Design and application of non-binary low peak factor signals for system dynamic measurement

Examines the role of non-binary multifrequency signals in system identification. The test stimulus primarily considered is the prime multifrequency signal due to its superior harmonic rejection capability over other multifrequency signals. The paper explores various algorithms for generating low peak factors for such a signal and presents a set of optimum phases. It considers the implication of mechanising such a signal within an instrument and shows that the aliasing problems associated with the digital techniques used for the storage and generation of multifrequency waveforms can be overcome using a software compensation algorithm. The theoretical basis for the new algorithm is presented along with a method by which the algorithm can easily be incorporated within a multifrequency response analyser. The paper then demonstrates how the prime-sinusoid waveform can be used to provide meaningful spectral estimates of systems that are non-linear