[phenixbb] refinement of peptide links in mmcif coordinates (multi character chain identifiers)

Dear PHENIX users, for a structure with approx. 30 instances of an N-terminal AYA (N-acetylalanyl) residue, I have noticed that PHENIX did not restrain planarity of the peptide bond with the next residue. What is the recommended way of restraining those peptide bonds' dihedral angles and bond distances? Thank you, Wolfram Tempel