Your right to expect a more stable system when using smaller time steps. I don't quite understand your observation. It is possible that the "1/2" time step used hits on a (calculation)harmonic in the equation set. If so, the instability should not appear when the calculations are...
