Can't join manifolds

[fc_tofu]

I don't think to cut the feet can help the offset feature because it's to the junctions the issue occurs.

Would you like share your model? I want have a try.
If FCStd is too large, STEP would be OK.
Also, you can use 7zip split your file into small parts, and upload in separate posts.

[Vincent B]

done. above the picture.

[raidho36]

Changing the dimensions or location of a profile enough to make the union work can be as minute as a couple of microns, below the resolution capacity of most manufacturing methods, imperceptible in reality.

Sure, but it's extra work, in certain scenarios it's a lot of extra work, it requires a bit of knowledge not covered in the manual, and all of that just to work around a bug - I shouldn't need to do this, the bug should be fixed instead.

The source of the issues lies in the third party OCC geometric kernel.

So I'm barking up the wrong tree, huh. Well, it's still in your power to make the interface more user-friendly I suppose. I liked the idea of selecting individual sketch components to use 3d tools on. It can be accomplished within current framework - if it doesn't already support this - by generating virtual surfaces bounded by sketch elements, which would be selectable from the 3d viewport, and then the selection is passed into boolean union and from the resulting surface a final virtual sketch is extracted, which is then used to generate 3d geometry by extrusion or lofting or what have you. Generating a virtual sketch consisting of a single closed profile when the user clicks on an edge and using that as the input is also an option that would be useful to have.

[fc_tofu]

done. above the picture.

Thank you. Your surfacing technique is great.

Finally I failed in FreeCAD. Workaround is Modifier >Solidify in Blender, though topo is poor quality.

fsc_2020-06-11_223639.jpg (50.94 KiB) Viewed 1396 times

fsc_2020-06-11_223502b.jpg (51.63 KiB) Viewed 1396 times

[tanderson69]

Not freecad, but occt, so close enough. Brep file is too big.

manifold_Blobfish01.png (99.95 KiB) Viewed 1366 times

[mrdic]

That one doesn't have knife-edge splitters in the ports, if we're going to observe notions of flow management. The most recent version of ACIS refuses to add the individually lofted runners as the original workflow requires, too, but its branched lofting tool has been around for 11 or 12 years.

[fc_tofu]

That one doesn't have knife-edge splitters in the ports, if we're going to observe notions of flow management. The most recent version of ACIS refuses to add the individually lofted runners as the original workflow requires, too, but its branched lofting tool has been around for 11 or 12 years.

Would you give a try? Offset and Thickness under ACIS with @GlouGlou model.
https://forum.freecadweb.org/download/f ... ?id=117575

[fc_tofu]

Interesting reply from a Fusion 360 dev.

https://forums.autodesk.com/t5/fusion-3 ... -p/6684491

jeff.strater
Software Architect - Fusion 360
in reply to: kurtgui
‎11-12-2016 06:59 PM
Re: 4 way pipe joint

Hi @kurtgui,
This would be challenging to do in BRep. Unfortunately, Fusion doesn't have a "branched loft" feature. So, I would be tempted to do this with a TSpline pipe.

[mrdic]

I tried GlouGlou's model, ACIS wouldn't shell/thicken it, citing internal boolean inconsistency. I tried another workflow, seemed brute-force, but whatever: lofting with the inlet profile circles initially offset, adding the lofts, then moving the inlet profiles to the intended location. I successfully added four of the branches together that way, but it crashed when the inlet profiles of the fifth were colocated with the others. Displacing the fifth minutely coaxially works. I could live with that if I didn't have access to other things. The Fusion dev's comment was made in 2016, so the OP's success shows that they've worked on it since then. ACIS wouldn't shell my FC model as .stp, but the usual thing would be a bit of slice'n'dice to find the area of greatest convolution, the show stopper, and fudge that.

[tanderson69]

... add the individually lofted runners as the original workflow requires

Where was that spelled out as a requirement?

That one doesn't have knife-edge splitters in the ports, if we're going to observe notions of flow management.

Yeah I didn't consider that, but it wasn't hard to change. Lots of bopalgo errors!

manifold_Blobfish02_front.png (113.45 KiB) Viewed 1298 times

manifold_Blobfish02_back.png (71.17 KiB) Viewed 1298 times