DP83848K: About 100 Base-TX UTP Differential Output Voltage test

Part Number: DP83848KDP83848K

Hello,

My customer tested the 100 Base-TX, UTP Differential Output Voltage (+Vout/-Vout) with DP83848K-MAU-EK EVM.

But these results ware failed.

(Test result:)

+Vout: Trial 1 (Worst): 1.0670 V, Trial 4: 1.0667 V, Trial 3: 1.0665 V (Total 6triais)

-Vout: Trial 5 (Worst): -1.0751 V, Trial 1: - 1.0749 V, Trial 2: -1.0698 V (Total 6triais)

Pass Limits:(950.0 mV to 1.0500 V)

(Test conditions:)

EVM: DP83848K-MAU-EK

Test equipment : DSO90804A(KEYSIGHT)

Probe: Model: 1130B(KEYSIGHT)

Disturber Source: 33250A (KEYSIGHT)

They tested other company PHY on same test conditions then both test was passed.

(Question)

(1)Does DP83848K confirmed to pass the 100 Base-TX, UTP Differential Output Voltage test with DP83848K-MAU-EK?

(2)About the center tap of pulse trans,

According to datasheet(65page), the decap is placed but it isn’t placed in DP83848K-MAU-EK.(Please refer below.)

 

Do you think this is the possible reason of test fail?

(3)Do you have any solution or idea? (ex. the change RBias value.)

(4)Could you please tell us possible reason?

 

Regards,

Tao_2199

  • Hi Tao,

    To answer some of your questions:

    1. DP82848K-MAU-EK is confirmed to pass the test.
    2. Not having those decaps should not be a major issue, since you have the other decaps that are closer to the PHY.

    I have a few questions of my own:

    1. What is the RBIAS resistor value you are currently using?
    2. Are you terminating the MDI lines with 100 ohms?
    3. Can you provide a block diagram of your test setup?

    Regards,

    Adrian Kam

  • Hello Adrian,

     

    Thank you for reply.

    To answer some of your questions:

     

    1. What is the RBIAS resistor value you are currently using?

    (Ans.)

    RBIAS resistor value is 4.87ohm. It isn't changed.

    2. Are you terminating the MDI lines with 100 ohms?

    (Ans.)

    Yes. It is set to 100ohm on test fixture.

    3. Can you provide a block diagram of your test setup?

    (Ans.)

    Please see attached file.

     

    Regards,

    Tao_2199

  • Hi Tao,

    Is the RBIAS resistor within 1% tolerance? If not, I would suggest getting a resistor with that tolerance. Otherwise, I can also suggest adjusting the RBIAS resistor value to see if it improves.

    Regards,

    Adrian Kam

  • Hello Adrian,

     

    They use DP83848K-MAU-EK and the RBIAS(R24:4.87k) isn’t changed.

    So, the RBIAS resistor is within 1% tolerance

    I attached block diagram of test setup in my first question.

    Could please give us some comment or advice about block diagram?

     

    Regards,

    Tao_2199

  • Hi Tao,

    I do not see any issues with the block diagram.

    Even though you are using the EVM, can you still try to adjust the RBIAS resistor value and see if there is a beneficial difference?

    Regards,

    Adrian Kam

  • Hello Adrian,

     

    Thank you for reply.

     

    About RBIAS resistor,

    As you know, if RBIAS resistor(4.87k±1%) is exceed, IEEE standard can’t be met.

    They need more high precision resistor to adjust within 1% precision.

    Also, they referred below thread.

    https://e2e.ti.com/support/interface/f/interface-forum/806498/question-about-dp83848/2984979?tisearch=e2e-sitesearch&keymatch=DP83848%252520rbias#2984979

    So, they are hard to adjust RBIAS resistor.

     

    They have some question.

     

    1)Could you please tell possible reason why the Vout Differential Output Voltage test is fail on their test environment.?

    For example;

    Difference in revision of DP83848 used for EVM

    Impedance mismatch of PCB pattern layout of EVM

    Difference in test environment between TI’s and my customer’s

     

    2)About impedance mismatch of PCB pattern layout of EVM

    They confirmed TD+/TD-(single end output) but it exceeds the ideal lower limit of amplitude (2.80V)

    by about 60mV. (Please see attached file.)

     

    Why does it exceed the ideal lower limit of amplitude? Could you please tell us possible reason?

    They guess it is impedance mismatch of PCB pattern layout of EVM as possible reason.

     

    Regards,

    Tao_2199

    【DP83848K-MAU-EK】Voltage swing of TD+ and TD-.pdf

  • Hi Tao,

    1. About RBIAS resistor, the value that is suggested in the datasheet is used to meet IEEE standard. We usually inform customers not to change it to avoid issues. However, if you are failing a test even though the RBIAS resistor is the same as the suggested value, it would be ok to try and adjust it to see if it passes.
    2. Let me consult with my team about exceeding the lower limit of amplitude issue and I can get back to you tomorrow.
    3. As for why the test is failing, I may have overlooked something on the block diagram. It seems you are using a link partner for the test. How are you ensuring that the signal is coming from the DUT and not the link partner? If you are not sure, would you be able to try the test without the link partner?

    Regards,

    Adrian Kam

  • Hello Adrian,

     

    Thank you for reply.

    (1)

    >However, if you are failing a test even though the RBIAS resistor is the same as the suggested value,

    >it would be ok to try and adjust it to see if it passes.

    (Question)

    If the test passed with changed RBIAS resistor value, is there no problem to use changed resistor value for their product?

     

    (3)

    (Answer)

    They are terminating the MDI lines with 100ohm on test fixture as differential output voltage test.

     

    Regards,

    Tao_2199

  • Hello Adrian,

     

    Additional question and answer.

    (1)

    (Question)

    Or the changing value is use for only test and after test, RBIAS resistor will be restored value that is suggested in the datasheet(4.87k±1%).

    Is there no problem about above?

    (3)

    (Answer)

    They tested without Link Partner on test fixture section, but it was failed similarly.

     

    Regards,

    Tao_2199

  • Hi Tao,

    I am looking into a couple of things regarding your issue. I will respond back by Tuesday at the latest.

    Regards,

    Adrian Kam