I have built a few flyback drivers from 555 controlled to the ZVS. I always have one problem, the power transistors keep burning out. For the ZVS driver, I used the IRFP250 and keep having the problem of transistor failure over about 24 volts for the input. Here's a schematic of the ZVS.
I recently ordered some FCA47N60F 600 volt 47 amp MOSFETs on sale at electronic goldmine. Would these transistors work with the circuit, or are there specific transistor properties needed for the circuit?
You haven't provided any measurements taken on the circuit with the original parts, so it's pure speculation why the old ones failed. Excessive power? Bad gate drive? Avalanche? Excessive dv/dt? If we knew what killed the old FETs, it makes choosing a replacement much less based on guesswork.
On paper, The Fairchild parts are "better" in several key areas: \$V_{DS}\$, \$R_{DS(on)}\$ and avalanche capability. However, the total gate charge is higher, meaning the device will switch slower than the original part.
In my experience, it's very difficult to judge solely based on paper. I've seen MOSFETs which "look" better in all areas end up burning more way power than the originals, or fail catastrophically. Since several parameters appear better, try the substitution and see if your reliability improves.
I don't know much about this type of circuit but it strikes me that there will be problems at high powers and high frequencies. For a start for the FETs to turn on the gate capacitance has to be charged quickly - this makes a switching power supply much more efficient. I've taken the liberty of dismantling your circuit to make it more readable: -
The problem are two two 470R resistors circkled in red - if gate capacitance is (say) 3nF (not unheard of by any means), the CR charging time for the gate will be: -
\$470\times 3\times 10^{-9}\$ seconds = 1.41 micro seconds. That's just 1x CR time and I suspect your circuit will take 3 times as long to turn fully on - maybe up to 5 us. Turning off is similar - it relies on the other FET turning on to kill the gate voltage and so this is going to be about the same. All this wastes a lot of power and heat. But it's worse than that because the signal on the drains I guess is going to be more resonant than switching due to L1 and the bank of capacitors.
There is also the \$\dfrac{dV}{dt}\$ voltage change on the drain affecting the gate voltage. Because the gate voltage is not powerfully driven, as the voltage on the drain starts to fall, it will, by internal drain-gate capacitance inject a signal into the gate that stops it turning on as quickly would the gate be driven by a proper drive circuit.
Like I say, I'm no expert on this type of design but it strikes me this type of switcher is good because it is simple but also inefficient also because it is too simple.
