Phase Difference Measurement by means of Digital Sampling

[johnsold]

At 100 kHz one cycle takes 10 microseconds.  Divide that by 360*1000 to get millidegree resolution = 27.8 picoseconds.  This requires a 36 GHz clock.

Lynn

[Ransu]

I'm curious. Why do you need to do this?

We are an accredited standards laboratory and our main calibration source for phase difference is a 5500-2. And it is not cheap.

If you are trying to build something that matches or is better in accuracy then that - you will have difficulties. Have you considered lowering your goal a little?

[Ransu]

Just out of fun I plugged in our 1GHz 8bit digizer to the 5500-2 and recorded a full cycle of 1kHz test signals with 500MSps.

There were about 800 samples at zero crossing - with fuzzy edges.

Guess what your problem really is

[johnsold]

Ransu,

Thanks for the information.  Although he has not admitted it, I think the original poster had a school assignment.  I suspect that someone took a quick look at the instrument specs and said "Measure what it puts out."

It could still be a good learning tool, but some realism should be applied.

Lynn