Transparent surface created - has anyone seen anything like this before?

Rhino Rhino for Windows
, , ,
1

I have a problem with a surface that looks ok from a side angle, but when you pan round you can see through it. Has anyone seen anything like this and/or know how to fix it? My fallback position is to replay the steps I did from the last known good position and check at each stage, but there is a few hours of work in that.

To get to my current point, I started out with a SubD solid, then converted ToNURBS and then did some boolean differences to create cutouts where other objects will fit. At some point along the way, in one of the cutout bits I ended up with a surface that looks like it’s there (because the cross lines on the surface can be seen) but then it’s possible to see through it and inside the shape. See screen shots below.

I’m able to create a curve around the border of this surface and generate a surface from planar curves, but unfortunately Rhino says it can’t do the Edit, Join when I try to join the new surface to the rest of the solid. I also tried to create a solid by extruding this curve a short distance, capping the planar holes and then doing a solid union but that failed too.

Rhino does not think it’s a bad object when I do SelBadObjects.

When I try to convert to mesh something interesting happens that may be a clue as to what is going on. In this bit of the model a partial surface is created that does not extend to cover the entire opening.

If I export to STL I get much the same result and the downstream slicing software can’t process it.

Angled view shows the lines crossing the expected surface
image

Straight on can see right through the surface
image

Mesh From NURBS object showing smaller surface in the middle of the opening
image

2

typically a bad trim that causes the render mesh to fail.

run selbad. whatever lights up run isolate.

then run extractbadsrf and click on whatever lit up previously. the surface(s) that come out, run isolate on them and you will be zeroed in on what the issue is.

typically you can untrim, keeping your trim curves and then isolate the trim curves. The examine them closely and look for any issues in the trim curve… a common issue is control points that are crossed up causing a bowtie or a very tiny segment that gets collapsed.

if you get stuck, post the part and I’ll take a peek at it, or email it to us a tech@mcneel.com

3

Thanks very much, Kyle. It’s actually reporting no bad surfaces or objects when I try those. Looking back through my model snapshots as I’ve been building it up, it looks like that surface was ok when I first boolean diffed it, but then looks to have gone see through when I differenced other cutout bits nearby.

I’ll email the part to see if you spot anything and it not then I’ll go back to the plan B to roll forward from that last good snapshot.

Thanks again

4

that happens when the boundary trimmed edges are selfintersecting. extract the surface, untrim it and use the object to trim it again

5

Thanks a lot Diego. That comment got me over the finishing line.

For the benefit of anyone reading, in slightly more detail the steps I used to solve were

  1. ExtractSrf then carefully click so as to select that jigsaw shaped surface and not the whole solid
  2. Untrim. It asked me to select a line, so I selected one of the curved lines toward bottom left. The surface then turned into a rectangular kind of shape.
  3. Create a Curve from Face Border, to create the jigsaw shaped outline.
  4. Add two points on the jigsaw curve at the bottom and top then Split the closed curve to obtain the single line I need for the trim
  5. Trim the surface from 2 to remove the protruding bit
  6. Now Edit, Join the surface from 5 and the rest of the solid. This created a closed polysurface.