[time-nuts] I am looking to measure phase jitter between two synchronized DDS signals

Bob Camp kb8tq at n1k.org
Wed Nov 30 09:44:06 EST 2016


The highest sensitivity approach to jitter measurement is to look at phase noise with a phase noise test set. You 
don’t mention jitter bandwidth or what sort of cleanup you have after the DDS. That will all impact the degree to which
phase noise is a useful indicator. 

Likely the quickest way to do that with your sources would be to run a regenerative divider to get the 3 GHz back down
to 750 MHz. Then compare the two signals at 750 with a double balanced mixer. Effectively you have built a vector
voltmeter to do what you need. If you have concerns about the regenerative dividers, build two and look at them 
against each other. If you can lock the 750 and 3 GHz, you can use the same setup for phase noise and get very
good resolution (-170 dbc / Hz) without a lot of trouble. 

Yes it is a bit of construction. You have an odd requirement and off the shelf test gear isn’t targeted at odd things …


> On Nov 30, 2016, at 4:47 AM, Jim Potter <jpotter at jpaw.com> wrote:
> I just posted this on the HP Equipment forum. Someone reading it there suggested I contact timenuts because what I want to do is effectively an accurate time stability measurement.
> I am attempting to look at the phase jitter between two synchronized DDS rf sources, one at 3 Ghz and one at 750 GHz. My 10 Gs/s scope has way to much jitter to be useful. A vector voltmeter might be just the thing. I have had an HP 8508A VVM and would buy another one if it was useful, but it is rated only to 2 GHz. I also have an old PRD 2020 with the 2021 head that is in working condition. It is rated to 2.4 Ghz, but syncs on the 750 MHz signal and apparently passes the 3 GHz signal somewhat. I get a reading of about -40 dBm for +10 dBm in at 3 Ghz. However, I do get a phase measurement that varies with the phase of the 3 Ghz signal corresponding to the changes I make in the output phase. The issue is the noise. I see about +/- 1 degree jitter in the phase reading on the panel meter. I also see a drift of 2 to 3 degrees. I don't think the DDS units are the source of the drift, and I'd like to show that the phase noise is lower than what I am observing.
> I don't know of any VVM that works at 3 GHz, beyond the out of spec use of my PRD unit. Using an oscilloscope is problematic because I have measured the trigger jitter on a 100 Gs/s Tektronix unit when it was brand new. It's about 3.5 ps RMS. A 3 GHz signal has a period of 333 ps so that corresponds to 10 degrees. I see that Tek has a new scope with 70 GHz bandwidth and 200 Gs/s, but I don't think that is a big enough improvement over the 100 Gs/s unit. In any case even renting one of these is probably too expensive.
> I would appreciate any suggestions for how to accurately measure the phase jitter. I have a Tektronix RSA 306A which has been handy, but I'm uncertain about how to use it to measure phase jitter. In any case, what really matters is not the jitter of the individual signals but the jitter in the phase difference between the two signals which are all derived from the same time base.
> One obvious question is, if I can't measure it how can it matter? The answer is that each rf unit is a signal source for a section of linear accelerator. If the phase jitter is excessive, it will show up as loss of beam quality and current. It's a bit hard to set up a proton accelerator to see if the jitter is acceptable. I really need able to make a bench measurement to qualify the performance.
> Thanks in advance for any suggestions.
> Jim
> James M. Potter PhD, Pres.
> JP Accelerator Works, Inc.
> 2245 47th Street
> Los Alamos, NM 87544
> TEL:505-690-8701
> FAX: 888-301-2833
> mailto:jpotter at jpaw.com
> http://www.jpaw.com
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.

More information about the time-nuts mailing list