Unexpected behaviour pulling T-Spline to surface

[ot_dsgn]

Like shown in the Screencast, I tried to convert a B-Rep to a T-Spline and then use the 'pull' command to match it to another B-Rep surface. The preview is correct and as expected, but when I confirm the action, the result is a mess.
When I reverse the action, the expected result shows, but in a different spot. This might happen because I moved the original, first B-Rep model earlier in the history tree and captured the position.

[jean.flower]

Is there a link to the screencast somewhere?  I can't spot a link but I'd like to watch it...

[jean.flower]

Thanks for taking the time to record this screencast and share here.  I can see it's unexpected!

I just tried the same workflow on a more simple case I made and things worked as expected.  So there's something specific about your example.  Is it possible for you to share the design file?

[ot_dsgn]

Not this specific one unfortunately 😕
But let me try to recreate this bug with a different file in the next couple of days.
Thank you for your effort!

[TrippyLighting]

@ot_dsgn the example you used to demonstrate the unexpected behavior does not demonstrate a case where one would usually use that function.

You first convert a trimmed NURBS surface into a T-spline patch and then it looks like you are trying to pull the vertices of that T-Spline patch to the very surface it was created from. That does not make a lot of practical sense.

Can you demonstrate this in a practical example?

[ot_dsgn]

It's a similar surface but not the same. I used this workflow to adjust certain features of a STEP-File to changing parameters while staying as close as possible to the given surface tension.

[ot_dsgn]

Hi @jean.flower !

So I finally found the time to narrow the problem down.

It's not the pull command per se, but any modification on the T-Spline body.

I import a Step file and move it and the capture then position.
When I convert one of the components surface to T-Spline the origin is conflicted.

[jean.flower]

OK yes I can reproduce,  Thanks very much for taking the time to explain this and record a screencast!
I'll make sure to pass this onto the relevant team(s) to take a deeper look.

[glenn-chun · Accepted solution]

Hi Olaf, many thanks for reporting this issue. Your simplified model and reproducible steps tremendously helped us narrow down the issue.

We found a defect against converting a B-Rep face to a T-Spline surface: The Convert operation stores an untransformed (wrong) T-Spline surface rather than transformed (correct) surface.

We logged FUS-69008 and expect it to be fixed soon. I will post here again when the fix is included in Fusion 360.

Glenn

[glenn-chun · Accepted solution]

@ot_dsgn Hi Olaf,

The above defect (FUS-69008) has been fixed in August 2020 Product Update (version 2.0.8749). Please let us know if you still see any T-Spline surface jumping to an unexpected location.

Glenn