A comment on the discrete describing function method

This correspondence points out the inadequacy of the discrete describing function method in predicting certain modes of oscillations in a class of relay sampled-data systems. These modes, however, do not exist in this class of systems. This must be ascertained before the previously mentioned method could be applied to determine all the possible modes of a system. Furthermore, the necessary change in critical regions is suggested for this class of systems to predict the correct mode of oscillation.