Boolean Union failing, no error message, no naked or non-manifold edges. Help!

entry wall.3dm (9.5 MB)

Model attached. I am making a fairly simple model which is why I’m so confused by this failure. The blue and green solids are the ones that will not Union with the overall shape, and it fails without telling me what’s wrong. I’ve checked for naked edges and non-manifold edges and am shown none.

I really need this to just be a single solid, and I’m about to bash my head into the keyboard. Please help!

Thank you!

Hello - Booleans will fail if there are coincident faces that are not planar.
They will also fail if there are non-manifold edges in the input.

Your objects have both…




Thanks for your response. I guess I’m having trouble understanding what is not planar about those faces? I simply extruded curves imported from illustrator. I don’t understand how that would result in 90% of them being non planar. Also, when I check for non-manifold edges in my model in rhino, it shows none. How are you seeing non-manifold edges? What’s different?



Edit to add: how do I fix these edges so I can union and be done with it? :smiley:

Hi Andrea -


Any of the curved faces in your part are non planar - when these exactly correspond to the faces in another input to the Boolean operation, there is no clean intersection - this can be handled on planar faces but not non-planar ones.

For the non-manifolds, ShowEdges will let you see them (dialog box setting)
These cannot be fixed really - the thing is, curves are infinitely thin - they have no actual width- the extrusions like you have there have, at these non manifold edge, zero thickness. To ‘fix’ you need to either make the corners overlap or pull them apart.


It’s all the arcs, those curved surfaces that overlap messes up the booleans. It can handle that(though it’s still not recommended) with perfectly flat surfaces, but it can’t handle curved ones. It doesn’t “know” that all those surfaces were made from the same arcs and automatically how to handle them, and due to the complexities of math on a computer where those surfaces won’t appear to Rhino to perfectly match even though they do, it freaks out and gives up.


Thanks. That explanation helps me understand what’s going on a little better. So I get why the two curved surfaces won’t union, but what about the entire simple rectangular backer that all these pieces penetrate. Shouldn’t they be able to union to that? Why do some of the pieces with curved edges union to the backer work and others fail? And is there any way I can create a single solid shape with what I have?

I don’t have license to alter the design that was given to me, I was tasked to make it a model consisting of a single solid so it could be put through a CNC mill. So if there’s no possible way to make those oddballs union, I have to ask the client for permission to alter the design.

Ok, I just tested unioning those pieces that are giving me problems to a fresh new rectangular backer and all but 2 unioned successfully. But I still can’t make the whole thing come together :upside_down_face:

Well if booleans fail you should run Intersect to check what you get, it should be one or more nice closed curves. And then depending on the problem manual splitting and trimming may be necessary.

Hi @aversteeg,

Sorry, late to the party here, but may I ask how the geometry on Layer 2 relates to the solid you are attempting to achieve? Is it for an overlay, or something to be incised into the surfaces, or what? I ask because it may have a bearing on the solution to your problem.


Hi @pascal,

This poster’s problem intrigued me when I came across it so I did some investigation. I successfully built a closed polysurface in spite of having several instances of what I believe you are describing as non manifold edges (i.e. four faces coming together at a single edge). Furthermore, ShowEdges shows zero non-manifold edges for this model. Here’s the object with some of the possible non-manifold edges indicated:

Is it possible that now Rhino is getting older the devs have made it handle this scenario without telling you? :wink: Or am I misunderstanding non-manifold edges?


Because we are discussing non-manifold edges, here’s one that’s been sitting in my files:

Problem (1.2 MB)

BooleanUnion RED+BLUE and you get non-manifold edges,

Do BooleanDifference (BLUE-RED) (No delete), then Boolean Union (RED+BLUE) and all is well in the world.

Alternatively, you can do RED+BLUE and manually extractsrf and get rid of redundant surfaces.

Hi Jeremy - I guess it is possible to join ‘in order’ and not get non-manifolds.


You might want to revisit that - I still see two non-manifold edges:

So both approaches fail the same way.

However, make the document tolerance one order of magnitude smaller (0.0001) and both approaches result in a good union with no manifold edges.

(Using Rhino for Windows, 7.9.21215.13001, 2021-08-03)