Generalized alpha time dependent solver

Hi,
if I switch from BDF to the generalized alpha time dependent solver the solver works with a timestep of 0 or near zero. I can set a manual time step and the solver calculates the time correctly, but the timestep is always 0, too.

That's why the algorithm crashes in cases of automatic time stepping.

Is there anybody who has some experiences with the generalized alpha time dependent solver? Are there any incompatibilities to other settings? BDF runs correctly. I want to test g. alpha to improve performance.

There you can see the log with time step 1:

Time-Dependent Solver 1 in Solver 3 started at 25-Jan-2013 12:06:34.
Time-dependent solver (Generalized-alpha)
Number of degrees of freedom solved for: 23185.
Nonsymmetric matrix found.
Scales for dependent variables:
mod1.u: 0.5
mod1.c2: 5e+002
mod1.mc2: 5.7
mod1.c1: 4.5e+002
mod1.mc1: 4.6
mod1.p: 0.36
Step Time Stepsize Res Jac Sol Order Tfail NLfail
0 0 out 3 4 3 0
1 0.1 0 5 6 5 2 0 0
1 0.1 0 out 16 17 5 2 0 10
Time-Dependent Solver 1 in Solver 3: Solution time: 19 s.


Thanks,

Best regards, Michael