Synchronization between multiple AD9371 boards

Hi,

I'm trying to synchronize between two AD9371 boards. I have gone through all the posts to related to it in the discussion forum and user guide (UG-992 page 53 for details on Multichip Synchronization) and the wiki page (https://wiki.analog.com/resources/eval/user-guides/ad-fmcomms5-ebz/multi-chip-sync).

1) First I have done the  base band synchronization using MCS (with the clock of 30.72MHz)

2)Then I have done the RF phase synchronization using external LO (twice of RF frequency).

After completing above steps then I have measured phase difference between Tx1 of one ad9371 board and Tx1 of other AD9371 board then their phase difference is almost constant in multiple time measurement during one power cycle.(let say phase difference between two boards of tx1 (theta_d) is around 3 degrees).

When I was sending out of phase signals(after compensating phase offset) through combiner then the signals are not cancelling. Please explain what was the reason for not cancelling?  and what could be the solution?

kindly help me.

  • 0
    •  Analog Employees 
    on Nov 2, 2020 8:18 AM 2 months ago

    Are you using EVAL board or custom board for this testing?

    When I was sending out of phase signals(after compensating phase offset) through combiner then the signals are not cancelling. Please explain what was the reason for not cancelling?  and what could be the solution?

    Are you checking the signal at RX output or TX output?

    When you are checking the signal for phase cancellation, did you check in the same power cycle(n which you measured the phase offset of 3 degree)or did you check after power off-on?

  • I'm using two AD9371 EVAL boards with ZC706 FPGA and I'm using IIO scope(GUI).

    I have connected Tx1 of 1st board & Tx1 of 2nd board  to combiner inputs & output of combiner connected to Rx1 of 1st board  for checking cancellation of signals.

    During first power cycle, I have measured phase difference between two boards multiple times which is around 3deg. Then in second power cycle(power off & on) phase difference is around 176 degrees in multiple measurements. 

    I'm trying to cancel two out of phase signals from different boards during any one power cycle after compensating with corresponding phase offset.

  • 0
    •  Analog Employees 
    on Nov 3, 2020 10:24 AM 2 months ago in reply to JDusari748

    Did you compensate for the phase difference between TX and RX path? Check the output of the combiner in the spectrum analyzer. The phase of the RX path will get added to the combined TX1 and TX2 signal.

    Also share the plots.

  • Did you compensate for the phase difference between TX and RX path?

    Yes. I have used the common receiver for both Tx1 & Tx2 while calculating their individual phases through combiner  & corresponding phase compensation has done.

    Check the output of the combiner in the spectrum analyzer.

    I have checked the output in the spectrum analyzer also but cancellation not occurring. Even if keeping one signal phase constant in the first board and changing phase from 0 to 360 of the same signal & sending through second board then also signals are not canceling at any one of the phases.

    Fig. combiner output in spectrum analyser.

    Note: I have calculated the phase difference between Tx1 & Tx2 of the same board & after compensating phase offset & transmitting out of phase signals then phase cancellation is occurring (I have verified it)

    But If I have used different boards of transmitters then phase cancellation not occurring even though the same procedure used for calculating phase difference.

      

    Fig.1: Tx1 output

    Tx2 output

    Combiner output at Rx1 in IIO

  • 0
    •  Analog Employees 
    on Nov 5, 2020 9:12 AM 2 months ago in reply to JDusari748

    Hope you have given the same ref_clk to both the eval boards and the FPGA as well. How are you making sure that the data is transmitted to the TX of both the boards at the same time instant?

    Check the TX output in an oscilloscope and see if both of them are properly aligned. Do the testing by giving a sinusoid signal at the input so that the phase cancellation can be viewed properly.