Skip to main content
Visitor II
January 18, 2024
Question

H723ZG ethernet fail

  • January 18, 2024
  • 4 replies
  • 1511 views

Hello,

I have made a simple project using NUCLEO-H723ZG to constantly transmit data via UDP with some low but non-trivial speed (~2300 1KB datagrams /s). Unfortunately, after about 3h of such transmission, the board crashed, and afterwards, the ethernet peripheral appears bricked -- it fails to start, it does not establish the link with other devices, only the orange LED on the port lights constantly. Otherwise, the MCU seems perfectly ok.

I tried with an another board and exactly the same thing happened; this time I was monitoring the temperature of the components but there was no sign of overheating.

Do you have any ideas what may be happening and how to prevent that?

Thank you in advance

    This topic has been closed for replies.

    4 replies

    Technical Moderator
    January 18, 2024

    Hello,

    You need to elaborate more about " the board crashed", is it a hardfault or the MCU is running but the transmission is stopped? That kind of issues need to be debug and see where it fails ..

    Super User
    January 18, 2024

    Get a eth. hub or switch with LEDs indicating connection and speed on ports. Connect your board to the hub. Will it see the connection? If not, the PHY chip may be fried up (or its crystal dead, or whatever).

    mbqAuthor
    Visitor II
    January 18, 2024

    As I wrote, it does not. Entire peripheral is dead, it doesn't respond to MCU neither to other devices via Ethernet. Yet I don't think it has physically been "fried", there is no sign of overheating and the situation is reproducible.

    Super User
    January 18, 2024

    Leave other devices for now. There are two "peripherals" - the ETH and PHY (whatever it is). Are the ETH registers readable? Where does ETH_init stuck? or is there a hardfault?

    What is the power source for your boards?

     

    mbqAuthor
    Visitor II
    January 19, 2024

    True, I was imprecise, by the "peripheral" I suspect I meant the PHY. On the MCU/MAC part, I see no problem, there are no hard faults or errors, it initiates and then behaves as if there is no link.