Performance robustness and stiffness analysis on a machine tool servo design

Abstract

The machine tool servo design is very different from the traditional high performance servo systems. Traditional servo design depends heavily on a precise system model so that frequency domain or time domain compensation techniques can be applied, and many synthesis tools such as LQG, H ∞, or LMI, … are available based on this concept. While these synthesis tools are becoming very useful in many situations, they are not used in the machine tool servo design. Most high performance machine tool systems still rely on the more primitive PID or PDF controllers in conjunction with complicated friction and temperature compensation algorithms. This paper investigates the different approaches based on a house-designed servo control board. With the ability to change the servo control algorithms on a test machine tool, the performance robustness and servo stiffness were tested both numerically and experimentally. The numerical tests conducted with PID, PDF, and LQG controllers showed that the PID controller is the easiest to achieve good performance, but the PDF controller contains the best performance robust margin. Experimental results also indicated that the PDF controller exhibits far superior robustness properties over the other two controllers.