Position error signal estimation at high sampling rates using data and servo sector measurements

Abstract

Track density in hard disk drives (HDD) is related to the performance characteristics of the HDD servo control system. Higher track density can be achieved by improving the performance characteristics of the servo controller that keeps the head close to the center of the track. The performance of the control system however is often limited by its bandwidth. High bandwidth implies high sampling rate. Since the sampling rate in most HDD is limited by the number of servo burst sectors that have prewritten position error signal (PES) data and by the spinning speed of the HDD, the controller bandwidth cannot be increased beyond a certain range of values. In addition to bandwidth considerations, a higher sampling rate implies more accurate control action that in turn improves performance. We present an algorithm that can generate estimates of the PES during read at high sampling rates by processing the dedicated servo burst measurements and the measurements of a nonlinear function of PES obtained from the data sectors. The nonlinear function of PES is generated from the mean square error (MSE) path metric data used to choose the correct bit sequence using the Viterbi algorithm in the partial response maximum likelihood (PRML) decoding process of the read signal. The nonlinear function is a parabola-like curve (bathtub) that relates the accumulated MSE data with the absolute value of the location of the written data. The algorithm is a nonlinear one and is based on several logic statements together with the use of two state observers. The estimation algorithm, referred to as the ACORN estimator is tested on an actual disk drive and demonstrated to provide values of the PES at a frequency four times higher than that generated by the servo burst measurements. The ACORN estimator is designed and tested for the read phase of the HDD.