發展QEP-CORDIC之伺服馬達19-bit弦波解碼器及電路實現

Abstract

伺服馬達控制系統中包含伺服馬達、感測器及控制器三部份，想要完成一個高效能的伺服馬達控制控制系統三者缺一不可，本論文著重於弦波編碼器的解碼演算法發展實現，及低速控制適用的估測法及控制法。在弦波編碼器解碼部分，文中首先使用線性化法進行解碼實驗，之後再提出一個QEP-CORDIC 演算法，與一般CORDIC 演算法不同處在於QEP-CORDIC 演算法特別適用於弦波編碼器的解碼，除了有自動線上及時的A/D轉換器之校正外，另外加入了輸入向量位置的對應、以及狀態轉換的QEP 計數方式等等，並以模擬及實驗找出最合適的遞迴次數，最後完成DSP-Based 線性化法19bit 解碼及FPGA-Based QEP-CORDIC 19bit 弦波編碼器之解碼電路實現，使用17521 個邏輯閘、 150ns 解碼計算效率。 本研究進一步將所研製的19-bit 弦波編碼器應用在伺服馬達低速控制，在速度方面首先比較三種不同的估測法，(1)差分法加低通濾波器、(2)最小平方近似法、及(3)卡曼濾波器，經過模擬及實驗結果證實卡曼濾波器在低速估測上擁有最佳的效果，能有效降低回授速度變異量；在低速控制法則方面首先使用傳統PI 控制器，並且調整一組適合低速控制的參數與另一組適合中高速的參數做比較，之後再以PI 控制器加上狀態空間控制法實現，經過實驗證實傳統PI 控制器在低速控制上有最低的速度變異量，效果最佳。最後將所有適用於低速控制的PI 控制器及卡曼低速估測法成功的應用於CNC 上。

This study focuses on the development and realization of the decoder algorithm of sinewave encoders for precision motion control, especially for low-speed operations. The linearization method for sinewave decoding has been implemented on a DSP microcontroller with high-level-programming realization based on the C language. Furthermore, The proposed QEP-CORDIC algorithm is realized on the FPGA-based 19-bit decoding. The present QEP-CORIDC algorithm which employs the CORDIC algorithm with the proposed techniques as : (1) auto-calibration of ADC converter, (2) vector mapping of QEP, and (3) the state counter of the QEP. Also, the best recursive procedure is also determined by both simulation and experimental results, with 17521 logic gates and 150ns decoding time. The developed 19-bit sinewave encoder is applied to the serve motor with a low-speed control. There are three methods are compared to estimate the velocity from the encoder output: (1) the differential approach with a lowpass filter, (2) the east-squares fit, and (3) the Kalman filter. The Kalman filter has been proven to be the best performance with the minimum velocity variance. Both the PI controller and the Kalman filter has been applied to the CNC to render satisfactory control performance in extremely low-speed motion control.