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