Why Mesh Split does not work?

Hello,

All of a sudden Mesh Split stopped working.
Link to the file: https://drive.google.com/file/d/1_eQHIemlwmn1c2VwOBtstNf0wRkzrqIZ/view?usp=drivesdk
And for the future: how to upload files to the Forum?
Why?

Drag and drop or use the link symbol and choose your file you want to upload.

model.3dm (6.3 MB)

Thanks :slight_smile:

1 Like

Thank you but this is not the solution. It is my four cutting meshes that must be used because they are not oriented vertically. Anyway your cutting with a rectangle does not work too. It must be due to a program fault. Before the Mesh Split worked properly - even for bigger files. Now it works only for very small files. What could have happened? And it is not a space shortage (I have 6.87 GB free on :C).

do a single meshintersect for each of the 4 cutters.
check the results.

it looks like there are some artefacts in the mesh - from an earlier attempt ?

so let s

_check (command, that will analyse the mesh)

Mesh has 328 pairs of faces that intersect each other.
  This can cause problems if you're doing mesh boolean operations with it.

see also:

1 Like

Hello!

Many thanks. Once it came to my mind that it might be the tolerance vs. scale setting but I forgot it and you luckily reminded me of that. Interestingly it worked fine with a wrong mesh for many times and finally it refused to work. Rhino does not seem to be the best one to manipulate big irregular meshes.

Regards

The trick is always if the intersection involves one of the poor overlapping faces in the mesh itself.

Running Rhino 8.6 on the sample model returned this for me that looks correct:

  1. I am interesting to know more about how you do manipulate the meshes?
  2. Is there certain tools that are needed?
  3. Are these meshes imported with the problems already? What tools are there to deal with them?

We have been adding a lot of tools to try and fix poor meshes. But at some the mesh structure can be a mess.

Self intersection is the most destructive to booleans:

Mesh has 328 pairs of faces that intersect each other.
  This can cause problems if you're doing mesh boolean operations with it.

Hi Scott,

Thank you for contacting me.
The trick is always if the intersection involves one of the poor overlapping faces in the mesh itself.
It must be due to algorithm not working properly. Before cutting relief mesh I reduce polygons 80% and recently even one more time the same as it lasted for ages to cut the relief before (sometimes without success). As I was lucky not to hit any ovelapping faces before (God knows why?), I forgot another Forum user’s remark on the same topic and all of a sudden I was not able to cut anything. Now back to the re-triangulating algorithm: I think it should manipulate vertices without separating them from faces. If this condition is fulfilled, I don’t see the possibility for any overlapping faces. It happens, however, so I suspect it is solved in a different way :frowning:

Ad 1,2,3. I import a point cloud. In Grasshopper I use Delaunay Mesh component to get the mesh. Then I apply the “Reduce mesh polygon count” that produces a faulty mesh apparently. So my mesh is created within Rhino.

What is very important: my meshes are topo meshes so the triangulating algorithm must always give a clear result contrary to meshes approximating a solid geometry. To get a solid mesh is not always a crucial issue for the3D printing slicing software but this time I wanted to have a perfect watertight mesh as I want to send the model for someone else to deal with.

I use 2 programs that always produce watertight meshes from open meshes (extrusion with flat bottom): Meshmixer and AccuTrans3D. But sometimes simple extrusion is not enough to get the desired effect.

I wonder what is the philosophy of creating mesh repairing algorithms. Do you combine vector and “pixel/voxel” approach combined contrary to only vector one (as I think I might be productive)?

Kind regards

Maciej

This was the process I used. The voxel stuff for us is Shrinkwrap. But it does take a few steps.

  1. Offset mesh using user direction 12mm down.
  2. Turn on points, select only the bottom ones and set z to the same height. That should give the model a flat bottom,
  3. Run Shrinkwrap with a value of 0.25 or so. The result is super nice.
  4. Boolean split.

The original model sent ran in Rhino 8 without a problem. So, I think even with the problems in the mesh it worked.

But the Extrude/Shrinkwrap trick may mimic the workflow described above? It does create watertight meshes. I know it is a few more steps, although I wonder if a extrude to flat might be in interesting option to add?

Flat bottom? Maybe rather as included in the GH component “Extrude Mesh” ?

Let me think how it might be done in Grasshopper.

But even as a manually process does that seem like it might all work in Rhino?

If it always works properly in both programs I mentioned, why not?

@amsarse I’m late to the party here. I’m trying to figure out if there’s anything especially different we should look into. Do you also have a part of the GH definition that we could look at?

I simply meant “Mesh Offset & Extrude” GH component could have a flat bottom option.

1 Like

Filed at RH-81474.

Where is the “user direction” setting located?
“Shrink wrap” is probably not Rhino 6 (which I use) feature.

Hi -

In Rhino 8, you can set the direction method:

DirectionMethod ( UseVertexNormals UserSelectedDirection ):

-wim

Shrinkwrap is new in Rhino 8: