generating VIN for LTC4269-2

Hi Adam,

This method is the simplest and cheapest way to provide an 18V VIN supply from a 37V – 57V PoE input, while also drawing low power from the system supply and improving short circuit performance.

The resistor RSTART limits the current into VIN, so that the CSTART capacitor can only charge when the IC is not switching (IC is OFF and VIN is less than VIN ON). The capacitor provides bias current while converter starts up (IC is ON and VIN is greater than VIN OFF). An 18V Zener also required to protect the VIN pin from PoE input voltages. Once VOUT is near regulation, a bias winding provides VIN. During short circuit, the bias winding collapses and the IC shuts down and restarts only after the VIN pin charges back up to VIN ON.

Using another regulator to provide VIN would be more expensive, less efficient, and may limit short circuit robustness.

Best Regards,

Eric

Hi Eric,

thank you for the explanation. That performance under short-circuit is in fact a good thing. I am concerned with minimal output current necessary to keep the device running, in case when somebody would like to put everything to sleep. But knowing that I will be able to figure something out, if necessary.

Best regards,
Adam

Hi Adam,

Yes, this behavior helps during short circuit.

Please be aware, a PD is required to provide a valid Maintain Power Signature (MPS) for the port to remain powered. If PD does not draw enough current, the PSE may decide the cable has been disconnected and stop powering the port. MPS current for an IEEE 802.3at PD is 10mA, for a minimum of 75ms every 325ms period. It is a good practice to make sure your design draws 10mA DC from the PoE input during normal operation.

Best Regards,

Eric

Thanks, Eric, that's good to know.

That looks like a terrible waste of energy, though. I wonder if they're already developing a standard in which PSE just asks PD - similarly to what's happening in Qi (wireless charging), or QuickCharge (yeah, that's a worse example).

Best regards,

Adam