• State
  • Replies 3 replies
  • Subscribers 33 subscribers
  • Views 43 views
  • Users 0 members are here
Support feedback
Options
Options
Related

DP83867IR: DP83867 Ethernet 1000BaseT issue

Felix Deininger
Prodigy 10 points
Part Number: DP83867IR


Dear TI-Experts,

while we where performing the ETH-compliance test we faced two issues with our current design of the ETH-Interface.
1. We are failing the Template test in two points (F and H).
Template_Fail_pointF.PNG

Template_Fail_pointH.PNG

2. We also are failing the Master Filtered Jitter by around 200 ps
Jitter_Fail.PNG

We have used the DP83867 in some of our designs already and had no issues with the compliance test before. 
The schematics in this design are taken over from our known-good designs.

Schematics_DP83867IRRGZ.PNG

Schematics_DETH_Connector.PNG

Troubleshooting:
We tried...
1. ...swapping DUTs, ETH-cables, compliance boards
2. ...changing values of center taps
3. ...following every step of the TI DP83867 Troubleshooting Guide (3.6 Compliance Debug)
    www.ti.com/.../snla246c.pdf

- All measures to this point including adjusting register values did not show any effect on the issues.

4. After that we tried adjusting the Rbias from 11K to 10K. This resulted in passing the Template test completely (including points F and H).
However the Jitter Issue is still unaffected.

Our questions to you:

1. Do you have any ideas/experiences on what could be the cause of the observed issues?

2. Is adjusting Rbias an acceptable solution for the first issue? Is the performance of the PHY affected by this change?

3. Can you recommed any other measures to fix the jitter issue?

  • 0
    TI__Mastermind 20566 points

    Hi, 

    2. The performance of the PHY will not be affected for changing RBIAS from 11k to 10k and this is actually one of workarounds for some of the compliance test failures (distortion test) that we suggest. However, other parts of the compliance test may be affected due to the change in the RBIAS value. 

    For jitter issue, I would check if the crystal is within our recommendation. This test usually fails because the crystal exceeds the ppm requirement. 

    Best,
    J

  • 0
    Prodigy 10 points

    Hello J,
    Thank you for your answere. We passed the complete compliance test (with the ecxeption of the Master Filterd Jitter) with a 10K Rbias equiped, so we will continue to use it.

    On paper the crystal seems to fit the requirements of the PHY. After your suggestion we started further investigations on the cryctal, that are still running.

    Greetings,
    FD

  • 0 in reply to Felix Deininger
    TI__Mastermind 20566 points

    Hi Felix, 

    As previously mentioned, there should be no issue with using 10k for RBIAS. 
    If crystal is giving out more noise than expected, this can cause the PHY to give out additional jitter causing the failure of master filter jitter test. Please keep me updated. 

    Best,

    J