• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 32 subscribers
  • Views 243 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DP83848Q-Q1: Output at "CLK_OUT" is missing when in "RMII" mode (sometimes)

Markus B.
Prodigy 10 points
Part Number: DP83848Q-Q1

Hello Texas instruments specialists,

I have some weird behavior with DP83848Q-Q1 which I can not understand.
I don't have 50MHz output at Pin "CLK_OUT".

Description:
We use the Pin "X1" to input a 50MHz oscillator clock reference (Signal from a EBRA53T2H-50.000M oscillator).
Furthermore Pin "CLK_OUT" is used as a reference for our MCU.

In most cases this works very well, we have the 50MHz signal output provided and the MCU can be reached via ethernet.
In some cases this "CLK_OUT" output is not given. I have no 50MHz signal output and therefore no connection via ethernet to the MCU.

Examination:
I checked the signals "X1" and "CLK_OUT" with an oscilloscope: 50MHz in, nothing relevant out.
I checked the connection of Pin "MII_MODE": It is connected to 3V3 via 1kOhms Pull-Up.

When I use the 50MHz signal directly from the oscillator instead, everything is working fine.
The only problem is that we do not have the output at "CLK_OUT".

In datasheet "DP83848Q-Q1" it says:
RMII 50 MHz CLOCK OUTPUT: This pin provides a 50 MHz clock output to the system.
For RMII mode, it is not recommended that the system clock out be used as the reference clock to the MAC.

Question:
Is this "CLK_OUT" something I can rely on even though it is 'not recommended' for 50MHz / RMII mode?
If not, why is it working in most cases (approx. 8 out of 10 are functional)?
Under which circumstances there will be no output at this pin?

Hope someone has an answer.
I'm looking forward to hear about that.

Thanks in Advance

Greetings Markus

  • +1
    TI__Mastermind 27885 points

    Hi Markus,

    As datasheet specifically states that this pin is not recommended for using as MAC reference clock, we strongly advise finding another workaround to providing 50MHz clock to both PHY and MCU.

    Sincerely,

    Gerome

  • 0 in reply to Gerome Cacho
    Prodigy 10 points

    Hi Gerome,

    thanks for your reply.
    The workaround will help. I tried it for a few pcb last week.
    But 'Not recommended' does not mean 'not working'.
    Why is it that phrase in the datasheet?

    Also in most cases the "CLK_OUT" works very well.
    I read the datasheet and the 'application note' and I haven't seen any description why it shouldn't work.
    What happens at 50MHz which does not occur at 25MHz?
    Why does this work different? What are the limitations?
    Why is that not working for only 20%-30% of the batch?
    Some physical restrictions / limitations / aversion of the chip or the RMII mode?

    Don't get me wrong: I was just wondering and I want to understand it in detail.
    I also will talk to our customer in the meanwhile and find out what they are thinking about redesigning the pcb.

    Greetings Markus

  • 0 in reply to Markus B.
    TI__Genius 17800 points

    Hi Markus, 

    Gerome is out of office today. Please expect a response later this week. 

    Thanks,

    David

  • +1 in reply to David Creger
    TI__Mastermind 27885 points

    Hi Markus,

    We have not characterized this device in this fashion and cannot recommend that the CLKOUT pin be used as the reference clock. In addition, in general, RMII is a very strict timing interface and we have not characterized the device to work as a clock source for the MAC. We would recommend instead sending the MAC a signal directly from the oscillator for its reference clock instead of the current configuration.

    Sincerely,

    Gerome

  • 0 in reply to Gerome Cacho
    Prodigy 10 points

    Hi Gerome,

    thanks for your reply.
    Although I'm not happy with the description in the datasheet, providing the oscillator signal directly to the MAC and the DP83848Q in parallel (not in series) solved the problem for now. We will change that in our project.

    However, in my opinion the phrase "not recommended" could easily be misunderstood. Especially if it's working while testing the prototypes and even in the series in most cases (it worked for 8 out of 10).
    This should be written more clearly in the datasheet.

    Thanks to you and your team for all your help with this issue.

    Best wishes to all

    Markus