Signal demodulation is a fundamental procedure in many situations during a spectral analysis. Through an envelope analysis, it becomes possible to identify fault frequencies that are embedded in a modulated signal and that are not clearly visible only by directly applying some signal processing techniques such as the Fourier transform and filtering. In this paper, a very simple and empirical technique for demodulation is proposed. It is based only on the analysis of local extremes from a modulated time sequence to find a new time sequence that carries the wanted relevant fault data. The method of analysis is a good alternative tool for electrical fault detection in induction motors. The numerical and experimental results demonstrate the effectiveness of the proposed technique.65 559569