Layer Strength Testing Jig and Test Chips

I thought you used temp towers to figure out the best temperature for a specific filament on your printer. Then I saw people claiming that lower temperatures, even if they look better, have poor layer adhesion. Some people claimed to test this by manually breaking the temp tower, but this is going by feel, not data, and the temp tower is more likely to break where the stress is highest (i.e. in the middle) rather than at the level with the presumably worst layer adhesion. So I tried to come up with a basic test of layer adhesion strength where instead of trying to measure anything, you test samples against each other in a series of “which one is stronger” tests. Just use a tournament-type bracket to find the best temp.

First I tried two different implementations of a three-point bending test, but the results were inconsistent and the test samples were pretty big and used more filament and time than I wanted. Then I ran across this “splitting test” (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9229671/) that was more consistent, repeatable, and concluded that “the splitting testing procedure that was the most effective tool for assessing layer adhesion”. It also works with much smaller test samples and a smaller test fixture than my prior attempts. I added notches in both ends of the test chip for alignment and as stress raisers to focus and redirect the compressive force on the layer bond. It requires 1/8” (3mm should work for you metric folks) steel rod to be cut, ground down, and glued on as pins to break the test chips without deforming the jig and you will need a vise of some kind to slowly apply the compressive force across the jig, but I’m sure someone could remix a 100% 3D-printed version if they wanted to.

Here’s how it works:

• Put a test chip in one side against the central pin then insert a slide to hold it in place against the slide pin
• Put the other chip in the other side against the central pin and insert the other slide to hold it in place against that slide pin.
• Put the jig assembly into a vise with flanges on the slides up to keep it from falling through and slowly compress until one test chip breaks or the slide pins bottom out in the center section.

The test chip that breaks loses and the one that doesn’t wins, but use an untested sample in the next round in case the winner was damaged. The test chips are so small and fast to print just make a ton and test the same combination multiple times to make sure. 

Suggested settings and print options for the test chips:

• Enough perimeters to make the test samples solid (2 for 0.6mm & 3 for 0.4mm nozzles). 
• No top or bottom layers or infill. 
• Put the seam in the middle and not near one of the notches on the ends.
• I print batches of 10 or so and color code the top surface of each batch using a paint pen to indicate printing temp before pulling them off the bed.
• Prusa Slicer won’t allow Gcode changes when printing sequentially, but I manually added M104 commands to the Gcode in between sequential printing of groups of 10 in the attached bgcode for an XL. This file will let you print 40 test samples at four different temps in 28 minutes. I also included the x10 sample chip stl for other printers. To edit the Gcode just add the M104 command after these lines when sequentially printing the groups:
  ;AFTER_LAYER_CHANGE
  ;0.4

Suggested settings and print options for the test jig (center and slides): 

• 3 perimeters using a 0.6mm nozzle or 4 using a 0.4mm nozzle
• 5 layers of 0.2mm top and bottom
• 30% infill. I prefer gyroid.

PLA looks best on temp towers printed on my XL at around 200 and I expected hotter temps to be stronger as others have suggested, but here’s what I’ve seen so far for different PLAs at temps of 190, 200, 210, and 220:

Overture PLA (white)

• 220 & 210 didn’t break or separate. They just squished down to max compression. Both temps equally good- does not fail at layer bond.
• 200 was similar to 210 but a little bit worse- it didn’t break but started to separate at max compression. Unless strength is really important 200 temp is probably good enough for this filament.
• 190 could be broken by all other temps. Don’t use.

Overture Eco PLA (black)

• 220 was strongest, but all temps broke very easily compared to the regular Overture PLA.
• The 220 samples of this filament were easily broken by the 200 samples of the regular Overture PLA.
• This is not a filament to use for anything where strength is important, even at the best temp of 220.

CC3D Silk Orange PLA

• I like using silk filaments for appearance so I thought I’d try one of them for comparison. This one also snaps very easily compared to regular PLA.
• 200 breaks 190 again.
• 210 broke 200 3 of 4 times. 
• 210 breaks 220.
• So the best temp for good layer bonding for this filament is 210, but 220 of the Eco PLA breaks 210 of this one. So this filament has the weakest layer bond of the three tested.

I store my filament in Ziploc bags with desiccant, but it's been a while since I've really dried this one. So I dried it a few hours, reprinted, and retested. It looks like drying the filament didn't make this one any stronger.

This test will be useless for TPU and I’m not sure how it’ll work on PET-G. Please let me (and everyone else) know in the comments about your results if you try it on other filaments. I plan to use this to test PCCF next after I get my air scrubber built, and then PACF. I also want to test the PCCF vs. PACF. I’ll post an update if I find anything interesting.

EDIT 10/30/24:

I tested 250-280 in my IEMAI PCCF and 270 was definitely best. 
I tried it against the other materials above and it beat all of them except the good white PLA, which mashed down instead of splitting at the layer line. But that's more of a different failure mode than better layer adhesion.