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.

DP83825I: MAC to PHY RMII PCB Layout

Part Number: DP83825I
Other Parts Discussed in Thread: DP83825

Hi,

I'm currently working to integrate a DP83825I into a design that uses and ESP32 as the main controller. The ESP has an integrated MAC which I am using to communicate and control the PHY. To be clear MDIO is working as expected. I can control and inspect the status of the PHY from the MAC but frames transmitted from the MAC to the PHY rarely seem to make it intact. I've tried a couple different revisions to get the MAC and PHY communicating properly over the RMII link, But thus far have come up empty. I initially tried designing the RMII link without termination resistors as they're supposed to be integrated into the PHY but I noticed a fair amount of over and undershoot not only on the data lines but also the clock line from the ESP to the PHY. The second revision includes some 33 ohm series termination which has helped with the extra noise a little but the RMII link still isn't working as expected. What's the recommended layout for the RMII signal traces? I've attached the relevant section of my layout which shows how the RMII link is laid out. Any help with this would be much appreciated. One interesting thing to note is that when I probe the clock line from the ESP to the PHY with my scope, the PHY starts to receive some of the frames from the MAC in the ESP, so there definitely seems to be something incorrect with my layout.

Thanks, Neil.

  • Hi Neil,

    Can you confirm that the RMII traces are routed with 50ohm impedance? Is the DP83825 configured as RMII Slave? How well matched are the MAC RX lines, MAC TX lines?

    You can find additional layout recommendations in section 10.1 of the DP83825I datasheet. 

    Regards,
    Justin 

  • Hi Justin,

    Thanks for the quick reply! After rechecking the design it looks like i'll have a serious impedance matching problem. The traces as routed are going to be 150ohms, this is particularly bad since I've routed on a two layer board. One more question I do have is do the RMII signal lines require series termination or with the correct trace impedance can I go directly from the MAC to the PHY?

    Thanks, Neil. 

  • Hi Neil,

    The RMII lines do not require additional series termination, that is incorporated in the DP83825.

    Regards,
    Justin 

  • Hi Justin,

    Thanks again for the reply. If you'll indulge me one more question. Is there a reason that the typical rules of thumb don't seem to apply here? My understanding was that transmission lines under 1/10th the wavelength would have negligible transmission line effects and that the need for impedance control was minimal. In this case all the traces are less than 4cm in length or 1/150th the wavelength of the 50MHz RMII signal. I assume there is something i'm not considering in this regard but i'm unsure of what it is at this point. I've tried a nearly identical layout without added series termination and had the same results.

    Thanks, Neil.

  • Hi Neil,

    The difference in impedance between transmitter, transmission line, and receiver will act as a boundary causing reflections at each point there is a mismatch. The length of the trace is not as important as matching impedance of the drivers. 

    Regards,
    Justin