increments

HI everyone,

I am actually pretty confused about how to determine the correct time step increment in ABAQUS in a non-linear analysis. I first faced this problem when I used Hashins damage to capture in-plane shear response of the composite . The response which is bilinear graph depends a lot on the initial time step I give for the analysis i.e., for some time steps there is no bilinearity , only a single straight line. This has led to great confusion in my mind whether my model has a problem or whether ABAQUS itself has this problem.