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

perseus GPS synch

Expand Messages
  • Christoph Mayer
    Hi all, did anyone connect a GPS-locked 80MHz signal source to the perseus? What signal source has been used? I guess that one has to match the reference
    Message 1 of 9 , Feb 4, 2014
    • 0 Attachment
      Hi all,

      did anyone connect a GPS-locked 80MHz signal source to the perseus?
      What signal source has been used?
      I guess that one has to match the reference signal to the impedance of
      the "input" in the perseus? What is the impedance? Is this an
      important point?

      Has anyone used the MKU XO 1 PLL from Kühne electronics for this purpose?

      I have found an old message on this group
      http://groups.yahoo.com/neo/groups/perseus_SDR/conversations/topics/6996
      discussing this topic. What is the latest state of these ideas? Have
      they been implemented?

      Thanks and best regards
      Christoph
    • Werner Karn
      Hi Christoph, I m using one of my Perseus with a GPS locked LO. Source is a Trimble 10 Mhz GPS controlled reference, multiplied by 8 through a PLL ICS501
      Message 2 of 9 , Feb 4, 2014
      • 0 Attachment
        Hi Christoph,

        I'm using one of my Perseus with a GPS locked LO.
        Source is a Trimble 10 Mhz GPS controlled reference, multiplied by 8 through a PLL ICS501 circuit. 
        I have disconnected the internal oszillator and am supplying the ICS501 80 MHz output instead.
        I did not experience a situation where increased phase noise would have been a problem.
        Frequency measurements in the sub milli-Hertz range are possible.
        Limiting factor seems to be the temperatur drift of the sound card oszillator.

        best regards

        Werner


        2014-02-04 Christoph Mayer <hcab14@...>:
         

        Hi all,

        did anyone connect a GPS-locked 80MHz signal source to the perseus?
        What signal source has been used?
        I guess that one has to match the reference signal to the impedance of
        the "input" in the perseus? What is the impedance? Is this an
        important point?

        Has anyone used the MKU XO 1 PLL from Kühne electronics for this purpose?

        I have found an old message on this group
        http://groups.yahoo.com/neo/groups/perseus_SDR/conversations/topics/6996
        discussing this topic. What is the latest state of these ideas? Have
        they been implemented?

        Thanks and best regards
        Christoph


      • Marco IK1ODO -2
        Christoph, I did thermally stabilize the original 80 MHz oscillator, using a Kuhne QH 40 A. I preferred to retain the very high spectral purity of the XO, and
        Message 3 of 9 , Feb 4, 2014
        • 0 Attachment
          Christoph,

          I did thermally stabilize the original 80 MHz oscillator, using a
          Kuhne QH 40 A.
          I preferred to retain the very high spectral purity of the XO, and
          accept a compromise on frequency accuracy.
          The QH 40 A works at 9V, I added a 7805 on the Perseus box and
          powered everything at 9V.
          The stability is in the order of 0.1 ppm, acceptable for my purposes.

          73 - Marco IK1ODO
        • Christoph Mayer
          Hi Marco, thanks for your reply. It is a nice idea to add thermal stabilization to the quartz crystal. However for my purposes I need
          Message 4 of 9 , Feb 4, 2014
          • 0 Attachment
            Hi Marco,

            thanks for your reply. It is a nice idea to add thermal stabilization to the quartz crystal.
            However for my purposes I need <10 ppb accuracy.

            Best regards
            Christoph


            2014-02-04 Marco IK1ODO -2 <ik1odo@...>:
             

            Christoph,

            I did thermally stabilize the original 80 MHz oscillator, using a
            Kuhne QH 40 A.
            I preferred to retain the very high spectral purity of the XO, and
            accept a compromise on frequency accuracy.
            The QH 40 A works at 9V, I added a 7805 on the Perseus box and
            powered everything at 9V.
            The stability is in the order of 0.1 ppm, acceptable for my purposes.

            73 - Marco IK1ODO


          • Marco IK1ODO -2
            ... You need another order of magnitude. For a while I considered doubling a 10 MHz standard (I have a Z3801A and several Rb standards) by three times, then
            Message 5 of 9 , Feb 4, 2014
            • 0 Attachment
              At 15:12 04-02-14, you wrote:
              >
              >
              >Hi Marco,
              >
              >thanks for your reply. It is a nice idea to add thermal
              >stabilization to the quartz crystal.
              >However for my purposes I need <10 ppb accuracy.
              >
              >Best regards
              >Christoph

              You need another order of magnitude. For a while I considered
              doubling a 10 MHz standard (I have a Z3801A and several Rb standards)
              by three times, then decided that all the HW needed to do a clean
              work (three amplifiers, doublers, filters) was too much.
              Another possibility would be to build a good 80 MHz XO and
              injection-lock it to a 10 MHz source. That may even be easier. I
              would avoid digital PLLs in the clock chain.

              Marco IK1ODO
            • Werner Karn
              Hi Christoph, please see attached picture. 10 Mhz in, 80 Mhz out; Metal box with BNC connectors; 3,3 V supply; 4,7 nF AC coupling capacitor. Regards Werner
              Message 6 of 9 , Feb 4, 2014
              • 1 Attachment
              • 1.7 MB
              Hi Christoph,

              please see attached picture. 10 Mhz in, 80 Mhz out; Metal box with BNC connectors; 3,3 V supply; 4,7 nF AC coupling capacitor.  

              Regards

              Werner


              2014-02-04 Marco IK1ODO -2 <ik1odo@...>:
               

              At 15:12 04-02-14, you wrote:
              >
              >
              >Hi Marco,
              >
              >thanks for your reply. It is a nice idea to add thermal
              >stabilization to the quartz crystal.
              >However for my purposes I need <10 ppb accuracy.
              >
              >Best regards
              >Christoph

              You need another order of magnitude. For a while I considered
              doubling a 10 MHz standard (I have a Z3801A and several Rb standards)
              by three times, then decided that all the HW needed to do a clean
              work (three amplifiers, doublers, filters) was too much.
              Another possibility would be to build a good 80 MHz XO and
              injection-lock it to a 10 MHz source. That may even be easier. I
              would avoid digital PLLs in the clock chain.

              Marco IK1ODO


            • Christoph Mayer
              Hi Werner, thanks a lot for the photo! I guess you connected the output signal to pin3 of the quartz in the perseus. Did you remove the quartz in the perseus
              Message 7 of 9 , Feb 8, 2014
              • 0 Attachment
                Hi Werner,

                thanks a lot for the photo!

                I guess you connected the output signal to pin3 of the quartz in the
                perseus. Did you remove the quartz in the perseus for that? It would
                be nice to keep it and use injection locking but I don't know if it
                works that easily.

                Also I think I will use the 3.3V from the perseus to power the IS501.

                Have you considered to use jitter cleaning, e.g., Si5317? It should
                reduce the clock jitter to <1ps.

                The corresponding evaluation board comes with a 6-layer PCB. I am
                wondering if one can use the Si5317 in a more simple way. What do you
                think about this?

                What about having a filter after the frequency multiplication, e.g.,
                made of 80MHz crystals?

                Best regards
                Christoph


                2014-02-04 Werner Karn <werner.karn@...>:
                >
                >
                > [Attachment(s) from Werner Karn included below]
                >
                > Hi Christoph,
                >
                > please see attached picture. 10 Mhz in, 80 Mhz out; Metal box with BNC connectors; 3,3 V supply; 4,7 nF AC coupling capacitor.
                >
                > Regards
                >
                > Werner
                >
                >
                > 2014-02-04 Marco IK1ODO -2 <ik1odo@...>:
                >>
                >>
                >>
                >> At 15:12 04-02-14, you wrote:
                >> >
                >> >
                >> >Hi Marco,
                >> >
                >> >thanks for your reply. It is a nice idea to add thermal
                >> >stabilization to the quartz crystal.
                >> >However for my purposes I need <10 ppb accuracy.
                >> >
                >> >Best regards
                >> >Christoph
                >>
                >> You need another order of magnitude. For a while I considered
                >> doubling a 10 MHz standard (I have a Z3801A and several Rb standards)
                >> by three times, then decided that all the HW needed to do a clean
                >> work (three amplifiers, doublers, filters) was too much.
                >> Another possibility would be to build a good 80 MHz XO and
                >> injection-lock it to a 10 MHz source. That may even be easier. I
                >> would avoid digital PLLs in the clock chain.
                >>
                >> Marco IK1ODO
                >>
                >
                >
              • Werner Karn
                Hi Christoph, I d cut the PCB trace from the internal clock with a sharp knife and connected the external signal to the SMD capacitor for the clock input to
                Message 8 of 9 , Feb 9, 2014
                • 0 Attachment
                  Hi Christoph,
                   
                  I'd  cut the PCB trace from the internal clock with a sharp knife and connected the external signal to the SMD capacitor for the clock input to the ADC. I have also considered using an anlog switch IC  to being able to switching from internal to external clock but didn't do so finally as I have other Perseii with internal clock for comparison etc. and using this one for exact frequency measurements without caring to much about higher phase noise levels.
                  I'm not using a 80 MHz filter, so I get signals at 10Mhz and harmonics at a max. level of -90dBm.
                  Thanks for the info on the Si5317. I had a quick look to the datasheet and was not impressed by the values of the specified typ. and max. phase noise levels. 
                  I'd be interested in hearing about your test results, though.
                   
                  Best Regards
                   
                  Werner 

                  2014-02-08 20:29 GMT+01:00 Christoph Mayer <hcab14@...>:
                   

                  Hi Werner,

                  thanks a lot for the photo!

                  I guess you connected the output signal to pin3 of the quartz in the
                  perseus. Did you remove the quartz in the perseus for that? It would
                  be nice to keep it and use injection locking but I don't know if it
                  works that easily.

                  Also I think I will use the 3.3V from the perseus to power the IS501.

                  Have you considered to use jitter cleaning, e.g., Si5317? It should
                  reduce the clock jitter to <1ps.

                  The corresponding evaluation board comes with a 6-layer PCB. I am
                  wondering if one can use the Si5317 in a more simple way. What do you
                  think about this?

                  What about having a filter after the frequency multiplication, e.g.,
                  made of 80MHz crystals?



                  Best regards
                  Christoph

                  2014-02-04 Werner Karn <werner.karn@...>:
                  >
                  >
                  > [Attachment(s) from Werner Karn included below]
                  >
                  > Hi Christoph,
                  >
                  > please see attached picture. 10 Mhz in, 80 Mhz out; Metal box with BNC connectors; 3,3 V supply; 4,7 nF AC coupling capacitor.
                  >
                  > Regards
                  >
                  > Werner
                  >
                  >
                  > 2014-02-04 Marco IK1ODO -2 <ik1odo@...>:
                  >>
                  >>
                  >>
                  >> At 15:12 04-02-14, you wrote:
                  >> >
                  >> >
                  >> >Hi Marco,
                  >> >
                  >> >thanks for your reply. It is a nice idea to add thermal
                  >> >stabilization to the quartz crystal.
                  >> >However for my purposes I need <10 ppb accuracy.
                  >> >
                  >> >Best regards
                  >> >Christoph
                  >>
                  >> You need another order of magnitude. For a while I considered
                  >> doubling a 10 MHz standard (I have a Z3801A and several Rb standards)
                  >> by three times, then decided that all the HW needed to do a clean
                  >> work (three amplifiers, doublers, filters) was too much.
                  >> Another possibility would be to build a good 80 MHz XO and
                  >> injection-lock it to a 10 MHz source. That may even be easier. I
                  >> would avoid digital PLLs in the clock chain.
                  >>
                  >> Marco IK1ODO
                  >>
                  >
                  >


                • df9ic
                  Hi Christoph and Werner, 1 ps jitter may sound low but is high compared to the quality of an ordinary crystal oscillator signal. You can only make the PERSEUS
                  Message 9 of 9 , Feb 10, 2014
                  • 0 Attachment

                    Hi Christoph and Werner,


                    1 ps jitter may sound low but is high compared to the quality of an ordinary crystal oscillator signal. You can only make the PERSEUS worse with this kind of stuff.


                    You would need a 80 MHz VCXO and a PLL to 10 MHz.


                    Regards


                    Henning

                  Your message has been successfully submitted and would be delivered to recipients shortly.