ADRV9002 coherence between the transmit and receive channel

I have ADRV9002 + ZC706. I'd love to use this kit for radar / lidar applications. For this purpose, it needs coherence between the transmit and receive channel.

I performed the first tests using the TES application (version 0.15.0). I chose the mode: FDD / Custom / Single Cannel. I set the same operating frequencies on the transmit and receive channels. I connected the Tx1 output to the Rx2 input with a cable through an attenuator. I recorded the signal received on the Rx channel and analyzed it in Matalab. It turned out that there is a frequency difference between the transmitted and received signal, about 0.1 Hz. Very small, but in my case unacceptable.

Could the difference be due to the fact that the Tx1 output is driven by LO1 and the Rx2 input is driven by LO2?

Is it possible to force coherence operation between the transmitter and receiver in the ADRV9002?

Can this type of work be forced in TES?

Best regards,


  • 0
    •  Analog Employees 
    on Jun 1, 2021 9:28 AM

    Hello Adam,

    Apologies for the delayed response, I wanted to verify a few bits of information before responding. Before going too deep into this response I would like to ask if you could send some pictures of your setup? That way I could tailor my response to your unique setup.

    As for the application, the ADRV9002 product should be well able to perform as a RADAR, depending on the application thereof. The frequency error you're seeing likely is due to the fact that each RF signal chain is being driven by a different LO. If you were to change to a TDD setup you could use the same LO to drive both the Tx and the Rx, which should remove the frequency error issue at the cost of being able to run both RF signal chains simultaneously. If this TDD setup is viable to you do let me know!

    Keep an eye out for the next release of TES also, it should be posted online sometime in the next few days. This next release will allow for one LO to drive multiple signal chains when in Fast Frequency Hopping mode. That's a flexibility you could exploit in your setup potentially.

    Additionally, you should also inspect the fine details of the Multi-Chip Sync functionality. Even though you are using only 1 ADRV9002 part the MCS system also has an impact on the latency of the RF signal chains, as well as on LO synchronization on device start-up. Well worth a look.

    However if an FDD setup is a must for you, I would also like to ask if an External LO is an option for you? If so you could ignore the internal LO's altogether, use a single LO source with a splitter and identical path delays from source to LO input. That should completely remove the frequency error problem at the cost of some additional hardware. Is this an viable option for you?

    Any more details you provide will be a big benefit to this discussion! Do let me know if these ideas are of any use to you.

    Best Regards,

  • +1
    •  Analog Employees 
    on Jun 4, 2021 8:41 AM

    Hello again Adam,

    Our applications engineers have had a bit of time with your setup, they've come back to me with some feedback.

    You are using two LOs, one each for Tx1 and Rx2. You observe 0.1 Hz difference in the frequency domain. I tried on my setup 900MHz, using narrow band. If I set the IF frequency to be 490K on Rx2, so the LO for Tx is 900MHz and Rx is 900.49 MHz, there will be certain difference between what we calculate(on air) vs what we expect. Using the two LOs will have 0.3 Hz difference in my case. I suggest trying 2 tests:

      1. Using the same LO see if there is any difference
      2. Using different LO but set IF to 490K and 0 to see if there is any difference.

    I tried both test, first test there is no difference, and second test, there is a difference. With 0 IF, the frequencies between Tx and Rx are exactly the same, with 490K I see 0.3 Hz difference.

    Give these tests a go and report here what you find. The more information your provide the better we can assist you.

    Best Regards,

  • Hello Oisin,

    Thank you for the information. I set the IF to 0 Hz and actually the frequencies on TX and RX are the same. I wouldn't have guessed it myself. I don't know how to set one LO in TDD mode. That's why I didn't do the first test, but from your description I understand that it doesn't matter if we use one or two LOs, the difference comes from the IF device other than zero.

    Best Regards,