DP83826AE: Communication failure

Hello,

If the PHY is not linking up it suggests an issue on the MDI path. I see you have two common mode chokes in series per pair, which is not a common implementation. Can you try removing one CMC per pair and see if the HPY can link up?

Furthermore please check the following:

1. Are the MDI trace lengths matched to within 50mils on your layout?
2. Is there a solid GND plane underneath the MDI traces between the PHY and magnetics?
3. Are the magnetics meeting the specs laid out in the DP83826A datasheet?
   1. 
4. Can you try with multiple different cables and link partners? This should remove a dependency on the cable used or link partner.

Best,

Shane

Hi,

1. Trace lengths are of equal length within 5mil.

2. yes, solid GND plane.

3. magnetics datasheet

变压器手册(2).pdf

4. Customer cannot do it, the product is sealed, the external interface is not standard RJ45, and the external cables cannot be replaced.

Hello,

Can you please try to link up 2 of the same designs against each other? As most of the characterization of the device is assuming RJ-45, there is a chance this connector or the cabling may be at fault.

Can you please probe the TRD_P/N pairs differentially while providing a 100 Ohm differential load?

Sincerely,
Gerome

Hi,

  During normal use, the modules are interconnected.
Could you please provide how to perform differential detection on TRD_P/N pairs under a 100 ohm differential load.
The question is how to measure it, the impedance has already been measured.

Hello,

Please use a differential high impedance scope across TRD_P and TRD_N lines while also providing a 100 Ohm load across the lines (surface mount is fine).

Sincerely,

Gerome