1

Topic: Just had my first slic3r screwup

I guess it is time to try and figure out how to get my settings into kisslicer :-).

I was trying to print what seemed like a simple enough part and a "ramp" I cut out of the inside of a cylinder was turned into a "step" by slic3r  (running slic3r-0.9.10b-4.fc19.noarch on fedora).

If I look at the stl preview in RH, the ramp is there, if I look at the gcode simulation, it becomes a step.

Very weird. (This is a little connector designed to grab the bar on the PCB Vise from thingiverse and fit inside a short piece of electrical conduit).

PCB Vise: http://www.thingiverse.com/thing:12756

Post's attachments

mountbar.stl 97.87 kb, 11 downloads since 2013-11-07 

You don't have the permssions to download the attachments of this post.

2

Re: Just had my first slic3r screwup

It is actually amazing how much this little part confused every slicer. I previously made it a cylinder and nothing had any problems, then a decided to make the top half very very slightly conical so it would squeeze more tightly into the conduit, and not only did I get the bad slic3r results, but when I tried kisslicer, it was differently confused. You wouldn't think such a simple thing would wreak so much havoc :-).

I did finally run it through netfabb cloud service and the resulting fixed STL was able to slice and print correctly with slic3r.

I don't know if this implies an openscad bug or slic3r bug or what.

More details at: http://home.comcast.net/~tomhorsley/har … bvise.html

3 (edited by adrian 2013-11-08 13:07:57)

Re: Just had my first slic3r screwup

The original model appears to have 6 self-intersecting triangles in some key positions and a degenerate face in a crucial joint and a few 'non-orientated normals' aka flipped-triangles.

Its hard to see all, but the bad self-intersecting ones are highlighted mainly in this view, and the orange intersected triangle leading to a degenerate face is the thing one down the front of the handlebar mount right where it steps out.

My attached Repetier-Host 0.91 with its improved model analysis and repair (aka 'deep analysis' button) manages to fix it up and slic3r goes on to slice it just fine.

So the issue is 100% the model input, but I can't say if its just openscad or how you modelled it in openscad. Its confusing the slicers quite fairly due to its aforementioned model construction issues, but at least the soon-to-be-released RH 0.91 with its built in analysis/repair works a treat wink

Attached is the image and my RH-0.91-dev that I posted elsewhere.
http://www.soliforum.com/misc.php?action=pun_attachment&item=3897

Same thing via RH-0.91 with model repair and straight into slic3r:
http://www.soliforum.com/misc.php?action=pun_attachment&item=3898

Anyway - you can definietely move away from the slicers and back to openscad in terms of finger pointing big_smile

Post's attachments

mountbar.jpg
mountbar.jpg 359.18 kb, file has never been downloaded. 

mountbar2.jpg
mountbar2.jpg 353.36 kb, file has never been downloaded. 

RepetierHost-v0.91dev.zip 1.99 mb, 2 downloads since 2013-11-08 

You don't have the permssions to download the attachments of this post.

4

Re: Just had my first slic3r screwup

Yea, I was sort of suspecting openscad. I don't think I did anything very exotic in openscad, just unioned my cone and cylinder shapes and subtracted off the PCB vise bar. The trouble all started when I added the cone. I'll check out the new RH. Thanks!