• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 33 subscribers
  • Views 216 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DP83848J: PHY troubleshooting help: no PHY seen by MCU

Matt Ooi
Prodigy 130 points
Part Number: DP83848J

Good day all,

I was hoping to get some TI feedback and troubleshooting my network design. I am using the STM32F4 lineup with includes out of the box support for the DP83848J. It must be noted that this MCU has a built in MAC. 

I am also using RMII layout for my board.

The prototype is sitting here on my desk and I seem unable to get the PHY to communicate. As I am validating the software settings I wanted to get started on troubleshooting potential physical hardware layout issues. It's worth noting that for some reason, the LED_SPD light is very faintly illuminated when the board has power but the LED_LINK light is not. The jack shown has integrated magnetics.

As early troubleshooting steps performed, I have verified the 50MHz clock is operational with a scope. 

Here is my schematic derived from the reference design for this product:

  • 0
    TI__Mastermind 20975 points

    Hi Matt,.

    thank you for sharing the information. On the magnetic side, I did not see any DC isolation. Is the connector an integrated RJ45 connector? Could you follow the schematic recommendation below on the connector side to see if that help with your performance:

    Did you have register access on DP83848? If so, could you share the register 0x000 to 0x001F?

    --

    Thank you,

    Hillman Lin

  • 0 in reply to Hillman Lin
    Prodigy 130 points

    Hi Hillman,

    I haven't had the chance to look into registers yet. I will report back with findings.

    Two updates worth noting:

    1.) I overlooked the requirement to connect PFBIN1/2 to PFBOUT. This has been corrected and PHY boots with nominal voltage.

    2.) My colleague was able to get LwIP functional on the RMII so there's comms to the PHY.

    3.) Connecting ethernet cable to port and laptop lights up LINK LED and there's DHCP discovery--however there appears to be no communication between PHY and the computer over ethernet. Scoping the back of the ethernet jack on the board shows activity on the line.

  • 0 in reply to Matt Ooi
    TI__Mastermind 31005 points

    Hello,

    Hillman is OoO today, and will look to respond on Monday when he returns.

    Sincerely,

    Gerome

  • 0 in reply to Gerome Cacho
    Prodigy 130 points

    Thank you for your help. I am providing additional information as it may be helpful to pinpoint the transmission failure. Provided below is a better screen-grab from the back of the RJ45 port on the TX+ line. A bit of overshoot but seemingly not enough to fail completely.

    I also understand there are TI-recommended discrete magnetics jacks specifications and the one I've chosen is close--no manufacturer seems to have an exact match for impedance and TI doesn't provide a list of specific jacks they recommend.

    Currently I am attempting to send out a packet and I never receive anything on WireShark. However, I do see the DHCP discovery for the PHY--just no throughput. 

  • 0 in reply to Matt Ooi
    TI__Mastermind 31005 points

    Thank you for this information.

  • 0 in reply to Gerome Cacho
    TI__Mastermind 20975 points

    Hi Matt,

    Thank you for sharing your magnetic information. As long as the specification fall within the following specification, it should be fine.

    From my understanding, it seems like the PHY is able to link up properly, but you did not see any communication between PC and DP83848. If possible, could you also change the link partner to DP83848 to see if you are able to read anything on Wireshark while you change the cable type?

    If possible, Could you also add a decaps on the 3.3V that goes into MDI lines to prevent any noise going into the MDI lines?

    Could you also double check on the impedance matching of your traces especially TD_M and TD_P to make sure they match 50 ohms

    Meanwhile, I will wait for your information on registers.

    --

    Regards,

    Hillman Lin