Oooh, I so want to fillet this edge. Why does it resist my desire?

Hi

To make it work i did two things first scaled it up by factor of 10 the object is really small (im not sure if you made it small or it is importing small I use mm as units). Then I cut it in 2 down the middle keeping the left side and then mirror and join it back up. This is to get rid of the seam on the right side. Without the seam it works fine.

regards

Jamie

p.s also works without the scaling, just cut in half mirror to get rid of the seam then it works

Hi Yenmonger, it actually does work on the original model you posted with a radius of up to 0.02 units.

But you can see where it is having a problem here:





Note the small slivery fragment that Fillet is creating there.

How were the surfaces there created, do they come from a Sweep? I think that there is possibly a very small tangent deviation at the closure point which Fillet is particularly sensitive to.

- Michael

Thanks for the help, guys.

I think it was created using a Network.

Hi Yenmonger, so I investigated your fillet problem here.

The closed surfaces are actually completely smooth where they are closed, there is no tangency break in either one of them.

That little extra fillet segment that I was showing earlier just comes from how Fillet works, part of the fillet calculation involves creating offset surfaces and then intersecting the offsets.

If you take these 2 surfaces where the surface seams are touching:



And then offset them (here by 0.05 units, with "Flip" enabled):



You can see there that the offset surfaces do not have their seam edges touching each other anymore, which is just part of having things meet at an angle.

Anyway, that's what is causing the fillet to be made in some segments.

The actual bug here seems to be either in surface/surface intersection or somewhere deep down in the edge filleter.

Often times in circumstances like this you can work around a problem in the edge filleter by separating things into individual surfaces and selecting 2 surfaces to do a surface/surface fillet calculation rather than an edge based one.

There was a problem though in this case where that was not working, the surface/surface fillet mechanism was bailing out because it thought the intersection curves were intersecting each other. I've removed that particular bail-out for the v2 final release, so in the v2 final surface/surface intersection will work for this case, here I show it on your original model (separated into surfaces), with a radius of 0.05:




Let me know if you still need help with that fillet, and if you let me know what radius you want I can produce it for you now with the surface/surface fillet.

- Michael