I’m attempting to create a network srf that matches tangency to a high degree. But when I turn up the angle tolerance, It creates a bunch of little bumps. Does anyone know a way to prevent this?
Below is a zebra analysis showing the bumps, a picture of the three curves and brep edges that I am using to form the surface, and a 3d model of the inputs.
Do you mean the Angle tolerance in NetworkSrf , or the Maximum angle setting when the display mesh is adjusted in Zebra, or the Maximum angle setting in Options > Mesh?
What do you mean by “turn up the angle tolerance” - increase or decrease the numerical value?
This looks like a render/display mesh issue to me, not actual bumps in the math data of the NURBS surface.
I’m talking about the angle tolerance on the Network Srf. I’m talking about decreasing this number to increase precision of the tangency. My display setting are cranked up to the maximum. Also, when I mesh the surface, the mesh has the weird bumps too. Any when I project a curve onto the surface, the curve has squiggles. All of this leads me to believe that it is an issue with the surface being generated, not just the display.
Thanks for looking into this:)
Can you upload a file with the input geometry including the surfaces used as edges, and the surface with bumps?
Here you go. The bumpy surface was generated with .001 for all networksrf tolerances. And I matched tangency with the other surface, not curvature.
bumpy.3dm (5.8 MB)
The problems originate with the trimmed surface edge used as input to create the surface. The edge curve has 1153 control points and and has oscillations. (I used DupEdge to extract the curve from the surface, and then analyzed the curve.)
How was the surface trimmed? With a fine mesh? Super tight tolerances?
Add the tight tolerance of 0.001 for position accuracy in NetworkSrf and the result is a surface with 1315 control points in the U direction. Further add the extremely tight angle tolerance of 0.001 and the result is bumps in the middle.
RebuildEdges the surface with a tolerance of 0.001 and the trim edge curve has 90 control points.
Amazing, you are a life saver. I feel like I understand the issue much better now. The surface doesn’t have many control points, but the trimmed edge does.
I’m not sure how the trim was created, this is a surface from a client, but with a fine mesh seems like a plausible explanation.
Thanks so much for the help!
