Boolean Intersection Mystery

Here’s the model. I can’t understand why the boolean operation is failing. Maybe it’s obvious to an experienced Rhino.

Chef’sPan_Metric copy.3dm.zip (1020.6 KB)

The two objects have parts of surfaces that occupy the exact same space (co-planarity). Booleans work by sensing difference between “self” and “other”, via intersecting surfaces. Surfaces which are shared by both objects don’t necessarily intersect, so the Boolean operations are challenged to function correctly. They are confused and give up.

If you go to your Top viewport, use your smallest Nudge to move the vertical component up or down one click. Currently, your smallest Nudge (Alt+Ctrl+ up or down arrow) = 0.01 mm. If you can live with that variance, you’re good to go. If not, you’ll need to re-order your thinking about your workflow so that you can deliberately make the objects not occupy the same surfaces, perform the Boolean operations, then trim/split/Boolean the object so that it looks like the coplanar version that didn’t work.

There’s a great intro to Boolean issues in the FAQ. Check it out.

Best regards,
Doug

1 Like

Thanks Doug. I appreciate your help. I’ll look into the FAQs.

Hi James - here’s how I’d do this:

Chef’sPan_Metric_pg.3dm (828.2 KB)

Some things I’d change from your approach:

  1. Simpler, cleaner curves.
  2. Don’t include the edge softening/filleting in the input curves - build to hard edges and add the fillets later.
  3. I left out the spherical surface till later - coincident surfaces in Boolean operations, apart from planes, tend not to work, as has been pointed out. I chopped that surface out of the sphere after the Boolean of the two objects forming the handle. The Boolean works because the input surfaces both end cleanly at the sphere (even though it is not there as far as the Boolean is concerned) as a result of them being Revolves on the sphere center.

-Pascal

Thanks Pascal,
I’ve redone the the curves to minimize the number of control points. But there’s an artifact I’m curious about as you can see in the attached - that is the double line through the middle of the revolved object. I thought it might be a c-plane issue but after projecting the curve to a c-plane I still get this. What is the cause of this, and does it matter?

I really appreciate the education here Pascal. I’ve a question about the nature of the curves you used and the number of isocurves that result in the solid. I’ve learned to reduce the number of control points to the minimum but I can see a great number on each of the construction curves used to revolve and I wonder how you end up with such clean objects.

Hi James - the point count does not necessarily exactly relate to isocurve count - isocurves are drawn per span, and higher degree curves (I probably used degree 5) will have more points per span than lower degree ones.

-Pascal

The 5 degree curves were applied using the rebuild tool?

Hi James - no - I probably made the initial shape curves using degree 5 curves (Curve command) - that is my usual choice for swoopy shapes. When curves are joined, the result takes on the highest degree of all the inputs.

-Pascal

Okay, thanks that’s a nice tip.