AD9371 TX QEC Performance

Hello

We are using the AD9371 IC on our board.
I expect the QEC perfmace can be get over 70dBc
but our board only reduce around 55dBc.

Please refer to this result
This result got to use "MYKONOS_enableTxNco"
LO : 1970MHz, tx1TonekHz : 1000kHz

Other information:

  • MYKONOS_getEnabledTrackingCals
    enableMask : 0x000000FF
  • MYKONOS_getAllTrackCalState
    trackCals : 0xFFFFFFFF
  • MYKONOS_getTxQecStatus
    errorCode : 0x0000
    percentComplete : 0x0064
    performanceMetric : 0x0008
    iterCount :0x0055
    updateCount : 0x0053

We want to improve the QEC performance
Please give us your suggestion.

Thanks!
Tango

  • From the MYKONOS_getEnabledTrackingCals API, looks like the TX QEC tracking cal is not enabled. Have you enabled the TXQEC cal in both tracking as well as init cal mask? Is the update count and the iteration count increasing with every run?

  • In my understanding, TX QEC is enable because TX1_QEC and TX2_QEC is assigned to bit [6] and [7].
    I referred to below table from mykonos.c 

     * trackCals[bit]  |  Bit description
     * ----------------|------------
     *       [0]       | TRACK_RX1_QEC
     *       [1]       | TRACK_RX2_QEC
     *       [2]       | TRACK_ORX1_QEC
     *       [3]       | TRACK_ORX2_QEC
     *       [4]       | TRACK_TX1_LOL
     *       [5]       | TRACK_TX2_LOL
     *       [6]       | TRACK_TX1_QEC
     *       [7]       | TRACK_TX2_QEC
     *       [8]       | TRACK_TX1_DPD
     *       [9]       | TRACK_TX2_DPD
     *      [10]       | TRACK_TX1_CLGC
     *      [11]       | TRACK_TX2_CLGC
     *      [12]       | TRACK_TX1_VSWR
     *      [13]       | TRACK_TX2_VSWR
    And, I had already confirmed that the update count and iteration count increase.
  • iterCount :0x0055
    updateCount : 0x0053

    Yes correct. and you are getting correction of -57dBc. this looks good.

    What is the tone power used.?

    Are you using any internal attenuator ?

    I don't think there is any thing more to do here as update count is incrementing. 

  • > Yes correct. and you are getting correction of -57dBc. this looks good.
    I think this result is not good.
    In the datasheet, QEC performance is grater than 70dBc.

    > What is the tone power used.?
    I think tone power can not be changed with MYKONOS_enableTxNco. Tone power is always 6dB.

    >Are you using any internal attenuator ?
    0dB

    >I don't think there is any thing more to do here as update count is incrementing.
    How to improve the QEC performance?

  • The tone power that you are getting at the output is around -9dBm. Instead of using internal NCO tone, give a tone from the FPGA DDS with a backoff of -3dBFs. the tone power should come as 3-4dBm. .Check the image power in dBC levels. If the tone power is coming lesser, you can check the RF front end matching. There is not much that can be done to improve the QEC performance as the QEC algorithm is working fine.