Velocities were not correctly handled, also not in fragments
First bug was loading config with velocities and saving it resulted in all but first component of velocity vecto
r being zero. Flag=2 was re-introduced in command line switching to prevent saving after fragmentation (as nothi
ng is done with the original molecule anyway). This removed the bug completely (which is thus kind of mysterious
in nature). Now, corresponding atoms in fragments have the correct velocities (also replaced hydrogens)