DP83826E: Possibly broken parts from supplier

Hi Jens,

May I ask how can you determine "appear to have been broken"?

• Are you seeing an link up?
• Could you probe the CLK_OUT pin to see if you are able to see clock signal
• Could you probe Rbias pin to see if you are able to see 1V?
• Are you able to access register with the broken part? If possible, could you give us a register dump from 0x0000 to 0x001F and 0x0467

--

Regards,
Hillman Lin

Hi Lin,

On the 'broken' parts they are all driving a small leakage current on pin 31 of -170uA, No clock output. I am attaching the schematic which is pretty simple. Once the parts are replaced there are no issues. The broken parts do respond to MDIO and they seem to be in enhanced mode despite the mode select pin being tied to GND. The magnetics interface of the broken parts appears to be active and shows plausible signalling although there is no link through to RMII.

Oscilloscope measurements of Rx and Tx lines.

Hi Jens,

Thank you for sharing the information. If you are not able to see output clock, that means the PHY is not power up properly.

Could you probe the following pins:

• Rbias pin
• Plot VDDA3V3, VDDIO, CEXT, and MDC at the same time. We would like to check the power sequence.

--

Regards,

Hillman Lin

Measurement 1 Startup 1ms/div 2V/div . Trace 1->VCCIO, 2->VDDA3V3, 3-RBIAS, 4->CEXT

Measurement 2 Startup 1ms/div 2V/div. Trace 1->VVCIO, 2-> MDC, 3-> RBIAS, 4-> CEXT

Measurement 3 Startup 10ms/div 2V/div. Trace 1->VDDIO, 2->VDDA3V3, 3->RBIAS, 4->Pin31(CLKOUT)

This is the DUT

Hello,

Hillman is OoO today, and will look to respond on Monday when he returns.

Sincerely,

Gerome

Hi Jens,

It seems like Rbias is not pull up to 1V from the waveform you provided. It is most likely be the damage part. Last couple of test to confirm the PHY is damaged by the following test:

• You mention that DP83826PHY is operated in Enhanced mode but we see that you have a pull down in mode selection which is in Basic mode. Have you ground this pin or pull this pin up by accident at any chance. This will might damage the PHY.
• Could you also probe Reset and PWDN pin to make sure the PHY is not operate in Reset mode?

• If you place the bad parts into another PCB, are you seeing this bad behavior?

You mention that this parts is good from different supplier (Digikey), May I ask where did you get the bad parts from? 

• Did you have any detail information such as LOT number?

--

Thank you,

Hillman Lin

Hi Lin,

• The parts are supposed to be in basic mode by design. When we read the registers it seems as if they are in enhanced mode. 
• The bad parts exhibit the same behaviour if moved to another PCB.
• The bad parts were not purchased by us. They are supplied by our manufacturer Kitron. https://kitron.com/ I will have to contact them to get batch information if it is available.
• The only information on lot number I have is what can be read from part markings. All bad parts on the current boards (4 parts on 3 boards) are marked  '826E, TI 198, ALSF G4'.

Trace 3 below is PWRDN/INT pin 21. 100ms/div

Hi Jens,

Thank you for sharing the information. We will wait for your information on Kitron. We will also check internally to see the part you are using a valid parts.

If possible, could you perform a quick test. Could you hold the reset to low for a while after power up and release it. Does it solve the issue?

Could you try to pull down Mode selection (Pin1) and see if that pull the PHY back to basic mode?

--

Thank you,

Hillman Lin

Hi Lin,

Information from our manufacturer: 

Lot/Batch: 2262887CD2

Date: 2229

Distributor: Mouser

Hi Jens,

Thank you for your information. We will discuss with the team internally and get back to you by the end of this week.

--

Regards,

Hillman Lin

Hi Jens,

Based on the picture you provided in Mouser and the Lot number that manufacturer provided. We see a mismatch on the Lot number. 

• The Lot number: 2262887CD2 we have in our system is the DP83826PHY revision B.
• The PHY that you show us the the picture above is the DP83826PHY revision A.

In revision A, we do have some strap resistor change internally. If possible, could you try to remove the ground connection on the Mode selection pin to see if you are able to see any voltage on Rbias and CLK_OUT? The Lot number that manufacture provide is revision B of DP83826PHY. Could you double check with the manufacture to make sure that is the correct Lot number that they provide? Because the Lot number is not matching with the chip you are using.

--

Thank you,

Hillman Lin

Hi Lin,

This is interesting indeed. I wonder if Mouser or our manufacturer is to blame? Anyway I think we will instruct Kitron to scrap the existing stock and buy fresh chips.

Un-strapping the mode pin will require some pcb surgery. I will give it a try.

Hello,

Hillman is OoO and will be returning on Wednesday 12/27.

Sincerely,

Gerome

Hi jens,

Thank you for confirming.

--

Regards,

Hillman Lin