Jittery readings
Martin,
The problem is covered in a little detail in the Audio Precision App. Note no.5, "Measurement Techniques for Digital Audio" by Julian Dunn (actually a book, rather than a short app note!).
The problem arises due to the interval between the 'scope trigger and the transition being examined. It's also affected by the spectral content of the jitter.
What this really displays is interval variations and whilst there is a relationship between these and jitter, it's not straightforward and gives rise to the problems I mentioned earlier.
This approach is especially insensitive to low frequency jitter.
A PLL (or multiple PLL's) can be used to generate an ideal clock (or in fact a clock with a low-pass jitter characteristic), or you can do comparisons with an ideal clock.
There's some good info here (www.wenzel.com) on jitter / phase noise measurement, and a good ref clock isn't so hard to build.
Once it's working you can iteratively improve the reference and the item under test, quite readily.
The 'scope measurement isn't useless, you just need to be aware of limitations.
Andy.
Martin,
The problem is covered in a little detail in the Audio Precision App. Note no.5, "Measurement Techniques for Digital Audio" by Julian Dunn (actually a book, rather than a short app note!).
The problem arises due to the interval between the 'scope trigger and the transition being examined. It's also affected by the spectral content of the jitter.
What this really displays is interval variations and whilst there is a relationship between these and jitter, it's not straightforward and gives rise to the problems I mentioned earlier.
This approach is especially insensitive to low frequency jitter.
A PLL (or multiple PLL's) can be used to generate an ideal clock (or in fact a clock with a low-pass jitter characteristic), or you can do comparisons with an ideal clock.
There's some good info here (www.wenzel.com) on jitter / phase noise measurement, and a good ref clock isn't so hard to build.
Once it's working you can iteratively improve the reference and the item under test, quite readily.
The 'scope measurement isn't useless, you just need to be aware of limitations.
Andy.
