Post Go back to editing

ADRV strange PHASE behaviour after calibration procedure

Hello,

We have 16 ADRV9026 mounted on a single board, everyone is clocked by the same signal and multi-chip sync procedure is successfully performed.
All chips, and all chip channels are fed by a CW signal @3 GHz with power tuned to have around 0dBFS on digital part.
RX samples are simultaneously dumped by our FPGA through JESD connection.

What I see is different phases on each chip, and this is normal as electric paths are slightly different on board itself,
but every time I perform the procedure described in user manual, "CALIBRATION GUIDELINES AFTER PLL FREQUENCY CHANGES",
phase differences between chips varies (up to +-15 deg), even if we don't modify RX pll frequency.

How can it be ?

BR

Pietro

Top Replies

Parents
  • All chips, and all chip channels are fed by a CW signal @3 GHz with power tuned to have around 0dBFS on digital part.
    RX samples are simultaneously dumped by our FPGA through JESD connection.

    Can you test with -5dBFS signal level?

    What I see is different phases on each chip, and this is normal as electric paths are slightly different on board itself,
    but every time I perform the procedure described in user manual, "CALIBRATION GUIDELINES AFTER PLL FREQUENCY CHANGES",
    phase differences between chips varies (up to +-15 deg), even if we don't modify RX pll frequency.

    Are you seeing a different phase at the same frequency from run to run? Hope you are calibrating only when there is a frequency change. Is your frequency change falliing in Type 1 or Type 2. For Type 2, you will have to run ADI_ADRV9025_LOOPBACK_RX_LO_DELAY and
    ADI_ADRV9025_TX_QEC_INIT calibrations additionally on top of Type 1 cals.

  • Hello

    1) "Can you test with -5dBFS signal level?" done, no luck!

    2) I am seeing different phase at the same frequency almost every run I perform;

    3) I am making a Type1 frequency change from the one set in initdata.c: 3.05GHz to 3.0GHz

    Can you please check my calMask?,

    maybe something is missing or wrong; initCals.calMask = 0xC0D97FF;

    of course I run ADI_ADRV9025_TX_LO_LEAKAGE_EXTERNAL once the internal calibrations are done and after enabling PA.

    BR

    Pietro

  • maybe something is missing or wrong; initCals.calMask = 0xC0D97FF;

    Can you test with cal mask of 0xD37FF and let us know the outcome?

  • Hello,

    It seems the problem was solved by changing calMask as you suggested.

    I tested bot type1 and type2 pll changes

    What was the matter ?

    Pietro

Reply Children
No Data