Weird striping effect


#101

Yes, there is curling happening. My supports are bad: too massive and not cross-linked. On the other hand castable resins are harder to print and therefore should show flaws in the setup/process better than conventional resins.

From the benchmarking/comparison point of view it’s not ideal, but better than nothing.


#102

actually, it could also be the angle. is it at 45°? Try with 50-60deg if you want to print more tests


#103

thanks for testing. Wax has its own set of problem of being softer than model resin and will have layer lines. The correct approach which you took is to lower pm motor speed/z follow speed. some users print wax by turning the z follow off (by change z follow from 10 to 0)


#104

I have something to test: I was able to get rid of most of the lines in Cura by making the option “Compensate wall overlaps” and both underoptions visible in “shell”. Then I turned both off.

Compared to them on:

I just can’t currently print it, because my resin is bad and also I think my machine has additional problems so it could mask it.


#105

I already tried that, it’s just very slightly better. I usually print without Compensate wall overlaps now.


#106

I think we are at a point where Peopoly should release a list gcode commands that the Moai can use, as well as A list of all cura options they use with an explanation of why they are used. I have the feeling we are flying blind regarding these errors.


#107

I don’t think cura can be actually used, unless it becomes more specifically crafted for SLA.

They say Prusa is releasing a SLA printer, but I don’t know if it’s LCD or galvo based. Anyway we could probably use some of their work on the software side (at least for supports).

In regards to gcodes I feel like moai only listens to G0 and G1 :slight_smile:


#108

Here is a previous post about this.


Consistent inconsistency in Z axis (thinking it's Z wobble)
#109

A new day, a new test.

So I took the latest Slic3r 1.3.x, recompiled it with 5 digits precision (instead of 3) to see if we could get any improvement… aaaaand…

I’d say, no. No improvement whatsoever. I really don’t know what other tests to run at this point.

All I can say is that latest Slic3r looks slightly better than Cura 2.x, I haven’t tested 3.x but I don’t think there would be any improvement.


#110

I also tested Cura 3.3. I spare you the picture because it changes absolutely nothing compared to cura 2.6. I tried the experimental “resolution” feature in cura 3 to decrease the minimum segment size (and theoretically get higher resolution paths) to no avail. I’m now printing a remeshed version of the same model to see if anything changes.


#111

Could it be possible that the Mirror & Laserunit are not working correctly?


#112

Cura 3.3 has a lot of issues with many of its new features. It is one of the reasons why we have not released it. We likely to do a 3.4 version


#113

Could try the Simplify 3D profile they released. Just make sure to print on supports because the first 2 layers end up weird.


#114

I learned a great deal about slicers in these last few weeks. I’m starting to understand how the topology of a model influences the way it is sliced.

I was able to slice at a 0.0001 grid resolution with a minimum segment size of 0.0001. To give you an idea the gcode for the benchmark I posted came out 300mb instead of about 60mb generated by Cura. The result is always the same, maybe better than the early prints, but still far from perfect.

@peopoly are you sure that it’s not something happening in the firmware or an hardware limit (in the DAC maybe)? I tried 5 slicers now and they all generate that kind of glitches, I start to believe that the problem is somewhere in the moai.

@Paul_Jack-Frederick I know it’s a long post to follow. We tried all slicers on the market, S3D included (here’s the result). Very little changes. Slic3r seems to be the best one so far.


#115

yea, sorry TLDR.

have you tried re-meshing the model?


#116

dude, you gotta read the whole thread :smiley: it’s not something we can easily fix (or fix at all). Apparently it’s just how the printer is at this moment in time.


#117

I wish we knew more Form2 people for comparison. I get little “shadow” indentations like that occasionally when I print objects that have a lot of planar surfaces. I can usually sand them off but it certainly isn’t a desired effect.


#118

I am not sure if that Cura resolution function works at all. Our modified Cura makes changes to how much it reduces polygons which affects resolution as well. In regular Cura, no matter how small the resolution is, it still eat away too much polygons for Moai to print.

A good test file is the ring-test stl file we use for calibration, the tip of the support does not slice at all in standard Cura as it is removed. This shows that the resolution does not do what we think it does.


#119

I am not using cura, I’ve come across a slicer called iceSL that is completely scriptable via LUA and I’m pretty in love with it :slight_smile: you can basically customize the output code in any way you want (it even supports LAYER: and LAYER_COUNT or have all E set to 1)


#120

we tested that back in the developmental stages. it was unstable at the time so we didn’t continue. do you have a setup config that prints?

another one we looked at was KISS. I personally likes it a lot but this has even more FDM stuff baked in like auto flow rate / print speed change which definitely not going to work. It is pretty rigid so we never tried to hack it to work