TPS23882B: TPS23882B1RTQR : Not Powering ON as expected

Hi Guru,

I will look into this and provide a response today. Thanks!

Regards,

Brandon

Hi Guru,

When probing the port you should be probing between RJ1_CT0 and RJ1_CT1. This will be your differential, you should not probe these two signal between ground. 

Also, from the schematic, looks like your magnetics or incorrectly wired. Your CMC of the ethernet magnetic should be on your data only side, not your power + data side which is what you currently have. 

Regards,

Brandon

Hi Sir,

Sorry for the late reply. Due to some unavoidable situations, I am unable from my end.

Regarding the query, please find the below image : 

As per the reference J2, in our schematics it is from the MAG1_D[3:0]P/N

MAGNETICS is placed correctly

As per the reference J3, in our schematics it is H11

Please find the schematics & image below.

I shall also probe & update you at the earlier.

Regards

Guru

7127.PDF-EPCB-9038-O-00_texas.pdf

Hi Guru,

Please let me know the results with waveform captures when you are able to capture. 

Regards,

Brandon

Hi Sir,

Sorry for the late reply. Please find our observations while probing RJ0_CT0 with RJ0_CT1

We have connected 48V supply & shorted the 12V GND & 48V GND. In open condition, without connecting anything  we found this observation 

Please find the below waveform.

When we have connected the RJ45 with the POE tester, we found the following observation, 

also in POE tester, we found the following image in POE Tester.

Please guide us.

Regards

Guru

Hi Sir,

Also, for debugging just to manually turning ON the mosfet or bypassing the mosfet, we got the following observation

Rework done :  R12 directly connected to ground / shorted to ground. 

Then we got the following observation.

The circuit works as per our requirement. Require your support.

Regards

Guru 

Hi Guru,

I have transitioned roles and have re-assigned this to one of our PoE experts to help resolve. 

Regards,

Brandon

Hi Guru,

Please check below items and let me know if it is still not working. Thanks.

1. The data transformer normally makes the PoE on the side without CM choke. Since the PoE current is CM. Please check with H6096NL vendor.

2. Bob Smith terminal normally has a >2kV cap between earth GND

3. /RESET better has a discharge loop. You may short D8.

4. Please short the unused Kelvin sense to AGND

5. You can leave OSS open (DNP R4) if you do not have a net connect to PSE_OSS

Best regards,

Diang

Hi Sir,

1. The data transformer normally makes the PoE on the side without CM choke. Since the PoE current is CM. Please check with H6096NL vendor.

We are communicating with the vendor. We shall update you at the earliest.

2. Bob Smith terminal normally has a >2kV cap between earth GND

We are using the 1210GC102KAT2A – 1000 pF ±10% 2000V (2kV) Ceramic Capacitor X7R 1210 capacitors at  C241, C242, C243, C244, C237, C238,  C239, C240

3. /RESET better has a discharge loop. You may short D8.

Rework done as per the suggestions.

4. Please short the unused Kelvin sense to AGND

As the component is QFN Package, we are unable to do it. Is there any alternative for that or is it mandatory to do the rework.

5. You can leave OSS open (DNP R4) if you do not have a net connect to PSE_OSS

Rework done as per the suggestions.

Even after the reworks, the POE section is not working.

Regards

Guru

Hi Sir,

1. The data transformer normally makes the PoE on the side without CM choke. Since the PoE current is CM. Please check with H6096NL vendor.

H6096NL  supports  4PAIR PoE 75W

Regards

Guru

Hi Guru,

Thanks for your reply. 

1. It is not about the power rating of the data transformer. You may need to know which side (left or right) is for PoE + Data and which side is for Data only.  

Can you get the waveforms of Q1 and Q2's Vds and Vgs when port is open and connected? It seems FETs behaviors are strange.

Best regards,

Diang 

Hi Sir,

Please find the waveforms for Q2 for Vgs & Vds when connected & unconnected conditions.

Vg_Vs when POE tester not connected.

Vg_Vs when POE tester connected

Vd_Vs When POE tester unconnected

Vd_Vs when POE tester connected

Also we are using the application as per the attached image

Regards

Guru

Hi Sir,

Please ignore the earlier waveforms.

In board we have GND & 48V_GND, The above waveforms are captured without shorting the grounds. Please find the waveforms with shorting the grounds.

Vg_Vs When POE tester unconnected

Vg_Vs When POE tester connected

Vd_Vs When POE tester unconnected

Vd_Vs When POE tester connected

RJ1_CT0 & RJ1_CT1 When POE tester unconnected

RJ1_CT0 & RJ1_CT1 When POE tester connected

Regards

Guru

Hi Guru,

Thanks for your quick reply. I will get back to you by this week.

Best regards,

Diang 

Hi Guru, 

I am out of office currently. Will get back to you later. Thanks for patience. 

Best regards,

Diang

Hi Guru, 

Thanks for updates. Yes it makes sense that FET Vds rating is the cause of abnormal voltage at the port.

Best regards,

Diang

Hi Sir,

Now we are removing the mounted FET Q1,Q2. And want to change the chipset from autonomous mode to I2C mode. Via I2C we want to manually enable the gate and want to check the voltage of Vg pin (in open condition at Vg pad) just to ensure in next revision it shall work properly and no other issues are there in the board. 

So as per datasheet & schematics, we are using address 0X20. But still we are unable to detect the Chipset. Please support us, that what shall be the address for the CHISET. We have pulled LOW for A4,A3,A2,A1 pins as per the schematics.

Regards

Guru

Hi Sir,

Please review the schematics and share the feedback.

Regards

GuruPOE_TEXAS.pdf

Hi Guru,

Thanks for your updated schematic.

We will get back to you in 2-5 business days. Thanks for your patience. 

Best regards,

Diang

Hi Guru,

You should be able to use 0x20 slave address when A4-1 are ground. Do you have a Saleae waveform about the SDA and SCL at TPS23882B. Also, can you try if you can access the boardcast address?

See the comments about the schematic but may not be related to the I2C address.

The data transformer normally makes the PoE on the side without CM choke. Since the PoE current is CM. Please check with H6096NL.

/RESET better has a discharge loop. You may short D8.

Best regards,

Diang

Hi Sir, 

Please find the screenshot from H6096NL

The power shall be connected to MCT1/ MCT2 pins. 

So shall we proceed for new Gerber by changing only the MOSFET (considering remaining circuitry is fine) 

Regards

Guru 

Hi Guru,

Hi Sir,

Thanks for the confirmation.

We are updating the symbol (alone) in schematics, so we are keeping the connections as they are.

Regards

Guru