When using a complementary pair of MOSFETS such as IRFP240/IRFP9240 for a class AB output stage (say, instead of complementary BJTs), the bias between the gates of the MOSFETs has to be roughly the sum of the gate threshold voltages, as opposed to the sum of the diode drops across the base-emitter junctions of the BJTs. I was playing around in a simulator where I swapped out the BJTs for the IRFP240/IRFP9240 and I am using a Vbe multiplier to set the bias point between the two gates. I noticed that the quiescent current in the MOSFETS, which needs to be in a narrow range as to not cause distortion if it is too low, or dissipate a lot of power if it is too high, is very sensitive to the bias value, and I was wondering how in a practical complementary MOSFET class AB amplifier the bias point is maintained so that the quiescent current is a few tends of milliamps, because this sensitivity might make the quiescent current not stable when the temperature changes, for example. I looked at some other examples of class AB mosfet amplifiers and they seem to use circuits that look like Vbe multipliers to be and so I suppose it could be made to work?
Attachments
Hi .You may add a diode in series with vbe multiplier ,also you will get some thermal compensation. I'm using in this role TL431 with voltage divider ,voltage is very stable then .
FWIW I use IRFP250 or 240 in my quasi complementary 300W Bass Guitar and PA amplifiers, bias them with a plain vanilla Vbe multiplier and never had a problem with that.
My thermal sensor is a humble BC55x epoxied to heat sink.
Not a simulation but hundreds of them used in the real world under grueling live stage conditions without a fault.
My thermal sensor is a humble BC55x epoxied to heat sink.
Not a simulation but hundreds of them used in the real world under grueling live stage conditions without a fault.