Burr size reduction in drilling by ultrasonic assistance

ARTICLE IN PRESS

Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450 www.elsevier.com/locate/rcim

Burr size reduction in drilling by ultrasonic assistance Simon S.F. Chang, Gary M. Bone McMaster Manufacturing Research Institute (MMRI), McMaster University, Hamilton, Ontario, Canada L8S 4L7 Received 30 August 2004; received in revised form 4 November 2004; accepted 5 November 2004

Abstract Accuracy and surface finish play an important role in modern industry. Undesired projections of materials, known as burrs, reduce the part quality and negatively affect the assembly process. A recent and promising method for reducing burr size in metal cutting is the use of ultrasonic assistance, where high-frequency and low-amplitude vibrations are added in the feed direction during cutting. Note that this cutting process is distinct from ultrasonic machining. This paper presents the design of an ultrasonically vibrated workpiece holder, and a two-stage experimental investigation of ultrasonically assisted drilling of A1100-0 aluminum workpieces. The results of 175 drilling experiments with uncoated and TiN-coated drills are reported and analyzed. The effect of ultrasonic assistance on burr size, chip formation, thrust forces and tool wear is studied. The results demonstrate that under suitable ultrasonic vibration conditions, the burr height and width can be reduced in comparison to conventional drilling. r 2005 Elsevier Ltd. All rights reserved. Keywords: Burr; Drilling; Metal cutting; Ultrasonic assistance; Ultrasonic assisted drilling; Vibration assisted drilling

1. Introduction Conventional metal cutting methods produce undesired projections of material that result from plastic deformation, known as burrs. Burrs reduce the accuracy of the parts and subsequent assembly processes. Typically deburring accounts for up to 25% of the total production cost [1]. To reduce or even eliminate the deburring effort, the burr size must be reduced. In this paper, burr size reduction in drilling will be considered. There are various methods to reduce the burr size. These include altering the cutting conditions and using suitable type of coolant. Dornfeld and Ko [2] showed that the influence of feedrate on burr size is not linear, and is dependent on other cutting conditions and on the material being machined. Varying the feedrate during drilling can also reduce burr size [3]. Special drill geometry, such as radial periphery drills, can produce smaller burrs than standard drills [4]. However, these Corresponding author. Tel.: +1 905 525 9140x27591; fax: +1 905 572 7944. E-mail address: [email protected] (G.M. Bone).

0736-5845/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.rcim.2004.11.005

special drill geometries are often expensive to manufacture. Using suitable coolant and tool coating to reduce the friction between the tool and the workpiece was found to produce smaller burrs [3]. However, coolants are expensive, hazardous to worker health, and pollute the environment. Kim et al. [5] have developed an empirical drilling chart to choose suitable cutting condition for different materials in order to reduce burr size. However, these drilling charts are only applicable to limited ranges of drilling. Drilling with a backup material can also reduce burr size [6]. However, this technique cannot be applied when the exit surface of the workpiece is not accessible. A recent and promising technique to reduce burr size is known as ultrasonic-assisted (UA) drilling. The principle of this technique is adding high-frequency (1–200 kHz) and low peak-to-peak (pk–pk) vibration magnitude (2–26 mm) in the feed direction to the tool or workpiece. This cutting process is distinct from ultrasonic drilling. Ultrasonic drilling, also known as rotary ultrasonic machining, is a specific class of ultrasonic machining. Ultrasonic machining is a machining process where a tool is vibrated ultrasonically and feed axially

ARTICLE IN PRESS S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450

into the work material. Abrasive slurry is fed between the vibrating tool and the work material, resulting in material removal by brittle fracture. This brittle fracture is caused by the impacts between the abrasive and the work material that are induced by the vibrating tool. Ultrasonic drilling is an ultrasonic machining process with a rotating cylindrical tool. The rotation of the tool enhances the abrasive process. Ultrasonic drilling has only been applied to brittle materials. On the other hand, UA drilling is a hybrid process of conventional drilling and ultrasonic oscillation. It is applicable to both ductile and brittle materials. The goal of this machining process is to reduce burr size and thrust force. Takeyama and Kato [4] have experimentally shown that burr size reduction in drilling aluminum is possible with UA drilling. Zhang et al. [7] theoretically and experimentally concluded that there exists an optimal vibration condition such that the thrust force and torque are minimized, which results in smaller burrs. Clearly, more work is required to understand the effect of vibration condition on burr size. This paper presents a two-stage experimental investigation of UA drilling of aluminum in terms of burr size reduction. In the first stage, the effect of vibration frequency, pk–pk vibration magnitude, spindle speed, feedrate, and drill diameter on burr size are investigated; in the second stage, the use of coated drills for tool wear reduction is studied. In Section 2, the designs of the actuated workpiece holder and drive circuit are presented. In Section 3 the experimental investigations are presented. Conclusions are given in Section 4.

2. Design of actuated workpiece holder and drive circuit In order to study UA drilling, an actuated workpiece holder and a drive circuit has been designed and built. The workpiece holder consists of a piezoelectric stack actuator, a preloading mechanism, an aluminum fixture, a stainless steel-shell and a base plate (See Fig. 1). The desired vibration conditions were chosen to be 20 kHz and 4 mm pk–pk. The actuator must be capable of producing sufficient force to drive the combined mass of the diaphragm, workpiece holder and the workpiece at this condition. The vibration of the combined mass can be modeled by simple harmonic motion. Fig. 2 shows the free body diagram of the combined mass. The vibration displacement X ðtÞ; velocity V ðtÞ and acceleration aðtÞ of the combined mass is: X ðtÞ ¼ Au sin ð2p f u tÞ,

(1)

V ðtÞ ¼ X_ ðtÞ ¼ 2p f u Au cosð2p f u tÞ,

(2)

aðtÞ ¼ X€ ðtÞ ¼ 4p2 f 2u Au sinð2p f u tÞ.

(3)

443

Bolts

Drill Workpiece

Aluminum Fixture

Actuator

Diaphragm

Base Plate Stainless Steel Shell

Fig. 1. Cross section of the workpiece holder.

X(t) mw

Fw

Fig. 2. Free body diagram of the combined mass.

Table 1 Required specification of the actuator Frequency range

Displacement range

Force delivery

0–20 kHz

0–4 mm

3.2 kN

In Eqs. (1)–(3), Au is the vibration amplitude (half of the pk–pk vibration magnitude) and f u is the vibration frequency. Hence, the force F w required to drive the combined mass mw and the corresponding maximum force magnitude F wMAX is given by F w ¼ mw aðtÞ ¼ 4p2 f 2u mw Au sinð2p f u tÞ,

(4)

F wMAX ¼ 4p2 f 2u mw Au .

(5)

The maximum weight of the workpiece was chosen to be 10 g, and the mass of the combined mass was chosen to be 100 g in the design criteria. The required specification of the actuator can then be defined as in Table 1. A stack actuator manufactured by Sensor Tech. Ltd. (BM532 series with 33 layers of piezoelectric disks) was chosen. The specifications of the actuator are summarized in Table 2. The chosen actuator requires a drive voltage of 200 V pk–pk to produce its maximum displacement. The power requirement can be computed by first considering the required charging current of the actuator. In general, a piezoelectric actuator can be electrically modeled

ARTICLE IN PRESS S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450

444

Table 2 Manufacturer’s specification of the actuator Number of layers (N)

Layers thickness

Max. applied voltage (Va)

Max. displacement (D)

Max. Force delivery

Charge coefficient (da)

Capacitance (C)

33

0.5 mm

200 V

4 mm

5 kN

280  1012 C/N

290 nF

Desired Actuator's Voltage

250

Approximated Actuator's Voltage Internal Capacitance (C)

Va

Actuator’s voltage (V)

Internal Resistance

200

ic

ir

i

150

100

50

Fig. 3. Electric model of a piezoelectric actuator. 0

as a capacitor and a resistor connected in parallel (see Fig. 3) [8]. The internal resistance (typically higher than 1 MO) of the actuator is significantly higher than its capacitance (typically 107 F). Most of the current will therefore pass through the capacitor at high driving frequencies. The required charging current can then be approximated as C

dV a ¼ iC , dt

0

2.5E-05

7.5E-05

0.0001

Fig. 4. Applied voltage vs. time.

Half bridge driver Diode Full-wave rectifer

(6)

where V a is the applied voltage, C is the capacitance, t is the time, and iC is the current passing through the capacitor. Fig. 4 shows the desired actuator voltage vs. time. The drive current can be estimated by using a triangular wave approximation. From this approximation, the potential difference of the actuator will rise from 0 to 200 V in 25 ms. According to Eq. (6), the required charging current is then 2.3 A. A custom drive circuit, using polarity switching, was designed and built to drive the actuator (see Fig. 5). The principle of this circuit is to supply a constant 100VDC to the actuator and switch its polarity at frequencies ranged from 0 to 20 kHz. The full-wave rectifier and capacitor together converts the AC supply voltage to a DC voltage. The half-bridge drivers, MOSFETs, and diodes work together to achieve the polarity switching, and the switching frequency is controlled by a square wave from the signal generator. Experiments confirmed that the designed circuit is capable of driving the actuator at the desired voltages and frequencies. At 20 kHz, a voltage difference from 0 to 200 V can be applied. Due to the nature of the piezoelectric crystal, the piezoelectric actuator cannot withstand a high tensile

0.00005

Time (s)

VB VCC H0 HIN

120VAC supply

VS LIN L0 com

+

18VDC supply Isolation Transformer

Actuator

VB VCC H0 HIN VS LIN L0 com GND

+

MOSFET

−

Signal generator

VB VCC H0 HIN

Fig. 5. Polarity switching circuit.

VS LIN L0 com

ARTICLE IN PRESS S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450

445

were conducted with a CNC drilling machine. Multiple tests were conducted at the same vibration and cutting conditions to allow the variation in the outcome to be studied. The testing specimens used were 25  25  1.6 mm A1100-0 aluminum. The average burr heights were taken by measuring the maximum burr height on each specimen with a toolmakers microscope under four different views (see Fig. 7). Burr widths were measured by first measuring the diameter at the base of the burr and the actual hole diameter using a Vernier caliper, then dividing the difference by two. 3.1. Stage 1 experimental investigation

Fig. 6. Completed workpiece holder.

load, and a preloading mechanism is required in the present application. An AISI 304 stainless steel circular diaphragm was designed and machined to preload the actuator. Using finite element simulation, an estimation of the required thickness was computed. In the current design, a circular disk with 76.2 mm diameter and 0.79 mm thickness was machined and used. Its static stiffness is 7.9  105 N/m (3.2 kN/4 mm). The stress on the diaphragm after deflecting 4 mm is 3.24 MPa, which is smaller than its yield stress (255 MPa), verifying that no plastic deformation will occur. Note that the resonance frequency of the diaphragm (computed using Eq. (7)) is well below the range of operating frequencies. Since the stiffness of the diaphragm will be higher under dynamic loading as long as the operating frequency is away from its resonance frequency, this static approximation is conservative. sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffi 1 k 1 7:9  105 N=m f n;diaphragm  ¼ 450 Hz, ¼ 2p mw 2p 0:1 kg

In the first stage, the effect of vibration frequency, pk–pk vibration magnitude, spindle speed, feedrate, and drill diameter on burr height and width were studied. The drills used were standard uncoated high-speed steel (HSS) drills. Each drill was used to drill five specimens for each test. Fig. 8 shows some photographs of the burrs produced after the drilling tests. The size of each field of view is 4.2 mm  3.2 mm.

Toolmakers microscope

Workpiece

Burr

Fig. 7. Equipmental setup for measuring burr height.

(7) where k is the stiffness and f n;diaphragm is the resonance frequency of the diaphragm. The rest of the components of the workpiece holder (refer to Fig. 1) include an aluminum fixture with a cavity at the center, which provides room for the drill to penetrate the workpiece, stainless-steel shell, and the base plate. The actuator is preloaded between the diaphragm and the base plate. The completed design is shown in Fig. 6.

3. Ultrasonic-assisted drilling experiments and results The actuated workpiece holder and drive circuit were used in a two stage experimental investigation of UA drilling without coolant (i.e. dry cutting). Experiments

Fig. 8. Workpiece samples: 3.18 mm drill 6000 RPM, 1.90 mm/s feedrate.

ARTICLE IN PRESS S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450

446

0.25

1.2

0.21

Ave. burr width (mm)

Ave. burr height (mm)

1

0.8

0.6 Conventional 0.4

0.17

0.13 Conventional 0.09

0.2

0

0.05

0

5

10

15

20

25

0

5

Frequency (kHz)

Figs. 9 and 10 show that the relationships between vibration frequency and burr height and width are not linear. In these figures, the dashed lines represent conventional drilling, and the solid lines (which were formed by calculating the least-squares fit through the data points for a second-order polynomial equation) represent UA drilling. The chip morphology was also examined. When the workpiece was excited at 10 and 15 kHz, long wavy chips were formed (see Fig. 11). At these frequencies, the cutting was continuous, and the vibration forms the waviness of the chips. When the vibration frequency reached 20 kHz, the chips became discontinuous. Because the axial feed of the tool per each vibration cycle was small, the cutting was discontinuous, and ultrasonic impact action (UIA) occurred. This action formed fine powdered chips, which were easier to remove. This caused a reduction of the thrust force acting on the workpiece, resulting in less plastic deformation and smaller burrs. Figs. 12 and 13 show the relationship between pk–pk vibration magnitude and burr size. The relationships are again non-linear. When the pk–pk vibration magnitude passed through a threshold, the burr height deceases while the burr width increases. When the pk–pk vibration magnitude is small, continuous cutting occurs, forming long wavy chips. When pk–pk vibration magnitude is large, the axial feed of the tool per each vibration cycle is small, the cut became discontinuous, forming fine, powdered chips, and smaller burrs. When the pk–pk vibration magnitude continues to increase, the cutting by UIA dominates, increasing the force normal to the rake face (see Fig. 14). This caused the material to begin to rollover earlier, resulting in a wider

15

20

25

Fig. 10. Average burr width vs vibration frequency (3.18 mm diameter drill, 4000 RPM, 1.90 mm/s feedrate, 4 mm).

Fig. 11. Wavy chips produced by UA drilling with low frequency.

0.9

0.7

Ave. burr height (mm)

Fig. 9. Average burr height vs. vibration frequency (3.18 mm diameter drill, 4000 RPM, 1.90 mm/s feedrate, 4 mm).

10

Frequency (kHz)

0.5 Conventional

0.3

0.1 0

1

2

3

4

5

pk-pk vibration magnitude (microns) Fig. 12. Average burr height vs. vibration magnitude (3.18 mm diameter drill, 4000 RPM, 1.90 mm/s feedrate, 20kHz).

ARTICLE IN PRESS S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450 0.25

1

0.2

0.8

447

Conventional

Ave. burr height (mm)

Ave. burr width (mm)

Conventional

0.15

0.1

0.05

0.6

0.4

0.2

UA 0 0

1

2

3

4

0

5

3000

pk-pk vibration magnitude (microns) Fig. 13. Average burr width vs. vibration magnitude. (3.18 mm diameter drill, 4000 RPM, 1.90 mm/s feedrate, 20kHz)

Thicker

5000

7000

9000

Spindle Speed (RPM) Fig. 15. Average burr height vs. spindle speed (3.18 mm diameter drill, 1.90 mm/s feedrate, 20 kHz, 4 mm). 0.4

Short and wide burr

0.35

Conventional 0.3

Thinner

Long and thin burr

Low pk-pk

Ave. burr width (mm)

High pk-pk

0.25 0.2 0.15 0.1

Fig. 14. Effect of vibration magnitude on burr formation. 0.05

burr. These wide burrs have higher stiffness, causing it to maintain a strong contact with the cutting lips, resulting in smaller height. Figs. 15 and 16 compare the burr height and width produced by UA (20 kHz and 4 mm pk–pk) drilling with conventional drilling under different spindle speeds. In conventional drilling, burr height and width increase with spindle speed, but the trends are not linear in UA drilling. There exists certain vibration condition such that the burr size can be effectively reduced for different cutting conditions. On the other hand, there exists a range of vibration conditions that produces larger burrs. It is also observed that at higher spindle speed, the reduction in burr size with UA drilling is more significant. This is because higher spindle speed reduces the uncut chip thickness, resulting in thinner burrs, which are effectively broken by UIA and removed from the hole efficiently.

0 3000

UA

5000

7000

9000

Spindle Speed (RPM) Fig. 16. Average burr width vs. spindle speed (3.18 mm diameter drill, 1.90 mm/s feedrate, 20 kHz, 4 mm).

Other findings included the effect of feedate and drill size. At normal cutting feedrate1, UA drilling performs well, but degrades rapidly at higher feedrates. This is because at higher feedrates, the uncut chip thickness increases, and the chip segmentation effect of the UIA reduces. Drill size was found to have insignificant effect on the efficiency of UA drilling. This is logical because the ultrasonic action is axial and therefore not significantly affected by changes in the radial direction.

1

The feedrate recommended by the manufacturer of the tool.

ARTICLE IN PRESS 448

S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450 1.8 1.6

Ave. burr height (mm)

1.4 1.2 1 0.8 0.6

UA

Conventional

0.4 0.2 0 0

10

20

30

40

# of drilled holes

The experiments concluded that UA drilling effectively reduced burr size when drilling A1100-0 aluminum at 6000 RPM spindle speed and 1.905 mm/s feedrate using 3.18 mm diameter drills. However, the tool wear and chipping using uncoated HSS drills became significant when UA was applied because of fatigue failure (see Fig. 17). 3.2. Stage 2 experimental investigation To reduce tool wear, drills with suitable coating and material should be used. In the second stage of the investigation, 3.18 mm uncoated and TiN-coated HSS drills were used in several drilling experiments without coolant. Each drill was used to drill 30 holes at 6000 RPM and 1.90 mm/s feedrate. Vibration condition applied was 20kHz and 4 mm pk–pk. Based on the results from stage 1, UA drilling performs best under these conditions. Note that these conditions were not necessarily the optimal ones. Figs. 18 and 19 show the average burr height versus number of drilled holes for uncoated and TiN-coated HSS drills, respectively, and Figs. 20 and 21 show the corresponding average burr width. In these figures, hollow data points represent UA drilling, while the solid points represent conventional drilling. Both results consistently showed that UA drilling effectively reduced burr height, while the reduction in burr width is less significant. For conventional drilling with uncoated drills, the burr height follows a slightly increasing trend as the number of drilled holes increases. This is due to the increase in tool wear. UA drilling not only produced an average of 70% smaller burr height than conventional drilling, the trend also shows insignificant size increase with number of drilled holes, suggesting that

Fig. 18. Average burr height vs. number of drilled holes for uncoated drills.

1.8 1.6 1.4

Ave. burr height (mm)

Fig. 17. Worn drills from tests performed at 8000 RPM and 3.81 mm/s feedrate (top two: ultrasonic assisted; Bottom two: conventional).

1.2 1 0.8 0.6

UA

0.4

Conventional

0.2 0 0

10

20

30

40

# of drilled holes Fig. 19. Average burr height vs. number of drilled holes for TiNcoated drills.

UA drilling is beneficial in terms of burr size reduction even with worn tools. TiN-coated drills provide higher wear resistance, and the burr heights produced by conventional and UA drilling show insignificant increase with number of drilled holes. It is however obvious that UA drilling produces smaller burr height (85% average reduction). Although burr width shows relatively minor reduction with UA drilling, in general it is beneficial (40% average reduction). The chip morphology was again examined. Conventional drilling produced long, continuous chips (see Fig. 22), which were difficult to remove. On the other hand, the cuts produced by UA drilling are discontinuous, allowing the occurrence of UIA. These actions produced fine, discontinuous chips (see Fig. 23), which were easy to remove. This reduced the average thrust

ARTICLE IN PRESS S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450

449

0.35

UA

Conventional

Ave. burr width (mm)

0.3 0.25 0.2 0.15 0.1 0.05 0 0

10

20

30

40

Fig. 22. Continuous chips resulted from conventional drilling.

# of drilled holes Fig. 20. Average burr width vs. number of drilled holes for uncoated drills.

0.35

UA

Conventional

Ave. burr width (mm)

0.3 0.25 0.2 0.15 0.1 0.05

Fig. 23. Discontinuous chips resulted from UA drilling.

0 0

10

20

30

40

# of drilled holes

90

Fig. 21. Average burr width vs. number of drilled holes for TiNcoated drills.

80

force on the workpiece, resulting in less plastic deformation and smaller burrs. Figs. 24 and 25 show the average thrust force for uncoated and TiN-coated HSS drills respectively. An average of 20% reduction in thrust force by using UA drilling was observed. Fig. 26 shows the flank wear progression at the outer edge of the cutting lips of four drills. UA drilling increases tool wear significantly because of fatigue (an average of 35% comparing with conventional drilling). Although the tool wear on TiN-coated drills used in UA drilling is larger than that used in conventional drilling, TiN coating provides wear resistance, and reduces tool wear caused by UIA by approximately 30% comparing with uncoated drills, and allows the practical uses of UA drilling without significant reduction in tool life. It

Ave. thrust force (N)

70 60 50 40 30 20 10

UA

Conventional

20

30

0 0

10

40

# of drilled holes Fig. 24. Average thrust force vs. number of drilled holes for uncoated drills.

ARTICLE IN PRESS S.S.F. Chang, G.M. Bone / Robotics and Computer-Integrated Manufacturing 21 (2005) 442–450

450 90 80

Ave. thrust force (N)

70 60 50 40 30 20 10

UA

Conventional

0 0

10

20

30

40

# of drilled holes Fig. 25. Average thrust force vs. number of drilled holes for TiNcoated drills.

0.35

0.3

Wear width (mm)

0.25

0.2

0.15

0.1

UA uncoated

produced. It is believed that when the vibration frequency is high enough for a given material and cutting condition, UIA becomes significant and causes chip segmentation. This results in discontinuous chips, reducing the thrust force and burr size. Insignificant UIA occurs when the vibration frequency is too small, producing long and partially segmented chips, which increase thrust force and burr size. When pk–pk vibration was increased, the burr height decreased but the burr width increased. It is believed that the UIA dominates when the pk–pk vibration increases, increasing the force normal to the rake face. This force causes the material to begin to rollover earlier, increasing the burr width. However, the deformed material is thicker and stiffer, causing it to maintain a stronger contact with the cutting lips, hence reducing the burr height. There exists a better vibration condition for each particular cutting condition (spindle speed and feedrate), where burr sizes can be reduced effectively. Carelessly chosen ultrasonic assistance can produce larger burrs. UA drilling also allows a higher spindle speed and feedrate to be used without increasing the burr size. UA drilling introduces challenges in the context of tool strength and tool life. Suitable drill material and coating must be used. TiN-coated HSS drill offers resistance to the wear caused by fatigue in UA drilling, and allows the practical use of UA drilling without significantly reducing tool life. Moreover, regardless of the increases in tool wear, UA drilling can consistently produce smaller burrs than conventional drilling.

References

Conventional uncoated 0.05

UA TiN coated Conventional TiN coated

0 0

10

20

30

40

# of drilled holes Fig. 26. Flank wear progression of the tools.

should also be noted that although tool wear is increased, UA drilling produces smaller burrs than conventional drilling even with worn tools.

4. Conclusions We have reported and analyzed the results of 175 UA and conventional drilling experiments. When the vibration frequency in UA drilling was above a certain threshold, burrs with smaller height and width were

[1] Bone GM, Elbestawi MA, Lingarkar R, Liu L. Force control for robotic deburring, ASME J. Dyn. Syst. Meas. Control 1991; 113:395–400. [2] Dornfeld DA, Ko SL. A study on burr formation mechanism, Transactions of the ASME. J. Eng. Mater. Technol 1991;113: 75–87. [3] Lin TR, Shyu RF. Improvement of tool life and exit burr using variable feeds when drilling stainless steel with coated drills. Int. J. Adv. Manuf. Technol 2000;16:308–13. [4] Takeyama H, Kato S. Burrless drilling by means of ultrasonic drilling. Ann CIRP 1991;40:83–6. [5] Kim J, Min S, Dornfeld DA. Optimization and control of drilling burr formation of AISI 304L and AISI 4118 based on drilling burr control charts. Int. J. Mach. Tools Manuf 2001;41:923–36. [6] Dornfeld DA, Guo Y. Finite element analysis of drilling burr minimization with a backup material. Trans. NAMRI/SME 1998;XXVI:207–12. [7] Zhang LB, Wang LJ, Liu XY, Zhao HW, Wang X, Luo HY. Mechanical model for predicting thrust and torque in vibration drilling fibre-reinforced composite materials. Int. J. Mach. Tools Manuf 2001;41:641–57. [8] Kim JD, Nam SR. Development of a micro-depth control system for an ultra-precision lathe using a piezo-electric actuator. Int. J. Mach. Tools Manuf 1997;37(4):495–509.