I have seen that ignoring hydrogens leads to higher R-factors and energy terms, particularly when solvent is used during the refinement. Not sure how much this weighs against computational efficiency. But turning off charges on Asp etc has always been useful during simulated annealing. Regards Raman