I have been developing on the PIC-Micro-Web for some time using the POE breakout lead connected to a 30V bench PSU and all has been well. In preparation for deployment I thought I'd better test at 48V as this is most likely encountered in the field. Within seconds the board failed and on closer inspection I see that Z1 has failed and most likely the BD901F regulator chip, a very difficult part to locate here in the UK. The data sheet says up to 50V is OK so why has this failed so quickly? I want to use these on 48V to minimise the current drawn down the wires but now I'm concerned that they will all blow. Looking at the BD901F and 1N4731 datasheets I see no problem with this. Any ideas? Any good supplier of BD901F chips?
MICRO-WEB should take 54V before failing - 48 are handled by the voltage regulator (the BD9001F), the rest by the zener - take this in account when replacing the parts. Try putting bigger zener diode when replacing the old one. 4735A or 4736A would do. It is hard to burn this zener with overvoltage so check if your input doesn't exceed 800mA.
Here I found the BD9001F for you: http://www.digikey.co.uk/product-detail/en/BD9001F-E2/BD9001F-E2DKR-ND/1950390 (http://www.digikey.co.uk/product-detail/en/BD9001F-E2/BD9001F-E2DKR-ND/1950390).
PIC-MICRO-WEB shouldn't burn at 50V. Let us know if it happens again.