I’m importing various lab equipment into a lab layout, and I was looking around for complete STL files for the High-res Openflexure Microscope v7 and the OpenFlexure Delta Stage. All the STLs I could find seemed to be of individual components. I found the glb files (links attached with the interactive viewers); however, these are on the order of 100’s MBs when converted, it comes in a huge scaling (likely due to the random online converter I used), and can’t be selected as a solid to scale it down in Fusion 360 due to “mesh not closed” etc.
You are looking for a single STL of all of the physical parts together and in place? Like the GLB, but with out the size? Rather than a sungle downliad of all of the STLs.
The GLB does import into Blender, and there you can group and delete parts to make it less detsiled. The electronics particularly have lots of bits. I don’t know what export options there would then be from Blender.
If you are able to use OpenScad, then you can make an STL from the complete_microscope_rms.scad in the rendering folder.
OpenSCAD often produces STLs which give errors, so Fusion 360 may still not like it. There are tools to fix STLs that help - slicers do that before slicing to print.
As @WilliamW said, those scad files can create the STLs. Quite a lot of the triangles are because the motor and LED boards are automatically converted. This means all of the circuit board silk screen are thousands of triangles.
I’d try some of the earlier assembly steps first or try deleting some of the electronics.
Thank you both! I think I will try importing to blender and deleting the electronics. Makes a lot of sense (and also cool that the GLB is accurate to that level - in a true spirit of open source)
I tried taking out the electronics components in the OpenSCAD. This is less trivial than it appears. It is easy enough to remove the code that renders the electronics, but for rendering we use an imported STL of the main_body.stl. OpenSCAD does not like its own import of the STL it generated. It is happy to preview it (which is what we use for coloured render images), but a full render to allow exporting the resulting assembly to STL does not work.
So I set it to render the whole thing from the base main_body.scad. After 7 hours 36 minutes it has created the render! Save to STL took a while as well, and is 31 MB which is too big to post here.
@WilliamW Have you tried the OpenSCAD nightly builds with the “Manifold” backend? I’ve seen massive speed increases for other renders, but haven’t tried rendering OFM with it yet.
I am under the impression that ‘manifold’ does not allow the minkowski() construction. Unfortunately we use that in a few key places in the main body structure. One of my longer term goals is to determine how we can eliminate minkowski(), as it is inherently quite slow anyway.
Edit: There is actually some work-around in the current OpenSCAD build allows it to use manifold mostly, and feed in and out of CGCAL for minkowski(). It may or may not work on the microscope github.com/openscad/openscad/issues/6297
It may be good for building to get visualisations, but I wouldn’t fully trust it for building our STLs. I’m a bit worried about OpenSCAD’s direction on the language front, lots of changes, some which could possibly alter geometry. Even when they finally make a new release it is going to take us some time to really validate that it isn’t changing geometry
Yes. I hope that for this visualisation any geometry changes would be either very noticeable or unimportant. For anything else using a different CAD engine has hidden dangers.
