Reduced-state nonlinear equalization for magnetic storage

We present a nonlinear depth-limited tree search for magnetic storage. Using a discrete Volterra series description, we outline a minimum-mean-square error solution for a finite length feedforward filter and nonlinear branch metrics. Feedback is used to subtract contributions from the current decision as well as previous ones. We evaluate the probability of error by averaging over many possible error events considering the effects of a mismatched channel. The performance using a minimum phase criterion and assuming no “d-constraint” is compared with a linear and nonlinear decision feedback equalizer (DFE)