The change of that line of code fixed the problem with the 63 error in output2. Unfortuantly, the divide by zero in normal.f on line 36 is still there. So qpw(ispin) must be the problem here. I'm trying to find this at the moment. Michael