DP83822I: Ethernet compliance test fail (jitter test)

Hello Muk,

Thanks for your question. Have you reviewed our application note on jitter analysis for the DP83822? Have you also confirmed that the MDI design is as recommended per our datasheet?

Thanks,

Cecilia

Hi Cecilia-san,

Thank you for your advice.

Customer checked the jitter as below,

Customer think, The clock jitter is no problem because u = 206.3ps in the waveform. It is lower than 375ps that is explained in your application note. 

Do you have any other idea for improvement method?

Thanks

Muk

Hi Muk,

Thanks for following up. Can you share the schematic of your design?

I'd like to review the MDI portion

Thanks,

Cecilia

Hi Muk,

Thanks for sharing schematics. I have some questions about the schematic

1. Have you confirmed that the R2876-2879 are placed on there and used for ground isolation? Also can you share the values because it looks like they are 0 ohms but I would like you to confirm. 

2. Are there separate magnetics being used for the Tx and Rx of the MDI side? 

3. Can you please share the script you are using to test the compliance? And, have you confirmed the setup looks similar to figure 1 with the short cable and fixture? I would just like to confirm that the set up is the same so that we can compare the same expected results.

Figure is shown below in this app note.

Thanks,

Cecilia

Hi Cecilia-san,

1. The ground is divided into EMI countermeasures, but it is set to 0Ω because it is not necessary for now.

2. Is it about L2801 and L2802? It is designed according to the recommended usage conditions of TDK.

3. The customer has done a different setup.
At the recommendation of Tektronix, the customer created the cable below and tested it with the setup below.
The reason for creating the cable is that it was suggested by Tektronix as a way to improve the situation where the waveform cannot be confirmed.

Thanks

Muk

Hello Muk,

Would it be possible to try the set up as recommended in our application note and see if that provides different results?

Thanks,

Cecilia

Muk,

Can you also share the full results with the details of the failure and full numbers recorded?

Thank you!

Cecilia