Angular damping lower threshold?

[Leadwerks]

When angular damping is set to any number less than about 0.000039, behavior seems to revert back to the default (higher) damping value. Why is this?

[Julio Jerez]

Are you saying that if you do this?

angularDamp (0, 0, 0)
inline void dgBody::SetAngularDamping (dgVector angularDamp)

then if you do this
GetAngularDamping (angularDamp) const
you get.

angularDamp (0.000039, 0.000039, 0.000039)

[Leadwerks]

No. It still returns the correct value. I am saying that if I set angular damping to 0.0, a significant damping effect is seen. I tested this by setting the omega to (45*60,0,0) (in radians) and updating with a time step of 1/60.


1. Set angular damping.

2. Set omega to 45*60,0,0 (converted to radians)

2. NewtonUpdate(1/60)

3. Get the body matrix, and retrieve the rotation (converted to degrees):

When angular damping=0
resulting rotation is about = 35,0,0

When angular damping=0.001
resulting rotation is = 45,0,0

[Julio Jerez]

what could you possible have spinning 54 radian in one sixtiest of a secund.
and object like cannot be simulated at 60 fps, you will need maybe few tousand fps.

[Leadwerks]

It shouldn't really matter, should it? I just set it to a number that was easy to test. That's actually degrees, and the routine converts it to radians. The point is below some number, the simulator treats the damping differently. I am guessing this is something like a default value that gets used if the damping is close to zero?