![The principle of large-scale 3D printer based on fused deposition of plastic pellets. [Color figure can be viewed at wileyonlinelibrary.com]](https://www.researchgate.net/profile/Jiao-Zhiwei-3/publication/315954363/figure/fig1/AS:874591664754690@1585530310650/The-principle-of-large-scale-3D-printer-based-on-fused-deposition-of-plastic-pellets_Q320.jpg)
Content uploaded by Jiao Zhiwei
Author content
All content in this area was uploaded by Jiao Zhiwei on Mar 30, 2020
Content may be subject to copyright.
A large-scale double-stage-screw 3D printer for fused deposition
of plastic pellets
Xiaojun Liu,
1
Baihong Chi,
2
Zhiwei Jiao,
1
Jing Tan,
1
Fengfeng Liu,
1
Weimin Yang
1
1
College of Mechanical and Electrical Engineering, Beijing University of Chemical Technology, Beijing 100029, People’s Republic
of China
2
Space Star Technology Co., Ltd., Beijing 100086, People’s Republic of China
Correspondence to: W. Yang (E-mail: yangwm1965@163.com) and Z. Jiao (E-mail: jiaozw@mail.buct.edu.cn)
ABSTRACT: Fused deposition molding (FDM) is the most popular technology in the fields of three-dimensional printing, but it is hard to
use a variety of plastic materials due to the limitation of filament form of material. Using plastic pellets as printing materials gives advan-
tages in cost, processing speed, and available materials. In this work, a large-scale double-screw FDM three-dimensional printer based on
plastic pellets has been designed. It is capable of printing large plastic products at a low cost and high speed. Using ABS110%GF as
printing material, this work is first focused on the effects of the pressure and speed of the metering screw on the flowrate of melt. The
equation for the relationship of these three parameters was established as well. Based on this equation, the effects of melt flow, printing
speed, and layer thickness on the width of fused filament were investigated with experiments. Furthermore, the effects of printing spacing
between fused filaments on surface accuracy and bonding strength were also explored. By printing models, it was revealed that the
designed printer is able to print products with plastic pellets. V
C2017 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2017,134, 45147.
KEYWORDS: bonding strength; FDM; melt flow; plastic pellets; surface accuracy; width of fused filament
Received 13 January 2017; accepted 23 March 2017
DOI: 10.1002/app.45147
INTRODUCTION
As an example of additive manufacturing (AM), three-
dimensional printing has become a well-known technique,
which can be used in aerospace,
1–3
tissue engineering,
4–6
fashion
design,
7,8
equipment manufacturing,
9,10
and other applica-
tions.
11–13
Currently, the processes of 3D printing are researched
and developed in various directions,
14
and are bringing huge
benefits to research institutions and enterprises. In order to
apply 3D printing to more fields and to change the traditional
manufacturing model,
15,16
more practical research should be
done for obtaining the superiority in the innovating wave of 3D
printing. In the casting and molding industry, the mold is the
core part of the process. As single-piece and small batch sizes
production have become a trend in modern production systems,
they require multistage processing molds in order to satisfy the
complex structure of the products, resulting in great wastage of
energy and raw materials.
17
Three-dimensional printing can
realize the advantages of processing parts with high geometrical
complexity, near net shape fabricating without using dies or
molds and reducing the cost and time of production.
18
Therefore, processing mold via 3D printing has become a major
concern.
19
Three-dimensional printing processes have a lot in common.
However, a wide range of discretely different processes are avail-
able under the AM umbrella. American Society for Testing and
Materials (ASTM) International Committee approved the AM
process categories: material extrusion, material jetting, glue
injection, UV curving, laminated object manufacturing, powder
bed melting, etc.
20
Fused deposition manufacturing (FDM) is
the most popular technology of 3D printing. In FDM, a fila-
ment of material is fed into the melting system of 3D printer
via a pinch roller. The feedstock is melted in a heated liquefier
with the solid portion of the filament acting as a piston to push
the melt through a print nozzle. A gantry moves the print noz-
zle in the horizontal x–yplane as the material is deposited on a
build surface that can be moved in the vertical zdirection.
21
However, the feedstock must meet the viscosity and rigidity
requirements to ensure that the melt squeezes out from the noz-
zle and to avoid salivation.
22
Due to the demands of stiffness
This article was published online on 6 April 2017. An updaet was subsequently identified for the affiliation and Acknowledgments.
This notice is included in the online and print versions to indicate that both have been corrected 13 April 2017.
V
C2017 Wiley Periodicals, Inc.
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (1 of 9)
and strength as well as the increased cost of the filament mak-
ing process, a large variety of materials are inapplicable to fila-
ment based printing. Therefore, directly using plastic pellets to
print has many advantages, such as low cost, fast processing
speed and widely available materials.
Efforts have been made to use plastic pellets for 3D printing.
Venkataraman et al.
23
designed a melt deposition 3D printer
with a screw extruder, which can print with polypropylene
(PP), polyethylene (PE), and polystyrene (PS). Volpato et al.
24
designed a plunger-type 3D printer which is capable of using
polymer pellets. They also analyzed the extrusion system from
terms of material degradation, fuse continuity, product scale
and surface precision. Wang
25
developed a high speed extrusion
equipment for fused deposition of pellet materials. In order to
use pellet polyurethane for 3D printing, Li et al.
26
designed a
series of experiments for the effects of processing parameters on
the forming process and mechanical properties of support
structure, and obtained the optimal range of these parameters
for manufacturing auricular cartilage scaffold.
The 3D printer using plastic pellets has been applied in some rele-
vant fields. The Arbug company developed the 3D printer-
Freeformer for manufacturing single or small batches of standard
particles without using a mold.
27
We previously developed the
polymer melt differential 3D printer that can use some plastic and
elastic pellet materials for printing, such as polylactic acid (PLA),
acrylonitrile butadiene styrene (ABS), and thermoplastic polyure-
thane (TPU).
28
However, both the above 3D printers are only able
to print products whose length and width are less than 30 cm,
which is much smaller than that of the molds in industry.
29
As for
ordinary 3D printers whose material is in filament form, it may
cost much more to print large product due to the preparation of
filaments. Therefore, developing a 3D printer which is capable of
printing large products with pellets would greatly reduce cost.
This work designed a large-scale melt deposition 3D printer
based on polymer pellet, which combines a double-stage screw
extrusion device with a large 3D platform. The processing
parameters were analyzed to determine its working conditions.
For practical applications, this printer was also tested by print-
ing models, which indicates that it can print large plastic molds
and products under the mutual action of the double-stages
screw to achieve the purpose of high speed and low cost
manufacturing.
DESIGN OF 3D PRINTER FOR LARGE PRODUCTS WITH
PELLET MATERIAL
Working Principle
Previous pellet 3D printing device has several disadvantages:
1. Material transfer is not stable, for example, “bridge” phe-
nomenon would appear.
2. The polymers are easy to be mixed with air in plasticizing
process.
3. Uncontrollable salivation at nozzle would cause overflow,
which may result in uneven quality of products.
These problems limit the extensive use of pellet-material 3D
printer. This work adopted the mechanism of double-stage-
screw extrusion as a pellet extrusion system. Such design has a
more reasonable distribution of power consumption than the
one screw system. The first-stage screw has a relatively large
diameter and is used to melt and convey materials as well as to
provide pressure. The second-stage screw (metering screw) is
smaller and is used for further plasticizing and measurement of
the melt. One of the pressure sensors is located between the
two screws to monitor and control the real-time pressure by
changing the rotating speed of the first-stage screw. There is a
vent hole located at the connection position of the two parts
from which gas can be squeezed out in case of bubbles being
embedded into the extrusion. As illustrated in Figure 1, the
whole printing process includes a melt generation unit, a 3D
model informing unit and a process control unit. The melt gen-
eration unit is mainly used for transforming plastic pellets into
melt and establishing pressure. The plastic pellets are heated to
melt by the heating sleeve fixed on the cylinder and then
Figure 1. The principle of large-scale 3D printer based on fused deposition of plastic pellets. [Color figure can be viewed at wileyonlinelibrary.com]
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (2 of 9)
transported to the front-end by the rotated screw. Because the
spiral groove depth of the front-end of the screw is shallower
than that of the back-end, a large and steady pressure is estab-
lished during the whole process. At the same time, another driv-
ing motor drives the metering screw to rotate and extrudes the
melt through the nozzle to print products.
Design of Experimental Device
According to the processing principle, a 3D printer incorporat-
ing double-stage screws is prepared. The prototype of the device
is shown in Figure 2. The hopper is 30 L, ensuring printing for
more than 3 h per feed. The diameter and length-diameter ratio
of the first-stage screw are 25 mm and 20:1, respectively. These
parameters of the second-stage screw (the metering screw) are
16 mm and 6:1 respectively. These two screws have maximum
speeds of 100 and 120 rpm respectively, and the second one is
able to rotate forward or afterward under the control of pulse
signal. A pressure sensor is arranged at the connecting point of
the two screws, and the speed of the first-stage screw is con-
nected with the setting pressure. The internal and external
diameter of the nozzle is 4 mm and 8 mm respectively.
The 3D motion mechanism is modified by a vertical milling
machine, and the extrusion equipment is fixed on the Zaxis.
The platform can move along Xdirection while the extruder
can move along Yand Zdirection, which forms the 3D motion
mechanism of the printer. The printing space is 800 3600 3
600 mm
3
, and the surface temperature of the platform can be
heated up to 120 8C.
Theoretical Analysis of Accumulation Process
According to Jin et al.,
30
as illustrated in Figure 3, the cross-
section of extruded filament can be regarded as the combination
of two semicircles and a rectangle, where d
i
is the inner size of
nozzle, his layer thickness in the printing process, xis the
width of deposited filament, Eis the printing spacing between
two lines, Dcis the overlapping part of two filaments. In the
process of deposition, the width of the filament and printing
spacing between them determine the precision and strength of
products. Therefore, it is of great importance to determine the
effects of process parameters on the width of the fused filament.
The main factors that affect the size of the filament are melt
flow, printing speed and layer thickness, among which the most
critical one is melt flow. The flow rate of melt is influenced by
the structural parameters of the device, the processing parame-
ters in printing process, and the rheological properties of the
printing material. According to Hagen-Poiseuille equation,
31
Figure 2. (a) The picture of two stages screw extrusion system; (b) The complete picture of large-scale 3D printer. [Color figure can be viewed at
wileyonlinelibrary.com]
Figure 3. The schematic diagrams of micro unit of extruded filaments
and their accumulation. [Color figure can be viewed at wileyonlinelibrary.
com]
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (3 of 9)
Q5pd4
128lLDp(1)
where Qis the flow rate of melt, dis the inner diameter of noz-
zle, lis the dynamic viscosity of melt, Lis the length of nozzle,
and Dpis the pressure difference between the inside and outside
of the nozzle. For non-Newtonian polymeric fluid, the viscosity
is greatly affected by temperature. The melt has a high viscosity
at a lower temperature, and requires a high pressure to make it
extruded out of the nozzle. If other conditions are fixed, the
flow rate of melt can be controlled by adjusting the pressure of
melt and the rotating speed of the metering screw.
DESIGN OF EXPERIMENTS
Thermal shrinkage and deformation are the main factors that
influence the accuracy of products.
21
For convenience, we chose
ABS for experiments, since it has low thermal shrinkage and
good adhesion. In this work, ABS757-GF10 (ABS mixed 10%
glass fiber of mass fraction, Zhenjiang Qimei Chemical Co. Ltd)
was used as printing material. Its melt index and density are
3.8 g/10 min (230 8C) and 1.26 g/cm
3
, respectively. Its thermal
expansion rate is 0.33% (baked at 80 8C for 3 h, Labthink Lan-
guang, Ji’nan Languang mechanical and Electrical Technology
Co. Ltd.), which satisfies the requirements of FDM process that
linear shrinkage rate should be less than 1%.
32
Before printing,
the pellets are dried in vacuum drying oven for 2 h, and the
drying temperature is about 90 8C. The design of the experi-
ments are listed as follows:
1. Under the process parameters: melting temperature 5230 8C,
nozzle diameter 54 mm, We first studied the influence of the
back pressure in the cylinder which was set to 4, 5, 6, 7, 8, 9,
10 MPa respectively on the melt flow rate when the rotating
speed of the metering screw is 0, and then studied the influ-
ence of the rotating speed of the metering screw which was
set to 218, 215, 212, 26, 23, 0.3, 3, 6, 9, 12, 15, 18 rpm
respectively on the flow rate under the pressure of 5, 7, 9
MPa respectively to establish the equation of the relationship
between these three parameters. In these experiments, the
back pressure was controlled by the first-stage screw.
2. This work also serves to analyze the influence of melt flow
(Q), printing speed (v) and layer thickness (h) on the width
of fused filament (x) by printing three layers of square con-
tour with different process parameters, and to establish the
corresponding equation for the precise control of the fila-
ment width, then to analyze the reasons of failed printing.
3. In order to analyze the effects of printing spacing on the
mechanical properties of products, the rectangular blocks with
different printing spacing such as 2 mm, 2.5 mm, 3 mm, 3.5
mm, 4 mm were printed under one set of better process
parameters of h52.0 mm, Q55.61 g/min, v520 mm/s, and
then were cut into tensile samples for the test of their strength
and surface quality, as shown in Figure 4. The gauge length of
tensile test is 50 mm, and the loading speed of it is 20 mm/
min, when the specimen broken, the test ends.
4. The research above provides data for manufacturing large
3D printing products, but its validity requires confirmation.
In this article, two models are printed and the defects of
them are analyzed too. The first model that built in Creo is
the casting mold, the size of which is 180 3100 350 mm.
Due to the fact that the printed model must be post-
processed before using, 5 mm machining allowances are
added onto the model before printing so as to make the
product maintain the original size after processing. The pro-
cess parameters are set as follows to print the model: P57
MPa, the speed of metering screw n58 rpm when extrud-
ing and the speed of it v
r
5230 rpm when restraining,
printing speed v520 mm/s, layer thickness h52.0 mm,
printing spacing E52.5 mm, heating temperature of cylin-
der T
c
5230 8C, heating temperature of platform T
p
580 8C.
The other model printed on the printer is a large hollow
product with thin wall, whose maximum size is 650 3650
3800 mm and the wall thickness of which is 20 mm. Pro-
cess parameters are set as follows to print it: P57MPa,
when extruding, the speed of metering screw n513 rpm,
when restraining, the speed of it v
r
5230 rpm, printing
speed v530 mm/s, layer thickness h52.0 mm, printing
spacing E53 mm, heating temperature of cylinder
T
c
5230 8C, heating temperature of platform T
p
580 8C.
RESULTS AND DISCUSSION
Analysis of Melt Flow
Effect of Back Pressure on Extrusion Flow. Figure 5 shows the
extruded flow rate of melt with the function of various back
pressure when the speed of metering screw is 0. It reveals that
the flow rate linearly increases with the increase of back pres-
sure in the considered range. Due to the fixed diameter of noz-
zle, the diameter of the fuse is stable at 3.2 to 3.3 mm, so the
melt velocity also keeps linear relationship with the increase of
Figure 4. Preparation of tensile samples. [Color figure can be viewed at
wileyonlinelibrary.com]
Figure 5. Relationship between melt pressure and flow rate at the heating
temperature of 230 8C. [Color figure can be viewed at wileyonlinelibrary.
com]
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (4 of 9)
melt pressure. When the pressure is less than 4 MPa, the melt
flow is unstable, which is not suitable for printing. Based on the
experimental data, the equation of back pressure and melt flow
rate is fitted. It can be illustrated with eq. (2), in which the Pis
back pressure, the Qis melt flow rate.
Q50:59P20:61 (2)
Effect of the Speed of Metering Screw on Melt Flow Rate. The
relationships between the speed of metering screw and melt
flow rate are illustrated in Figure 6.
It can be seen from the results that the melt flow increases line-
arly with the increase of the speed of the metering screw at dif-
ferent pressure within a certain range due to the state of no slip
between melt and screw. The relationship among these three
parameters can be illustrated with eq. (3), in which the Qis
melt flow, the Pis the back pressure of cylinder, the nis the
speed of metering screw.
Q50:58P10:28n20:71 (3)
When a printing path ends, the nozzle needs to move quickly
from the end point to the start point of the other path. In the
process, the melt flow must be restrained to ensure mass unifor-
mity and surface accuracy of the printed products. Through the
reversal rotation of metering screw, the flow rate can be reduced
and controlled rapidly and effectively. So we set the speeds of
the metering screw to 23, 26, 212, 215, 218 rpm respectively
under various pressures to research the changes of melt flow
rate so as to confirm a better reverse rotating speed. The result
can be illustrated in Figure 7, from which it can be seen that
the reversal rotation of metering screw has a good effect on
restraining melt flow, but the relationship between them is non-
linear due to the gravity effect of the remaining melt in the noz-
zle. When the reversal rotation speed is higher than 18 rpm, the
slope of the curve decreases, which means that the melt flow
can be easily controlled. When the speed reaches up to 30 rpm,
the melt flow turns to stable, which means the effect of restrain-
ing melt is almost the best, though there is still little amount of
melt flowing out the nozzle, it has almost no effect on the final
product. The extruded samples under various speeds of
metering screw under the pressure of 7 MPa are illustrated in
Figure 8. The samples 1 to 14 are extruded with the speeds of
18, 15, 12, 9, 6, 3, 0.3, 23, 26, 29, 212, 215, 218, 230 rpm
of the metering screw respectively.
Width of Deposited Filament
By printing three layers of square contour with different param-
eters, the influence of h,v, and Qon xare researched. The
relationships among these four parameters are illustrated in Fig-
ure 9. It can be concluded that the xincreases with the increase
of Q, and decreases with the increase of hand v. When
h52.5 mm,Q57.29 g/min,v530 mm/s or Q55.61 g/min,
v530 mm/s, due to the higher printing speed and smaller
melt flow, the wire drawing phenomenon appears in the print-
ing process and it causes an unstable deposition of extruded fil-
aments, which results in poor bonding effect of the interface
between two filaments and failed printing, as illustrated in Fig-
ure 10. Therefore, to control the process parameters reasonably
is the key to gain good products.
Figure 6. Relationship between the speed of metering screw and flow rate
at the heating temperature of 230 8C. [Color figure can be viewed at
wileyonlinelibrary.com] Figure 7. Relationship between the reversal speed of metering screw and
melt flow rate at the heating temperature of 230 8C. [Color figure can be
viewed at wileyonlinelibrary.com]
Figure 8. The samples extruded with various speeds of the metering
screw. [Color figure can be viewed at wileyonlinelibrary.com]
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (5 of 9)
The details of printed models are illustrated in Figure 11(a)
shows the picture of printed model. The xin (b) shows the
width of deposited filament and it is measured by vernier cali-
per. Figure 11(c) shows the model printed under the process
parameters of v530 mm/s, Q57.29 g/min, h52.0 mm.
Though the model was formed, many defects were present such
as large spacing between layers and numerous gaps, resulting in
low strength and possibility of separation. So the parameters of
the model showed in (c) are not suitable for printing. Figure
11(d) shows the model printed under the process parameters of
v520 mm/s, Q58.97 g/min, h52.5 mm, it can be seen that
the extruded filaments are uniform, and the interface between
two filaments has bonded well with no gap, which is suitable
for printing. But when v510 mm/s, Q58.97 g/min,
h52.0 mm, the xis 8.4 mm, which is wider than the external
diameter of nozzle due to low printing speed, high melt flow
rate and low layer thickness, causing overflow and an uneven
surface.
Based on the experimental results above, it can be concluded
that to ensure products with good bonding strength, enough
bonding area between two filaments is of great importance, so
the layer thickness should be smaller than the diameter of
extruded filament. Furthermore, the layer thickness should be
less than 2.5 mm due to cooling and shrink of the filaments,
otherwise, it is difficult to ensure a good adhesion between fila-
ments. In addition, in order to ensure good surface accuracy
and strength of products, the width of deposited filament
should be more than the internal diameter and less than the
external diameter of the nozzle so as to avoid overflow but to
form an appropriate overlap between the two filaments.
Printing Accuracy and Bonding Strength
In the process of accumulation, there exists a temperature dif-
ference between the deposited filament and the newly extruded
molten filament. Therefore, the temperature of their interface
can reach melting temperature by heat conduction, which can
make the two filaments bond together when the temperature
drops.
22
Therefore, the strength of the products can be
enhanced by increasing the thermal conductivity of the inter-
face, which can be achieved by reducing the printing spacing
between two lines. However, smaller printing spacing will affect
the surface accuracy of the products, so the effects of different
printing spacing on surface accuracy and strength are discussed
below.
The rectangular models printed with the printing spacing of
4 mm, 3.5 mm, 3 mm, 2.5 mm, 2 mm are shown from (a) to
(e) in Figure 12, respectively. It can be concluded that when the
printing spacing is 4 mm, there is a clear gap between filaments
of the model, and it is hard to form tensile samples. When the
printing spacing is 3.5 mm, though the filaments can connect
with each other to a degree, some gaps still exist which weaken
the bonding strength and make it impossible to form tensile
samples. When printing spacing is 3 or 2.5 mm, both the filling
effect and surface precision are of good state. When the spacing
is 2 mm, filaments are bonded tightly with each other, but a
large number of irregular wrinkles appear on the surface of the
top layer due to the narrow distance between two lines, result-
ing in irregular shape of the model. The longitudinal sections of
the models are illustrated in Figure 13, from which it can be
Figure 9. Relationships between xand h, v,Q. [Color figure can be
viewed at wileyonlinelibrary.com]
Figure 10. Failed printing model due to unreasonable process parameters.
[Color figure can be viewed at wileyonlinelibrary.com]
Figure 11. The details of printed model. [Color figure can be viewed at
wileyonlinelibrary.com]
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (6 of 9)
seen that both internal filling rate and the height of model
increased with the reduction of printing spacing, and with the
smooth effect of nozzle, the surface accuracy was improved. But
when the spacing decreased to 2 mm, the surface quality
became worse due to the accumulation of redundant overflow.
Due to when the printing spacing is 3.5 mm or 4 mm, it is too
large to produce tensile samples, so we give up these two
groups. The tensile samples cut from other rectangles are tested
on a tensile testing machine and the results are illustrated in
Figure 14, from which it can be concluded that when the print-
ing spacing is 3 mm, the average tensile strength and tensile
modulus of the samples are 5.62 MPa and 534.6 MPa respec-
tively, and the tensile strength is only 13% of the injection
molding parts.
33
When the printing spacing is 2.5 mm, the ten-
sile strength is slightly less than that of the sample with the
spacing of 2 mm, which indicates that when the printing spac-
ing is less than 2.5 mm, the reduction of it cannot improve the
strength effectively anymore. In the case of similar tensile
strength, the modulus of the samples printed with 2 mm spac-
ing is higher than that printed with 2.5 mm, and it can be
explained that small printing spacing results in an dense
arrangement of filaments and a high crystallinity, which
decreases the tensile strain and increases the tensile modulus.
Based on the research above, it can be concluded that under the
process parameters of h52.0 mm, Q55.61 g/min, v520 mm/
s, considering both accuracy and strength, the best printing
spacing is 2.5 mm.
Printing Test of Products
As shown in Figure 15, section (a) is the casting male mold
built in Croe, sections (b) and (d) are the printed model. It can
be seen from (b) that there exists overflow material in the cor-
ner and the positions where the size is small, which can be
explained by the fact that when the nozzle is around the corner,
the printing speed is reduced, while at the same time, due to
the lag of flow control, redundant material remains on the
Figure 12. Models printed with various printing spacing. (a–e) Models printed with the printing spacing of 4 mm, 3.5 mm, 3 mm, 2.5 mm, 2 mm.
[Color figure can be viewed at wileyonlinelibrary.com]
Figure 13. The longitudinal sections of the models printed with various
spacing. (a–e) Longitudinal sections with the printing spacing of 4 mm,
3.5 mm, 3 mm, 2.5 mm, 2 mm. [Color figure can be viewed at wileyonli-
nelibrary.com]
Figure 14. Relationships among printing spacing, tensile strength, and
tensile modulus. [Color figure can be viewed at wileyonlinelibrary.com]
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (7 of 9)
surface of the model. Section (c) in Figure 15 is the printed
model dealing with the milling process. Though the surface
roughness and the overall size of the model can meet design
requirements, there are some tiny cracks in the local area, rea-
sons of which are analyzed as follows: (1) unstable melt flow
which was caused by the pressure fluctuation of melt in the cyl-
inder resulted in sparse fill in local position of the model. (2)
In the milling process, local location position cracked under the
milling force due to the weak bonding. (3) The model is small
and has some small details, which are not suitable to be printed
with the nozzle of 4 mm diameter. Therefore, controlling melt
flow precisely and selecting process parameters reasonably are of
great importance to manufacture large scale products.
In addition, as shown in Figure 16, the large hollow product with
thin wall was printed and it is found that the overall size meet the
demands of design requirements by measurement. The facial qual-
ity is better by adjusting process parameters, however, there is a
small collapse on the surface, the possible reasons of which are
explained as follows: (1) the instability of melt flow results in
facial defect. (2) The overhang angle inside the model is greater
than 458, which makes it hard to stack in the vertical direction.
Summarizing the research above, we can conclude the models
can be printed successfully on the printer by parameter optimiza-
tion, but there are still some problems such as unstable melt flow
and lag of flow control. Therefore, choosing a pressure stabilized
extruder and extruding system that can control melt flow precise-
ly has become the key point of FDM for large products. The
pressure of cylinder has a great influence on the melt flow due to
the fact that there is no seal between the end of the extruder and
the nozzle when the metering screw is stationary, so the pressure
fluctuation will result in an uneven mass distribution of products.
Therefore, it is better to cut off the relationship between the pres-
sure of cylinder and melt flow. That means, when metering screw
is stationary, there is no material flowing out of the nozzle, mak-
ing the pressure only provide power to feed material to the
metering screw, thus the melt flow can be controlled well by the
metering screw independently.
CONCLUSIONS
In this article, a large-scale 3D printer based on fused deposi-
tion of plastic pellets was designed. It consisted of a double-
stage screw extruder and a large 3D forming platform. The
machine developed was found capable of printing large plastic
molds and products at a low cost and rapid speed. In this arti-
cle, the working principle and main structure of the printer was
introduced, and the physical model of the accumulation process
was analyzed. In addition, based on the accumulation model,
some process parameters and their effects on surface accuracy
and strength were researched and analyzed. Finally, two models
were printed to test the feasibility of the printer and the reason
of defects in products were analyzed. In conclusion, the paper
can be summarized as follows:
1. By using ABS757-GF10 as the printing material, factors affect-
ing melt flow rate were studied, and the relationship among
melt flow (Q), pressure (P), and the speed of the second-
stage screw (n) was analyzed. Through experimental trials,
equation among these three parameters were established:
Q50:58 P10:28n20:71, Pranged from 4 to 10 MPa,
nranged from 0 to 18 rpm. When the second-stage screw
was reversed, the melt flow could be reduced greatly, which
provides accurate flow control with the methods.
2. On the base of the flow control equations, the effects of layer
thickness (h), printing speed (v), and melt flow (Q) on width
of fused filament (x) were investigated. The experimental
result showed that the xincreased with the increase of Q,
and decreased with the increase of hand v.Toensurethe
products with good bonding strength, enough bonding area
between two filaments is needed, therefore, selecting reason-
able process parameters to create an overlap between fila-
ments is vital. The xshould be more than internal diameter
and less than external diameter of the nozzle so as to avoid
overflow and to form overlapping part of two filaments.
3. The effects of printing spacing EðÞon surface accuracy and
bonding strength were researched, the results showed that
under the process parameters of h52.0 mm, Q55.61 g/min,
v520 mm/s, both internal filling rate and the height of mod-
el increased with the reduction of E, and with the smooth
effect of nozzle, the surface accuracy was improved. But when
Edecreased to 2 mm, the surface quality became worse due to
the accumulation of redundant overflow. It also showed that
Figure 15. The casting male mold (a) and its 3D printed model (b, d)
and postprocessed model (c). [Color figure can be viewed at wileyonlineli-
brary.com]
Figure 16. (a) Large hollow product with thin wall and (b) its 3D printed
model. Inner part of the oval in (b) is the collapse on the surface of prod-
uct. [Color figure can be viewed at wileyonlinelibrary.com]
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (8 of 9)
when Ewas greater than x, there was no bonding strength
between filaments, but when Ewas less than x, both tensile
strength and tensile modulus increased with the decrease of E.
The strength of samples printed with E52 mm was a little bit
greater than that printed with E52.5 mm, that means when E
is less than 2.5 mm, the reduction of it cannot improve the
strength effectively anymore.
4. Two casting molds were printed to test the properties of the
printer, and results showed that models could be printed
successfully on the printer by parameter optimization, but
there were still some problems such as unstable melt flow
and lag of flow control. Therefore, choosing a pressure stabi-
lized extruder and extruding system which can control melt
flow precisely has become the key point of pellets FDM 3D
printer for forming large products.
This article proved the feasibility of printing plastic products
with pellets materials on a large FDM 3D printer, which can
achieve printing large casting molds with fast speed and low
cost. But there exist some shortcomings, so the technology in
this field requires to be researched further.
ACKNOWLEDGMENTS
This work was financially supported by the National Natural Sci-
ence Foundation of China (Grant No. 51403014) and Guangdong
Provincial Plan Projects of Science and Technology (Grant No.
2016B090915001)
REFERENCES
1. Gockel, J.; Fox, J.; Beuth, J.; Hafley, R. Mater. Sci. Technol.
2015,31, 912.
2. Palm, F. Presented at the 27th Advanced Aerospace Materi-
als and Processes (AeroMat) Conference and Exposition,
May 23-26, ASM, 2016.
3. Gockel, J.; Beuth, J.; Taminger, K. Additive Manuf. 2014, 1: 119.
4. Wang, X.; Xu, S.; Zhou, S.; Xu, W.; Learyb, M.; Choongc, P.
Biomaterials 2016,83, 127.
5. Kolesky, D. B.; Truby, R. L.; Gladman, A. S.; Busbee, T. A.;
Homan, K. A. Adv. Mater. 2014,26, 3124.
6. Ravari, M. K.; Kadkhodaei, M.; Badrossamay, M.; Rezaei, R.
Int. J. Mech. Sci. 2014,88, 154.
7. Hudson, S. E. Presented at the Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems.
ACM, 2014; p 459.
8. Xu, H.; Li, Y.; Chen, Y.; Barbic, J. ACM Trans Graph. 2015,34,1.
9. Rayegani, F.; Onwubolu, G. C. Int. J. Adv. Manuf. Technol.
2014,73, 509.
10. Sachs, E.; Wylonis, E.; Allen, S.; Cima, M.; Guo, H. Polym.
Eng. Sci. 2000,40, 1232.
11. Bartlett, N. W.; Tolley, M. T.; Overvelde, J. T. B.; Weaver, J.
C.; Mosadegh, B.; Bertoldi, K. Science 2015,349, 161.
12. Zhang, D.; Chi, B. H.; Li, B. W.; Gao, Z. W.; Du, Y.; Guo, J.
B. Synth. Met. 2016,217, 79.
13. Walters, P.; McGoran, D. Soc. Imaging Sci. Technol. 2011,
2011, 185.
14. Wohlers, T.; Caffery, T. Annu. Worldwide Prog. Rep. 2015.
15.Petrick,I.J.;Simpson,T.W.Res. Technol. Manage. 2013,56, 12.
16. Lipson, H.; Kurman, M. Fabricated: The New World of 3D
Printing; John Wiley & Sons, 2013.
17. Liang, J.-C.; Gao, S.; Teng, F.; Yu, P.-Z.; Song, X.-J. Int. J.
Adv. Manuf. Technol. 2014,71, 1939.
18. Campbell, I.; Bourell, D.; Gibson, I. Rapid Prototyping J.
2012,18, 255.
19. Bassoli, E.; Gatto, A.; Iuliano, L.; Violante, M. G. Rapid Pro-
totyping J. 2007,13, 148.
20. Caffrey, T.; Wohlers, T. Wohlers Rep. 2015.
21. Turner, B. N.; Gold, S. A. Rapid Prototyping J. 2015,21, 250.
22. Turner, B. N.; Strong, R.; Gold, S. A. Process Des. Model.
2014,20, 192.
23. Venkataraman, N.; Rangarajan, S.; Matthewson, M. J.;
Harper, B.; Safari, A.; Danforth, S. G. Rapid Prototyping J.
2013,6, 244.
24. Volpato, N.; Kretschek, D.; Foggiatto, J. Int. J. Adv. Manuf.
Technol. 2015,81, 1519.
25.Wang,T.HighSpeedExtrusionEquipmentBasedon
Fused Deposition of Bulk Material in Granulated Form
and Study on the FDM Process; Shanghai Jiaotong Uni-
versity, 2006.
26. Li, S.; Yan, Y.; Qin, H. J. Mech. Eng. 2010,46, 139.
27. Batra, R. C. Popular Plast. Pack. 2014, 18.
28. Yang, W. M.; Chi, B. H.; Gao, X. D.; Tan, J.; Jiao, Z. W.
Plastic 2016,45, 70.
29. Zhu, J. L.; Zhao, H. G.; Wu, G. Ind. Appl. Commun. 2016,
35, 115.
30. Jin, Y. A.; Li, H.; He, Y.; Fu, J. Z. Additive Manuf. 2015,8,
142.
31. Shu, X. Y.; Zhang, H. H.; Liu, H. Y.; Xie, D.; Xiao, J. F. Mas-
ter Thesis, Huazhong University of Science and Technology,
2005.
32. Wang, T. M.; Xi, J. T.; Jin, Y. Int. J. Adv. Manuf. Technol.
2007,33, 1087.
33. Chi, B. H.; Xie, L. Y.; Gao, X. D.; Jiao, Z. W.; Yang, W. M.
Plastic 2015,44, 40.
ARTICLE WILEYONLINELIBRARY.COM/APP
WWW.MATERIALSVIEWS.COM J. APPL. POLYM. SCI. 2017, DOI: 10.1002/APP.4514745147 (9 of 9)
