Model-independent constraints on ΔF= 2 operators and the scale of new physics