FilletEdge_oneGo_noRevolve_.3dm (2.5 MB)
Version 8 (8.28.26041.11002, 2026-02-10)
do a simple box 0 300 100 200
_filletEdge with R40 and R30
if i do the _filletEdge in one go (left) the revolve is not found at the lower cornes
if i do the _filletEdge in 2 passes - the revolve is found.
somehow related:
thanks for a fix.
kind regards - tom
Not a FilletEdge user, but when I fillet that box in one go, it works and I don’t see any Revolve in the mix. What exactly is the problem?
FilletEdge_oneGo_noRevolve_.3dm (2.8 MB)
from my point of view / experience - it makes a difference wether a torus-like-surface is exactly represented by a rational Degree 2x2 Surface or if the revolve direction is approximated by multi-span Degree 3. Less accuracy but more CVs. And if there are workflows on top / depending on the fillet surfaces, the results are better for true revolve surfaces - see the linked topic above.
_explode and _list both surfaces to see the complexity vs a surface of revolution.
There is a difference that is relevant in the 1-step vs 2-step procedure. In the 1-step procedure, the surfaces that are not simplified to a revolve originate from the input corners, while in the 2-step procedure they originate from input edges.
I agree that the corner surfaces (left) should also be found as revolved surfaces in this case, but currently this is only done for edges. I have made this YT for the request to also find revolved surfaces for corners: RH-93845
ok - thanks for the explanation - even if geometrically the same it s 2 different parts of the algorithm…
thanks for all your effort.
