Save body as STL and save component as STL give different file sizes

[edgemarston]

Posed this question to Autodesk tech support and got some theories, but no definitive answer. Opening it up to the community. Here's the issue:

You have a component with only one active body. Selecting "Save as STL" for the component results in a different STL than if you select "Save as STL" for the active body. The file size is different, and if you check the number of faces of the resulting STL, it's also different. Why? I would think these two files should be identical.

I've noticed instances with greater than 2x the resulting file size when saving from the component vs. saving from the body.

This raises a few questions:

1. Why are they different?

2. Is the bigger file "better"? Or does it contain extraneous information?

3. If the goal is high-resolution 3D-printing, which saving method is the recommended one?

Thanks!

You can see the difference below. The component and the body were both saved as STL. Then they were re-imported back into Fusion. The part from the saved component is on the left, and the part from the saved body is on the right. The fidelity is clearly different, even though the exact same settings were used in the "Save as STL" dialog box.

[HughesTooling]

Can you share the Fusion design you're exporting from please? I just did a quick test and didn't see any difference so might be something in your design.

Thanks Mark

[HughesTooling]

Also what settings did you use for the export, one of the standard like Medium, Fine or did you use custom.

This is what I see exporting a body and a component then importing the STLs into Rhino.

Mark

[edgemarston]

Sure, here are the STL settings:

And here's the project: https://a360.co/2J5V50A

Try saving Body1 as STL vs. the top-level component. I see a greater than 2x difference in STL size.

Thanks for your help!

[jeff_strater]

the difference can be seen if you expand the Refinement Options area of the dialog.  When selecting the Component, it looks like this:

while when you select just the body, it looks like this:

you can see here that the Surface tolerance for the Component version is much higher than for the Body version.  The "high" setting is there just to give an easy UI way to switch between pre-defined refinements.  I think that the settings for these are based on bounding box of the object, and the Component will have a larger bounding box, which you can see in the preview.

you can see the difference if you re-import the two meshes side-by-side:

the one on the bottom is from the Component export.  You can tell that it is much denser, which is why is is larger on disk.

[edgemarston]

Thanks @jeff_strater! That's really helpful. I have two follow-up questions:

1. I'm probably seeing this wrong, but I'm confused why a larger bounding box would lead to a higher-resolution STL. Intuitively, I would have thought it's the other way around. It seems like selecting the body (smaller bounding box that's just concentrated on the relevant area) would actually result in higher resolution, since it encompasses only the relevant region, and excludes a large extraneous region. I would have thought that would concentrate the total face count into the relevant area. Does that make sense?

2. What do you suggest is the takeaway from this? If the goal is high-resolution printing, is it better to always save from components than bodies? Or do I need to learn how to fiddle around with the advanced settings?

[jeff_strater]

to be honest, I'm not really sure how all that works.  It looks to be based on bounding box, but it may also take into account surface complexity, and other stuff I don't know about.

As to the takeaway, my opinion is not to worry too much about it.  "High" is always going to be more precise than "Medium" is always going to be more precise than "Low".  Those are just convenience settings for those that don't want to expand that Refinement Options section.  I would probably turn on the preview in the command, and take a look at the faceting.  If it looks precise enough, then you are probably OK.  Otherwise, as you say, learn to use those sliders and get the exact faceting that you are after.  If you are doing 3D printing, there is probably not much benefit in going more precise than your printer can produce.