Hi Forum. I'm trying one of Tom Meeks' cool project ideas for a first 3D print (I've linked his project video below). Instead of using text on the globe like example I've imported vector lettering with closed curves. When I perform the boolean difference I'm getting very inconsistent behavior. Sometimes the result looks nothing like it should and other times there are small inconsistencies like the center of vector letters . In my latest attempt ( attached ) the boolean only seems to perform correctly for the first line of the first hashtag lettering instance where the relief in globe is visible either in 3D view or with faces made invisible. As far as I can tell the "200 Working Text" object is consistent and should result in the consistent boolean diff on "000 Working Globe". What am I missing ? Thanks, Sven

https://youtu.be/uJqxvO45gbY

I'm unable to upload attachment at the moment. Just says "wait" after I push upload. I'll try again in a bit.

Hi Sven, if your file is pretty large (like say more than 5 MB) you may need to upload it to a file sharing service like Dropbox / We Transfer / OneDrive / etc... and post the link here.

A couple of general things to look out for - it is not too unusual for there to be anomalies in font geometry like spots where outlines cross over themselves and stuff similar to that. Those things don't really matter too much when the font is used just for 2D printing but 3D booleans are more sensitive to that type of thing.

So there may be some cleanup needed on the font curves.

Another thing that can be good is when you are going to boolean some text out of an object, it's a lot more difficult for booleans to handle cases where there is a lot of coincident overlapping surface area. So position it so that it pushes a little bit through the level instead of being flush with it. Like this:



- Michael

Hi Michael. Thanks for the insights, always helpful. The font curves have been extruded so there should be adequate overlap. I looked at the vector letters a bit looking for something that could explain, but don't see anything. Flow can result in different results without changes to what is being flowed ? I've uploaded the .3dm file to google drive below.

https://drive.google.com/file/d/1ft0DPWiWj6PtQcIzxRYjSk-k0xyRsS5z/view?usp=share_link

Hi Sven,

re:
> Flow can result in difference results running without changes to what is being flowed ?

The flow result is dependent on where you click on the base surface and target surface.

If you click in different areas that can give different results.

You need to click near the corner of a surface but not quite exactly on the corner, a little bit along one edge as described here:
https://moi3d.com/4.0/docs/moi_command_reference8.htm#flow

Looking at the file I think the self-intersections are in the B letter here:



- Michael

And there's a little zig zag in the I that would be good to get rid of too:





- Michael

Hi Sven, try the boolean with this attached version, it should be ok now.

I edited the curves in those 3 spots, 2 on the "B" and that zig zag which is only on one I it seems?

- Michael

Wow. I now see how carefully one needs to look at these things. Thanks for showing the actual problem areas. Helps explain where the flow stopped working so that should have been the area where I started looking for problem. Learning. Thanks Michael

Sven