Micromagnetics-Based Analysis of Multi-Track Detection With Simplified 2-D Write Precompensation on Shingled Magnetic Recording

Abstract

In shingled magnetic recording systems, a shingled writer is used to write narrow tracks by overlapping the previous tracks, which also leads to severe inter-track interference, jitter, and nonlinear transition shift (NLTS) in the cross-track direction. To deal with these challenges, we propose a simplified 2-D write precompensation (WPC) scheme that adjusts the write current magnitude instead of the timing, to mitigate the 2-D NLTS in the writing process. In the proposed WPC scheme, the WPC is only implemented for the few specific patterns that seem to cause large media noise mean and variance. During the training period, the optimum write field for current transition being written is obtained by subtracting demagnetizing field from the previous neighboring bits in the current and nearest previous tracks from a fixed write field, and the optimum writing current is determined by minimizing the mean-squared error between the needed field and the available fields within a writer field library. Then, the specific bits producing the dominant demagnetizing field and the obtained optimum writing current are stored in a lookup table during the training period. During the writing process, we can obtain the optimum writing current by matching the specific previous written bits against the bit patterns in the lookup table, avoiding the need for computing the demagnetizing field and optimum writing current for every written bit. Here, two simplified WPC schemes are investigated corresponding to precompensating two sets of data patterns. For readback, a joint pattern-dependent noise-predictive (PDNP) detector is investigated to further whiten the 2-D media noise. Simulations indicate that at the target bit error rate of $10^{-2}$ , the proposed simplified WPC schemes combined with the joint PDNP detector can provide the density gains of as high as 22% compared with the system without WPC and detected with a joint Bahl–Cocke–Jelinek–Raviv detector.