Optimizing Equalizations of FFE, CTLE, and DFE Jointly Through a Single Pulse Response

Abstract

With the increasing data rate of high-speed digital systems, equalization techniques have been widely applied to counteract intersymbol interference and maximize eye opening in today's high-speed serial links. Channel simulations with optimal equalization settings allow engineers to explore design tradeoffs, reduce the need for costly prototypes, and ensure robust performance before manufacturing. Existing equalization optimization methods require either iterative interaction with commercial channel simulators during the optimization process, developing models with extensive training data before optimization, or separately optimizing linear equalizers in the frequency domain. In this article, the method to jointly optimize feedforward equalization, continuous-time linear equalization, and decision feedback equalization at the transmitter, receiver, or both is proposed. The optimization is achieved through the analysis of a nonequalized pulse response, eliminating the computational cost of channel simulations during optimization. Practical examples using non-return-to-zero (NRZ) and pulse amplitude modulation four-level (PAM4) signaling schemes demonstrate the effectiveness of the proposed method. A comparison with the Bayesian optimization approach, a widely discussed method for equalization optimization, shows that while both methods achieve nearly identical optimal eye openings, the proposed method offers significantly higher optimization efficiency.