|Thread Links||Date Links|
|Thread Prev||Thread Next||Thread Index||Date Prev||Date Next||Date Index|
A few more thoughts on the Tx jitter generator for TP3 -
I suggested that a filtered PRBS generator may make some sense for the higher jitter frequencies. By filtered, I mean that the PRBS output would be filtered to spread what would otherwise be a bimodal probability function into something more uniform or normal. This signal could be applied to a phase modulation input, much the same as SJ is applied now.
Advantages of this approach are that it generates a wider spectrum that should be more realistic and testing would be done with a single test (no sweeping or multiple tests).
As I mentioned on the call, it could be done with something like PRBS7 running at ~1Gbd. The fundamental frequency in this case would be ~8 MHz with harmonics spaced at ~8 MHz up to the bandwidth of the filter. A figure is attached that shows this with a single-pole filter corner frequency at 60 MHz. Note the shape of the pdf and that it is clearly bounded.
On the other hand, if SJ does not cause EDC to do something that it would not do in application (which I don't think would happen), then it is probably preferred because it is familiar and has a lot of history.
Filtered PRBS7 jitter example.png