HTMAA 2025 CBA section - Week 4 Group Assignment

Group Session 1

Team: Tyler, Ceci

Tyler:

In order to test the capabilities of the printers, I started with the files provided on the class site. In this section I printed and tested the angle and clearance files on Prusa MK4S printers with 1.75mm PETG filament. I left the settings on default values for the PrusaSlicer software: 0.20mm SPEED, generic PETG, HF0.4 nozzle, and 15% infill.

From the clearance test, we can see that we need at least a 0.3mm clearance in order for parts to be separated. The angle test print is a bit more subjective due to poorer print quality. The 0° column definitely had some issues and demonstrates the difficulty with printing that angle without supports. The 10° and 20° portions also showed signs of collapsing filament, and one could argue the 30° may have a little bit of deformation as well.

Ceci:

I worked on the overhang for supported and unsupported files. I used the Prusa MK4S printers with 1.75 PLA filament. The settings were left in default for the unsupported version: 0.20mm SPEED, generic PLA, HF0.4 nozzle, and 15% infill. In this test, you can see how short overhang might work with no support for very short distances, under 3mm.

In addition, I tested the supported version by changing the support type. To change this, you have to click in Print Settings > Support Material and toggle "generate support material". Once it's checked, the dropdown for "Style" under "Options for support and raft" should be available. If you do not see it, you have to change the mode in the right hand corner from "Beginner mode" to "Normal" or "Expert". Grid makes a column with straight lines going back and forth that go even beyond the area in overhang. Snug makes the same but it fits exactly the overhang area. Organic makes tree-like structures that take less material and are easier to remove, leaving behind less residue.

Then I printed this test print in the Prusa XL with 1.75 PLA to quickly go over several settings. The results are very similar to the MK4s.

Group Session 2

Team: Sun, Ben, Matti, Miranda, Yufeng | TA: Dan

Ben created an initial overhang test model, but the team identified a more comprehensive alternative model for testing. To optimize print time and material usage, the print area was reduced by eliminating angles that had already been validated.

The team selected bgcode format over gcode for export based on what was already on the flash drive. Initial printing attempts failed due to inadequate base plate preheating. We had to run the machine’s built-in preheating routine. In the meantime, Yufeng demonstrated the filament loading procedure.

The printing process proceeded normally for approximately 20 minutes before encountering a nozzle clog that required print termination. The premature failure prevented completion of the overhang test, indicating the need for nozzle maintenance or different print parameters for future attempts.

After re-exporting and double checking our settings (and the material used on the printer), we were able to run a successful print. The results generally reflect the advice we got in class that overhangs up to 30 degrees are doable without supports. In this particular print, we determined that an overhang of 20 degrees still looks ok, but that the print starts to degrade from there. The 10 degree overhang shows signs of sagging filament, the 0 degree overhang has disconnected filament threads, resulting in "spaghetti".

Group Session 4

Team: Sara, Eitan, Jacqueline, and Edward | TA: Dan

To test dimensions, Sara remade the example test block in Fusion since the original example was much too large. The file was then exported and prepared for printing in the PrusaSlicer with the following settings: 0.15mm structural, Prusament PETG, Original Prusa MK4S 0.4 nozzle, no supports, and 15% infill, resulting in a 16-minute print time. The numbers were unfortunately designed too small to read since they were smaller than the minimum dimensions producible by the 0.4mm nozzle. The measurement of the external 20mm diameter design was spot on at exactly 20.00mm, but the internal 10mm diameter was slightly off at 9.96mm. This indicates that the dimensional accuracy is pretty high but not perfect. This should be taken into account for future designs.

Edward also tested the Infill at 0%, 15%, 50%, and 100% using gyroid patterns. We tested by 3D printing a 20mm * 20mm * 20mm cube and adding a pause about halfway through to see the inside print pattern of the cube. For sturdiness, 15% infill is able to give enough strength to the print. Each print took approximately 20 minutes and the higher the infill the longer the print was.

Group Session 5

Team: Jacqueline and Dimitar
We tested Wall Thickness and Anisotropy. First, we downloaded the stl files from the class site, then printed them with PETG on a Prusa MK4S.

Wall Thickness:

Part:		Measured:
3		3
2		2.01
1		1
0.9		0.95
0.8		0.86
0.7		0.74
0.6		0.56
0.5		0.54
0.4		0.47
0.3		0.46
0.2		0.46
0.1		too small to print

Hole Thickness:

Part:		Measured:
3		3.15
2		2.18
1		1.19
0.9		1.08
0.8		1.03
0.7		0.88
0.6		0.79
0.5		too small to use a standard caliper
0.4		too small to use a standard caliper
0.3		too small to use a standard caliper
0.2		too small to use a standard caliper
0.1		too small to use a standard caliper

Next, we downloaded the anisotropy design from the HTMAA website and sliced it using the Prusa Slicer. The filament was Geeetech-brand PLA. The settings we used are shown below:

The print took about 4 minutes and turned out like this:

Next, we tried to determine if orientation and size make a difference. We used the same printer and settings. The print turned out like this:

Unfortunately, we could not reproduce the issue in the positive control. In the original dimensions and orientation, the resulting lengths were as follows:

Horizontal leg	19.96mm
Vertical leg	20.18mm

Additional testing

Team: Eitan, Matti | TA: Alphonso, Dan

We wanted to conduct some additional anistropic testing, and see if there was any way to quantify how strong the different printing angles would be. I found the paper "Understanding the Anisotropic Characteristics of 3D Printed Parts" by Richard Joseph Williams Jr and Dr. Mehmet Emre Bahadir at Southeastern Louisiana University, which showed them using "dogbones" at different angles to conduct similar tests.

So I went ahead and tried to recreate these. Here's how my first attempt turned out:

It not only failed to fully print, but they were also VERY ugly and also, as Alphonso kindly informed me, were not actually useful for testing since they were not really up to testing standards. I needed to re-create these to better fit the official sizes. Alphonso helped tremendously with this, and showed me how to use different circle radii in Fusion to get the correct taper angles.

Here's a gallery of the printing process. After some tests, we decided that the 90-degree vertical print would definitely fail, so we'll leave that aside for a follow-up test. We found that even the 60-degree vertical print was having trouble, so we painted on supports, but it was so brittle that it broke multiple ways in the process of removing it from the supports. This happened multiple times, so we also left that aside for the time being. That leaves us with 0, 30-x, 60-x, 90x, and 30-y to test.

Now it was time to test! Alphonso brought us down to the CBA lab, where we used the Instron Tensile Tester to pull each dogbone apart. Here's the before and after:

Here are our findings:

Somewhat to be expected, the control/0-degree rotation held out the longest, and experienced the most "stretch", so to speak – there was a longer period between peak force and the final snap. 30-y lasted the least by far, which was also somewhat to be expected given how the filaments are arranged and how the 60-y and 90-y dogbones fared in the original printing process, and also had the shortest period between peak force and the final snap – barely any "stretch" or deformation at all. It also seems like 60-x experienced the highest actual applied force, nearing 0.7 kN, even compared to the control/0-degree rotation.

All in all, this shows the importance of orienting prints correctly – the angle really matters when it comes to how the filaments are arranged, and therefore the strength of the print, so make sure to optimize for that. That said, this experiment would be worth repeating using at least 5 of each angle to average out, as well as official tests of the 60-y and 90-y orientations. We were also left with further questions about how the white plastic deformation was happening – some of the weaker ones had a gradient from one side (seemingly the side touching the print tray) while the control and stronger ones deformed from the center out. A great experiment!

MAS.863/4.140/6.9020

Design Rules