Chamfer issue

This felt like a relatively simple chamfer:

However, it fails spectacularly (yes, Rhino is very eager to shrink surfaces, but seldom extends them):

image

Here’s a closeup of a surface produced with the RollingBall setting:

The DistBetweenRails setting at least produces somewhat clean intersections:

image

However, I must confess that I discovered that other CAD packages also do have a bit of trouble with this (weirdly, increasing the size turned out to be the solution), but it really should be a 1-click operation.

Skärmbild 2022-09-25 125911

Doing the manual work, I did discover that splitting edges projects very well which was nice:

In the end, I just did a revolve to get the job done (with lots of manual trimming, though):

image

But anyway, in case someone at McNeel wants to improve the chamfer command in the future so it’s really just 1-click, here’s an example you can throw on the pile:

chamfer-fail.3dm (429.9 KB)

PS. While doing the above, I noticed quite a slowdown in Rhino… the entire thing is just an 80mb file with little history and I’m working in isolation mode anyway, so unsure where that is coming from…

I split the brep for illustration purpose. The chamfer on the right sidee is created without problems.

The yellow segment is what causes the whole chamfer to fail.

If the chamfer is larger than the segment, tho process fails.

I’d try to rebuild the surfaces so the seams do not interfere with the fillets or chamfers.

Right, that’s why the other CAD package I tried was able to do the chamfer in one go once the size was enough to consume that small segment. Rhino should be able to do the same.

1 Like

If the intersecting material removals are done after the material-addition chamfer, Rhino can create the latter:
image

1 Like

was the other software a solids package? cuz from what I understand, the math for solids and surface is quite different.

Great, and here I tried so hard to avoid mentioning it… :stuck_out_tongue_closed_eyes: (It was NX.)

I get that the math is different, but I feel that this is still a good “real world” example that McNeel could use to improve Rhino.

1 Like

agreed