How to constrain to circular paths
 

How to constrain to circular paths  

  RSS
craigbot
(@craigbot)
Active Member

I just can't figure this out. I'm working with ColorFabb XT, and it's a little tricky. Especially in colder weather it clogs easily even when I ramp up the nozzle temperature. The design I have is a half sphere, and it lays filament well with circular traverses (first photo). However, no matter what I try, PrusaSlicer decides to make partial circular traverses with embedded infill (second photo) on some layers, and with the first layer like that with rapid small movements, the nozzle clogs. Is there any way that I can constrain those layers to circular traverses only? Thanks in advance!

 

Best Answer by Neophyl:

Mr stoned is correct.  It doesn't matter what thickness shell you use until it gets to a thickness that allows infill.  At the edges where your triangles meet the shell is minimally thinner.   You can see the faceting of the slice where the circles arc has been broken into straight lines.  If it gets to a boundary where it cant quite fit an entire perimeter in it breaks it up into sections of perimeter and gapfill (the yellow and white colour coding).  This causes the 'dotted line' effect you are seeing.  Changing the models shell thickness varies where this happens as when you go higher or lower in the slice the apparent thickness of the shell varies too.

Once you get to a thickness that allows for your set number of perimeters AND normal infill  then the internal gap is no longer being done via gapfill but by infill all the way around so the effect goes away.

Yes Prusa Slicer could do it better probably.  Its a known problem, see https://github.com/prusa3d/PrusaSlicer/issues/2666   

As to fixing it, well probably not completely untill they do a software change on PS.  Things to try are setting your extrusion widths to something smaller (but not smaller than your nozzle).  That way you might get to fit a full perimeter in instead but with a sphere its doubtfull that would work for the entire shell.  Turn off gapfill (set gapfill speed to zero to do that).  Beyond that you are basically looking at using a different slicer.

Posted : 12/01/2021 3:59 pm
mrstoned
(@mrstoned)
Trusted Member

I'm just going to mostly copy-paste knowledge i learnt from others;

Those are a mix of perimiter and gapfill sections.

Circles are a mesh of triangles that vary in thickness around the curved face.

I would try and adjust the extrution width so you end up with 4 perimiters on the half-circle.

Also try enable/disable detect thin walls and see if that helps with the gaps.

Perhaps even make internal perimiter wider only alongside thin walls enables it may make it 4 lines, with the inner lines varying width instead of the 5th line/gapfill.

 

Hope my ramblings makes some kind of sense.

Flsun QQ-s "pro". Tevo Nereus. Prusa Mini+ KIT on order....
Posted : 12/01/2021 8:02 pm
craigbot
(@craigbot)
Active Member

@mrstoned

Thanks: I had tried varying extrusion width significantly, no luck, and I had changed the detect thin walls with no luck.

I'm not sure what you mean by "make internal perimiter wider only alongside thin walls enables it may make it 4 lines, with the inner lines varying width instead of the 5th line/gapfill." Can you be more explicit?

Posted : 12/01/2021 8:56 pm
mrstoned
(@mrstoned)
Trusted Member

If for example external perimiter is 0.45 you can try perimiter to 0.5 to make the internal perimiters thicker. With detect thin walls enable, it should vary thickness of the internal perimiter instead of making gap fills.

Same with solid infill if this is depentant of the gap fill. Unsure how it calculates this.

 

This is just theory so far.

Also i remember something of the quality of stl exports from the cad software (the higher mesh count translates to smoother curves with the tradeoff at much larger files and gcode instructions).

Flsun QQ-s "pro". Tevo Nereus. Prusa Mini+ KIT on order....
Posted : 13/01/2021 9:00 am
craigbot
(@craigbot)
Active Member

@mrstoned

Ah, yes, I had tried increasing the perimeter as well. The problem isn't the perimeter if you look at the picture: the problem is the innermost traverse. The perimeters are all circular.

I'm using Solidworks. The meshes are very high density. As I vary parameters, where the innermost traverse starts to break up changes layers, so the problem is PrusaSlicer.

Posted : 13/01/2021 11:31 am
mrstoned
(@mrstoned)
Trusted Member

Could you try and increase the thickness of the half-circle by 0.1 or .2mm just to see what the slicer does with it?

I'm not very well traversed with the inner workings of the slicer, perhaps others have more things to try.

Oh, also which version of the slicer are you on?

Flsun QQ-s "pro". Tevo Nereus. Prusa Mini+ KIT on order....
Posted : 13/01/2021 1:54 pm
Neophyl
(@neophyl)
Noble Member

Mr stoned is correct.  It doesn't matter what thickness shell you use until it gets to a thickness that allows infill.  At the edges where your triangles meet the shell is minimally thinner.   You can see the faceting of the slice where the circles arc has been broken into straight lines.  If it gets to a boundary where it cant quite fit an entire perimeter in it breaks it up into sections of perimeter and gapfill (the yellow and white colour coding).  This causes the 'dotted line' effect you are seeing.  Changing the models shell thickness varies where this happens as when you go higher or lower in the slice the apparent thickness of the shell varies too.

Once you get to a thickness that allows for your set number of perimeters AND normal infill  then the internal gap is no longer being done via gapfill but by infill all the way around so the effect goes away.

Yes Prusa Slicer could do it better probably.  Its a known problem, see https://github.com/prusa3d/PrusaSlicer/issues/2666   

As to fixing it, well probably not completely untill they do a software change on PS.  Things to try are setting your extrusion widths to something smaller (but not smaller than your nozzle).  That way you might get to fit a full perimeter in instead but with a sphere its doubtfull that would work for the entire shell.  Turn off gapfill (set gapfill speed to zero to do that).  Beyond that you are basically looking at using a different slicer.

Posted : 13/01/2021 2:08 pm
Neophyl
(@neophyl)
Noble Member

Just found this one that might explain it better than me https://github.com/prusa3d/PrusaSlicer/issues/3025   I remember seeing it when it was posted originally but couldn't find it earlier

Posted : 13/01/2021 3:14 pm
craigbot liked
craigbot
(@craigbot)
Active Member

@neophyl

You're a hero. Really appreciate the great explanation. Your first link led me to this https://github.com/supermerill/Slic3r/releases, and the most recent push just eliminated the problem entirely, at least on the computer. Running the print out now to see if it actually works. Many thanks!

Posted : 13/01/2021 5:54 pm
Neophyl
(@neophyl)
Noble Member

I really like Super Slicer, I'm active over on Merills github too, as Ive been using it for a couple of years now when it was called Slic3r++. However be VERY careful slicing and checking the results with the 2.3.55 releases.  They are based on the PS Alpha releases.  As such they have ALL the bugs from the alphas as well as issues with the merge of PS and Super Slicer.  Has a tendency to crash or do weird things.

Merill mentioned he was going to wait until a full release of PS 2.3 before going through the merge process again and sorting out bugs from that.  As that's now released he should be doing the merge now.  

I'm currently swapping between PS 2.3 and Susie 2.2.54.4 (stable) but really wanting a new stable release of 2.3.55.x

Posted : 13/01/2021 7:11 pm
craigbot liked
craigbot
(@craigbot)
Active Member

@neophyl

Thanks super for the words of warning! Do you know  if Susie 2.2.54.4 corrects this hemisphere bug in PS? The most recent version definitely does, but I haven't run it out yet.

Posted : 13/01/2021 9:27 pm
Neophyl
(@neophyl)
Noble Member

@craigbot

I think the gapfill changes are in 2.2.54, but I'm going from memory on that.  

Posted : 13/01/2021 10:29 pm
craigbot liked
craigbot
(@craigbot)
Active Member

@neophyl

I pulled the last stable version 2.2.53.4 and it does fix the gapfill issue. I ran a print off it, and it worked. So many thanks!

Posted : 17/01/2021 6:51 pm
Share:

Please Login or Register