For digital oscilloscopes, there are 2 main typical jitter test methods:
1) Adopt the equivalent sampling mode of digital storage oscilloscope or directly use the sampling oscilloscope to measure the timing jitter through histogram statistics. The disadvantage of equivalent sampling is that it cannot eliminate the effect of the oscilloscope's own trigger jitter on the test results, and because it adopts the operating mode of multiple triggers, multiple acquisition, and cumulative display, it is subject to more limitations for circuit design and debugging, and it is impossible to carry out in-depth jitter analysis.
(2) The more popular method is to use the real-time capture mode of digital storage oscilloscope, single trigger, continuous acquisition of a large amount of data, with the corresponding jitter test software for jitter testing. Compared with the equivalent sampling method, it eliminates the impact of the oscilloscope's own triggering jitter on the test results, and is able to carry out complex jitter analysis and jitter decomposition to obtain each jitter component, helping designers and testers to analyse the causes of jitter, and even estimate the system's BER through jitter decomposition. For example, in the U.S. National Committee for Information Standards (INCITS) under the T11.2 organisation in the jitter and signal integrity methodology (MJSQ), recommended Tektronix real-time oscilloscopes with TDSJIT3 jitter analysis software for jitter testing and analysis. Figure 1 shows the TDSJIT3 real-time jitter test results.
Jitter Test
Jitter can be described as a timing change in the period or phase of adjacent pulse edges, or even non-adjacent pulse edges. These metrics are suitable for checking long-term and short-term clock and data stability. By analysing jitter metrics in more depth, jitter test results are used to predict the data transfer performance of complex systems.
Cycle jitter is used to measure the edge-to-edge timing of clock or data cycle sample points. For example, by measuring the time between the rising edges of 1,000 clock cycles, you can sample a statistical period and the statistics will tell you the quality of the signal. The standard deviation becomes the RMS cycle jitter, and the maximum cycle is subtracted from the minimum cycle to get the peak-to-peak cycle jitter. The accuracy of each different cycle measurement determines the accuracy of the jitter measurement.
