Multichannel signal processing for multiple-head digital magnetic recording

Abstract

This paper analyzes the performance of a multiplehead, digital, magnetic recording system using multichannel signal processing theory. A multi-input, multi-output (MIMO) channel that models the multiple-head recording process provides the basis for this analysis. This model characterizes the write process, the read process and both electronics and media noise. In particular, the model accounts for side fringing between a read head and adjacent tracks; side fringing is a major cause of intertrack interference that is often neglected during signal processing analysis. Optimal decision feedback and partial response equalizers for the MLMO channel are derived using a minimum mean-squared error criterion. Numerical evaluations show that a magnetic recording system comprising a multiple-head array and equalizer combats intertrack interference much better than does the corresponding single-head system. As a result, the multiple-head system achieves a gain of up to a factor of 3.5 in radial density over the single-head system for a fixed value of maximum head misalignment.

Constructive inter-track interference (CITI) codes for perpendicular magnetic recording

“…Previous work that looked at this for longitudinal recording channels include Refs. [1][2][3][4][5][6][7][8][9][10]. Multi-track recording codes that control the type of adjacent transition can significantly affect error performance in the presence of ATI [11].…”

Low-complexity multi-head detection for multi-track partial response and two-dimensional recording channels

“…Theoretical investigations have revealed that multi-head detection for multi-track partial response (PR) recording channels is capable of increasing storage density and combating intertrack interference (ITI) in high-density systems much better than single-head detection [1][2]. Maximum-likelihood sequence detection (MLSD) is an optimal approach [3] but its computational complexity grows exponentially with both the partial response order and the total number of tracks due to the use of the multi-dimensional Viterbi algorithm.…”

Signal Processing for Present and Future Magnetic Recording Systems

“…(iii) Improvements in system performance: Servos, track following, multiple head configurations [9][10].…”