Z axis distortion with Dental SG

I printed some prototypes of a large part in both clear and white resins and the dimensions were accurate. When I finally printed the final device with Dental SG resin, there was a significant size inaccuracy in the Z axis of 0.5mm in a 50mm part, or about 1% distortion. I used more supports with larger touchpoints with the Dental SG than with the prototypes. The printing position was the same.

The only difference (besides the resin) was the wash, which was 10 minutes in the FormWash on clean IPA and then a 10 minutes soak (in also clean IPA), as recommended in the datasheet.

The part in question needs very high accuracy, which is what I was expecting from a material made for surgical guides, so I’m lost at why this happened.

Hi there,

First up: I have no experience with Dental Surgical Guide resin. I’m working with High Temp V2 to print small plastic injection molds.
But: I noticed that Dental SG is also a translucent resin, just like High Temp V2 is.
I noticed that my parts are 0,5mm too high (Z-direction). Also my parts swell but that’s a different story, I believe.
Anyways, my current theory is that when the part is printed, some resin stays on top of the part and unintentionally hardens, as some light passes through the translucent material. I’m currently preparing a test to find out whether that theory holds up to real life.

Is your overall part too big or is there a “0,5mm thick extra layer” where the support structure was?
Or rather, to not start on the wrong foot: is your part too big or too small?

Hi, thank you for your reply. Unfortunately my problem is the opposite (sorry, I should have mentioned it in the main post). The part is 0.5mm shorter. This measure is consistent among different features, both internal and external. I checked in all these points and all showed about 0.5mm less than it should have been:

There is a separate “cover” that goes around and inside that main print with a 0.1mm total clearance (0.05mm in each side). This “cover”, printed in both white and clear standard resin, fits perfectly each other’s complement (I can switch the clear and white parts and it will fit anyway). The issue only appears when using the Dental SG resin.

I haven’t checked if the distortion is present before curing, but the part if probably too soft to get good measures with the caliper. Could the problem be only that the PreForm does not compensate for shrinkage when curing the Dental SG as accurately as the other resins (which are made by FormLabs, while the Dental SG is a third party one)?

The part was also “hanging” in the printer for a full day before being removed, but I have done that many times for even longer and it has never caused any inaccuracy.

Hm. The only idea that comes to my mind right now is scaling the part up by 50/49,5% in Z to compensate for the 1% shrinkage, but my colleague just told me, that might fix the overall size but throws off internal holes and structures.

I think if FormLabs considers dimensional changes in PreForm, they would do it with every resin in their repertoire…
And if it’s a small part, it should be able to take some hanging.

Maybe if you print your part almost flat on the buildplate, then the shrinkage in Z direction becomes negligible.
My colleague wrote his thesis also about printing holes and he found out, that the steeper the hole was oriented, the more likely it was filled with resin and became smaller. So the holes should be oriented in Z direction too, to make them as accurate as possible.

Just out of curiosity: are you using a Form 3 or a Form 3B?

Sorry, I just realize my problem is the same you are having. I printed 3 test “T-beams” of 10mm in the X, Y, and Z orientations yesterday. All of them ended up exactly with 10.5 mm! I guess that because I was measuring inside diameters, I assumed my parts were smaller when they were actually “swollen”.

Did you find a fix for it?

I’m using a Form 3 with a modified firmware to accept the Form 3B consumables. I don’t use it for medical purposes, of course.

Right now I only derived a usable fix for the added 0,5mm in Z-direction, with data to back it up: And that’s simply to make the part 0,5mm shorter in Z… (Man, I figured to find out more from a 24h print and two days worth of measuring.)

I printed this huge part that pretty much takes up the entire buildplate:

XY-Calibration global vs. local

It was an overkill, as I figured. Printing smaller but more parts would have given me more data and maybe even made it significant data. Well, that’s considered for next time.

I tried to create a basic f(x) = ax+b formula to describe the deviation in length of my parts and this is what I got:
(x is the specified length in mm; X-Axis: left to right, the wiping direction; Y-Axis: Front to back, Z: up and down)

Outside length in X: f(x) = 0,999x-0,0501
Outside length in Y: g(x) = 0,9992x-0,0672
Outside length in Z: h(x) = 1,0073x-0,4869
Inside length in X: i(x) = 0,9971x+0,0493
Inside length in Y: k(x) = 0,9979y+0,0676

The input (x) is the desired length, and the function output is the length the model should have to get the desired length.

So as of now I would say:
Make the part 0,05 to 0,06 mm shorter in X and Y.
Make the part 0,49mm shorter in Z.
For holes and cavities: Make the feature 0,2 to 0,3% bigger and add 0,05 to 0,06mm in X and Y for printing.
And don’t scale the whole part to compensate the… swelling?
Of course you’d have to do all this in the CAD-software and then print the part exactly as oriented in the CAD…

I’m curious if this holds up in a test :thinking:

Are all your T-beams 10,5mm instead of 10mm long? Or is it “only” the “Z-Axis-Phenomenon”? And what about the width and height of the beams?