Spheres in a pile spinning in place?

[graceful nexus]

I've created a pile of spheres (with some walls around them so they can't role away) using regular SphereColliders and Rigidbodies, and then dragged them into a subscene, when I hit play, this pile of spheres behaves pretty weirdly. Rather than the spheres resting in a pile nice and still, many of the spheres seem to spin in place, as if they are all infinitely slippery. Their positions remain constant, but their rotations are changing at a constant rate. I don't have any code interacting with these entities.

This doesn't happen with regular PhysX gameobjects, but I'm seeing this for entities with both UnityPhysics and Havok... what gives?

I've tried adding a PhysicsStep component, increasing the solver iteration count seems to make them spin slower (but doesn't eliminate the spin entirely), changing "synchronise collision world" or "EnableSolverStabilizationHeuristic" doesn't seem to have any effect, neither does toggling multithreaded

I've attached a pic (just imagine this, but some balls spinning in place!), easily reproducable in a fresh project

[wet pasture]

Honestly I'm surprised you see it with havok but would expect it under unity physics due to its stateless nature
If it was just Unity physics I would have suggested enabling the stability profile option or whatever it's called and see if it helps.

[graceful nexus]

Yeah, doesn't seem to be any different with havok, really strange? By "stability profile" do you mean EnableSolverStabilizationHeuristic? That specifically mentions the stability of piles, but just turning it on doesn't seem to fix it...that said, I've dug around in the docs and found some additional physics heuristics you can toggle programmatically, these look promising, will try and report back

https://forum.unity.com/threads/unity-physics-0-4-0-contact-solver-stabilization-heuristic.914957/

[wet pasture]

I've found bumping iterations up by just 1 can help certain physics quirks as well

[graceful nexus]

Yeah, increasing iterations definitely helps, but the jank is definitely still noticeable. Changing the heuristics hasn't seemed to have made a difference either :/

[worn lion]

Looks like a potential scale issue to me. How big are those spheres?

I've also seen this sort of behavior in other engines with iterative solvers before. Adding a rolling resistance torque (to mimic rolling friction) helps a lot.
You can try adding a torque to the spheres as explained in section 4.3 here: https://diglib.eg.org/bitstream/handle/10.2312/PE.vriphys.vriphys09.021-030/021-030.pdf?sequence=1&isAllowed=y
See equation 3 and figure 5.
That worked well for simulating piles of spheres in the engine I was using at the time this was published.

@graceful nexus What could help here also is just filing a bug (https://unity.com/releases/editor/qa/bug-reporting) and attaching this project to help reproduce it.

[warped sonnet]

(and then post the IN-***** number here so we don't lose it)