Union issue

roger_m_kolaks

In the linked document I began what I thought would be a simple task, making a couple of triangular holes through a polyhedron.  Began about noon. It is now 18:00.

I attempted several methods to create the holes: Lofting Remove, Lofting to solids then Subtracting, Slicing the solids into chunks and Subtracting separately, Slicing through the polyhedron with planes, and Unioning everything but the holes, slicing the target to create a plane surface then reattaching after making the holes.

In case I could create one of the holes but not the other.  It didn't matter which I created first, the second would always fail.  The closest I came to success was to Loft the target holes to Solids performing a Union on the them, making a flat target surface by slicing away the stellated vertex, Subtracting the joined solids from the target polyhedron, and removing the appropriate parts of the stellated vertex.  But when I attempted to Union those two parts to the remains of the polyhedron the Union always failed with the second part.  Again one Union would work and the other would fail, in whichever order I chose to make the Unions.

Could someone offer me a clue as to the cause of this behavior?


https://cad.onshape.com/documents/dfd6aaeb68733003b3cc920d/w/eea8447ca66f41174e99bf62/e/a0fd435e6e18c68ff059e7f8

_anton

Possibly a non-manifold geometry issue.

Could you clarify what you're trying to do? Are you trying to just remove one triangle out of that triangle fan at the vertices? In which direction do you want to cut the hole? Should it join to a triangle at a neighboring tetrahedron vertex?

A picture would be worth at least 100 words here. 

roger_m_kolaks

It isn't obvious.  🤣  🤣 Sorry.  I spent hours trying to find out what is happening.

Okay.I was looking at intersections and subtractions from polygons.  In this case the triakis tetrahedron and the tetrahedron.  There is a green pentagonal  face in the drawing.  Opposite it is a face composted of six triangles meeting at a vertex.  Adjacent the green pentagonal face are two triangular faces.  

There are two shapes (named prism1 & prism2)  that connect the triangular faces adjacent the pentagon with two of the triangular faces in the group of six opposite.  There are also two shapes (named triakis green & triakis blue).  These two shapes were created by slicing off the vertex with the six triangles by a plane through the bases of the triangles then removing the two sections that correspond to prism1 & prism2. Looking into the truncated triakis tetrahedron normal to the cutting plane are two red faces.

In the current state triakis green & triakis blue cannot both be Unioned with the large shape though either shape can be.  An attempt to union the remaining element will always fail.

This is the current state of the investigation.

Initially I attempted to individually Loft Delete between the two triangles adjacent the pentagon and the respective opposites.  That failed.

Then I Lofted the same faces to New shapes (prism1 & prism2) and attempted Boolean Subtracts.  That also failed.

I then then sliced the prisms at the various planes where the Booleans failed. That's how I cut the two holes.But I could not successfully cut through the section with the six triangles.

I then sliced off that vertex and removed the two target triangular solids.
 I was then able to cut through the remaining portion of the truncated triakis tetrahedron.  I then tried to Union the two elements (triakis green & triakis blue) with the larger shape. That failed.

I then deleted the two red faces and rebuilt them and again tried the Union.  It failed. 

This is the current state.

You'll need to play with the document to clearly understand the problem.  Attempt to Union both trakis green and triakis blue with tracks trunc stellate.  No matter which order you choose the first item will make a Union but the second will fail.

There are two Named  Views that focus on the issue.  Rolling back history  also might provide some insight.

Ergo, I didn't post any images because the document is the images and I was too tired at the time to create a video.  In any case, I would have had to have begun the video when I first encountered the problem.  I guess I just lack foresight.  🤓  😄  🤣 

     ..rmk

S1mon

Take a look at this:
https://forum.onshape.com/discussion/20585/better-error-handling-and-training-around-zero-thickness-parasolid-limits

Basically Parasolid can’t handle two shapes joining at a point or a line.

roger_m_kolaks

It isn't obvious.  🤣  🤣 Sorry.  I spent hours trying to find out what is happening.

Okay.I was looking at intersections and subtractions from polygons.  In this case the triakis tetrahedron and the tetrahedron.  There is a green pentagonal  face in the drawing.  Opposite it is a face composted of six triangles meeting at a vertex.  Adjacent the green pentagonal face are two triangular faces.  

There are two shapes (named prism1 & prism2)  that connect the triangular faces adjacent the pentagon with two of the triangular faces in the group of six opposite.  There are also two shapes (named triakis green & triakis blue).  These two shapes were created by slicing off the vertex with the six triangles by a plane through the bases of the triangles then removing the two sections that correspond to prism1 & prism2. Looking into the truncated triakis tetrahedron normal to the cutting plane are two red faces.

In the current state triakis green & triakis blue cannot both be Unioned with the large shape though either shape can be.  An attempt to union the remaining element will always fail.

This is the current state of the investigation.

Initially I attempted to individually Loft Delete between the two triangles adjacent the pentagon and the respective opposites.  That failed.

Then I Lofted the same faces to New shapes (prism1 & prism2) and attempted Boolean Subtracts.  That also failed.

I then then sliced the prisms at the various planes where the Booleans failed. That's how I cut the two holes.But I could not successfully cut through the section with the six triangles.

I then sliced off that vertex and removed the two target triangular solids.
 I was then able to cut through the remaining portion of the truncated triakis tetrahedron.  I then tried to Union the two elements (triakis green & triakis blue) with the larger shape. That failed.

I then deleted the two red faces and rebuilt them and again tried the Union.  It failed. 

This is the current state.

You'll need to play with the document to clearly understand the problem.  Attempt to Union both trakis green and triakis blue with tracks trunc stellate.  No matter which order you choose the first item will make a Union but the second will fail.

There are two Named  Views that focus on the issue.  Rolling back history  also might provide some insight.

Ergo, I didn't post any images because the document is the images and I was too tired at the time to create a video.  In any case, I would have had to have begun the video when I first encountered the problem.  I guess I just lack foresight.  🤓  😄  🤣 

     ..rmk

roger_m_kolaks

This was the answer.  A bit of reading on non-manifold solids helped my understanding of the issue.  Then some experimental elimination of suspect vertices made it even more clear. 

Thank you for the reference.