Hi Rudy,

There was something really weird with this geomtry ...

In this area, the thickness of this edge is suspect,


There are even 3 edges in fact


So I separate everything, to isolate this part


The strange thing is that this geometry here is unselectable ... so I only copy/paste the valid geometry in a new file ... rejoin everything


and now the boolean works perfectly

So weird construcion :)

Hi Rudy, looks like a couple of things were just slightly off on your initial curves:



Zooming in on the area marked #1 above, the profile from the side view is not actually aligned to the world axis, it is slightly off to the side from it:



But actually the bigger problem is the area marked #2 above - the rail path curve is not closed, it is an open curve that has its ends crossing over each other, kind of like an X at the end instead of it being closed:



That causes the resulting object to not be a solid, instead it is a non-closed surface that has some part of it intersecting itself, that is probably what is making the booleans work strangely in that area.

Hope this helps,

- Michael

Hello Michael,
Thank you very much!

Sometimes I give for granted the operations I do (mirror, etc) while I should really pay more attention.

I have noticed those 3 tiny edges, but though since the shape is closed, there are no problems....Anyway, I got your points. Beautifullly explained.
Thank you to take your precious time for me.
APPRECIATED very much!

Best,
Rudy

Hi Rudy, no problem!

Getting the mirror slightly off can tend to cause a bunch of problems later on.

Sometimes the cause is not that the mirror line is off, but rather that the curve itself is slightly off the world axis. Any combination of the mirror line or the curve placement can make that happen.


When I get a chance to add some more kind of warning messages and error diagnostics, that one will probably be a good specific area for me to target first.

It can be sometimes difficult to add in comprehensive warning messages because it can take a lot of processing time to very thoroughly check everything for some conditions like self-intersections. But I think for curve endpoints the processing time would not be a problem and would probably catch the largest number of things.

It may be a while before I can focus on that area though. But there is another thing that is coming up shortly though that will probably help, which is this object properties panel that will give some information about the selected object. I'm hoping to set it up so that it will let you check at a glance whether a curve or surface model is actually closed or not. It can be hard to tell by just looking at the curve directly because you can't really see the little crossing-over ends like in this case without zooming in a fair bit.

- Michael

Hi PaQ,

> The strange thing is that this geometry here is unselectable ...

Just a quick note - this typically means that the object has a surface in it that has degenerate trims - that is that the trims are totally squished down flat together making the active part of the surface inside the trim a kind of infinitely thin slice.

That kind of thing will cause the mesher to fail, and the regular mouse-picking selection uses that mesh data so that interferes with selecting those squished things.

But they can be selected by doing Select All, or Invert.


Anyway, that's what is going on with pieces like that, that look like curves but are actually squished degenerate surfaces.

- Michael