BlendSrf fails on simple geometry

i am not sure whats happening to Rhino… or with me? i made a circle copied resized and extruded both, but BlendSrf acts up and creates a twisted surface. no matter how i rotate the input geometry.

Loft with match ends works. can anyone tell me please that i am dumb and that i missed something extremely obvious, because i really cant believe that anymore.

BlendSrfFail.3dm (45.1 KB)

@encephalon

your surface normals are facing opposite directions.

I tried rotating the top cylinder 180 degrees and the blend surfaces seemed to work. I am guessing this is because the v normal were facing the opposite direction, the blend surface was having to self intersect to make the shape possible. It is the same as when a loft curve needs to be flipped.

Hopefully this helps!

BlendSrfFail.3dm (147.9 KB)

@Parker_Gillespie

it is not about fixing it, it is that blendSrf should not act up in this case, i simply extruded two circles. the input geometry could not have been much easier.

but thanks for looking into it, i guess its a bug.

@encephalon

I understand. I just tried fallowing the steps that you mentioned in the original post. I used the original curve, copied it and created the top and bottom surfaces. When using the blendsrf command I did not have the same result. There was no error with the blend surface. I tried several different methods of creating the top and bottom surfaces but none of them changed the outcome of the blend surface.

I’m curious, what exact steps did you take to create the top and bottom surface? In my case, the surface normals were aways positive on the Z-Direction (green arrow).

only what i wrote above, nothing else.

Left: Extruded degree 3 circles, U and V swapped. Right: Cylinders, U and V identical.

Briefly, I rebuilt it and it’s fixed. I didn’t understand why it did that; I guess Blend’s working logic is different from Loft

If kink-free degree 3 or 5 circles, cylinders, revolves, etc. are used, the U and V direction does not affect the result, and BlendSrf builds as expected.

one different method is to split the edges at the midpoints, BlendSrf then behaves again.

still, i dont think it should happen that i have to start fiddling around with the input geometry at all. most certainly if i did 2 straight forward operations and nothing else to create input geometry for that task. loft does not care how the u v or normal directions are and blend should also not, specifically when the edges are clearly facing each other.

further more what some might not have noticed that it does not ask to drag/set the seam point here, it jumps straight to creating a twisted blend instead, though i can rotate and move the seams after that but the twist keeps rotating along with it no matter how i set it.

just to clear up more doubts that this is a bug, when i do the exact same operations once more, creating a circle copying it resizing it and gumball extruding both curves then blend it all works suddendly as expected, so why and what happened here. @Gijs can you have a brief look? i assume its a bug.

Yes this issue is known and listed as
RH-90770 BlendSrf Odd Twist

I’ve added this thread

RH-90770 is fixed in Rhino 8 Service Release 30

I tried BlendSrf in a similar case, but at the moment it still doesn’t work well.

Sure you have the build with the fix?
I have (8.30.26069.2001, 2026-03-10) in here, and it works fine.

Ah no, I’m still on (Rhino 8, 8.29.26063.11001).
At the moment it says that mine is the latest version, so I’ll wait for the next update.

You are on final Service Releases. Charles is on the Candidates.

If all goes well, that will be April 14th.
-wim