When not get any better convergence in amorphous solids with ions or molecular crystals and the 'energy change' is oscillatory. I have found that solves recalculating the loads of the groups of the final structure. The recalculation of this structure again converge without oscillations in the 'energy change'.