Introduction
Benchmarking for
comparative evaluation
of RP systems and
processes
M. Mahesh
Y.S. Wong
J.Y.H. Fuh and
H.T. Loh
The authors
M. Mahesh Y.S. Wong, J.Y.H. Fuh and H.T. Loh are all
based at the Department of Mechanical Engineering, Faculty
of Engineering, National University of Singapore, Singapore.
Keywords
Benchmarking, Advanced manufacturing technologies,
Standardization, Rapid prototypes
Abstract
A geometric benchmark part is proposed, designed and
fabricated for the performance evaluation of rapid
prototyping machines/processes. The benchmark part
incorporates key shapes and features of better-known
benchmark parts. It also includes new geometric features,
such as freeform surfaces, certain mechanical features and
pass-fail features that are increasingly required or expected
of RP processes/systems. The part is suitable for fabrication
on a typical RP machines. In this paper, the application of the
benchmark part is demonstrated using relatively common RP
processes. The ability of the benchmark part to determine
achievable geometric features and accuracy by the
aforementioned RP processes is presented and discussed.
Electronic access
The Emerald Research Register for this journal is
available at
www.emeraldinsight.com/researchregister
The current issue and full text archive of this journal is
available at
www.emeraldinsight.com/1355-2546.htm
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · pp. 123–135
q Emerald Group Publishing Limited · ISSN 1355-2546
DOI 10.1108/13552540410526999
123
Suitable benchmark parts can be designed for
performance evaluation of RP systems and
processes, and enable to obtain helpful data for
use in decision support systems. Besides the
process and the material, other factors (such as
the building style and specific process
parameters) may also affect the accuracy and
finish of the part. In RP benchmarking, it is
necessary not only to standardise the design of
the benchmark part, but also the fabrication and
measurement processes. This paper presents
issues on RP benchmarking and aims to identify
factors affecting the definition, fabrication,
measurements and analysis of benchmark parts.
Literature review
Earlier, Gargiulo (1992), Ippolito et al. (1994),
Ippolito et al. (1995), Juster and Childs (1994a,
b), Kruth (1991), Lart (1992), Reeves and
Cobb (1996), Shellabear (1999) and Xu (1999)
have reported on benchmarking of RP systems.
Table I shows a comparison of the features
between some of the benchmark parts to the
proposed benchmark part.
Proposed benchmark part design
The RP&M system performance evaluation is
assumed to be based on a benchmark part
consisting of three-dimensional (3D) part
features of different sizes at different locations,
and orientations. Figure 1 shows the details
of the proposed benchmark. The detailed
descriptions of the proposed benchmark part
and its comparison with other benchmarks
have been presented earlier by Wong et al.
(2002).
The geometric features of the benchmark
part shown in Figure 1 are identified by
two-letter names, such as CH, SB, SL, etc.
for referencing in the succeeding table and
results. The geometric features of the
benchmark part aim to test the capability of
an RP process or system. The features on
the proposed benchmark are summarised in
Table II.
Received: 26 October 2002
Revised: 10 October 2003
Accepted: 07 November 2003
 Table I Comparison of the benchmark parts
Gargiulo
Juster and Childs Ippolito et al. (1992) 3d
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh

Properties Kruth (1991) (1994a, b) Shellabear (1999) (1994) Lart (1992) systems Proposed part
Size Small Large Small Large Medium Large Medium
Benchmarking for comparative evaluation of RP systems and processes

Dimensions 100 £ 50 mm 250 £ 250 mm 71 £ 75 mm 240 £ 240 mm – 240 £ 240 mm 170 £ 170 mm
Features Simple: cylindrical shell, Comprehensive: features to Simple: planar Simple: used the Complex: rich in Simple: Comprehensive: typical

124
incorporated inclined cylinders, pegs test linear accuracy and surfaces, which 3d systems fine- and medium- features are geometric shapes, mechanical
and overhangs feature repeatability include various benchmark part sized features planar features incorporated
angles
Complexity of Simple, but not Simple Easy, but not Simple Difficult: features Simple Simple with CMM
measurement standardised standardised not accessible to programmed measurements
CMM
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · 123–135
 Benchmarking for comparative evaluation of RP systems and processes
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh
Figure 1 The proposed benchmark part
Benchmark part correlated with the
existing standards
Several of the typical geometric features
incorporated in the benchmark part are
designed for easy reference to the existing ISO
standards etc. for example, to measure
straightness (ISO 12780), roundness (ISO
12181), flatness (ISO 12781), cylindricity (ISO
12180), etc. EN ISO 10360 specifies standards
for the use of CMM in measurement. Use of the
benchmark part and standardised measurement
techniques can reduce variability in the
determination of geometric accuracy of the
various RP processes and systems.
Experimental study
An experimental study has been conducted to
demonstrate the use of the benchmark for the
performance evaluation of RP processes and
machines. The four well-known RP processes
stereolithography apparatus (SLA), selective
laser sintering (SLS), fused deposition
modelling (FDM) and laminated object
manufacturing (LOM) have been used in the
experimental study. The benchmark parts were
first fabricated with the default machine
parameters and settings. In fabricating the
benchmark part, effort was also made to attain
the best performance to identify a possible
process benchmark.
Fabrication of the benchmark part on the
SLA machine
A benchmark part of epoxy resin was fabricated
on the SLA-190/250 from a 3D system. The
125
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · 123–135
benchmark part (Figures 2 and 3) was first
visually inspected and then measured using a
CMM to determine the geometric accuracy of
the fabricated part. SLA is able to build parts
with a variety of resins that have different
working temperatures. The parameters were set
according to the type of the resin used. In the
case of Cibatool SL 5170, the layer thickness
was 0.15 mm, hatch type was “box”, hatch
spacing was 0.10 mm, etc. and also with specific
support and recoat parameters.
The thin cylinders and thin walls of the built
benchmark part were warped as seen in Figure 3.
Part shrinkage is a common problem in the SLA
process due to the resins used. The software
program has a compensation factor of 2 5 to 5
per cent for shrinkage in the X, Y, and Z
directions. The following post-processing
techniques have been found from experience to
affect the part accuracy.
(1) Prolonged exposure of the green part to air
before curing.
(2) Long solvent bath.
(3) Less/over curing in the ultraviolet (UV)
chamber.
It was observed that the base of the benchmark
part was not really bottom-flat and had some
undulations. It was later attributed to the lower
initial preheating (for example, the
recommended build temperature is 288C for SL
5170) of the resin before the actual building,
and possibly also to a drop in the laser power
(for example, from 30 to 29 mW) during the
building of the first few layers. Hence, it is
important to ensure that the resin reaches the
operating temperature before building the part.
Thus, it is important that the process is
optimally tuned to obtain the best attainable
performance of the process.
With regard to the geometric accuracy of the
features built, SLA was found to be the best
compared to the other processes. The square
bosses, solid cylinders, cones, cylindrical holes,
spheres, slots, circular holes and the pass-fail were
distinctly built. However, the hollow squares, thin
walls, slim cylinders were warped. Dimensions of
these features are listed in Table AI.
Fabrication of the benchmark part on the
SLS machine
Two parts were fabricated on the DTM
Sinterstation 2500. The material used was
 Table II Summary of the features and purposes
Features Purpose Number and nominal size
Square base (SB) Flatness and straightness (a base for the other features) 1 (170 £ 170 £ 5 mm) (size likely to fit build size of most machines)
Cube (CB) Flatness, straightness, linear accuracy, parallelism and repeatability 8 (15 £ 15 £ 15 mm)
Flat beam (FB) Overhang, straightness and flatness 1 (120 £ 15 £ 3 mm)
Cylindrical holes Accuracy, roundness, cylindricity and repeatability of radius 4 (10 mm diameter)
(z-direction) (CH)
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh

Cylindrical holes (CH) Accuracy, roundness and concentricity (X-direction) – 2 (10 mm diameter)
(Y-direction) – 2 (10 mm diameter)
Spheres (SP) Relative accuracy, symmetry and repeatability of a continuously changing 4 (15 mm diameter)
Benchmarking for comparative evaluation of RP systems and processes

sloping surface

126
Solid cylinders (SC)
Hollow cylinders (HC)
Cones (CN)
Slots (SL)
Hollow squares (HS)
Brackets (BR)
Circular holes (CR)
Mechanical features
Pass-fail features
Roundness cylindricity and repeatability
Accuracy, roundness, cylindricity and repeatability of radius
Sloping profile and taper
Accuracy of slots, straightness and flatness
Straightness and linear accuracy, also thin wall built
Linear accuracy, straightness and angle built
Cylindricity, relative position, roundness and repeatability
Efficiency of machine to build special features
Ability of machine to build certain features
4 (5 mm diameter)
2 (10 mm diameter)
2 (outer diameter 10 mm; and inner diameter 5 mm)
2 (base diameter 10 mm)
11 (length 10 mm, height 5 mm, varying width) arranged in two rows
2 (first: 25 £ 25 £ 15 mm, wall thickness: 2 mm; second: 60 £ 60 £ 15 mm, wall
thickness: 1 mm)
4 (length 10 mm, height 15 mm, width 10 mm, and thickness 1 mm)
5 (15 mm diameter 1; and 10 mm diameter 4)
Fillet (FL), chamfer (CF), blending (BL), free-form (FF) features
Thin walls (TW), thin slots (TS), slim cylinders (SC), small holes (SH)
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · 123–135
 Benchmarking for comparative evaluation of RP systems and processes
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh
Figure 2 Benchmark part fabricated on SLA 250
Figure 3 Ability of SLA to build all features including pass-fail features
ProtoForm Composite (LNC- 7000), which
was a mixture of recycled and new ProtoForm
powder in the ratio of 1:3. The first part was
built with the default parameter setting of the
machine. The X scale factor was 1.03, Y scale
factor was 1.03 and Z scale factor was 1.015.
The fill laser power was about 8 W. It was found
that the part was warped. In addition, the thin
cylinders, 0.5 mm holes, 0.5 mm slots, and thin
wall could not be completely built. This was
attributed to the laser power and parameter
setting. The benchmark part was built again
with a new set of parameters and the laser power
was raised slightly to about 9 W. The new set of
parameters was obtained after a number of trial-
and-error experiments. It is noteworthy that the
thin walls and thin cylinders were better built
when no scaling factor was applied (or set at
1.00). The reason could possibly be due to the
127
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · 123–135
mixture of the recycled and the new ProtoForm
powder used in the experiment.
In the second case, most of the features of the
part could be built. But there were problems to
separate the specific geometric features from the
trapped powder. The unsintered powder had to
be carefully brushed off from the fine features,
such as from the 0.5 mm-diameter cylinder, as
the increase in the part bed temperature caused
the unsintered powder to adhere together and to
the built parts.
Effects of warpage on the surface accuracy
Warpage has an adverse effect on the geometric
accuracy of the fabricated parts. It could be in a
form due to upwards curling of the part at both
edges (Figure 4) or only one edge, such as that
shown in Figure 5. In the first part, fabricated
using the default parameter setting, the warpage
was observed on one side of the part, similar to
the one in Figure 5. It was distorted from the
centre towards the corner. Figures 6 and 7 show
that, the shape and size of the square boss
towards the corner have been affected due to the
warpage. The second part was fabricated with
Figure 4 Part warpage drifting towards the edges
Figure 5 Part warpage drifting towards a side
Figure 6 Side showing the distinct warpage
 Benchmarking for comparative evaluation of RP systems and processes
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh
Figure 7 Effect of warpage on the features built
a different orientation and was placed more
towards the side near the heater, rather than
being centred on the build platform. The laser
power was also increased by a watt (from 8 to
9 W). The part bed temperature and the powder
bed temperature were 192 and 1018C,
respectively. The warpage of the fabricated part
was less when compared to the earlier case, as
seen in Figures 8 and 9.
Hence, the benchmark part can be
appropriately used to optimise the machine.
This involves initial trial experiments to
fabricate the benchmark part and fine-tune the
parameters till the best attainable features can
be built. This typically is a trial-and-error
method that needs to be repeated with different
types of materials, and to a certain extent, relies
on the experience of the operator.
Figure 8 New benchmark part showing a better built and
reduced warpage
128
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · 123–135
Figure 9 Part showing the features built including pass-fail features
Fabrication of the benchmark part on the
LOM machine
The benchmark part was fabricated on the
Helisys LOM 1015 system and the material
used for fabrication was a laminated paper of
0.09652 mm thickness. As shown in Figures 10
and 11, most of the features could be built.
It was practically impossible to separate some
of the features from the support structures.
For examples, some of the pass-fail features,
such as the thin cylinders, could not be
separated from the crosshatches. In creating
crosshatches, it has been found that it is better
to make the crosshatches smaller than necessary
for easy post-processing. In our experiment, the
crosshatches were 0.125 in. (3.175 mm) and
the fine crosshatches were 0.06 in. (1.524 mm).
The various speeds used in the experiment
include: cutting speed of 16 in./s, platform
speed of 2 in./s (50.8 mm/s), feeder speed of
Figure 10 Benchmark part built on the Helisys LOM system
 Benchmarking for comparative evaluation of RP systems and processes Rapid Prototyping Journal
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh Volume 10 · Number 2 · 2004 · 123–135

Figure 11 Highlighted part showing some results of fabrication like the consequently, broke away readily during the
delamination, air holes, thin walls and brackets post processing.
In general, the dimensional accuracy of the
LOM part is quite satisfactory, although it is
not as accurate as the SLA part. Post
processing is most delicate and time-consuming
in LOM.

Fabrication of the benchmark part on the

FDM machine
The benchmark part was fabricated on the
Stratasys FDM 3000 using ABS-400 as the part
material and ABS-400R as the support material.
All part features, except the pass-fail features,
could be fully built (Figures 12 and 13). The
FDM process was least suitable to build fine
features as compared with the other processes.

Figure 12 Benchmark part built on the FDM 3000

4 in./s (101 mm/s), and heater speed of 2 in./s
(50.8 mm/s). The tile size was 1 in. (25.4 mm),
and the laser power setting was 1.6 per cent.
Delamination is a severe problem in the case of
the LOM systems. The benchmark part could
help in the identification and evaluation of
delamination, and its influence on the geometric
features. The features, such as the wedge and
the freeform features, appeared to be separated
from the base because of the delamination.
Delamination limits the geometric accuracy of
a part built by LOM. It generally occurs if the
bonding between the layers is not strong enough
to hold them together and separation between
the bonding layers could cause the part to Figure 13 Highlighted areas on the benchmark part showing the warpage,
distort, thereby affecting the geometric and failure to build – very thin walls, cylinders
dimensional accuracy of the part. Any of the
following could influence delamination: the
material selection, heating, weak bonding
between the base of the part and platform,
humidity, and induction of air through the layers.

Effect of delamination on the benchmark part

The LOM-fabricated benchmark part was
delaminated more towards the base as seen in
Figure 11. An analysis on the possible cause of
delamination indicates that it was caused by
the induction of air during the fabrication
process. Humidity of the environment caused
the induced air pockets in the model and
ultimately its delamination. The features
above the base of the benchmark part were
affected as well by delamination and
129
Benchmarking for comparative evaluation of RP systems and processes
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh
In addition, the surface finish and dimensional
accuracy were not as good as those from the
other processes.
Warpage
The base was warped, but to a lesser extent
when compared with that of the SLS part.
The warpage was seen to be more progressive
towards the corners to result in an upward
curling of the base. The warpage could have
been due to the temperature setting in the
build area or late removal of the part from
the FDM machine after fabrication. The model
liquefier temperature was at 2708C, the
support liquefier temperature was at 2358C,
and the envelope temperature was at 708C.
Using proper temperature setting for
particular materials is important to fabricate
a good part.
Measurement
For comparison of parts built by various
RP processes, there is a necessity to have
standardised measurement technique for
consistent evaluation. It is noteworthy that
although Shellabear (1999) used the same
benchmark part of Reeves and Cobb (1996)
their measurement results could not be directly
correlated due to different measurement
methods. A standardised measurement system
and technique would reduce variability. The
CMM machine is a suitable measurement
equipment because of its versatility, speed and
high accuracy compared to other measurement
methods and can be programmed to carry out a
variety of automatic measurements, ranging
from simple to complex, to measure dimensions
and shape errors, based on ISO standards
such as straightness (ISO 12780), roundness
(ISO 12181), flatness (ISO 12781), cylindricity
(ISO 12180), etc.
Measurement of the benchmark part on the
CMM
A CMM was programmed to carry out automa-
tic measurement on the benchmark part to
minimise inconsistency that would be incurred
with manual measurement. The measurement
points picked on the benchmark part would
be the same in the programmed measurement
for the parts from different processes after
appropriate referencing on the part.
130
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · 123–135
Measurement of the geometric features
There are various standardised measurements
that can be conducted on the benchmark part.
These include:
(1) measurements of elements or individual
features, such as points, straight lines, holes,
boss, spheres, cylinders, cones, slots, etc.;
(2) relations between the elements;
.
distance relation, for example, the
distance between centres of circular
features,
.
intersection relationships such as the
angle between two planes,
.
checking geometric tolerances,
.
parallelism tolerances,
.
perpendicularity tolerances,
. angularity tolerances,
.
checking coaxiality,
.
checking concentricity,
.
symmetry tolerances, and
.
position tolerances.
Table III compares the features built by the
various RP processes and Table IV compares the
associated relative measurements. The results
are tabulated and rated in the order of the best
to the worst performance, respectively. It can be
observed that the SLA is best in terms of
accuracy and surface finish, followed by SLS,
LOM and FDM. FDM and LOM are least
suitable in building very fine features. For
medium-sized features, the order of
performance is SLA, LOM, SLS and FDM.
The order of best performance in terms of
surface roughness is SLA, LOM, SLS and
FDM, as seen in Appendix 1. Additionally,
Appendix 1 present the warpage analysis on the
fabricated benchmark parts. Tables AI and AII
in Appendix 2 show the details of the
programmed CMM measurements on the
geometric features and some of their relative
measurements.
Process benchmarks
A process benchmark aims to facilitate the
consistent fabrication of the best part that can
be achieved with the process. In the
aforementioned experiments, the fabricated
geometric benchmark part has been used to
identify suitable process benchmarks. Evolving
 Benchmarking for comparative evaluation of RP systems and processes Rapid Prototyping Journal
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh Volume 10 · Number 2 · 2004 · 123–135

Table III Comparison of the various RP process based on the fabrication of the Shi (1997), Kulkarni et al. (2000), Lee and Choi
geometric benchmark part (2000), Muller and Schimmel (1999), Rosen
et al. (2001), Shellabear (1998), Xu et al.
Features SLA SLS LOM FDM
(1997), Zhao and Laperriere (2000) and Zhou
SB A A A A
et al. (2000). In addition to a suitable process
CB A B B B
planning, there are also several other parameters
FB A B B C
that need to be properly tuned during the
CH (z-direction) A C A A
fabrication of a part, such as hatching, scan-
SP A D A A
speed, laser, etc. depending on the RP process.
SC A D D E
Obviously, the best method for the fabrication
HC A B B B
of the RP part and the measurement technique
CN A D B C
for it also need to be identified to develop a
SL A E D E
successful process benchmark.
HS A B A B
Performance indices targeted in achieving the
BR C B C C
best process are typically accuracy and surface
CR A C A A
finish, as well as minimum build time. The use
Mechanical features (FL, BL) A C B E
of the benchmark part can facilitate the
Pass-fail features TW, SL, SH, SC, FF A C E F
optimisation of key process parameters to evolve
Surface roughness 0.4 12.1 2.6 18.4
a suitable process benchmark to achieve these
Index: per cent Deviation from nominal dimensions
indices.
< 5 per cent Very good A
< 10 per cent Good B
< 15 per cent Satisfactory C Conclusion
< 20 per cent Poor D
< 25 per cent Worse E Various standardisation issues have been given,
< 50 per cent or not built Fail F presented and discussed with regard to
appropriate geometric benchmarks. The
geometric benchmark can be suitably fabricated
Table IV A comparison on the relative measurements
by different RP processes and inspected/tested
Relative measurements SLA SLS FDM LOM using the best practice or controlled process.
Distance (HS1, HS2, SB) A D C B The results captured and stored in a web-based
Flatness (SB, FB) B C A D database will serve as an easily accessible central
Symmetry (SP) A D C B database (Fuh et al., 2002). This benchmark
Coaxiality (HC) A D B C database can then be used by appropriate RP
Perpendicularity (SB-CB) A D B C decision support systems to identify suitable RP
Angularity (wedge) A D C B machines, materials and processes to meet the
Parallelism (CB) B D A C specific requirements.
Note: A-D: ranking in the order of best to worst

a process benchmark could be lengthy, but the
result would be highly beneficial as it can
subsequently reduce the time and resources
needed to optimise the building of the parts.
Process planning is an essential part of
evolving the process benchmark and this
includes: part orientation, support structure
design, slicing of the part, path planning, and
process parameter selection. RP process
planning or related research work has been
extensively carried out by Allen and Dutta
(1995), Cahlasani et al. (1995), Cheng et al.
(1995), Dolenc and Makela (1994), Gibson and
131
References
Allen, S. and Dutta, D. (1995), “Determination and evaluation
of support structures in layered manufacturing”,
Journal of Design and Manufacturing, Vol. 5,
pp. 153-62.
Cahlasani, K., Jones, L. and Roscoe, L. (1995), “Support
generation for fused deposition modelling”, Solid
Freeform Fabrication Symposium, pp. 229-41.
Cheng, W., Fuh, J.Y.H., Nee, A.Y.C., Wong, Y.S., Loh, H.T. and
Miyazawa, T. (1995), “Multi-objective optimisation of
part-building orientation in stereolithography”, Rapid
Prototyping Journal, Vol. 1 No. 4, pp. 12-23.
Dolenc, A. and Makela, I. (1994), “Slicing procedures for
layered manufacturing techniques”, Computer-Aided
Design, Vol. 26 No. 2, pp. 119-26.
Benchmarking for comparative evaluation of RP systems and processes
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh
Fuh, J.Y.H., Loh, H.T., Wong, Y.S., Shi, D.P., Mahesh, M. and
Chong, T.S. (2002), “A Web-based database system for
RP machines, processes and materials selection”, in
Gibson, I. (Ed.), Software Solutions for RP, PEP Ltd, UK,
pp. 27-55.
Gargiulo, E.P. (1992), “Stereolithography process accuracy:
user experience”, Proceedings of the 1st European
Conference on Rapid Prototyping, 6-7 July, University
of Nottingham, pp. 187-207.
Gibson, I. and Shi, D. (1997), “Material properties and
fabrication parameters in selective laser sintering
process”, Rapid Prototyping Journal, Vol. 3 No. 4,
pp. 129-36.
Ippolito, R., Iuliano, L. and de Filippi, A. (1994), “A new user
part for performances evaluation of rapid prototyping
systems”, Proceeding of 3rd European Conference on
Rapid Prototyping and Manufacturing, 6-7 July,
Nottingham, UK, pp. 327-39.
Ippolito, R., Iuliano, L. and Gatto, A. (1995), “A
benchmarking of rapid prototyping techniques in terms
of dimensional accuracy and surface finish”, CIRP
Annals, Vol. 44 No. 1, pp. 157-60.
Juster, N.P. and Childs, T.H.C. (1994a), “Linear and geometric
accuracies from layer manufacturing”, Annals of the
CIRP, Vol. 43 No. 1, pp. 163-6.
Juster, N.P. and Childs, T.H.C. (1994b), “A comparison of rapid
prototyping processes”, Proceedings of the 3rd
European Conference on the Rapid Prototyping and
Manufacturing, 6-7 July, University of Nottingham,
pp. 35-52.
Kruth, J.P. (1991), “Material incress manufacturing by rapid
prototyping techniques”, CIRP, Annals. 41st General
Assembly of CIRP, Vol. 40, No. 2, Stanford, CA, USA,
pp. 603-14.
Kulkarni, P., Marsan, A. and Dutta, D. (2000), “A review of
process planning techniques in layered
manufacturing”, Rapid Prototyping Journal, Vol. 6
No. 1, pp. 18-35.
Lart, G. (1992), “Comparison of rapid prototyping systems”,
Proceedings of the 1st European Conference on Rapid
Prototyping, 6-7 July, University of Nottingham, UK,
pp. 243-54.
Lee, K.H. and Choi, K. (2000), “Generating optimal slice data
for layered manufacturing”, International Journal of
Advanced Manufacturing Technology, Vol. 16,
pp. 277-84.
Muller, H. and Schimmel, A. (1999), “The decision dilemma –
assessment and selection of rapid prototyping process
chains”, Proceedings of the 8th European Conference
on Rapid Prototyping and Manufacturing, Nottingham,
pp. 177-92.
Reeves, P.E. and Cobb, R.C. (1996), “Surface deviation
modelling of LMT processes – a comparative analysis”,
Proceedings of the 5th European conference on Rapid
Prototyping and Manufacturing, 4-6 June, Helsinki,
pp. 59-75.
Rosen, D.W., West, A.P. and Sambu, S.P. (2001), “A process
planning method for improving build performance in
stereolithography”, Computer Aided Design, Vol. 33,
pp. 65-79.
132
Rapid Prototyping Journal
Volume 10 · Number 2 · 2004 · 123–135
EOS GmbH Shellabear, M. (1998), “Model manufacturing
processes – state-of-the-art in rapid prototyping”,
RAPTEC, Task 4.2, Report 1.
EOS Gmbh Shellabear, M. (1999), “Benchmark study of
accuracy and surface quality in RP models”, RAPTEC,
Task 4.2, Report 2.
Wong, Y.S., Fuh, Y.H., Loh, H.T. and Mahesh, M. (2002),
“Rapid prototyping and manufacturing (RP&M)
benchmarking”, in Gibson, I. (Ed.), Software Solutions
for RP, PEP Ltd, UK, pp. 57-94.
Xu, F. (1999), “Integrated decision support for part
fabrication with rapid prototyping and manufacturing
systems”, PhD Thesis, National University of Singapore.
Xu, F., Wong, Y.S., Loh, H.T., Fuh, J.Y.H. and Miyazawa (1997),
“Optimal orientation with variable slicing in
stereolithography”, Rapid Prototyping Journal, Vol. 3
No. 3, pp. 76-88.
Zhao, Z. and Laperriere, L. (2000), “Adaptive direct slicing of
the solid model for rapid prototyping”, International
Journal of Production Research, Vol. 38 No. 1,
pp. 69-83.
Zhou, J., Herscovici, D. and Chen, C. (2000), “Parametric
process optimization to improve the accuracy of rapid
prototyped stereolithography parts”, International
Journal of Machine Tools and Manufacture, Vol. 40
No. 3, pp. 363-79.
Appendix 1. Surface roughness
The surface roughness is measured using the
Rank Taylor Hobson’s surface texture
measuring equipment. The procedures followed
are based on ISO 468.
Z L
1
Ra ¼ j yðxÞj dx
L 0
where Ra is the arithmetic mean of the
departures of the roughness profile from the
mean line, L, assessment length, is defined as
the length of the profile used for the
measurement of surface roughness.
The result of the surface roughness
measurement is shown in Figure A1.
Figure A1 Surface roughness, Ra
Benchmarking for comparative evaluation of RP systems and processes Rapid Prototyping Journal
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh Volume 10 · Number 2 · 2004 · 123–135

Warpage analysis on the benchmark parts Figure A5 First SLS part – warpage measurement
The warpage analysis on the benchmark parts
was done with the aid of the CMM machine.
The objective perform a quantitative
mesurement of warpage on the base, SB.
A matrix ð5 £ 5Þ of points was chosen on the
base of the benchmark part. With a fixed
datum, the deviation of the plane was
determined. The deviations are plotted
graphically in Figures A2-A6. The graphs
show the extent of warpage on the
benchmarks from the various RP Figure A6 Second SLS part – warpage measurement
processes.The SLA part shows the least

Figure A2 SLA – warpage measurement

deviation (within 2 mm from the nominal

plane), the FDM part is the next with
deviation of less than 2.5 mm and the LOM
part with a deviation of within 5 mm. The
first SLS part built without the machine being
appropriately tuned showed the maximum
Figure A3 LOM – warpage measurement deviation of around 7 mm, however the
second part showed a lower deviation of
around 4.5 mm.

Appendix 2

Roundness is like profile of a line except that

the curve is closed to itself. Flatness is the 3D
equivalence of straightness. Cylindricity is the
3D equivalence of roundness. Concentricity is a
tolerance where the axis of a feature is
required to be coaxial to a specified datum
Figure A4 FDM – warpage measurement regardless of the datum’s or feature’s size.
Parallelism is a tolerance, which controls
independent surfaces and axes, which are to be
equal distances from a datum plane or axis.
Perpendicularity is a tolerance, which controls
surfaces and axes, which are 908 from the
datum axis. Angularity is the tolerance of an
axis surface, or centre plane at a specified
angle from another feature, datum plane or
axis.
133
Benchmarking for comparative evaluation of RP systems and processes Rapid Prototyping Journal
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh Volume 10 · Number 2 · 2004 · 123–135

Table AI Geometric feature measurements

Geometric features SLA SLS FDM LOM
CH 1 DM 10.0833 DM 8.9668 DM 10.0024 DM 9.9089
RD 0.0042 RD 0.3907 RD 0.0073 RD 0.0391
CH 2 DM 10.0950 DM 8.9725 DM 9.9759 DM 9.9204
RD 0.0038 RD 0.2037 RD 0.0821 RD 0.0108
CH 3 DM 10.0765 DM 8.9881 DM 10.0030 DM 9.8299
RD 0.0066 RD 0.1857 RD 0.0333 RD 0.0484
CH 4 DM 10.0920 DM 9.2271 DM 10.0055 DM 9.8783
RD 0.0083 RD 0.2356 RD 0.0085 RD 0.0476
CR 1 DM 10.0621 DM 9.2545 DM 9.9803 DM 9.8214
RD 0.0297 RD 0.2900 RD 0.0069 RD 0.0027
CR 2 DM 10.0609 DM 8.8200 DM 9.9582 DM 9.8079
RD 0.0029 RD 0.4012 RD 0.0048 RD 0.0098
CR 3 DM 10.0771 DM 8.8254 DM 10.0122 DM 9.7998
RD 0.0102 RD 0.3126 RD 0.0095 RD 0.0290
CR 4 DM 10.0833 DM 8.7569 DM 10.0063 DM 9.8457
RD 0.0184 RD 0.2690 RD 0.0106 RD 0.0122
CR 5 DM 15.1281 DM 14.0473 DM 15.0151 DM 14.8728
RD 0.0055 RD 0.2611 RD 0.0518 RD 0.0062
SP 1 DM 14.8131 DM 17.0193 DM 14.9896 DM 15.2477
SPR 0.0167 SPR 0.3076 SPR 0.0075 SPR 0.0204
SP 2 DM 14.9588 DM 17.1279 DM 14.9556 DM 15.2621
SPR 0.0384 SPR 0.1183 SPR 0.0355 SPR 0.0882
SP 3 DM 14.8575 DM 17.5853 DM 14.8421 DM 15.1629
SPR 0.0196 SPR 0.1870 SPR 0.0304 SPR 0.0732
SP 4 DM 14.8954 DM 17.6982 DM 14.8352 DM 15.1495
SPR 0.0091 SPR 0.2150 SPR 0.0589 SPR 0.0698
SC 1 DM 9.6566 DM 7.4202 DM 7.5251 DM 7.7351
CYN 2.7930 CYN 2.6043 CYN 2.0692 CYN 2.2595
SC 2 DM 4.5793 DM 4.2680 DM 3.5959 DM 3.4897
CYN 1.1759 CYN 1.4285 CYN 1.1888 CYN 1.4700
SC 3 DM 4.9095 DM 3.7845 DM 3.0930 DM 3.8761
CYN 1.4740 CYN 2.1065 CYN 1.7969 CYN 1.6755
SC 4 DM 8.4497 DM 8.4434 DM 8.0537 DM 8.1104
CYN 1.9517 CYN 2.4248 CYN 2.4296 CYN 2.9025
SC 5 DM 3.5727 DM 5.2572 DM 3.6848 DM 3.8603
CYN 2.0019 CYN 1.4967 CYN 1.3674 CYN 1.8887
SC 6 DM 3.9235 DM 4.5556 DM 3.3157 DM 3.5200
CYN 1.3723 CYN 1.6636 CYN 1.8819 CYN 2.3731
HC 1 DM 10.7810 DM 9.0160 DM 8.1143 DM 8.7872
CYN 1.0227 CYN 2.4199 CYN 2.6523 CYN 2.4081
HC 2 DM 8.9863 DM 8.7845 DM 8.3199 DM 9.1013
CYN 2.7811 CYN 2.3168 CYN 2.3107 CYN 1.5369
CN 1 ANG 36.820035 ANG 38.687633 ANG 36.920307 ANG 35.222364
CON 0.0114 CON 0.0984 CON 0.0158 CON 0.0241
CN 2 ANG 37.201382 ANG 38.881123 ANG 36.992861 ANG 35.772145
CON 0.0162 CON 0.0451 CON 0.0321 CON 0.0643

Notes: DM: diameter, RD: roundness, CYN: cylindricity, SPR: spherity, CON: concity, FLT: flatness, CCN: concentricity, SQR:
squareness, PAR: parallelism, and ANG: min. angularity

134
 Benchmarking for comparative evaluation of RP systems and processes Rapid Prototyping Journal
M. Mahesh, Y.S. Wong, J.Y.H. Fuh and H.T. Loh Volume 10 · Number 2 · 2004 · 123–135

Table AII Relative measurements

Relative measurements SLA SLS FDM LOM
Distance
HS 1 61.4881 £ 60.5284 62.0734 £ 60.6094 60.9950 £ 60.0947 61.1436 £ 60.3210
HS 2 25.1545 £ 25.1210 26.1230 £ 25.5292 25.0024 £ 25.0910 25.1909 £ 25.2777
SB 170.1038 £ 170.3304 170.8658 £ 170.1549 170.1738 £ 169.8054 170.1867 £ 170.9608
Flatness
SB 0.1468 0.1357 0.1397 0.2731
FB 0.0414 0.0680 0.0184 0.0287
Symmetry
SP 145.8939,145.7701 146.4939,146.7701 145.3584,145.3284 146.8965,146.7832
109.8494, 109.7736 110.3554,110.3855 109.8965,109.7894 110.2500,110.1925
Coaxiality
HC RD 0.0099 RD 0.2737 RD 0.0530 RD 0.0285
CCN 0.1843 CCN 0.2598 CCN 0.2062 CCN 0.1645
RD 0.0065 RD 0.3025 RD 0.0219 RD 0.0371
CCN 0.1939 CCN 0.1723 CCN 0.2444 CCN 0.2917
Perpendicularity FLT 0.0224 FLT 0.0075 FLT 0.0052 FLT 0.0970
SB-CB SQR 0.1136 SQR 0.5572 SQR 0.0358 SQR 0.3066
FLT 0.0381 FLT 0.0026 FLT 0.0554 FLT 0.0409
SQR 0.1147 SQR 0.7023 SQR 0.1758 SQR 0.5448
Angularity ANG 153.419278 ANG 155.115186 ANG 153.797042 ANG 153.641372
Wedge ANG 148.903473 ANG 150.472539 ANG 149.247875 ANG 148.010241
Parallelism FLT 0.0018 FLT 0.0374 FLT 0.0050 FLT 0.0078
CB PAR 0.0772 PAR 0.4744 PAR 0.0933 PAR 0.2333
FLT 0.0024 FLT 0.0304 FLT 0.0022 FLT 0.0014
PAR 0.0576 PAR 0.5213 PAR 0.0162 PAR 0.0568

135