N8823B Datasheet, PDF (11/21 Pages) Keysight Technologies – Vertical Noise Analysis Software

11 | Keysight | EZJIT Complete Jitter and Vertical Noise Analysis Software for Infiniium Oscilloscopes - Data Sheet
Accurate Random Jitter Separation using Spectral and Tail Fit Algorithms
The problem
Common spectral algorithm for RJ
extraction over-reports RJ.
Why?
Any Aperiodic Bounded Uncorrelated Jitter
(ABUJ) on the signal is factored into RJ.
A common source of ABUJ is crosstalk
from adjacent signals.
How does this affect me?
First, the RJ measurement can be
inaccurate. Second, when measuring Total
Jitter (TJ) vs. BER, TJ can be dramatically
overestimated due to the RJ multiplier. For
example, TJ = DJ + 14*RJ_RMS for a BER of
10^^-12. This means the scope will provide
an unfairly negative report of the overall
jitter performance of your serial data link.
Figure 14. Jitter separation with artificially extreme crosstalk using spectral algorithm for RJ
extraction. Notice the discontinuity in the BER bathtub.
The solution
EZJIT Complete provides two distinct
algorithms for RJ extraction to ensure you
achieve accurate measurements under all
signal conditions.
The Tail Fit algorithm accurately extracts
RJ, even in extreme ABUJ/crosstalk
environments, by fitting directly to the
tail of the RJ,PJ histogram.
Figures 14 and 15 contrast the
two algorithms in a contrived jitter
measurement with extreme crosstalk.
You can see the tail fit algorithm in Figure
5 provides a much better fit to the BER
bathtub curve in the upper left and reports
significantly lower RJ and TJ.
Figure 15. Jitter separation with artificially extreme crosstalk using tail fit algorithm for RJ
extraction. Notice the good fit to the BER bathtub and dramatically lower RJ and TJ results.