How to measure time variation of 2 signals using an oscilloscope?

[Nate Duong]

I have 2 rubidium clocks which both are generating 10MHz sine wave. Some reason my system is giving bad data. So, I am deciding to inspect every single part of the system.

I am trying to check the rubidium clock and make sure it have to be in nanosecond time variation, because: t=1/f = 1/10e6 = 100 (ns).

I am using this oscilloscope: https://www.atecorp.com/products/tektronix/tds3034

and this rubidium clock: http://www.thinksrs.com/products/PRS10.htm

Is the time variant between of 2 sine wave also known as the delay on the scope? if I say right, the time variant is display on the screen is about 38.74ns. If I say wrong, how to calculate or how to have time variation from the scope?

I am not experience in oscilloscope and hope anyone can help.

Thank you.

 

[analogdesign]

The two clocks are out of phase which makes sense unless you have some system that is phase locking them.

You will need to trigger on one of the clocks and verify that the phase shift is constant to see if the clocks are working as advertised.

 

[Nate Duong]

The two clocks are out of phase which makes sense unless you have some system that is phase locking them.

You will need to trigger on one of the clocks and verify that the phase shift is constant to see if the clocks are working as advertised.

@analogdesign : so, If I want to trigger 1 of those signal, I just need to choose "CH1" (yellow button) and press "B TRIG" button?

 

[analogdesign]

You can try that... every scope is different but they are similar. Try the different trigger modes until you have the output you want.

 

[Nate Duong]

You can try that... every scope is different but they are similar. Try the different trigger modes until you have the output you want.

I just want to confirm again that the time variation measurement is the similar of "delay" function on the scope, right?

 

[f95toli]

It is not at all clear to me what you are trying to do.
Assuming your PRS10 are working reasonably well there is way you would be able to see anything with the oscilloscope directly; the internal clock of the oscilloscope will be several orders of magnitude worse than that of the PRS10. Hence, checking them by connecting them to the two inputs is not going to tell you anything (unless they are REALLY not working).

The best way to do a measurement like this would be to beat the two clocks. Get a cheap mixer from e.g. Minicircuits that works at 10 MHz and connect one PRS10 to the RF and one to the LO input. You can then look at what comes out of the IF port with your oscilloscope. There are many, many more advanced ways of doing this but it would be a good starting point (note that this is essentially how new types of atomic clocks are tested)
There should be plenty of material about this on the web. A good starting point would be some of the free articles on Enrico Rubiola's website (he is well-known guru when it comes to measuring phase and amplitude noise of frequency references)

 

[jim hardy]

In the days of analog scopes
we'd select the trigger to trigger on only one channel, not composite

your scope''s trigger section offers a menu
meaning you're at the mercy of what made sense to some programmer

let's just say you select trigger from channel A only at zero volts positive slope
what will to happen is the dot to start across the screen drawing a wave every time channel A crosses zero going up.
So channel A should appear stationary on the screen, always starting at zero volts on the way up
If channel B remains stationary on the screen they are running at exact same frequency
and difference in times of zero crossings on the two traces you see on the scope is their phase difference.

Probably channel B will drift very slowly, count the number of seconds it takes to drift by one cycle .

Dont know quite what else to tell you. Most of becoming skillful with the oscilloscope lies in judicious application of its trigger settings. Learn your 'scope's capabilities.

 

[f95toli]

If channel B remains stationary on the screen they are running at exact same frequency

I don't this would work here (unless the PRS10are really not working), The problem is that the OP is trying to measure the stability of a fairly decent frequency reference; the AVAR is something like 1 part in 10^11 which is very likely to be orders of magnitude better than the stability of any of the internal digital circuity of the oscilloscope. The only way you would be able to measure the stability directly was if you had access to instruments which were even more stable; which in reality would mean either an even better rubidium reference or a hydrogen maser. However, by downmixing (beating the two references) you dramatically reduce the requirements on the measurement instruments to the point where even a normal oscilloscope should be able to tell you if there is a problem.

 

[jim hardy]

I don't this would work here (unless the PRS10are really not working), The problem is that the OP is trying to measure the stability of a fairly decent frequency reference; the AVAR is something like 1 part in 10^11 which is very likely to be orders of magnitude better than the stability of any of the internal digital circuity of the oscilloscope.

Well, that's why you trigger ONLY on one of the sine waves not both.
Shouldn't nominal 10⁷ hz sinewaves that differ by one part in 10¹¹ slip phase by one whole cycle every 10⁴ seconds, about three hours ? That'd be readily observable on an analog 'scope by looking at it just a few times every hour.
Were it lower frequency we'd use chop mode sweep not alternate.

Or is my thinking muddled ? (again..)

I have 2 rubidium clocks which both are generating 10MHz sine wave. Some reason my system is giving bad data. So, I am deciding to inspect every single part of the system.

 

[rbelli1]

Or is my thinking muddled ? (again..)

I think you are spot on.

That would tell if one of them is a bit off but not which one.

Some sort of correlation with a GPS receiver might help tell which is the problem. That would need a fairly wide counter and some careful programming.

BoB

 

[Baluncore]

I have 2 rubidium clocks which both are generating 10MHz sine wave. Some reason my system is giving bad data. So, I am deciding to inspect every single part of the system.

What is this "bad data" you have? We do not know the experimental constraints.
What are the clocks used for, why do you need two separate clocks ?
What is your “system” and how accurately do the frequency reference clocks need to track ?

The digital TDS 3034 oscilloscope should be able to measure the difference time due to phase shift sufficiently accurately to monitor the drift of the two frequency references.