TPS2375: PD does not start. The operational mode does not reach PD powered mode

Hi Robert,

Thanks for reaching out. I will get back to you in 2-4 days.

Best regards,

Diang

Hi Robert,

Thanks for your patience. 

In the 1st waveform, looks like the hotswap between VSS-RTN is not turned on after the PoE handshake. UVLO of TPS2375 should already be reached at 53V. In the schematic it is TS2376, and I assume TPS2375 was used actually. 

In 2nd waveform looks like the hotswap on either PD or PSE side reached the inrush current limitation, even the hotswap turns on eventually. Is R321 = 182k correct? And may I know if PG pin is used in TPS23755 to disable downstream circuit during inrush? 

Besides, is it possible to make D302 DNP and solder R329 before D302? So that we can isolated the downstream circuit influence.  

Best regards,

Diang

Hey Diang,

Thanks for the suggestions. Yes I am using TPS2375 instead of TPS2376 as the schematic suggest. And I forgot to mention on case 1 (the non working new design), the load circuit was disconnected. The only load is the R329 4.7kohm resistor. But I can remove the D302 and solder a 0R resistor inplace of the D302. Or do you want to exclude the 10uF capacitor as well?

R321 is 182kohm, but I try different values, like 110kohm, 160kohm, 220kohm and non of them were working. On case 2 (old design) the POE was functional with each of the resistors.

For case 2, where in-rush current liimit was reached, a DCDC converter is connected to the POE voltage. The PG signal is not used. 

See this:

The D302 is needed because of the Option 2 Auxiliary power oring (so the board can be powered from external 12V or POE voltage):

Hi Robert,

Thanks for your updates. 

May I know if you could measure the current flow into TPS2375? It will help to know if the issue is due to overcurrent or current less than the maintain power signature requirement. 

"And I forgot to mention on case 1 (the non working new design), the load circuit was disconnected. The only load is the R329 4.7kohm resistor. But I can remove the D302 and solder a 0R resistor inplace of the D302. Or do you want to exclude the 10uF capacitor as well?"

If so I think it hints the PD circuit causes the issue. The R329 4.7kohm resistor is needed for maintain power signature when you disconnected the load. It offers ~53V/4.7k = 11.3 mA > 10 mA minimum requirement. If you consider 11.3mA has less margin you may try to use 4k.

The way you disconnect the DC/DC is by DNP F201 or FB201? 

“For case 2, where in-rush current liimit was reached, a DCDC converter is connected to the POE voltage. The PG signal is not used. ”

The PG pin is normally recommended to be used to disable the downstream DC/DC converter when PoE PD is in inrush time (~80 ms from PD hotswap turning on ). PG is an open-drain pin, which will set low when PD is in inrush time.

Best regards,

Diang

Thanks. In the next iteration I will try to use the PG. However the issue related to something different. I am still wondering how it is working on a prototype which was build before and it is not working on the new design... If I remove the load circuit and only the 4.7kohm + 10uF is the load it doe snot work, while it should.

There is a jumper at the output of the 10uF capacitor, R330 is not connected, so the VCC_POE is not connected:

I did additional measurements.
CH3: output of the D302 + 10uF + 4.7kohm load
CH4: RTN-VSS voltage

The output has negative 344mV voltage level. It is not close to 53V. It seems like the IC does not want to switch.
What is that ripple after the classification, in the middle of the 53V on the RTN-VSS measurement?

Then, I removed the D302 and measured the output of the IC.

CH3: VDD-GND
CH4: RTN-VSS voltage

On PIN7-VSS I can measure 10ohm in circuit resistance on the old board. There are no external component connected to PIN7. ON the new boards (10 of them), I measure in the range of  >10Mohm, so it is disconnected. Why do I measure 10ohm on the old board?

These chip marking are present:

2375 9CM AOT7G4 - old PCBA - working
2375 9CM AOT7G4 - new PCBA - not working
2375 91M AELYG4 - new PCBA - not working
2375 2BM A6E464 - new PCBA - not working

The R321, R322, R323 resistors are 1% resistors.

So any idea what is the issue?

UPDATE:
On the old PCBA, I desoldered the TPS2375 which has 10ohm on PIN7-VSS, and resoldered a new TPS2375. And it is still working. I got this measurement:
CH3: output of the D302 + 10uF + 4.7kohm load
CH4: RTN-VSS voltage

So the question, both hardware has same components, same circuit, same layout. Old design working, 10pcs of PCBA with the new design does not. I have already exluded soldering issues, manufacturing related issues. 

I measured the in-circuit resistances relative to the GND as well:
PINx: old PCBA / new PCBA
PIN1: 896kohm / 679kohm
PIN2: 717kohm / 481kohm
PIN3: 29.96kohm / 29.91kohm
PIN4: 732kohm / 474kohm
PIN5: 0R / 0R
PIN6: >10Mohm / >10Mohm
PIN7:  >10Mohm / >10Mohm
PIN8: 4.94kohm / 4.94kohm
R321: 182kohm / 182kohm
R323: 25.04kohm / 25.13kohm
R322: 357.7ohm / 357.7ohm

The BR301, BR302, D301 has the same chip marking.

It is strange why the old PCBA has different in-circuit resistances on the pins of TPS2375.

Hi Robert,

Thanks for your updates. 

Sorry that I got busy recently.... I will check your test data and get back to you early next week.

Best regards,

Diang

Hey Diang, 

Thanks! Any help would be welcomed. Currently the design validation is blocked and we cannot proceed because of this issue.

Best Regards,

Robert

Hi Robert,

Thanks for your new tested waveforms. For your measurements, may I know if you used differential probes when you are referring different grounds like VSS and RTN. 

One thing I guess is there could be some open-circuit damage on the PCB caused by a short-circuit since the 30V diode may not sustain 57V PSE voltage. It is less likely since PSE has current limitation too, but I could not think about other reasons that PD could not be turned on when its VDD-VSS is high enough.

Could you measure the resistance on the old and new PCB to see if there is any short circuit on these two locations?

What is that ripple after the classification, in the middle of the 53V on the RTN-VSS measurement?

The output of the D302 + 10uF + 4.7kohm load voltage is low, which could be some coupled voltage though the off hotswap. The RTN-VSS voltage shown as the CH4, I am also not sure why there is a time of oscillation. It might be PSE is doing some measurement on this port.

By the way, is the time 500ms/div or 100ms/div? If it is 500ms/div, looks like the PSE turns off the port due to the absence of MPS (300 - 400ms delay time in normal operation) since the cases like short circuit should react faster. 

Why do I measure 10ohm on the old board?

In TPS2375 Pin7 is NC. But there could still be diodes for ESD protection at Pin7 (it may refer to VSS or RTN). So it is possible that you could measure a low impedance when you forward bias the internal ESD diode. You may see different resistance values when you use the +/- probes of the voltmeter. 

Best regards,

Diang

Hi Robert,

Glad that you found the issue!

I will close this thread for now and please reply or open a new thread if you have further questions.

Best regards,

Diang