Hello,
I've noticed that in the test script you mixed up your dielectric constants: eps_out is the parameter for the dielectric constant between the plates, the "epsilon [list]" specifies your boundary. For metallic BC, it should be eps_out << epsilon [list]. Luckily, this mistake was canceled out by a second error, namely wrong signs in the normals of the ICC particles...
Further, we've found that in the force.cpp, few lines after the call calc_long_range_forces(); there is ghost_communicator(&cell_structure.collect_ghost_force_comm);
This ghost communication is missing in the void force_calc_iccp3m(). Could this be part of the problem?
Regards,
Konrad