Using Measure-Based Modeling to Analyze Backplane Deterministic Jitter
Alfred P. Neves, TDA Systems -- Test & Measurement World, 6/1/2004
|
Download a PDF of the full paper, " Using Measure-Based Modeling to Analyze Backplane Deterministic Jitter." |
Recently, much has been written regarding jitter analysis with the goal of improving BER. Decomposition of the random and deterministic jitter components for the Active Interconnect system, which includes the PHY, is an excellent method for understanding jitter sources, but falls short when only dealing with the Passive Interconnect System. Separating the total jitter into random and deterministic components will not provide insight into the signal integrity specifics for the physical layer itself. An example of this would be how much will a single via, separate from all other physical structures, contribute to NRZ random data eye closure, timing jitter, or amplitude noise? This paper specifically deals with using measure-based modeling of deterministic jitter contributors in a concise and verifiable methodology.
Introduction
Low Bit Error Rate, low cost, and high scalable speed designs are the typical objectives of backplane design. Jitter is used as a useful methodology tool to improve BER characteristics. Engineering BER by itself is intractable, making it impossible to practically point to specific problems. Total jitter, the peak-peak jitter measured at a specified BER, is composed of random (RJ) and deterministic (DJ) components. Much has been developed in the signal integrity community regarding decomposing jitter into RJ and DJ and is very useful for evaluating system designs and PHY technology, or the Active Interconnect System. It lacks as a method, however, when dealing with the Passive Interconnect Platform. A method of relating specific physical structures to eye degradation, amplitude noise and peak-peak DJ is the method illustrated. These simple structures can be integrated into a complete model for a backplane where each jitter and AM noise contributor can be identified topologically. The driver-receiver PHY is precluded as part of this method, where the focus is only on the passive platform for this technique.























