Moment of Inertia possible issue

[Starfox]

I have a bunch of dominoes with a mass of 1 unit and for some reason they insist on reverting to the upright position - see video: https://vimeo.com/79563905

Why do they revert to upright position like that? For collision they use a btCompoundShape with one child, a btBoxShape (basically a shifted box collider since the origin is at the domino's bottom). The moment of inertia calculated by Bullet seems to match the one I calculated by hand and using either doesn't seem to make a difference. What could be wrong here?

[Jonathan]

I believe the origin of the btCompoundShape becomes the center of mass. (So you've got a low center of gravity in your situation.)

[Starfox]

I see. How would I fix that?

[Jonathan]

I'm not sure how to change the rotation origin to differ from the center of mass... maybe someone around here knows?

Then again, maybe that's not a reasonable request, because isn't that the purpose of the center of mass? Naturally, an object would always rotate around its center of mass unless modified by another force? (Right? I don't claim to have a great knowledge of physics. )

[c6burns]

The center of mass is the local origin of the rigidbody. There's no "fixing" that. Just offset the object in your renderer instead of altering the origin in bullet.