• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 33 subscribers
  • Views 1352 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.

DP83822H: TM4C129XNCZAD and DP83822 transformerless configuration unreliable

Joseph Orender
Intellectual 260 points
Part Number: DP83822H

Hello all,

I have implemented a DP83822 tm4c129x combination and connected the DP83822 in rmii mode to its mac. The tiva is using a stock lwip config from tivaware and we have coded the configuration to be force 10 full duplex and commented any autonegotiation register calls. The links are unreliable at times. Autonegotiation has never worked and forcing the speed seems to be the only way to connect. I have followed TI's app note on transformerless operation and although the app note suggests that the devices will perform in 100TX mode, we have never actually been able to achieve a 100TX link. 

In the schematic, I have 12 DP83822 phys connected to my fpga. The phys have a 51 ohm pullup on the pairs followed by a series 0.1uF capacitor. They connect to the tiva with another 51 ohm pullup on the pairs. The link length is very short - on the order of about a couple inches from phy to tiva. 

I have tried various combinations of register settings and forcing vs autonegotiation and was wondering if the hardware configuration is possibly an issue or if there is a preffered method to connect them. One caveat is that the tx pairs are connected to tx on the other side and rx as well. This requires that one of the phys be configured to swap the pairs internally. 

Thank you in advance. 

-Joe

  • 0
    TI__Mastermind 18425 points
    Hi Joe,

    Are all the PHYs on the board showing link instability? Do you force MDI or keep Auto-MDIX active? Lastly, would you be able to share your schematics? Only the ethernet PHY related circuit should be fine and if sharing over public forum is a concern then we can share over private message.

    -Regards,
    Aniruddha
  • 0 in reply to Aniruddha Khadye
    Intellectual 260 points

    Hello Aniruddha

    Sometimes all the phys appear to drop out and sometimes it seems isolated to one or two phys. The amount of data we are sending seems to affect the link significantly. Keeping our data rate low helps. 

    I have been forcing MDI and allowing Auto-MDIX and experimenting with both. The PHYs are connected TX to TX and RX to RX. I have experimented with the Tiva Lwip opts and set force 10 full without mdix and on the dp82822 I have allowed auto negotiation and this seems to be the most reliable. I can also set the MDI on the dp83822 manually and this is stable as well. 

    I can share the schematics over a private message. 

    -Joe