• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 182 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: Crystal Characterization and Magnetics

Mark Ackerson
Genius 10190 points
Part Number: DP83867IS

Customer has a few questions regarding the DP83867IS:

1. Customer is changing  PHY IO voltage from 1.8V to 3.3V. Will the crystal drive circuit be affected by the different IO voltage at all? We’re guessing not, but figured we’d double check just in case. Is it good practice to get our crystal re-tested with this change or can we say with confidence there will not be any change in the crystal drive circuit?

2. We might also have a possible situation coming up where we could have two sets of magnetics on the same signal path/board for the PHYs. Are there any glaring concerns at the operability of this situation? I think we’ll run it through a compliance test to see what we find but if there are any thoughts or specific things to look for in this situation that would be good to know.

3. It could also be possible to have the opposite scenario occur. Other than isolation/ESD issues, are there any other concerns with having no magnetics in the signal path between two PHYs at all?

Please advise.

Thanks,
Mark

  • 0
    TI__Mastermind 18425 points

    Hi Mark,

    For the change in IO voltage, the crystal schematic remains the same. The internal oscillator will automatically adjust to the change without needing any edits to the crystal or load capacitors. 

    In your second question, I am assuming there will be two phys on the same board connected to each other via differential traces with each having its own magnetics. In this case, the configuration should work. However, there might be difficulty in performing IEEE compliance testing. For compliance testing, the PHY would need to be isolated from the other PHY and the signals need to connected to the appropriate test fixture.

    For the third scenario, magnetics need to be placed between the PHYs. So having not magnetics may cause packet errors during communications.

    -Regards

    Aniruddha

  • 0 in reply to Aniruddha Khadye
    TI__Genius 10190 points

    Hi Aniruddha,

    Thanks for the reply.  One clarification on question #2, this particular situation we were investigating is still between 2 PHYs on different boards, connected by an Ethernet cable. So same as a normal situation.

    Normal Connection:
    PHY1 -> Magnetics1 -> cabling -> Magetnics2 -> PHY2

    Double magnetics situation:
    PHY1 -> Magnetics1 -> Magetics2 -> cabling -> Magnetics3 -> Magnetics4 -> PHY2

    So essentially when talking from board to board there will be double the number of magnetics in the signal path.

    Any concerns?

    Regards,
    Mark 

  • 0 in reply to Mark Ackerson
    TI__Mastermind 18425 points

    Hi Mark,

    Double magnetics is a unusual configuration, are they all magnetics with integrate CMCs? Or is this a configuration where 1:1 transformer and CMCs are separate? If its the latter, then this is ok. Having 1:1 transformer and CMCs discrete is ok as long as they are rated for 1000Base-T application and they meet the electrical spec requirement. However, if all 4 magnetics are complete solutions (transformer + CMCs), then we have not tested in this configuration. 

    Since the transformers are 1:1 I would expect the signal to scale correctly. However adding more passive components in the signal path will introduce possibilities of potential loss of signal quality. These effects will be more apparent at long cable lengths (~100m+). Since this is non-standard way of implementing ethernet PHY, it would need to be verified on the application level.

    -Regards

    Aniruddha