Page 2 of 2
Re: [bug] the revolution of a face gives an unxepected result
Posted: Mon Jun 20, 2022 2:48 pm
by onekk
Probably marking a post with [bug] prior to have a confirm that this is a "real wrong behaviour" it is not the best way to proceed.
It is very good that people will collaborate to make FC better.
Regards
Carlo D.
Re: [bug] the revolution of a face gives an unxepected result
Posted: Fri Jun 24, 2022 4:07 pm
by mwlandolsi
onekk wrote: ↑Mon Jun 20, 2022 2:48 pm
Probably marking a post with [bug] prior to have a confirm that this is a "real wrong behaviour" it is not the best way to proceed.
It is very good that people will collaborate to make FC better.
Regards
Carlo D.
It is obvious that my posting here is because I want to participate to make FreeCAD better and better. I like FreeCAD and it is my favourite CAD software.
According to Yorik post related to bug submission:
Make sure that what you found is a bug. The definition of a bug is this: Something that is not working, but for which you are sure that it should be working (for example it worked in the past, or it is advertised on the FreeCAD documentation wiki). Otherwise, it might be simply that what you want is not implemented, or maybe you are doing something wrong
So I think that I didn't contradicted the definition. Such simple feature that can't give a good result, on my opinion, is a big that I hope to see it solved.
I really appreciate all the precious efforts and time that are making developpers.
Best regards.
Re: [bug] the revolution of a face gives an unxepected result
Posted: Fri Jun 24, 2022 4:34 pm
by onekk
But you have been told that is a known problem and this is not related to FC.
Probably there are workarounds, or to be more precise a specific way to avoid the problem changing the order of the operations.
Regards
Carlo D.
Re: The revolution of a face gives an unxepected result
Posted: Sun Jun 26, 2022 9:46 am
by marioalexis
Perhaps the problem is OpenCascade calculating surface normals for some types of geometry, in particular, at the point where the vertical edge of the cylinder and the toroidal edge of the torus are tangent. There was already a problem related to the calculation of normals from tangents:
https://github.com/FreeCAD/FreeCAD/pull/5382
And maybe that's why it doesn't fail if the angle of revolution is a little different from 360°, or if the diameter of revolution is changed a little or if an offset is added to the cylinder.
A possible workaround is to generate another sketch for the revolution with a circumference equal to that of the cylinder, but with a rotation applied on the sketch attachment, so that the edge of the revolution does not coincide tangentially with the vertical edge of the cylinder.