You can create PSFs with CHARMM, either directly or via eg the CHARMM-GUI web interface; I would recommend starting with CHARMM since in this way you learn what is going on - there are examples in the Script Archive forum.
Yes, that's what I have decided to do after all.
I guess I have successfully created the PSF file but
I encountered something strange. Namely while the
number of bonds in CHARMM PSF and Gromacs TOP files is
the same for my protein, the number of angles and
dihedrals is by merely 1 lower in the former case.
Besides, number of CMAP and improper dihedrals is in
agreement according to both packages. I have
checked the protonation state of all residues and
terminating residues and they are exactly the same
when I use either of the packages.
Could you give me a clue on where this discrepancy may
come from, please? My PSF generation does not finish
with either warning or error.