Minimum Feature Size Test

Remember we are testing a whole manufacturing process, not one single part of it. The methods, resin, hardware, and software is all one system that we are testing. It either makes a hole or it doesn’t.

True that, i guess.

Edit: It just occurred to me that it might be interesting to see a cross section of a print bisecting the two smallest holes. No idea on how to do it without mangling the holes, though. It’d have to be cut with a VERY sharp cutting tool.

Hi JoshK, I mostly print rather smal stuff so I don’t have anything tall and thin to show. I´ll be happy to print a tall test - which file do you suggest?

Just put my 100 micron print in the mill/router and cut away 1 mm on front and back and then cut away down to half the round holes. I made this a bit angled to not miss the tiniest holes. There is no trace of the 300 micron which could just barely be seen on the surface to begin with.


I’m finding those random voids interesting too.

That was a very good idea to cut into it like that. Very interesting.
There’s prior prints to compare of of Kevin’s test part. If you want to give it a go I found a link on Steve’s post here.

I’m thinking the issue with printing small details, small objects, small holes has less to do with the “minimum feature size that the laser is capable of”, and more with the resin being used, layer thickness and the laser cure time per layer and area.

Having one setting for printing delicate details and large cross-sections/large items at the same time is not going to give the best result for either print. High-end expensive printers for example treat various parts of the model differently when curing. Thin areas need less energy than thick areas, etc.

So until there is some sort of variable curing algorithm that controls the laser intensity/scan speed based on the area type, this is the best we can get. It’s highly possible that these methods are patented by patent trolls such as 3ds.

Another huge factor is the peeling mechanism. I think a lot of those bumps and tears are because of the peeling, and the condition of the PDMS layer.

Another way to get really sharp detalis/holes, is to use a highly pigmented resin that blocks light more efficiently and is optimized for a certain slice thickness.

Yea I agree. And of course the path offset PreForm creates is a guess at that.
It’s interesting to see the performance of each unique machine.

I am really liking this idea. That would be an amazing feature request. But yea it might be forbidden.

I know this or something similar is being done by some DLP printer software (perhaps the b9 software) and there are even free software out there that do similar things with DLP. Although it’s much easier to do this with DLP projectors because you can just use variable intensity masking on each layer.

From what i’ve seen, it’s already doing this… Kinda.
From my (limited) testing, it seems castable uses different exposure for infill, but the same (or almost the same) for perimeter compared to black. And, of course, it changes based on layer height.

Both are modifications to laser output power / exposure based on parameters of the print.

So i doubt anyone’s got the patent on that (it’s too broad and it would’ve already been breached), but there might be patents on algorithms to determine curing profile / laser power based on feature size and surface.

Also, it might not be as trivial as it seems.

Consider the following - let’s say you have a flat plane. And then a flat plane with almost a fractal surface as it’s texture. How do you determine which one requires the “detailed” curing profile? Total surface? Total surface enclosed in a fixed volume? Number of faces per ml? Number of edges per square mm? It’s not that simple.

What would make things utterly simple (but still offer the possibility to improve the print quality) would be a slider you can drag yourself:

Fine <------ Universal ------> Coarse

1 Like

Here are mine (black at castable, 50um):


And now the same, but black printed at the black setting, 50um:

Interestingly, despite the botched surface, the smallest pillars survived:


Here is the test print of Kevins test part. 50 micron setting, black resin with black setting. Front side OK, left and right side rather OK but back side not good. Hinge side part to the left in pictures. Laser spot also attached. I would say the bad back surface is due to the laser spot…? Also rather much “gelly” in the resin after this print. The non-hinge side part is slightly better, could be more movement in the resin stirring away some of the partially cured resin around that part better than on the hinge side.

@Ante_Vukorepa, What size does that test part go down to?
@KjellNilsson, Those look nice except for the back side. I assume that is because of the flare. That’s real interesting because your flare does not affect the min feature test.

I have a theory about that tall object and why the back is so bad and every other side is perfect.

I don’t think it’s caused by the flare, but by “PDMS fatigue” and by the peel process. When each layer peels off the pdms layer, over and over, the very end of the peel, depending on orientation can cause bad surfaces, especially when you use the same area, over and over again for a thousand peels.

I imagine if you twist that model and make it more like a helix, it will print just fine. It’s worth a try.

Actually the blobby surface on the support side is due to extra exposure on the draining resin as the laser passes beyond the blockers. The smooth side is always the pdms side where the layer is compressed and the laser first hits the surface.

slightly cloudy pdms probably does not help and especially with lasers that have large flares.

Wonder if it is possible to develop a resin that uses a dye blocker instead of a pigment based blocker reducing the settling effect and chance of uneven grind of pigments.

Of course no flare on the laser would be nice too.

I’m talking about the tall tower prints. There is no support side on those, except the bottom.

On the tower prints that is determinately from the flare and as you mentioned condition of the pdms.

I wonder if a resin that blocked better allowing for more precise exposure on each layer with so light doesn’t refract through would help?

A resin that has lower viscosity and higher pigment should give better results. Unfortunately, for large prints, if there is no mixing going on, then the pigments will settle and that opens up a new set of problems.

That’s an interesting hypothesis, but already disproved pretty well.
@KevinHolmes demonstrated quite clearly the flaking side changes with the orientation of the laser (by rotating the laser).

That would make the test useless for what it’s supposed to show. Namely that the flakes are due to repeated exposure of same volume of resin to the lower intensity, unwanted parts of the beam, that - over time - accumulate enough to cause curing or partial curing and end up as a flake hanging off the print.

Making the object helical would remove the repeated exposure of the same area to the flare from the equation, making the test useless.