Dear Bruce and Josh,

Many thanks for your help! I have followed Josh's suggestion and increased the limits of the nx and npx parameters to very large values (nx=10^6, npx = 800.000 in path_tot.f90, npx = 100.000 in ff2x_tot.f90), and recompiled.

The npatx parameter was set to 8 already, so I didn't change it.

In addition, in order to get enough paths included in the fine structure, I had to increase the np1x parameter from 60.000 to 200.000. I am not quite sure what kind of parameter this is, but by checking the path listing in the xmu.dat file, I find that all important paths up to 8 A are included now. The fine structure looks much more like the experiment now – yay! Unfortunately there are so many paths, that a real EXAFS fitting analysis is impossible. 

Thanks again & best wishes,

Renske