AD9361 Power fluctuation

Hi:

We are working on AD9361 frequency hopping test. The TX1/2 output are direct connected to RX1/RX2 through attenuators. The TX LO frequency sweeps from around 400MHz up to 4GHz, RX LO sweeps together with TX LO. The baseband were modulated by DDS on TX. On the RX side, we expect to see constant demodulated waveforms according to the frequency sweep. On each frequency, we expect to get the same waveform amplitude between sweeps. However, we find that the demodulated signal amplitude varies between sweeps at the some frequencies especially on those frequencies under around 1.7GHz, above 1.7GHz, the demodulated signal amplitude look consistent between sweeps. The RX gain mode is set at 0 which I think is manual mode. I appreciate if you could give us some hints to trouble shoot this issue.

thanks in advance.

Jason

  • 0
    •  Analog Employees 
    on Oct 14, 2021 3:19 PM

    Are you using custom board or eval board?

    How much fluctuation in power are you seeing at the output? If the decrease in power with frequency  sweep is high, then connect the TX output to a spectrum analyser and then check the power difference. The external circuit that is used between TX and RX can also be a contributing factor for loss of power.

  • Hi Srimoyi:

    Thanks for your feedback, we are using our custom board, but the external circuit between TX and RX are the same as FMCOMM3 board, only baluns and cables. We don't have power loss, just the amplitude fluctuation bothers me. We are using fast hopping, the demodulate signal amplitude changes between different sweep. For example, at 1GHz, the baseband signal amplitude could be between 0.6Vpp to 1Vpp, the baseband frequency is single tune at 1MHz, it doesn't vary between sweeps, pretty consistent. I wonder if there are some calibrations we missed.

    In the manual it mentioned, DC tracking or quadrature calibration have to be done if the frequency changes larger than 100MHz, is it true in the fast-hopping mode as well?

    thanks,

    Jason

  • 0
    •  Analog Employees 
    on Oct 18, 2021 11:20 AM in reply to njjz2007

    Yes in fast hoping mode also, you need to run these callibrations. Fast lock mode eliminates the need for doing  only VCO calibrations, hence eliminating only the VCO cal time. So if the frequency change is more than 100MHz, then you need to run these cal even in fast hop mode.

    So, are you seeing this variation in amplitude even when you keep the LO fixed at 1GHz?  If you change the Lo to say 2.4GHz, and stay fixed to that particular LO are you still seeing the variation in amplitude?

  • Thank you Srimoyi, our application need fast hopping over 4GHz bandwidth. We are sweeping from 100MHz up to 4GHz. It seems the power fluctuation only happened in the low frequency range like below ~1.6GHz, beyond that the power fluctuation is minimum. During the scan, we use the same gain setting, so I guess we don't need the RF DC calibration. For the TX/RX quad Cal, shall we wait the RF locked signal first before starting calibration? Do you have an estimated time of how long it takes for each TX/RX calibration? 

    Jason

  • 0
    •  Analog Employees 
    on Oct 26, 2021 10:19 AM in reply to njjz2007

    The RFDC calibration needs to be run at the start and also need to be run every time  the frequency change is more than 100MHz. So you need to run RFDC,RX quad and TX quad cal since you are hoping over a wide frequency range.

    The cal times depends on the CLKRF rate and the configuration settings. Check the UG for more details.