Dear Professor Hess,
thanks to all of you, instead. I am sharing another bug about the free energy code, regarding the decoupling from the environment of a charged solute (i.e., switching off its charges and vdWs): Free energy, annihilation: charged ligand - #4 by Jacopo
In that very case, the problem concerns the fact that no warning about net charge of the box in any intermediate (and the ending) state is ever raised. I shared examples in there and to be crystal clear and not waste your time.
Hoping this helps, I am looking forward to any news, wish you a good day.
Sincerely,
Jacopo