Offset going wrong. Can't figure out why?

Hi Guys,

I got this shape which is causing an issue in one of my grasshopper scripts and I’m trying to figure out why so I can build a solution into the script.
image
When this curve gets offset it results in a super tiny triangle which appears to be open.
image

image

Can anybody explain why this happens so I can come up with a solution to deal with these curves? It seems like it’s somehow related the to Seam of the curve, when that gets moved it sometimes does give a correct offset result. I’d rather not change the seam of every curve as that will cause issues in other parts of my definition.
211108 OffsetWrong.gh (7.3 KB)

Hi @Siemen
You are out of the modeling space, or here known as to far from World O (Original)
Skale it down in to modelling Space.
WO

Surely that’s not the problem. I’m only 5000 units away from origin. When I move to the origin I get the same problem.


If I scale it down and back up I don’t get the same tolerance as specified in the file.

Strangely, if you explode and rejoin the curve, it seems to work.

Exploding this curve gives a combination of lines and several arc-like curves with extremely large radii, and hence centers which are way out from the origin. Maybe that’s a contributing factor. Also, according to the IsKinky component in Pufferfish, it is kinky.

Hmm interesting. It doesn’t do that for me here, though:

Also doesn’t do it if I manually explode and join in Rhino and then load it in GH again. I have my units set to millimeters and tolerance to 0.001.

Good thinking about the large radius, would that be it?
Not sure what “IsKinky” would tell me, though? Curves with Kinks should be able to be offset, no?

Edit>
The large arc radius is replaced by smaller ellipse and smaller distance.

I think that is true

I don’t know if such shallow arcs are important to your project or are just artifacts of however this curve was generated. If not, you could try replacing them by lines.

The curve with lines does look a little different, but rebuilding the arcs as degree 2 nurbs with 2 segments works with offset and doesn’t look too bad – to me at least.

1 Like

Edit

Yeah I need to end up with a curve made up of arcs and lines. Guess I’ll need to replace the arcs that have a large radius and can be changed only if they’re within a certain tolerance. Thanks for your help.

Would be nice if anybody from Mcneel could confirm that this is the issue here.

What’s even weirder is that when I mirror the curve along X-axis, the offset seems to work fine.