• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 62 subscribers
  • Views 1148 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS23753A: MOSFET not connecting RTN to VSS after detection/classification

Sebastian Floss
Prodigy 50 points
Part Number: TPS23753A
Other Parts Discussed in Thread: TPS23753

Hi,

I have a PoE PD setup based on the power-part of this reference design: https://www.ti.com/tool/TIDM-TM4C129POE.

When I connect an external 12V adapter, everything works perfectly fine. A cheap PoE midspan power supply also works well. Only when I attach the device to a PoE Switch (lancom G2326p+) I see the following behaviour:

The detection and classification works and the PSE supplies 53V (measured between VDD and VSS). 
Unfortunately the TPS doesn't switch open the MOSFET to connect RTN to VSS. after a couple seconds the DC/DC converter gets very hot (I assume because there's no feedback from the secondary side and the duty cycle approaches 100%). Some seconds later the either the overheat-protection kicks in or the switch detaches since there's no valid MPS.

The only difference I could measure between the midspan power supply and the switch so far, is that with the former I get 48V between VDD and VSS and with the latter 53V.

Any suggestions would be greatly appreciated!
P.S.: I triple checked the schematic, layout and parts.

Thanks in advance, Sebastian

  • 0
    Prodigy 50 points
    One thing I forgot to mention: When connecting to the midspan supply, sometimes the classification occurs multiple times (3-4) in a row before the PSE delivers the 48V - dunno if that helps in narrowing the issue down.
  • 0 in reply to Sebastian Floss
    TI__Genius 9230 points

    Hi Sebastian,

    By classification occuring multiple times, are you referring to individual fingers or multiple sets of fingers?

    Also, what kind of PSE are you using?

    Lastly, how are you ORing the 12V adapter into your circuit? See ORing options here: http://www.ti.com/lit/an/slva306a/slva306a.pdf

    I'll probably have more questions as I think of them.

    Thanks,
    Tom A.

  • 0 in reply to Thomas Amlee
    Prodigy 50 points

    Hi Tom,

    thanks for the quick reply!

    First of all: sorry for assuming the PSE did a "multi-classification" - turned out, I was wrong. 
    What happens with the cheap midspan PSE is this (yellow is VDD/VSS, cyan is the secondary side of the transformer):


    To answer your other questions:

    - The adapter is ORed with adapter preference over the APD pin (Option 2, Chapter 3.3 from the doc you linked).
    - The midspan PSE is a "PSE151", a cheap 48V/16W device from Taiwan.
    - The endspan switch is a Lancom Systems GS-2326p+ 

    I see if I can post an oscilloscope-screenshot of the PD attaching to the switch. 

    Sebastian 

  • 0 in reply to Sebastian Floss
    Prodigy 50 points

    Okay, so I solved my problem, I guess.

    Look at the screenshot of the PD with the switch as PSE: notice the negative voltage?


    Now here's the thing:

    I had to invert the signal, because I didn't have a differential probe at hand so I ref'd the scope's GND (PE) to VDD while probing VSS. The other way 'round it wouldn't work.

    Looking a little closer I saw, the RJ45's shield on my prototype was accidentally (I know I said I triple checked - well, sometimes it takes a fourth time, I guess ;) )
    connected to the TPS23753's RTN ground plane. 

    Since the switch has PE connected to its RJ45's shield (as it's supposed to), when plugging in a shielded CAT5 patch cable RTN got pulled to mains earth.
    The midspan supply didn't have a PE connector (I said it was cheap), so the problem couldn't occur.

    Now, to put an end to this:
    can you help me understand, why it is a problem when VSS (or RTN) gets pulled to mains earth (PE), but not VDD (as I did during probing with the scope)?

    Thanks in advance, Sebastian 

  • 0 in reply to Sebastian Floss
    Prodigy 50 points
    Marked it as resolved, since I found the flaw myself ;)