How to model this

Hi everyone, I wnat to model something like that and im getting crazy with booleans.
Everything I try fails…

Any ideas?

Boolean differences are definitely one way to think about this. I’d need to see what you modeled to say for sure why it was failing. One possibility is that you need to do the differences in separate steps.

Here’s another approach too using surfaces and trims…
IntersectingGrooves_bjames.3dm (944.5 KB)

Hello, thank you very much for the anwer.
Your example is not exactly what i want…
When i make the cut or the intersection I obtain an incorrect curve just where the two pieces become tangent
I attach a photo and a rhino file to show you my problem

test.3dm (1.3 MB)

This was partly due to the unit tolerance of your file being changed. You had it set at .1 mm but I think an object of this size with junctures like this will work better at the defaults of .001. The first thing I noticed was the bad edges in the screenshot below. This likely resulted from the unit tolerance change too but I’m not sure when it happened. I exploded that polysrf and used RebuildEdges then I joined again after setting the document to .001.

That won’t be enough to get a good intersection curve here though. Coincident surfaces are known for this type of headache and the best you can do is decrease the area in which they overlap perfectly and also use surfaces with enough control point info to get an answer. Here I split out a section of the straight surface shrinking it as I did so, then I rebuilt that new trimmed surface to have more control points and a degree of 3 in the long direction. This allowed me to get an intersection curve that I could use for the trim. There were still a few pesky naked edges in other areas present after the rebuildedges and join at a higher tolerance. I just deleted these and did a couple Sweep2 surfaces…
I hope this helps.
test_bjames.3dm (2.0 MB)

1 Like

Thank you Brian, It was very useful for me.
I understand that when a surface has more control points, its easier to get the intersection…, isn´t it?

Thank you again!

Yes, but this isn’t always the case. There are many variables involved such as the file tolerance, the size of the objects in question in relation to the model units, edge tolerances of the surfaces involved and coincident overlapping intersections. I wish there were a set of steps that always worked on all cases but in my experience the best you can do is learn more tricks so there’s always something to try next.

I don’t know everything though! @pascal is smarter and might have more to add.