[ODE] Sphere getting stuck on a single contact

[Michael Lacher]

Michael Lacher wrote:
> Hi!
> 
>   I am trying to integrate ODE (0.035) with our 3D engine to provide for 
> built in physics. So far everything worked very well, but now i have a 
> small problem:
> 
>   I create a sphere body (body with dMassSetSphere) and drop it onto a 
> bigger static sphere. The sphere bounces off the static sphere as 
> expected, but then, when it comes to rest after a few bounces and it 
> should start to roll down it doesn't. It just sits there and spins very 
> slowly in the opposite direction ....
> 
>   I am using my own collision detection library and it works pretty 
> good, also i watched the contact position/normal in the debugger and it 
> looks perfect so I assume that the problem is somewhere else.
> 
>   The global ERP/CFM parameters are 0.2 and 0.0001 respectively.
> The contact has only dContactBounce enabled (with bounce of 1.0f and 
> bounce_vel of 0.0f), and inifinite friction. As the bigger sphere is 
> static, the contact is created between the body of the smaller sphere 
> and 0 (the environment).
> 
>   Any help is appreciated :)
> 
> Mucki

Oh well ... the mystery is solved. For some reason I have not found out 
yet (but has nothing to do with ODE whatsoever) the bounding volume for 
the dropped sphere is a sphere with radius 0 instead of 0.5 ... 
resulting of course in the contact point be generated right at the 
center of the body. This of course cannot build up any momentum to make 
it rolling ...

shame on me for not checking everything before posting stupid problems.
Mucki