• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 99 subscribers
  • Views 339 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM6548: Power Domain Issue for MCU_MDIO0_MDC and MCU_MDIO0_MDIO signals

Johnathan Feuerstein
Intellectual 340 points
Part Number: AM6548

Hello,

While probing the MCU_MDIO0_MDC and MCU_MDIO0_MDIO signals (at ball L1 and L4), I noticed that they were being driven at 1.8V on the AM65x EVM.

According to the datasheet, these signals are both powered by the VDDSHV0_WKUP domain (which corresponds to balls U8, V7, W8, and Y7). Looking at the EVM schematic, the VDDSHV0_WKUP pins are powered by 3.3V. Based on the datasheet I would assume the MCU_MDIO0 signals would then be driven at 3.3V, which matches the schematic labeling convention of "MCU_MDC_RGMII_3V3" and "MCU_MDIO_3V3".

However, these MDIO signals are consistently driven at 1.8V before the voltage translating buffers. I came across this issue in our own design, and this prevented communication with the ethernet phy since we used a different buffer scheme to translate 3.3V to 1.8V. 

Are these signals actually driven by another domain, or maybe the MDIO module internally regulates the power? Maybe just a typo in the documentation? Please advise. We just want to make sure we revise our design appropriately to workaround this.

Thank you,

John Feuerstein

  • 0
    TI__Guru**** 161600 points
    Hi John,

    In case of differences between datasheet & EVM you should always refer to the Datasheet recommendations. EVMs are usually done with preliminary specifications before the documentation is released and may have errors, I think this is the case here.

    I've notified the design team to elaborate, if needed.

    Best Regards,
    Yordan
  • 0 in reply to Yordan Kovachev
    Intellectual 340 points
    Hello Yordan,

    Thanks for the quick response.

    I see, good to know. I might not have been completely clear, but the datasheet and EVM seem to line up in this case. They both indicate these signals should be at 3.3V. But when I physically measure and probe these signals, they are at 1.8V.

    Please let me know if you find out any more details.

    Thank you,
    John Feuerstein
  • 0 in reply to Johnathan Feuerstein
    TI__Mastermind 42545 points
    Hi John,
    I have confirmed your observation on the EVM which was designed based on the information provided to us by the chip designers. This may indicate a change that was made to the design which was not communicated to us. I have posted an email to the engineer with the responsibility for the IO buffers in the AM6548. He will check the design database and let me know how these buffers are powered. I will get back to you as soon as I have an answer.
    Regards, Bill
  • 0 in reply to Bill Taboada
    TI__Mastermind 42545 points
    Hi John,
    Initial checks in the SoC database have confirmed that the MCU_MDIO0_MDC and MCU_MDIO0_MDIO are in the VDDSHV2_WKUP power domain instead of the VDDSHV0_WKUP. This was unexpected and we are checking on the required changes to the documentation. This will be listed as a known issue for the E3 version of the EVM. We will attempt to bypass the level translators and ensure that the EVM continues to communicate with the PHY. I will provide an update once we have an understanding on the changes needed to the documentation.
    Regards, Bill
  • 0 in reply to Bill Taboada
    Intellectual 340 points

    Hello Bill,

    Thank you for looking into this. Good to know; we will note this and implement an appropriate workaround for now.

    Regards,

    John