PYNQ: PYTHON PRODUCTIVITY

RFSoC2x2 Clock Frequency Error

Hi

I’ve been experiencing large frequency offsets when running the base Spectrum Sweep example on my new RFSoC2x2 Board. With the results below:

Tx Freq. (MHz) 600.0 900.0 1400.0 2600.0 2800.0
Rx Fundamental Freq. 596.0 894.0 1390.75 2609.75 2780.75
SFDR 82.2 79.8 68.3 72.7 72.5

I think this look like there is a problem with the clocking on the board?

Further, looking into the schematics LEDs D22-D24 turn on when the relevant PLL chips are frequency locked but only D22 and D24 are currently illuminated. Is this related?

Thanks for the help.

Hi Lawrence,

It seems something has went wrong during the frequency lock of an LMX chip on the board. Since you have indicated that D23 won’t switch on, you are probably having an issue with the LMX chip named U13. This chip is responsible for driving clocks to the RF ADCs, so there should be issues on both receive channels.

Are you able to provide me more information as follows:

  • Are you using a fresh install of PYNQ v2.6 for the RFSoC2x2?
  • Can you confirm the results for the other channel by providing the spectrum sweep table?
  • Have you noticed other parts of self-test fail in any other way?

You may need to have your board checked out, and I will let you know how to do this shortly,

Kind Regards,
David.

Hi David,

Thanks for the quick response!

  1. I have tested this with the SD card provided with the board and have also tried a fresh install on a new SD card. Both have the same problem.
  2. Below is the channel sweep output:
    image002
  3. I’ve run the Display Port, USB, Ethernet and Buttons/LEDs notebooks and they have all ran successfully. Attached is the output from the self-test package giving similar results to the spectrum sweep.
    Self-test-log.txt (17.2 KB)

Connecting the board to an external SDR I can see that the DACs are functioning correctly i.e. single tone at desired frequency. However the RFSoC ADCs are both measuring input tones incorrectly. Again suggesting a problem withe the ADC clocks.

Best,
Lawrence

Hi Lawrence,
This looks like a hardware failure unfortunately. I’ll contact you directly with next steps to resolve this.

Cathal