Using FeynArts with FeynCalc gives incorrect results for 3- and 4-point loop integrals when the scalar products are predefined. The problem goes right back to version 5.1 (2006). A simple "fix" is to disable the automatic loading of FeynArts in the FCConfig.m file.
===========
ScalarProduct[p1, p1] = m^2;
ScalarProduct[p1, q] = ScalarProduct[q,q]/2;
SetOptions[OneLoop,Prefactor->-I/Pi^2];
db=FeynAmpDenominator[PropagatorDenominator[k,lam1]]*
FeynAmpDenominator[PropagatorDenominator[k-q,lam2]]*
FeynAmpDenominator[PropagatorDenominator[k-p1,m]];
OneLoop[k,db]
===========
gives C0[m^2,m^2,q^2,lam1^2,lam2^2,m^2], which is the wrong answer.
If loading of FeynArts is disabled, then we get the correct answer,
C0[m^2,m^2,q^2,lam1^2,m^2,lam2^2].
This archive was generated by hypermail 2b29 : 09/04/20-12:55:05 AM Z CEST