Self balancing multiple MOSFET output transistors.

[nigelwright7557]

I have read a few times on here about matching MOSFET's for their gate switch on voltages.

My amplifier uses 3 off IRFP240 and 3 off IRFP9240.

I have built a couple of MOSFET amplifiers for disco and guitar use and never really had a problem despite being driven very hard at times.

I put a bit of thought into why this was.

At idle it is possible to get one MOSFET pair taking all the current.

However at typical running voltages the MOSFET's balance out due to the higher voltage required to drive higher current through the MOSFET's. This means the MOSFET with the lower switch on voltage will still conduct current but not quite as much.

Unless the switch on voltages are wildly different then I cants see a problem.

[CBS240]

One of the reasons we use source resistors.

[Mr Evil]

Look at the IRFP240 datasheet. The graph of Id vs Vgs shows the problem with current sharing. Note that current at a given voltage increases with temperature, which means that the FET that has the highest current heats up the most, drawing more current, heating up more etc. leading to thermal runaway.

And yes, Vgs can be significantly different between devices even at the same temperature. They are often very close if they come from the same batch, but you can't bet on that happening. You don't have to match them by hand though; that's what source resistors are for.

[nigelwright7557]

Originally posted by Mr Evil
Look at the IRFP240 datasheet. The graph of Id vs Vgs shows the problem with current sharing. Note that current at a given voltage increases with temperature, which means that the FET that has the highest current heats up the most, drawing more current, heating up more etc. leading to thermal runaway.

And yes, Vgs can be significantly different between devices even at the same temperature. They are often very close if they come from the same batch, but you can't bet on that happening. You don't have to match them by hand though; that's what source resistors are for.

I use 0R22 source resistors.
I usally find at idle the voltage across the source resistors varies by quite a bit for each output transistor.

I set the bias for minimum croosover distortion.

I found it is quite important to have the output DC offset as close to zero as possible too to save wasted power.