Vous êtes sur la page 1sur 9

Precision Engineering 37 (2013) 399407

Contents lists available at SciVerse ScienceDirect

Precision Engineering
journal homepage: www.elsevier.com/locate/precision

Prole error compensation in high precision 3D micro-EDM milling


Minh Dang Nguyen , Yoke San Wong, Mustazur Rahman
National University of Singapore, Department of Mechanical Engineering, 9 Engineering Drive 1, 117576, Singapore

a r t i c l e i n f o a b s t r a c t

Article history: Together with the product miniaturization trend, manufacturing technologies to fabricate small size
Received 18 June 2012 shapes are highly demanded for micro-mold and die industry. Among the versatile machining processes,
Received in revised form micro-EDM milling is superior owing to its negligible cutting force. By using layer by layer removal
20 September 2012
strategy in association with CAM software for tool path generation, intricate 3D micro-shapes could be
Accepted 3 November 2012
produced. However, there exist some systematic errors which affect the dimensional accuracy and the
Available online 16 November 2012
nal form of obtained shapes. This study attempts to identify and analyze the error components of 3D
micro-EDM milling process. It is found that aside from the inherent machining gap and the indispensible
Keywords:
3D micro-EDM milling
electrode wear, the corner radius of virtual electrode is also of prime importance in determining the
Compensation machining accuracy. In this study, geometric models are presented to simulate its effects. It is detected
Corner radius that the prole error could be reduced by implementing the corner radius into virtual electrode model.
Electrode wear For verication, typical 3D micro-shapes formed by the inclined planes and a partial sphere are fabricated
Machining gap with and without the new model for virtual electrode geometry.
Virtual electrode geometry 2012 Elsevier Inc. All rights reserved.

1. Introduction die. Due to tool wear issue, the electrode needs to be replaced
with a fresh one after a certain usage time [8]. However, when the
With the high demand of smaller equipments, the product electrode size falls within micro-scale, continuous substitution and
miniaturization trend is indispensible. In order to manufacture alignment of tiny electrode are problematic. In order to overcome
parts with feature size smaller than few hundred microns for micro- this issue, micro-EDM and micro-milling have been combined in
mold and die industry, innovations of manufacturing processes are a process called contour micro-EDM milling [9]. In this technique,
required [1]. Therefore, conventional metal cutting processes have an electrode in simple form moves along the preset tool path to
been developed for micro-scale applications. Among these machin- remove the materials like conventional milling. Nevertheless, the
ing processes, micro-milling is a highly feasible method to fabricate removal mechanism is still based on the discharges, not the shear-
intricate micro-molds and dies. In association with computer-aided ing of material. Although this approach is effective in reducing the
manufacturing (CAM) systems, freeform micro-mold could be pro- electrode replacement routines, the electrode wear still occurs.
duced [2,3]. However, one drawback of micro-milling process is This inherent characteristic changes the electrode shape during
the occurrence cutting forces during machining. These forces may machining, affecting the dimensional accuracy and the nal form
lead to tool deection which affects the accuracy of machined parts. of fabricated workpiece [8].
Especially, when the milling tool diameter or the feature size is less In that scenario, the uniform wear method, a prominent idea,
than few hundred microns, micro-milling seems to be ineffective has been proposed to resolve the tool wear issue in micro-EDM
because there is high probability of tool breakage and permanent milling [10]. Electrode wear in micro-EDM has been classied into
workpiece deformation [46]. two types: the corner wear and the bottom wear. The corner wear
Micro-EDM is a superior process in machining micro-features changes the electrode shape whereas the bottom wear reduces its
owing to the negligible cutting force yielded by its non-contact length. Because the wear phenomenon is inescapable, this uniform
material removal mechanism [7]. For that reason, the electrode wear method attempts to eliminate the corner wear and thus main-
as small as a couple tens of microns in diameter could be tain the original shape of electrode. Consequently, the electrode
used. Notwithstanding this remarkable advantage, micro-EDM die- wear is merely the bottom wear thereby it could be completely
sinking alone is impractical in manufacturing complex mold and compensated in the longitudinal direction. To fulll this require-
ment, the material needs to be removed layer by layer and the
tool path of micro-EDM milling must comply with certain motion
Corresponding author. Tel.: +65 16 4644. rules. Furthermore, in order to fabricate intricate 3D shapes for
E-mail addresses: dang@nus.edu.sg, nmdang85@yahoo.com (M.D. Nguyen), micro-mold, CAM systems must be applied to automatically gener-
mpewys@nus.edu.sg (Y.S. Wong), mpemusta@nus.edu.sg (M. Rahman). ate the motion code for machine tool. However, most current CAM

0141-6359/$ see front matter 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.precisioneng.2012.11.002
400 M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407

software is mainly used for conventional milling and it cannot be


directly applied for micro-EDM milling due to electrode wear issue
[11]. Therefore, there have been few attempts to develop the CAM
system in-house for 3D micro-EDM milling. The results obtained
from these reports show that intricate 3D micro-shapes could be
successfully produced by 3D micro-EDM milling [1113].
In the light of the reported studies, it could be seen that
micro-EDM milling is capable of fabricating different complex 3D
micro-shapes and features for micro-mold. However, the dimen-
sional accuracy of this versatile process is also of prime importance.
During machining, certain errors exist, affecting the dimension and
the nal form of obtained shape. Hence, this study attempts to
identify some error components and apply the appropriate com-
pensation to obtain micro-shapes with higher precision by 3D Fig. 2. Tool path offset in micro-EDM milling.
micro-EDM milling process.

gap. Although this ne gap is as small as few microns in micro-


2. Compensation of error components EDM, it is essential to take this issue into account for achieving
higher dimensional accuracy. Therefore, aside from the radius of
In 3D micro-EDM milling process, the original shape of elec- electrode, the offset distance in micro-EDM milling also involves
trode need to be maintained during the entire machining process the machining gap, as interpreted in Fig. 2.
to effectively compensate the electrode wear. As aforementioned, On the other hand, the machining gap exists not only on the lat-
the bottom wear makes the electrode length shorter and the cor- eral surface of the electrode but also on its bottom side. In order to
ner wear changes its shape. Because of the electro-erosion nature, resolve this machining gap issue, the concept of virtual electrode
electrode wear is indispensible. Therefore, the unique way to keep is proposed. This virtual electrode is dened to be larger than the
the tool shape unmodied is to entirely eliminate the corner wear. real electrode by the machining gap in all sides, as illustrated in
From this point of view, the discharges are intentionally held to Fig. 3. Consequently, the diameter of the virtual electrode will be
occur merely on the bottom side of the electrode. Consequently, bigger than the real electrode by two times of machining gap and
only electrode length is shorter and the tool shape is preserved its length is longer than that of the real electrode by the machining
whereby the electrode wear could be effectively compensated in gap. By considering that the virtual electrode is used as the cutting
longitudinal direction. In order to fulll this requirement, layer tool instead of the real electrode, the process is considered to be
by layer material removal strategy is exploited and thus the layer similar to conventional micro-milling. Therefore, the electrode off-
depth is of prime importance. The depth of each layer must be set could be automatically calculated and thus the machining gap is
smaller than a critical value which is the machining gap, as illus- compensated properly. This method is also exible in case of using
trated in Fig. 1a. In Fig. 1b where the layer depth is higher than different discharge pulse energies. The machining gap varies when
the machining gap, the sparks occur not only on the bottom side the discharge pulse energy is changed. In that case, the correlative
but also on the lateral surface of tool electrode, resulting in the machining gap could be accordingly compensated just by selecting
alteration of both electrode length and shape. In addition, high new dimension of the virtual electrode.
layer depth induces poorer atness of machined surface. Hence,
the layer depth is usually kept much smaller than the machining
2.2. Electrode wear compensation
gap.

In conventional CAM software which mainly caters for conven-


2.1. Machining gap compensation tional milling, tool length and tool shape are treated unchanged
during the entire machining process. Thus, electrode wear com-
In contrast to conventional micro-milling, micro-EDM is a non- pensation needs to be added in the tool path generation system
contact machining process so there is a short distance between for micro-EDM milling. In this system, the electrode wear would
the workpiece surface and the tool electrode. Hence, the gener- be compensated during post-processing step. Prior to performing
ated prole is apart from the electrode surface by the machining that step, the electrode wear data for compensation must be deter-
mined. The longitudinal wear of each layer would be calculated
based on the layer volume and the volumetric electrode wear ratio.

Fig. 1. Illustration of electrode wear in micro-EDM milling: (a) only bottom wear
(d < machining gap), (b) both bottom wear and corner wear (d > machining gap) Fig. 3. Illustration of 3D view and cross-sectional view of the virtual electrode (g is
where d is the depth of each layer. the machining gap).
M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407 401

Firstly, the 3D CAD model of removed volume is built and split 3. Experimental details
into a series of layer. Then, the volume of different layers could
be obtained. Accordingly, the longitudinal wear after machining a 3.1. Developed tool path generation system
certain layer i could be calculated using the following equation:
As aforementioned, the current CAM software could not be
Vi
li =  directly applied for micro-EDM milling. Hence, the machining code
Ae
for performing 3D micro-EDM milling is obtained using the tool
where is the volumetric wear ratio between the workpiece and path generation system shown in Fig. 4. It is in-house developed
the electrode, Vi is the volume of the layer i, Ae is the cross section based on the 3D CAD software SolidWorks 2011 and the ESPRIT
area of the electrode. CAM 2010 provided by DP Technology Corporation [14,15]. The
Secondly, after the preliminary machining program has been compensation of machining gap and the electrode wear in this
generated by the CAM software, the iteration is used to compen- system are highlighted in Fig. 4.
sate for the longitudinal wear after each layer by modifying the Z
coordinate of next layer. The new Z coordinate of layer i + 1 could
be determined as follows:
3.2. Sample micro-shapes
Zi+1 = Zi d li
For demonstrating the effect of the systematic errors, two typical
where d is the layer depth and li is the longitudinal wear after micro-shapes shown in Fig. 5 are chosen. Fig. 5a is the micro-dome
layer i. formed by a partial sphere whereas Fig. 5b exhibits the truncated

Fig. 4. Tool path generation system.


402 M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407

3.4. Observation and measurements

After machining, the scanning electron micro-graphs of fab-


ricated micro-shapes were captured using the JEOL JSM-5500
scanning electron microscope (SEM). In order to acquire the dimen-
sional data, the 3D proles of obtained shapes were collected using
the non-contact optical proler NANOVEA PS-50 equipped with
confocal imaging sensor CHR-150 [18].
Fig. 5. The 3D CAD models of sample micro-shapes: (a) micro-dome and (b) trun-
cated square micro-pyramid.
4. Results and discussions

4.1. Prole errors induced by the machining gap and the


longitudinal electrode wear

Fig. 7 shows the scanning electron micro-graphs of the obtained


micro-shapes. It can be seen that the obtained micro-cavities are
similar to the 3D CAD models of the desired micro-shapes in Fig. 5.
This indicates that the developed tool path generation is capable of
realizing the machining code for 3D micro-EDM milling.
Fig. 8 shows the measured 3D proles of the micro-dome fab-
ricated without the compensation for both machining gap and
Fig. 6. Scheme of experimental apparatus. electrode wear. It can be observed that the width of the machined
cavity is around 410 m, larger than the desired width by 10 m.
Table 1 This 10 m deviation is found to be two times of the 5 m machin-
Machining conditions. ing gap which was obtained in the previous study [19]. Hence, this
Machining uid DI water 0.5 M  cm is result of no compensation for machining gap during micro-EDM
milling process. On the other hand, it is also noted that actual depth
Voltage (V) 60
Resistor () 220
of machined micro-shape is approximately 35 m only, which is
Capacitor (pF) 265 less than the desired depth by 15 m. This stems from the electrode
Pulse frequency (kHz) 500 wear during machining. Although the layer by layer removal strat-
Duty cycle (%) 30 egy is effective in preserving the original shape of the electrode, the
Electrode rotation speed (rpm) 500
wear still occurs and reduces its length for a certain amount after
every layer. As a result, the actual tool tip coordinate is altered and
Table 2 it induces dimensional error. This error cumulates after each layer
Tool path parameters. and becomes as high as 15 m for the entire process. Hence, it is
Electrode diameter (m) 90 conrmed that the machining gap and the electrode wear are two
Machining gap (m) 5 systematic errors of micro-EDM milling process.
Layer depth (m) 0.5 For improving the dimensional accuracy of fabricated shapes,
Scanning feedrate (m/s) 50 the compensation for machining gap and longitudinal electrode
Tool path overlap (m) 10
wear is implemented during the tool path generation. The tool is
dened as the virtual electrode shown in Fig. 3. In addition, the
square micro-pyramid constructed by four planes inclined at 45 . new Z coordinate of the tool tip is also updated after removing each
Both the micro-cavities has the size of 0.4 mm 0.4 mm 0.05 mm. layer. With this approach, the machined micro-shapes are found to
have better prole accuracy. Fig. 9 shows the cross-sectional pro-
3.3. Machine set-up les of the obtained micro-dome and micro-pyramid. In both cases,
the width of the cavities is observed to be around 400 m and the
Micro-EDM milling process was conducted on the set-up shown actual depth is found to be nearly 50 m. This substantiates that the
in Fig. 6. The multi-process machine tool DT-110 manufactured by prole errors induced by machining gap and electrode wear could
Mikrotools was used [16]. Its resolution is 0.1 m and the repeat- be effectively compensated during tool path generation to improve
ability is 1 m for all axes. In this apparatus, deionized water was the dimensional accuracy of fabricated shapes.
used as the machining uid. However, the stray material disso-
lution during micro-EDM in deionized water could lead to the 4.2. Prole errors induced by the corner radius of virtual electrode
unanticipated material removal, affecting the dimensional accu-
racy. Hence, an in-house developed short pulse generator was In order to analyze the prole accuracy, Fig. 10 compares the
used to suppress the electrochemical reaction and thus ensure the measured cross-sectional proles of the fabricated shapes with
machining accuracy [17]. their ideal proles. It could be seen that the side and the bottom
Table 1 shows the main machining conditions of micro-EDM sections of micro-cavities have the relatively good t with the ideal
milling process whereas Table 2 exhibits the parameters den- prole. However, it is realized that the curved section of micro-
ing the tool path and removal strategy. Prior to machining, the dome are apart from the ideal prole by a small distance (Fig. 10a).
tungsten electrode was on-machine fabricated by horizontal block- A similar observation is also found in the slanting sections of micro-
electro-discharge-grinding to eliminate the eccentricity induced pyramid, as shown in Fig. 10b.
by clamping error. The bottom surface of the electrode was also Fig. 11 exhibits the closer view of these two remarkable sections
dressed to make it at. Then, the experimental works are carried with the additional superimposition of the theoretical proles.
out to fabricate two typical micro-shapes on stainless steel SUS 304 Because the material is removed layer by layer, the theoretical pro-
workpiece. le would be in the form of a stepped line. When the layer depth is
M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407 403

Fig. 7. SEM images of fabricated micro-shapes: (a) micro-dome and (b) micro-pyramid.

Fig. 8. Cross-sectional prole of the micro-dome fabricated by micro-EDM milling without the compensation for machining gap and electrode wear.

Fig. 9. Cross-sectional prole of fabricated micro-shapes with the compensation of both machining gap and electrode wear: (a) micro-dome and (b) micro-pyramid.

Fig. 10. Measured proles vs. ideal proles of fabricated micro-shapes without corner radius compensation: (a) micro-dome and (b) micro-pyramid.
404 M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407

Fig. 11. Comparison between measured proles, ideal proles and theoretical proles of fabricated micro-shapes: (a) micro-dome and (b) micro-pyramid.

the generated overlapping prole is separated from the theoretical


prole by a signicant gap. Moreover, it should be highlighted that
the separated distance is seen to be even for the slanted line and it
increases gradually for the curved section, which are in accordance
with the above-mentioned experimental results. Because the issue
stems from the corner radius, this prole error only appeared when
the intricate 3D shapes such as the inclined plane or curved surface
are machined. Hence, it reveals that the corner radius of virtual
electrode likely accounts for the prole error of these 3D complex
shapes.

Fig. 12. Illustration of virtual electrode geometry: (a) without corner radius (old
model) and (b) with corner radius (new model). 5. Geometric models to simulate the prole error induced
by the corner radius of virtual electrode

0.5 m, it can be seen that the theoretical proles and the ideal pro- Fig. 14 shows a geometrical model which manifests the prole
les have the close t. However, the measured prole is observed error in case of no compensation for corner radius of virtual elec-
to be apart from the theoretical by a signicant distance. It is also trode. The theoretical error ht is determined by the step height of
observed that the deviation is quite uniform for the inclined plane each layer:
while it gradually increases for the partial sphere. Hence, this devi-
ation is hypothesized to be attributed to the corner radius of virtual
ht = d cos  (1)
electrode. The corner of the virtual electrode is not perfectly sharp
and it should be rounded off with a llet radius due to the limited
distance of machining gap and the imperfect sharpness at the edge where d is the layer depth and  is the slope angle of the incline.
of real electrode, as illustrated in Fig. 12. When there exists the corner radius and it is not compensated,
Following this hypothesis, the prole generated by the over- the prole error will be:
lapped virtual electrode corner is further superimposed as plotted
in Fig. 13. It could be observed that when the corner radius exists, e = H + h (2)

Fig. 13. Illustration of proles generated by the electrode with virtual corner radius: (a) micro-dome and (b) micro-pyramid.
M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407 405

Fig. 16. The two situations of intersection points corresponding to different elec-
trode position.

Fig. 14. Prole error of slanting section without the compensation of virtual corner
Consequently, the theoretical prole error would be:
radius.

 d
2
where h is the scallop height formed by overlapped electrode prole h = r |zA | = r r2 (6)
and H is the deciency stemming from the corner radius: 2 sin 

H = r( 2 1) (3) It could be seen that the H is deducted in theoretical prole error
at which r is the corner radius of the virtual electrode. as compared to Eq. (2), which is the result of corner radius com-
From this, it can be seen that with the existence of the corner pensation.
radius, the theoretical error ht is no longer valid and it is superseded
by prole error e.
5.2. On a spherical surface
5.1. On an inclined plane
On a spherical surface, the intersection point of the overlapped
electrode prole depends on the location of the electrode. Again,
Fig. 15 shows the geometric model when the corner radius of
the layer depth is much smaller than the machining gap (d  r) to
virtual electrode is taken into account during tool path generation.
eliminate the corner wear, there are two situations determining
Instead of the sharp comer, the tool tip coordinate is now deter-
the intersection point as shown in Fig. 16. At the same layer depth
mined based on the contact of the ideal prole and the corner prole
d, when the x is larger than or equal the critical distance, the
of virtual electrode. Because the layer depth is much smaller than
intersection point will be dened by the dotted line (L) and the
the machining gap (d  r) to eliminate the corner wear, the theo-
circle (C2 ), as illustrated in Fig. 16a. On the contrary, when the x
retical prole error h would be the scallop height formed by the
is smaller than the critical distance, it is the intersection between
overlapped corner radius.
the circle (C3 ) and (C4 ), as shown in Fig. 16b.
Let the O1 XZ be the rectangular coordinate system as shown in
In order to determine the theoretical prole error for these two
Fig. 15. The coordinate of intersection point A of two circles (O1 )
circumstances, the geometric model shown in Fig. 17 is used.
and (O2 ) will be the result of a set of two equations: 2
2 2 2 Case 1, x = xO2 xO1 r 2 (d r)
x +z = r
  d 2 (4)
x + z2 = r 2
sin 
Solving for zA , it gives:

 d
2
zA = r2 (5)
2 sin 

Fig. 15. Prole errors of slanting section with the compensation for corner radius Fig. 17. Prole errors of curved section with the compensation of virtual corner
of virtual electrode. radius.
406 M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407

Let the O1 XZ be the rectangular coordinate system as shown in 6. Experimental verication of virtual electrode corner
Fig. 17. The coordinate of intersection point A will be the result of radius compensation
a set of two equations:
The above simulation shows that the prole error could be
z  = r reduced when the corner radius is taken into account in new virtual
(7)
2
(x x) + (z  + d) = r 2
2 electrode geometry. In order to validate this point, the same typi-
cal micro-shapes are fabricated with the implementation of corner
Then radius compensation during tool path generation. Based on the
 machining condition used, the modied virtual electrode is dened
(xA , zA ) = (x
2
r 2 (d r) , r) (8) to have 6 m corner radius which consists of 5 m machining gap
So and 1 m corner radius of the real electrode. The cross-sectional
proles of obtained micro-shapes are plotted in Fig. 18. Compared
(xA , zA ) = (xA + xO1 , zA + zO1 ) (9) with the proles in Fig. 10, the measured proles in Fig. 18 are
observed to be closer to the ideal proles. This afrms that the
where
corner radius of virtual electrode is also a key factor determining
(xO1 , zO1 ) = (q sin 1 , q cos 1 ) (10) the dimensional accuracy of fabricated shape. The prole errors
induced by this corner radius essentially appears when intricate
d = q cos 1 q cos 2 (11)
shapes such as the curved surface or the inclined plane are fabri-
The resulting height of the scallop is: cated.

With the compensation of the corner radius and the setting of
h= xA2 + zA2 R (12) 0.5 m layer depth, the theoretical prole error is determined to
be 0.011 m for slanted section (Eq. (6)) and it ranges from 0.006

Case 2, x = xO4 xO3 < r 2 (d r)
2 to 0.3 m for curved section (Eqs. (12) and (18)). Meanwhile, the
Let the OXZ be the rectangular coordinate system. The coordi- measured average prole error is found to be 1.55 m for machined
nate of intersection point B will be the result of a set of equations: inclined plane and 0.89 m for curved section. These deviations

are probably attributed to the inherent surface roughness formed
x =0 by overlapped discharge craters (as can be seen in Fig. 7), the ran-
(13) dom error caused by the detection of contact point between the
(x + q sin )2 + (z  q cos )2 = r 2 electrode and workpiece prior to machining, and the random error
where of measurements. However, it does show that the compensation
of the corner radius of virtual electrode is effective in reducing the
q=R+r (14) prole error. The average prole error e for curved section in Fig. 10a
is 1.49 m while it reduces to 0.89 m for Fig. 18a. A similar obser-
Then
 vation is found for slanted section. It is observed to decrease from
zB " = q cos r 2 (q sin )2 (15) 4.06 m in Fig. 10b to 1.55 m in Fig. 18b. It should be highlighted
that this reduction amount 2.51 m is consistent with the H in
where Eq. (3), which is determined to be 2.485 m.
1  d
 For the partial spherical shape, in order to isolate the random
= arccos cos 3 3 (16)
2 q error in detecting the contact point between the electrode and the
workpiece, the curved surface is intentionally designed to be lower
which is derived from
than the top surface by 5 m. By setting the top of spherical surface
4 3 = 2 as a common point, the measured proles of obtained micro-shapes
(17) are then superimposed in Fig. 19. It could be observed that the
d = q(cos 3 cos 4 ) obtained prole with the compensation of corner radius has better
The resulting height of the scallop is: t to the theoretical prole than that without the electrode corner
compensation. Hence, it is revealed that the compensation for the
h = zB " R (18)

Fig. 18. Measured proles vs. ideal proles of fabricated micro-shapes with corner radius compensation: (a) micro-dome and (b) micro-pyramid.
M.D. Nguyen et al. / Precision Engineering 37 (2013) 399407 407

Acknowledgements

The authors would like to express their sincere thanks to


Machine Tool Technologies Research Foundation (MTTRF) and DP
Technology Corporation for providing the ESPRIT CAM 2010.

References

[1] Uriarte L, Herrero A, Ivanov A, Oosterling H, Staemmler L, Tang PT, et al. Com-
parison between microfabrication technologies for metal tooling. Proceedings
of the Institution of Mechanical Engineers. Part C: Journal of Mechanical Engi-
neering Science 2006;220:166576.
[2] Krajnik P, Kopac J. Modern machining of die and mold tools. Journal of Materials
Processing Technology 2004;157158:54352.
[3] Lazoglu I, Manav C, Murtezaoglu Y. Tool path optimization for free form surface
machining. CIRP AnnalsManufacturing Technology 2009;58:1014.
[4] Masuzawa T. State of the art of micromachining. CIRP AnnalsManufacturing
Technology 2000;49:47388.
[5] Kim GM, Kim BH, Chu CN. Estimation of cutter deection and form error in ball-
end milling processes. International Journal of Machine Tools and Manufacture
2003;43:91724.
[6] Ratchev S, Liu S, Huang W, Becker AA. Milling error prediction and compensa-
tion in machining of low-rigidity parts. International Journal of Machine Tools
Fig. 19. Comparison of generated proles with and without the compensation for and Manufacture 2004;44:162941.
electrode corner radius. [7] Kunieda M, Lauwers B, Rajurkar KP, Schumacher BM. Advancing EDM through
fundamental insight into the process. CIRP AnnalsManufacturing Technology
2005;54:6487.
corner radius of virtual electrode is effective in reducing the prole [8] Tsai YY, Masuzawa T. An index to evaluate the wear resistance of the elec-
error of intricate 3D shapes. trode in micro-EDM. Journal of Materials Processing Technology 2004;149:
3049.
[9] Uhlmann E, Piltz S, Doll U. Machining of micro/miniature dies and moulds
7. Conclusions by electrical discharge machiningrecent development. Journal of Materials
Processing Technology 2005;167:48893.
In this study, some systematic errors of 3D micro-EDM milling [10] Yu ZY, Masuzawa T, Fujino M. Micro-EDM for three-dimensional cavities
development of Uniform Wear Method. CIRP AnnalsManufacturing Technol-
which induce the prole error of obtained shapes were analyzed
ogy 1998;47:16972.
and discussed. The following conclusions could be drawn: [11] Rajurkar KP, Yu ZY. 3D micro-EDM using CAD/CAM. CIRP
AnnalsManufacturing Technology 2000;49:12730.
Machining gap, electrode wear and corner radius of virtual [12] Zhao W, Yang Y, Wang Z, Zhang Y. A CAD/CAM system for micro-ED-milling
of small 3D freeform cavity. Journal of Materials Processing Technology
electrode are identied as three systematic errors of 3D micro- 2004;149:5738.
EDM milling which affect the dimensional accuracy of fabricated [13] Dimov S, Pham DT, Ivanov A, Popov K. Tool-path generation system
micro-shapes. for micro-electro discharge machining milling. Proceedings of the Institu-
tion of Mechanical Engineers. Part B: Journal of Engineering Manufacture
Prole errors induced by machining gap and electrode wear could 2003;217:16337.
be effectively reduced by adjusting the offset distance during [14] 3D CAD Design Software Solidworks. <http://www.solidworks.com/> [last
tool path generation and updating the new Z coordinate of the accessed 15.09.12].
[15] ESPRIT CAM. <http://www.dptechnology.com/>; 2010 [last accessed 15.09.12].
electrode tip after each layer. [16] Rahman M, Asad ABMA, Masaki T, Saleh T, Wong YS, Kumar AS. A multiprocess
Geometric models are presented to simulate the effect of corner machine tool for compound micromachining. International Journal of Machine
radius of virtual electrode on the dimensional accuracy. The pro- Tools and Manufacture 2010;50:34456.
[17] Nguyen MD, Rahman M, Wong YS. An experimental study on micro-EDM in
le error induced by this corner radius essentially appears when low-resistivity deionized water using short voltage pulses. The International
intricate shapes such as curved surfaces or inclined planes are Journal of Advanced Manufacturing Technology 2012;58:53344.
involved. [18] Prolometers 3D Non Contact Metrology. <http://www.nanovea.com/
By implementing the corner radius into new virtual electrode Prolometers.html> [last accessed 15.09.12].
[19] Nguyen MD, Rahman M, Wong YS. Enhanced surface integrity and dimensional
model, the prole error has been demonstrated to be further accuracy by simultaneous micro-ED/EC milling. CIRP AnnalsManufacturing
reduced. Technology 2012;61:1914.

Vous aimerez peut-être aussi