Timing Jitter Analysis Based On Emd Decomposition Method

In recent years, many theories of jitter analysis and decomposition have been developed based on the past study contents of jitter. The deep researches on jitter analysis and decomposition can make people to understand, measure and eliminate jitter more reasonablely and completely. And, in the areas of both theory and practice, the deep researches further have great meaningful to evaluating and designing communication systems. On other hand, Empirical Mode Decomposition (EMD) method is one of many new technologies which have appeared in signal processing field in recent years. But EMD is a new technology, and further researches still has to be done for testifying its validity in more applications. Nowadays EMD is also a study focus in signal processing field.The two issues above are connected together for study in this paper originally and scientifically. It not only provides a new method for jitter analysis and decomposition but also explores a new application of EMD algorithm. The main contents are illustrated as follows:(1)The basic concepts of jitter are formulated in this paper firstly. It includes three different jitter definitions and their relations, noise models of jitter, definition of jitter function,jitter categories and subcomponents, jitter statistics and models, jitter quantification, the relationship between jitter and BER, sources of jitter, and etc.(2)Many techniques for jitter measurement and analysis are discussed. There are new contents of jitter measurement proposed, many collection techniques for jitter function, many ways to jitter analysis and some equipment based techniques. Besides, a simple ADC jitter collection method is simulated in this paper. Then jitter decomposition is investigated, in which some conventional solutions are formulated and testified their validities by simulation results. The limits of these methods are also illustrated. Furthermore, in terms of TLC method's limitations,such as instability, inaccuracy, time consumption, etc., an improved TLC algorithm based on FFT and filtering processing is proposed in this paper. The simulation results show that the proposed method improves accuracy and efficiency of TLC method.(3)Based on EMD, a new method for jitter analysis and decomposition is proposed in this paper. It can not only analyze and measure jitter frequencies with adaptive and high resolution in the time-frequency plain but also decompose and estimate jitter components of short sample length accurately in the time domain. Furthermore, by the comparison experiments the proposed method is testified more accurate than other techniques in the case of separation between RJ and low frequency PJ.