Created attachment 515 [details] Proton spectra in n+Bi reaction at 63 MeV (test30) If an interface to the Geant4 native pre-compound de-excitation is used by the Bertini cascade there is a problem of energy balance check. In the case of internal conversion an electron may be emitted instead of gamma. Electron mass is not taken into account in the balance computation and as a result re-sampling is forced several times. This increases CPU and damage de-excitation spectra. One of the plots demonstrating the problem is attached.
Many problems found and fixed in interface to native G4PreCompound de-excitation. 1) unit problems in going from MeV in G4Fragment to GeV in Bertini 2) electron charge and mass subtracted 3) used a copy of G4Fragment from G4PreCompound instead of const ref because precompound can change the fragment 4) for momentum only, increased allowed non-conservation from 1 keV to 10 keV
