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.

DP83826E: difference between hardware revision 0 and 1

Part Number: DP83826E

Hello,

we have a board with DP83826E hardware revision 1 (PHYIDR2 register field revision number = 1) running.

A second PCB with DP83826E hardware revision 0 (PHYIDR2 register field revision number = 0) and the same software is not running. We get no link.

What is the difference between these two revisions?

What must be modified in software to get the revision 0 running?

  • Hi Sven,

    The only different between revision 0 and revision1 on DP83826E is the internal strap resistor of the pin. If possible, could you double check on the strap resistor and see if you are seeing the right strap mode between Revision 0 and Revision 1?

    The strap register is 0x0467.

    May I also confirm one more time all the schematic are the same between two boards right?

    --

    Regards,

    Hillman Lin

  • Hello Hillman,

    thank you for your quick response.
    The schematic for both boards is the same. The only difference is the revison of the DP83826E.

    We just read the strap register. The value is the same for both revisions: 0x0086 corresponding to the settings in the schematics. But we discovered that the revision 0 starts up as 10Mbit/s halfduplex instead of the wanted 100Mbit/s full duplex. We now initalised the chip to 100Mbit/s full duplex in software. Checking the registers showed that this worked correct. When sending an ethernet frame from another ECU, the revision 0 looses the link and jumps to 10Mbit/s halfduplex. It recovers automatically to 100Mbit/s full duplex after some seconds, but fails again when we send the next frame...

    The revision 1 chip doesn't show this behavior.

    Regards

    Sven

  • Hi Sven,

    Thank you for sharing the observation. It seems like it is an signal quality issue that is making the PHY drop the speed. If possible, could you read register 0x0218 for both of the register and provide a register dump for Register 0x0000 to 0x001F for both Revision 0 and Revision 1 PHY?

    --

    Regards,

    Hillman Lin

  • Hello Hillman,

    I just read out all registers you asked for (see below) directly after startup before starting communication. For the rev 0 chip I read the registers with our standard initialisation we use for rev 1, which worked there without any problems, and with additional setting of the registers 0x00 and 0x04 as shown below:

    register rev 1 rev 0 rev 0
    address standard init standard init init
          reg 0x00 = 0x3100
          reg 0x04 = 0x01E0
    0x00 0x3100 0x3000 0x3100
    0x01 0x786D 0x786D 0x786D
    0x02 0x2000 0x2000 0x2000
    0x03 0xA111 0xA110 0xA110
    0x04 0x01E1 0x00A1 0x01E1
    0x05 0x41E1 0x41E1 0x41E1
    0x06 0x0005 0x0005 0x0005
    0x07 0x2001 0x2001 0x2001
    0x08 0x0000 0x0000 0x0000
    0x09 0x0000 0x0000 0x0000
    0x0A 0x0102 0x0100 0x0100
    0x0B 0x0009 0x0000 0x0000
    0x0C 0x0000 0x0000 0x0000
    0x0D 0x0000 0x0000 0x0000
    0x0E 0x0000 0x0000 0x0000
    0x0F 0x0009 0x0000 0x0000
    0x10 0x4615 0x0611 0x4615
    0x11 0x010B 0x010B 0x010B
    0x12 0x0000 0x0000 0x6C00
    0x13 0x0000 0x2800 0x2800
    0x14 0x0000 0x0000 0x0000
    0x15 0x0000 0x0000 0x0000
    0x16 0x0100 0x0100 0x0100
    0x17 0x0045 0x0041 0x0041
    0x18 0x0400 0x0400 0x0400
    0x19 0x8C01 0x8401 0x8C01
    0x1A 0x0000 0x0000 0x0000
    0x1B 0x007D 0x007D 0x007D
    0x1C 0x05EE 0x05EE 0x05EE
    0x1D 0x0000 0x0000 0x0000
    0x1E 0x0102 0x0102 0x0102
    0x1F 0x0000 0x0000 0x0000
           
    0x218 0x2B … 0x069   0x2B … 0x035
           
    0x467 0x0086 0x0086 0x0086
           

    If it is a signal quality issue, why is it working with the rev 1 chip and why does the rev 0 shows totaly different values right after startup?

    How can we distinguish these two revisions by their ordering number for future PCBs?
    Where in the datasheet are the differences between these two revisions described?

    Regards

    Sven

  • Hi Sven,

    Between rev1 and rev0, it is mainly changes on the strapping option. In revision 1, we also have enhanced mode for EtherCAT application. From the register read on 0x000A and 0x000B, I did see that your odd nibble detection is enable which is will always bring unexpected issue on EtherCAT application. 

    If possible, could you also change the register 0x000A and 0x000B to see if that help with your application.

    --

    Regards,

    Hillman Lin 

  • Hello Hillman,

    we just tested it with register 0x0A set to 0x102 and register 0x0B to 0x09. These are the values which are working in rev 1. There is still no communication with the rev0 chip possible.

    Please provied us an errata sheet and a detailed description of the differences between rev 0 and rev 1 and a workaround to get rev 0 working.

    Regards

    Sven

  • Hi Sven,

    Could you ask FAE to send me a message? I can share the documentation through the email but those information is under NDA which require send those information through the email.

    --

    Regards,

    Hillman Lin