DP83867IR: Link fails

John Ford

Part Number: DP83867IR

Hi all.

I've been working with this part on a custom board, and suddenly the port does not link up to my host. I tested the host port. It's fine.

The configuration hasn't changed. Measuring the 4 RJ-45 pairs with the ieee test #1 mode shows that three of the 4 pairs look correct as far as the ieee test #1 waveform, but the waveform on the 4th line has an amplitude that is way too high. The peak voltage on the square wave part. are about 1.5 volts, and the pulses are over 1 volt. The terminator on the wild pair is good.

It appears to me that the PHY is faulty, but I would like to hear from someone with more experience with this part. Any other ideas?

over 5 years ago

0 Nikhil Menon over 5 years ago

TI__Genius 11725 points

Hi John,

I just want to confirm, you said it was it working before? Is the 4th line you mentioned channel D? If so, please try using 10/100 speed to see if channels A and B are still functional.

Thank you,

Nikhil

0 John Ford over 5 years ago in reply to Nikhil Menon

Intellectual 540 points

Thanks. Yes it was working.

That's a great idea. A and B are OK, I think. I'll try that tomorrow.

John

Nikhil Menon said:

Hi John,

I just want to confirm, you said it was it working before? Is the 4th line you mentioned channel D? If so, please try using 10/100 speed to see if channels A and B are still functional.

Thank you,

Nikhil

0 Nikhil Menon over 5 years ago in reply to John Ford

TI__Genius 11725 points

Hi John,

Are you still seeing any issues? Are channels A and B still functional?

Thank you,

Nikhil

0 John Ford over 5 years ago in reply to Nikhil Menon

Intellectual 540 points

Hi Nikhil.

Yes, I still have the problem. Channels A and B work fine, i.e. if I force both enss of the link to 100 Mb full duplex, the link comes up.

So probably the PHY is damaged?

Is an ESD/TVS part needed or recommended on the MDI lines of this part? The eval modules that I've looked at don't have one.

0 Nikhil Menon over 5 years ago in reply to John Ford

TI__Genius 11725 points

Hi John,

ESD protection/TVS diodes are not a requirement, but for applications that may involve exposure to an ESD event, they are recommended. Based on what we see in this situation, there may have been an ESD event. Please try testing another device to confirm.

Thank you,

Nikhil

0 John Ford over 5 years ago in reply to Nikhil Menon

Intellectual 540 points

Hi Nikhil. This board had its TVS device removed for testing due to some other issues with CRC errors on the link. Removing it did not help the CRC errors, but I guess we killed the PHY with ESD while experimenting.

We have 2 TPD4E05U06DQAR devices on the 4 differential pairs from the PHY to the magnetics. Do you think that's a good part for this use?

John

0 Nikhil Menon over 5 years ago in reply to John Ford

TI__Genius 11725 points

Hi John,

Yes, we have used this part with our PHYs, including the DP83867, in other designs.

Thank you,

Nikhil

Interface

Interface forum

DP83867IR: Link fails