• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 33 subscribers
  • Views 663 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.

DP83867IS: RGMII, 802.3 Preamble/SFD

John Booth64
Prodigy 170 points

Part Number: DP83867IS

Tool/software:

Does the RGMII standard or the DP83867 guarantee that an 802.3 preamble will always begin on the rising edge of rx_clk?  It's not an issue when in 10 or 100 Mb modes, but it seems that it should be case when in 1000 Mb mode.  The 802.3 standard defines everything in terms of octets, so things are perfectly clear for a GMII interface.  However the RGMII standard was developed outside of 802.3 and it doesn't seem to show any indication of Preamble or SFD relationship to clock edge.  This seems to be an important detail since a SFD (per the 802.3 definition of 10101011) could be missed if the wrong clock edge is used to re-assemble the octet (e.g 1011xxxx). 

Thanks,

John

  • 0
    TI__Mastermind 19621 points

    Hi John,

    What is the primary concern here? Are we seeing any functional issues at the moment?

    Regards,

    Alvaro
     

  • 0 in reply to Alvaro (Al-vuh-roe) Reyes
    Prodigy 170 points

    We have system where we have new components on either end of the ethernet link.  The symptoms are the receiving end of the link seem to be missing packets.  We can cycle power on that end but the condition persists. If we switch ports on the transmission end the problem resolves itself.  So we are trying to understand what might cause such a condition.

    John

  • 0 in reply to John Booth64
    TI__Mastermind 19621 points

    Hi John,

    Thank you for clarifying, but I have a few more questions:

    1. Could you provide a block diagram of your set up please
      1. Please include the link partner as well
    2. In your setup, where is our DP83867?
      1. Is it on the receiving end?
      2. What is the communication speed?
      3. What MAC interface is being used?

    To answer your initial question, I went into lab and measured RX_CLK vs RX_CTRL (which functions as RX_DV on the rising edge of the clock). It behaved as expected, on the rising edge (RGMII spec document linked below). Below is a scope shot, keep in mind that RX_CLK is delayed by 2.0 ns, so you can see RX_CTL triggered off the rising edge.

    Figure 1 - RX_CTL vs RX_CLK

    0511.RGMIIv1_3.pdf

    Regards,

    Alvaro