Post Go back to editing

ad9631 rx I/Q signal peak clipping

Hi ad9361 experts:

    When I used vivado to observe the rx signal, and found that I/Q signal was too large with peak clipping. As the picture below shows:

   The value of the I/Q signal sometime is 0xffff, which reaches the peak. I am using FAGC. I tried to lower the overload threshold, but the problem still existed. And I changed gain control  mode to MGC, and lower the gain, still not solved. Could you give me some advice? 

Thanks a lot.

Parents Reply Children
  • Are you looping back tx to rx? Instead of looping back can you give a fixed amplitude tone at rx input, set the gain mode to fast attack and then see the Rx output.

  • Thanks for your reply. Forgive me for not answer your question. I am not looping back tx to rx. I inputed a fixed amplitude signal at rx input, and set the gain mode to fast attack. Still the Rx output peak clipping, any suggestion please? Thanks a lot.

  • Hi srimoyi:

      When I set parameter ctrl_outs_index to 0x0A, I found that the control output 1 is always high, which means that Rx FIR filter always overflows, even when there is no signal input into the ad9361. My RXFIRConfig/TXFIRConfig is as below:

    AD9361_RXFIRConfig rx_fir_config = {
    1, // rx
    -6, // rx_gain
    2, // rx_dec
    26,0,-40,-41,-8,22,22,10}, // rx_coef[128]
    128, // rx_coef_size
    {1280000000,320000000,160000000,80000000,40000000,20000000}, // rx_path_clks[6]
    20172411 // rx_bandwidth

    AD9361_TXFIRConfig tx_fir_config = {
    1, // tx
    0, // tx_gain
    2, // tx_int
    -11,11,14,-1,-16,-17,-9,-2}, // tx_coef[128]
    128, // tx_coef_size
    {1280000000,160000000,80000000,40000000,40000000,20000000}, // tx_path_clks[6]
    19611989 // tx_bandwidth

    Dose my setting affect the performance of ad9361 RX FIR? Or the peak clipping has no connection with these RXFIRConfig/TXFIRConfig?


  • Thanks for your help. This problem is due to my error ILA setting. The signal is OK. AD9361 works well. Thanks.