Hello, I have a PD design for a 5.0V isloated (flyback) power converter, primarily cribbed from the Evaluation board HPA108. I have inserted an aux power source as per option 3. The aux source is approximately 5.0V, but will be lower than the 5.0V output of the PD due to the ORing diode. Normally the system works properly, in that the PoE source is the preferred power source, even when the aux power is connected. The problem comes in when both sources are connected and the PoE network cable is removed. The system will quickly enter a continuous power-up/reset/power-down loop. My initial debug suggests that after the cable is removed, the residual charge on the V48 PoE input keeps the regulator operational (FET switching). As the V48 depletes, it reaches a point whereby the aux 5V power is higher than the PoE output voltage, and the switching will back-drive through the fly-back transformer and charge up the V48. The aux input doesn't have enough power to back-drive fully, so it will cut out, then the PoE will kick back in, but will deplete the V48 again. When the aux recovers, the cycle will continue. Would somebody be able to explain the cause and possibly suggest a remedy?
I am not sure the best way to address this issue. One option I considered is to attempt to prevent back-drive by putting a series blocking diode in the primary side of the transformer, but this had disastrous results in that something was damaged on the board, although I haven't diagnosed which part yet. I was also wondering if I could put an alternate UVLO circuit on the primary side, to shut down the switching if V48 drops below a pre-set value. I understand the device already has UVLO, but it appears to not kick in soon enough. If I add an additional UVLO, could someone suggest the best way to add, without in preventing startup under normal conditions?
I look forward to your reply.