crucial settings with respect to fvSchemes and fvSolution to achieve convergence
 

I am trying to do RANS of pulsating flow in a 2-D diffuser.
I use pimpleFoam solver and the settings are copied from pitzDaily case.

While giving uniform inlet velocity I have a converged solution with a very small deltaT.
However when I run a simulation with pulsating inflow the simulation diverges.

I am using kEpsilon with average y+ 22.

What are the most crucial settings with respect to
fvSchemes and fvSolution
to achieve convergence (for different solvers as well)?