Equalization and detection in storage channels

Abstract

In this work, a model for digital magnetic recording on thin-film media is developed, and the performances of several detector/run-length-limited code combinations are compared under different densities, media noise, offtrack interference, and channel mismatch conditions. The study is based on computer simulations using an experimentally extracted transition response of a magnetoresistive read head. The detection algorithms considered include the partial response maximum likelihood detectors, decision feedback equalizer (DFE), and fixed-delay tree search with decision feedback (FDTS/DF) detector. It was found that the DFE with a (0,k) run-length-limited (RLL) code and the FDTS/DF with a (1,7) RLL code show the least bit error rate degradation, as the user density increases beyond 2.5 bits/PW50. The PRML and DFE suffer less from offtrack interference. The (0,k) coded DFE and the (1,7) coded EPRML and FDTS/DF channels are relatively robust under PW50 mismatch. At a user density of 2.5, FDTS/DF (1,7) and DFE (0,k) are found to yield the best bit error performance under varying intensities of white Gaussian noise and media noise.