DP83822I: Transient Protection in POE Application

Balz Odermatt

Part Number: DP83822I
Other Parts Discussed in Thread: TIDA-01463

We use the DP83822I in an POE application. During rapid disconnect/connect events of the ethernet cable, high voltages may occur on the PHY side of the magnetics in the PSE as well as in the PD.

As an ESD protection, a TVS diode TVS3V3L4U is placed between the magnetics and the PHY DP83822I.

Measurements on the TX and RX lines (pins 9, 10, 11, 12) still show voltages going below -1.0 V for > 200 ns, which is the lower limit listed in the absolute maximum ratings.

No destroyed parts have been reported so far, but the customer wants to be sure that there is no risk of damage for the DP83822I.

Looking at the ESD ratings for the pins affected, the device should be able to withstand voltages above or below the maximum ratings for a very short time. But the observed transients may well have longer durations compared to the ESD pulse specified in the IEC61000-4-2 standard.

Is additional protection needed? What is your suggestion on this issue?

over 2 years ago

0 Rahul over 2 years ago

TI__Expert 8721 points

Hi Balz,

Here is a reference design using DP83822 in PoE application. Please have a look at it:
https://www.ti.com/tool/TIDA-01463

The PHY will meet the ESD specs mentioned in the datasheet. We haven't used clamping diodes as done in the schematic shared by you.

Regards,
Rahul

0 Balz Odermatt over 2 years ago in reply to Rahul

Prodigy 10 points

Hi Rahul

Thanks for your reply!

The reference design TIDA-01463 contains a TPD4E05U06Q (U7) to protect the PHY, which is very similar to the TVS3V3L4U in my design. This clamping diodes limits negative transient voltages to approx. > - 0.9 V. This still leads to a violation of the absolute maximum rating of the DP83822I, where -0.5 V is the limit.

In the meanwile, I have actually managed to destroy a DP83822I with rapid off/on cycling of the ETH w POE lines .

Best regards,

Balz

0 Rahul over 2 years ago in reply to Balz Odermatt

TI__Expert 8721 points

Hi Balz,

Is it possible to measure the voltage being reached to the MDI pins and if that is not violating the absolute max ratings ?

Regards,
Rahul

0 Balz Odermatt over 2 years ago in reply to Rahul

Prodigy 10 points

Hi Rahul

I measured on the two pins MDIO (19) and MDC (20) and could not see any violation of the max ratings (no higher transients at all during the short off/on cycle).

I'm no longer sure whether the above-mentioned defect was really caused by the transient tests. I cannot reproduce the defect with a second sample. Despite excessive testing the device seems to be able to withstand the transients. Anyway, we would like to have some kind of guarantee that no device will die in the field.

Best regards,

Balz

0 Rahul over 2 years ago in reply to Balz Odermatt

TI__Expert 8721 points

Hi Balz,

Understood, thanks for the update.

Another thing I would recommend is to have a strong RC isolation between connector GND and digital GND.

Regards,
Rahul

0 Balz Odermatt over 2 years ago in reply to Rahul

Prodigy 10 points

Hi Rahul,

1.5 kV isolation of GND and digital GND is ensured in the design.

I got in contact with FAE Jon Bilbao on this topic. I also shared our schematics.

Best regards,

Balz

0 Rahul over 2 years ago in reply to Balz Odermatt

TI__Expert 8721 points

Hi Balz,

This issue is being handled over the email with the FAE Jon, should I continue this over the email and close this thread ?

Regards,
Rahul

0 Balz Odermatt over 2 years ago in reply to Rahul

Prodigy 10 points

Hi Rahul,

We can close this thread, thanks for your help!

Best regards,

Balz

Interface

Interface forum

DP83822I: Transient Protection in POE Application