(bug) _filletSrf

Rhino
1

_filletSrf (Extend = no, R=8) fails to build the cyan fillet

image
image1300×938 114 KB

splitting away small stripes (red) (_split isocurve shrink = yes) the command works.
Same when moving the right surface 0.001 mm up in world z.

image
image1300×938 143 KB

might just be a tolerance issue but somehow its unexpected to not get any surface / result / error-Message

EDIT

sorry here is the file

fillet_bug_2023-06.3dm (3.2 MB)

2

Assigned to Rhino category

1 Like
3

@Tom_P can you post the file as well?

4

ups - forgot… update initial post
fillet_bug_2023-06.3dm (3.2 MB)

5

Hi Tom - thanks - in our latest in V7, I get the (nearly) expected surface -
image

with a little extra noodle at the end (I re-extended the surfaces)- but it does not fail. In V8 the correct surface is made but the preview is ugly for some reason, I am adding that to the bug list.

-Pascal

6

here - no result at all.
mac os x
Version 7 (7.30.23130.11002, 2023-05-10)

7

OK - I see- if I stop both surfaces at the lower edge intersection, it fails.
RH-75134 FilletSrf fails
thanks,

-Pascal

1 Like
8

thanks for having a look.

9

A few more items to note:

  • V5 can _FilletSrf the green surfaces with Extend=No
  • V7 can _FilletSrf them with Extend=Yes
  • V7 can _FilletSrf them at a 0.0007 model tolerance (Extend=No)
  • The lower edges of the green surfaces miss intersecting by ~8e-6. When they intersect, V7 can _FilletSrf them (Extend=No)
10

Thanks for your detailed examination / report and more or less solution.

11

All of those methods produce defective fillets. If you incorporate inaccurate fillets like that into a model, it will guarantee that downstream operations like shelling will fail.

The only way to make the edges perfectly intersect is to split (by isocurve) the ends of both fillets and then move the final 13.5 mm arc in both fillets the microscopic amount so that the edges have the same end point. That is often the best way to trick Rhino into making an almost good fillet. You still usually need to use setpt to fix the pointy end.
fillet_bugx.3dm (156.7 KB)