Filliting annoying

engine-6.3dm (985.0 KB)
I’m trying to get a fillet on the four spokes where they touch the rim (rolling ball- value = 2) but it won’t give me anything clean. Tried moving the seam, different values and settings other than rolling ball. I can’t see why this should fail. Windows 10 64bit, Rhino 5 and latest update. This isn’t vital, it’s just annoying.

Hi Tone - the ‘back’ edge does not have enough room for a fillet at 2 without its crossing itself. Try setting aradius of 2 at the more open end and .7 at the more crowded end.

Also, I think I like the result from 'DistanceBetweenRails better than RollingBall in this case, still at 2 and .7 - in fact you can boost the smaller radius a bit without its crossing up.

-Pascal

Hello Pascal
It’s the crossing of itself that puzzles me. Why would it do that? A 2cm circle can surly be tangent (in theory at least) to the two surfaces where they cross one another. Never mind, I’ll see if I can fiddle this some other way.

This would be quite a programing feat, but it would be wonderful if the user could set the largest radii, and then click the “best” option, and the object would be filleted using the largest radii it can.

Apparently, some foundation code for this does exist because the radii that are too big are marked in red.

1 Like

Hi Tone - if you make a FilletSrf at 2 and look at it in wireframe, you can see from the isocurves there what happens - the turn is to tight that the arcs cross one another there.

-Pascal

Hello Pascal
I understood what was happening and this maybe impossible in Rhino but I don’t believe it is in reality. Anyway I got there in the end. Not perfect but good enough for my purposes.

engine-6.3dm (1.7 MB)

If geometry goes wrong and they look like a catastrophe, actually, it sort of does :

Had a quick look at Fusion 360 to see what it returned for that fillet at 2mm (and some others)

Rims.3dm (903.5 KB)

Andy