Documente Academic
Documente Profesional
Documente Cultură
doi: 10.14355/ijics.2015.04.001
www.iji-cs.org
University of Maribor, Faculty of Mechanical Engineering, Department of Textile Materials and Design
Abstract
In this article, the shape of fused deposition modelling (FDM) 3D printed testing objects is verified with 3D scanning technique.
Leapfrogs Creatr is used as a manufacturing device. The whole process of a materialization is explained from preparation of a
model to preparation of a printer. The study is meant to evaluate dimensional errors of printed parts and to give some advices
on how to minimize them.
Keywords
FDM 3D Printing; Leapfrog Creatr; Comparison; Parameter Confirmation; Verification
Introduction
Fused deposition modelling (FDM) or an equivalent term fused filament fabrication (FFF) is entry level additive
manufacturing technology of 3D printing that has widespread over the world, since 2009 the patent for that
technology expired [7], [12].
RepRap was the first promising open-source project, which idea was to build the first general-purpose selfreplicating manufacturing machine [9]. It wanted to bring 3D printing to the masses and to start new thinking of
production to download a virtual model and materialize it with a home 3D printer. To materialize means to make
a real model that can be touched and held in the hand. Nowadays, low cost FDM 3D printers become even cheaper
and more affordable, however, mostly enthusiasts and geeks use them. Due to the fact, that the information on
how to print and what to print is scattered over the internet and is time consuming to find it. Usually it takes much
of technical knowledge to understand and prepare the machine. For example, different parameters for different 3D
printing devices can be found. Sometimes some parameters work for some users machines and for others simply
do not. That makes this technology still experimental. On the other side, it is possible to buy systems that should be
more reliable but they are also more expensive. These are closed systems with closed working chamber, where
only manufacturers filament material can be used. That means same input quality, but it is usually more expensive
because only one provider has monopole to deliver the filament.
For an experiment in this report, Leapfrogs Creatr 3D printer was used. It is a pre-calibrated, well build system.
With the price of about 1500, it is not the cheapest, but it offers much for that price.
The deviation of the printed part is sometimes crucial for a functional prototype. Therefore, the accuracy or rather
the exactness of 3D printed parts was verified and reported in this paper. The results show the tendency of
deviations, and present recommendations for further modelling and research for more accurate 3D prints.
Methodology
Process of materialization of the object demands several preparation steps. The term preparation will be used
because the whole process is observed; from the idea to the real object, where the following steps are included:
3D modelling
www.iji-cs.org
3D printer preparation.
Polygonal modelling
Digital sculpting
Subdivision modelling
Detailed mathematical background for several techniques can be found in [3]. It is important to know that
modelling techniques should produce 3D models, suitable for 3D printing. However, this suitability can be verified
in the next step.
Printing Preparations stl
Newly created virtual 3D model has to be saved in the proper format. Usually, computer file formats .stl, .obj or
.3ds are used. That are meshes that describe 3D modelled object. Such mesh has to be watertight, without gaps or
overlaps. All the triangle normal vectors should be directed outward of the object. If those requirements are not
fulfilled, a .stl file has to be repaired [6]. For this purpose, freeware program Netfabb Basic [6] can be used. Open
source program Meshlab [5] can repair or even rebuild the .stl file into the correct mesh as well. The main geometry
of the main mesh can be changed. If errors of the input mesh are grave and at delicate places, this change is
unacceptable. Then modelling preparation step has to be repeated to improve the 3D model for printing.
Printing Preparations G-Code
Corrected .stl mesh has to be sliced and converted to G-code. G-code is a numerical control programming language
[10], used to control XYZ robot, which 3D printer is a kind of. There are many programs to produce that code from
the 3D object, but plenty of them are quite expensive. Slic3r [11] is a free available G-code generating program that
works implemented in Leapfrog Repetier [8] and fits to the 3D printer. Makerbots Cura [13] is a freeware as well.
It was also used, but some adjustments for proper printing with Creatr were needed. Repetier was used for direct
3D printer control and both Cura and Slic3r were used for G-code creation. Their user interfaces (UI) are shown in
figures 1 and 2.
www.iji-cs.org
Caliper and
Model One
3D model having several construction shapes, shown in figure 3, was created using the 3D modelling software
Solid Works and printed with the 3D printer with different layer properties. Then, prints were 3D scanned and
compared to the original virtual model. Color maps of deviations and parameters are shown in table 1.
Our first test model is shown in figure 3. On a cylindrical plate having the diameter of 50mm and 5mm high were:
www.iji-cs.org
Cura 14.01 was adjusted with printing parameters from Repetier Leapfrog V0.90C. Mainly, the layer height, wall
thickness and number of bottom and top layers were adjusted. Manufacturer provided two types of printing with
Repetier:
In this showcase, the manufacturers printing parameters have been proved, and two finer layer thicknesses; 0,1
and 0,06 mm were added. However, the most important part is the shape validation. Test models were 3D scanned
and compared to the CAD model. Gom ATOS II 400 3D optical scanner with measuring volume 135mm was used.
It means that the package of lens to measure the objects with approximation of the virtual measuring cube with
edge length 135mm was placed. One of the researchers [2] already tested its usage and accuracy.
Predicted time for the same G-code differs in programs and also in versions of programs. First, Cura 14.01 was
used, and printing times are shown in table 1. New version, Cura 14.03, was released, where the predicted printing
times differed for the same object. However, there are almost the same predicted times in Repetier for ready to
print G-code in both versions of Cura, which means mere time calculation was adjusted in different Cura versions.
Measured printing times were different at all printing sessions, but they were all in the range under +-9%.
The same object was sliced with both programs as shown in figure 5. Repetiers Slicr produces G-code with more
dead moves presented with light blue lines on the left side of the figure 5. Curas slicer Steam Engine minimizes
dead moves, which is why the printing process was finished in a shorter time. Cura also generates the G-code in a
shorter time. However, it works well only on objects without small details, which is shown in figure 6, where both
covers are from the same virtual object, but different programs produced the G-code.
www.iji-cs.org
Layer
Thickness
[mm]
Predicted
time (Cura
14.01)
Predicted
time (Cura
14.03)
0,06
1h 49 min
1h 58 min
Predicted
time
(Repetier
0.90c)
1 h 47min
Measured
time [min]
Time difference
(Repetier/measured)
Chamber
temperature
[C]
1h 51min 27s
+8%
26,2
Fig. 4a: Deviation at layer 0,06 mm, Shell thickness: 0,3mm and Bottom/Top thickness: 0,3mm
0,1
59min
1h 4min
58min
53min 7s
-9%
25,5
Fig. 4b: Deviation at layer 0,1 mm, Shell thickness: 0,4mm and Bottom/Top thickness: 0,4mm
0,2
0,35
41min
46min
41min
45min 14s
+9%
26,5
Fig. 4c: Deviation at layer 0,2 mm, Shell thickness: 0,8mm and Bottom/Top thickness: 0,8mm
29min
31min
29min
30min 6s
+3%
26,5
Fig. 4d: Deviation at layer 0,35 mm, Shell thickness: 0,7mm and Bottom/Top thickness: 0,7mm
Model Two
The second model was a simplified pipe, shown in figure 7, because measurements could have been repeated, and
its arithmetic average calculated. Unlike the first test model, the exactness of 3D print inner geometry the hole
and outer geometry the cylinder can be determined. The results are shown in table 2.
www.iji-cs.org
FIG. 5: REPETIER SLIC3RS G-CODE ON THE RIGHT AND CURAS G-CODE ON THE LEFT.
FIG. 6: SAME OBJECT, LEFT 3D PRINT CREATED WITH CURAS G-CODE AND RIGHT WITH REPETIERS G-CODE.
Print No.:
II
III
IV
VI
VII
VIII
G-code software
Cura
Cura
Cura
Cura
Repetier
Repetier
Repetier
Repetier
0,35
0,35
0,35
0,06
0,2
0,35
0,35
0,35
215
215
215
215
190
190
190
190
42min
12min 19s
29min 17s
16min 51s
16min 51s
16min 51s
Printing time
46min 36s
12min 32s
28min 26s
16min 42s
16min 51s
16min 36s
Outer diameter
[mm]
19,96
19,44
19,83
19,91
19,71
19,79
19,84
19,74
Deviation [%]
-0,22
-2,82
-0,86
-0,44
-1,47
-1,04
-0,78
-1,28
Printing
temperature [C]
Theoretical printing
time (Repetier)
-1
Inner diameter
[mm]
9,34
9,42
9,24
9,65
9,51
9,42
9,40
9,41
Deviation [%]
-6,62
-5,76
-7,58
-3,46
-4,92
-5,84
-5,98
-5,88
-5,9
Height [mm]
20,65
19,75
20,16
20,16
20,07
19,90
19,90
19,92
Deviation [%]
3,25
-1,26
0,8
0,79
0,35
-0,51
-0,48
-0,38
www.iji-cs.org
www.iji-cs.org
deflated. Several 3D programs are capable to sculpt a 3D mesh. With a plug-inn or script this reparation can be first
detected and then adjusted automatically.
FDM or FFF technology is still a kind of experimental way of product materialization. Usually, the enthusiasts are
those who build and use these machines. Anyway, people who do not know much about the stepper motors or
Arduino boards but they want a result a 3D print. Moreover, they are talking about 3D printing, which makes
this technology popular. In the future, this technology will be more and more accessible. Accessibility does not
merely mean availability at stores and low prices, but also easy to use, quality products and reliability of the whole
machine. Ready to print desktop 3D printers can be already bought, but still many unexpected things can happen
during the printing process.
In the future, some improvements are provided. Firstly, constant conditions in a building volume are needed,
which can be achieved with a closed chamber that allows precise temperature control. Second are structural
improvements of extruders that have to be separated from the chamber with flexible barrier, so they do not
overheat. Third is the use of a new G-code generating program and controller called Simplify 3D, but this program
is not a freeware.
Entry-level additive manufacturing is very useful in the field of education and industrial design, where price and
time are important factors. It is also important for individuals who want to build something on their own. They can
start to build the 3D printer first, or they can buy assembled 3D printers and then make different things. If that
things are accurate enough depends on the purpose of the use of the parts. However, precise work always requires
some experimental prints to improve sufficient import-export equality.
REFERENCES
[1]
[2]
Brajlih, T., Testing the accuracy of ATOS 3D optical scanner measuring volumes, Annals of DAAAM & Proceedings, 2007.
[3]
Farin G., Curves and Surfaces for CAGD: A Practical Guide. Morgan Kaufmann, 2002.
[4]
Farin, G., J. Hoschek, and M.-S. Kim, Handbook of computer aided geometric design. Amsterdam [etc.]: Elsevier, 2002.
[5]
[6]
[7]
[8]
[9]