Thread 1013339

This is rough sketch/blockout of something I need in my scene. Rings have to be empty inside ofc. What do you think are the best ways to model this? The first thing that comes to my mind is to keep arcs and rings separate and simply boolean subtract inner space of rings to get rid of arc segments inside of rings. This approach is clunky, inaccurate and I think it looks and shades problematically in places where rings connect to arc. But I'm not sure of better ways. Maybe do it in lowpoly and try to manually connect arc mesh and ring meshes by heavily messing with geometry by hand? Sounds doable, but quite inaccurate and complex. How would you do it?

>>1013339 (OP)
took about 15 mins
>circle
>extrude to ring
>cut it down to one quadrant
>use the cutter locked to one axis to get one of the end's verts in line and snap the others to it
>duplicate to make a half-ring
>duplicate that to make the next ring along
>dupe two more rings
>extrude the connecting piece
>modifier curve deform along a circle
>add ring cuts until its smooth

>>1013344
Thanks, I explained it poorly ("Rings have to be empty inside") I guess since that inner part of rings isn't needed, but in your picture it looks like it's really easy to just remove it so it still looks great.
> use the cutter locked to one axis to get one of the end's verts in line and snap the others to it
Wdym by "cutter"? Do you cut the big ring here?
> duplicate to make a half-ring
Do you refer to small rings as half rings? I made rings in OP by doing cylinder, selecting top and bottom, inset and then bridge faces. Not sure if there is problem with this, or it's just two ok ways to do the same thing.

>>1013345
here, this might clear things up a bit
>first thing, you can see im using the knife tool
>use it locked to the X axis to cut across the right quadrant on its far right segment
>now i can snap to that point, which will line up nice square corners for when i run the bar through the rings
the other cut is just to show the tool in case you havent seen it before or its not clear what im talking about

>>1013347
Hmmm, I think I get it, is it related to those parts?

>>1013348
yeah thats the idea. its about getting it square so you can just extrude or add faces later

>>1013349
Thanks, I get the idea. Another big missing piece for me here was curve deform, solves those alignment inaccuracy issues.

>>1013351
yeah, array and curve modifiers solve a lot of problems. glad i could help!

>>1013339 (OP)
100% procedural

>>1013339 (OP)
I can think of 3 ways of doing it.
1. With 100% geometry nodes.
2. With geometry nodes and warp/deform modifier
3. With applied modifiers from above to get the base shape + manual tweaking.

But I would need to actually try that out and confirm if/which methods works best. I might do some experiments and post results later.

>>1013339 (OP)
>>1016005
It worked out quite nicely and is fully procedural. Everything can be changed and animated - size, ring count, offset, rotation, resolution etc.
This is probably not the best method, because it is susceptible to glitches due to the mesh boolean inside geometry nodes. Obviously this method doesn't produce the best topology, but you can get away with it. :D

>>1016041
I am tired of topology purists, if it works it isn't stupid and if you're not going for hyper-realism those minor lighting glitches are negligible. Looks good enough anon thanks for sharing

>>1015999
this is making me go huh