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.

TPS23754: POE OR-ing

Part Number: TPS23754

I have a customer that would like to follow the example on page 13 of this app note (Advanced Adapter ORing Solutions using the TPS23753) to have two power supplies OR'ed together (PoE and DC/DC). 

  1. Is this the best approach?  
  2. What signal from the DC/DC would they use to disable the POE PS? (Per this statement: Applications using option 3 with adapter priority may want to disable the PSE source)
  3. Can a PGOOD flag from the adapter DC/DC be used to disable PSE source once the adapter is plugged in?

Regards,
Ryan B.

  • Hi Ryan,

    1. Typically Option 3 ORing is best when the adapter voltage is the same as the output voltage of the converter. For example, a PoE converter that outputs 12V will have a 12V adapter OR'd on the secondary to save on efficiency.
    2. Like shown on page 13, it uses the output voltage (through an opto) as the signal to know to remove the PSE. When DEN goes LOW, the PSE will remove power and correct re-detection such that it will not try to reconnect.
    3. Since the DEN pin is on the secondary, an opto will be needed. They can use a PGOOD flag but will still need a way to drive the opto.

    Thanks!
    Regards,
    Darwin
  • Hi Darwin,

    Great, thank you for that explanation.

    A few other questions have popped up: do I need an OR-ing diode on both the output of the adapter AND the PoE output?

    Looking at Figure 9 in the app note, I see D(A) on the adapter to keep the PoE voltage from appearing on the adapter. Do we need anything to block the adapter voltage from the PoE power stage?

    Also, can you further explain the need for an output blocking FET if using a synchronous converter? If using a synchronous gate drive like in Fig27 of the datasheet, wont the rectifying FET (M2) always be off when the adapter is present? The app notes states: "When used with a secondary-side synchronous rectifier, this circuit needs special attention to turn the rectifier off when the adapter takes over. This prevents power from flowing backwards and powering the primary side of the converter". I figured the GAT2 signal would turn M2 off once the adapter disables the chip via DEN.

    Regards,
    Ryan Bishop

  • Hi Darwin,
    To be clear, the "output blocking MOSFET" is only mentioned in Table 1 of the app note.
    Later in the app note, Section 3.5, it only mentions to take special care to turn the sync-rectifier off.
    Thank you,
    Ryan B.
  • Hi Ryan,

    If the converter is a synch solution (like synch flyback or synch active clamp forward), yes you need a blocking diode. If the converter is a simple diode rectified flyback, then you do not need an extra diode.

    Thanks!
  • Hi Darwin,

    Thanks again, I think this answers my question enough to continue advising the customer.

    However, for my own understanding, what current path are we worried about if a diode wasn't placed on the PoE output?

    When DEN is pulled low, does GAT2 not turn off the sync-rec FET? I figured with this FET off then maybe there's no conduction path to worry about.

    Maybe it takes too long for FET to fully turn off? Sorry for dragging this on...

  • Hi Ryan,

    That's not a problem! Actually you're question makes sense and to clarify, for Driven flyback converter you won't have this problem. A lot of designs use the self-driven flyback topology  which will have this issue.

    Thanks!