Transistor dead again

Is the sketch right?. Seems it powers up and the bulb must come ON. The 4meg and "T" in the right place?.

is

I am not sure if the circuit is correct, but assuming it is, then Q1 is not likely to take a hit, but Q2 is - if that voltmeter connection were to experience a momentary short circuit.

Oops. There are mistake here. Q1's emitter goes to ground and the 4.7M is connected to the collector instead. It's based on the third circuit here:

The voltmeter is actually a voltage comparator, but I ran into problems just connecting a 10k in its place. Initially the FET is off and thus minimal current goes through the lamp. Briefly toggling Q1 then raises the gate voltage and turns the light on. At least that's what it did yesterday but not any more..

It's a nice circuit arrangement and will work fine. There's no way that Q1 can be blown by overvoltage or overcurrent. Sounds like there's a wire shorting somewhere.

Could it be happing when you remove the power? It seems that if its working while the power is on and if theres no power then it can't fail(except maybe caps but doubtful) then it must happen when the power is changed.

You might try running it with the power on for a while and put it through some stress test and see what happens. If it lasts for a long time then try turning the power off for a split second and then put it through another test for a few mins. Repeat this a few times. Then turn it off for a longer period of time and try. You'll be able to determine which stage the problem is comming from.

Never mind. It wasn't the transistor that was dead (again) but a partially broken trace causing the problem.