A semiempirical quantum chemical method combined with the Cyclic Cluster Model provides the opportunity to optimise hydrogen positions in organic crystals. Quantification of their interactions is enabled by the correct geometry within the framework of the same method. Test calculations on systems containing only weak C–H⋯O contacts describe their relative stability reasonably well. Selectivity in a diastereomeric salt formation, where both stronger and weaker X–H⋯Y interactions contribute to crystal formation, is studied and rationalised on an energetical basis.