any idea why this does not work?
it should be fixed at the top, but the rest of the chain should straighten out vertically if i press the play button, or not?
I think your issue is that the objects aren’t members of the “RigidBodyWorld” collection. When enabling rigid body they should be added to that collection automatically. You can check in the object properties panel.
More generally though, it’s not advisable to use the ‘Mesh’ collision shape as it’s unstable and performs poorly. Instead, when you need to simulate non-convex geometry you can create a compound using child meshes.
you changed the size of the chain to get the child meshes done.
it also works when i put all in the “RigidBodyWorld” collection as you said and scale my mesh up 10 times and apply scale…but quite slow…
so with small elements blender as it seems has got a problem?
would be better it that works it my scene at the original scale…BAD…
hopefully someone knows a “scale problem” solution
to make all these additional child meshes does not work in my case…
I didn’t put any thought into the scale of the demo scene, but its main relevance is in believably (appropriate gravity) and potentially needing additional sim substeps to account for objects moving very fast. Also, if the chain isn’t the main focus of your scene, do the simulation in a different file and import it as a mesh cache.
That’s entirely because you’re not using proxy geometry like I demonstrated. I cannot overstate how awful using mesh collision shapes (concave geometry) is. It’s not a Blender issues, it applies to all CG software.
thanks for looking in and giving a good solution.
i clearly understand all the points.
i think about your idea. so if no other person comes along with a fast and easy solution i will maybe reconstruct with that additional geometry.
…worked good with original geometry. exported lowres passive rigids and my “highres” active rigid chains and just scaled up 10 times. small sensitivity margin of 0.001 to get fast calculation. in scene properties i screwed the substeps up to 40 and the solver iterations up to 100. so i got stable animation without exploding or braking chain. calculation was only about 2 to 3 minutes. then i baked all dynamics and then “baked all to keyframes” and linked that collection back in my original file. there downscaled 10 times…looks good. always the same in blender…without upscaling 10 times blender has problems. i hope they will fix this as soon as possible…
thanks for feedback, but i have architectural scenes with all real scale elements, much of them appended. so all the hundreds of elements have real world sizes. i cant have chain elements with 4x4x0,8 mm , i dont make jewelry animations. or maybe i dont understand you exactly…for me calculations should definitely work in every size and scale !
yes, exactly, and thats the exact size the chain elements should have , 30mm x 50mm x 10mm not 4mm x 4,5mm x 0,8mm …and with that size i dont get a working calculation.
that this animation works with other sizes and scales is nothing new, i showed already in my fiirst post. maybe you want to check the file chainanimb.blend (4.0 MB)
the first problem was that the chain elements objects were not members of the “RigidBodyWorld” collection. solved by [init_pixel]
second was that calculation gets faster (if you cant reduce the sensitivity margin ) with reduced collision geometry. solved by [init_pixel]
third is element size and working physics calculation…
maybe you post a blender file with a working calculation with my chain elements and my scene scale. maybe i can learn something then.