Differences in neutron + proton spectra between Fluka/DPMJET v Phojet 1.12

Hello,

I am generating 7 TeV pp collisions in Fluka using the SPECSOUR card and the flukadpm3 executable
(compiled using ldpmqmd), and comparing the particle spectra with events generated using standalone Phojet 1.12 (fed into Fluka
using source.f). I see differences in the normalisation for neutron and proton spectra between the two
(attached plots are for 50000 collisions each). For other particles (charged pions, photons, electrons, muons) I get reasonable
agreement.
Can anyone explain why these differences would occur? Thanks for any assistance.

Cheers,
Paul