There can be almost the full input DC voltage on T1
, but not on the PNP
current limit transistor (or the NPN
, for that matter).
I found that the simulation was unstable (high frequency oscillation) until I added some capacitors to speed up the feedback loop, as well as the capacitor on the op-amp. The emitter resistor R4
also helped, but it reduces the gain of the system. I did not simulate the PMOS
series pass transistor, and it is probably a lot faster than the Darlington 2N6287
for the MOSFET is largely based on the power dissipation. The FQP4P40
is rated 85 watts, while 350V at 500 mA is 175W. Looking closely at the chart, it is more like 250 mA at 350V, which is about 85 watts
. It can withstand higher current pulses of short duration. Your 1.5 ohm current sampling resistor limit is about 450 mA
, so a short circuit would cause 0.45*280 = 126 watts
A "cheap and dirty" way to limit short circuit current and power dissipation might be to add a power resistor of about 200 ohms
, which will drop 100 volts at 500 mA and dissipate 50 watts
, so the MOSFET would only see about 75 watts
. A thermistor on the heat sink could be used to shut down (or throttle) the output in the case of extended overload. A 10k thermistor across R1
might do the job.
I am considering a PIC-based power supply where it will monitor the current through the series pass element as well as the voltage, and limiting power dissipation accordingly. It could also monitor temperature of the heat sink. Perhaps a PIC16F1764
, which includes an op-amp as well as other useful peripherals.