Why decompose jitter when you can get a true BER?

When high-speed serial or optical engineers have access to a BERT (bit-error-rate tester), they can effectively determine design performance by connecting their device under test to the BERT and get a true bit error ratio and a TJ (total jitter) number, or TJ@BER. These results typically are suitable for most compliance test specifications on the market today.

So why would you want to break down jitter into a series of components? Here are a couple of the main reasons:

• Extrapolation: You may want to extrapolate system performance to cases that would be difficult or time-consuming to measure directly with a BERT. In some cases, counting billions of bits can tie up lab time for hours, if not days.
• Analysis: For root cause, an understanding of the jitter model can provide insight into the precise cause or causes of excessive jitter. This insight can lead to quicker resolution of the “aggressor” in your design causing noncompliance.

A high-performance oscilloscope with the right jitter analysis toolset can recreate a model that places jitter into different components, of which the primary ones are random and deterministic. While random jitter has a Gaussian shape to its PDF distribution, deterministic jitter differs across the spectrum and can come in different forms.

An understanding of these deterministic types can lead to better analysis and root-cause determination. As you might expect, with every “model” there are assumptions made and results will not be as repeatable as a true BER count. But the upside is two-fold: You achieve faster results (extrapolation) and deeper insight (analysis). Moreover, the models are continuing to evolve and the latest algorithms can even identify the presence of BUJ (bounded uncorrelated jitter) resulting from crosstalk.

What tools are you using to troubleshoot timing errors on your digital communication designs?