TPS2373: PoE PD with OnSemi FDMQ8205A

Hi Diang,

Does the DC-DC part of the TPS2373 influence the identification part between the PD and the PoE Switch? Just wondering how the TPS2372 will simplify my design.

Thanks for your quick answer!

Jeremy

Is the circuit you explained mandatory for the TPS2373 or it could work without it? My question is, could it be the reason my board doesn't work?

Thanks,

Jeremy

Hi Jeremy,

Thanks for your quick reply. 

TPS2373 will have start-up failure if its VC_IN and VC_OUT do not have the circuit I talked in last post. It can cause your board failure. 

Best regards,

Diang

Hi Diang,

I've received and assembled my TPS2372 PCB. I've followed the schematic shown in the eval board datasheet but still can't seem to make it work. Do I need to put any of the AUTCLS, MPS_Duty, IRSHDL_EN or AMPS_connect jumpers?

Hi Jeremy,

Thanks for your updates. You can refer the below image and red cycled parts can be jumped. Let me know if it can work or not. Thanks!  

Best regards,

Diang

Hi Diang,

Thanks for your reply, it doesn't seem to work.. Do you have any test I can perform so that you can diagnose what's going on?

Best regards

Hi Jeremy,

Could you share your schematic?

For the testing, it would be great if you can measure the Ethernet interface current using a current gun and RTN-VSS voltage. It will help to see if there are any abnormal during detection/classification/port on. Below is an example. 

Best regards,

Diang 

Hi Diang,

We made three different boards to test them separately and are testing them on a breadboard, here's the schematic for the part with the TPS2372

Heres the ethernet input and the FDMQ8205A  schematic

Hi Jeremy,

Thanks for your schematic. The FDMQ8205A page looks not very clear but other schematics look good. 

Can you double check the jumpers are shown like below? And RV1 could be reduced to minimum resistance (0 Ohm) if possible.  

Best regards,

Diang

Hello Diang,

I have set RV1 to 0 ohm and probed VSS and RTN with my oscilloscope. It seems like i'm not getting anything. Any other things I should check?

Hi Jeremy,

This looks like a sign that the PSE - PD interface is not conductive. 

Is it possible to also check voltage between PR12 - PR36, and PR45 - PR78? If also no signal could be observed, PSE - PD does not create a pass way and could not do handshake. Then you may need to double check the Ethernet wire / RJ45 port connections. 

Best regards,

Diang

Hi Diang,

So even if the handhshake hasn't happened yet, there should be voltage between PR12 and PR36 and also between PR45 and PR78?

Hi Diang,

When connecting my ethernet to the board I get these waves:
 This is pair 45 and 78 at 2.16v pk2pk. 500mv/div and 20ms/div

And this is pair 12 and 36:
 at 1.14v pk2pk. 500mv/div and 20ms/div

The pulses also come out of my FDMQ8205A board at 800mV pk2pk, same oscilloscope settings as previous pictures

Hi Jeremy,

Thanks for your waveform and it looks like the PSE is trying to send signal to PD for handshake (if it is not a noise coupling).

While the magnitude of voltage is not in a valid detection range (2.7V - 10.1V). This could be not wrong since some PSE may use current source for detection try instead of voltage source. 

Strange that the R_detection and C_detection are both fit the requirement in standard. 

Can you send a FDMQ8205A board with higher resolution and are you able to read the registers of your PSE to check what is the cause of detection failure? 

Best regards,

Diang

Hello Diang,

Here's a better picture of the FDMQ8205 circuit:

The PSE doesn't seem to have logs or alarms on the cause of the detection failure.

Hi Jeremy,

Thanks for your new rectifier circuit. I will take a look and get back to you by this week.

Best regards,

Diang

Hi Jeremy,

Some PSE can show the detection R = low or R = High / Open-circuit.... This will help use zoom in the potential issue.

By the way, for the rectifier can you double check the pins number are matched as below? 

And could you use a muti-meter diode mode to measure if the VSS -> VDD is diode forwarding? I tested one PD board on my side, VSS->VDD voltage drop is ~0.6V, while VDD->VSS is non-forwarding. 

Best regards,

Diang

Hello Diang,

We've finally made it work. Our problem was that we had the wrong ethernet connectors so we were missing 2 pairs out of the 4 pairs needed for poe. There's one last thing that doesnt work and its that my PD can only identify up to 51W. I'm using the 3 LEDs on TPH, TPL and BT and the datasheet to know the max power my PD can get. Do you have any idea what could be the cause?

Best regards,

Jeremy

Hi Jeremy,

Glad that you made it work! CAT5 or higher level cable is needed for PoE usually.

TPH/TPL/BT can tell the PSE type. Type-3 is up to 51W, and Type-4 at power demotion mode (limited to Type-3 for example) could also be 51W. Could you let me know what is your PSE rating and is it BT Type-4?  

And did you use below combination to have 71.3W class-8 setting?

Best regards,

Diang

Hi Diang,

So i'm not sure which table to check but, with my jumpers positioned like you asked, the TPH led and the BT led light up. TPL stays off. So I'd say it's LOW HIGH LOW?

Best regards

Hi Jeremy,

Thanks for your reply. Yes it is TPH/TPL/BT = L/H/L, which means only Class-6 or -5 power is allocated. 

Could you share your PSE information? In this case, your PSE could be only Type-3 rated or in a power demotion mode.

Best regards,

Diang

Hi Diang,

our switch is the  UTP5328S-PSD2000. I aready confirmed with Utepo that the switch i have is type-4 rated.

Best regards

Hi Jeremy,

Can you measure your VDD-VSS voltage during classification?

You should see 5 fingers like below for Type-4 handshake (TPH/TPL/BT = L/L/L). 

If you see 4 fingers that means Type-3 handshake (TPH/TPL/BT = L/H/L). Either the PD CLB resistance is not 90.9 Ohm, or the PSE has limited power allocation to 60W.

Best regards,

Diang