Trebuchet #2 – Solving for thermal effects

This post will just document some of the issues we’ve had with uneven shrinkage and warpage and compare different parts we’ve had made.

After doing print tests with two different material, ABS and PLA, we initially decided to work with ABS because it gave us more geometrical resemblance to the CAD models and the fitting tolerances for the assembled parts were more accurate. But we run in to a problem while removing the parts from the build platform. Since the process temperatures for ABS are relatively high, by removing the parts right after the print was finished we introduced a high thermal gradient in to the system which generated uneven shrinkage and in turn warpage.

We also noticed, just as the theory suggests, that the bigger the relation between the area of the part and it thickness the more warping effects we got. And this is evident just by looking at the pictures of the arm and counterweight bellow.

oo2       oo5

Since from class we know that the fundamental cause for this warping effects are the residual stresses induced by uneven cooling, we thought about solving the problem by annealing the parts, we also placed some weight in the parts to force the geometry back to their desired shape.

oo4

The initial impression is that the annealing had worked, the warpage had been reduced to a minimum. But after further inspection we realized we’d done more harm than good by annealing; all the specific geometries that needed to be assembled were distorted.

After some analysis it was clear to us that there were way more residual stresses than we were accounting for, and these stresses were inherent to the FDM process regardless of the material used. As the nozzle deposits the material it also stretches it and the bead enters a state of tension, by annealing the parts the material started relaxing and the general dimensions were reduced; this can be seen in the picture bellow: The throwing arm printed in ABS ( white), that underwent the annealing end up being 1 cm shorter than the PLA(gold) arm.

oo1

This final picture shows in more detail how the relaxation of the stresses affected the shape of the part; the two parts were generated with the same CAD file and it is evident how the geometry  changed from one to the other.

oo