cl!!! W CL < n MROZ-159 Experimental Analysis of Chip Formation in Micro-Milling author(s) abstract conference terms

Transcription

1 cl!!! W CL < n Experimental Analysis of Chip Formation in Micro-Milling author(s) CHANG-JU KIM MATTHEW BONO JUN NI University of Michigan Ann Arbor, Michigan MROZ-159 abstract Z This study investigates the mechanism of chip formation when milling 360 brass with a 2-flute flat end-mill of diameter 635 urn. Experiments reveal how chips are formed in the micro-milling process, which differs from conventional milling in that the feed per tooth is often smaller than the cutting edge radius of the tool. Several slots are milled in the workpiece using a speed of 80,000 rpm and feeds per tooth ranging from urn to 6 urn. For each set of cutting conditions, the volumes of the resulting chips and the feed marks on the machined workpiece surface are analyzed. conference NAMRC XXX May 21-24,2002 West Lafayette, Indiana terms Micro-milling Chip Formation Chip Volume Feed Society of Manufacturing Engineers Sponsored by the North American Manufacturing Research Institution of the Society of Manufacturing Engineers One SME Drive Dearborn, MI Phone (313) 271-l 500

2 SME TECHNICAL PAPERS This Technical Paper may not be reproduced in whole or in part in any form without the express written permission of the Society of Manufacturing Engineers. By publishing this paper, SME neither endorses any product, service or information discussed herein, nor offers any technical advice. SME specifically disclaims any warranty of reliability or safety of any of the information contained herein.

3 EXPERIMENTAL ANALYSIS OF CHIP FORMATION IN MICRO-MILLING Chang-Ju Kim, Matthew Bono, and Jun Ni Department of Mechanical Engineering University of Michigan Ann Arbor, Michigan ABSTRACT This study investigates the mechanism of chip formation when milling 360 brass with a 2-flute flat end-mill of diameter 635 urn. Experiments reveal how chips are formed in the micro-milling process, which differs from conventional milling in that the feed per tooth is often smaller than the cutting edge radius of the tool. Several slots are milled in the workpiece using a speed of 80,000 rpm and feeds per tooth ranging from urn to 6 pm. For each set of cutting conditions, the volumes of the resulting chips and the feed marks on the machined workpiece surface are analyzed. The data indicate that when the feed per tooth is relatively small compared to the cutting edge radius of the tool, the volume of the chips is significantly larger than the nominal volume, and the spacing between feed marks is significantly larger than the feed per tooth. The data indicate that for a small feed per tooth, a chip is not formed with each pass of a cutting tooth, meaning that chips are produced intermittently. INTRODUCTION The development of miniaturization technologies is a rapidly emerging field of engineering. The benefits and potential of miniature scale devices have been documented in a number of studies, and considerable resources are currently being dedicated to the development of miniaturized technologies for applications in many fields of science and engineering. The major obstacle to the development of these smallscale devices is the difficulty in manufacturing them. Most traditional manufacturing processes are unable to create the required feature sizes with adequate accuracy or precision. Thus, the development of appropriate fabrication processes for manufacturing miniaturized devices and systems is critical to realizing their full potential. Many different technologies have been used to fabricate small-scale devices. The traditional MEMS processes are well established methods of producing a variety of different types of devices. However, these processes have a major drawback, because they are limited in their ability to produce true, three-dimensional features. In addition, MEMS processes can only be used with a narrow class of materials, and their accuracy is limited even though their feature sizes can be at sub-micron level. An alternative manufacturing process that can potentially overcome these problems is the micro-milling process. The potential of the micro-scale milling process has been reported by several researchers, but the mechanics involved with this process are not very well understood. Thus, there are many problems with current micromilling practices, including uncertainty in the

4 optima! cutting conditions, frequent tool breakage, and poor surface finish. Many of the current problems encountered in the micro~milling process result from attempts to simply reduce conventional milling processes to 3 smaller length scale. Unfortunately, different principles are at work in the micro-milling process, so the development of improved micro-milling toois and processes will require some basic research to identify the fundamenta! mechanisms of chip formation. Micro-scale milling differs from conventional scale milling in several important ways. In conventional scale milling, the feed per tooth is usually much larger than the cutting edge radius of the tool, but the feed per tooth in micro-milling is often comparable to or even less than the cutting edge radius. In addition. the small size of micro-milling cutters makes thorn very weak and results in a smali stiffness And finally, at the conventional scale, polycrystalline workpieces essentially act as isotropic, homogeneous materials, but in micro-ngiling, the grains of the workpiece material can bc comparable in size to the tool. These differences complicate the microscale milling process. Research on the micro-mechanical milling process as a method of fabricating miniature structures is still in its early stages. One problem that has received some attention is the ability to create miniature tools. Vasile et al. (1996) fabricated 25 pm diameter steel milling tools using a focused ion beam. These tools machined trenches of width 24 pm. Many other studies have revealed the ability of the micro-milling process to fabricate micro-components (Week et al., 1997, Friedrich et al., 1998, and Schaller et al., 1999). Other issues that have been studied qualitatively include tool wear (Baa and Tansel, 2000) and burr formation (Schaller et al., 1999). Very little research has focussed on the mechanism of chip formation in micro-milling. Some research has been done on chip formation in conventional scale milling, which shares some similar characteristics with milling at smaller scales. Some of the first detailed studies of conventional milling by Martellotti (1941 and 1945) provided a basic understanding of chip formation in the general milling process. Recent efforts have revealed how specific process parameters affect chip formation, such as the study reported by Spiewak (1995), who modeled chip thickness in milling by considering the geometrical cutting path, tool runout, and tool deflection. These studies of conventional scale milling provide a basis for research on the formation of chips in micro-milling. Several researchers have studied simple orthogonal cutting using a depth of cut that is smaller than the edge radius of the tool. Lucca et al. (1991, 1993, and 1994) performed extensive experiments and have contributed to an understanding of energy dissipation and surface generation in ultra-precision machining. The effect of a rounded cutting edge and the size effect in machining have been analyzed by Abdelmoneim and Scrutton (1974), Taminiau and Dautzenberg (1991), and Manjunathaiah and Endres (2000). It has been reported that there exists some minimum depth of cut, below which chips are not generated (Ikawa et al., 1992). Yuan et al. (1996) analyzed this phenomenon and proposed that chips will not be formed when the depth of cut is less than about 20% - 400/u of the cutting edge radius. The current research studies the mechanism of chip formation in micro-milling and reveals some important information about how the tool interacts with the workpiece material. This type of fundamental study contributes to a comprehensive understanding of the micro-milling process that will allow engineers to develop improved tools and processes. To study the formation of chips, slots are machined using micro-milling cutters using feeds of up to 6 pm per tooth. The chips produced in the cutting tests are collected, and their volumes are estimated from scanning electron microscope (SEM) photographs. In addition, the marks made by the tools on the machined surface of the workpiece are analyzed. The information gained from these experiments has led to a better qualitative understanding of how chips are formed in the micro-milling process. CHIP FORMATION IN MILLING In both conventional scale milling and micro,- milling, material removal is achieved through two concurrent motions: feed motion, and rotation of the tool, as illustrated in Figure 1. In the conventional milling process, the radial feed per tooth is usually much larger than the cutting edge radius, and the tool is relatively stiff. Thus, each pass of the tool produces a chip whose thickness varies along the length, as described by Equation (I). t = t,,, sine (1)

5 t is the instantaneous uncut chip thickness tmbx is the maximum uncut chip thickness, and C) is the angular position of the tooth (Martellotti 1941). The experiments are performed on a CNC micro-drilling machine equipped with hydrostatic slides and a pneumatic spindle supported on an air bearing. The maximum operation& spindle speed of 80,000 rpm is used for all of the experiments. The feeds range from urn to 6 pm per tooth. The accuracy of the motion of the slides is measured with a laser interferometer (HP 5528A) to ensure the smoothness of the feed motion. The measurements show that the table of the CNC machine moves with a velocity error of less than 15% at the smallest feed per tooth of pm, and a velocity error of less than 3% at 6 urn per tooth. FfGURE 1. SCHEMA3 IC OF THE MILLING PROCESS I TABLE 1. EXPERIMENTAL CONDITIONS Workpiece i Brass 360 _I 1 ESI micro-drilling machine Hydro-static Air-turbine slides spindle - ~_ In conventional milling, the radial feed per tooth usually coincides with tzlx. However, in micro-scale milling, the structural weakness of the tool often requires a radial feed per tooth that is smaller than the cutting edge radius. The theory of the minimum depth of cut (Yuan et al., 1996) predicts that a chip wilt not form when the depth of cut is significantly smaller than the edge radius of the tool. Moreover, the small diameter of a micro-milling tool results in a small stiffness, so the tools deflect more easily. The relatively small bending stiffness of the tool causes the chip formation process to differ even further from the conventional case. Thus, in micro-milling, the maximum uncut chip thickness, tmox, can differ from the feed per tooth. The current study performs an experimental investigation of this premise. EXPERIMENTAL Cuttinq Conditions APPROACH To investigate the formation of chips in the micro-milling process, a series of experiments is performed using the conditions listed in Table 1. The workpiece is a rectangular bar of brass 360 with a width of 12.7 mm and a thickness of 3.2 mm. The tools are commercially available miniature tungsten carbide 2-flute flat end-mills. The diameter of the flute portion of the mills is 635 urn, and the length of the flute section is 1.5 times the diameter. [Axial , depth of cut Radial depth of cut; Feed per tooth [ 2 flute flat end-mill 25.4 pm Full cut pm - Before performing each cutting test, the surface of the workpiece is milled to a smooth finish using a 2 mm diameter flat end-mill. For each feed rate tested! nine grooves of length 12.7 mm are cut in the workpiece, and the resulting chips are collected. Measurement of Chip Volume The chips collected in each experiment are analyzed and photographed using a SEM. For each set of cutting conditions, the pictures are used to estimate the volume of typical chips. Naturally, not all of the chips produced by a particular combination of feed and speed are identical, and there is some variation in their width, thickness, and length.

6 4 For the most part. chip thickness and width tend to be similar among the chips. However, the lengths of the chips vary considerably, probably because the chips are fragile and can break easily. Thus, while some of the coilectt?d chips have a fuli length that corresponds to initial engagement of the cutting tooth to disengagement, many of the chips are shorter, because they have broken. To determine the typical chip voiume fur each experiment, six of the largest available chips are selected, and the average value of their measured volumes is calculated. The measurement of chip volume from a picture of the chip is complicated by the fact that the thickness of a chip produced in a milling operation varies along its length. For example, in an ideal milling process, the uncut chip thickness is zero at the end of the chip. increases to a maximum value at its midlength. and then decreases back to zero at the other end, as shown in Figure 1. Therefore, the cut chip thickness also varies continuotisly along the length of the chip. Unfortunately, the thickness of the chip can only be determined at tl~ose points in the photograph where the edge of the chip is visible, such as points 0 through 5 in Figure 2. i is the index of the points along the chip at which the thickness is known. n is the number of sections into which the chip is divided. For example, for the chip in Figure 2, n is 5. ~1 is the measured width of the chip, which is relatively uniform along the length. ti is the measured thickness of the chip at point i. Ic,.,,.); is the measured length of the section of chip between point i-7 and point i. xeasurement of Marks on the Machined Surface After performing each cutting test, the surface of the machined workpiece is photographed with a magnification of between IO0 and 400. From these pictures, the distances between the machining marks at the center of the milled slots are measured using the image analysis software ImageJ 1.23~. EXPERIMENTAL Measured Chip Volume RESULTS Figure 3 contains pictures of some typical chips produced using feeds ranging from pm per tooth (Figure 3A) to 6 pm per tooth (Figure 3F). These pictures illustrate that each set of cutting conditions produces long, ribbon-like chips. Each of the chips is curled, so there are several places along the length of the chip where the thickness can be measured. The pictures also reveal that larger feeds tend to produce thicker chips. FIGURE 2. PICTURE OF A CHIP SHOWING POINTS AT WHICH THE THICKNESS CAN BE ML4SURED Using pictures like the one shown in Figure 2, the thickness of each chip is determined at several points along its length. The chip volumes are then estimated using the trapezoidal numerical integration formula on the right side of Equation (2).

7 5 (A) FPT = vn-1 (8) FPT = pm (F) FPT = 6 urn FIGURE 3. SEM PICTURES OF TYPICAL CHIPS (C) FPT = 0.X urn Some interesting trends can be identified from the calculated volumes of the chips produced in the different cutting tests. The chip volumes calculated from Equation (2) are graphed as a function of feed per tooth in Figure 4. Recall that for each set of cutting conditions, six typical chips are analyzed and measured. Each dot on the graph represents the average value of volume obtained from the six chips, and the bars represent the standard deviation of volume. (D) FPT = 1.5 urn Feed per tooth [pm] FIGURE 4. MEASURED CHIP VOLUMES Figure 4 shows that as the feed rate increases, the volume of the resulting chips

8 increases as well. This same data is presented a different manner in Figure 5 jr> The following figure shows some SEM pictures of the machined workpiece surfaces at the center of the slots. The tool moves in the direction from the bottom to the top of the pictures. Note that feed marks are generated by both the leading and trailing edges of the milling cutter, and several of the pictures reveal mars on the machined surfaces. However, in each picture, the feed marks are clearly distinguishable Feed per tooth [pm] FIGURE 5. RATIO OF MEASURED CHIP VOLUME TO NOMINAL CHIP VOLUME (A) FPT = pm (B) FPT = pm Figure 5 graphs the ratio of the measured chip volume to the nominal chip volume as a function of feed per tooth. The measured chip volume is the volume presented in Figure 4; it is calculated from the pictures of the chips using Equation (2). The nominal chip volume is a function of the feed rate of the tool; it is simply the product of the feed per tooth, the axial depth of cut, and the tool diameter. In an ideal milling process, the measured chip volume is equal to the nominal chip volume. Figure 5 illustrates that in the micro-milling process, when the feed per tooth is very small, the measured chip volume is much larger than the nominal chip volume. As the feed per tooth increases, the measured chip volume approaches the nominal chip volume, and in this case, for a feed per tooth of 6 pm, the measured and nominal chip volumes are essentially the same. This data indicates that for a small feed per tooth, a chip is not formed with each pass of a cutting tooth. (C) FPT = 0.75 pm (E) FPT = 3 pm (D) FPT = 1.5 pm (F) FPT = 6 pm FIGURE 6. FEED MARKS ON THE MACHINED SURFACES The distances between the feed marks measured from these pictures are graphed in Figure 7. Distance Between MachininqjMarks.- Further evidence that a chip is not formed with each pass of a tooth is obtained by examining the machined surface of the workpiece. In most milling processes, the tool creates feed marks on the bottom surface of the workpiece, and the distance between these feed marks is generally equal to the maximum uncut chip thickness (Sutherland and Babin, 1988). By examining the feed marks on the bottom surface of the workpiece, it is possible to determine the interval at which chips are formed Feed per tooth [urn] FIGURE 7. MEASURED FEED MARK SPACING

9 7 Figure 7 shows that for a large feed per tooth, the distance between feed marks is approximately equal to the feed per tooth. This data indicates that a chip forms with each pass of the tool when the feed per tooth is large. However, for a small feed per tooth, the distance between feed marks is much larger than the feed per tooth, which indicates that chips do not form with each pass of the tool. Thus, the feed marks on the bottom surface provide further support to the conclusion that if the feed per tooth is too small, then a chip is not formed with each pass of the tool. DISCUSSION The observed variation in the mechanism of chip formation with the feed per tooth is probably caused by the combined effect of 1) the ratio of the cutting edge redius to the feed per tooth and 2) the small bending stiffness of the cutting tool. As reported by Yuan et al. (1996): if the depth of cut is significantly smaller than the cutting edge radius of the tool, a chip will not be formed. Instead, the tool will either run along the surface of the workpiece without removing any material, or a type of plastic deformation similar to burnishing may occur. Figure 8 illustrates these phenomena. For this data point, the measured volume of the chips is approximately 9 times the nominai chip volume. It is possible that during this milling process, the cutting teeth pass over the workpiece surface 8 times without performing any cutting, and on the ninth tooth pass, a chip is formed. This process is illustrated in Figure 9. During the 8 non-cutting tooth passes, the teeth may perform some sort of burnishing or plowing process. Note that the table of the machine continuously feeds the milling cutter into the workpiece, and due to the small bending stiffness, the tool may simply deflect during these 4 revolutions. After these 8 tooth passes, the table of the machine has moved far enough (or, alternatively, the transverse force acting on the tool has reached a sufficient value) to achieve conditions sufficient for the formation of a chip. Thus, during the ninth tooth pass, the tool cuts the workpiece material and forms a chip. / Workpiece I,-, _... ;../&&i&\ 4th tooi pass I..-+I 8th tooi pass FIGURE 8. MECHANISMS OF TOOL-WORKPIECE INTERACTION FOR VARICUS RATIOS OF DEPTH OF CUT TO EDGE RADIUS The minimum depth of cut phenomenon is relevant to the micro-milling process, because the feeds per tooth used in the process are typically comparable to the cutting edge radii of the tools. Measurements of the commercial milling cutters used in the experiments have revealed that the radii of the cutting edges are on the order of several microns. Thus, the theory of the minimum depth of cut predicts that if the feed per tooth is less than a few microns, then the milling cutter will not produce a chip each time a cutting tooth passes through the workpiece. Consider the data point in Figure 5 corresponding to a feed per tooth of pm. _. / / A FIGURE 9. BENDING OF THE TOOL DURING THE MULTIPLE NON-CUTTING TOOL PASSES PRECEEDING THE FORMATION OF A CHIP This process repeats every 9 tooth passes, so the chips produced with a feed of pm per tooth have a volume of 9 times the nominal chip volume. For a larger feed per tooth, there are fewer non-cutting tooth passes between chipproducing tooth passes. And for a sufficiently large feed per tooth, a chip is formed with each pass. /

10 8 CONCLUSIONS This study investigates the formation of chips in the micro-milling process. After milling several slots in a brass workpiece using a variety of feeds per tooth, analyses are performed on the volumes of the resulting chips and the feed marks on the machined workpiece surface. The results indicate that for a mrcro-milling system with a particular stiffness and cutting edge radius, the mechanism of chip formation varies with the feed per tooth. If the feed per tooth of a milling cutter is relatively small compared to the cutting edge radius of the tool, or if the stiffness of the system is small, a chip may not form with each pass of the tool. Instead, the tool can rotatc several times without performing any cutting, meaning that chips are produced intermittently. ACKNOWLEDGEMENT This work was supported in part by the Engineering Research Program of the Office of Basic Energy Sciences at the Department of Energy. REFERENCES Abdelmoneim, M. Es., and Scrutton, R. F., (1974), Tool Edge Roundness and Stable Buildup Formation in Finish Machining, Journal of Engineering for Industry, Vol. 96, pp Bao, W. Y. and Tansel, I. N., (2000), Modeling Micro-end-milling Operations. Part III: Influence of Tool Wear, International Journal of Machine Tools and Manufacturing, Vol. 40, pp Friedrich C., Coane P.: Goettert J., and Gopinathin N., (1998), Direct Fabrication of Deep X-ray Lithography Masks by Micromechanical Milling, Precision Enginecring, Vol. 22, pp Ikawa, N., Shimada, S., and Tanaka, H., (1992), Minimum Thickness of Cut in Micromachining, Nanotechnology, Vol. 3, pp Lucca, D. A., and Seo, Y. W., (1993), Effect of Tool Edge Geometry on Energy Dissipation in Ultraprecision Machining, Annals of the CIRP, Vol. 42, No. I, pp Of Lucca, D. A., and Seo, Y. W., (1994), Aspects Surface Generation in Orthogonai Ultraprecision Machining, Annals of the UKP, Vol. 43, No. 1, pp Lucca, D. A., Rhorer, R. L., and Komanduri, R.. ( l991), Energy Dissipation in the Ultraprecision Machining of Copper, Annals of the CIRPS Vol. 40, No. 1, pp Manjunathaiah, J., and Endres, W. J., (2000), A New Model and Analysis of Orthogonal Machining with an Edge-Radiused Tool, Journal of Manufacturing Science and Engineering, Vol. 122, No. 3, pp Martellotti, M. E., (1941), An Analysis of the Milling Process, Transactions of the ASME. Vol. 63, pp Schaller, Th., Bohn, L., Mayer, J., and Schubert, K., (1999), Microstructure Grooves with a Width of Less Than 50 pm Cut with Ground Hard Metal Micro End Mills, Precision Engineering, Vol. 23, pp Spiewak, S., (1995), An Improved Model of the Chip Thickness in Milling, Annals of the CIRP, Vol. 44, No. 1, pp Sutherland, J.W., and Babin, T. S., (1988), The Geometry of Surfaces Generated By the Bottom of an End Mill, Proceedings of NAMRC, Vol. 16, pp Taminiau, D. A., and Dautzenberg, J. H., (1991), Bluntness of the Tool and Process Forces in High-Precision Cutting, Anna/s of the CIRP, Vol. 40, No. I, pp Vasile, M. J., Friedrich, C. R., Kikkeri, B., and Mcelhannon, R., (I 996), Micrometer-Scale Machining: Tool Fabrication and Initial Results, Precision Engineering, Vol. 29, pp Week, M., Fischer, S., and Vos, M., (1997), Fabrication of Microcomponents using Ultraprecision Machine Tools, Nanotechnology, Vol. 8, pp Yuan, Z. J., Zhou, M., and Dong, S., (1996), Effect of Diamond Tool Sharpness on Minimum Cutting Thickness and Cutting Surface Integrity in Ultraprecision Machining, Journal of Materials Processing Technology, Vol. 62, pp