Data-Transition Adaptive Decision Feedback Equalizer for 2/4PAM Serial Links

Conventional decision feedback equalizers (DFEs) suffer from the fundamental drawback of shrinking rather than increasing data eyes when consecutive 1s or 0s are present in data. To combat this drawback, a new data-transition adaptive DFE is proposed. The proposed DFE takes into account the dependence of post-cursors on the polarity of data and searches for optimal tap coefficients using a sign-sign least-mean-square (SS-LMS) with consideration of the state transition of data. To validate the effectiveness of the proposed algorithm, both data-state DFE and data-transition DFE are employed to equalize impaired channels with known characteristics. Simulation results demonstrate that the proposed data-transition DFE outperforms data-state DFE equalization with improved vertical eye-opening, reduced jitter, and shorten adaptation.