Hard Disk Drive Servo Systems, 2nd Edition

by B.M. CHEN, T.H. LEE, K. PENG, and V. VENKATARAMANAN Iwas first introduced to disk drive servo systems as a graduate student under William Messner at Carnegie Mellon University. While my graduate work focused on control for distributed manipulation systems, most of my colleagues worked on projects for the Data Storage Systems Center. Our group meetings invariably focused on disk-drive-relevant topics, such as dual-stage control, windage disturbances, and servo patterns and demodulation. With this background, I was able to obtain a summer internship at Seagate Research working on spinstand microactuator integration, and I continued to work at Seagate part time while I finished my Ph.D. What drew me to disk drive control as a career was the wide range of control problems inherent in the industry. Robust control must be used to overcome manufacturing variability, but adaptation is also needed to maintain performance. Iterative learning control is needed to correct for repetitive disturbances due to eccentric mounting of the disk to the spindle. Nonlinearities arise due to actuator saturation during seeks, bearing hysteresis in the voicecoil motor (VCM), and position-error signal (PES) demodulation. Dual-stage actuators move the control problem from single-input, single-output (SISO) to double-input-single-output (DISO), or double-input, double-output (DIDO), and bring additional saturation nonlinearities. All of these problems must be solved while pushing performance at approximately a 40% cumulative annual growth rate. The scale of the problem, with track density quickly approaching 200,000 tracks per inch (200 kTPI) and track-following errors constrained to be within 15% of track pitch 3 σ , leads to fundamental challenges in position detection, while pushing bandwidths ever higher as higher frequency disturbances become relevant.