Advanced Digital Controls for Disk Drives

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

0072752 Messner The rapid increase in the complexity and size of software in business, science and engineering, and the need to store increasingly large amounts of information in data bases for commercial, government, and medical applications has propelled the growing demand for data storage devices that have higher date transfers rates, have higher capacity, and are physically smaller. In future years this trend is expected to rise dramatically as the use of disk drives in portable and wireless devices becomes widespread. To respond to this demand, the current annual rate of increase in data storage density is 100% per year. The majority of this increase comes from increases in track density, because of physical limitations on magnetic recording media and read and write heads. This places disproportionate burden on the servo system to keep the head near the data track center for good read/write performance of the drive. Reduced latency times to reach data on different data tracks also requires higher performance servo systems. To improve the performance of disk drive servo systems to support the rate of increase in data storage density, new and advanced analytical techniques must be developed that specifically address the complex dynamics of the disk drive mechanics. Furthermore, there are many issues concerning the hybrid nature of the control system which must simultaneously respond to both discrete-time and continuous-time signal, since the controllers for disk drives are inherently discrete-time. The construction of such methods requires a spectrum of research effort ranging from the development of fundamental theory to the creation of domain specific computational algorithms to the implementation of these algorithms on appropriate test facilities. The broad objectives of this research proposal are (1) to develop high performance design methods for control of disk drives, which have further general use; and (2) to demonstrate and apply these methods successfully to testbed disk drive systems. Two major features of the work proposed here are that the PI's will directly address the hybrid nature of disk drive control where both continuous and discrete signals must be directly confronted and that the PI's will unify the approach to control for the two separate functions of the disk servo, namely track following and track-to-track seeking. The proposed methods and approach will assist the competitiveness of the US dominated $60 billion data storage industry by providing systematic methods for control design for current and future disk drives. Furthermore, many of the analytical tools developed will be applicable to alternative mechanical applications, and insights gained from this complete treatment and evaluation of disk drive control design will be parlayed into other related technologies.