Advanced detection for high density perpendicular magnetic recording channels

Perpendicular recording is believed capable of deferring the superparamagnetic limit to a substantially higher areal density as compared to longitudinal recording, therefore it is becoming the new standard for next generation's commercial hard disks. However, its novelty also brings up new issues, one of which is the new noise profile that may degrade the performance of current read channels. Both simulation and experimental results reveal that transition jitter, which deviates magnetization transition centers from their nominal positions, constitutes the dominant source of noise for high density perpendicular recording systems. Due to the signal dependent, non-additive and non-Gaussian natures of this noise, development of optimal channel detector becomes a challenging problem.; This dissertation presents the development of new detection schemes for transition noise dominant perpendicular recording channels. The basic idea is that transition noise, which is signal dependent and thus contains useful information about the data pattern, should be treated differently from electronic noise in the channel detector. Based on this intuition, several jitter sensitive detection schemes have been proposed, where the model of transition jitter noise is utilized for the joint estimation of discretized jitters and recorded data using modified Viterbi algorithms. This idea is also extended to develop a post processing scheme that locates the residual error events using the jitter sensitive soft output Viterbi algorithm, as well as a joint timing error and transition jitter detector that effectively tracks correlated timing errors in the jitter sensitive sequence detection. Finally the idea is used to derive a simulation based algorithm to compute the achievable information rate for future high density perpendicular recording channels with non-Gaussian transition noise.