cfdemSolverPiso: decreasing particle size induces divergence of CFDEM coupling procedure

Dear all,

I would like to submit to your attention an annoying problem I am not not able to solve.

I built a simple geometry, similar to the cylindrical domain used in ErgunTestMPI, where the initial conditions
are represented by a certain amount of steel particles settled at the bottom of the cylinder. The CFD flow enters
the bottom side of the domain, and exits from the upper side.
What happens to me is that with a particle size of 0.008 mm (diameter) the coupled simulation performs well, but
reducing the particle size a little bit (passing from 0.008 to 0.006 mm), the simulation blows up almost immediately,
with particles which accelerate enormously and goes out from the domain.
I tried to reduce the Young modulus, and at the same time to reduce the timestep, in order to satisfy the relation:
timestep < 0.2*Rayleigh time, but nothing changed.

I would like to ask you if someone could give me any hint about how to solve my problem.

I thank you in advance for the attention.