGE 9

CHAPTER 1 INTRODUCTION This thesis documents the design, assembly and testing of an ultrahigh-speed micro-milling spindle used for creating micro and meso-scale features in soft as well as hard engineering materials. This chapter briefly describes the micro-milling process and describes difficulties encountered in micro-milling technology which motivates the need for a new spindle. 1.1 Micro-Milling There is a continuous demand for miniaturization of components for consumer and other products. Micro-milling is a potential technology for the production of miniaturized parts and components. There exists a wide variety of important applications for microand meso-scale mechanical systems in the commercial and defense sectors, which require high-strength materials and complex geometries that cannot be produced using current MEMS fabrication technologies. Micro-milling has the potential to fill this void in MEMS technology by adding the capability of free form machining of complex 3D shapes from a wide variety and combination of traditional, well-understood engineering alloys, glasses and ceramics. Micro-milling is an enabling technology that may allow rapid and economic fabrication of meso-scale components with micro-scale features. These types of parts are a vital link between surface micro machined MEMS systems and the outside macroscopic world, and potentially have great implications for how MEMS components interact with each other as well as how they are packaged. 1

PAGE 10

2 Micro-milling can be defined as a traditional milling process that has been scaled down to micron level. However due to the small diameter tools used in the process, micro-milling is significantly different than conventional milling in a variety of factors, viz. cutting forces, tool run-out and metal removal rate etc. 1.2 Difficulties in Scaling Milling Processes Inefficiencies in milling small features result directly from the relationships between a cutting tools breaking strength, the applied cutting force, and the metal removal rate (MRR). To the first order, end mills act like long, thin, end-loaded, cylindrical cantilever beams; where the maximum bending stress (x) is directly proportional to the tool length (L) and the cutting force (Fc), while inversely proportional to the cube of the diameter (D) (Equation 1.1). 332DLFcx (1.1) In slot milling the maximum tangential cutting force (Equation 1.2) is proportional to the workpiece material cutting stiffness (Ks), axial depth of cut (dc), and feedrate (fr), while it is inversely proportional to the spindle speed (n) and number of teeth on the cutter (N). ChipLoadrcscrevteethNrevnmmfmmdmmNKNF)/(min)/(min)/()()/()(2 (1.2) The MRR (Equation 1.3) is proportional to the feed rate, the depth of cut and the radial depth of cut of the tool (rc). mmrmmdmmfmmMRRccrmin)min(3 (1.3)

PAGE 11

3 In light of Equations 1.1-1.3, we consider a macro-scale part and associated tooling that one wishes to homogenously scale in size by a scaling factor, k. If the spindle speed is constant for both operations, the axial and radial depths of cut will scale by k, and thus the feed rate must be scaled by a factor of k to maintain a constant bending stress in the tool. Since the MRR is proportional to the feed rate, the depth of cut and the radial depth of cut of the tool (rc) the MRR scales by k3. However, cutting tool edge radius effects and spindle run-out become more important at micro-milling scales. Commercially available micro-milling tools typically have a cutting edge radius on the order of 2 to 3m. In micro-milling, to avoid tool breakage the chip thickness is typically less than 1m, so the tool edge effectively has a large negative rake angle. This poor cutting geometry increases the effective cutting stiffness, Ks, resulting in even higher forces, and thus requiring an even smaller chip thickness, further slowing the feed rate and increasing the processing time. Typical milling spindles used for these small tools employ either rolling element bearings or air bearings to support the spindle shaft. Tools are typically clamped in such spindles using a collet or set screw. The combination of asynchronous spindle bearing error motions and clamping errors often result in tool run-out 3 to 20 times the nominal chip thickness. This will drastically increase the cutting force and may lead to tool breakage unless the axial depth of cut is reduced, thus further reducing the MRR and increasing the processing time. As a result, the tool spends most of its time under-loaded to protect against the few times when the spindle and clamping errors cause excessively thick chips to be cut.

PAGE 12

4 1.3 Feasibility and Need For Very High Speed Micro-Milling Spindles In conventional machining processes there is a preferred range of cutting speeds for a given tool and workpiece material. Therefore, to maintain the preferred cutting speed, the spindle speed should increase as the tool diameter decreases (Equation 1.4). mmrrevnmmVtoolcutmin/602min/ (1.4) Since cutting forces in milling do not increase significantly with the cutting speed, one approach to improving the efficiency of micro-milling operations is develop higher speed spindles with low radial error motions to decrease machining time. Micro-milling operations should not be more time consuming than conventional milling for a given part of similar geometry, if sufficiently high spindle speeds can be achieved and the problem of high tool run-out at such speeds can be solved. Consider a macro-scale steel part machined using a 10mm diameter two-flute end mill, rotating at 5000RPM with a chip load of 0.18mm/tooth (yields a feed rate of 1800mm/min). Scaling the part with k =0.01 requires a 0.1mm diameter two-flute end mill, rotating at 500,000 rpm with a chip load of 0.0018 mm/tooth, which yields a feed rate of 1800mm/min. If these parameters can be achieved, the machining time should be the same. The required spindle speed for the micro-milling case is far beyond the range of currently available spindles, and points to the need to develop a very high-speed spindle for micro-milling operations. Additionally, any such spindle should have sub-micron run out to avoid overloading the tool, and should include a sensing system of some sort to aid in tool setting, and detecting tool breakage.

PAGE 13

5 1.4 Ultra-High Speed Micro-Milling Spindle System The proposed spindle design uses the tool shank itself as the spindle shaft, supported in an air bearing. A friction drive transmits torque from a drive motor to the tool shaft as shown in the Figure 1-1. Figure 1-1. Solid model of ultra-high spindle micro-milling spindle system. The drive ratio of the friction drive is 1:8, making it is possible to achieve more than 500K RPM when combined with commercially available high speed motors capable of speeds in the 10K-90K RPM range. Commercially available micro-milling cutters typically have a 0.125-inch diameter shank, approximately 1.5 inches long. The actual milling cutter is ground on the end of the shank. Instead of clamping the tool shank into

PAGE 14

6 the rotating shaft of the spindle, the proposed spindle will support the tool shank radially in a porous carbon air bearing. The tool shank effectively becomes the spindle shaft and should greatly eliminate tool run-out. 1.5 Organization of Remainder of Thesis Literature reviews and background information are provided in Chapter 2. Chapter 3 describes the motivation and conceptual design of the micro-spindle. Chapter 4 describes the detailed design of the micro-spindle components and parts. Chapter 5 describes design validation analyses using finite element analysis to understand the dynamics of the micro-spindle and calculations for a reliability check of the design. Chapter 6 documents the results of the initial testing of the prototype spindle. Chapter 7 provides conclusions and suggestions for future work.

PAGE 15

CHAPTER 2 BACKGROUND AND LITERATURE REVIEW This chapter gives a summary of the available literature relevant to the micro-milling process. 2.1 Development of Micro-Milling Process Conventional milling is known to be a very versatile process capable of creating three-dimensional features and structure. Adaptation of this process to create micro-scale features is termed micro-milling. Friedrich and Vasile [1] describe the development of the micro-milling process for fabricating molds and mask structures in PMMA. The process uses focused ion beam [2] micro machined tools [3] and utilizes a high-precision milling machine designated for the milling process. Friedrich, Coane, Goettert and Gopinathin [4] have shown that micromechanical milling is also a rapid and direct method for fabricating masks for deep x-ray lithography with lateral absorber features down to 10m. Adams et al. [2] demonstrate micro-milling as a first step in fabricating metal alloy micro components. The metal alloys that were tested are 6061-T4 aluminum, brass and 4340 steel. For all tests 15-25 m deep trenches were cut several millimeters in length as commanded. 2.2 Micro-Milling Tool Fabrication Vasile et al. [3] describe a process for making very small milling cutters using focused ion beam machining to create tools with a desired number of cutting edges and tool end clearance. Micro-end mills having ~25m diameter are made by sputtering 7

PAGE 16

8 cobalt M42 high-speed steel and C2 micro-grain tungsten carbide tool blanks. Cutting edge radii of 0.1 m can be achieved. Inexpensive micro-end mills are also commercially available. They typically have a 1/8 inch diameter shank, 1.5 inches in length. The cutting portion is ground on the end of the shank, and is available in sizes ranging from 0.005 to 0.030 inch diameter, and in 2 and 4 flute geometries. 2.3 Modeling Micro-End-Milling Operations As discussed in Chapter 1 micro-milling behaves very differently than conventional milling in a number of ways. The ratio of feed per tooth to tool edge radius is much smaller for micro-milling than for conventional milling. Stress variations on the shaft of a micro-tool are much higher than on conventional tools. These extreme operating conditions can drastically shorten tool life. Less than one hundred inches of tool life is common when hard metals such as stainless steel are machined. If the cutting conditions are not selected properly, micro-tools will be broken in a few seconds. Because of their small size, it is very difficult to notice the damaged cutting edges and even the broken shaft. Many hours of machining time may be wasted if the tool failure is not detected in time. 2.3.1 Analytical Cutting Force Model Cutting force prediction is very important to achieve optimal cutting conditions in machining operations. Several cutting force models have been proposed for accurately measuring the cutting forces for better control in milling operations. In 1975 Tlusty and MacNeil [5] gave the first analytical expressions for cutting forces for conventional milling. It was improved by Gygax [6] in 1979 and by Kline and DeVor [7] in 1983. Kline and DeVor [7] considered the effect of cutter run-out on cutting geometry and

PAGE 17

9 hence affecting the cutting forces. Again a database of cutting force coefficients for different tool and material combination was created considering different aspects of milling like up and down-milling, symmetric and asymmetric cuts. Later Tlusty and Ismail [8] in 1983 and Ismail and Bastami [9] in 1986 concentrated on the dynamics of the cutting operation and development of chatter. In these studies, cutting forces were calculated numerically to be able to consider the influence of the present and previous tool vibrations to the uncut chip area. Tlustys cutting force model which considers the tool tip path, as circular arcs that are mutually shifted by ft (feed per tooth) is no longer valid for micro-milling as the ratio of feed per tooth and tool radius is not very small in micro-milling. Bao and Tansel [10-12] proposed new analytical models in order to perform micro-machining operations at the optimal cutting conditions. Their method calculates the chip thickness by considering the trajectory of the tool tip while the tool rotates and moves ahead continuously. This model considers the trajectory of the tool tip as cycloids not circles as proposed in Tlustys model. This gives a different expression for chip thickness and takes into account the difference between the up and down-milling. As opposed to Tlustys model, where chip thickness is zero when cutter angle is zero or 1800 in a slotting cut, here the chip thickness is never zero for any real value of cutter angle. 2.3.2 Influence of Tool Run-Out The effect of tool run-out creates negligible changes in cutting force profile of conventional end milling operations whereas in case of micro-milling it leads to drastic force variations. It is very common to see that only one cutting edge of a two-flute micro-end-mill performs the machining operations alone while the other edge doesnt touch the workpiece at all. When one of the cutting edges starts to perform all or most of the cutting

PAGE 18

10 operations, the force variation increases significantly. The tool wears out much more quickly, and the probability of tool breakage increases. To select the optimal cutting conditions in micro-milling, it is very helpful to be able to estimate the cutting force characteristics in a wide cutting parameter range if the tool run-out is known. To consider the tool run-out, the chip thickness expression derived by Tlusty and MacNeil [5] in 1975 was modified by Gu Kapoor and DeVor [13] in 1991. Again a computer-based numerical model was introduced by Sutherland and DeVor [14]. Bao and Tansel [11] gave the first compact expression to calculate the cutting forces for micro-milling with tool run-out. The cutting force expressions were simplified by eliminating the insignificant components of the chip thickness expression to obtain the results in a more compact form. This can be also used in with optimization algorithms to estimate the tool run-out, to select the optimal cutting conditions, to monitor the operating conditions, and to estimate the surface quality from experimental cutting force data. The validity of the proposed model is evaluated by comparing the simulated and experimental cutting force profiles. Again the expressions are reducible to the cutting force expressions without tool run-out if the run-out term is set to zero. 2.3.3 Influence of Tool-Wear In addition to tool run-out as discussed in the previous section tool wear also plays a significant role in accurate prediction of cutting forces in micro-milling. The increase of cutting forces with tool wear was observed in turning operations in 1975 by Cook, Subramanian and Basile [15]. However estimation of tool wear from cutting force is not easy since the cutting forces continuously change even in typical turning operations when the cutting conditions change. The number of parameters increases in end-milling operations and estimation of the tool condition becomes much more complicated. To

PAGE 19

11 detect tool breakage and estimate wear, characteristics of various signals were studied by Altintas, Yellowley and Tlusty [16] in 1988 and the effectiveness of empirical models including the time series analysis and neural networks was demonstrated by Takata, Ogawa, Bertok, Ootsuka, Matushima, and Sata [17] in 1985, Liang and Dornfeld [18] in 1989 and Tansel and McLaughlin [19] in 1993. Bao and Tansel [12] in 2000 modified their own cutting force models [10-11] to represent tool wear at any stage of usage. Modification of the analytical models [10-11] to represent the tool wear is straightforward; however, knowledge of different tool conditions requires estimation of the parameters of these non-linear models. Genetic algorithms were proposed for estimation of the parameters of the model. 2.3.4 Cutting Force Model for Micro-Milling of Heterogeneous Materials Bao and Tansels [10-12] model of cutting forces assumes the workpiece material to be homogeneous. But actually as the depth of cut and feed rates are reduced, the chip load encountered in the process becomes the same order of magnitude as the grain size of many alloys such as steel. For such materials the workpiece material must be modeled as heterogeneous. Vogler et al. [20], in 2002 developed a mechanistic model for cutting force prediction that explicitly accounts for the different phases while machining steel by using different cutting force coefficients for the different material phases. The model is validated by using calibration tests for ferrite and pearlite, the major components of ductile iron and many other steels, in order to determine the machining force-chip load relationship for individual phases of heterogeneous material. 2.4 Current High-Speed Spindle Technology Some literatures were reviewed on current miniature high speed spindle technology. The first one to be investigated in this respect was National Jet made high

PAGE 20

12 precision micro-milling machine [1]. This machine is composed of 1500 kg of granite for vibration and thermal stability and air bearing axes for added stability of travel. Interferometric position control gives a resolution of 1.25 nm for x and y-axes. The micro-milling head uses a vee-block bearing arrangement that has a total of four spherically convex diamond surfaces arranged as two sets of two each. Tool rotation is highly concentric about the axis in space with no measurable radial run-out of the tool itself. The second spindle that was investigated is Zindles [21] 250,000 RPM drilling mechanism. High RPM and zero run-out of this spindle makes it ideal for drilling holes in the 0.018 to 0.003 diameter range. The primary mechanics of this spindle system consists of three high precision, hardened tool steel wheels arranged in triangle fashion. Two of these wheels are mechanically driven, while the third one is an idler. The drill rides against the faces of the two drives wheels and is held against them by the idler wheel. The high RPM is accomplished by way of a simple mechanical gear-up, using the faces of the drive wheels and pulleys in the system. The large diameter of the drive wheels relative to the small diameter of the drill shank makes it possible to rotate the drill at 250,000 RPM while the wheel rotates at 45,000.

PAGE 21

CHAPTER 3 MOTIVATION AND CONCEPTUAL DESIGN This chapter describes the design goals of the micro-milling spindle and provides details of the proposed design. 3.1 Motivation The literature review in Chapter 2 summarized some of the challenges in micro-milling. These include: Low chip thickness. High tool run-out. Low metal removal rate. High cutting force coefficients. Low cutting speed. Unpredictable tool life. 3.1.1 Low Chip Thickness The micro-milling process is characterized by milling tools that are currently in the range of 22-100m. Because of the small diameter of the tools, the cutting forces must be kept very small so as not to exceed the bending stress fatigue limit of the tool at the root. In order to keep the forces sufficiently small, the chip thickness must be very small. Typical chip load values reported for machining of metallic workpieces are on the order of 0.5 to 1.0 m. 13

PAGE 22

14 3.1.2 High Tool Run-out Typical milling spindles used for micro tools employ either rolling element bearings or air bearings to support the spindle shaft. Tools are typically clamped in such spindles using a collet or set screw. The combination of asynchronous spindle bearing error motions and clamping errors often result in tool run-out 3 to 20 times the nominal chip thickness. This means that in ordinary operation, some teeth on the cutter may not contact the workpiece at all during rotation, while others are forced to cut chips up to several times the desired thickness. This leads to overloads of the tool and premature failure. 3.1.3 Low Metal Removal Rate The low chip thickness characteristic of micro-milling leads to low feed rates. For example, with a 1 m chip thickness and a 2 fluted cutter, the feed rate is only 2 m per revolution. Therefore, if the spindle speed is 20,000 rpm (a typical maximum spindle speed for a high speed milling spindle), the feed rate of the tool through the work is only 40 mm/min. This results in very low material removal rates. Also tool run-out becomes more significant at higher speeds, which drastically increases the cutting force and may lead to tool breakage unless the axial depth of cut is reduced, thus further reducing the MRR and increasing the processing time. 3.1.4 High Cutting Force Coefficients Commercially available micro-end mills typically have cutting edge radii on the order of 2 to 3m, or 2 to 6 times the chip thickness. This means that the effective rake angle of the cutting edge is highly negative, on the order of to degrees. In this situation, the conventional models of the mechanics of chip formation do not apply, and

PAGE 23

15 the cutting force coefficients are typically 20 to 40 times higher than in conventional milling. 3.1.5 Low Cutting Speed In order to achieve efficient cutting there exists preferred ranges of tangential speeds of the cutting edge for different work piece and tool material combinations. Spindle speed must increase as tool diameter decreases in order to achieve the desired cutting speed. For instance, when using carbide tools to machine aluminum, the recommended cutting speed is on the order of 500 meters/minute. To achieve this cutting speed with a tool diameter of 0.25 mm, the required spindle speed is over 600,000 rpm. This speed is unachievable with current machine tool spindle technology. 3.1.6 Unpredictable Tool Life Unpredictable tool life and premature tool failure are the major concerns in micro machining using micro grain carbide cutters. The reasons for these premature failures are believed to be related to the low chip thickness relative to the tool edge radius, the large tool run-out relative to the chip thickness, and the low cutting speeds achievable with current micro-milling spindles. 3.2 Conceptual Design Based on the challenges outlined above, it appears that the key bottleneck to efficient use of micro-milling for metallic materials appears to be related to the spindle. Spindles with low run-out and speeds on the order of 500,000 rpm are desired for this application. 3.2.1 Conceptual Design The conceptual design for the ultra-high speed micro-milling spindle is shown in Figure 3-1.

PAGE 24

16 Figure 3-1. 3D model of ultra-high spindle micro-milling spindle system.

PAGE 25

17 The spindle system uses an air bearing to support the 1/8th inch tool shank. The tool shank itself is driven by a friction drive with the drive wheel driven by a commercial high speed spindle. The drive ratio of the system is 1:8 and so it is possible to achieve speeds over 500k rpm with the commercially available high speed spindles rotating as speeds of 60-90 k rpm. Commercially available micro-milling cutters typically have a 0.125-inch diameter shank, approximately 1.5 inches long. The tool shank effectively becomes the spindle shaft and should greatly reduce tool point radial run-out. 3.2.2 Design Description The major components of the proposed spindle system are: A custom porous carbon air-bearing to support 1/8 diameter shank micro tools. A 3-axis force sensor capable of sensing in-situ milling forces. A high speed drive motor capable of producing a torque of 0.01 N-m at 50K rpm. A split mount for mounting the high speed drive motor. A friction wheel coated with suitable polymer, interfaced to the high speed drive motor. A pair of spherical washers and fine adjustment set screws for adjusting the motor axis relative to the tool shank axis in two perpendicular directions. A frame for mounting the components of the system. A computer interface to control and display spindle speed, monitor tool forces during milling, and provide means to interpret and feedback milling

PAGE 26

18 forces to identify tool breakage and identify and control tool over/under-loading.

PAGE 27

CHAPTER 4 DETAILED SPINDLE DESIGN AND ASSEMBLY Detailed design of the major components and the assembly of the parts are discussed in this chapter. The order of these discussions follows the logical sequence required to assemble and align the machine. Figure 4-1 shows the complete assembled machine. Following good design and manufacturing practices from industry, part numbers are assigned to all parts purchased, manufactured, or borrowed, for the purposes of record keeping. A list of all parts and assemblies appears in Appendix A. Only assemblies of particular relevance are described in detail, since this thesis is not meant to be an exhaustive assembly manual. The following scheme is used to designate part numbers (P/N): Manufactured parts: These parts were designed from scratch and required detail engineering drawings so as to be fabricated by outside vendors. The nomenclature that was used to designate these parts has the format MSDXX. Here MSD stands for Micro-spindle Design and two XXs stands for two digited numbers between 00-99. For example a valid part number is MSD01 or MSD99 etc. Purchased parts: These parts are purchased directly from vendors. A specification sheet from the manufacturer or vendor replaces a detailed drawing. This apart all major purchased parts are supported by an outline engineering drawing. The naming convention for the purchased parts follows 19

PAGE 28

20 the format MSDPXX. Here MSD stands for Micro-spindle Design, P stands for Purchase and two XXs stand for two digited numbers between 00-99. For example a valid part number is MSDP01 or MSDP99 etc. Components already belonging to the university were treated as purchased parts when assigning part numbers. Figure 4-1. The assembled micro-milling spindle.

PAGE 29

21 Fasteners: Although purchased, in order to distinguish them from other purchased parts, these parts have naming convention of MSDFXX. Here the first three letters have the same usual meaning whereas F stands for Fasteners. The other two XXs are as before numbers. Assemblies: Sub-assemblies and the consequent final assembly have a nomenclature of MSDAXX where A stands for assembly and other letters are defined as before. All the parts may be found in the appendix according to their logical order. Part numbers are given next to part names throughout this thesis. Figure 4-2. The assembled micro-milling spindle system mounted to HSM-2.

PAGE 30

22 4.1 Base Frame The L shaped base frame (P/N MSD02) supports the whole spindle system. The following figure shows the CAD model of the base frame. This has a recessed hole on top to accept the female half of the spherical washer set (P/N MSDP10). The recess is designed to provide a light press fit of the spherical washer into the base frame. This design makes sure the axis of the motor (P/N MSDP01) is located precisely with respect to the base frame. Figure 4-3. 3D model of the base frame. 4.2 Air Bearing The air bearing (P/N MSD03) is a special part which is custom designed specially for this spindle system and fabricated by New Way Bearings Inc. Figure 4.3 shows the solid model of the air-bearing. The external dimensions were supplied to the vendor, who then designed all of the internal details. Figure 4.4 shows the actual drawing of the air bearing.

PAGE 31

23 Figure 4-3. 3D model of the air bearing. The bearing housing is made of aluminum. The air bearing system has a porous carbon insert with a central hole for 1/8 diameter tool shank and also a thrust bearing on top for the head of the tool to seat in. The radial and axial air bearings are intended to ensure that there is effectively no friction between the tool and the bearing surface. Figure 4-4. Isometric view of the air bearing.

PAGE 32

24 For the air bearings to work properly the manufacturer has recommended a shaft tolerance of +/-0.00015 inches. However, the commercially available micro-tools have a shaft tolerance of +/-0.0005 inches. Therefore, a collection of micro-tools was purchased and the shaft diameters measured. The air bearing bore was sized to fit these tools. The micro-tools float freely inside the air-bearing when it is supplied with air, indicating that the bearing is sized appropriately. Lastly, the bearing must be supplied with filtered and moisture free air pressurized to 100 psi. It must be free from oil vapor, water and dust particles. Figure 4.5 shows the air filtration system. From top to bottom the components are two pressure regulators, a compressed air-filter, a desiccant air dryer, a coalescing filter, an oil-vapor removal filter and finally a valve with two 1/16th inch hose fittings. The direction of air flow is shown by the arrow in the figure. Figure 4-5. Air cleaning system and valve.

PAGE 33

25 4.3 Force Sensing System In order to measure the dynamic or quasistatic cutting forces during machining a Kistler 3-axis miniature force sensor (P/N MSDP06) is sandwiched between the air bearing and the base frame. The piezoelectric force sensor has a resolution of 0.01N and a measuring range of +/-1kN. Technical data about the sensor is available in manufacturers specification sheet. When a force is exerted on the sensor, the piezo-electric material generates an electric charge. A PCB charge amplifier (P/N MSDP08) converts charge developed by the sensor into proportional DC voltage. This voltage is read through three input channels of a LabView DAQ card. Figure 4.6 shows the picture of the force sensor. Figure 4-6. Three axes force sensor.

PAGE 34

26 4.4 Assembly of Base Frame, Air-Bearing and Force Sensor Kistler [22] recommends special mounting instructions for the force sensor which are incorporated in the design and assembly. The sensor is mounted under preload because the shear forces Fx and Fy must be transmitted through static friction from the bearing and the base frame to the surfaces of the force sensors. A 13 kN preload is recommended in order to use the full measuring range of the forces, however we preloaded it to 11.2 kN as the expected range of forces that we will be measuring is less than 20 N. The preloading is carried out per Kisler recommendations with the help of a torque wrench set to 10 N-m, corresponding to a force of 11.2 kN. The relationship between torque and axial preload is given by: T = rF (4.1) where, T is the torque in N-m, F is the preload force, is coefficient of friction and r is the preloading bolt diameter. The preloading bolt diameter was measured and be 5.952 mm. The coefficient of friction suggested by Kistler was 0.15, yielding a preload force of 11.2 kN corresponding to an applied torque of 10 N-m. Figure 4-7 shows the picture of the assembly. Figure 4-7. Assembly of base frame, air-bearing and force sensor.

PAGE 35

27 4.5 Motor Mount A single piece, split ring motor mount (P/N MSD01) is used to mount the Precise drive spindle (P/N MSDP01). The mount is designed according to the manufacturers recommendations [23]. Figure 4.8 shows the CAD model of the motor mount. Appendix B shows the detailed engineering drawing of the mount. Figure 4-8. CAD model of the motor mount. The mount has 2 clamping screws and one spacer screw. The clamping screws are used to assemble the motor to the mount whereas the spacer screw helps in dissembling it from the mount. The torque wrench is used to provide the necessary torque during clamping according to the manufacturers data sheet for the motor. Figure 4.9 shows a photograph of the motor mount.

PAGE 36

28 Figure 4-9. Motor mount. 4.6 Assembly of Motor Mount with Base Frame The male central spherical washer is seated on the base frame and then the motor mount is placed on top it. The four smaller sets of spherical washers (Figure 4-10) are inserted in the four holes of the motor mount from the top. Figure 4-10. Spherical washer set.

PAGE 37

29 The four holes of the motor mount are aligned with corresponding holes in the base frame and they are secured with the help of the 4 fine adjustment (80 threads per inch) set screws (Figure 4-11). Figure 4-11. Fine adjustment set screw. Final assembly is shown in Figure 4-12. Figure 4-12. Assembly of motor mount and base frame.

PAGE 38

30 4.6 Assembly of Precise Spindle with Base Frame and Friction Wheel The motor arbor (P/N: MSDP02) is secured to the Precise motor (P/N: MSDP01) according to the manufacturers instructions in the data sheet [23]. Then the clamping screws of the motor mount are loosened and the spacer screw is tightened to spread the clamp, until the motor can slide inside the inner hole of the motor mount. In this position the friction wheel (P/N MSDP05) is secured at the motor arbor end. Figure 4-13 shows the photo of the friction wheel and corresponding CAD model. The detailed engineering drawing of the friction wheel is available in Appendix B. Figure 4-13. Friction wheel. The motor is positioned properly inside the clamp and, the spacer screw is loosened and the clamping screw is tightened using a torque wrench to 2.8 N-m of torque as recommended by the manufacturer [23]. This procedure ensures that the bearings inside the drive motor are not loaded improperly by clamping forces on the motor housing. Finally, the clamping screw of the friction wheel is secured to the arbor of the drive motor using a special wrench provided by the manufacturer. The final Assembly is shown in Figure 4-1. 4.7 Assembly of Micro-Spindle with HSM-2 An interface plate (P/N MSD10) was designed to mount the micro spindle on HSM-2. Figure 4-14 shows the CAD model of the interface plate. A narrow shoulder on

PAGE 39

31 the plate ensures that the plate is vertically aligned to one of the vertical faces of HSM-2. The micro-spindle is secured to the mounting plate using the four bottom holes. Figure 4-14. Interface plate. 4.8 Micro-Spindle System Connections The spindle drive system consists of a solid state frequency converter (P/N MSDP03) and associated electrical connections, liquid cooling connections, etc. These are discussed in detail in the manufacturers instruction manual [24-25]. The frequency converter basically controls the motor speed. It can be operated in two different modes, manual remote control. The spindle drive is connected to the National Instruments DAQ card channels using the remote control connector, and is interfaced with appropriate LabView VIs. Figure 4-15 shows the general schematic diagram of the spindle system.

PAGE 40

32 Figure 4-15. Schematic of system connections.

PAGE 41

33 4.9 Computer, Electronics and Software A 300 MHz Pentium laptop computer and National Instruments LabView and DAQ card [26] was used to control and monitor the spindle system, record cutting forces and display results. National Instruments NI DAQ -1200 card [27] was chosen, which features digital triggering capability; three 16-bit, 8 MHz counters/timers; two 12 bit analog output channels; 24 digital I/O lines and four12 bit differential analog input channels. Three of the input channels are used for force measurements in x, y and z direction. One output channel is used to provide a variable DC voltage from zero to ten volts to command the Precise Motor to rotate from zero to 90k rpm. The two frequency counters are used for measuring the actual speed of the Precise Motor. Five of the digital lines of port A are used for creating virtual LEDs in LabView VI for various warnings as recommended in the Precise Frequency Converters instruction manual. A breadboard circuit is used for all electrical connections. It consists of a 7404 inverter chip, which is a hardware aid for measuring the frequency of rotation of the Precise motor. There are five registers of 100k ohms each for electrical tuning of the circuit designed for setting the warning signals during the operation of the motor. Figure 4-16 shows the picture of the circuit. The National Instruments data acquisition card is connected to an I/O connector block. The breadboard and the I/O connector are mounted inside an enclosure which has a 15 pin connector for the Precise motor speed control and measurement, three BNC jacks for force measurements, and one on/off switch. Figure 4-17 shows the components mounted in the enclosure.

PAGE 42

34 Figure 4-16. Bread board circuit. Figure 4-17. Enclosure containing i/o connector and bread board.

PAGE 43

35 LabView VIs are developed for spindle speed control and measurement, cutting force measurements and warning signals during spindle rotation through virtual LEDS. Figure 4-18 shows the LabView interface. Figure 4-18. Labview interface for spindle control and measurement. In the above figure the speed control knob can set a variable voltage between zero to ten volts and the corresponding commanded speed is visible from the waveform chart. The actual rpm is also shown in the waveform chart. There is a digital display for the both the commanded and actual rpm. LED K1 glows red when the actual spindle speed is zero although the commanded is not. LED K2 glows red when the set load limit is exceeded for the motor. LED K3 turns red in the event of a load change. Similarly K4 is green while the unit is operational and K5 becomes green when the actual speed is nearly

PAGE 44

36 equal to commanded speed. Separate sub VIs were developed for these functions. Figure 4-19 shows the picture of the force measurement VI. Figure 4-19. Labview interface for cutting force measurement. Once the machine was fully assembled and aligned, it was ready for testing. Chapter 6 describes initial testing of the machine at the University of Florida Machine Tool Research Center.

PAGE 45

CHAPTER 5 DESIGN VALIDATION This chapter discusses the validity of the design. The design is validated by two finite element analyses carried out to better understand the system dynamics. Other calculations are also performed to analyze stresses and strains in rotating parts at high rpm, in order to avoid material failure. The results are incorporated in the design. 5.1 FEA of Motor and Friction Wheel Assembly To analyze the natural frequencies of the drive spindle, a finite element analysis of the motor, motor arbor, and friction wheel assembly was carried out by Mr. Paul Frederickson of Precise Corporation, who was kind enough to do this analysis. The first analysis uses a steel friction wheel attached to the end of the motor arbor (P/N MSDP02), which is again assembled to the motor (P/N MSDP02), as shown in the Figure 5-1. The analysis predicts the first natural frequency occurs at 74,000 rpm. In order to achieve higher speed, the friction wheel material was replaced by aluminum, resulting in a lighter part. The second analysis using the aluminum friction wheel predicts the first natural frequency to be 84,000 rpm. Figure 5-2 shows the results of the second analysis. The 2nd and 3rd natural frequencies are also shown, as well as the calculated shaft deflection with a 10 N radial load applied to section #2. 37

PAGE 46

38 Motor Friction W h eel Motor A rbo r Figure 5-1. Motor, motor arbor and friction wheel assembly for FEA. Figure 5-2. Results of FEA of motor and friction wheel assembly1 1 Courtesy Paul Frederickson, Precise Corporation, Racine,Wisconsin, USA.

PAGE 47

39 5.2 FEA of the Micro-Tool, Air-Bearing and the Friction Drive In order to know the dynamic response of the micro-tool rotating at high rpm in the air-bearing, a separate finite element analysis was carried out. Figure 5-3 shows the FE model. Figure 5-3. FEA model of the tool, air-bearing and friction wheel. The tool inside the air-bearing is modeled as a continuous beam supported on springs with composite stiffness equivalent to the air bearing stiffness. The stiffness of the custom air-bearing was predicted to be approximately 5e5 N/m by extrapolating from published stiffness values of other air bearings produced by the vendor. This value was further verified by conducting stiffness measurements after assembly and testing of the real bearing when it was available. Chapter 6 discusses the stiffness measurements in detail. The beam is also contacted by another spring with stiffness equivalent to the Hertzian contact stiffness between the tool and the friction wheel. For two cylinders in contact the Hertzian contact stiffness is given by the formula: )1()1(333.121222121EEErEKeff (5.1) Also, 2121rrrrr (5.2) where

PAGE 48

40 1E = Modulus of elasticity of first cylinder. 2E = Modulus of elasticity of second cylinder. 1= Poissons ratio of first cylinder. 2= Poissons ratio of second cylinder. Although the exact material properties of the ML6 polymer coating on the friction wheel were not available from the vendor, it was possible to find the material properties of urethane. The polymer ML6 is a urethane class material. The modulus of elasticity of urethane was found to be 2e7 N/m2, and Poissons ratio was 0.4. The properties of tungsten carbide tool are E = 6.8e11 N/m2 and = 0.24. Using equation 5.1 and 5.2 the contact stiffness is calculated to be 1.68e6 N/m. The FEA analysis predicts the lowest natural frequency of the system to be 4857 Hz. This corresponds to a rotational speed of approximately 292,000 rpm. The analysis was also performed for stiffness values of the air bearing increasing by up to 100%. As expected, the first natural frequency increased as the stiffness of the air bearing increased. This points to the need to increase the air bearing stiffness in future versions. Figure 5-4 to Figure 5-7 shows the results with different mode shapes in different frequencies. Figure 5-4. First mode: 4857.3 Hz.

PAGE 49

41 Figure 5-5. Second mode: 1.35e5 Hz. Figure 5-6. Third mode: 2.65e5 Hz. Figure 5-7. Fourth mode: 3.89e5 Hz. 5.3 Stress/Strain Analysis The friction wheel rotates at a very high rpm. So a stress/strain analysis was pertinent to make sure the design is safe. Hoop stress and radial strain is calculated in the friction wheel made of aluminum and rotating at 75,000 rpm, which is suppose to be the maximum operating speed. From the theory of thick cylinders and rotating disks [28] the equation for hoop stress for a solid disk is given by: = 2/8[(3 + )ro2 (1 + 3)r2] (5.3) where, is hoop stress. is the density of the material.

PAGE 50

42 is the Poissons ratio. ro is the outer radius of the disk. Now, putting r = ro in the above equation and solving for gives: = 2/4(1 )ro2 (5.4) Now again putting = 7850 rads/sec (75k rpm), = 2698.9 kg/m3 (density of aluminum) and = 0.36 and ro = 12e-3 m gives = 38.31 MPa, (5.5) which is 50% of the tensile strength of aluminum. Hence the friction wheel is safe at this rpm. The radial displacement is also calculated using the following formula: urr = r/E( rr) (5.6) where E is the modulus of elasticity of aluminum. rr is radial stress (= 0 in this case). So, at 75k rpm urr = 6.76 m. At 50k rpm this value reduces to 3 m. Based on these analyses, the design is safe.

PAGE 51

CHAPTER 6 TESTING AND RESULTS This chapter describes the testing of the first generation ultra-high-speed micro-spindle. All initial testing is carried out at University of Florida Machine Tool Research Center. Figure 6.1 shows the experimental set-up. Run-out measurements, cutting force measurements, and air-bearing stiffness measurements comprise the sections of this chapter. All the cutting tests are carried out using aluminum as the workpiece material. Other materials such as mild steel were also tried but gave unsatisfactory results. It is believed this is due to insufficient stiffness in the air-bearing. 6.1 Run-out Measurements Radial run-out measurements were carried out using a Lion Precision capacitance probe and amplifier setup. The amplifier was interfaced with PCScope. The data was recorded at a sampling rate of 1x105 Hz. The run-out measurements were carried out at 3 different places. 6.1.1 Run-out Measurement at Motor Arbor Initially, the radial run-out of a point on the drive spindle arbor was measured at various spindle speeds both with and without the friction wheel mounted. The results are shown in Figure 6-2. Without the friction wheel attached to the arbor the run-out remains less than two microns, which is the value given in the manufacturers specification sheet for the motor. With the friction wheel installed and run-outs measured at the same point on the arbor, there is a clear parabolic increase in radial run-out with increasing speed. 43

PAGE 52

44 This indicates an unbalance in the friction wheel with centrifugal forces causing increased run-out at higher speeds. Figure 6-1. Experimental set-up. Drive spindle arbor radial runout02468101220000300004000050000Spindle speed (rpm)Radial runout (micrometers) Without friction wheel With friction wheel Figure 6-2. Run-out of drive spindle arbor.

PAGE 53

45 6.1.2 Run-out Measurements at Friction Wheel and Gage Pin The radial run-out of the friction wheel surface was measured using a capacitance probe reading against the wheel surface through the polymer coating. These results are somewhat unreliable since they contain effects both due to the non-circularity of the metal portion of the friction wheel, and thickness variations of the polymer coating. The yellow line in the Figure 6-3 shows the run-out at the friction wheel. Finally, a gage pin was inserted in the air bearing and driven by the friction wheel to measure run-out near the location of the tool point, approximately 6mm from the end of the air bearing. The results of these measurements are shown in Figure 6-3 by a blue line. Radial runout of friction wheel surface and gage pin02040608010012020000300004000050000Spindle speed (rpm)Raidal runout (micrometers) Friction wheelsurface Gage pin tip Figure 6-3. Run-out of friction wheel and gage pin. These results show that the friction wheel surface is far too irregular for this application and the imperfections in its surface are transmitted into the tool shank and amplified at the tool tip. Further analysis of the air bearing and friction wheel, and improved manufacturing methods for the friction wheel are needed to rectify this problem.

PAGE 54

46 6.2 Force Measurements Although the tool tip run-out was exceedingly high as discussed above, preliminary cutting tests were performed to evaluate the cutting performance of the micro spindle. The initial cutting tests were performed at a drive spindle speed of 10,000 rpm which yielded a nominal tool speed of approximately 79,000 rpm. Slotting cuts were performed using tool diameters ranging from 0.03(762m) inches to 0.005(127m) inches. Commanded slot depths of one-sixth of the tool diameter, and one third of the tool diameter were used. Subsequent measurements of the slots using a measuring microscope showed a significant variation in the slotted depths. For each tool diameter, depth, and width of cut, the feed rate was increased until the tool failed. Figure 6-4 shows the feed direction force record for 127m cutter with a 42.33m slot depth and the largest achievable feed rate without tool breakage of 0.045 mm/min (0.376 m/tooth). The passage of individual teeth is clearly evident in the force record, and shows that the actual tool speed is approximately 67568 rpm, indicating significant slippage in the friction drive. Significant variation in the peak cutting force is also evident, and may be due to asynchronous error motions of the tool. Typical cutting test results are shown in table 6-1. Both Tlusty and MacNeil [5] and Bao and Tansel [10] cutting force models were used to calculate the cutting force coefficient, Ks, using the average maximum feed direction force. The average maximum feed direction force was found using a Matlab code to average the maximum force for each tooth passage over a substantial number of tooth passages. For milling with macro tools, Ks for this material will be on the order of 850 N/mm2. These results are somewhat unreliable due to fact that measured forces are different than the actual cutting forces. The air bearing sitting on top of the force sensor affects the

PAGE 55

47 dynamics of the system. Figure 6-5 shows the mathematical model of the tool/air-bearing/force sensor assembly. A hammer testing experiment needs to be done to actually get the transfer function of the system. This transfer function needs to be inverted and then multiplied with measured forces in frequency domain to get cutting forces in frequency domain. Again these cutting forces in frequency domain need to be converted to forces in time domain which will be actual cutting forces. Figure 6-4. Force plot for micro-milling of Al using 127m micro-tool.

PAGE 56

48 Figure 6-5. Mathematical model of the tool, air-bearing and force sensor system. Table 6-1. Cutting test results Aluminum 6061 Dia Width Depth Speed Feed Feed Avg Max Cutting Cutting of cut of cut /tooth Force Coeff. Coeff. (m) (m) (m) (rpm) (mpm) (m) (N) (N/mm^2) (N/mm^2) (Tlusty) (Tansel) 762 762 70.22 60976 0.144 1.18 2.8157 58200 4630 762 762 73.13 60976 0.203 1.66 6.0838 85600 7100 508 508 91.44 66195 0.296 2.24 5.0212 42000 6530 508 508 93.98 59524 0.313 2.63 6.3418 43900 7010 254 254 104.1 67568 0.0593 0.439 0.8811 33000 11700 254 254 106.7 65789 0.0762 0.579 1.1787 32600 11900 127 127 42.33 67568 0.0253 0.187 0.6678 144000 41600 127 127 42.33 67568 0.0508 0.376 1.1292 121000 35000

PAGE 57

49 The actual speed of the tool rotation is back calculated from the tooth passing frequencies obtained from the force plot. Figures 6-6 to 6-7 show photographs of some of the machined slots. Figure 6-6. Slots made from 508m micro-tool. Figure 6-7. Slots made from 254m micro-tool.

PAGE 58

50 The actual depth of cut and width of cut of these slots was also measured using a measuring microscope and table 6-1 shows the actual values rather than commanded ones. 6.3 Machining of Hexagonal Test Part In order to test the capability of the spindle to create miniature features a hexagonal feature was machined in aluminum using a 254m micro tool with a wall thickness of 80 m. Figure 6-8 shows the close up photograph of this feature. Figure 6-8. Hexagonal feature with a wall thickness of 30m. The total depth of the feature was created in two layers, with each layer having a depth of cut equal to half of the diameter of the tool. A constant feed rate of 0.0593 mpm was used. A number of tests were carried out to determine the minimum wall thickness that can be machined. To accomplish this, adjacent slots were widened in steps of 10m using 254m micro tools. Using this procedure, it was determined that the minimum

PAGE 59

51 thickness of the wall that can be machined is 30m. Figure 6-8 shows results from some of these trials. We were not able to achieve this thickness in the hexagonal feature due to non repeatable error motions of the X and Y axes of HSM-2. The machined part has a commanded wall thickness of 80 m. Measurements the machined walls showed thickness variations ranging from 30m to 120m. 6.4 Air-Bearing Stiffness Measurements The stiffness of the air-bearing was measured to compare with values predicted by extrapolation of manufacturers data for larger diameter bearings. The set-up was similar to the one used for the run-out measurements. For the stiffness measurement, one of the PCScope channels was used to measuring the Z-direction force. The other channel was used for displacement measurement in the same direction using the capacitance probe. The charge amplifier was set to the highest time constant (5 sec) making it suitable for quasi-static force measurements. The bearing was pressurized and the tool was mounted in the air-bearing. The tool shaft was manually loaded near the friction wheel using a screw driver while the force and displacement were recorded. The experiment was repeated a number of times. The recorded data was transferred to Matlab and force versus displacement plots were created. Figure 6-9 shows a typical graph.

PAGE 60

52 Figure 6-9. Force vs. displacement plot to measure the stiffness. The slope of the above curve shows the stiffness to be approximately 3.8e5 N/m. The measured values of air bearing stiffness were very close to the values predicted by extrapolating from manufacturers data, as described in the previous chapter. Stiffness measurements were also carried out without air pressure in the air bearing and resulted in values approximately one order higher (6e6 N/m to 7.5e6 N/m), indicating that the measured stiffness of the pressurized bearing was primarily due to displacements of the tool shaft within the air bearing bore.

PAGE 61

CHAPTER 7 CONCLUSIONS AND FUTURE WORK This chapter summarizes the results from initial testing of the micro-spindle, and provides recommendations for future work. The friction wheel shows excessive run-out which increases with increasing RPM, due to high centrifugal forces resulting from wheel unbalance. At drive spindle speeds greater than 20k rpm it appears that the tool shank begins to contact the inner surface of the air bearing, which leads to heat generation, restricting the spindle to lower speeds in order to ensure that the tool floats. This is believed to be related to the problems with friction wheel run-out and unbalance, which are exacerbated at higher speeds. Run-out measurements using the gage pin gives values much higher than the target values for the spindle. It is believed that this is due to excessively high contact stiffness between the tool and the polymer coated friction wheel, leading to transfer of the error motions of the friction wheel into the tool. Despite this, the spindle is capable of operating at speed in excess of 500,000 rpm for short periods. We believe that design improvements outlined below will allow the micro-spindle to operate successively at much higher speeds and with low tool run-out. Suggestions for the future designs of the spindle are: Redesign the friction wheel to achieve proper balance when mounted to the motor arbor. Change the friction coating on the friction wheel to reduce the contact stiffness between the tool and the friction wheel. 53

PAGE 62

54 Improve manufacturing methods for the coating (such as abrasive finishing) to achieve improved concentricity with respect to the friction wheel. Redesign the air bearing to increase its stiffness. Commercially available micro-tools have large shaft diameter variations which may limit achievable air bearing stiffness, since high stiffness requires small air gaps. This may require special tools with precision ground shaft diameter. Alternatively, special precision ground collars should could be used to support the commercial tools. Incorporate a tool speed sensor in order to know the actual speed of the tool rotation. The current design senses the rotational speed of the drive motor. However, the actual speed of tool rotation is unknown if slip in the friction drive. Replace the current charge amplifier with charge amplifiers which have a large range of adjustable time constant. This will allow the force sensor to provide better quasistatic force measurements for setting the contact between the friction wheel and the tool. Incorporate sensors to measure the error motions and run-out of the tool shaft during high speed rotation.

PAGE 63

APPENDIX A PART LIST

PAGE 64

56 P/N DESCRPTION QTY VENDOR VENDOR P/N COST EA. COST MSD01 SPLIT MOUNT, MOTOR 1 PENNSYL. TOOL 1000.00 1000.00 MSD02 BASE FRAME 1 PENNSYL. TOOL 800.00 800.00 MSD03 CUSTOM AIR BEARING ASSEMBLY 1 NEWWAY 3000.00 3000.00 MSD05 FRICTION WHEEL 1 SANDIA 1000.00 1000.00 MSD06 POLYMER COATING 1 MERIDIAN LABS. 320.00 320.00 MSD10 INTERFACE PLATE 1 PENNSYL. TOOL 500.00 500.00 MSDP01 SC40 SPINDLE, WITH POWER CABLE 1 PRECISE CORP CAT # 550279 3600.00 3600.00 MSDP02 MOTOR ARBOR 1 PRECISE CORP CAT # 400510 245.00 245.00 MSDP03 FREQUENCY CONVERTOR, TYPE PCF 310 1 PRECISE CORP CAT # 330916 3600.00 3600.00 MSDP04 COOLING SYSTEM, 115V, TYPE 7136 1 PRECISE CORP CAT # 332370 1900.00 1900.00 MSDP05 BASE TUBING KIT 1 PRECISE CORP CAT # 332004 45.00 45.00 MSDP06 3 AXIS FORCE SENSOR 1 KISTLER 9018A 3168.00 3168.00 MSDP07 FORCE SENSOR CONNECTING CABLES 1 KISTLER 1694A 366.00 366.00 MSDP08 CHARGE AMPLIFIER, PCB 3 MTRC LAB 443B101 0.00 0.00 MSDP09 SPHERICAL WASHER, LARGE, MALE 1 J.W. WINCO 43NG40/CNI 54.68 54.68 MSDP10 SPHERICAL WASHER, LARGE, FEMALE 1 J.W. WINCO 49NG40/DNI 55.78 55.78 MSDP11 SPHERICAL WASHER, SMALL, MALE 4 J.W. WINCO 64NG40/CNI 1.70 1.70 MSDP12 SPHERICAL WASHER, SMALL, FEMALE 4 J.W. WINCO 71NG40/DNI 1.79 1.79 MSDF13 FINE ADJUSTMENT SET SCREW, 1/4-80, 2'' LONG 4 THORLABS FAS200 8.20 32.80 MSDP14 MICRO TOOLS, DIA. 0.005'' 15 NATIONALTOOL 15.75 236.25 MSDP15 MICRO TOOLS, DIA. 0.010'' 15 NATIONAL TOOL ITEM # 32352 15.50 232.50 MSDP16 MICRO TOOLS, DIA. 0.015'' 15 NATIONAL TOOL ITEM # 32357 15.50 232.50 MSDP17 MICRO TOOLS, DIA. 0.020'' 15 NATIONAL TOOL ITEM # 32362 15.00 225.00 MSDP18 MICRO TOOLS, DIA. 0.030'' 15 NATIONAL TOOL ITEM # 32372 12.50 187.50 MSDP19 CAPACITANCE PROBE, 3/8'' CYLINDRICAL 1 MTRC LAB. C1C 0.00 0.00 MSDP20 CHARGE AMPLIFIER, CAPACITANCE PROBE 1 MTRC LAB. DMT20 0.00 0.00 MSDP21 GAGE PIN, 0.1247'', CLASS XX 1 MEYER GAGE 0.1247X CLX 19.00 19.00 MSDP22 GAGE PIN, 0.1248'', CLASS XX 1 MEYER GAGE 0.1248X CLX 19.00 19.00 MSDP23 GAGE PIN, 0.1249'', CLASS XX 1 MEYER GAGE 0.1249X CLX 19.00 19.00 MSDP24 GAGE PIN, 0.1251'', CLASS XX 1 MEYER GAGE 0.1251X CLX 19.00 19.00

PAGE 65

57 MSDP25 GAGE PIN, 0.1252'', CLASS XX 1 MEYER GAGE 0.1252X CLX 19.00 19.00 MSDP26 GAGE PIN, 0.1253'', CLASS XX 1 MEYER GAGE 0.1253X CLX 19.00 19.00 MSDP27 GAGE PIN, 0.1254'', CLASS XX 1 MEYER GAGE 0.1254X CLX 19.00 19.00 MSDP28 GAGE PIN, 0.1255'', CLASS XX 1 MEYER GAGE 0.1255X CLX 19.00 19.00 MSDP29 GAGE PIN, 0.1256'', CLASS XX 1 MEYER GAGE 0.1256X CLX 19.00 19.00 MSDP30 GAGE PIN, 0.1257'', CLASS XX 1 MEYER GAGE 0.1257X CLX 19.00 19.00 MSDP31 HOSE NIPPLE, 1/4'' HID,1/4'' PIPE 14 MCMASTER 5346K14 0.45 6.30 MSDP32 HEX COUPLING, 1/4'' X 1/8'' 1 MCMASTER 9171K72 3.90 3.90 MSDP33 COMP. AIR FILTER 1 MCMASTER 4274K13 20.74 20.74 MSDP34 DESICCANT AIR DRYER 1 MCMASTER 5164K77 41.17 41.17 MSDP35 COALESCING FILTER, OIL REMOVAL 1 MCMASTER 59185K11 83.10 83.10 MSDP36 OIL VAPOR FILTER 1 MCMASTER 5757K61 33.94 33.94 MSDP37 REGULATOR, WITH GAGE 2 MCMASTER 9892K11 21.00 21.00 MSDP38 LEVER CONTROLLED REGULATOR 1 MCMASTER 4158K72 22.26 22.26 MSDP39 FITTING, 10-32 THREAD TO 1/16'' ID HOSE 2 MCMASTER 5454K61 0.48 0.96 MSDP40 BRAID-REINFORCED PVC TUBING 50 MCMASTER 52375K12 0.37 18.50 MSDP41 POLYURETHANE TUBING, 1/16'' ID, 1/8'' OD 30 MCMASTER 5184K61 0.26 7.80 MSDP42 CABLE TIES, LONG 50 MCMASTER 7130K16 0.0804 4.02 MSDP43 CABLE TIES, FINE 100 MCMASTER 7130K12 0.0180 1.80 MSDP44 BLACK BOX 1 RADIO SHACK 6.23 6.23 MSDP45 PUSH BOTTON SWITCH 1 RADIO SHACK 3.23 3.23 MSDP46 7404 HEX INVERTOR CHIP 1 ELECTRONIC PLUS 0.78 0.78 MSDP47 BREAD BOARD 1 MTRC LAB 0.00 0.00 MSDP48 REGISTORS 5 MTRC LAB 0.00 0.00 MSDP49 NI, DATA ACQUISITION CARD, 1200 SERIES 1 MTRC LAB 0.00 0.00 MSDP50 RIBBON CABLE 1 MTRC LAB 0.00 0.00 MSDP51 I/O CONNECTOR 1 MTRC LAB 0.00 0.00 MSDP52 15 PIN CONNECTOR CORD 1 ELECTRONIC PLUS 11.75 11.75 MSDP53 15 PIN FEMALE JACK 1 MTRC LAB 0.00 0.00 MSDP54 BNC JACKS 3 MTRC LAB 0.00 0.00 MSDP55 BNC CONNECTING CABLES 3 MTRC LAB 0.00 0.00 MSDP56 LAPTOP 1 MTRC LAB 0.00 0.00

PAGE 66

58 MSDF57 HEX BOLT, M8 X 1.0 4 MSC 0.42 1.68 MSDF58 HEX BOLT, 13, 3'' LONG 2 MTRC LAB 0.00 0.00 MSDF59 WASHER, 1/2-13 BOLT 2 MTRC LAB 0.00 0.00 MSDF60 SOCKET HEAD SCREW, M5 X 0.8 3 MTRC LAB 0.00 0.00 MSDF61 SOCKET HEAD SCREW, 1 MTRC LAB 0.00 0.00 TOTAL 21283.66

PAGE 67

APPENDIX B DETAIL DRAWINGS

PAGE 68

60

PAGE 69

61

PAGE 70

62

PAGE 71

63

PAGE 72

64

PAGE 73

1. .XXX :+/.005 2. .XX :+/.01 3. .X :+/.1 4. Material: Steel 5. Dimensions in mm. Note: 65

PAGE 74

LIST OF REFERENCES 1. Friedrich, C. R. and Vasile, M. J., 1996, Development of the micro milling process for high aspect ratio microstructures, Journal of Microelectromechanical Sytems, v 5, n 1, pp. 33-38. 2. Adams, D. P., Vasile, M. J., Benavides, G., and Campbell, A. N., 2001, Micromilling of metal alloys with focused ion beam-fabricated tools, Precision Engineering, v 25, pp. 107-113. 3. Vasile, M. J., Friedrich, C. R., Kikkeri, B., and McElhannon, R., 1996, Micrometer-scale machining: tool fabrication and initial results, Precision Engineering, v 19, pp. 180-186. 4. Friedrich, C. R., Coane P., Goettert J., and Gopinathin N., 1998, Direct fabrication of deep x-ray lithography masks by micromechanical milling, Precision Engineering, v 22, pp. 164-173. 5. Tlusty, J. and MacNeil, P., 1975, Dynamics of cutting forces in end milling, Annals of the CIRP, v 24, n 1, pp. 21-25. 6. Gygax, P. E., 1979, Dynamics of single-tooth milling, IWF/ETH Zurich, Annals of the CIRP, v 28, n 1, pp. 65. 7. Kline, W. A. and DeVor, R. E., 1983, The effect of cutter runout on cutting geometry and forces in end milling, International Journal of Machine Tool Design and Research, v 23, pp. 123. 8. Tlusty, J. and Ismail, F., 1983, Special aspects of chatter in milling, Trans. of ASME, Journal of Engineering for Industry, v 105, pp. 24. 9. Ismail, F. and Bastami, A., 1986, Improving stability of slender end mills against chatter, Trans. of ASME, Journal of Engineering for Industry, v 108, pp. 264. 10. Bao, W. Y. and Tansel, I. N., 2000, Modeling micro-end-milling operations, Part I: analytical cutting force model, International Journal of Machine Tools & Manufacture, v 40, pp. 2155-2173. 11. Bao, W. Y. and Tansel, I. N., 2000, Modeling micro-end-milling operations, Part II: tool run-out, International Journal of Machine Tools & Manufacture, v 40, pp. 2175-2192. 66

PAGE 75

67 12. Bao, W. Y. and Tansel, I. N., 2000, Modeling micro-end-milling operations, Part III: influence of tool wear, International Journal of Machine Tools & Manufacture, v 40, pp. 2193-2211. 13. Gu, F., Kapoor, S. G., and DeVor, R. E., 1991, An approach to on-line cutter run-out estimation in face milling, Transactions of the North American Manufacturing Research Institution of SME, pp. 240-247. 14. Sutherland, J.W., and DeVor, R.E., 1986, An improved method for cutting force and surface error prediction in flexible end milling systems, Journal of Engineering for Industry, v 108, pp. 269. 15. Cook, N.H., Subramanian, K., and Basile, S.A., 1975, Survey of the state of the art of tool wear sensing techniques, Materials Processing Laboratory, Department of ME, MIT, Cambridge. 16. Altintas, Y., Yellowley, I., and Tlusty, J., 1988, The detection of tool breakage in milling operations, Trans. of ASME, v 110, pp. 271. 17. Takata, S., Ogawa, M., Bertok, P., Ootsuka, J., Matushima, K., and Sata, T., 1985, Real-time monitoring system of tool breakage using Kalman filtering, Robotics and Computer-Integrated Manufacturing, v 2, n 1, pp. 33. 18. Liang, S., and Dornfeld, D.A., 1989, Tool wear detection using time series analysis of acoustic emission, Trans. of ASME, Journal. of Engineering for Industry, v 111, pp. 199. 19. Tansel, I. N., and McLaughlin, C., 1993, Detection of tool breakage in milling operations: Part 2 The neural network approach, International Journal of Machine Tools and Manufacturing, v 33, n 4, pp. 545. 20. Vogler, M. P., Liu, X., Kapoor, S. G. and DeVor, R. E., 2002, Development of meso-scale machine tool (mMT) systems, Transaction of the North American Manufacturing Research Institution of SME (NAMRI), 30, pp. 653-662. 21. IDCT Micro-drilling Systems, 1999, 250,000 RPM Zindle Drilling Mechanism, IDCT Torrance, CA. 22. Kistler, 2002, Operating Instructions, 3 component force sensor, Kistler Instrument Corporation, New York. 23. Precise, 2002, Operating and maintenance instructions for spindle series: SC 40, SC 60, SC 77, Precise Corporation, Racine, WI. 24. Precise, 2002, Operating instructions for solid state frequency converter, type PCF 310, Precise Corporation, Racine, WI.

PAGE 76

68 25. Precise, 2002, Operating and maintenance instructions for coolant circulation systems, types 7136, 7137, and 7138, Precise Corporation, Racine, WI. 26. National Instruments, 2003, LabView data acquisition basics manual, National Instruments, Austin, Texas. 27. National Instruments, 2003, DAQ hardware overview guide, National Instruments, Austin, Texas. 28. Budynas, R. G., 1999, Advanced Strength and Applied Stress Analysis, WCB McGrawHill, New York.

PAGE 77

BIOGRAPHICAL SKETCH The author was born in Sultanpur, UP, India, on December 16th, 1977. He grew up primarily in Calcutta, and in school he used to enjoy thinking, problem/puzzle solving, working with different ideas and mathematics as his most favorite subjects. This developed a fancy for a career in engineering and he joined mechanical engineering stream at Kamla Nehru Institute of Technology, Sultanpur, India. During his undergraduate study he focused more in the area of designing systems. He earned his Bachelor of Science degree in July, 2001 and came to pursue masters degree in August, 2001. The author plans to go to industry after graduation preferably in the area of precision system design. 69


Permanent Link: http://ufdc.ufl.edu/UFE0002847/00001

Material Information

Title: Design, Assembly, and Testing of an Ultra-High-Speed Micro-Milling Spindle
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0002847:00001

Permanent Link: http://ufdc.ufl.edu/UFE0002847/00001

Material Information

Title: Design, Assembly, and Testing of an Ultra-High-Speed Micro-Milling Spindle
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0002847:00001


This item has the following downloads:


Full Text












DESIGN, ASSEMBLY, AND TESTING OF AN ULTRA-HIGH-SPEED
MICRO-MILLING SPINDLE















By

JAY PRAKASH PATHAK


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2003















ACKNOWLEDGMENTS

The author would like to acknowledge with thanks the contribution of the

following people who helped him during this project.

Dr. John Ziegert provided distinguished and expert tutelage as the chair of

supervisory committee throughout the development of the spindle system right from

discussing idea during the design phase to the end with actual assembly and testing.

Dr. John Schueller provided support during assembly of the spindle system.

Dr. Ashok Kumar provided help during the model development and finite element

analysis.

Dr. Sanjay Ranka was kind enough to be part of my committee.

Dr. Tony Schmitz provided relevant feedback during assembly and force

measurements.

Mr. Paul Frederickson of Precise Corporation was kind enough to do the finite

element analysis for analyzing the natural frequencies of the spindle system as well as

provided help in the design of the mount for the Precise motor.

Mr. Bernhard Jokiel and David Gill of Sandia National Labs instigated this project

and provided funding for the development of the spindle system.

Mr. Andrew Dewitt and Mr. Tim Claffey of New Way Machine Components, Inc.,

extended their cooperation to make custom designed air bearing relevant to our needs.

Mr. Ron Brown, machinist at University of Florida, helped in making some final

changes during assembly of the machine.









The author would like to thank his friends, lab-mates and others who one way or

the other helped him throughout this project.

Lastly, the author would like to thank his parents for moral support and motivation.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S .................................................................................................. ii

L IST O F FIG U R E S .... ...................................................... .. ....... ............... vi

A B STR A C T ..................... ................................... ........... ................. viii

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

1.1 M icro-M killing .............................................................................. 1
1.2 D difficulties in Scaling M killing Processes .................... .... ..............................2
1.3 Feasibility and Need For Very High Speed Micro-Milling Spindles.................4
1.4 Ultra-High Speed M icro-M killing Spindle System ...............................................5
1.5 Organization of Remainder of Thesis.......................... .............. ... .......... 6

2 BACKGROUND AND LITERATURE REVIEW ...............................................7

2.1 Development of Micro-Milling Process........... .......... ..........................7
2.2 M icro-M killing Tool Fabrication ........................................ ........................ 7
2.3 M odeling M icro-End-M killing Operations ........................................ ...................8
2.3.1 Analytical Cutting Force Model ..... ............. ...................8
2.3.2 Influence of Tool Run-Out ...............................................9
2.3.3 Influence of Tool-W ear ......................................................... ................ 10
2.3.4 Cutting Force Model for Micro-Milling of Heterogeneous Materials .......11
2.4 Current High-Speed Spindle Technology.................. .. ................ ....................11

3 MOTIVATION AND CONCEPTUAL DESIGN............... ........ ............... 13

3 .1 M o tiv atio n ....................................................................................................... 1 3
3.1.1 L ow C hip T hickness........................................................ ............... 13
3.1.2 H igh T ool R un-out ............................................................. .. ................ 14
3.1.3 Low M etal Rem oval Rate.................................... ......................... 14
3.1.4 High Cutting Force Coefficients ..................................... ..................... 14
3.1.5 L ow C cutting Speed .......................................... ................. ............... 15
3.1.6 U unpredictable Tool L ife ........................................ ......... ............... 15
3.2 C conceptual D design .... ...................................................... .. ... ... .... .. .. .... 15
3.2 .1 C onceptual D esign .......... ........................................ ............ ... ......... 15
3.2.2 D esign D description .............. .... .. ........... ............... ...............17











4 DETAILED SPINDLE DESIGN AND ASSEMBLY ............................................19

4 .1 B a se F ra m e ..................................................................................................... 2 2
4 .2 A ir B e a rin g ..................................................................................................... 2 2
4.3 Force Sensing System .................................................. ................... ............... 25
4.4 Assembly of Base Frame, Air-Bearing and Force Sensor..............................26
4.5 M otor M ount .. ............. ................... ................................ 27
4.6 Assembly of Motor Mount with Base Frame ....................................................28
4.6 Assembly of Precise Spindle with Base Frame and Friction Wheel ..................30
4.7 Assembly of Micro-Spindle with HSM-2........................................ ...............30
4.8 Micro-Spindle System Connections ..... .................... ...............31
4.9 Com puter, Electronics and Softw are ........................................ .....................33

5 D E SIG N V A L ID A T IO N ......................................... .............................................37

5.1 FEA of Motor and Friction Wheel Assembly...............................................37
5.2 FEA of the Micro-Tool, Air-Bearing and the Friction Drive .............................39
5.3 Stress/Strain Analysis ................................. .... ........ ............ 41

6 TESTIN G AND RESULTS.............................................. .............................. 43

6.1 R un-out M easurem ents .................................................... ................ ............... 43
6.1.1 Run-out M easurement at M otor Arbor............................ .....................43
6.1.2 Run-out Measurements at Friction Wheel and Gage Pin...........................45
6 .2 F force M easu rem ents .............................................................................. ...............4 6
6.3 M achining of Hexagonal Test Part.................................................................... 50
6.4 Air-Bearing Stiffness M easurem ents......................................... ............... 51

7 CONCLUSIONS AND FUTURE WORK.....................................................53

APPENDIX

A P A R T L IST .............................................. ........... ............... 55

B D E TA IL D R A W IN G S ....................................................................... ..................59

LIST OF REFEREN CES ............................................................................. 66

B IO G R A PH IC A L SK E TCH ..................................................................... ..................69
















LIST OF FIGURES

Figure page

1-1 Solid model of ultra-high spindle micro-milling spindle system.............................5

3-1 3D model of ultra-high spindle micro-milling spindle system............................16

4-1 The assembled micro-milling spindle ........................................ ............... 20

4-2 The assembled micro-milling spindle system mounted to HSM-2.....................21

4-3 3D m odel of the base fram e. ...................................... ............. ..........................22

4-3 3D model of the air bearing. ............... ........... ............... 23

4-4 Isometric view of the air bearing. .............. ............ ........... ...................23

4-5 Air cleaning system and valve. ........................................................................24

4-6 Three axes force sensor ........... ................. ............................ ............... 25

4-7 Assembly of base frame, air-bearing and force sensor.......................................26

4-8 CAD model of the motor mount. ............. ............................................... .. 27

4-9 M otor m ount. ................................. ... .. ............ .............. .. 28

4-10 Spherical washer set ................. ...... .............. ....... ..... .......... 28

4-11 Fine adjustm ent set screw ...... ........................... ......................................29

4-12 Assembly of motor mount and base frame. ....................................29

4-13 Friction w heel. ................................................ .. ........... .... .. .....30

4 -14 In te rfa c e p late ................................................................................................... 3 1

4-15 Schematic of system connections. .............................................. ............... 32

4-16 Bread board circuit ............. ........ .......... ................. 34

4-17 Enclosure containing i/o connector and bread board.........................................34









4-18 Labview interface for spindle control and measurement.......................................35

4-19 Labview interface for cutting force measurement. ................................................36

5-1 Motor, motor arbor and friction wheel assembly for FEA. ..................................38

5-2 Results of FEA of motor and friction wheel assembly ..........................................38

5-3 FEA model of the tool, air-bearing and friction wheel.........................................39

5-4 F irst m ode: 4 857 .3 H z ......................................... .............................................4 0

5-5 Second m ode: 1.35e5 H z. ........................................................... .....................4 1

5-6 Third m ode: 2.65e5 H z. ............................................... .............................. 41

5-7 Fourth m ode: 3.89e5 H z. .............................................. ............................. 41

6-1 E xperim mental set-up. ..................................................................... ...................44

6-2 Run-out of drive spindle arbor ........................................ .......................... 44

6-3 Run-out of friction wheel and gage pin. ..................................... ............... 45

6-4 Force plot for micro-milling of Al using 127[im micro-tool............................. 47

6-5 Mathematical model of the tool, air-bearing and force sensor system ................48

6-6 Slots made from 508pm micro-tool ....................................................................49

6-7 Slots made from 254[m micro-tool ....................................................................49

6-8 Hexagonal feature with a wall thickness of 30[im...............................................50

6-9 Force vs. displacement plot to measure the stiffness......................................... 52















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

DESIGN, ASSEMBLY, AND TESTING OF AN ULTRA-HIGH-SPEED
MICRO-MILLING SPINDLE

By

Jay Prakash Pathak

December 2003

Chair: Dr. John C. Ziegert
Major Department: Mechanical and Aerospace Engineering

Micro-milling is an emerging fabrication technology that broadens applicable

material ranges including metals and plastics. It is potentially the technology of choice to

create complex three-dimensional micro- and meso-scale components in hard engineering

materials. The micro-milling process is characterized by the use of milling tools 100 jtm

or less in diameter. Presently, these tools are used to create miniature features in plastics

and some soft metals, typically with a very low material removal rate. Spindles for micro-

milling often show excessive run-out relative to the chip thickness required by these

small tools. This thesis documents the design, assembly and testing of an ultra-high speed

micro-milling spindle; aimed at improving the efficiency of micro-milling operations.














CHAPTER 1
INTRODUCTION

This thesis documents the design, assembly and testing of an ultra- high-speed

micro-milling spindle used for creating micro and meso-scale features in soft as well as

hard engineering materials. This chapter briefly describes the micro-milling process and

describes difficulties encountered in micro-milling technology which motivates the need

for a new spindle.

1.1 Micro-Milling

There is a continuous demand for miniaturization of components for consumer and

other products. Micro-milling is a potential technology for the production of miniaturized

parts and components. There exists a wide variety of important applications for micro-

and meso-scale mechanical systems in the commercial and defense sectors, which require

high-strength materials and complex geometries that cannot be produced using current

MEMS fabrication technologies. Micro-milling has the potential to fill this void in

MEMS technology by adding the capability of free form machining of complex 3D

shapes from a wide variety and combination of traditional, well-understood engineering

alloys, glasses and ceramics.

Micro-milling is an enabling technology that may allow rapid and economic

fabrication of meso-scale components with micro-scale features. These types of parts are

a vital link between surface micro machined MEMS systems and the outside macroscopic

world, and potentially have great implications for how MEMS components interact with

each other as well as how they are packaged.









Micro-milling can be defined as a traditional milling process that has been scaled

down to micron level. However due to the small diameter tools used in the process,

micro-milling is significantly different than conventional milling in a variety of factors,

viz. cutting forces, tool run-out and metal removal rate etc.

1.2 Difficulties in Scaling Milling Processes

Inefficiencies in milling small features result directly from the relationships

between a cutting tool's breaking strength, the applied cutting force, and the metal

removal rate (MRR).

To the first order, end mills act like long, thin, end-loaded, cylindrical cantilever

beams; where the maximum bending stress (cx) is directly proportional to the tool length

(L) and the cutting force (Fe), while inversely proportional to the cube of the diameter (D)

(Equation 1.1).

32F= (1.1)


In slot milling the maximum tangential cutting force (Equation 1.2) is proportional

to the workpiece material cutting stiffness (Ks), axial depth of cut (do), and feedrate (fr),

while it is inversely proportional to the spindle speed (n) and number of teeth on the

cutter (N).



F (N) = K, (N / mm2) d ( mm) ( mm / (1.2)
n(rev min)N(teeth/ rev)
1 ChipLoad j

The MRR (Equation 1.3) is proportional to the feed rate, the depth of cut and the radial

depth of cut of the tool (re).

MRR(mmmin) = f (mm/ (mm)rn (mm) (1.3)









In light of Equations 1.1-1.3, we consider a macro-scale part and associated tooling

that one wishes to homogenously scale in size by a scaling factor, k. If the spindle speed

is constant for both operations, the axial and radial depths of cut will scale by k, and thus

the feed rate must be scaled by a factor of k to maintain a constant bending stress in the

tool. Since the MRR is proportional to the feed rate, the depth of cut and the radial depth

of cut of the tool (re) the MRR scales by k3. However, cutting tool edge radius effects and

spindle run-out become more important at micro-milling scales. Commercially available

micro-milling tools typically have a cutting edge radius on the order of 2 to 3[tm. In

micro-milling, to avoid tool breakage the chip thickness is typically less than 1l m, so the

tool edge effectively has a large negative rake angle. This poor cutting geometry

increases the effective cutting stiffness, K,, resulting in even higher forces, and thus

requiring an even smaller chip thickness, further slowing the feed rate and increasing the

processing time.

Typical milling spindles used for these small tools employ either rolling element

bearings or air bearings to support the spindle shaft. Tools are typically clamped in such

spindles using a collet or set screw. The combination of asynchronous spindle bearing

error motions and clamping errors often result in tool run-out 3 to 20 times the nominal

chip thickness. This will drastically increase the cutting force and may lead to tool

breakage unless the axial depth of cut is reduced, thus further reducing the MRR and

increasing the processing time. As a result, the tool spends most of its time under-loaded

to protect against the few times when the spindle and clamping errors cause excessively

thick chips to be cut.









1.3 Feasibility and Need For Very High Speed Micro-Milling Spindles

In conventional machining processes there is a preferred range of cutting speeds for

a given tool and workpiece material. Therefore, to maintain the preferred cutting speed,

the spindle speed should increase as the tool diameter decreases (Equation 1.4).


V,, (mm / min) = 2- n(rev / min)rtool (mm) (1.4)
60

Since cutting forces in milling do not increase significantly with the cutting speed,

one approach to improving the efficiency of micro-milling operations is develop higher

speed spindles with low radial error motions to decrease machining time.

Micro-milling operations should not be more time consuming than conventional

milling for a given part of similar geometry, if sufficiently high spindle speeds can be

achieved and the problem of high tool run-out at such speeds can be solved. Consider a

macro-scale steel part machined using a 10mm diameter two-flute end mill, rotating at

5000RPM with a chip load of 0.18mm/tooth (yields a feed rate of 1800mm/min). Scaling

the part with k =0.01 requires a 0.1mm diameter two-flute end mill, rotating at 500,000

rpm with a chip load of 0.0018 mm/tooth, which yields a feed rate of 1800mm/min. If

these parameters can be achieved, the machining time should be the same. The required

spindle speed for the micro-milling case is far beyond the range of currently available

spindles, and points to the need to develop a very high-speed spindle for micro-milling

operations. Additionally, any such spindle should have sub-micron run out to avoid

overloading the tool, and should include a sensing system of some sort to aid in tool

setting, and detecting tool breakage.









1.4 Ultra-High Speed Micro-Milling Spindle System

The proposed spindle design uses the tool shank itself as the spindle shaft,

supported in an air bearing. A friction drive transmits torque from a drive motor to the

tool shaft as shown in the Figure 1-1.











Motor



Motor Mount


Spherical
Washer
Air Bearing

Motor Arbor

ase Structure
Friction
wheel coated
with ML6
Micro Tool 3 Axis Force Sensor

Figure 1-1. Solid model of ultra-high spindle micro-milling spindle system.

The drive ratio of the friction drive is 1:8, making it is possible to achieve more

than 500K RPM when combined with commercially available high speed motors capable

of speeds in the 10K-90K RPM range. Commercially available micro-milling cutters

typically have a 0.125-inch diameter shank, approximately 1.5 inches long. The actual

milling cutter is ground on the end of the shank. Instead of clamping the tool shank into









the rotating shaft of the spindle, the proposed spindle will support the tool shank radially

in a porous carbon air bearing. The tool shank effectively becomes the "spindle shaft"

and should greatly eliminate tool run-out.

1.5 Organization of Remainder of Thesis

Literature reviews and background information are provided in Chapter 2. Chapter

3 describes the motivation and conceptual design of the micro-spindle. Chapter 4

describes the detailed design of the micro-spindle components and parts. Chapter 5

describes design validation analyses using finite element analysis to understand the

dynamics of the micro-spindle and calculations for a reliability check of the design.

Chapter 6 documents the results of the initial testing of the prototype spindle. Chapter 7

provides conclusions and suggestions for future work.














CHAPTER 2
BACKGROUND AND LITERATURE REVIEW

This chapter gives a summary of the available literature relevant to the micro-

milling process.

2.1 Development of Micro-Milling Process

Conventional milling is known to be a very versatile process capable of creating

three-dimensional features and structure. Adaptation of this process to create micro-scale

features is termed micro-milling. Friedrich and Vasile [1] describe the development of

the micro-milling process for fabricating molds and mask structures in PMMA. The

process uses focused ion beam [2] micro machined tools [3] and utilizes a high-precision

milling machine designated for the milling process. Friedrich, Coane, Goettert and

Gopinathin [4] have shown that micromechanical milling is also a rapid and direct

method for fabricating masks for deep x-ray lithography with lateral absorber features

down to 10tm. Adams et al. [2] demonstrate micro-milling as a first step in fabricating

metal alloy micro components. The metal alloys that were tested are 606 1-T4 aluminum,

brass and 4340 steel. For all tests 15-25 [tm deep trenches were cut several millimeters in

length as commanded.

2.2 Micro-Milling Tool Fabrication

Vasile et al. [3] describe a process for making very small milling cutters using

focused ion beam machining to create tools with a desired number of cutting edges and

tool end clearance. Micro-end mills having -25[tm diameter are made by sputtering









cobalt M42 high-speed steel and C2 micro-grain tungsten carbide tool blanks. Cutting

edge radii of 0.1 [im can be achieved.

Inexpensive micro-end mills are also commercially available. They typically have a

1/8 inch diameter shank, 1.5 inches in length. The cutting portion is ground on the end of

the shank, and is available in sizes ranging from 0.005 to 0.030 inch diameter, and in 2

and 4 flute geometries.

2.3 Modeling Micro-End-Milling Operations

As discussed in Chapter 1 micro-milling behaves very differently than conventional

milling in a number of ways. The ratio of feed per tooth to tool edge radius is much

smaller for micro-milling than for conventional milling. Stress variations on the shaft of a

micro-tool are much higher than on conventional tools. These extreme operating

conditions can drastically shorten tool life. Less than one hundred inches of tool life is

common when hard metals such as stainless steel are machined. If the cutting conditions

are not selected properly, micro-tools will be broken in a few seconds. Because of their

small size, it is very difficult to notice the damaged cutting edges and even the broken

shaft. Many hours of machining time may be wasted if the tool failure is not detected in

time.

2.3.1 Analytical Cutting Force Model

Cutting force prediction is very important to achieve optimal cutting conditions in

machining operations. Several cutting force models have been proposed for accurately

measuring the cutting forces for better control in milling operations. In 1975 Tlusty and

MacNeil [5] gave the first analytical expressions for cutting forces for conventional

milling. It was improved by Gygax [6] in 1979 and by Kline and DeVor [7] in 1983.

Kline and DeVor [7] considered the effect of cutter run-out on cutting geometry and









hence affecting the cutting forces. Again a database of cutting force coefficients for

different tool and material combination was created considering different aspects of

milling like up and down-milling, symmetric and asymmetric cuts. Later Tlusty and Ismail

[8] in 1983 and Ismail and Bastami [9] in 1986 concentrated on the dynamics of the cutting

operation and development of chatter. In these studies, cutting forces were calculated

numerically to be able to consider the influence of the present and previous tool vibrations to

the uncut chip area.

Tlusty's cutting force model which considers the tool tip path, as circular arcs that

are mutually shifted by ft (feed per tooth) is no longer valid for micro-milling as the ratio

of feed per tooth and tool radius is not very small in micro-milling. Bao and Tansel [10-

12] proposed new analytical models in order to perform micro-machining operations at

the optimal cutting conditions. Their method calculates the chip thickness by considering

the trajectory of the tool tip while the tool rotates and moves ahead continuously. This

model considers the trajectory of the tool tip as cycloids not circles as proposed in

Tlusty's model. This gives a different expression for chip thickness and takes into

account the difference between the up and down-milling. As opposed to Tlusty's model,

where chip thickness is zero when cutter angle is zero or 1800, in a slotting cut, here the

chip thickness is never zero for any real value of cutter angle.

2.3.2 Influence of Tool Run-Out

The effect of tool run-out creates negligible changes in cutting force profile of

conventional end milling operations whereas in case of micro-milling it leads to drastic

force variations. It is very common to see that only one cutting edge of a two-flute micro-

end-mill performs the machining operations alone while the other edge doesn't touch the

workpiece at all. When one of the cutting edges starts to perform all or most of the cutting









operations, the force variation increases significantly. The tool wears out much more quickly,

and the probability of tool breakage increases. To select the optimal cutting conditions in

micro-milling, it is very helpful to be able to estimate the cutting force characteristics in a

wide cutting parameter range if the tool run-out is known. To consider the tool run-out, the

chip thickness expression derived by Tlusty and MacNeil [5] in 1975 was modified by

Gu, Kapoor and DeVor [13] in 1991. Again a computer-based numerical model was

introduced by Sutherland and DeVor [14]. Bao and Tansel [11] gave the first compact

expression to calculate the cutting forces for micro-milling with tool run-out. The cutting

force expressions were simplified by eliminating the insignificant components of the chip

thickness expression to obtain the results in a more compact form. This can be also used

in with optimization algorithms to estimate the tool run-out, to select the optimal cutting

conditions, to monitor the operating conditions, and to estimate the surface quality from

experimental cutting force data. The validity of the proposed model is evaluated by

comparing the simulated and experimental cutting force profiles. Again the expressions

are reducible to the cutting force expressions without tool run-out if the run-out term is

set to zero.

2.3.3 Influence of Tool-Wear

In addition to tool run-out as discussed in the previous section tool wear also plays

a significant role in accurate prediction of cutting forces in micro-milling. The increase of

cutting forces with tool wear was observed in turning operations in 1975 by Cook,

Subramanian and Basile [15]. However estimation of tool wear from cutting force is not

easy since the cutting forces continuously change even in typical turning operations when

the cutting conditions change. The number of parameters increases in end-milling

operations and estimation of the tool condition becomes much more complicated. To









detect tool breakage and estimate wear, characteristics of various signals were studied by

Altintas, Yellowley and Tlusty [16] in 1988 and the effectiveness of empirical models

including the time series analysis and neural networks was demonstrated by Takata,

Ogawa, Bertok, Ootsuka, Matushima, and Sata [17] in 1985, Liang and Dornfeld [18] in

1989 and Tansel and McLaughlin [19] in 1993. Bao and Tansel [12] in 2000 modified

their own cutting force models [10-11] to represent tool wear at any stage of usage.

Modification of the analytical models [10-11] to represent the tool wear is

straightforward; however, knowledge of different tool conditions requires estimation of

the parameters of these non-linear models. Genetic algorithms were proposed for

estimation of the parameters of the model.

2.3.4 Cutting Force Model for Micro-Milling of Heterogeneous Materials

Bao and Tansel's [10-12] model of cutting forces assumes the workpiece material

to be homogeneous. But actually as the depth of cut and feed rates are reduced, the chip

load encountered in the process becomes the same order of magnitude as the grain size of

many alloys such as steel. For such materials the workpiece material must be modeled as

heterogeneous. Vogler et al. [20], in 2002 developed a mechanistic model for cutting

force prediction that explicitly accounts for the different phases while machining steel by

using different cutting force coefficients for the different material phases. The model is

validated by using calibration tests for ferrite and pearlite, the major components of

ductile iron and many other steels, in order to determine the machining force-chip load

relationship for individual phases of heterogeneous material.

2.4 Current High-Speed Spindle Technology

Some literatures were reviewed on current miniature high speed spindle

technology. The first one to be investigated in this respect was National Jet made high-









precision micro-milling machine [1]. This machine is composed of 1500 kg of granite for

vibration and thermal stability and air bearing axes for added stability of travel.

Interferometric position control gives a resolution of 1.25 nm for x and y-axes. The

micro-milling head uses a vee-block bearing arrangement that has a total of four

spherically convex diamond surfaces arranged as two sets of two each. Tool rotation is

highly concentric about the axis in space with no measurable radial run-out of the tool

itself.

The second spindle that was investigated is Zindle's [21] 250,000 RPM drilling

mechanism. High RPM and zero run-out of this spindle makes it ideal for drilling holes in

the 0.018" to 0.003" diameter range. The primary mechanics of this spindle system

consists of three high precision, hardened tool steel wheels arranged in triangle fashion.

Two of these wheels are mechanically driven, while the third one is an idler. The drill

rides against the faces of the two drives wheels and is held against them by the idler

wheel. The high RPM is accomplished by way of a simple mechanical gear-up, using the

faces of the drive wheels and pulleys in the system. The large diameter of the drive

wheels relative to the small diameter of the drill shank makes it possible to rotate the drill

at 250,000 RPM while the wheel rotates at 45,000.














CHAPTER 3
MOTIVATION AND CONCEPTUAL DESIGN

This chapter describes the design goals of the micro-milling spindle and provides

details of the proposed design.

3.1 Motivation

The literature review in Chapter 2 summarized some of the challenges in micro-milling.

These include:

Low chip thickness.

High tool run-out.

Low metal removal rate.

High cutting force coefficients.

Low cutting speed.

Unpredictable tool life.

3.1.1 Low Chip Thickness

The micro-milling process is characterized by milling tools that are currently in the

range of 22-100m. Because of the small diameter of the tools, the cutting forces must be

kept very small so as not to exceed the bending stress fatigue limit of the tool at the root.

In order to keep the forces sufficiently small, the chip thickness must be very small.

Typical chip load values reported for machining of metallic workpieces are on the order

of 0.5 to 1.0 [m.









3.1.2 High Tool Run-out

Typical milling spindles used for micro tools employ either rolling element

bearings or air bearings to support the spindle shaft. Tools are typically clamped in such

spindles using a collet or set screw. The combination of asynchronous spindle bearing

error motions and clamping errors often result in tool run-out 3 to 20 times the nominal

chip thickness. This means that in ordinary operation, some teeth on the cutter may not

contact the workpiece at all during rotation, while others are forced to cut chips up to

several times the desired thickness. This leads to overloads of the tool and premature

failure.

3.1.3 Low Metal Removal Rate

The low chip thickness characteristic of micro-milling leads to low feed rates. For

example, with a 1 itm chip thickness and a 2 fluted cutter, the feed rate is only 2 itm per

revolution. Therefore, if the spindle speed is 20,000 rpm (a typical maximum spindle

speed for a high speed milling spindle), the feed rate of the tool through the work is only

40 mm/min. This results in very low material removal rates. Also tool run-out becomes

more significant at higher speeds, which drastically increases the cutting force and may

lead to tool breakage unless the axial depth of cut is reduced, thus further reducing the

MRR and increasing the processing time.

3.1.4 High Cutting Force Coefficients

Commercially available micro-end mills typically have cutting edge radii on the

order of 2 to 3 tm, or 2 to 6 times the chip thickness. This means that the effective rake

angle of the cutting edge is highly negative, on the order of-45 to -60 degrees. In this

situation, the conventional models of the mechanics of chip formation do not apply, and









the cutting force coefficients are typically 20 to 40 times higher than in conventional

milling.

3.1.5 Low Cutting Speed

In order to achieve efficient cutting there exists preferred ranges of tangential

speeds of the cutting edge for different work piece and tool material combinations.

Spindle speed must increase as tool diameter decreases in order to achieve the desired

cutting speed. For instance, when using carbide tools to machine aluminum, the

recommended cutting speed is on the order of 500 meters/minute. To achieve this cutting

speed with a tool diameter of 0.25 mm, the required spindle speed is over 600,000 rpm.

This speed is unachievable with current machine tool spindle technology.

3.1.6 Unpredictable Tool Life

Unpredictable tool life and premature tool failure are the major concerns in micro

machining using micro grain carbide cutters. The reasons for these premature failures are

believed to be related to the low chip thickness relative to the tool edge radius, the large

tool run-out relative to the chip thickness, and the low cutting speeds achievable with

current micro-milling spindles.

3.2 Conceptual Design

Based on the challenges outlined above, it appears that the key bottleneck to

efficient use of micro-milling for metallic materials appears to be related to the spindle.

Spindles with low run-out and speeds on the order of 500,000 rpm are desired for this

application.

3.2.1 Conceptual Design

The conceptual design for the ultra-high speed micro-milling spindle is shown in

Figure 3-1.


























Spherical
Washer

Motor Arbor


Friction
wheel coated
with ML6


Motor




Motor Mount



Air Bearing



"--B ase Structure


Micro Tool


3 Axis Force Sensor


Figure 3-1. 3D model of ultra-high spindle micro-milling spindle system.









The spindle system uses an air bearing to support the 1/8th inch tool shank. The tool

shank itself is driven by a friction drive with the drive wheel driven by a commercial high

speed spindle. The drive ratio of the system is 1:8 and so it is possible to achieve speeds

over 500k rpm with the commercially available high speed spindles rotating as speeds of

60-90 k rpm. Commercially available micro-milling cutters typically have a 0.125-inch

diameter shank, approximately 1.5 inches long. The tool shank effectively becomes the

"spindle shaft" and should greatly reduce tool point radial run-out.

3.2.2 Design Description

The major components of the proposed spindle system are:

A custom porous carbon air-bearing to support 1/8" diameter shank micro

tools.

A 3-axis force sensor capable of sensing in-situ milling forces.

A high speed drive motor capable of producing a torque of 0.01 N-m at 50K

rpm.

A split mount for mounting the high speed drive motor.

A friction wheel coated with suitable polymer, interfaced to the high speed

drive motor.

A pair of spherical washers and fine adjustment set screws for adjusting the

motor axis relative to the tool shank axis in two perpendicular directions.

A frame for mounting the components of the system.

A computer interface to control and display spindle speed, monitor tool

forces during milling, and provide means to interpret and feedback milling






18


forces to identify tool breakage and identify and control tool over/under-

loading.














CHAPTER 4
DETAILED SPINDLE DESIGN AND ASSEMBLY

Detailed design of the major components and the assembly of the parts are

discussed in this chapter. The order of these discussions follows the logical sequence

required to assemble and align the machine. Figure 4-1 shows the complete assembled

machine.

Following good design and manufacturing practices from industry, part numbers

are assigned to all parts purchased, manufactured, or borrowed, for the purposes of record

keeping. A list of all parts and assemblies appears in Appendix A. Only assemblies of

particular relevance are described in detail, since this thesis is not meant to be an

exhaustive assembly manual.

The following scheme is used to designate part numbers (P/N):

Manufactured parts: These parts were designed from scratch and required

detail engineering drawings so as to be fabricated by outside vendors. The

nomenclature that was used to designate these parts has the format

"MSDXX". Here MSD stands for "Micro-spindle Design" and two XXs

stands for two digited numbers between 00-99. For example a valid part

number is MSD01 or MSD99 etc.

Purchased parts: These parts are purchased directly from vendors. A

specification sheet from the manufacturer or vendor replaces a detailed

drawing. This apart all major purchased parts are supported by an outline

engineering drawing. The naming convention for the purchased parts follows









the format "MSDPXX". Here MSD stands for "Micro-spindle Design", P

stands for Purchase and two XXs stand for two digited numbers between 00-

99. For example a valid part number is MSDP01 or MSDP99 etc. Components

already belonging to the university were treated as purchased parts when

assigning part numbers.


Figure 4-1. The assembled micro-milling spindle.








Fasteners: Although "purchased", in order to distinguish them from other
purchased parts, these parts have naming convention of"MSDFXX". Here the
first three letters have the same usual meaning whereas F stands for
"Fasteners". The other two XXs are as before numbers.
Assemblies: Sub-assemblies and the consequent final assembly have a
nomenclature of"MSDAXX" where A stands for assembly and other letters
are defined as before.
All the parts may be found in the appendix according to their logical order. Part
numbers are given next to part names throughout this thesis.


"-. .I


Figure 4-2. The assembled micro-milling spindle system mounted to HSM-2.









4.1 Base Frame

The L shaped base frame (P/N MSD02) supports the whole spindle system. The

following figure shows the CAD model of the base frame. This has a recessed hole on top

to accept the female half of the spherical washer set (P/N MSDP10). The recess is

designed to provide a light press fit of the spherical washer into the base frame. This

design makes sure the axis of the motor (P/N MSDP01) is located precisely with respect

to the base frame.















Figure 4-3. 3D model of the base frame.

4.2 Air Bearing

The air bearing (P/N MSD03) is a special part which is custom designed specially

for this spindle system and fabricated by New Way Bearings Inc. Figure 4.3 shows the

solid model of the air-bearing. The external dimensions were supplied to the vendor, who

then designed all of the internal details. Figure 4.4 shows the actual drawing of the air

bearing.































Figure 4-3. 3D model of the air bearing.

The bearing housing is made of aluminum. The air bearing system has a porous

carbon insert with a central hole for 1/8" diameter tool shank and also a thrust bearing on

top for the head of the tool to seat in. The radial and axial air bearings are intended to

ensure that there is effectively no friction between the tool and the bearing surface.


Figure 4-4. Isometric view of the air bearing.









For the air bearings to work properly the manufacturer has recommended a shaft

tolerance of +/-0.00015 inches. However, the commercially available micro-tools have a

shaft tolerance of +/-0.0005 inches. Therefore, a collection of micro-tools was purchased

and the shaft diameters measured. The air bearing bore was sized to fit these tools. The

micro-tools float freely inside the air-bearing when it is supplied with air, indicating that

the bearing is sized appropriately.

Lastly, the bearing must be supplied with filtered and moisture free air pressurized

to 100 psi. It must be free from oil vapor, water and dust particles. Figure 4.5 shows the

air filtration system. From top to bottom the components are two pressure regulators, a

compressed air-filter, a desiccant air dryer, a coalescing filter, an oil-vapor removal filter

and finally a valve with two 1/16th inch hose fittings. The direction of air flow is shown

by the arrow in the figure.


Figure 4-5. Air cleaning system and valve.






4.3 Force Sensing System
In order to measure the dynamic or quasistatic cutting forces during machining a
Kistler 3-axis miniature force sensor (P/N MSDP06) is sandwiched between the air
bearing and the base frame. The piezoelectric force sensor has a resolution of 0.01N and
a measuring range of +/-lkN. Technical data about the sensor is available in
manufacturer's specification sheet.
When a force is exerted on the sensor, the piezo-electric material generates an
electric charge. A PCB charge amplifier (P/N MSDP08) converts charge developed by
the sensor into proportional DC voltage. This voltage is read through three input channels
of a LabView DAQ card. Figure 4.6 shows the picture of the force sensor.

SFx


I Fz




kN
^*-IPQ^Ik


Figure 4-6. Three axes force sensor.









4.4 Assembly of Base Frame, Air-Bearing and Force Sensor

Kistler [22] recommends special mounting instructions for the force sensor which

are incorporated in the design and assembly. The sensor is mounted under preload

because the shear forces Fx and Fy must be transmitted through static friction from the

bearing and the base frame to the surfaces of the force sensors. A 13 kN preload is

recommended in order to use the full measuring range of the forces, however we

preloaded it to 11.2 kN as the expected range of forces that we will be measuring is less

than 20 N. The preloading is carried out per Kisler recommendations with the help of a

torque wrench set to 10 N-m, corresponding to a force of 11.2 kN. The relationship

between torque and axial preload is given by:

T = rF (4.1)

where, T is the torque in N-m, F is the preload force, i is coefficient of friction and r is

the preloading bolt diameter. The preloading bolt diameter was measured and be 5.952

mm. The coefficient of friction suggested by Kistler was 0.15, yielding a preload force of

11.2 kN corresponding to an applied torque of 10 N-m. Figure 4-7 shows the picture of

the assembly.

ri aC


Figure 4-7. Assembly of base frame, air-bearing and force sensor.









4.5 Motor Mount

A single piece, split ring motor mount (P/N MSD01) is used to mount the "Precise"

drive spindle (P/N MSDP01). The mount is designed according to the manufacturer's

recommendations [23]. Figure 4.8 shows the CAD model of the motor mount. Appendix

B shows the detailed engineering drawing of the mount.


Figure 4-8. CAD model of the motor mount.

The mount has 2 clamping screws and one spacer screw. The clamping screws are

used to assemble the motor to the mount whereas the spacer screw helps in dissembling it

from the mount. The torque wrench is used to provide the necessary torque during

clamping according to the manufacturer's data sheet for the motor. Figure 4.9 shows a

photograph of the motor mount.

































Figure 4-9. Motor mount.

4.6 Assembly of Motor Mount with Base Frame

The male central spherical washer is seated on the base frame and then the motor

mount is placed on top it. The four smaller sets of spherical washers (Figure 4-10) are

inserted in the four holes of the motor mount from the top.


Figure 4-10. Spherical washer set.









The four holes of the motor mount are aligned with corresponding holes in the base

frame and they are secured with the help of the 4 fine adjustment (80 threads per inch) set

screws (Figure 4-11).


Figure 4-11. Fine adjustment set screw.

Final assembly is shown in Figure 4-12.


Figure 4-12. Assembly of motor mount and base frame.









4.6 Assembly of Precise Spindle with Base Frame and Friction Wheel

The motor arbor (P/N: MSDP02) is secured to the Precise motor (P/N: MSDP01)

according to the manufacturer's instructions in the data sheet [23]. Then the clamping

screws of the motor mount are loosened and the spacer screw is tightened to spread the

clamp, until the motor can slide inside the inner hole of the motor mount. In this position

the friction wheel (P/N MSDP05) is secured at the motor arbor end. Figure 4-13 shows

the photo of the friction wheel and corresponding CAD model. The detailed engineering

drawing of the friction wheel is available in Appendix B.











Figure 4-13. Friction wheel.

The motor is positioned properly inside the clamp and, the spacer screw is loosened

and the clamping screw is tightened using a torque wrench to 2.8 N-m of torque as

recommended by the manufacturer [23]. This procedure ensures that the bearings inside

the drive motor are not loaded improperly by clamping forces on the motor housing.

Finally, the clamping screw of the friction wheel is secured to the arbor of the drive

motor using a special wrench provided by the manufacturer. The final Assembly is shown

in Figure 4-1.

4.7 Assembly of Micro-Spindle with HSM-2

An interface plate (P/N MSD10) was designed to mount the micro spindle on

HSM-2. Figure 4-14 shows the CAD model of the interface plate. A narrow shoulder on









the plate ensures that the plate is vertically aligned to one of the vertical faces of HSM-2.

The micro-spindle is secured to the mounting plate using the four bottom holes.


Figure 4-14. Interface plate.

4.8 Micro-Spindle System Connections

The spindle drive system consists of a solid state frequency converter (P/N

MSDP03) and associated electrical connections, liquid cooling connections, etc. These

are discussed in detail in the manufacturer's instruction manual [24-25]. The frequency

converter basically controls the motor speed. It can be operated in two different modes,

manual remote control. The spindle drive is connected to the National Instruments DAQ

card channels using the remote control connector, and is interfaced with appropriate

LabView VIs. Figure 4-15 shows the general schematic diagram of the spindle system.













M~ 1or- Ein


Type 7136, 7137, 7113
Cooung Systsm


:'~ 1A~j4~


', 5C 40
5C E0
5sc


* i~b~iBk~ '.;I !~i


r~' C


T)yu PCF 2f07 310
Frequency Con otor


Figure 4-15. Schematic of system connections.


F~i~ *S*rrn. i


I
... .., .......


230 ',AC 1'p!aoa, Hf









4.9 Computer, Electronics and Software

A 300 MHz Pentium laptop computer and National Instruments LabView and DAQ

card [26] was used to control and monitor the spindle system, record cutting forces and

display results.

National Instrument's NI DAQ -1200 card [27] was chosen, which features digital

triggering capability; three 16-bit, 8 MHz counters/timers; two 12 bit analog output

channels; 24 digital I/O lines and fourl2 bit differential analog input channels. Three of

the input channels are used for force measurements in x, y and z direction. One output

channel is used to provide a variable DC voltage from zero to ten volts to command the

Precise Motor to rotate from zero to 90k rpm. The two frequency counters are used for

measuring the actual speed of the Precise Motor. Five of the digital lines of port A are

used for creating virtual LEDs in LabView VI for various warnings as recommended in

the Precise Frequency Converter's instruction manual. A breadboard circuit is used for all

electrical connections. It consists of a 7404 inverter chip, which is a hardware aid for

measuring the frequency of rotation of the Precise motor. There are five registers of 100k

ohms each for electrical tuning of the circuit designed for setting the warning signals

during the operation of the motor. Figure 4-16 shows the picture of the circuit.

The National Instrument's data acquisition card is connected to an I/O connector

block. The breadboard and the I/O connector are mounted inside an enclosure which has

a 15 pin connector for the Precise motor speed control and measurement, three BNC

jacks for force measurements, and one on/off switch. Figure 4-17 shows the components

mounted in the enclosure.





34










!Z: W!^ ^,
Figure 4-16. ...Bread board ci t.






Figure 4-16. Bread board circuit.


Figure 4-17. Enclosure containing i/o connector and bread board.









LabView VIs are developed for spindle speed control and measurement, cutting

force measurements and warning signals during spindle rotation through virtual LEDS.

Figure 4-18 shows the LabView interface.




i I| 1 3pt Applic hon F ont | iji


Speed Control
4.0 6.0

0-.
2.0 -8.0

0.0 10.0


ISTOP0


Commanded RPM
0 00
Commanded/Actual RPM
90000 0-
80000 0-
70000 0-
60000 0-
50000 0-
i 40000.0-
30000 0-
20000 0-
10000.0-
00 -,


LED K1 LED K2 LED K3 LEDK4
SWo S=


LED K5
S'


Actual RPM
Io.o


Figure 4-18. Labview interface for spindle control and measurement.

In the above figure the speed control knob can set a variable voltage between zero

to ten volts and the corresponding commanded speed is visible from the waveform chart.

The actual rpm is also shown in the waveform chart. There is a digital display for the

both the commanded and actual rpm. LED K1 glows red when the actual spindle speed is

zero although the commanded is not. LED K2 glows red when the set load limit is

exceeded for the motor. LED K3 turns red in the event of a load change. Similarly K4 is

green while the unit is operational and K5 becomes green when the actual speed is nearly










equal to commanded speed. Separate sub VIs were developed for these functions. Figure

4-19 shows the picture of the force measurement VI.

-"I x


File Edir Uperate Tools Browve Wirndow Help


ll I'1 *I 1 pt Applic i1on Font L| 1 *


Fx
IG-00 Newton:



0.0X I NetMons


Fz
1000 I Newtons

stop
flaP


FrWhite/Fy: Red
60-
4 ni-
2 0-


20-
40-
.60-
-1


01



00-
; a-



01- ,
Neg


II I

Figure 4-19. Labview interface for cutting force measurement.

Once the machine was fully assembled and aligned, it was ready for testing.

Chapter 6 describes initial testing of the machine at the University of Florida Machine

Tool Research Center.


20 00.01 00
12/31/1903


....


I














CHAPTER 5
DESIGN VALIDATION

This chapter discusses the validity of the design. The design is validated by two

finite element analyses carried out to better understand the system dynamics. Other

calculations are also performed to analyze stresses and strains in rotating parts at high

rpm, in order to avoid material failure. The results are incorporated in the design.

5.1 FEA of Motor and Friction Wheel Assembly

To analyze the natural frequencies of the drive spindle, a finite element analysis of

the motor, motor arbor, and friction wheel assembly was carried out by Mr. Paul

Frederickson of Precise Corporation, who was kind enough to do this analysis. The first

analysis uses a steel friction wheel attached to the end of the motor arbor (P/N MSDP02),

which is again assembled to the motor (P/N MSDP02), as shown in the Figure 5-1. The

analysis predicts the first natural frequency occurs at 74,000 rpm. In order to achieve

higher speed, the friction wheel material was replaced by aluminum, resulting in a lighter

part. The second analysis using the aluminum friction wheel predicts the first natural

frequency to be 84,000 rpm. Figure 5-2 shows the results of the second analysis. The 2nd

and 3rd natural frequencies are also shown, as well as the calculated shaft deflection with

a 10 N radial load applied to section #2.
















h 25,0 U_
25.0
Friction Wheel Motor Arbor


Figure 5-1. Motor, motor arbor and friction wheel assembly for FEA.


Figure 5-2. Results of FEA of motor and friction wheel assembly1


1 Courtesy Paul Frederickson, Precise Corporation, Racine,Wisconsin, USA.









5.2 FEA of the Micro-Tool, Air-Bearing and the Friction Drive

In order to know the dynamic response of the micro-tool rotating at high rpm in the

air-bearing, a separate finite element analysis was carried out. Figure 5-3 shows the FE

model.










Figure 5-3. FEA model of the tool, air-bearing and friction wheel.

The tool inside the air-bearing is modeled as a continuous beam supported on

springs with composite stiffness equivalent to the air bearing stiffness. The stiffness of

the custom air-bearing was predicted to be approximately 5e5 N/m by extrapolating from

published stiffness values of other air bearings produced by the vendor. This value was

further verified by conducting stiffness measurements after assembly and testing of the

real bearing when it was available. Chapter 6 discusses the stiffness measurements in

detail. The beam is also contacted by another spring with stiffness equivalent to the

Hertzian contact stiffness between the tool and the friction wheel.

For two cylinders in contact the Hertzian contact stiffness is given by the formula:

Kf= 1.333rEE2 (5.1)
K 2 = (5.1)
e V2) E( v V+E 12)


Also,


r = r2 (5.2)
r, +r2


where









E, = Modulus of elasticity of first cylinder.

E2 = Modulus of elasticity of second cylinder.

vi= Poisson's ratio of first cylinder.

v2= Poisson's ratio of second cylinder.

Although the exact material properties of the ML6 polymer coating on the friction

wheel were not available from the vendor, it was possible to find the material properties

of urethane. The polymer ML6 is a urethane class material. The modulus of elasticity of

urethane was found to be 2e7 N/m2, and Poisson's ratio was 0.4. The properties of

tungsten carbide tool are E = 6.8el 1 N/m2 and v = 0.24. Using equation 5.1 and 5.2 the

contact stiffness is calculated to be 1.68e6 N/m.

The FEA analysis predicts the lowest natural frequency of the system to be 4857

Hz. This corresponds to a rotational speed of approximately 292,000 rpm. The analysis

was also performed for stiffness values of the air bearing increasing by up to 100%. As

expected, the first natural frequency increased as the stiffness of the air bearing increased.

This points to the need to increase the air bearing stiffness in future versions. Figure 5-4

to Figure 5-7 shows the results with different mode shapes in different frequencies.


Figure 5-4. First mode: 4857.3 Hz.
















Figure 5-5. Second mode: 1.35e5 Hz.


Figure 5-6. Third mode: 2.65e5 Hz.


Figure 5-7. Fourth mode: 3.89e5 Hz.

5.3 Stress/Strain Analysis

The friction wheel rotates at a very high rpm. So a stress/strain analysis was

pertinent to make sure the design is safe. Hoop stress and radial strain is calculated in the

friction wheel made of aluminum and rotating at 75,000 rpm, which is suppose to be the

maximum operating speed. From the theory of "thick cylinders and rotating disks" [28]

the equation for hoop stress for a solid disk is given by:

ee = po2/8[(3 + i)ro2 (1 + 3))r2] (5.3)

where,

0ee is hoop stress.

p is the density of the material.









u is the Poisson's ratio.

ro is the outer radius of the disk.

Now, putting r = ro in the above equation and solving for 5ee gives:

0ee = po2/4(1 u)ro2 (5.4)

Now again putting co = 7850 rads/sec (75k rpm), p = 2698.9 kg/m3 (density of aluminum)

and u = 0.36 and r = 12e-3 m gives

c00= 38.31 MPa, (5.5)

which is 50% of the tensile strength of aluminum. Hence the friction wheel is safe at this

rpm.

The radial displacement is also calculated using the following formula:

Urr = r/E(5ee uC5) (5.6)

where

E is the modulus of elasticity of aluminum.

,rr is radial stress (= 0 in this case).

So, at 75k rpm Ur = 6.76 am. At 50k rpm this value reduces to 3 am.

Based on these analyses, the design is safe.














CHAPTER 6
TESTING AND RESULTS

This chapter describes the testing of the first generation ultra-high-speed micro-

spindle. All initial testing is carried out at University of Florida Machine Tool Research

Center. Figure 6.1 shows the experimental set-up. Run-out measurements, cutting force

measurements, and air-bearing stiffness measurements comprise the sections of this

chapter. All the cutting tests are carried out using aluminum as the workpiece material.

Other materials such as mild steel were also tried but gave unsatisfactory results. It is

believed this is due to insufficient stiffness in the air-bearing.

6.1 Run-out Measurements

Radial run-out measurements were carried out using a Lion Precision capacitance

probe and amplifier setup. The amplifier was interfaced with PCScope. The data was

recorded at a sampling rate of lx105 Hz. The run-out measurements were carried out at 3

different places.

6.1.1 Run-out Measurement at Motor Arbor

Initially, the radial run-out of a point on the drive spindle arbor was measured at

various spindle speeds both with and without the friction wheel mounted. The results are

shown in Figure 6-2. Without the friction wheel attached to the arbor the run-out remains

less than two microns, which is the value given in the manufacturer's specification sheet

for the motor. With the friction wheel installed and run-outs measured at the same point

on the arbor, there is a clear parabolic increase in radial run-out with increasing speed.









This indicates an unbalance in the friction wheel with centrifugal forces causing

increased run-out at higher speeds.






j-













o 8i





Figure 6-1. Experimental set-up.

Drive spindle arbor radial runout





10




20000 30000 40000 50000
Spindle speed (rpm)


Figure 6-2. Run-out of drive spindle arbor.










6.1.2 Run-out Measurements at Friction Wheel and Gage Pin

The radial run-out of the friction wheel surface was measured using a capacitance

probe reading against the wheel surface through the polymer coating. These results are

somewhat unreliable since they contain effects both due to the non-circularity of the

metal portion of the friction wheel, and thickness variations of the polymer coating. The

yellow line in the Figure 6-3 shows the run-out at the friction wheel.

Finally, a gage pin was inserted in the air bearing and driven by the friction wheel

to measure run-out near the location of the tool point, approximately 6mm from the end

of the air bearing. The results of these measurements are shown in Figure 6-3 by a blue

line.


Radial runout of friction wheel surface and gage
pin

120
100 -
S 80 Friction wheel
S80 surface
E 60
4R 2 Gage pin tip
2 40
E 20
0
20000 30000 40000 50000
Spindle speed (rpm)


Figure 6-3. Run-out of friction wheel and gage pin.

These results show that the friction wheel surface is far too irregular for this

application and the imperfections in its surface are transmitted into the tool shank and

amplified at the tool tip. Further analysis of the air bearing and friction wheel, and

improved manufacturing methods for the friction wheel are needed to rectify this

problem.









6.2 Force Measurements

Although the tool tip run-out was exceedingly high as discussed above, preliminary

cutting tests were performed to evaluate the cutting performance of the micro spindle.

The initial cutting tests were performed at a drive spindle speed of 10,000 rpm which

yielded a nominal tool speed of approximately 79,000 rpm. Slotting cuts were performed

using tool diameters ranging from 0.03(762pm) inches to 0.005(127pm) inches.

Commanded slot depths of one-sixth of the tool diameter, and one third of the tool

diameter were used. Subsequent measurements of the slots using a measuring

microscope showed a significant variation in the slotted depths. For each tool diameter,

depth, and width of cut, the feed rate was increased until the tool failed.

Figure 6-4 shows the feed direction force record for 127 im cutter with a 42.33 im

slot depth and the largest achievable feed rate without tool breakage of 0.045 mm/min

(0.376 tm/tooth). The passage of individual teeth is clearly evident in the force record,

and shows that the actual tool speed is approximately 67568 rpm, indicating significant

slippage in the friction drive. Significant variation in the peak cutting force is also

evident, and may be due to asynchronous error motions of the tool.

Typical cutting test results are shown in table 6-1. Both Tlusty and MacNeil [5] and

Bao and Tansel [10] cutting force models were used to calculate the cutting force

coefficient, Ks, using the average maximum feed direction force. The average maximum

feed direction force was found using a Matlab code to average the maximum force for

each tooth passage over a substantial number of tooth passages. For milling with macro

tools, Ks for this material will be on the order of 850 N/mm2

These results are somewhat unreliable due to fact that measured forces are different

than the actual cutting forces. The air bearing sitting on top of the force sensor affects the







47



dynamics of the system. Figure 6-5 shows the mathematical model of the tool/air-


bearing/force sensor assembly. A hammer testing experiment needs to be done to actually


get the transfer function of the system. This transfer function needs to be inverted and


then multiplied with measured forces in frequency domain to get cutting forces in


frequency domain. Again these cutting forces in frequency domain need to be converted


to forces in time domain which will be actual cutting forces.


Feed Direction Force Plot


---- -- -i--------------r----- -
I -
I






I I


13.26 13.261 13.262 13.263
Timerri(Sie:)


Figure 6-4. Force plot for micro-milling of Al using 127lpm micro-tool.


13.264


--------- ~---- -------------- ------





-- j-- ----- -- -
I


,
.. . T - - - -












Transfer Function


Fmeasured


Figure 6-5. Mathematical model of the tool, air-bearing and force sensor system.

Table 6-1. Cutting test results Aluminum 6061
Dia Width Depth Speed Feed Feed Avg Max Cutting Cutting
of cut of cut /tooth Force Coeff Coeff.
(mr) (um) (um) (rpm) (mpm) (um) (N) (N/mm^2) (N/mm^2)
(Tlusty) (Tansel)
762 762 70.22 60976 0.144 1.18 2.8157 58200 4630
762 762 73.13 60976 0.203 1.66 6.0838 85600 7100
508 508 91.44 66195 0.296 2.24 5.0212 42000 6530
508 508 93.98 59524 0.313 2.63 6.3418 43900 7010
254 254 104.1 67568 0.0593 0.439 0.8811 33000 11700
254 254 106.7 65789 0.0762 0.579 1.1787 32600 11900
127 127 42.33 67568 0.0253 0.187 0.6678 144000 41600
127 127 42.33 67568 0.0508 0.376 1.1292 121000 35000










The actual speed of the tool rotation is back calculated from the tooth passing

frequencies obtained from the force plot. Figures 6-6 to 6-7 show photographs of some of

the machined slots.

I-


.I. -


Figure 6-6. Slots made from 508im micro-tool.


Figure 6-7. Slots made from 254im micro-tool.









The actual depth of cut and width of cut of these slots was also measured using a

measuring microscope and table 6-1 shows the actual values rather than commanded

ones.

6.3 Machining of Hexagonal Test Part

In order to test the capability of the spindle to create miniature features a hexagonal

feature was machined in aluminum using a 254im micro tool with a wall thickness of 80

nm. Figure 6-8 shows the close up photograph of this feature.


Figure 6-8. Hexagonal feature with a wall thickness of 30m.

The total depth of the feature was created in two layers, with each layer having a

depth of cut equal to half of the diameter of the tool. A constant feed rate of 0.0593 mpm

was used. A number of tests were carried out to determine the minimum wall thickness

that can be machined. To accomplish this, adjacent slots were widened in steps of 10lm

using 254rim micro tools. Using this procedure, it was determined that the minimum









thickness of the wall that can be machined is 30m. Figure 6-8 shows results from some

of these trials. We were not able to achieve this thickness in the hexagonal feature due to

non repeatable error motions of the X and Y axes of HSM-2. The machined part has a

commanded wall thickness of 80 im. Measurements the machined walls showed

thickness variations ranging from 30im to 120im.

6.4 Air-Bearing Stiffness Measurements

The stiffness of the air-bearing was measured to compare with values predicted

by extrapolation of manufacturer's data for larger diameter bearings. The set-up was

similar to the one used for the run-out measurements. For the stiffness measurement, one

of the PCScope channels was used to measuring the Z-direction force. The other channel

was used for displacement measurement in the same direction using the capacitance

probe. The charge amplifier was set to the highest time constant (5 sec) making it suitable

for quasi-static force measurements. The bearing was pressurized and the tool was

mounted in the air-bearing. The tool shaft was manually loaded near the friction wheel

using a screw driver while the force and displacement were recorded. The experiment

was repeated a number of times. The recorded data was transferred to Matlab and force

versus displacement plots were created. Figure 6-9 shows a typical graph.














6



5



-' 4




E
aC

C2







,


Force Vs Displacement Plot for Air Bearing Stiffness


F.:.rc e(N)


Figure 6-9. Force vs. displacement plot to measure the stiffness.


The slope of the above curve shows the stiffness to be approximately 3.8e5 N/m.


The measured values of air bearing stiffness were very close to the values predicted by


extrapolating from manufacturer's data, as described in the previous chapter.


Stiffness measurements were also carried out without air pressure in the air


bearing and resulted in values approximately one order higher (6e6 N/m to 7.5e6 N/m),


indicating that the measured stiffness of the pressurized bearing was primarily due to


displacements of the tool shaft within the air bearing bore.


-------------- -------------- ------------ ---- ------







------------------------- ----'"--------------~-----------



--------- ---- ------- r------------ -------- - - -

----


-


)














CHAPTER 7
CONCLUSIONS AND FUTURE WORK

This chapter summarizes the results from initial testing of the micro-spindle, and

provides recommendations for future work.

The friction wheel shows excessive run-out which increases with increasing RPM,

due to high centrifugal forces resulting from wheel unbalance. At drive spindle speeds

greater than 20k rpm it appears that the tool shank begins to contact the inner surface of

the air bearing, which leads to heat generation, restricting the spindle to lower speeds in

order to ensure that the tool floats. This is believed to be related to the problems with

friction wheel run-out and unbalance, which are exacerbated at higher speeds.

Run-out measurements using the gage pin gives values much higher than the target

values for the spindle. It is believed that this is due to excessively high contact stiffness

between the tool and the polymer coated friction wheel, leading to transfer of the error

motions of the friction wheel into the tool. Despite this, the spindle is capable of

operating at speed in excess of 500,000 rpm for short periods. We believe that design

improvements outlined below will allow the micro-spindle to operate successively at

much higher speeds and with low tool run-out.

Suggestions for the future designs of the spindle are:

Redesign the friction wheel to achieve proper balance when mounted to the
motor arbor.

Change the friction coating on the friction wheel to reduce the contact
stiffness between the tool and the friction wheel.









* Improve manufacturing methods for the coating (such as abrasive finishing) to
achieve improved concentricity with respect to the friction wheel.

* Redesign the air bearing to increase its stiffness.

* Commercially available micro-tools have large shaft diameter variations
which may limit achievable air bearing stiffness, since high stiffness requires
small air gaps. This may require special tools with precision ground shaft
diameter. Alternatively, special precision ground collars should could be used
to support the commercial tools.

* Incorporate a tool speed sensor in order to know the actual speed of the tool
rotation. The current design senses the rotational speed of the drive motor.
However, the actual speed of tool rotation is unknown if slip in the friction
drive.

* Replace the current charge amplifier with charge amplifiers which have a
large range of adjustable time constant. This will allow the force sensor to
provide better quasistatic force measurements for setting the contact between
the friction wheel and the tool.

* Incorporate sensors to measure the error motions and run-out of the tool shaft
during high speed rotation.















APPENDIX A
PART LIST














VENDOR P/N COST EA. COST


MSDO1
MSD02
MSD03
MSD05
MSD06
MSD10
MSDP01
MSDP02
MSDP03
MSDP04
MSDP05
MSDP06
MSDP07
j, MSDP08
MSDP09
MSDP10
MSDP11
MSDP12
MSDF13
MSDP14
MSDP15
MSDP16
MSDP17
MSDP18
MSDP19
MSDP20
MSDP21
MSDP22
MSDP23
MSDP24


SPLIT MOUNT, MOTOR
BASE FRAME
CUSTOM AIR BEARING ASSEMBLY
FRICTION WHEEL
POLYMER COATING
INTERFACE PLATE
SC40 SPINDLE, WITH POWER CABLE
MOTOR ARBOR
FREQUENCY CONVERTOR, TYPE PCF 310
COOLING SYSTEM, 115V, TYPE 7136
BASE TUBING KIT
3 AXIS FORCE SENSOR
FORCE SENSOR CONNECTING CABLES
CHARGE AMPLIFIER, PCB
SPHERICAL WASHER, LARGE, MALE
SPHERICAL WASHER, LARGE, FEMALE
SPHERICAL WASHER, SMALL, MALE
SPHERICAL WASHER, SMALL, FEMALE
FINE ADJUSTMENT SET SCREW, 1/4-80, 2" LONG
MICRO TOOLS, DIA. 0.005"
MICRO TOOLS, DIA. 0.010"
MICRO TOOLS, DIA. 0.015"
MICRO TOOLS, DIA. 0.020"
MICRO TOOLS, DIA. 0.030"
CAPACITANCE PROBE, 3/8" CYLINDRICAL
CHARGE AMPLIFIER, CAPACITANCE PROBE
GAGE PIN, 0.1247", CLASS XX
GAGE PIN, 0.1248", CLASS XX
GAGE PIN, 0.1249", CLASS XX
GAGE PIN, 0.1251", CLASS XX


PENNSYL. TOOL
PENNSYL. TOOL
NEWWAY
SANDIA
MERIDIAN LABS.
PENNSYL. TOOL
PRECISE CORP
PRECISE CORP
PRECISE CORP
PRECISE CORP
PRECISE CORP
KISTLER
KISTLER
MTRC LAB
J.W. WINCO
J.W. WINCO
J.W. WINCO
J.W. WINCO
THORLABS
NATIONAL TOOL
NATIONAL TOOL
NATIONAL TOOL
NATIONAL TOOL
NATIONAL TOOL
MTRC LAB.
MTRC LAB.
MEYER GAGE
MEYER GAGE
MEYER GAGE
MEYER GAGE


CAT # 550279
CAT # 400510


1000.00
800.00
3000.00
1000.00
320.00
500.00
3600.00
245.00


CAT # 330916 3600.00
CAT # 332370 1900.00
CAT # 332004 45.00
9018A 3168.00
1694A 366.00
443B101 0.00
43NG40/CNI 54.68
49NG40/DNI 55.78
64NG40/CNI 1.70
71NG40/DNI 1.79
FAS200 8.20
15.75
ITEM # 32352 15.50
ITEM # 32357 15.50
ITEM # 32362 15.00
ITEM # 32372 12.50
C1C 0.00
DMT20 0.00
0.1247X CLX- 19.00
0.1248X CLX- 19.00
0.1249X CLX- 19.00
0.1251X CLX 19.00


1000.00
800.00
3000.00
1000.00
320.00
500.00
3600.00
245.00
3600.00
1900.00
45.00
3168.00
366.00
0.00
54.68
55.78
1.70
1.79
32.80
236.25
232.50
232.50
225.00
187.50
0.00
0.00
19.00
19.00
19.00
19.00


P/N DESCRIPTION


QTY VENDOR













MSDP25
MSDP26
MSDP27
MSDP28
MSDP29
MSDP30
MSDP31
MSDP32
MSDP33
MSDP34
MSDP35
MSDP36
MSDP37
MSDP38
MSDP39
MSDP40
MSDP41
MSDP42
MSDP43
MSDP44
MSDP45

MSDP46
MSDP47
MSDP48
MSDP49
MSDP50
MSDP51

MSDP52
MSDP53
MSDP54
MSDP55
MSDP56


7404 HEX INVERTOR CHIP
BREAD BOARD
REGISTORS
NI, DATA ACQUISITION CARD,
RIBBON CABLE
I/O CONNECTOR

15 PIN CONNECTOR CORD
15 PIN FEMALE JACK
BNC JACKS
BNC CONNECTING CABLES
LAPTOP


1200 SERIES


GAGE PIN, 0.1252", CLASS XX
GAGE PIN, 0.1253", CLASS XX
GAGE PIN, 0.1254", CLASS XX
GAGE PIN, 0.1255", CLASS XX
GAGE PIN, 0.1256", CLASS XX
GAGE PIN, 0.1257", CLASS XX
HOSE NIPPLE, 1/4" HID,1/4" PIPE
HEX COUPLING, 1/4" X 1/8"
COMP. AIR FILTER
DESICCANT AIR DRYER
COALESCING FILTER, OIL REMOVAL
OIL VAPOR FILTER
REGULATOR, WITH GAGE
LEVER CONTROLLED REGULATOR
FITTING, 10-32 THREAD TO 1/16" ID HOSE
BRAID-REINFORCED PVC TUBING
POLYURETHANE TUBING, 1/16" ID, 1/8" OD
CABLE TIES, LONG
CABLE TIES, FINE
BLACK BOX
PUSH BOTTON SWITCH


MEYER GAGE
MEYER GAGE
MEYER GAGE
MEYER GAGE
MEYER GAGE
MEYER GAGE
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
MCMASTER
RADIO SHACK
RADIO SHACK
ELECTRONIC
PLUS
MTRC LAB
MTRC LAB
MTRC LAB
MTRC LAB
MTRC LAB
ELECTRONIC
PLUS
MTRC LAB
MTRC LAB
MTRC LAB
MTRC LAB


0.1252X CLX -
0.1253X CLX -
0.1254X CLX -
0.1255X CLX -
0.1256X CLX -
0.1257X CLX -
5346K14
9171K72
4274K13
5164K77
59185K11
5757K61
9892K11
4158K72
5454K61
52375K12
5184K61
7130K16
7130K12


19.00
19.00
19.00
19.00
19.00
19.00
0.45
3.90
20.74
41.17
83.10
33.94
21.00
22.26
0.48
0.37
0.26
0.0804
0.0180
6.23
3.23

0.78
0.00
0.00
0.00
0.00
0.00

11.75
0.00
0.00
0.00
0.00


19.00
19.00
19.00
19.00
19.00
19.00
6.30
3.90
20.74
41.17
83.10
33.94
21.00
22.26
0.96
18.50
7.80
4.02
1.80
6.23
3.23

0.78
0.00
0.00
0.00
0.00
0.00

11.75
0.00
0.00
0.00
0.00














HEX BOLT, M8 X 1.0
HEX BOLT, 2 13, 3" LONG
WASHER, 1/2-13 BOLT
SOCKET HEAD SCREW, M5 X 0.8
SOCKET HEAD SCREW,


4 MSC
2 MTRC LAB
2 MTRC LAB
3 MTRC LAB
1 MTRC LAB


MSDF57
MSDF58
MSDF59
MSDF60
MSDF61


0.42
0.00
0.00
0.00
0.00
TOTAL


1.68
0.00
0.00
0.00
0.00
21283.66















APPENDIX B
DETAIL DRAWINGS












































I C>
















LIST OF REFERENCES


1. Friedrich, C. R. and Vasile, M. J., 1996, "Development of the micro milling
process for high aspect ratio microstructures," Journal of Microelectromechanical
Sytems, v 5, n 1, pp. 33-38.

2. Adams, D. P., Vasile, M. J., Benavides, G., and Campbell, A. N., 2001,
"Micromilling of metal alloys with focused ion beam-fabricated tools," Precision
Engineering, v 25, pp. 107-113.

3. Vasile, M. J., Friedrich, C. R., Kikkeri, B., and McElhannon, R., 1996,
"Micrometer-scale machining: tool fabrication and initial results," Precision
Engineering, v 19, pp. 180-186.

4. Friedrich, C. R., Coane P., Goettert J., and Gopinathin N., 1998, "Direct fabrication
of deep x-ray lithography masks by micromechanical milling," Precision
Engineering, v 22, pp. 164-173.

5. Tlusty, J. and MacNeil, P., 1975, "Dynamics of cutting forces in end milling,"
Annals of the CIRP, v 24, n 1, pp. 21-25.

6. Gygax, P. E., 1979, "Dynamics of single-tooth milling, IWF/ETH Zurich," Annals
of the CIRP, v 28, n 1, pp. 65-70.

7. Kline, W. A. and DeVor, R. E., 1983, "The effect of cutter runout on cutting
geometry and forces in end milling," International Journal of Machine Tool Design
and Research, v 23, pp. 123-148.

8. Tlusty, J. and Ismail, F., 1983, "Special aspects of chatter in milling," Trans. of
ASME, Journal of Engineering for Industry, v 105, pp. 24-32.

9. Ismail, F. and Bastami, A., 1986, "Improving stability of slender end mills against
chatter," Trans. of ASME, Journal of Engineering for Industry, v 108, pp. 264-268.

10. Bao, W. Y. and Tansel, I. N., 2000, "Modeling micro-end-milling operations, Part
I: analytical cutting force model," International Journal of Machine Tools &
Manufacture, v 40, pp. 2155-2173.

11. Bao, W. Y. and Tansel, I. N., 2000, "Modeling micro-end-milling operations, Part
II: tool run-out", International Journal of Machine Tools & Manufacture, v 40, pp.
2175-2192.









12. Bao, W. Y. and Tansel, I. N., 2000, "Modeling micro-end-milling operations, Part
III: influence of tool wear," International Journal of Machine Tools & Manufacture,
v 40, pp. 2193-2211.

13. Gu, F., Kapoor, S. G., and DeVor, R. E., 1991, "An approach to on-line cutter run-
out estimation in face milling," Transactions of the North American Manufacturing
Research Institution of SME, pp. 240-247.

14. Sutherland, J.W., and DeVor, R.E., 1986, "An improved method for cutting force
and surface error prediction in flexible end milling systems," Journal of
Engineering for Industry, v 108, pp. 269-279.

15. Cook, N.H., Subramanian, K., and Basile, S.A., 1975, "Survey of the state of the
art of tool wear sensing techniques," Materials Processing Laboratory, Department
of ME, MIT, Cambridge.

16. Altintas, Y., Yellowley, I., and Tlusty, J., 1988, "The detection of tool breakage in
milling operations," Trans. of ASME, v 110, pp. 271-277.

17. Takata, S., Ogawa, M., Bertok, P., Ootsuka, J., Matushima, K., and Sata, T., 1985,
"Real-time monitoring system of tool breakage using Kalman filtering," Robotics
and Computer-Integrated Manufacturing, v 2, n 1, pp. 33-40.

18. Liang, S., and Domfeld, D.A., 1989, "Tool wear detection using time series
analysis of acoustic emission," Trans. of ASME, Journal. of Engineering for
Industry, v 111, pp. 199-205.

19. Tansel, I. N., and McLaughlin, C., 1993, "Detection of tool breakage in milling
operations: Part 2 The neural network approach," International Journal of
Machine Tools and Manufacturing, v 33, n 4, pp. 545-558.

20. Vogler, M. P., Liu, X., Kapoor, S. G. and DeVor, R. E., 2002, "Development of
meso-scale machine tool (mMT) systems," Transaction of the North American
Manufacturing Research Institution of SME (NAMRI), 30, pp. 653-662.

21. IDCT Micro-drilling Systems, 1999, 250,000 RPM Zindle Drilling Mechanism,
IDCT Torrance, CA.

22. Kistler, 2002, Operating Instructions, 3 component force sensor, Kistler Instrument
Corporation, New York.

23. Precise, 2002, Operating and maintenance instructions for spindle series: SC 40,
SC 60, SC 77, Precise Corporation, Racine, WI.

24. Precise, 2002, Operating instructions for solid state frequency converter, type PCF
310, Precise Corporation, Racine, WI.






68


25. Precise, 2002, Operating and maintenance instructions for coolant circulation
systems, types 7136, 7137, and 7138, Precise Corporation, Racine, WI.

26. National Instruments, 2003, LabView data acquisition basics manual, National
Instruments, Austin, Texas.

27. National Instruments, 2003, DAQ hardware overview guide, National Instruments,
Austin, Texas.

28. Budynas, R. G., 1999, Advanced Strength and Applied Stress Analysis, WCB
McGrawHill, New York.















BIOGRAPHICAL SKETCH

The author was born in Sultanpur, UP, India, on December 16th, 1977. He grew up

primarily in Calcutta, and in school he used to enjoy thinking, problem/puzzle solving,

working with different ideas and mathematics as his most favorite subjects. This

developed a fancy for a career in engineering and he joined mechanical engineering

stream at Kamla Nehru Institute of Technology, Sultanpur, India. During his

undergraduate study he focused more in the area of designing systems. He earned his

Bachelor of Science degree in July, 2001 and came to pursue master's degree in August,

2001. The author plans to go to industry after graduation preferably in the area of

precision system design.