Convolutional coding for parallel decision-feedback equalization

A decision-feedback equalizer (DFE) is one of the simplest detectors for use in a magnetic recording channel. An extension of DFE, termed parallel DFE (PDFE), is shown to offer significant performance improvements with a relatively minor complexity increase. To offer still greater performance gains, a convolutional code is added to the system. Although PDFE suffers less error propagation than conventional DFE, the error bursts remain problematic when codes are part of the overall system. Therefore, a convolutional interleaver is added to randomize the error bursts. Simulation results for a Lorentzian channel are presented for various PDFE structures, with and without convolutional codes, with various levels of interleaving, for both hard and soft sequence detection.