Query about max quality

[johnnyshield]

Thanks for your work on this tool, generally I'm getting great results, but occasionally I'm having issues with rough geometry on the output, despite the file size being quite small (3MB)

Checking the NIST SFA gui - low medium high options correspond to --quality flag of 4, 7 & 9 vs. range 1 - 10 available in STP2X3D.cpp.

Can you please advise if it output precision is limited by this range (and could be modified), or if this is a limitation of OCCT?

NB: I haven't tried using STP2X3D (not permitted to install Cmake at work), so just using the NIST SFA gui.

[robert-lipman]

Johnny, can you share the STEP file here or send it to robert.lipman@nist.gov. I'm the developer of SFA and would like to test the quality with your file.

[johnnyshield]

Funnel.zip

Posting the example file here as well (emailed to Robert), highlighted the sawtooth geometry on output

https://github.com/usnistgov/SFA

[soonjokwon]

Hi. Sawtooth mesh is not actually a problem. If something breaks, then that is the problem.
This bad geometry might be happening due to inadequate tessellation parameters.

Currently, we use the same tessellation parameters for all STEP files regardless of dimension.
We plan to adjust tessellation parameters per body or model in the near future.
However, this doesn't ensure that sawtooth-like geometry would never happen.