With knowing what issues we ran into in the first prototype, we worked on optimizing the designs in multiple ways. The first was print times, we used the data that was gathered on print times based on infill density and compared that to the desired part strength in order to choose infill density for each of the sides. Since we were able to achieve the correct quality print with the textured side, we will not be reprinting this side. Below is the results of our decisions on choosing infill density for each print and the required print time.
Table 1. Final part infill density and required print time.
|
|
Final Part Infill Density |
Print Time |
|
Exoskeleton |
100% |
16.9 hours |
|
Button Side |
20% Side, 100% Buttons |
6.25 hours |
|
Roller Ball Side |
20% |
6.75 hours |
|
Joystick Side |
100% |
7.9 hours |
|
Textured Side |
20% |
3.0 hours |
|
Dumbbell Side |
20% |
5.43 hours |
|
Flat Side Pen Holder |
20% 20% |
3.1 hours2Hr43Min |
Exo-skeleton
The skeleton that holds the sides together needed to be worked on because of the very tight fitment of all of the sides into the skeleton. With that in mind, we worked on extending the fin pockets 0.005 inches as well as widening the pockets by 0.0075 inches. This should allow room for all of the fins to slide in easily.
Figure: Skeleton model showing new fin pocket dimensions.
Roller ball
As was mentioned in the previous blog post, the print of the roller ball assembly was fused to the mounting plate. This print was initially given less than 0.1mm clearance between the ball and the plate due to a mistake in measuring the part. Because of this, it is not surprising that the printed assembly did not have the desired movement. To ensure that the ball will spin inside the assembly, we have altered the angle of the print to allow for increased clearance between the two. This print has the maximum allowable clearance so that the ball will still be contained in the part. This increased clearance should also allow the ball to spin even with surface imperfections.
Joystick
In order to fix the delamination and improve the durability of the joystick, the infill was increased. This will improve the strength between layers for the bending. However, this will increase the build time by 3 hours.
|
Infill Density |
15% |
20% |
50% |
100% |
|
Material Usage |
1.79 m |
1.9 m |
2.51 m |
3.52 m |
|
Print Speed |
60 mm/s |
60 mm/s |
60 mm/s |
60 mm/s |
|
Print Time |
4.1 hours |
4.1 hours |
4.5 hours |
7.9 hours |
Another option would be to increase the number of shells to increase the surface strength of the joystick without having a significant effect on the build time. Also, this would help retain some of the flexibility of the joystick by having a less dense infill.
Figure: TPU Joystick Problem and Solution
Slider
Figure: Remodeled dumbbell design with a cylindrical base
The sliding dumbbell side was modified to have a cylindrical base. The previous design was functional, but it had an unavoidable tilt when sliding. This new design aims to remove any tilt because the bottom of the shape is flat.
Button side
With the inability to get the TPU and PLA to print simultaneously the design for the button side was modified in order to print the TPU buttons and the PLA face of the side separately. The new design for the side can be see below. The buttons will be forced over the “pins” that have a type of “barb” on it to secure the buttons in place while maintaining the functionality of the side.
Figure: Back of new button side design.
Figure: Faceplate of new button side design.
Figure: Button insert for new design.
It can be seen in the above pictures that the new design will allow for post-print assembly and will keep the components together permanently. A 100% infill was used for each of the components to ensure their integrity especially on the posts and barbs for the faceplate as well as the actual buttons.
Pen holder side
The infill was chosen to 20% as that seemed reasonable without loss of print time.
|
Infill Density |
15% |
20% |
50% |
100% |
|
Material Usage |
1.47m |
1.53m |
2.83m |
2.33m |
|
Print Speed |
55 mm/s |
55 mm/s |
55 mm/s |
55 mm/s |
|
Print Time |
2h 41min |
2h 43min |
2h 56min |
4h 24min |
We also changed that layer height from 0.2mm to 0.1mm to further increase the print time and reduce the model inaccuracies due to slicing. We suspect thus would have been the cause of the bad print.