Analysis of three methods for sensorless speed detection in DC motors

Abstract

There are several techniques to measure the speed of DC motors without sensors, some techniques are based on the ripple component of the current. These techniques estimate the frequency of that component, which is related with the speed of the motor. This document describe three methods for speed measurement based on the previous techniques: spectral analysis through FFT, adaptive filtering with double derivative block and commutation detection with comparator. The accuracy of each method is also evaluated, and a comparative analysis between them is done as well. The results clearly show that the method with better performance is the spectral analysis through FFT, while the two other methods have different limitations for the speed detection.