Re: more X-rays Silvertel PoE Supply

On 2024-10-02 22:33, Edward Rawde wrote:

"Arie de Muijnck" <noreply@ademu.nl> wrote in message news:nnd$7f138845$4f992b3a@6f6ac280155cacc4...

On 2024-10-02 20:02, Edward Rawde wrote:

"Don Y" <blockedofcourse@foo.invalid> wrote in message news:vdjvmh$3aipj$1@dont-email.me...

On 10/2/2024 8:46 AM, Edward Rawde wrote:

I did PoE with LTC4267 some years ago such that 12V could also be supplied from a wart.
Fully isolated with a PA1138 transformer.

>
How did you inform the PD of the fact that it was no longer
dependant on (and controllable by) the PSE?  Or, did you forego
in-band reconfiguration and power control?

>
It wasn't necessary for the fairly specialized product which was configured at install time with power from PoE or wart but not
both.
I seem to recall at least one accidental connection of both with no detrimental effect but it wasn't investigated further.

>
>
I've seen a problem when the PoE module had a TL431 + optocoupler at it's output as feedback circuit.
A limiting resistor was lacking, and +12.3V from a wart turned the TL431 fully on and destroyed the optocoupler.
Subsequent PoE operation then had an open loop and destroyed the supplied circuit.

 A seperate smaller schottky and capacitor for just the TLV431 seems to work fine in the LTC4267 simulation I just did and works
around wall wart power getting connected to the TLV431.
 To disconnect wart power when PoE is operating, put a low power voltage doubler on the transformer secondary and use the 24V to
swich off a p-fet in series with wart power.
 

>
Arie
>

Way too complex. A series resistor between TL431 and optocoupler was all that was needed.
Arie