Difficulty achieving accurate print


I rarely reach out for advice, but I am at the end of my rope trying to print this model! It requires a relatively high degree of dimensional accuracy, as it is designed to align a memory chip with 1mm pitch pins inside a burn-in socket. I’m hoping someone may have some advice on something I’m missing that will solve my problems.

Here is the STL file, but it is not yet public due to Thingiverse new user rules:

My Setup:

  • Moai 130 (Q4 2018)
  • Standard build plate
  • Peopoly PDMS vat (only used with this resin for this specific print job)
  • Peopoly white standard resin
  • Using SD card included with Moai kit

Printer Settings:

  • X Size: 900
  • Y Size: 900
  • X deviation: 100
  • Y deviation: 100
  • Smooth 50
  • Z moto speed: 2
  • PM moto speed: 15
  • Laser Power: 58
  • XY Speed Set: 4
  • Z Reset Position: 1877
  • Z Follows: 2
  • CompensateX: 100
  • CompensateY: 200
  • Z Initial Speed: 2
  • PM Initial Speed: 10

Asura Settings:

  • Version: “Asura 2.22 latest”
  • Moai 130 20um profile
  • Currently using my latest distortion calibration(from Monday)
  • Support tip size: 1.5 (this was a test for latest run, normally 1)
  • Support Density: 6
  • Support Angle: 46
  • Support Shape: Round with no additional base

Things I have tried:

  • Perform galvo calibration check (circle test)
  • Bed leveling test (4 pegs) - I used the spreadsheet uploaded by @Apotessar
  • XY distortion test and adjustment - despite 10+ runs (print without calibration, measure, reprint with distortion), no reprint has ever come back accurate to within more than .5mm
  • Wait for warm weather - it is over 29C where I house the printer
  • Strain resin through a 190 micron strainer after several prints
  • Gently scraped the PDMS vat with a plastic razor to ensure no layers were left, making sure not to scratch PDMS
  • Polished bottom of vat with microfiber cloth to remove any smudges
  • Print at 45 degree angle with support
  • Scaling between 102% and 105% in Asura
  • Printing in different orientations relative to the peel/vat tilt direction
  • Varying support density, angle, shape, and tip width

Here is the latest setup in Asura:
Each piece is scaled a different percentage: starting at 102% through 105%. Results of this run show several distortions that I have also seen in other attempts, where hard edges are wavy, but not consistently across prints or axes.

But, overall, nothing comes close enough to fit in the adapter, with distortions among multiple axes, varying from one side to the other, i.e. the rectangles are trapezoidal. My “best” print so far is still off by what I would consider a significant magnitude: the correct outer dimensions should be 27x27mm. but even the best print was 27.83x26.88mm.
Does anyone have any ideas? What other information can I provide? Are my expectations just too high for the Moai?



Some tips:

  • Print in the middle of the plate.
  • Use dark resin (grey or black) for better details.
  • Try not to tilt the model - print on 5 mm supports.

Edit: CompensateY should also be 200.
Edit #2: Try to print with ”round islands” around the support feet.
Edit #3: Don’t give up! Your expectations are not too high - With correct settings on printer and resin temp (the resin should be ~ 28c) you will get much more accurate results. I constantly print the 20x20mm test cube within 0.05mm.

1 Like


Thanks, @Whazaa. I would have included more images of other things I have tried (I have been trying to print this on and off for over 2 months now), but I cannot as a new user. Actually, nearly all of my prior tests were single units in the center of the plate.

I will have to get a liter of black standard resin - I only have NEX in grey - unless that is what you are suggesting.

I have tried not tilting, and after searching this forum recently found that most of the recommendations for flat objects are print it at angle. Example thread: Printing failure



You may try as well to print it with the flat side set directly on the build plate.

Sadly, Moai in general is not very accurate. You may keep the same printing conditions, the same g-code and the prints may come out with different dimensions… Another aspect is that the resin changes it’s shape and dimensions after UV treatment and over time… (it deforms). It’s difficult to keep accuracy even if the print was in the expected range. After a while a plane element may become curved…



These were supposed to remain squares… There is a different resin in the two prints.



I have tried this, but found my results were consistent with the reasoning behind using supports: the first few layers are done under pressure against the PDMS vat, so the initial layers get squished and deform the overall shape. Do you know of a way of offsetting this in the model?



I would set the printing bed to such a distance from the surface of the vat that the four corner cylinders would be at their maximum height (the measured virtual height) and not squished. The resin still sticks to the building plate if the distance between the building plate and the surface of the vat is between 0 and less than 0.1mm when printing with 0.1mm resolution. The build plate must be very clean (rubbed with EP or IP alcohol). You don’t have to keep the plate to push into the surface of the vat.



So you are suggesting changing the Z Reset Position to a height 1 layer’s width above the PDMS?



Between 0 and less than 1 layer. You should be able to slide a small sheet of paper between them.



I am retrying with a new vat (which I was reserving for another type of resin) and a brand new 500ml bottle of white. I will try your print recommendations first, @Whazaa.

Edit: To clarify, I am trying new XY distortion calibration using 60um height on the “resin saver” calibration file, sliced with the newest Asura 2.23 that was released within the past day. Once distortion calibration has been done, I will reprint the recommendations here.

Anyone ever experience anything similar? I’d love more things to consider while the calibration jobs finish.



The Moai isn’t print well on geometric shape, even though you can still get successful prints
1.if the print is small enough, you can just place the flat surface directly on the build plate
2.the print should be scaled up to 107% to prevent shrinkage from changing the dimensions



Hello octa,

if the precision should be better than 0.1 mm, I recommend milling the part by CNC. I would choose the material POM or aluminum. This achieves a long-term accuracy of <0.1 mm and a high degree of robustness.

How many pieces do you need?