Loading ...
Sorry, an error occurred while loading the content.
 

Using AFEDRI SDR-net x2 for Frequency-Standard Phase-Detector

Expand Messages
  • larryweberplasma
    It recently occurred to me that my AFEDRI SDR-net x2 might make a very good phase detector to measure the Allan Deviation of GPS disciplined crystal
    Message 1 of 5 , 19 Mar


      It recently occurred to me that my AFEDRI SDR-net x2 might make a very good phase detector to measure the Allan Deviation of GPS disciplined crystal oscillators or rubidium frequency standards in order to test their frequency stability.  These high accuracy standards are essential for generating frequency stable signals in the GHz range.  This measurement needs to precisely measure the phase difference between two 7 dBm 10 MHz signals and record this at typically one second intervals for many hours up to a few days.  The output would be a text file with the once a second phase delay floating point numbers usually expressed as very small fractions of a second (nanoseconds of phase difference over the one second interval).  This text file could then be sent to free software such as TimeLab http://www.ke5fx.com/timelab/readme.htm to do the statistical analysis such as the Allan Deviation plot.

      If such a measurement would work well on the AFEDRI SDR-net x2 then it would be really valuable with the AFE822x4.   This would allow the simultaneous phase comparison of up to 4 frequency standards.  Comparison of 3 or more standards is quite useful since a single misbehaving standard can easily be identified when its errant phase is compared to the normal phase of the other two.  This cannot be done when measuring only 2 standards since it is not clear which one is misbehaving.


      My question to the group: what is the best software approach to accomplish the generation of such a phase difference text file?  I assume I will need to do some software work but I would like to use as many of the existing tools as possible.  I am not familiar with Linrad but is this a direction I should explore?  Or perhaps someone has already done this and can tell me how they did it.


      I would be very happy to hear any ideas or suggestions.


      Larry

      K9ZBU


    • Leif Asbrink
      Hello Larry, You could use Linrad. To get the text file you would have to add a file users_extra.c (linux) or wusers_extra.c (Windows) and compile from
      Message 2 of 5 , 20 Mar
        Hello Larry,

        You could use Linrad. To get the text file you would
        have to add a file "users_extra.c" (linux) or wusers_extra.c
        (Windows) and compile from source which is trivial under
        Linux as well as Windows.

        There is a file extra.c that can give you a starting point.

        You should then run Linrad with adaptive polarization
        and print pg_pol_angle once a second. The header screendef.h
        has to be included. You might print the time and the phase
        angle between the two signals because the time would not become
        exactly 1 second apart. For that there is the function current_time()

        Set users_extra_update_interval=1.0;

        73

        Leif





        >
        >
        > It recently occurred to me that my AFEDRI SDR-net x2 might make a very good phase detector to measure the Allan Deviation of GPS disciplined crystal oscillators or rubidium frequency standards in order to test their frequency stability. These high accuracy standards are essential for generating frequency stable signals in the GHz range. This measurement needs to precisely measure the phase difference between two 7 dBm 10 MHz signals and record this at typically one second intervals for many hours up to a few days. The output would be a text file with the once a second phase delay floating point numbers usually expressed as very small fractions of a second (nanoseconds of phase difference over the one second interval). This text file could then be sent to free software such as TimeLab http://www.ke5fx.com/timelab/readme.htm http://www.ke5fx.com/timelab/readme.htm to do the statistical analysis such as the Allan Deviation plot.
        >
        > If such a measurement would work well on the AFEDRI SDR-net x2 then it would be really valuable with the AFE822x4. This would allow the simultaneous phase comparison of up to 4 frequency standards. Comparison of 3 or more standards is quite useful since a single misbehaving standard can easily be identified when its errant phase is compared to the normal phase of the other two. This cannot be done when measuring only 2 standards since it is not clear which one is misbehaving.
        >
        >
        > My question to the group: what is the best software approach to accomplish the generation of such a phase difference text file? I assume I will need to do some software work but I would like to use as many of the existing tools as possible. I am not familiar with Linrad but is this a direction I should explore? Or perhaps someone has already done this and can tell me how they did it.
        >
        >
        > I would be very happy to hear any ideas or suggestions.
        >
        >
        > Larry
        > K9ZBU
        >
        >
      • larryweberplasma
        Hi Leif, Thanks for the good suggestion. Does Linrad support 4 channels or is it limited to 2? My initial tests will be with 2 channels but if that works well
        Message 3 of 5 , 21 Mar
          Hi Leif,

          Thanks for the good suggestion. 

          Does Linrad support 4 channels or is it limited to 2?  My initial tests will be with 2 channels but if that works well I will soon want to use 3 or 4.

          Larry
          K9ZBU
        • Leif Asbrink
          Hi Larry, Linrad is limited to two channels. It would be non-trivial to change thar. Sorry:-( I think you could run two instances of linrad, one on channels 1
          Message 4 of 5 , 21 Mar
            Hi Larry,

            Linrad is limited to two channels. It would be non-trivial
            to change thar. Sorry:-(

            I think you could run two instances of linrad, one on channels
            1 and 2 and the other on channels 3 and 4. You could feed one
            signal to 1 and 3, a second signal to 2 and a third signal to
            channel 4.

            Comparing three will be enough for you to know which one is not so good.

            As an alternative, you could feed 4 different signals into the channels,
            run the Afedri to produce 4 single channel receivers. Set 4 instances
            of Linrad to 10.000000 MHz and print the phase of the carrier in the
            baseband often enough to avoid any 360 degree ambiguity.

            You could use Airspy or SDR-IP or some other radio that can
            be locked to a 10 MHz reference and run any number of receivers
            all with its own instance of Linrad to compare any number of
            oscillators.

            For the purpose of generating good microwave signals I guess you
            are looking for very low sideband noise very close to the carrier
            (sideband noise and Allen variation is the same thing.) Maybe thiese
            pages are of interest to you:
            https://www.youtube.com/embed/ERrql2wOvjA

            https://www.youtube.com/embed/kQmOCl482zs

            A low noise amplifier with a crystal filter can be used as a
            Q multiplier to provide a very sharp filter that might be a good
            thing to clean up the 10 MHz signal from a GPS clock.



            73

            Leif






            > Hi Leif,
            >
            > Thanks for the good suggestion.
            >
            > Does Linrad support 4 channels or is it limited to 2? My initial tests will be with 2 channels but if that works well I will soon want to use 3 or 4.
            >
            > Larry
            > K9ZBU
          • larryweberplasma
            Hi Leif, You have given me a lot of food for thought . I had not thought of all of the different ways to hook things up with multiple instances. Looks like
            Message 5 of 5 , 23 Mar
              Hi Leif,

              You have given me a lot of "food for thought".   I had not thought of all of the different ways to hook things up with multiple instances.  Looks like there are a lot of good possibilities.  I am leaving today for over two weeks of vacation with my family and so I will think about these in the spare moments.

              Yes low sideband noise in the frequency standard is very important for microwaves.  In the jargon of the frequency or time standard world ( See http://leapsecond.com/time-nuts.htm ) the term Phase Noise is used to describe the sideband noise in the frequency domain from perhaps 1 Hz away from the carrier out to perhaps megahertz.   The Allan Deviation is a time domain metric used to describe stability in frequency or phase over longer periods typically from 1 second out to days.  You are correct that they are really both measuring the same thing.  They are simply two different ways of looking at this same thing with perhaps the only distinction being that Phase Noise is usually looking at relatively fast variations and Allan Deviation is usually looking at very slow variations.  Both are important for microwaves.

              I like the idea of using a Q multiplier to clean up the phase noise.  I remember using one of those in the early 60s on my novice receiver and it was quite effective at digging those CW signals out of the mud.

              Thanks for your help.

              Larry
              K9ZBU
            Your message has been successfully submitted and will be delivered to recipients shortly.