Pointed cone

Hello everybody,

I want to define a 3D equivalent of the 2D-pacman geometry,
i.e., a ball where a pointed cone is cut out.

I use EllipticCone from netgen.csg (see attached file).
To have a pointed cone, one would choose r=0.
However, as the help() function already tells me
"when r = 0, the top part becomes a point(tip) and meshing fails!"
--> And it does fail, when I try to mesh the 3D-pacman.

If choosing r small but positive, e.g, r=0.02,
then that small surface circle has to be resolved by the (surface) mesh,
and the mesh size at the back of pacman's mouth becomes really small.
(Of course, the same problem applies in 2D as well, if pacman's mouth is not pointed, see 2D-meshes in attachment)

I see the problem with meshing a pointed cone.
But I don't want to mesh the cone, but a geometry, where the cone is cut out.
This should not be a problem, because the "infinitely small circle", which is a point, does not have to be resolved by the (surface) mesh.
Moreover, like the perfect 2D pacman, the perfect 3D-pacman should allow for uniform mesh size as well.

Summary: Is it possible to mesh a perfect 3D pacman?
(Perfect, meaning that it's mouth is a cut out pointed cone.)

Any help is welcome,
Best, Carl

No reply in 5 weeks --> up.

you have observed it: meshing the inverse of the cone has the same difficulties as meshing the cone. Curvature is unbounded, at the tip. You have to cut the tip.