Estimating Angular Speed From Resolver Signals via Improved PLL With Noise Reduction

Both phase-locked loop (PLL) and angle-tracking observer (ATO) are commonly used to obtain the information of rotor position and speed from resolver signals. However, they are faced with different problems in estimating speed. PLL cannot overcome the pollution of inevitable measurement noises on the estimated speed, and the type number of ATO will be decreased by one for the estimation of speed. In this article, a noise-reduction PLL (NRPLL) is proposed to suppress measurement noises without type number degradation for software-based resolver-to-digital converters (RDCs). First, a first-order low-pass filter is introduced into a typical type-II PLL to prevent measurement noises in the phase detection error from propagating to the estimated speed directly. Second, the additional pole due to the filter is designed as an almost canceled zero-pole pair together with the zero in the closed-loop transfer function. Thus, the proposed PLL still remains type II for the estimation of speed while it can be approximated as a second-order system to tune the performance conveniently. Theoretical analysis indicates that the proposed method combines the advantages of ATO in noise reduction and PLL in steady-state accuracy for speed estimation. Simulation and experimental results demonstrate the effectiveness of the proposed method.