Comparison of Jitter Analysis Techniques
白書
DesignCon - February 2, 2005
Ransom Stephens, Brian Fetz, Steve Draving, Joe Evangelista, Michael Fleischer-Reumann, Greg LeCheminant, Jim Stimple
Different jitter analysis techniques yield results that can vary by hundreds of percent.
At Keysight Technologies, Inc. we were aware of the discrepancies and invested into the research necessary to understand the situation before introducing a solution. We built a data transmitter with a complete set of applied jitter levels precisely calibrated, in most cases, to traceable standards1. We assembled jitter analysis equipment, three 6 GHz bandwidth real-time oscilloscopes and a 4 GHz bandwidth time interval analyzer, from the major vendors as well as a Keysight bit error ratio tester (N4901B SerialBERT) and a Digital Communication Analyzer (86100C DCA-J) equipped with a 20 GHz electrical receiver and then applied a wide variety of signals with known TJ(10–12) and known levels of different types of jitter to determined which analyzers are accurate and why
This paper assumes an understanding of total jitter defined at a bit error ratio, TJ(BER) and how it can be decomposed into random and deterministic sub-components as indicated by Figure 1.