Prints Out of Round

A lot of the parts I print are small and cylindrical or near cylindrical in shape. Unfortunately I have noticed a consistent “out of round” condition on all my prints so I started doing some experiments to track down the issue. I tried doing the custom calibration and got good results out of the calibration part (see details below), but my parts are still not printing circular.

I generated a test part in Solidworks with the following dimensions (dimensions in [] are mm).

Using Asura (v 2.2.4) I added my parts in the following layout and printed them using the calibration file generated earlier.

The prints were made using the standard grey resing and I have a Peopoly heater in the chamber. All 3 prints show an out of round condition with the long axis appearing roughly in the same direction (long axis running roughly 45deg from back/left to right front corners of the build plate.

Using dial calipers, I made the following measurements. Note the diameter of the part is nominally 0.300" [7.62mm]. I would expect a small amount of shrinkage from the resin ~.002" [0.05mm], but I see out of round conditios on the order of .006-.018" [0.15-0.46mm] or about 5% of the diameter.

image

In ideas on how to correct this issue or what might be the problem are appreciated.

the resin has a 6% shrinkage, and it’s not the same everywhere even if your model is a circle like shown above.
for more exact measurements, it’s best to measure right after the print is done, before post curing

Hmm, in my experience the shrinkage is a lot less. I had to recently adjust another print by 0.5% for it to come out to size. Another square test part showed roughly 0.6%. Even taking the numbers from my test above and averaging them, I see roughly 1.8% shrinkage.

In either case, I can compensate for global shrinkage, but I care more about final dimensional stability. I typically post cure with a UV lamp over top of the parts in a water bath. Should I consider using a turntable for more even post cure?

The 6 % shrinkage @peopolysupport is talking about is volumetric shrinkage. You measure linear shrinkage, which is (1.06)^1/3=1.0196. So the 1.8% linear shrinkage is pretty close to the 1.96 % calculated.
I doubt that the effect you are seeing here is predominantly caused by shrinkage distortion. There are alot of threads in the forum about people printing cylinders and getting oval cross sections. In my opinion this is related to galvo control. There are some nonlinearities in the system that the firmware doesn’t fully compensate for. Look at this picture, a cyanotype of a calibration raster made with my Moai:


As you can see the “D” in the top right corner is skewed upwards on its right side, even though in the corresponding gcode the top and bottom lines of the “D” are parallel to the grid lines. I don’t know what is the exact cause of this, maybe galvo backlash or jitter.

Nevertheless using a turntable will improve post curing.

Thanks “nbltsgndpfrdbrms”, for the clarification on the shrinkage.

I have to concur with your assessment of the galvo control being problem. I just printed the same test parts and verified before removing them from the build plate that I get nearly the same dimensions on the parts (within a few thousands of an inch) as I did in my prior tests. All the parts as skewed in the same general direction. I don’t seem to have any issue printing straight linear edges, either square to the primary axes or at 45deg. It seems to specifically be an issue with curves or circular paths.

If would be great if Peopoly could look at this issue a little more closely and develop some compensation for the problem. I am not convinced the precision calibration file has any impact on this effect, since the test pattern is linear in nature. Perhaps a circular calibration process could be developed.

If there is no plan to fix this, I may have to go buy another cheap 4k LCD printer. At least the pixels can’t get this distorted. :crazy_face:

I think you can read here very good background information about the distortion problems. Bernhard did a lot of tests ans spent plenty of time for this.

I appreciate the detailed link and good background. Based on the work Bernhard has done, I don’t really have the time or skill to spend to fix this myself. @peopolysupport, It doesn’t look like a fix has been implemented yet. Is a fix for this issue something planned in the future?

Hi Eric,
you can follow the ongoing firmware update process here …

or here …