Does it make sense to multiply 10MHz reference?

  • Hello,

    I'm currently building a GPSDO with a 10MHz output and have a Rev. D ADALM-Pluto.


    My question is if there is any advantage in multiplying the 10MHz reference to 40MHz (either through PLL or harmonic multiplier) vs feeding the 10MHz directly and adjusting the frequency in software?


    I ask this as there is some info about phase noise performance being better when feeding with 40MHz reference, however I'm not sure if the extra phase noise caused by multiplication wouldn't be greater than Pluto's added noise.


    73,

    PY3PK

  • Hello Pedro,


    the Pluto will most likely not work properly with a 10 MHz reference. A friend tested 4 or 5 Plutos and only one of them worked properly with reference frequencies of 20 Mhz and below. Therefore it is advised to use reference frequencies higher than 20 MHz, 40 MHz or even 50 MHz is strongly recommended.


    If you generate your reference by using frequency doublers e.g. 40 MHz out of 10 MHz you will loose the same amount of phase noise performance as running the PLL with a lower reference frequency. Of course this is only true if the doubler is ideal and also the PLL.


    In a latest design Kurt DJ0ABR has developed a simple unit using a 100MHz OCXO and 2 dividers :2. With this we generate 50 MHz for the Pluto and 25 MHz for the reference frequency of the LNB. Simple and very good. The circuit of Kurt allows also to lock the OCXO additionally to a GPSDO. I have buil this up and it works very fine. The GPSDO is not really needed as the OCXO is very stable.


    Kind regards


    Matthias


    http://www.dd1us.de

  • Hi Pedro.


    Great project !


    Link:

    https://projects.dj0abr.de/doku.php?id=en:sat:plutolnb


    73!

    PY1SAN

  • Hi,

    Lets suppose that we have two OCXO one in 10 MHz and one in 100 MHz with the same phase noise at the same offset.The phase noise is increased by 20log(N) dB) where N is multiply factor. In our case N=10, so we have 20 times dB larger noise. It meens in simple words that if we create 10 MHz from 100 MHz it will have 20 time dB less phase noise or if we create 100 MHz from 10 MHz it will have 20 dB times larger phase noise. It is obvious that when we produce 25 MHz from 100 MHz the phase noise will be less than produced from 10 MHz. One calculation that I have done is that phase noise produced by multiplying by 390 (9750/25) is about 50 dB times bigger!!!

    The ECOC-2522-100.000 existis in various versions with similar prices and different ppB, so a small ppB is prefered as has better stability... PA3FYM AOCJY1-100.000MHZ seems a good choise for GPSDO available from Mouser at half the price.

  • Hi,

    the AD document UG-570 says, that the scaled frequency should be as close to 80 MHz as possible, i.e. 40 MHz x 2. But you can't take 50 MHz x 1.6!

    Detlev


    REFERENCE CLOCK SETUP AND OPERATION

    If the DCXO is not used, an external reference clock needs to be ac-coupled to XTALN (Pin M12). XTALP (Pin M11) is not connected (leave floating). The clock frequency must be between 5 MHz and 320 MHz, and can be scaled by 1×, ½×, ¼× and 2× using BBPLL, Rx and Tx reference dividers. The valid frequency range for the RFPLL phase detectors is 10 MHz to 80 MHz, and the scaled frequency of the reference clock must be within this range. For optimum phase noise it is recommended to operate the scaled clock as close to 80 MHz as possible. The selection between DCXO and external reference clock is made in the ad9361_init function.

  • Hi,

    Lets suppose that we have two OCXO one in 10 MHz and one in 100 MHz with the same phase noise at the same offset.The phase noise is increased by 20log(N) dB) where N is multiply factor. In our case N=10, so we have 20 times larger noise. It meens in simple words that if we create 10 MHz from 100 MHz it will have 20 time less phase noise or if we create 100 MHz from 10 MHz it will have 20 times larger phase noise. It is obvious that when we produce 25 MHz from 100 MHz the phase noise will be less than produced from 10 MHz. One calculation that I have done is that phase noise produced by multiplying by 390 (9750/25) is about 50 times bigger!!!

    The ECOC-2522-100.000 existis in various versions with similar prices and different ppB, so a small ppB is prefered as has better stability... PA3FYM AOCJY1-100.000MHZ seems a good choise for GPSDO available from Mouser at half the price.

    George, it's 20 dB more / less phase noise. Anyway .. that's exactly why I -from the beginning (end 2018)- beated on the drum concerning PN of the LNB reference and use(d) crystal filters for purification.

  • Hi Matthias,


    it is not the difference between 40 and 50 MHz, it is the difference between 80 (40 x 2) and 50 Mhz.


    Because I have no 50 MHz clock I can't measure it, but I believe AD knows the facts.


    73 Detlev

  • As an update, I was able to feed directly 10MHz into my Rev. C/D Pluto. I used a 0.9Vpp square signal from a ENE3311B. Had to run, as per https://wiki.analog.com/university/tools/pluto/devs/booting:


    Code
    fw_setenv refclk_source 'external'
    fw_setenv ad936x_ext_refclk_override '<10000000>'
    pluto_reboot reset

    This way, 10MHz was enabled from the external input.

  • As an update, I was able to feed directly 10MHz into my Rev. C/D Pluto. I used a 0.9Vpp square signal from a ENE3311B. Had to run, as per https://wiki.analog.com/university/tools/pluto/devs/booting:


    Code
    fw_setenv refclk_source 'external'
    fw_setenv ad936x_ext_refclk_override '<10000000>'
    pluto_reboot reset

    This way, 10MHz was enabled from the external input.

    Congratulations PY3PK .

    Does it work fine with 10 MHz?


    PY1SAN

    Resende - Rio de Janeiro.

    GG77SN