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.

DP83867CR: Establish Link 1Gbit

Part Number: DP83867CR

Hi,

I have a problem with the DP83867 Phy.

In my case, ther is no Link to a PC established.

Auto Negotioation is enabled, Pulses were verified via oscilloscope.

Test mode with pattern comes on all 4 channels.

RBIAS is 11k and 1.0V are measured

The MDIO interface works well. Register read/write is no problem.

The unknown Part is the RGMII interface. And here my question:

Is it necessary, that the RGMII interface works well (timing etc) to establish a link to a PC (without higher Layer Information like mac address...)?

Then the link status could be read via MDIO, right?

The PHY LED show no activity.

Thank you very much

Richard

  • Hi Richard,

    Can you share the register dump that you are seeing when there is no link to the PC?

    There could be link issues if you have an unstable clock.

    Thanks,

    Cecilia

  • Hi Cecilia,

    here is the first block of register map in hex.

    R00:1140    R01:7949    R02:2000    R03:A231
    R04:0181    R05:0000    R06:0064    R07:2001
    R08:0000    R09:0300    R0A:0000    R0F:3000
    R10:5048    R11:0002    R12:0000    R13:0000
    R14:29C7    R15:0000    R16:0000    R17:0040
    R18:0050    R19:4000    R1A:0002    R1E:0002

    If you need others registers please let me know this.

    Do you mean a unstable clock from the crystal 25MHz?

    Thank you very much,

    Richard

  • Hi Richard,

    Yes you will need to make sure the crystal at your XI input is within the PPM spec we recommend in our datasheet. 

    Is your RGMII clock skewed or aligned? You may need to review that as well since RGMII needs to have a skewed clock.

    Thanks

    Cecilia

  • Hi,

    I checked the crystal, its 20ppm. To get shure, I changed it to an other one which worked fine in past, but with no working result.

    Is a working RGMII interface necessary to get up a link?

    I tested on an older design, with no RGMII interface and straps for enabling Autonegotiation, there comes up a link (PC recongnized there is somebody on the end of the line and the LED on Network card is blinking )

    In the actual design, with nearly the same configuration nothing works.

    Habe you seen any abnormal thing in the register values?

    Thank you very much

  • Hi Richard,

    A working RGMII interface is not necessary for the link since this is only the link status on the cable side.it looks like in the register values you've sent, on 0x1, bit 2 it shows that the link is down. 

    Is your link partner also set to auto-negotiation? Have you tried testing at a different speed other than 1G and seeing if that works?

    Can you also provide your schematic if possible?

    Thanks,

    Cecilia

  • The partner is a Win10 PC, and sends autonegotiation pulses. I have measured the pulses on all 4 channels when the ethernet cable is connected.

    I don't tested it with other speed settings on PC side.

    The schematic is attached below.

  • Hi Richard,

    Your schematic looks to be okay. Can you test and confirm whether the PHY is out of reset by probing on the RESET pin? Can you also check if you see signal on RXCLK? 

    That is usually another item we test to confirm functionality on the PHY. 

    Thanks,

    Cecilia

  • Reset is probed and ok (2.5V)

    RXCLK is 2.5MHz square wave

    Thanks a lot

    Richard

  • Hi Richard,

    Ok, then it looks like your signals are ok and the PHY is running properly. When you mentioned that you measured the pulses on the 4 channels, what did you see? Could you please share those results with us? If possible can you also share the pulses that the link partner is transmitting as well?

    Have you tried connecting two PHYs together on the cable side to see if you can see link? I am trying to isolate the issue to see if we can at least see two PHYs linking together vs PHY + PC. 

    Thanks,

    Cecilia

  • Hello Cecilia, I  have found the problem.

    I dont't regcognized that while changing the polarity of LED0 from old to my actual design the strap for port mirroring changes.

    So port mirroring is now enabled by startup und nothing worked. When I change this in register als is fine.

    Thank you a lot

    Richard