Channel equalization methods for magnetic storage

Abstract

The authors introduce design criteria and algorithms for adaptive equalization in systems using peak detection and run-length-limited (RLL) codes. They compare these methods with systems using adaptive equalizers with sampling detection methods. It is found that, at moderate densities, the adaptively equalized RLL coded systems achieve good performance at the lowest cost. However, as densities are increased, adaptive decision feedback equalization (DFE) provides significantly better performance at a slightly higher cost of implementation. Partial response methods are also investigated, but they are not favored for implementation since they require a Viterbi detector to achieve better performance than the DFE at the higher densities and their performance is very sensitive to the exact choice of the partial response polynomial.