Spectral analysis of signals is the main tool in the development of new functional methods for monitoring asynchronous electric motors. The article describes a spectrum analyzer based on an autocorrection algorithm of time recording. The algorithm allows to minimize the effect of spectral leakage on the results of spectral analysis, which significantly increases the accuracy. The analyzer is designed to accurately determine the parameters of harmonics characteristic for the signals generated by asynchronous motors. A comparison is made with the traditional methods of spectral analysis based on the fast Fourier transform and the Windowed Fourier transform. The developed spectrum analyzer provides a higher accuracy of determining the amplitudes and frequencies of harmonic components of the signal spectrum in comparison with traditional methods. Another advantage is the high stability of the results, which is retained even with the reduction of the signal recording time up to one second.