0.1 points will be 0.035mm - that’s pretty fine, I haven’t tested but I wonder if the PDF format has a lower limit for line width sizes given that its original purpose was for printed documents.
The laser I used to work with (Epilog) would switch from vector to raster at line width values above 0.17mm or 0.5 points. I used to assign line widths of 0.01mm to vector curves to sent to the cutter (directly from Rhino, not via PDF) and it worked fine.
My general experience with PDF vector documents is that the precision is horrible - I would never use them to directly fabricate anything accurate.
I tried my own test here with Rhino PDF and Adobe PDF (very old Acrobat though) and the results were both very bad, the RhinoPDF being somewhat worse. The line width sizes are inaccurate relative to the original and seem to increase in specific steps. I have no idea if this is a characteristic of the PDF format itself, the PDF print (export) functions in Rhino, or the import function or all of the above…
The Rhino V7 file below has 24 squares with line widths from 0.01 to 0.24. The Rhino PDF starts at 0.042333 and has the same value all the way up to the square with 0.06, then it jumps to 0.084667 (i.e. the double) for the next 4 up to 0.10, then goes to 0.127, etc.
The Adobe PDF starts somewhat finer at 0.0211667 but exhibits the same “steppiness”, just that the steps are about half the size of those made by Rhino PDF. Therefore there is something being done by Rhino PDF that is different than Adobe PDF and that is not caused by the file format.
LinewidthPDFTest.3dm (113.9 KB)
LinewidthTest-RhinoPDF.pdf (22.4 KB)
LinewidthTest-AdobePDF.pdf (24.5 KB)
This is all on Windows though, I no longer have a Mac capable of running V7.