The Command Boolean-2-Objects–and similar Boolean operations with these objects—fails for no obvious reason:
Failed to Boolean. Nothing Done.
The surface I have dubbed the Shield is an Open Surface with no errors.
The second, complex shape is a closed surface, also with no errors.
This, croissant-like, shape was formed as a one-rail-sweep, with the rail lying across the surface of the shield and therefore not in a single plane.
The cross-section curves were: (1) a point on the shield surface; (2) a circle, outside the rail except for one point of touching; and (3) a second point at the end, also on the shield surface.
The shape of the shield is fixed but the croissant shape could be edited if necessary.
I’ve attached a file with the two shapes. Any suggestions appreciated, thanks.
A similar exercise with an unbroken, oval rail in a single plane, that roughly illustrates the desired result, did work (image below):
Uset the Intersect command between the two objects.
inspect the resulting curve looking for problems like gaps, small loops, and overlaps.
That curve is used to Split the objects so it’s flaws should be instructive in figuring out the underlying issues.
If the curve is good, then use it to Trim your surfaces, Join the surfaces, to manually complete the Boolean you need.
Unfortunately Intersect doesn’t give a full intersection so you cannot use it to split the surfaces as they stand.
However you will be able to union them if you first rebuild them both with a lower point count - I used 32 by 32 and that allowed me to get this union:
You will have to judge whether the rebuilt shield surface is close enough to the original to suit your purpose. If not, experiment with different point counts and see if you get a compromise that will union. However, as a general principle, it pays to keep surfaces as simple as possible.
John, thanks for your prompt reply. You were right. There are small gaps in the curve close to the points. A previous attempt had a single small gap from the (upper) point.
I haven’t struck this problem before and would much like to use these surfaces. As the problem is clearly with the points, I am fairly confident that the resolution will be to replace them with tiny circles, thus all the cross sections would be circles.
