Seven trials were done for this project. FFF was used for the first three trials and SLA for the rest . Table 1 summarizes all the trials and the print time. Trial 7 had the best properties since it lasted 40 cycles before failure, however it won’t be the last material to be tested. Digital ABS from Stratasys (Polyjet)and COR Alpha 30 from Polyspectra (Digital Light Projection) will be the next materials to test (Figure 1).
Table 1. Print trials description
| Trial |
Material and Process |
Print time (hours:minutes) |
| 1 | PLA – FFF (20% infil, 0.1 mm layer height) | 3:20 |
| 2 | PLA – FFF (20% infil, 0.1 mm layer height) | 3:20 |
| 3 | PLA – FFF (80% infil, 0.15 mm layer height) | 5:10 |
| 4 | High Temp – SLA | 6:00 |
| 5 | High Temp – SLA | 6:00 |
| 6 | Tough – SLA | 6:00 |
| 7 | Tough – SLA | 6:00 |
Figure 2 shows an insert made specifically for Engineering Expo using the same material and process as trial 7.
The parts that are being injected had some defects such as sink marks (Figure 3), warpage and shrinkage. A possible explanation of this defects might be that there is a non-uniform cooling in the part. The cooling channels in the mold are closer to one side of the part and only one half of the part is surrounded by the polymer insert (Figure 4-middle). Since half of the part is surrounded by the polymer insert, the cooling rate is lower in this section. In order to fix this issue, the entire part should be surrounded by polymer insert(Figure 4-right). To confirm hypothesis a simulation in Moldex3D was done.
For the initial simulations, a control simulation was used were the entire mold was made out of P20 steel, using the automated parameters Moldex3D. The simulations of the mold insert was then compared to two other simulations; one in which one mold insert covered the part and half of the runner system, and another one having another mold insert covering the rest of the part and runner system; this can be seen from Figure 4. The material properties used for the mold insert were those of ABS, as it resembles the material that we will be receiving from Stratasys and PolySpectra. Table 2 shows the following parameters that were used for all three simulations. It is important to note by Figure 5 that the mesh of the runner system and the part were warped as a result of having the surfaces of the part touching the surfaces of the mold insert. Although this might decrease the accuracy of the results, we are mostly using these simulations to observe the differences between using one mold insert and using two mold inserts, in comparison to having no mold insert at all.
The results, shown below in Table 3 and Figure 6, show that the major problem, which we theorized, was indeed the uneven cooling. The results show that although having two mold inserts increases the time it takes for the part to cool down, the part has better quality. It also shows that the part with the two mold inserts most closely resembles the part with no mold insert. What would be the next step for us in this project, regarding simulations, would be to test out how having the cooling channels run through the mold insert would result in part quality. It is also important to note that the exact properties of the plastic we are using are not known as both Stratasys and PolySpectra do not give exact properties of their materials; therefore the next step would be to contact them and ask if it were possible to obtain the material properties. Otherwise, the inserts will need to be thermally and mechanically characterized.
Table 3. Simulation results
In order to run the experiments in the injection molding machine, an additional aluminum plate that will hold the polymer insert in place will be machined.