Shell command still unreliable even for simple geometry?

Hi,

I’m using the Shell command more and more. It’s rather handy for both architectural and 3D-printing applications. It’s a neat feature, however even after the latest update, which promised more reliability, it’s still rather flimsy even with simply geometries.

In the above example, none of the operations where successful.

I know how to fix the geometries manually! No assistance is required.

shelll-fail.3dm (3.9 MB)

Hello- thanks, that is a good one for the developer to look at.
RH-69324 ViewCapture: Some materials transparent

-Pascal

2 Likes

dear @diff-arch
what detail do you expect, the _shell command should deliver for those corners ?
(red points)


just for others that need a simple _shell - workarround and dig in the forum:

(1 - simple)

sub-select (Mac: cmd-shift-klick, Pc: ctrl-shift-klick) the surface for the opening
_offsetSrf (solid=yes)
(the flange is perpendicular on the faces surrounding the opening)

(2 - maybe nicer flange)

_extractSrf (copy=yes) (the face to open)
_offsetSrf (solid = yes , above face, Distance as much as possible)
_offsetSrf (the input geometry, solid = no, the shell-thickness, flip normals inward)
_createSolid ( initial closed Surface + both offsetSrf’s results)
delete unneeded geometry

if createSolid fails - try trim join.
if further problems try with corner setting of _offsetSrf round or sharp

the flange is perpendicular to the removed face / opening

-tom

Two of the three examples could be closed with the cap command and I’d expect the shell command to try to cap open edge loops.

(i edited my post above)

maybe the _shell command needs an option how the flange (“edge of thickness”, connection between outer and parallel inner surfaces)
flange option

  • automatic (let the algorithm decide from below, for each single edge - i think this is what happens now)
  • use removed face / perpendicular to removed face
  • perpendicular to surrounding faces

(perpendicular to removed face)

(perpendicular to surrounding faces)

but i think evaluation should also be done on more complex, curved (poly) surfaces.

1 Like