By Leah Bartlett, Eric Grunden, Rachmat Mulyana, and Jose Castro Department of Industrial & System Engineering The Ohio State University Columbus, Ohio, USA 43210
Tooling for injection molding is expensive and the time it takes to manufacture a tool is also a concern, especially for companies who are on a tight production schedule. The introduction of Additive Manufacturing (AM) tooling for injection molding is an attractive option for cutting cost and time for not only prototype designs, but also for short production runs. The objective of this research is a preliminary study on two AM tooling questions: How long will the plastic tool survive, and will the parts look similar to the parts produced from a traditional steel tool? In this paper, we compare the mechanical integrity of ribs of different aspect ratio (length to thickness), both experimentally and via computer simulation. We show that there is good agreement between both. The rib with the larger aspect ratio (10 to 1) breaks as predicted by the simulation and the one with the smaller ratio (5 to 1) survives several moldings as expected. In the second case, if the cycle time is adjusted to allow the mold to cool down between cycles, the rib survived a large number of moldings. The effect of tool wall thickness under different packing pressures is also evaluated.
Introduction
Injection molding is the most widely used manufacturing process for thermoplastic parts1. Tooling is expensive and takes a long time to manufacture, so there could be a benefit to using AM technologies to manufacture tools for certain situations. Currently, AM is being utilized mostly for direct print design confirmation. This capability has been beneficial to the design process; however, with the current technologies available these direct print designs do not match mechanical properties of the final design 2. The next step is to direct print molds for injection molding 3. Currently, injection molding is economical for only large production quantities 1. With the integration of the current AM tooling technology, the initial tool cost of injection molding will decrease, opening a new economical option for low production quantities and prototypes. These molds require little to no post processing and are capable of delivering a prototype part in hand that is from the same material as the final design. This new manufacturing method, due to its lower cost and much shorter manufacturing times, can lend itself to multiple design iterations to be tested and verified on a timeline that has never been available without a large price tag.
Experimental Set Up and Materials
In the present work, two sets of experiments were carried out. First, what we will call experiment A, measured tool survivability by examining the impact of flow rate and the mold open phase time on tool rib aspect ratios. The rib analysis will compare the lifetime of two rib aspect ratios under varied flow rates.
This study could assist in determining a pass/fail critique of the AM tool design under the specified molding parameters 4. Secondly, what we will call Experiment B, explored part similarity by measuring the impact of holding pressure and the mold open phase time on tool dimensional integrity. In this experiment, we will measure the deflection of different wall thicknesses with several holding pressures.The tooling used for both experiments was selected because it was able to integrate an AM insert into an available mold base. The AM insert was printed and pressed into the mold base. The mold base available was a U-Frame Master Unit Die (MUD) manufactured by DME. The geometry of the pocket design is a 2 mirrored Ohio shape shown in Figure 1. The pockets inside this mold base allow for customization with different inserts.
The AM material used during both the rib analysis and tool thickness experiments was Digital ABS from Stratasys’ Objet1000 3D printer that prints with Polyjet technology 5. The Objet1000 3D printer uses a liquid UV curable resin with layer thicknesses between 16-30 microns. Stratasys recommends printing with Digital ABS for AM tools because the printer’s capability of printing within 1 mm of the design specifications 5.
Another reason to choose this AM material is the increased mechanical properties of Digital ABS compared to other AM materials. The heat deflection temperature (HDT) and tensile strength are capable of handling the injection molding process parameters 6. At the HDT, Digital ABS will start to deform under a load of 0.45 MPa 3.
The molding material used was thermoplastic polyolefin (TPO) with recommended melt temperature range between 180-220°. All of the injection molding experiments were conducted on a Sumitomo 50-ton
injection molding machine. This machine can develop a maximum hold pressure of 167 MPa and maximum injection speed of 160 mm/sec.
Experiment A Set Up
Experiment A consisted of using printed inserts with ribs of two different height-to-width ratios. The height to width ratios used were 10:1 and 5:1. Both ribs had the same height of 3.75 mm, but the 10:1 rib was 0.375 mm thick and the 5:1 rib was twice as thick (0.75 mm). The two rib designs have the rib located in the same place relative to the mold cavity and each rib had a one-degree draft angle. Figure 2 shows the insert geometry modeled in SolidWorks, and Figure 3 shows the inserts after they were pressed into the mold base.
The rib inserts were tested by molding using different injection rates. Flow rate was varied in order to determine the critical rib failure point. The flow rates used were 66.15, 15.04 and 6.15 cm3/second for the 10:1 rib. The flow rates for the 5:1 rib were 54.88, 40.16, 18.29 and 6.59 cm3/second.
To validate the experimental, the software Moldex3D was run using the same molding parameters. Pressure data at the ribs were obtained using sensor nodes during the filling phase. Placement of sensor nodes is shown in Figure 4.
The Finite Element Analysis (FEA) Software, Abaqus, was used to predict when the rib will fail under thepressures during the filing phase. Failure in the rib is defined as the melt front applying a net pressure difference large enough to cause the Von Mises stress at the base of the rib to exceed the yield strength of Digital ABS 4. When the Von Mises stress exceeds the yield strength of the rib, plastic deformation is predicted to occur. The simulation predictions were then compared to the experimental results.
Experiment B Set Up
Experiment B used flat inserts with 4 holes of 3 different depths, but the same diameter (Figure 5). These holes represent different tool thickness values of 3.2 mm, 5.2 mm and 7.2 mm.
The inserts were tested by varying the packing pressure during injection molding. Packing pressure was varied in order to determine critical tool thickness failure points. The three holding pressures used were 7, 48, and 57 MPa. These levels were chosen based on the maximum packing pressure this particular mold geometry was capable of without flashing. Surface profile values were taken to measure the changing deflection on the surface of the parts due to tooling deflection during molding. The surface profile was measured with a Mitutoyo Surftest SJ-500/P, SV-2100 profilometer.
Results
Experiment A:
Effect of Flow Rate
The 10:1 rib failed during the first shot for all flow rates tried (66.15, 15.04 and 6.15 cm3/second ). Failure is defined as the rib breaking off the insert base. Figure 6 shows an example of the broken rib on the insert. The 5:1 rib did not fail during the first shot like the 10:1 rib. The 5:1 rib insert produced 9 good shots at 54.88 cm3/second, 10 good shots at 40.16 cm3/second , 14 good shots at 18.29 cm3/second and 4 good shots at 6.59 cm3/second. The last shot that ultimately broke the rib was not included in the count of number of good shots produced.
The predicted pressure from two opposing sensor nodes (SN3& SN5) during filling is shown in Figure 8.
The results for the other pairs of sensor nodes is similar. It is assumed that the rib has the largest potential for failure when the pressure is largest in the forward face of the rib before pressure develops in the back face. Figure 10 shows the Abaqus model of a uniform net pressure difference applied on the face on the rib with the base fixed. Table 1 shows the Von Mises stresses at the base of the rib from the Moldex3D simulation during filling. If the Von Mises exceeds the tensile strength of Digital ABS, then it is predicted that the rib will fail during filling for the first shot. The 10:1 rib was predicted to fail for all cases. The simulated stress at the base of the 5:1 rib does not exceed the tensile strength of Digital ABS, so the 5:1 rib is not predicted to fail during the first shot, and it doesn’t. The rib however does fail after a certain number of shots. This is most likely due to the temperature increase during continued molding. The temperatures of the inserts at the time of the failure shots are recorded to be above 58°C which is the lower limit of the HDT of Digital ABS given by Stratasys. Therefore, the larger rib is most likely failing due to the increase in temperature rather than the injection pressure difference.
Effect of Temperature
The temperature was recorded on the insert at the run when the rib failed. Temperatures were recorded on the face of the insert using a type K thermocouple. The 10:1 rib did not produce any good parts, so no conclusions can be made. The 5:1 rib as discussed above did not fail at the first shot. In all experiments when the 5:1 rib did fail, the temperature was consistently recorded on the last shot to be above the HDT lower limit of 58°C An additional experiment was done to examine the effect of temperature on the survivability of the 5:1 rib. A longer mold open phase time of 270 seconds was determined by recording the time it takes for the temperature of the rib to drop below 5°C, the lower limit of the HDT 6. A longer cycle time is necessary for plastic molds due to the lower thermal conductivity of Digital ABS compared to a traditional mold 6. The mold open phase time was increased to allow the insert to cool down between runs longer than in the previous experiments. The insert temperature dropped below 58°C before the start of the next molding. During this
experiment, the holding pressure and cooling time remained the same: 8.25 MPa and 70 seconds, respectively. The flow rate tested was 6.59 cm3/second. The 5:1 rib did not break with the longer mold open time. With a longer mold open phase time, we were able to mold at least 40 shots without the 5:1 rib being damaged.
Experiment B:
Effect of Holding Pressure
Holding pressure affects the different tool thicknesses by applying a uniform normal force on the surface of the insert. Deformation results are shown in Table 2. The deformation values shown are measured off of the last molded part in each experiment, which consistently had the largest deformation measurement. 150 shots were completed with 7 MPa of holding pressure and an 80 second cycle time. The 7 MPa of holding showed a maximum of 0.042 mm of deformation on the 3.2 mm wall, 0.005 mm on the 5.2 mm wall, and 0.002 mm on the 7.2 mm thick wall. Figure 7 shows an example of the insert deformation at the different tool thickness areas on the insert.
100 shots were completed with 48 MPa of holding pressure and an 80s cycle time. With this elevated holding pressure, the 3.2 mm tool thickness (thinnest wall) showed immediate plastic deformation. The
largest deformation measured with the profilometer was 0.107 mm on shot number 80. The 5.2 mm wall showed a maximum of 0.046 mm deformation and the 7.2 mm thick wall showed a maximum deformation of 0.005 mm.200 shots were completed with the 57 MPa of holding pressure and an 80s cycle time. Figure 9 shows a graph of the increasing deformation from the 3 different wall thicknesses during molding as measured on the parts. The 3.2 mm wall thickness displaced a maximum of 0.218 mm. The 5.2 mm wall thickness displaced a maximum of 0.069 mm. The 7.2 mm wall thickness displace a maximum of 0.048 mm.
Effect of Temperature
An additional set of experiments was run on the tool thickness inserts with the same holding pressures, but a longer cycle time of 160 seconds. At 48 MPa of holding pressure and a longer cycle time of 160 seconds, the maximum deformation measured was 0.051 mm on the 3.2 mm thick wall, 0.044 mm on the 5.2 mm thick wall, and 0.032 mm on the 7.2 mm thick wall.
With the maximum holding pressure for this mold geometry of 57 MPa, the maximum deformation measured was 0.069 mm on the 3.2 mm thick
wall, 0.036 mm on the 5.2 mm thick wall, and 0.027 mm on the 7.2 mm thick wall.
A longer cycle time with lower packing pressure caused less deformation on the inserts than the shorter cycle time with higher packing pressure. Comparing experiments 57 MPa/80s and 48MPa/160s, there was 0.167 mm less deflection on the 3.2 mm wall. Similarly, on the 5.2 mm wall there was 0.046 mm less deflection and on the 7.2 mm wall there was 0.016 mm less deflection.
Conclusions and Future Work
Tool survivability was measured by analyzing the effect of different flow rates on different sized ribs. The predicted Von Mises stress at the base of the 10:1 rib were larger than the tensile strength of Digital ABS, so it failed during all molding trials. For the 5:1 rib, Von Mises stresses did not exceed the tensile strength in the simulation, which matched the experimental results of the rib not breaking during the first shot. The 5:1 rib failed after several shots.
Part dimension was studied by measuring the effect of packing pressure and high temperatures on different tool thicknesses. Results showed the thinnest wall deforming more than the thicker walls. Results also determined that a longer cycle time will help decrease part deformation.
Additional studies with AM tooling are currently being performed on the mechanical properties of parts being produced from AM tools. Preliminary results show the Ultimate Tensile Strength, Yield Strength, and Elastic Modulus are similar to parts molded in steel molds. Surprisingly, results show a significant decrease in percent elongation at break in parts molded from in a plastic mold compared to a traditional steel mold. More work is being performed to understand this decrease in ductility.
Acknowledgement
This work was supported by the Center for Design and Manufacturing (CDME) in Columbus, Ohio.
References
1. Groover, Mikell P. Fundamentals of Modern Manufacturing: Materials, Processes, and Systems. Upper
Saddle River, N.J: Prentice Hall, 1996. Print.
2. Kay, Ryan “Effect of Raster Orientation on the Structural Properties of Components Fabricated by Fused Deposition Modeling.” Electronic Thesis or Dissertation. Ohio State University, 2014.
3. Rajitha Aluru, Michael Keefe, Suresh Advani, (2001) “Simulation of injection molding into rapid‐prototyped molds”, Rapid Prototyping Journal, Vol. 7 Iss: 1, pp.42 – 51
4. Grunden, Eric. “Examination of Rapid Prototype Tooling.” 30 Sep 2016.
5. Stratasys. “Object1000 Plus”, Accessed September 30,2016. http://www.stratasys.com/3dprinters/production-series/objet1000-plus
6. J. Kovács, F. Szabó, N. Kovács, A. Suplicz, B. Zink, T. Tábi, H. Hargitai, Thermal simulations and measurements for rapid tool inserts in injection molding applications, Applied Thermal Engineering, Volume 85, 25 June 2015, Pages 44-51, ISSN 1359-4311
Reprinted SPE ANTEC® Anaheim 2017
