Behavioral Modeling of Pre-emphasis Drivers Including Power Supply Noise Using Neural Networks

Abstract

This paper addresses the nonlinear behavioral modeling of pre-emphasis drivers including power supply noise. The proposed multiple-port model relies on the use of power-aware weighting functions that control the driver’s output stage to model the pre-emphasis behavior with non-ideal power supply accurately. The weighting functions are implemented using feed-forward neural networks (FFNNs), and the dynamic memory characteristics of driver’s ports are captured using recurrent neural networks (RNNs). Practical industrial driver example demonstrates that the proposed modeling method offers good accuracy, flexibility and significant simulation speed-up to facilitate signal integrity and power integrity analysis without compromising intellectual property (IP).