[ODE] Joint Stops and Instability

Gene Ruebsamen gene at erachampion.com
Sun Feb 10 10:41:02 2002


Hello All,

I'm having a problem with instability in my simulation.  I have several
different bodies connected via joints (ball & socket and hinge joints).
Right now, I have no forces acting on the bodies except for gravity. All
joints have friction (dParamVel = 0, dParamFMax = 0.01). 

The problem I am experiencing is when my simulation starts, I am
dropping my bodies from a height of (z=4). When they fall and hit the
ground, some of the joints reach their Lo or Hi stops.  When they do
this, the simulation becomes unstable, and the bodies start to spin 
violently and shoot off into the sky!!

If I drop my bodies from a lower altitude (z=1.5).  The joint stops seem
to work properly, and my bodies end up in a heap on the ground (expected
behavior).

Based on observation, it seems that when the joint stops are hit hard,
the simulation becomes unstable.  I often get the following error
message when the simulation becomes unstable:
ODE Message 3: LCP internal error, s <= 0 (s=0.0000e+00)

However, when the joint stops are reached without much force, everything
seems to work properly.  Can anyone help?  I tried playing around with
the dParamStopERP to no avail. I tried 0.1 all the way up to 0.9 w/o any
luck.

Thanks!

Gene