Constraining shell particles

Hi all,

GROMACS version:2019.1
GROMACS modification: No

I’m trying to work with polarization on my system. Something would go wrong while trying to run the simulation with a constantly increasing RMS force after 50 iterations for 20 steps. I am trying to restrain the shell particles in position at a distance of 0.02 nm as suggested in previous literature, but I keep getting a floating point exception. I have two questions:
1- Has anyone ever tried to constrain shell particles and what is the process to do so.
2- does GROMACS have a minimum value for constraints that I cannot go below?

Thanks for the support,