Analysis of Self Noise in Symbol Timing Recovery Schemes with the Gardner Detector

Abstract

In this paper, we propose a new method to analyze pattern-dependent jitter or self noise in symbol timing recovery schemes with the Gardner detector. We introduce a Markov chain model of data patterns and thereby evaluate the statistical properties of the timing jitter caused by the random nature of the data patterns. Numerical results of analysis in various scenarios are presented along with simulation results in order to show the validity and accuracy of the proposed method. The proposed method reveals more accurately the statistical properties of the timing jitter compared to the conventional method with a linear time-invariant model. It provides the average time the timing jitter remains in specific regions and mean timing jitter as well as mean-square timing jitter.