Advanced channel signal processing for multi-track or multi-wordline data storage systems

Abstract

In this paper, advanced channel signal processing schemes are proposed for multi-track or multi-wordline read architecture in the future hard disk drive and NAND flash memory systems. As the capacity increases with aggressive scaling of the data bit or memory cell, the interference or coupling between neighboring information causes severe inter-track (ITI) or inter-cell interference (ICI), and advanced signal processing accounting the neighboring tracks or wordlines is needed to alleviate the effect of these interference. In magnetic recording (MR), array-reader-based MR (ARMR) technology has shown a potential gain in the areal density capability (ADC) and the data transfer rate by employing multi-track channel signal processing, however skew effect causes complications in reliable data retrieval by widely varying the reader to reader distances. In thins study, a novel multi-track read channel signal processing approach that differentiates linear densities of jointly processing tracks is suggested and numerically evaluated. For an ARMR with the 3 reader 2 track setup shows around 4% of potential ADC gain over the constant linear density configuration under aggressive ITI condition, providing the doubled read throughput. Likewise, the multi-wordline signal processing schemes are investigated for NAND flash memory channel under severe ICI. Joint equalization can be modified for processing two or more wordlines' soft information together, while it requires additional read operations with a channel customized set of slicing levels. Evaluations with numerical flash channel model show that the joint processing of two wordlines provides bit error rate (BER) performance gain around 0.1 order with 6 additional read operations per wordline.