[ODE] CG solver makes contacts sticky

[Patrick Enoch]

To answer the question myself:

the CG solver does not clamp the lambda values (which is limited to  
lambda>0 to drive the bodies apart). If I put this into the loop,  
everything is fine, though contacts are really jumpy then.

Patrick

On 15. Apr 2007, at 1:35 Uhr, Patrick Enoch wrote:

> Hello,
>
> the CG solver makes the contacts very sticky, even if I set them to
> "perfectly sliding". How is that?
>
> The SOR solver returns with an "lambda error" of about 1e+4. The CG
> solver returns with an error of 1e-8. (Both with a setting of 100
> iterations)
> When I set the iterations to 1..5, the CG solver produces "slippy
> contacts".
>
>
> Is it necessary NOT to solve the system of equations precisely?
>
> Patrick
> _______________________________________________
> ODE mailing list
> ODE at ode.org
> http://ode.org/mailman/listinfo/ode