Nonlinear Error Correction for Magnetic Encoders
Hsiang-Chun Chuang, Chung‐Wei Kung, Li‐Wu Chen, Shyh-Biau Jiang
Abstract
Magnetic encoders are used as angle sensors to control motors for precise positioning in many industrial manufacturing processes. However, nonideal factors such as installation misalignment lead to angle measurement errors. This study analyzed a nonlinear model of the measurement error due to the eccentricity of a magnetic encoder and developed a method of compensating for this error. An experimental calibration system consisting of a two-phase stepping motor was used to provide a standard reference angle for calculating the measurement error of the encoder. The fast Fourier transform (FFT) algorithm was used to demonstrate that the measurement error can be modeled as a multifrequency composite cosine wave with low-frequency components. Compensation for the measurement error was performed using the calculated experimental parameters of each frequency component in the nonlinear error model. The proposed method was experimentally verified, and the accuracy of the 14-bit AS5047D magnetic encoder was improved from 8–9 to 11–12 bits.