A nonlinear model and equalizer for magnetic recording channels
Yungsoo Kim, Kiho Kim · 2002
We propose a novel nonlinear model and a nonlinear equalizer for high density magnetic recording channels. The model is based on the superposition of data transition dependent pulses. The proposed nonlinear equalizer utilizes the proposed nonlinear model. When tested with actual read back signals, the model using 2 and 32 pulses achieved gains of 4.6 dB and 29.4 dB, respectively, over the linear model. Also the proposed nonlinear equalizer attained a gain of 1.5 dB over the RAM-DFE at a bit error rate of 1E10/sup -/6 when the nonlinear signal was generated by a finite state machine model.