Adaptive Continuous-Time Equalization Followed By FDTS/DF Sequence Detection

This talk will present fhepeifiomance and low power advantages of employing an adaptive analog continuous-time equalizing filter followed by a fixed-delay tree search with decision feedback (FDTSIDF) sequence detector. Although an adaptive continuous-time filter of modest complexity cannot achieve the degree of control over the equalized pulse shape offered by a long FIR equalizer, its power consumption can he dramatically lower. By employing an FDTS/DF rlctector which can be etlsily adapted to widely varying equalized pulse re,sponses we cm maintain channel performance approaching that of an FIR equalizer followed by an F D T S D F detector. Simulation results will be presented that compare this new approach with both an FIR equalizer followed by FDTSDF detectcir and with an FIR equalizer followed by an EPR4-ML detector. We will present and compare several architectures for the adaptive continuous-time equalizer that offer improved convergence when using the lemtmean-squares (LMS) adaptation algorithm. The use of an adaptive continuous-time equalizer of the proper architecture coupled directly with an FDTSLIF sequence detector can achieve perjlormance closely approachin$ that of sequence detectors preceded by adaptive FIR equalizers, while requiring a power dissipation that is only slightly greater than that of n standard peak detection channel.