Measuring jitter using an eye pattern/diagram
时间：04-05
整理：3721RD
点击：
For high speed signals an eye pattern can provide a lot of information. The data eye diagram is constructed from a digital waveform by folding the parts of the waveform corresponding to each individual bit into a single graph with signal amplitude on the vertical axis and time on horizontal axis.
One of the properties measured using eye pattern is the signal jitter. Many communication protocols will have a tolerance for how much jitter is acceptable.
The scope used for this purpose uses its internal clock for sampling the required signal. I am not sure if it is possible to use an external clock for this purpose. Let me know.
My question is that since the scope uses its own internal clock for sampling, I am sure that this clock will have its own jitter value also. Now, if we find that the eye pattern is showing jitter, how do we know how much of it is coming from the actual signal and how much from the sampling clock?
Seeing this is the case, how do we even measure jitter in the first place? Every scope we use to measure jitter will have jitter in its own internal sampling clock!
One of the properties measured using eye pattern is the signal jitter. Many communication protocols will have a tolerance for how much jitter is acceptable.
The scope used for this purpose uses its internal clock for sampling the required signal. I am not sure if it is possible to use an external clock for this purpose. Let me know.
My question is that since the scope uses its own internal clock for sampling, I am sure that this clock will have its own jitter value also. Now, if we find that the eye pattern is showing jitter, how do we know how much of it is coming from the actual signal and how much from the sampling clock?
Seeing this is the case, how do we even measure jitter in the first place? Every scope we use to measure jitter will have jitter in its own internal sampling clock!
How much precision do you need?
I am asking theoretical question only. Is my theory correct?
I believe that the eye pattern specified for some electrical standard, implicitly includes the portion of variance due to measuring itself. In other words, would be somewhat tricky to separate the effect of each one above a certain frequency range, but this should occur much above the range specified by the oscilloscope manufacturer.