Post Go back to editing

How to configure AD9988 when RF signals are located in the 2nd Nyquist zone of ADC

Product Number: AD9988

Hi, 

According the section "RECEIVE INPUT AND DIGITAL DATAPATH" in UG-1578, the register 0x2110 has to be set to 1 for an even Nyquist zone.

I had set the register 0x2110 to be 0x1f when RF signal at 3GHz located in 2nd Nyquist zone of ADC (Fs=4G). However, ADC signal is still error.

Are other configurations needed for the even Nyquist zone?

BRs,

Clement

Parents
  • Hello Clement, did you also set 0x2100[0]=1 after the write to 0x2110 to latch these settings? If so, can you provide a screenshot of the error you are seeing, and let us know if you are using ACE or the API with the ADS9 or some other setup? Thanks.

  • Thanks for your reply. I didn’t use ADS9 evaluation board . May I evaluate our system architecture by other ADI tool, e.g. ADI IIO tool?

    Actually, I first verified RF received signal at 1.5GHz, and the ADC output signal is meet my expectation.

    Under the same configurations, I used this api "adi_ad9081_adc_nyquist_zone_set" to set even zone and variable IF mode with NCO at 1GHz or -1GHz is tried to reconfigure. No other error log is shown.

    However, RF signal at 3GHz is not meet.

    I don't know if other configurations are needed to change.

    BRs, thanks.

Reply
  • Thanks for your reply. I didn’t use ADS9 evaluation board . May I evaluate our system architecture by other ADI tool, e.g. ADI IIO tool?

    Actually, I first verified RF received signal at 1.5GHz, and the ADC output signal is meet my expectation.

    Under the same configurations, I used this api "adi_ad9081_adc_nyquist_zone_set" to set even zone and variable IF mode with NCO at 1GHz or -1GHz is tried to reconfigure. No other error log is shown.

    However, RF signal at 3GHz is not meet.

    I don't know if other configurations are needed to change.

    BRs, thanks.

Children
  • Hi Clement,

    Can you explain how you are setting up your system? If you are not using the ADS9 are you using another carrier FPGA board? 

    What ADC sampling rate and decimation settings are you using? It sounds like you have:

    -ADC samping rate = ?

    -ADC input signal 1.5 GHz

    -cddc shift 1 GHz

    -even nyquist zone setting

    -decimation = ?

    Just need to see the full context. Please send your configuration settings so I can understand what you are trying to see, and if you have screenshots of your captured spectrum FFT showing the working and not working case that may be helpful as well. You should set the nyquist zone before bringing up the jesd links. Thank you.

  • Hi kberry, 

    I am using zcu102 + AD9988 to set up the system, and petalinux environment is built-up. 

    Because the 5G NR waveform is used for the received signal, both FFT spectrums are hard to be discriminated the difference. The api "ad9081_nyquist_zone_write" is set in device-tree file. 

    The configuration of the first working case is as follow.

    -ADC samping rate = 3.93216GHz

    -ADC input signal 1.5 GHz

    -cddc shift 1.5 GHz

    -odd nyquist zone setting

    -decimation = 16

    The configuration of the non-working case is as follow.

    -ADC samping rate = 3.93216GHz

    -ADC input signal 3 GHz

    -cddc shift 0.93216 GHz

    -even nyquist zone setting

    -decimation = 16

    The configuration of the other non-working case is as follow.

    -ADC samping rate = 3.93216GHz

    -ADC input signal 3 GHz

    -cddc shift -0.93216GHz

    -even nyquist zone setting

    -decimation = 16

    BRs, thank you.

  • Hello,

    For the non-working case, would it be possible to provide an FFT for a near full-scale, filtered CW input tone (-1 dBFS) at 3 GHz.  The FFT shows notable spectral regrowth compared to working case so it would be interesting to see any notable harmonics are present in the CW input tone case.

  • Hi both,

    I borrowed the ADS9 board and then compared my setting with the memory map generated from ACE.

    ADC input signal at even Nyquist zone can be demodulated correctly for now. 

    Thanks for your support.

    BRs,