Fusion
Good luck with that!
Your object is constructed from T-Splines that where converted into BReps and there hacked up and stitched in the Patch environment. That results in very complex trimmed NURBS surfaces with all sorts broken edges and undesirable curvature.
Another program won't help you here. I just tried this with another very powerful (and more expensive) software and it did not work there either. As expected 😉
You'll have to diet this object (unstitch) and manually offset and trim individual surface patches if you really need exactly 1mm.
Perhaps you can try to thicken just the main objects and then combine the smaller details with it. This will likely need a little experimentation.
In general if you have to use the word "complex" and "shell" in one sentence then that already indicates the problem.
@danolson1 I've looked at your model and the problem you are encountering might result in the same error message but the root cause is likely a different one.
However, going through the timeline of your model I am having difficulties finding where you ran into trouble shelling or what object this was related too.
@Anonymous wrote:
If the goal is to create a mold for dough...I've seen some pretty complex dough molds on the market...what do you suppose these people did to create those molds?
I have no idea what these people do, but a number of things come to mind. I personally doubt that modeling dough moulds is a market lucrative enough to yield using a multi thousand dollar CAD system, so they likely use a Sub-D modeler in combination with real sculpting software such a ZBrush, 3DCoat and Blender etc. Once the sculpts are done these models are re-meshed, or re-topologized and then the mesh is dense enough that it can be shelled/thickened very consistently using a quad-mesh thickening similar to what can be done with the T-Spline thickening method.
The reason the thickness in your T-Spline is inconsistent is because the model does not have enough geometry or really subdvisions. I take such meshes out of Fusion 360 and apply one or 2 levels of Catmul-Clark subdivision in Blender, apply a solidify modifier in Blender to thicken the shell and then that goes back into Fusion 360.
Here is a tutorial I made recently to addresses this T-Spline thickening problem:
That shape shell fine for me at a reasonable thickness. Reasonable meaning that the wall thickness cannot exceed the smallest radius on that curved surface, which is likely why shelling failed in your case.
I've been following along on this thread. Getting back to the original design problem - if you are making a bread mold, do you really need a precise offset? It seems to me that the only side that needs to be precise is the one that is touching the material to be molded - the inside. The outside doesn't matter. You have a very nice surface model of a seahorse, all you need is for that to be the "A side" of the mold. Are we making this design problem harder than it needs to be?
The screencast below is probably not the ideal result - would be too much material. If I had time, I'd make a simplified TSpline version of the B side, instead of a simple extrude. But, hopefully you get the idea and can make it work.
Jeff Strater
Engineering Director
OK, that makes sense. Thanks for the thoughtful reply. I did play around a bit with Scale - if you are very careful, you can copy that body, scale it just the right amount, and position it just as carefully, and get a pseudo-shell that way. It won't be uniform thickness, but you might be able to get that to be close.
Alternatively, if the entire model had been done in T-Splines, the Offset in Sculpt mode works pretty well, and can be an alternative to the more precise B-Rep shell.
Beyond that, the techniques that @TrippyLighting recommends are the only ones that I think might have a chance to work.
Jeff Strater
Engineering Director
I have no doubts that the toy industry has enough resources. For no good reason - silly me - I had assumed this might just be a private venture. Well .. never mind then 😉
I am officially jealous at the resources you have at hand!
Work flows and tools that help combine, bridge and enhance T-Splines and Sub-D modeling are very interesting to me.
What I would try is to change the order of operation a little. You created several individual T-Spline bodies/meshes. Those converted into surfaces, or solid bodies. Then you split them and stitched them into the final surface you were hoping to shell or thicken.
While it is more work I would try to thicken these individual surfaces individually and then do the same thing you did before but simply twice, once for the inner surface and once for the outer and then stitch the result.
The individual T-Splines have a much higher chance of getting thickened with the thicken in the patch workspace than the combined stitched object.
This is more work but likely much more robust.
Offsetting the surface and thickening will generate the same geometry in the file you attached.