[time-nuts] Phase measurement of my GPSDO

Fri Apr 3 16:12:56 UTC 2020

Hi

Your 5335 resolves 1 ns, that is what limits it’s performance. 
If you have a gate time of 1 second, you will get 9 digits in a 
second, regardless of frequency. That’s the advantage of a
“computing counter”.

If you had a 10 Hz signal with fast enough edges, you could read
it out to 9 digits. Simply put, the ADEV you plotted would be
*identical* at 10 Hz. You would get 7x10^-10 at 1 second off of
the 10 Hz signal. 

Next you get the x 1,000,000 because you did subtraction to
get to the 10 Hz. That is totally independent of anything else
going on. It’s like putting an amplifier in front of your system.

Take that million and put it on top of the 7x10^-10 and you are
at 7x10^-16. The only limits are the standard you compare to 
and how “quiet” you can get the edges.

Bob

> On Apr 3, 2020, at 11:59 AM, Tobias Pluess <tpluess at ieee.org> wrote:
> 
> Hi John
> 
> Yes, I totally agree with you and I also understand the difference.
> But what I still don't understand is the following:
> Obviously, my 5335A is not accurate/precise enough to measure below 1e-9
> for short tau. Currently I am comparing the 1PPS signals, but when I change
> that and use the DMTD method, I will still compare some 1Hz signals, and
> the counter is still not able to resolve stuff that is lower than 1e-9. So
> why would the DMTD work better?
> I totally see that the error is somehow multiplied, but if my GPSDO is good
> (which I hope it is :-)) the error will still be very small - perhaps in
> the 1e-9 or 1e-10 region, so too low for my 5335A. Not?
> 
> 
> Tobias
> 
> On Fri, Apr 3, 2020 at 5:34 PM John Ackermann N8UR <jra at febo.com> wrote:
> 
>> I think the difference is between *mixing* or *dividing* down to a low
>> frequency.
>> 
>> When you divide, you divide the noise along with the carrier frequency.
>> 
>> When you mix, you "translate" the noise.  If the signal bounces around
>> 0.1 Hz at 10 MHz (awful, I know), when you divide to 1 PPS the noise is
>> also divided by 1e7 so the ratio remains the same.
>> 
>> But if you mix via a 9.999 999 MHz local oscillator, now your output at
>> 1 Hz still has 0.1 Hz of noise on it.  i.e., it's the same absolute
>> value of noise as you started with.  So you measure that absolute value
>> but don't compare it to the mixed down 1 Hz frequency, compare it to the
>> original 10 MHz frequency.  It's basically an error multiplier.
>> 
>> John
>> ----
>> 
>> On 4/3/20 11:25 AM, Tobias Pluess wrote:
>>> Hi again Bob,
>>> 
>>> yes you describe a simple DMTD measurement. But could you tell me what
>> the
>>> difference is between that and comparing the 1PPS pulses?
>>> I mean, I could set the 10811 high in frequency by just 1Hz, and then it
>>> would result in two 1Hz signals which are then compared.
>>> Which is essentially the same as comparing two 1PPS signals, isn't it?
>>> Ok there is a minor difference: since the 1PPS signals are divided down
>>> from 10MHz, their noise is also divided down, which is not the case for
>> the
>>> DMTD.
>>> However, in the end I am comparing signals in the 1Hz to 5Hz or 10Hz
>>> region, and apparently, the 5335A is not suitable for those, at least not
>>> with the desired stability, is it?
>>> 
>>> 
>>> Tobias
>>> 
>>> On Fri, Apr 3, 2020 at 1:45 PM Bob kb8tq <kb8tq at n1k.org> wrote:
>>> 
>>>> Hi
>>>> 
>>>> The quick way to do this is with a single mixer. Take something like an
>> old
>>>> 10811 and use the coarse tune to set it high in frequency by 5 to 10 Hz.
>>>> 
>>>> Then feed it into an RPD-1 mixer and pull out the 5 to 10 Hz audio tone.
>>>> That tone is the *difference* between the 10811 and your device under
>>>> test.
>>>> If the DUT moves 1 Hz, the audio tone changes by 1 Hz.
>>>> 
>>>> If you measured the 10 MHz on the DUT, that 1 Hz would be a very small
>>>> shift
>>>> ( 0.1 ppm ). At 10 Hz it’s a 10% change. You have “amplified” the change
>>>> in frequency by the ratio of 10 MHz to 10 Hz ( so a million X increase
>> ).
>>>> 
>>>> *IF* you could tack that on to the ADEV plot of your 5335 ( no, it’s not
>>>> that
>>>> simple) your 7x10^-10 at 1 second would become more 7x10^-16 at 1
>>>> second.
>>>> 
>>>> The reason its not quite that simple is that the input circuit on the
>>>> counter
>>>> really does not handle a 10 Hz audio tone as well as it handles a 10 MHz
>>>> RF signal. Instead of getting 9 digits a second, you probably will get
>>>> three
>>>> *good* digits a second and another 6 digits of noise.
>>>> 
>>>> The good news is that an op amp used as a preamp ( to get you up to
>> maybe
>>>> 32 V p-p rather than a volt or so) and another op amp or three as
>> limiters
>>>> will
>>>> get you up around 6 or 7 good digits. Toss in a cap or two as a high
>> pass
>>>> and low pass filter ( DC offsets can be a problem ….) and you have a
>>>> working
>>>> device that gets into the parts in 10^-13 with your 5335.
>>>> 
>>>> It all can be done with point to point wiring. No need for a PCB layout.
>>>> Be
>>>> careful that the +/- 18V supplies to the op amp *both* go on and off at
>>>> the
>>>> same time ….
>>>> 
>>>> Bob
>>>> 
>>>>> On Apr 3, 2020, at 5:13 AM, Tobias Pluess <tpluess at ieee.org> wrote:
>>>>> 
>>>>> hi John
>>>>> 
>>>>> yes I know the DMTD method, and indeed I am planing to build my own
>> DMTD
>>>>> system, something similar to the "Small DMTD system" published by
>> Riley (
>>>>> https://www.wriley.com/A Small DMTD System.pdf).
>>>>> However I am unsure whether that will help much in this case, because
>> all
>>>>> what the DMTD does is to mix the 10MHz signals down to some 1Hz Signal
>> or
>>>>> so which can be measured more easily, and I already have 1Hz signals
>> (the
>>>>> 1PPS) which I am comparing.
>>>>> Or do you suggest to use the DMTD and use a higher frequency at its
>>>>> outputs, say 10Hz or so, and then average for 10 samples  to increase
>> the
>>>>> resolution?
>>>>> 
>>>>> Thanks
>>>>> Tobias
>>>>> HB9FSX
>>>>> 
>>>>> 
>>>>> On Fri, Apr 3, 2020 at 12:53 AM John Miles <john at miles.io> wrote:
>>>>> 
>>>>>>> b) if I want to measure 1e-11 or even 1e-12 at 1sec - what resolution
>>>>>> does
>>>>>>> my counter need? If the above was true, I would expect that a 1ps
>>>>>>> resolution (and an even better stability!) was required to measure
>> ADEV
>>>>>> of
>>>>>>> 1e-12, The fact that the (as far as I know) world's most recent,
>>>>>>> rocket-science grade counter (some Keysight stuff) has "only" 20ps of
>>>>>>> resolution, but people are still able to measure even 1e-14 shows
>> that
>>>> my
>>>>>>> assumption is wrong. So how are the measurement resolution and the
>> ADEV
>>>>>>> related to each other? I plan to build my own TIC based on a TDC7200,
>>>>>> which
>>>>>>> would offer some 55ps of resolution, but how low could I go with
>> that?
>>>>>> 
>>>>>> That sounds like a simple question but it's not.  There are a few
>>>>>> different approaches to look into:
>>>>>> 
>>>>>> 1) Use averaging with your existing counter.  Some counters can yield
>>>>>> readings in the 1E-12 region at t=1s even though their single-shot
>>>> jitter
>>>>>> is much worse than that.  They do this by averaging  hundreds or
>>>> thousands
>>>>>> of samples for each reading they report.  Whether (and when) this is
>>>>>> acceptable is a complex topic in itself, too much so to explain
>> quickly.
>>>>>> Search for information on the effects of averaging and dead time on
>>>> Allan
>>>>>> deviation to find the entrance to this fork of the rabbit hole.
>>>>>> 
>>>>>> 2) Search for the term 'DMTD' and read about that.
>>>>>> 
>>>>>> 3) Search for 'direct digital phase measurement' and read about that.
>>>>>> 
>>>>>> 4) Search for 'tight PLL' and read about that.
>>>>>> 
>>>>>> Basically, while some counters can perform averaging on a
>> post-detection
>>>>>> basis, that's like using the tone control on a radio to reduce static
>>>> and
>>>>>> QRM.  It works, sort of, but it's too late in the signal chain at that
>>>>>> point to do the job right.  You really want to limit the bandwidth
>>>> before
>>>>>> the signal is captured, but since that's almost never practical at RF,
>>>> the
>>>>>> next best thing to do is limit the bandwidth before the signal is
>>>>>> "demodulated" (i.e., counted.)
>>>>>> 
>>>>>> Hence items 2, 3, and 4 above.  They either limit the measurement
>>>>>> bandwidth prior to detection, lower the frequency itself to keep the
>>>>>> counter's inherent jitter from dominating the measurement, or both.
>>>> You'll
>>>>>> have to use one of these methods, or another technique along the same
>>>>>> lines, if you want to measure the short-term stability of a good
>>>> oscillator
>>>>>> or GPSDO.
>>>>>> 
>>>>>> -- john, KE5FX
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> _______________________________________________
>>>>>> time-nuts mailing list -- time-nuts at lists.febo.com
>>>>>> To unsubscribe, go to
>>>>>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>>>>> and follow the instructions there.
>>>>>> 
>>>>> _______________________________________________
>>>>> time-nuts mailing list -- time-nuts at lists.febo.com
>>>>> To unsubscribe, go to
>>>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>>>> and follow the instructions there.
>>>> 
>>>> 
>>>> _______________________________________________
>>>> time-nuts mailing list -- time-nuts at lists.febo.com
>>>> To unsubscribe, go to
>>>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>>> and follow the instructions there.
>>>> 
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts at lists.febo.com
>>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>> and follow the instructions there.
>>> 
>> 
>> 
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
>> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.