I was having a hard time offseting faces of a polysurface. This is why I want to share the issue with you to ask for your approach and ask McNeel whether it is possible to implement a different “Surface offset” library. Or what is the reason for such a huge difference between Fusion360 and Rhino? Maybe it’s just me and I am doing the whole thing wrong…?
The goal - create a casting mould. Shape created in Rhino, received as STEP. I have Rhino 8, exported as Rhino 7 file: OffsetSurfaceRhinoFusion.3dm (477.7 KB)
Steps to do:
Offset sides +5 mm.
Offset the whole surface +5 mm and work with these shapes further (not important).
Fusion approach - straightforward, easy, two steps.
The reason it fails for an offset value that is any bigger than about 2mm is that to do the offset to a sharp corner Rhino has to offset the surfaces and extend them then trim everything. The problem with your model is the trimmed spheres are arranged so that the extending must go past the singularity of the sphere and Rhino chokes on that.
If you untrim the spheres and rotate them 90 degrees and retrim so that the singularities are no longer involved, then offsetting by 5mm will work. offsetx.3dm (146.4 KB)
Yet mi question is why Fusion can do this operation with ease, but Rhino can’t on the first try without sphere surfaces modification? This is what I would love to know and also hopefully see in a near future enhanced in Rhino…
By the way, I like using both softwares even tho my number one is Rhino. But for instance, making threads or simple mechanical assemblies Fusion is better to do in Fusion.
Well, I already did explain. This is a special case. Its only on rare occasions that the extension of a surface runs into a singularity. Rhino doesn’t handle this special case very well. Fusion handles it better.
That said, I don’t particularly like how Fusion handles this special case. I don’t like what Fusion does either.
The fact is that the result you get in Fusion chops the sphere’s offset surface into multiple pieces. That is completely unnecessary as the model I posted earlier demonstrates. Chopping the sphere up into pieces may or may not make any difference in the long run, but when it does cause problems down the road I like to have software that can deal with that contigency.
It is also a special case that the surfaces in your model with singularities happen to be spheres. If they were not spheres I very much doubt Fusion would do any better than Rhino in this scenario. Knowing what is beyond a singularity on a sphere is dead simple. In the general case what lies beyond a singularity is somewhere between difficult to impossible.
interesting how it solved it tho 
> confusion 360 