[ODE] jitter

Nate W coding at natew.com
Sun Jan 19 22:33:01 2003


What are you trying to accomplish with the code below?  What do the
->rotate_[xyz] functions do?

Crashing with a .25 second time step may be because during that
quarter-second, things evolve to a situation where they cannot be
resolved.  Four frames per second is a *really* low frame rate, even if it
didn't crash I'd expect things to behave very strangely.

On Sat, 18 Jan 2003, Michael Powers wrote:

> thanks for the postings - i tried the same settings as
> chris' code but still get the non-stop spinning of
> boxes after they land on the ground - the contact
> points seem to be generated correctly - i am getting 3
> contact points
> 
> i am using the quaternion calls to get the rotations -
> then converting them to euler angles - it couldnt be
> that the quaternion code is returning the last
> recorded rotation everytime could it? here is my
> conversion code
> 
> sqw = q1[0]*q1[0];
> sqx = q1[1]*q1[1];    
> sqy = q1[2]*q1[2];    
> sqz = q1[3]*q1[3];  
> block_ptr->rotate_y((float)atan2(2.0 * (q1[1]*q1[3] +
> q1[2]*q1[0]),(sqx - sqy - sqz + sqw)));
> block_ptr->rotate_z((float)asin(-2.0 * (q1[1]*q1[2] -
> q1[3]*q1[0])));
> block_ptr->rotate_x((float)atan2(-2.0 * (q1[3]*q1[2] +
> q1[1]*q1[0]),(-sqx - sqy + sqz + sqw)));
> 
> other odd behavior - when i set the sim time step to
> 0.25 rather than 0.05 i get a crash (in the LCP
> solver)
> 
> so that means i get crashes when i try to change the
> timestep or when i try to turn soft CFM on
> 
> 
> 
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
> http://mailplus.yahoo.com
> _______________________________________________
> ODE mailing list
> ODE@q12.org
> http://q12.org/mailman/listinfo/ode
> 

-- 

Nate Waddoups
Redmond WA USA
http://www.natew.com