Hello 3dhubs Members,
I am optical engineer for an optical design software and we are dealing with freeform optical surfaces. Therefore we have our own STL export of such surface data. Unfortunately, we have a problem, where we do not know exactly what the reason for this problem is. After exporting the data to transfer it to our manufacturer, he said the data is not closed (non-manifold) but we do not know why. Maybe anyone of you has an idea, what the problem is and how we can fix it. A very simple example imported and analyzed in netfabb is attached…
Thank you
Rob
ex10_c1.stl (7.21 KB)
2 Likes
Hi, We imported your test stl file into Meshlab, flattened the visible layers, and exported it as another stl file (attached). Any better? (Both versions sliced fine in Cura.)
ex10_c1_test.stl (5.65 KB)
Hi,
thank you very much. The holes are still there. Please note that the file uploaded is a very simplified surface profile than we actually have. I can provide a bit more complex version of our actual surface. Our goal is to improve our STL Export of the data in our software such that the surface can be printed easily without repairing because that may influence negatively the function of our surfaces. Therefore we want to understand, where the surface problems are coming from.
Best regards,
Rob
section_pca_investigatingstlexport.stl (355 KB)
Perry_1
4
If you are writing your own software, you are in for a treat hunting this down. The issue is inverted normals.
You have to make sure your software always creates the surface of each piece, without having a piece EVER GO THROUGH another piece. Doing so confuses the stl export, as it now has a misdirected surface, that is, a surface that is pointing in the wrong direction.
Hello rob,
This may happen if your software is rounding floating points, so it may look perfect in your software but not compatible with the manufacturer that may have better precision.
You can solve that by applying a magic fix. See - YouTube
1 Like
