[ODE] 3 out of 4 contacts box->plane?

[Jon Watte (ODE)]

Those values look totally normal for an "almost resting" situation. If 
you really want it to settle, make sure body sleeping is turned on, and 
that you apply rotational inertial dampening each time step.

If there are no flags set on the contact, then Mu is a force limit (in 
whatever your force unit is -- likely Newtons if you're 
meters-kilos-seconds).

If the contact friction model flag is set, then Mu is a dimensionless 
friction coefficient.

The manual has more information.

Cheers,

			/ h+


Kevin Quigley wrote:
> This is what I'm seeing, looking down (-Z axis):
> 
> got 3 contacts
> contact 0 at {0.235, 0.1, 0}                            X-------X
> contact 1 at {0.235, -0.1, 0}                           |          |
> contact 2 at {-0.235, 0.1, 0}                           0-------X
> world time: 0.01
> got 3 contacts
> contact 0 at {-0.235, -0.1, -1.0791e-11}         X-------0
> contact 1 at {-0.235, 0.1, -5.39551e-12}	|          |
> contact 2 at {0.235, -0.1, -5.39551e-12}	X-------X
> world time: 0.02
> got 3 contacts
> contact 0 at {0.235, 0.1, -1.18701e-11}	       X-------X
> contact 1 at {0.235, -0.1, -1.0791e-11}         |          |
> contact 2 at {-0.235, 0.1, -9.7119e-12}	        0-------X
> world time: 0.03
> got 3 contacts
> contact 0 at {-0.235, -0.1, -1.7913e-11}        X-------0
> contact 1 at {0.235, -0.1, -1.40283e-11}        |         |
> contact 2 at {-0.235, 0.1, -1.33808e-11}       X-------X
> world time: 0.04
> ....
> 
> I tried setting mu = dInfinity, but didn't change anything.  This is
> just a static box sitting on a plane. e.g.  box -Z side is == plane.
> 
> What is the unit for mu btw?  Is there an SI used there?
> 
> thx,
> quig
> 
> 
> On 10/10/07, Jon Watte (ODE) <hplus-ode at mindcontrol.org> wrote:
>> 3 Contacts will give stability. However, depending on stability, you may
>> end up with 2 contacts on one side one frame, then 2 contacts on the
>> other side next frame, then bouncing back and forth -- usually too small
>> a movement to be actually noticed on the screen.
>>
>> So... are you sure what you're seeing is wrong?
>> Btw: you're setting mu to 0, which means no friction in contacts. That's
>> usually a bad idea for stability.
>>
>> Cheers,
>>
>>           / h+
>>
>>
>> Kevin Quigley wrote:
>>> I'm seeing 3 contact points generated by a box geometry sitting on (or
>>> through) a plane.  I can't figure out what's gone wrong in my
>>> callback.  This is a mixture of various bits I found in the
>>> wiki/manual.  Any ideas?
>>>
>>> The three contact points are not always the same, either.  When drawn,
>>> some or all of them flicker as one contact is not drawn during that
>>> step.
>>>
>>>
>>
> 
>