Journal of Manufacturing Processes 21 (2016) 39–45

Contents lists available at ScienceDirect

Journal of Manufacturing Processes

journal homepage: www.elsevier.com/locate/manpro

Technical Paper

Mechanical behaviour of ABS: An experimental study using FDM and

injection moulding techniques
Michael Dawoud ∗ , Iman Taha, Samy J. Ebeid
Mechanical Design and Production Engineering Department, Ain Shams University, El-Sarayat St. 1, Cairo 11517, Egypt

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

Article history: Fused deposition modelling (FDM) in contrast to injection moulding was studied to investigate the effect
Received 2 December 2014 of processing technique on the mechanical behaviour of virgin ABS. FDM parameters were further altered
Received in revised form 23 August 2015 in terms of varying raster angle and gap to further explore the potential of this technique. Results show
Accepted 3 November 2015
that an adequate selection of FDM parameters is able to reach mechanical properties comparable to those
Available online 9 December 2015
of injection moulded parts in both static and dynamic loading modes. Here, a negative raster gap proved
to be most significant in enhancing mechanical behaviour. A raster angle layup of −45◦ /+45◦ proves to
Keywords:
offer maximum tensile and impact strength, whereas highest flexural strength was recorded for a 0/90◦
Fused deposition modelling
DIY 3D printers
scaffolding system. In contrast, a positive gap drastically reduces the performance. Dimensional analysis
Injection moulding further show no significant alterations of dimensions are to be expected with varying raster angle and
Acrylonitrile butadiene styrene gap.
© 2015 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

1. Introduction In recent years, FDM technology has gained a public interest

3D Printing (3DP) is a technology used to directly convert 3D
Computer-Aided Design (CAD) data into physical prototype. Most
3D printing techniques are based on slicing a CAD model into a
number of 2D layers and building them layer by layer to form the
prototype.
One of many available 3DP techniques is Fused Deposition Mod-
elling (FDM) [1,2]. In this process a filament of material is melted
in a heated nozzle and deposited on a build platform. By using a
3-axis motion system, this nozzle is moved in the XY plane prin-
ting a layer of the prototype. When this layer is finished, the build
platform is moved down one step (known as the slice thickness) in
the Z direction and the cycle is repeated for the next layer until the
complete model is built-up [3].
FDM technology is primarily used for rapid prototyping poly-
mer parts [3,4]. The choice of material depends on the type
of application and desired properties. Nowadays, commonly
applied materials include Polylactic Acid (PLA) as a stiff and
environmentally-friendly material, Nylon for soft applications (e.g.
bracelets), high density polyethylene (HDPE) for the production of
food-compatible parts and Acrylonitrile Butadiene Styrene (ABS) as
a general solution for tough parts with acceptable strength.
∗ Corresponding author. Tel.: +20 1221271808.
E-mail address: michael.dawoud@eng.asu.edu.eg (M. Dawoud).
for its ease of use as well as the simplicity of the machinery itself.
Today, many FDM machines are designed and constructed by hob-
byists as a Do-It-Yourself (DIY) machine for personal use [5]. These
even assume that in the coming years, many of our daily-use plas-
tic products, especially spare parts that are produced by Injection
Moulding (IM) will be replaced by CAD models that would be home
printed for cost reductions and personalization. A problem, how-
ever, remains the expected loss in mechanical properties in contrast
to the monofilament or the injected material [2,6].
Whereas layer thickness, part orientation and raster width were
proven to have little influence on mechanical properties [2,7], other
printing parameters such as raster angle and air gap between two
successive rasters were observed to significantly affect the perfor-
mance of 3DP parts [2,7–9]. Sood et al. [2] observed that tensile
strength is enhanced by increasing the raster angle. In contrast,
Vega et al. [8] reported a maximum tensile strength at a raster ori-
entation of 0◦ to the tensile load. This angle of zero additionally
proved to be most suited for maximum flexural strength [2,8]. Fur-
ther investigations conducted by Sood et al. [7] show failure of 3D
parts due to buckling or debonding of adjacent rasters upon the
application of compressive loads.
A positive air gap was proposed to allow material flow towards
adjacent layers thus contributing to the enhancement of mechani-
cal properties [2]. In contrast, Ahn et al. [9] describe that a negative
air gap of 0.003 (0.08 mm) between rasters increases part density
thus enhancing tensile behaviour. A more negative air gap leading

http://dx.doi.org/10.1016/j.jmapro.2015.11.002
1526-6125/© 2015 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
40 M. Dawoud et al. / Journal of Manufacturing Processes 21 (2016) 39–45

Table 1
Overview of previous work in FDM technology.

Reference Machine Raster angle Other printing parameters Mechanical properties Form accuracy

[2,7] Stratasys 0, 30, 45, 60 Raster gap, raster width, layer thickness, layer Tension, compression Dimensions
orientation bending, impact
[8] Stratasys 0, 45, 90, 45/−45, 45/0 Layer orientation Tension, bending, impact –
[9,15] Stratasys 0, 90, 0/−90, 45/−45 Raster gap, raster width, printing temperature Tension, compression –
[6,16] Stratasys 0, 90 Raster gap, raster width, printing temperature Tension, torsion –
[10] Stratasys – No. of contours, layer orientation Tension –
[17] Stratasys 0/-90, 30/-60, 45/-45 Bending –
[18] Stratasys – Raster width, no. of contours, envelope – Voids
temperature, printing temperature
[19] Stratasys – Raster width, layer thickness, deposition speed – Roughness
[20] Stratasys – Raster width, layer thickness – Dimensions,
roughness
[21,22] Stratasys 0, 30, 60, 90, 0/−90, – Tension –
15/−75, 30/−60, 45/−45
Study DIY 0/−90, 15/−75, 30/−60, Raster gap Tension, bending, impact –
45/−45

to raster overlap would, however, lead to dimensional inaccuracy.
Similarly, Croccolo et al. [10] suggest improved strength through
increased number of contours.
Table 1 gives an overview of the parameters and mechanical
properties of FDM products intensively investigated in literature.
However, literature lacks an in-depth study of the criss-cross (+/−)
meshing system, where the raster angle in each layer is orthogonal
to the preceding layer. This meshing system is the main system used
by open source slicing software used in DIY 3D printer projects.
Studies are based, however, on commercially available FDM equip-
ment, relying on a single raster angle that is constant for all the
layers as a meshing system. In this context, based on the use
of DIY machines the authors propose a comparison of mechani-
cal properties for the commonly applied Acrylonitrile Butadiene
Styrene material processed by injection moulding and fused
deposition modelling in order to highlight the difference in
mechanical properties encountered by the different processing
techniques. Further, the effect of FDM printing parameters on the
performance of printed parts is investigated, targeting the selec-
tion of appropriate parameters that enable the fabrication of parts
of comparable behaviour to injection moulding.
2. Experimental work
2.1. Materials
Virgin ABS pellets supplied by Grand Pacific Petrochemical Corp
(D-150) [14] were used to extrude ABS filament for subsequent
FDM and injection moulding processing. Pellets were first dried at
80 ◦ C for 24 h prior to extrusion into 1.7 ± 0.05 mm filaments using

Table 2
Dimensions of tensile, flexural, and impact specimens.

Tensile specimens

Parameter Injected specimen 3D printed specimen

Gauge length L0 [mm] 50 ± 0.5 50 ± 0.5

Length of parallel section L1 [mm] 60 ± 0.5 60 ± 0.5
Distance between shoulders L2 [mm] 107.5 120
Total length L3 [mm] 195 220
Width of the parallel section B1 [mm] 10 ± 0.2 10 ± 0.2
Width at ends B2 [mm] 20 ± 0.2 20 ± 0.2
Thickness H1 [mm] 4 ± 0.2 4 ± 0.2
Fillet Radius R [mm] 40 125

Bending and impact specimens

Overall length L [mm] 80 mm ± 0.2 80 mm ± 0.2

Width B [mm] 10 mm ± 0.2 10 mm ± 0.2
Thickness H [mm] 4 mm ± 0.2 4 mm ± 0.2
 M. Dawoud et al. / Journal of Manufacturing Processes 21 (2016) 39–45 41

Table 3 2.3. Testing methods

Injection moulding parameters.

Parameter Value Density of specimens was measured using an Adam Equipment

Screw speed 180 rpm PW254 sensitive scale of 0.0001 g resolution and further equipped
Injection pressure 90 kg/cm2 with a density measuring kit.
Clamping pressure 105 kg/cm2 The effect of varying FDM parameters on dimensional accuracy
Holding time 35 s was investigated by measuring sample width and thickness using
a Mitutoyo Vernier Caliper at 0.01 mm accuracy. Average of 3 mea-
Table 4 surements was compared to the respective dimensions of the CAD
Sample designation and scaffolding setup. models to make statements about eventual dimensional variations.
Both injection moulded and FDM samples were tested for their
Designation Processing Raster angle Raster angle Gap (mm)
technique (even layers) (odd layers) mechanical behaviour. Tensile testing was conducted in compli-
ance to DIN EN ISO 527-2 using a Lloyd LRX Plus Universal Testing
I IM – – –
−P0 FDM 0 +90 −0.05 Machine equipped with a 5 kN load cell at a crosshead speed of
−P30 FDM +30 −60 −0.05 5 mm/min. 3-point bending was carried out according to the ASTM
−P45 FDM +45 −45 −0.05 standard 790-03 using the same Universal Testing Machine and test
−P75 FDM +75 −15 −0.05 speed.
+P0 FDM 0 +90 +0.05
Charpy impact strength was determined using un-notched sam-
+P30 FDM +30 −60 +0.05
+P45 FDM +45 −45 +0.05 ples in accordance to the ASTM standard 6110-04 applying a
+P75 FDM +75 −15 +0.05 BEIJING JINSHENGXIN digital impact testing machine, where a 2-
Joule hammer edgewise strikes the specimen.
SEM microscopy was performed using a CamScan Series 4
a Beier Machinery Co. Ltd. TSK 35 twin screw extrusion machine. scanning electron microscope using a 15 keV acceleration voltage.
Extruded filaments were partially pelletized into 3 mm pellets for Samples were gold sputtered prior to investigation.
injection moulding and the remainder wound on a drum for FDM.

2.2. Sample preparation 3. Results and discussion

Two sets of samples were prepared to perform a compari- 3.1. Density

son between injection moulded and filament deposition modelled
parts. Fig. 1 shows that injection moulded samples have the highest
Specimens for tensile, flexural and impact testing were prepared density of 1.054 g/cm3 , while the +P layups have the lowest den-
in compliance to DIN EN ISO 527-2, 178 and 179-1 standard, respec- sity of 0.943 g/cm3 . This result is conceivable in view of the fact that
tively having dimensions shown in Table 2. Injection moulding was injection moulding involves the application of high pressures that
performed using a Fu Chun Shin 60 ton, single screw hydraulic compact the material upon injection into the die cavity. Moreover,
injection moulding machine applying the injection parameters applied holding pressure during injection moulding additionally
listed in Table 3. compensates material shrinkage, resulting in more compact sam-
The effect of scaffolding angle, as well as air gap between rasters ples compared to FDM parts, where the individual rasters are
was studied by varying these parameters according to the experi- simply deposited next to each other.
mental setup summarized in Table 4. However, it can also be depicted from Fig. 1 that using negative
FDM samples were produced using a DIY FDM machine of air gaps between individual rasters is able to reach 98% of the max-
280 × 380 × 120 mm print volume, equipped with Marlin firmware imum IM density. In contrast, FDM parts produced at a positive air
that communicates with the user interface software Repitier-Host gap are much lower in density. This can be related to the fact that
version V1.0.4 obtained from [11]. STL files were generated using a positive air gap causes adjacent rasters to barely touch along a
Solid Works 2012 CAD software and further sliced for printing using
Slic3r program version 1.1.7 [13].
Similar specimens were prepared using the FDM technique. For
printed tensile samples, recommendations proposed by Croccolo
[10] were adopted, where the fillet radius was increased to 250 mm
to avoid stress concentrations. The CAD models for the various spec-
imens were sliced into layers of 0.5 mm thickness, resulting into 8
layers.
To prevent variation in FDM sample properties due to tempera-
ture differences on the build platform, each sample was printed
individually at the centre of the printing bed according to the
parameters given in Table 5.

Table 5
FDM parameters.

Parameter Value

Nozzle diameter 1 mm
Layer height 0.5 mm
No. of contours 1
Printing speed 30 mm/s
Nozzle temperature 250 ◦ C ± 5
Bed temperature 120 ◦ C
Fig. 1. Density of injection moulded and 3D printed parts.
42 M. Dawoud et al. / Journal of Manufacturing Processes 21 (2016) 39–45
Fig. 2. Tensile strength of injection moulded and 3D printed parts.
line (resulting from the cylindrical raster shape), whereas a neg-
ative raster gap, designed for overlapping rasters causes adjacent
filaments to partially squeeze and fill in the air gap in between.
3.2. Tensile strength
Fig. 2 shows the effect of production technique and 3DP parame-
ters on tensile strength. It becomes evident that there is a difference
in tensile strength between injection moulded (37.7 MPa) and FD
modelled (34.3 MPa) specimens, the latter being the highest tensile
strength for printed samples observed at −P45. This is expected
in view of the more compact injection moulded part, as proven
by density measurements. In addition, the nature of the injection
moulding process, in which molten material is introduced to a
chilled cavity, allows the formation of crystals, which in turn brings
along increased mechanical properties. In contrast, in the FDM pro-
cess, the molten filaments are deposited on a warm bed kept at
120 ◦ C where it is then left to air cool — a condition which favours
an amorphous, and thus less strong polymer structure. The observa-
tions made at this stage lie in complete agreement with the findings
reported by Ahn et al. [9].
Comparing tensile strength of FDM samples produced at pos-
itive and negative air gaps, Fig. 2 shows that a negative air gap
results in higher tensile strength. Again, this can be attributed to
the increased density of these parts, where rasters are allowed to
slightly overlap resulting in stronger interfacial bonding between
Fig. 3. SEM of the fracture surface of tensile samples of (a) injection moulded, (b) −P0 and (c) +P0 samples.
adjacent rasters [18]. In contrast, a positive air gap leads to the
formation of in-plane neighbouring cylindrical rasters that barely
touch, leading to week axial bonds. Thus, part integrity is primarily
achieved through bonding to subsequent layers and not within the
same layer. This concept is supported by SEM micrographic anal-
ysis, as illustrated in Fig. 3, where the fracture surfaces of tensile
samples of printed +P0, and −P0 are shown in contrast to injection
moulding.
Further, it can be depicted from Fig. 2 that varying raster angle
seems to have a more significant effect on tensile strength when
deposited with a positive gap, compared to raster angle variation
at a negative gap. These observations were additionally supported
by ANOVA [12] at a 95% confidence level and can be related to the
clear position of the individual rasters, which signify the effect of
the printing angle thus directly affecting its mechanical behaviour.
At this point it is worth mentioning, that behaviour observed in this
study come opposing to the observations reported by Sood et al. [2],
where a negative air gap was assumed to cause thermal stresses to
the material by reheating the adjacent layers, thus concluding that
increased air gap favours increase in strength.
Ziemian et al. [21,22] report that the effect of raster angle is of
high influence on the tensile properties of FDM samples. A single
90◦ angle was reported to exhibit highest mechanical properties
while a 0◦ raster orientation yields lowest mechanical properties.
This is due to the fact that at a raster angle ≤45◦ the sample strength
is mainly dependant on the inter-raster bonding. In the criss-cross
layup, however, sample strength is a result of the interaction of
two orientations, where only the layers >45◦ orientation positively
contribute to the behaviour.
In this respect, it becomes apparent that with a positive air gap
a scaffolding angle of zero results in least tensile strength val-
ues of 28.1 MPa. At this angle the load is only partially carried
by the individual rasters, where only rasters parallel to the load-
ing direction positively contribute to load bearing, whereas layers
composed of transverse rasters have negligible effect. A further
increase in raster angle up to 45◦ further increases tensile strength,
where raster alignment results in a maximum resolved load bearing
area. At an angle of 75◦ the effectively load subjected area is again
reduced, thus decreasing part tensile strength. In view of Fig. 3c,
this behaviour can be regarded in compliance to the behaviour of
fibrous structures or composites.
3.3. Flexural strength
Fig. 4 shows the flexural strength of bending samples produced
by injection moulding and FDM at different conditions. Following
a similar trend to tensile behaviour, samples prepared by injection
 M. Dawoud et al. / Journal of Manufacturing Processes 21 (2016) 39–45
Fig. 4. Flexural strength of injection moulded and 3D printed parts.
moulding proved to have highest flexural strength of 72.5 MPa. In
contrast, printed samples witness a flexural strength reduction of
at least 12% (−P0), depending on scaffolding angle and raster gap.
Next to this drop in flexural strength, Fig. 4 also shows that
filament deposition at a positive gap further drastically reduces
flexural strength. Considering the effect of air gap, a raster angle of
zero shows a significant reduction from 63.8 to 41.8 MPa. Similar to
tensile strength findings varying raster angle has insignificant effect
on flexural behaviour when a negative gap is applied. In contrast,
additionally supported by ANOVA at a 95% confidence level, a pos-
itive gap proves to bring along a significant effect of raster angle on
flexural strength. Again following tensile strength trend, increased
scaffolding angle up to 45◦ readily increases flexural strength. At
this angle all printed layers contribute to load bearing. A minimum
of 32.1 MPa is however recorded at an angle of +P75◦ , where both
even and odd layers are unable to carry the full bending load.
Unlike tensile strength, the maximum flexural strength is
recorded at −P0 configuration, an observation which complies with
the findings reported by Sood et al. [2]. Fig. 5a shows a schematic
representation of the P0 scaffolding layup subjected to bending
load. The bottom layer, representing the odd layers within the sam-
ple, comprises filaments in a sample axis direction, which are able
to carry induced tensile stresses. In contrast, the top (even) lay-
ers are deposited at 90◦ to the axis of the sample thus carry minor
loads in the case of a −P scaffolding. A positive air gap carries lower
loads since the bonding between adjacent filaments in the even
layers are weaker than in the above case. At an angle of 45 a max-
imum resolved shear stress is observed causing filaments to slip
Fig. 5. Schematic representation of a P0 scaffolding subjected to (a) flexural load
and (b) impact load.
43
Fig. 6. Impact strength of injection moulded and 3D printed parts.
apart. This explains the poor flexural behaviour compared to the
P0 layout.
3.4. Impact strength
Results of the impact testing, as presented in Fig. 5b, are shown
in Fig. 6 for both injection moulded and FDM samples at the
various conditions under consideration. It can be noted that injec-
tion moulded samples have highest impact strength of 79.7 kJ/m2 .
Printed samples reach a maximum of 41.8 kJ/m2 at a raster gap
of −0.05 mm and a scaffolding angle of 45◦ while a minimum of
18.3 kJ/m2 is recorded at +P75. This comes in contrast to Es-Said
et al. [8], who reported that the 0◦ orientation has the highest tough-
ness over his sample composition. However, this can be related to
the fact that only a single scaffolding orientation per sample was
used, i.e. no alteration between even and odd layers.
Unlike tensile and flexural strength analysis, ANOVA testing
proves a significant effect of scaffolding angle on impact strength
for both positive and negative raster gaps.
According to Sood et al. [2], the FDM technology causes ther-
mal stresses between the rasters due to the deposition of a newly
molten raster next to a partially solidified raster. These internal
stresses cause the crack to easily propagate between the deposited
filaments. By altering the raster angle, the direction of induced
residual stresses is altered, consequently significantly affecting
sample toughness.
3.5. Dimensional accuracy
Fig. 7 shows exemplary photographs of bending specimens and
their fracture surface after testing. The images confirm the state-
ments made through density measurements. It can be observed
that the use of a negative gap results in denser material, where
the individual filaments are squeezed together resulting in a struc-
ture similar to that of injection moulded parts. In contrast, positive
raster gap results into obvious voids between the individual rasters,
leading to reduced density and reduced flexural properties. It is
further to be noted that the cross section of the rasters, which
is only recognizable in layups of positive gap, remains insignifi-
cantly changed with the alteration of the various angular layups.
The apparent oval cross section of the filaments in the layups +P30,
+P45 and +P75 is based on the angular view of the sample cross sec-
tion in addition to the partial flattening of semi-solid filaments on
a flat bed. However, it is notable that contour rasters still appear in
44 M. Dawoud et al. / Journal of Manufacturing Processes 21 (2016) 39–45

Fig. 7. Photographs of bending specimens and their cross section for the various layups under investigation.

a circular shape. This is because the contour layer is first deposited
forming the borders, after which the filling rasters are deposited
according to the predefined angles.
Fig. 8 further shows the average dimensions measured for the
various layups of tensile, flexural and impact specimens in contrast
to the nominal sample size given by the applied standards. It can
be generally noted that the dimensional variations all lie within the
standards’ acceptable range of ±0.2 mm. Analysis of results using
ANOVA techniques revealed that there is no significant effect of
varying raster gap or angle on sample dimensions. These findings
are in agreement with observations stated by Sood et al. [23]. This
can be understood in view of the printing mechanism, where gap
width is realised by controlling the amount of material extruded
per raster while the extruding nozzle unaffectedly continues its
motion according to the foreseen G-code generated for the part.
Based on this, similar observations were made for thickness vari-
ation, where no significant difference amongst the various layups
has been detected.

Fig. 8. Average dimensions of tensile and flexural/impact specimens in (a) width and (b) thickness.
 M. Dawoud et al. / Journal of Manufacturing Processes 21 (2016) 39–45
Fig. 9. Summary of −P results in contrast to injection moulding.
4. Conclusion
The mechanical properties of ABS prepared by injection mould-
ing generally came superior in all the conducted tests if compared
to those of 3D printing. This is related to the nature of the injection
moulding process which results in higher material compaction in
addition to the enhancement of crystalline structure, thus enhanc-
ing mechanical strength.
However, this study shows that the adequate selection of FDM
parameters can yield parts of acceptable properties. Here, primar-
ily the raster gap is a crucial parameter where at a negative gap of
−0.05 mm, the material becomes denser almost reaching density of
injected parts. This high packing of the material makes the effect of
changing raster angle insignificant for static testing. Thus, a proper
selection of the raster angle with respect to the loading direction
becomes especially important when printing at positive raster gaps.
Here, it is to be noted that rasters lying parallel to the loading
direction will be able to carry all load in contrast to filaments in
transverse direction that insignificantly contribute to load bearing.
In contrast, however, dynamic behaviour is significantly affected
by the raster angle regardless of the raster gap direction. This
behaviour was related to internal stresses caused by thermal gra-
dients during the printing stage.
As depicted in Fig. 9, parts fabricated according to the −P45 prin-
ting conditions are able to reach 91% of the tensile strength and
86% the flexural strength of injection moulded products. This slight
reduction in strength often lies within the acceptable range for sim-
ple DIY parts, proving the applicability of self-made 3D printers for
personal use. Summarizing, a negative raster gap in combination
with a 45◦ raster angle is recommended if a high density part with
enhanced mechanical properties is desired.
Next to mechanical performance, the FDM technique applied in
this study proved acceptable dimensional accuracy, lying within
the nominal range of the specified size. Further, no significant
dimensional variation was observed with altering raster angle and
gap.
References
[1] Kamrani AK, Abouel Nasr E. Engineering Design and Rapid Prototyping. New
York, US: Springer; 2010.
[2] Sood AK, Ohdar RK, Mahapatra SS. Parametric appraisal of mechanical property
of fused deposition modelling processed parts. Mater Des 2010;31:287–95,
http://dx.doi.org/10.1016/j.matdes.2009.06.016.
[3] Pham D, Gault R. A comparison of rapid prototyping technologies. Int
J Mach Tools Manuf 1998;38:1257–87, http://dx.doi.org/10.1016/S0890-
6955(97)00137-5.
45
[4] Kumar S, Kruth JP. Composites by rapid prototyping technology. Mater Des
2010;31:850–6, http://dx.doi.org/10.1016/j.matdes.2009.07.045.
[5] Reprap http://reprap.org/ (accessed September 5, 2014).
[6] Rodríguez JF, Thomas JP, Renaud JE. Mechanical behavior of acrylonitrile
butadiene styrene fused deposition materials modeling. Rapid Prototyp J
2003;9:219–30, http://dx.doi.org/10.1108/13552540310489604.
[7] Sood AK, Ohdar RK, Mahapatra SS. Experimental investigation and empirical
modelling of FDM process for compressive strength improvement. J Adv Res
2012;3:81–90, http://dx.doi.org/10.1016/j.jare.2011.05.001.
[8] Vega V, Clements J, Lam T, Abad A, Fritz B, Ula N, et al. The effect of
layer orientation on the mechanical properties and microstructure of a
polymer. J Mater Eng Perform 2011;20:978–88, http://dx.doi.org/10.1007/
s11665-010-9740-z.
[9] Ahn S-H, Montero M, Odell D, Roundy S, Wright PK. Anisotropic material prop-
erties of fused deposition modeling ABS. Rapid Prototyp J 2002;8:248–57,
http://dx.doi.org/10.1108/13552540210441166.
[10] Croccolo D, De Agostinis M, Olmi G. Experimental characterization and
analytical modelling of the mechanical behaviour of fused deposition
processed parts made of ABS-M30. Comput Mater Sci 2013;79:506–18,
http://dx.doi.org/10.1016/j.commatsci.2013.06.041.
[11] Repitier Host http://www.repetier.com/ (accessed February 1, 2014).
[12] Montgomery DC. Design and Analysis of Experiments. 7th ed. New Jersy: John
Wiley & Sons; 2008.
[13] Slic3r http://slic3r.org/ (accessed February 1, 2014).
[14] Grand Pacific Petrochemical Corporation http://www.gppc.com.tw (accessed
May 28, 2014).
[15] Montero M, Roundy S, Odell D. Material characterization of fused deposition
modeling (FDM) ABS by designed experiments. Proc Rapid Prototyp Manuf Conf
2001;(1):1–21.
[16] Rodrıı́guez JF, Thomas JP, Renaud JE. Design of fused-deposition ABS compo-
nents for stiffness and strength. J Mech Des 2003;125:545, http://dx.doi.org/
10.1115/1.1582499.
[17] Lee BH, Abdullah J, Khan ZA. Optimization of rapid prototyping parameters for
production of flexible ABS object. J Mater Process Technol 2005;169:54–61,
http://dx.doi.org/10.1016/j.jmatprotec.2005.02.259.
[18] Agarwala MK, Jamalabad Vikram R, Langrana Noshir A, Ahmad Safari,
Whalen Philip J, Danforth Stephen C. Structural quality of parts processed
by fused deposition. Rapid Prototyp J 1996;2:4–19, http://dx.doi.org/10.1108/
13552549610732034.
[19] Anitha R, Arunachalam S, Radhakrishnan P. Critical parameters influencing the
quality of prototypes in fused deposition modelling. J Mater Process Technol
2001;118:385–8, http://dx.doi.org/10.1016/S0924-0136(01)00980-3.
[20] Bakar NSA, Alkahari MR, Boejang H. Analysis on fused deposition modelling
performance. J Zhejiang Univ Sci A 2010;11:972–7, http://dx.doi.org/10.1631/
jzus.A1001365.
[21] Ziemian C, Sharma M. Anisotropic mechanical properties of ABS
parts fabricated by fused deposition modelling. Mech Eng 2012;2:3,
http://dx.doi.org/10.5772/34233.
[22] Ziemian C, Cipoletti DE, Ziemian SN, Okwara MN, Haile K V. Monotonic and
Cyclic Tensile Properties of ABS Components Fabricated by Additive Manufac-
turing:525–41.
[23] Sood AK, Ohdar RK, Mahapatra SS. Improving dimensional accuracy of Fused
Deposition Modelling processed part using grey Taguchi method. Mater Des
2009;30:4243–52, http://dx.doi.org/10.1016/j.matdes.2009.04.030.