Adaptive nonlinear decision feedback equalization with channel estimation and timing recovery in digital magnetic recording systems

In high density digital magnetic recording systems, the nonlinear effects in the write process appear in the readback waveform as shifts in the peak positions and changes in the amplitude. The pulse shift causes a nonlinear intersymbol interference (ISI) on the readback signal. In this paper, a Volterra-DFE, in which a Volterra filter is used in the feedback section of decision feedback structure, is proposed to equalize the nonlinear ISI. By theoretical analysis, the minimum mean squared error (MSE) of the Volterra-DFE is found to be a function of sampling phase. This will affect the output signal-to-noise ratio (SNR) of the playback system. A baud-rate timing recovery system in conjunction with a channel estimator is used to operate the sampling clock on the optimum phase. By computer simulation, Volterra-DFE is found to outperform conventional DFE in both output SNR and BER in a high density magnetic recording system