Low-Speed Sensorless Control of a Surface Mounted Permanent Magnet Motor in an e-Bike Application

This paper presents a low-speed sensorless control implementation on a surface-mounted permanent magnet (SPM) motor for an e-bike application, using a low-cost fixed-point DSP. A novel dynamic PLL gain algorithm is proposed to improve the inherent reduced performance of the sensorless control under transient response. The angular offset caused by the magnetic field distortion is verified experimentally and using Finite Element Analysis (FEA). The acoustic noise and system stability are analyzed with different amplitudes and frequencies of the high-frequency (HF) voltage signal injection, and an adaptive frequency selection algorithm is proposed to reduce the acoustic noise production during a wide load condition operation.