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.

Ethernet Phy's compatible with AM1705

Other Parts Discussed in Thread: AM1705, AM1707

Hello,

I'm working on a design with the AM1705. I'm looking right now for a PHY chip that is compatible with this processor. I have seen discussions where some people say that the AM1705 is incompatible with RMII. Is this true?

Can you please recommend me a PHY chip that is compatible with AM1705?

Thank you in advance

Best regards,

Patricio

  • Dear Patricio,
    AM1705 does support RMII.
    I think, you can refer OMAPL137 Spectrum digital EVM which is using RMII PHY.
    OMAPL137 (ARM+DSP),AM1707 (ARM) and AM1705 (ARM) are almost same.
    In OMAPL137 EVM, they have used MICREL KSZ8893
  • Hello Titus,

    Thank you for your help. This is my first ARM design and I have just been gathering info on how to design the interfaces, ethernet among them.

    As I wrote in the first message, I thought that RMII wasn't compatible with this processor because of the note:

    that is shown in the features for the latest PSP for the AM1705 processor at:

    http://processors.wiki.ti.com/index.php/DaVinci_PSP_03.20.00.14_Device_Driver_Features_and_Performance_Guide#Ethernet_Driver

    Please correct me if I'm wrong, but it seems that all the problems related to ethernet interface with these processors is because of the reference clock signal. If the clock is supplied by the AM1705 problems could arise due to the fact that the generated 50 MHz clock is not compliant with the RMII specification, which requires a max freq deviation of +/- 50ppm. This is also true when the clock is generated from the PHY chip.

    This is suggested in the AM1705 literature and in PHY IC's literature as well:

    In page 104 of the Tech Reference Manual SPRUH93B for the AM1705 processor there is a note warning this:

    If we take the for example the Microchip LAN8720a, its datasheet also shows a note describing that if the clock generated from this PHY there could be potential problems as well:

    So in conclusion it seems, that the best alternative (that is, to avoid problems) is to use an external clock generator with the required +/- 50 ppm stability.

    I will really appreciate your opinion about my conclusion.

    Best regards,

    Patricio Cohen

  • Dear Patricio Cohen,

    Yes, it seems we should use external clock for RMII.

    I have looked into OMAPL137 schematics which is similar to AM1705, in that, external crystal used for RMII reference clock.

  • Hello Titus,

    Fantastic. Thank you very much.

    Best regards,

    Patricio