So, because Shell didn’t seem to work, and OffsetSrf didn’t seem to work, I decided to see if I could do a shell and miter of this polysurface manually.
Here’s what I figured out: there’s no way to make perfect miters at the vertices where four edges meet. A miter must bisect the crease angle or it isn’t a miter, and the opposing faces will have different widths, so there is no wiggle room or fudging to be done here.
What can be done is to compromise by setting priorities.
If you offset all these faces individually, and then scale them 2d, and then split them with planes that bisect the crease angles, you’ll get surfaces that extend too far or not far enough in the locations where you would want them to meet. You could just lop off the points that stick out too far, but if you do that the wrong way, and you’re working with sheet stock, you’ll end up exposing the interior of the sheet stock.
But if you do it right, you’ll get some odd looking angles while maintaining the veneer all the way around the object, like this:
What this has illuminated for me is that the reason _OffsetSrf and _Shell glitch out so often is pretty straightforward: the tasks they do are hard and have a lot of variables. This is also probably why we don’t see a lot of complex plywood origami out there in the real world. Also: the document tolerance for the original file may be too big.
Anyway, here’s the file, with all the miters done:helpMe.3dm (255.6 KB)
CNC routing this will require more than a 3 axis machine, and even on a 5axis machine, you’ll need steep v-bits to get the inside miter corners right.
IF this had been all quads, it would’ve been much, much easier.
Also, if anybody knows of any grasshopper definitions or threads that try to tackle this kind of task, please let me know. Plywood origami interests me very much.