Available online at www.sciencedirect.

com

Procedia Engineering
00 (2012)19
000000
Procedia Engineering
(2011) 3 8

Procedia
Engineering
www.elsevier.com/locate/procedia

1st CIRP Conference on Surface Integrity (CSI)

Productivity and workpiece surface integrity when WEDM

aerospace alloys using coated wires
M.T. Antara, S.L. Soo a*, D.K. Aspinwalla, D. Jonesb, R. Perezc
a

Machining Research Group, School of Mechanical Engineering, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK
b
Rotatives, Rolls-Royce plc., Bristol, UK
c
GF Agie Charmilles, Rue du Pr-de-la-Fontaine 8-10, CH-1217 Meyrin 1, Geneva, Switzerland

Abstract
Following a brief review of recent minimum damage EDM pulse generator developments, experimental data is
presented for workpiece productivity & integrity when WEDM Udimet 720 nickel based superalloy and Ti-6Al-2Sn4Zr-6Mo titanium alloy, using Cu core coated wires (ZnCu50 and Zn rich brass). Up to a 70% increase in
productivity was possible compared to when using uncoated brass wires with the same operating parameters.
Surfaces measuring ~ 0.6m Ra, with near neutral residual stresses and almost zero recast were produced following
two trim passes. Cross-sectional micrographs of specimens following rough machining (~180mm2/min) showed the
recast to be < 7m thick (up to 11m for uncoated wire), with comparable results for both alloys. Surface cracking,
when evident, was restricted to within the recast layer.

2012 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.

Selection and peer-review under responsibility of Prof. E. Brinksmeier
Keywords: WEDM minimum damage, Surface roughness, Residual stress

1. Introduction
Received wisdom suggests that in terms of material removal rate (MRR), EDM is a slower process
compared to conventional machining processes and results in significant damage to the machined
workpiece surface / subsurface, producing a recast layer, heat affected zone (HAZ), cracks and tensile
residual stresses etc. Over the last 10-15 years however, developments in EDM technology in general and
WEDM in particular, have been realised which contradict this picture. This has mainly been due to
improved generators employing MOSFET devices with ultra fast recovery diodes able to operate at ultra
high frequency (MHz) with pulse durations of 50-1000ns and full control of pulse shape, bipolar pulses /

* Corresponding author. Tel.: +44-121-414-44196; fax: +44-121-414-4201.

E-mail address: s.l.soo@bham.ac.uk.

1877-7058 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
doi:10.1016/j.proeng.2011.11.071

4 2

M.T. et
Antar
et al. / Procedia
Engineering
19 (2011)
38
M.T. Antar
al./ Procedia
Engineering
00 (2012)
000000

anti-electrolysis systems, the use of coated wires, automatic threading / rethreading, enhanced process
control / monitoring and improvements in reliability [1-3]. Currently, cutting rates for WEDM roughing
operations on straight cuts are quoted up to 500mm2/min (when using 330m diameter coated wires).
With optimised roughing and trim pass strategies, very fine surface roughness (~0.5m Ra) with near
zero recast layers, almost neutral residual stress and negligible changes to workpiece microstructure are
possible [4].
Typically, wire electrodes used in WEDM should ideally have high electrical conductivity to minimise
energy loss over the distance from power feed contact to sparking gap, as well as sufficient tensile
strength to withstand the necessary tension and maintain good dimensional accuracy. In addition, a low
melting temperature would also assist in shortening dielectric ionisation time and improving spark
formation. Due to its excellent electrical conductivity and use in die sinking, copper was the first
electrode material employed in WEDM, however its low strength and poor tensioning-ability which
produced machining instability, limited its scope and application. In contrast, brass (Cu63%Zn37%) wire
provided an excellent combination of low cost, good electrical conductivity and high tensile strength.
Since the early 1970s, brass has become the most commonly used electrode material for WEDM [5-6].
Increasing the zinc content in brass wire to provide enhanced gap conductivity generally causes drawing /
fabrication problems. The solution therefore was to produce zinc coated wires, where the low melting
point zinc particles vaporise quickly following spark ignition and contribute to increased gap conductivity
and size and thereby better flushing. Typically the thickness of the coating layers is ~20-30m and varies
depending on the wire diameter. Additionally the boiling off of zinc helps in cooling the wire core
which reduces the chances of wire breakage, thus allowing use of more intensive pulse power and a 40100% increase in cutting speed [7]. The following work was undertaken to evaluate the impact of utilising
coated wires on productivity and workpiece surface / subsurface integrity, including surface roughness,
recast layer thickness, heat affected zone formation and residual stresses, with comparison to standard
uncoated brass wire.
2. Experimental work
Workpiece materials were Ti-6Al-2Sn-4Zr-6Mo (hereafter called Ti6246) annealed to provide a bulk
hardness of ~HV390, and Udimet 720 nickel based superalloy solution treated and aged to ~HV490.
Machining was carried out on an Agie-Charmilles Robofil 240cc five axis wire machine employing
Clean Cut generator technology. During machining the de-ionised water dielectric conductivity was
maintained at ~5S/cm. Two types of coated wires; Bedra Broncocut Type X, copper core with CuZn50
coating (coated wire A) and Bedra Topas Plus X, copper core with a double layer zinc rich (60% wt)
outer coating (coated wire B) were employed for the tests. In addition, extra specimens were machined
using a standard uncoated brass wire for comparison purposes. Workpiece surface roughness was
measured using a Taylor Hobson Form Talysurf 120L with a 0.8mm cut-off length and 4mm evaluation
length in accordance with standard practice. Six measurements were taken for each surface and the results
averaged. Selected workpiece samples were sectioned and subsequently set in bakelite using a Buehler
mounting press and then ground and polished. Samples of Ti6246 were etched using Krolls reagent (2ml
HF, 10ml HNO2 and 88ml water) while Kallings No.2 etch (40-80ml ethanol, 40ml HCl and 2g CuCl2)
was used for the Udimet 720 samples. Recast layer evaluation and microstructural analysis were
conducted on a JEOL 6060 scanning electron microscope (SEM). Surface residual stress measurements
were performed using a Proto iXRD machine based at the University of Manchester. Residual stress
levels were calculated using the 2-sin2 method via computer software integrated with the
diffractometer. Furthermore, measurements were taken only in the transverse direction as little or no

35

M.T. Antar
Antar et
et al./
al. /Procedia
ProcediaEngineering
Engineering00
19(2012)
(2011)000000
38
M.T.

variation in the longitudinal direction was recorded. Test samples were wire machined using a sequence
of a roughing pass (R) followed by two finishing / trim cuts (T1 & T2) to provide 3 different surfaces on
each workpiece specimen. Udimet 720 test coupons were cut from cast blocks of 30mm height, while the
Ti6246 samples were machined from a scrap compressor disc of the same height. Generator operating
parameters were selected to provide specific surface roughness values based on the workpiece material
and the number and sequence of finishing passes employed, and to comply with aeroengine requirements
(0.8m Ra for finish machined parts). Table 1 shows the main operating parameters used for machining
both the Ti6246 and Udimet 720 specimens. No replications were performed due to workpiece cost and
time constraints.
Table 1. Main operating parameters
Parameter

T1

T2

Voltage (V)

80

200

120

Ignition current (A)

10

On-time (s) *

0.6

0.05

0.4

Off-time (s)

5.4

0.6

Flushing pressure (bar)

16

Wire speed (m/min)

12

10

10

Wire tension (daN)

1.2

2.0

1.8

Wire offset (mm)

0.160

0.144

0.135

*On-time value for T2 represents a cluster of pulses where the voltage signal is changed every 0.2s

3. Results and discussion

Table 2 details the cutting rates (data normalised due to commercial considerations) achieved when
using coated and uncoated wires for both workpiece materials. An increase in productivity of ~40% for
Udimet 720 and ~70% for Ti6246 during roughing was recorded when utilising coated wires compared to
the uncoated brass wire. This was in agreement with Dauw and Albert [7] who suggested that the use of
coated wires could increase cutting speed by 40 to 100% depending on the workpiece thickness.
Table 2. Cutting rates for roughing and finishing passes
W.P material

Wire

R (mm2/min)

T1 (mm2/min)

T2 (mm2/min)

Udimet 720

Coated A

101

102

122

Coated B

104

95

122

Uncoated brass

71

80

122

Coated A

129

109

122

Coated B

131

105

122

Uncoated brass

79

105

122

Ti6246

The enhanced productivity is primarily attributed to the increased conductivity within the sparking gap
caused by rapid melting / vaporisation of the zinc particles and also to the improved flushing conditions
resulting from the enlarged gap width. The servo speed during T2 was fixed / controlled by the generator

6 4

M.T. et
Antar
et al. / Procedia
Engineering
19 (2011)
38
M.T. Antar
al./ Procedia
Engineering
00 (2012)
000000

(to obtain the required surface roughness / quality) and hence the cutting rate was constant for the
different wires / materials. Similar surface roughness values (as detailed in Fig 1) were measured for
specimens machined using both coated and uncoated wires and also for the two workpiece materials. The
variation in surface roughness (represented by error bars) were due to imperfect sparks cause by wire
erosion however, the Ra values achieved after T2 (~0.6m) is acceptable for the majority of aeroengine
components. Sub-surface integrity evaluation of test samples highlighted up to 25% thinner recast layers
on Udimet 720 and ~40% on Ti6246 when rough machining using coated wires against uncoated brass
wire. Similar trends were also observed following finishing operations, see Table 3. This may have been
due to improved flushing in the gap with more molten material being carried away after sparking.

Fig. 1. Ra values for (a) Udimet 720 and (b) Ti6246

Fig 2 details sample cross-sectional SEM micrographs for both materials after roughing and finishing,
where almost no recast layer was detected following T2 operation (particularly with Udimet 720
specimens). In all evaluated samples microcracks, when evident, terminated within the recast layer and
did not penetrate into the bulk material. No microstructural alterations or thermal banding effects were
observed in the subsurface of any of the machined specimens. In addition, more microcracks were seen
within the Ti6246 recast layer than Udimet 720, probably due to the lower thermal conductivity of the
former, as microcrack intensity decreases with increasing thermal conductivity [8] (where the material
has more ability of dissipating heat from the heating zone).
Table 3. Average recast layer thickness
Workpiece material
Udimet 720

Ti6246

Wire type

Average recast layer thickness (m)

R

T1

T2

Coated A

5.5

2.5

~0

Coated B

6.8

3.0

~0

Uncoated brass

7.6

3.5

~0

Coated A

6.7

1.9

~ 1.0

Coated B

6.6

2.5

~ 1.0

Uncoated brass

11.4

3.4

~2.0

M.T. Antar
Antar et
et al./
al. /Procedia
ProcediaEngineering
Engineering00
19(2012)
(2011)000000
38
M.T.

Fig. 2. Cross-sectional micrographs for (a) Udimet 720 R, (b) Udimet 720 T2, (c) Ti6246 R and (d) Ti6246 T2

Fig 3 details surface residual stress measurements for the Udimet 720 and Ti6246 specimens machined
using coated and uncoated wires. The workpiece specimens were fully annealed prior to EDM testing in
order to eliminate any existing stress fields within the material. Therefore the residual stress values
detailed were due entirely to the EDM operation. No measurable peaks however were detected for the
Ti6246 roughing surfaces and therefore no stress readings were obtained. It is likely that the presence of
surface microcracks in the roughing samples for Udimet 720 led to a degree of stress relieving [9],
compared to the results for T1 where significantly higher tensile levels were measured. Some differences
in the results were apparent for T1 and T2 however whether this was statistically significant could not be
assessed due to cost considerations associated with multiple measurements. Also, the slightly
compressive residual stresses recorded in specimens machined using uncoated brass wire following T2,
could be due to inherent / pre-induced stresses in the material, since wire machining does not involve
physical contact between the wire and workpiece and any mechanical / plastic deformation. A best case
scenario would therefore be a neutral stress regime.

Fig. 3. Surface residual stresses for (a) Udimet 720 and (b) Ti6246

57

8 6

M.T. et
Antar
et al. / Procedia
Engineering
19 (2011)
38
M.T. Antar
al./ Procedia
Engineering
00 (2012)
000000

4. Conclusions
An increase in productivity of ~40% for Udimet 720 and ~70% for Ti6246 was possible when
replacing standard uncoated brass wire with diffusion annealed coated wires under the same operating
parameters. This could be further raised to 90% for Ti6246 and 60% for Udimet 720 by increasing
pulse on-time while still maintaining an acceptable level of surface roughness.
The variation in surface roughness values (represented by the error bars), mainly following roughing,
between the top, middle and bottom sections of the specimens could be related to imperfect sparks
generated due to erosion of the wire, leading to wider machining gaps.
In terms of recast layer thickness, better results were achieved using the coated wire for both roughing
and trim 1 operations (~25% thinner recast for Udimet 720 and ~ 40% for Ti6246), as a result of
improved cooling and flushing conditions.
A dramatic drop from ~700MPa (Udimet 720) and ~150MPa (Ti6246) tensile stresses (following T1)
to near neutral levels was seen after a T2 finishing pass.
Acknowledgements
We are indebted to the School of Mechanical Engineering, University of Birmingham and
Universities UK for the provision of an Overseas Research Student Award Scheme (ORSAS) scholarship
and to Rolls-Royce Plc. and GF Agie Charmilles for additional financial and technical support. Thanks
are also due to Dr. Andrew Winn, University of Manchester for undertaking residual stress measurements
and Dr. Tobias Noethe from Berkenhoff GmbH for supply of the coated wires.
References
[1]

Han F, Jiang J, Yu D. Influence of discharge current on machined surfaces by thermo-analysis in finish cut of WEDM,
Int J Mach Tool Manuf 2007; 47(7-8):1187-1196.

[2]

Aspinwall DK, Soo SL, Berrisford AE, Walder G. Workpiece surface roughness and integrity after WEDM of Ti-6Al-4V

[3]

Yan MT, Lai YP. Surface quality improvement of wire-EDM using a fine-finish power supply. Int J Mach Tool Manuf

and Inconel 718 using minimum damage generator technology, CIRP Ann 2008; 57/1:187-190
2007; 47(11):1686-1694
[4]

Antar MT, Soo SL, Aspinwall DK, Cuttell M, Perez M, Winn AJ. WEDM of Aerospace Alloys Using Clean Cut
Generator Technology. In: Proceedings of the 16th International Symposium on Electromachining (ISEMXVI); April 1923, 2010, Shanghai, China, p. 285-290.

[5]

Kapoor J, Singh S, Khamba JS. Recent developments in wire electrodes for high performance WEDM. In: Proceedings

[6]

Prohaszka J, Mamalis AG, Vaxevanidis NM. The effect of electrode material on machinability in wire electro-discharge

[7]

Dauw DF, Albert L. About the Evolution of Wire Tool Performance in Wire EDM. CIRP Ann 1992; 41(1):221-225.

[8]

Lee HT, Tai TY. Relationship between EDM parameters and surface crack formation, J Mater Process Technol 2003;

[9]

142(3):676-683.
Kruth J, Bleys P. Measuring residual stress caused by wire EDM of tool steel. Int J Elect Mach 2000; 5:23-28.

of the World Congress on Engineering (WCE2010); June 30-July 2, 2010, London, UK, p.1065-1068.
machining, J Mater Process Technol 1997; 69(1-3):233-237.