Class A vs Class AB design

[ervinl]

Hi, I'm still confused how to identify the circuit of Power Amp is pure Class A or just Class AB (A if driven with lower wattage).

I know that SE may become class A, but how about regular PP (complementary BJT or MOSFET) and Quasi Complement?

Could we just increase the bias with lower voltage supply to reach Pure Class A from any Power Amp Circuit (so that BJT or MOSFET will be "on" in idle/without input, or at its peak dissipation)?

Can Quasi Complement work into pure Class A?
Can Complement BJT work in pure Class A (how much Iq in basis and/or output)?
Can Complement MOSFET work in pure Class A?

Thanks,

Ervin L

 

[wahab]

Could we just increase the bias with lower voltage supply to reach Pure Class A from any Power Amp Circuit (so that BJT or MOSFET will be "on" in idle/without input, or at its peak dissipation)?

Can Quasi Complement work into pure Class A?
Can Complement BJT work in pure Class A (how much Iq in basis and/or output)?
Can Complement MOSFET work in pure Class A?

Thanks,

Ervin L

Yes for all the questions above..

 

[event horizon]

Hi Ervin, as above basically 🙂

If an amplifier has a single ended output stage then it'll definately be class A. I understand there are some fancy systems to increase the standing current when it needs more to drive a speaker, however these are few & far between.

The only real difference between a class A & class AB amplifier is the quiescent current accross the output stage, this might not be evident by simply looking at a circuit diagram. What you need to do is measure the voltage accross the emitter resistors on the output stage to get an idea of the current flowing.

Massive heatsinking & a lowish supply voltage may well suggest a class A amp, there will also likely be quite a few more output transistors on a class A amp than virtually any class AB.

Any kind of complimentary or quasi complimentary output stage could be run class A as long as there is enough heatsinking & dissipation in the devices to cope with the power, be it bipolar or mosfet. Interestingly the amplifier (assuming full class A operation) will run cooler when driving a loudspeaker at a decent volume as power that is dissipated solely by the output stage when it was idling is now fed to the speaker 😀

 

[ervinl]

On complementary, I know that to start working in Class A, the BJT or MOSFET must be turned on at idle, so that if BJT, VBe is set to around 0.6-0.7 volt (for MOSFET, set to On threshold/VGS thershold), so that at least 2 dioda (or equivalent transistor) is used to separate the voltage. But what about quasi, since the other is to separate V BC, not directly V BE (depend on power follower or power source design). Is it set so that VBe in both transistor or MOSFET is near to On threshold (that is somewhat difficult in quasi, unless it was driven by separate/complex circuit)?

In Class A, what is the minimum Iq in each of final BJT/MOSFET to setup?
What is the Class A reference, Iq only, or also V On threshold (BJT, MOSFET)?

Thanks,

Ervin L

 

[kenpeter]

In pure linear class A, the sum of both currents is a constant.
Iq is exactly one half the maximum current for either device.

There are many non-linear subclasses between A and B that
do not sum to a constant. And class AB does not adequately
describe some of them. Those that never turn completely off
for example, cannot be "B". So what do you call them?

 

[event horizon]

In Class A, what is the minimum Iq in each of final BJT/MOSFET to setup?
What is the Class A reference, Iq only, or also V On threshold (BJT, MOSFET)?

Ok, here is an example. Lets assume we have a fairly standard complementary push pull output stage with a driver & output transistor. I'll not go into Vas & front ends.

Lets assume the amplifier is going to be designed to do 10W RMS into 8 ohms in class A.

To drive 10W into 8 ohm the amplifier will be required to swing +/- 12.64V As there will be voltage drops of about 0.7V on each output transistor & driver & then some on the emitter resistor we'll need a power supply of about +/- 17V allowing for ripple in the PSU capacitors.

Now as speakers don't stick at exactly the same impedance i'd suggest that you set the quiescent current so as to do full class A into a 6 ohm load. The RMS output voltage of the amp will be 8.94V RMS, the peak voltage will be 12.64V as mentioned before. Now divide that by the 6 ohms & you get 2.11, divide it in two again & you have the value to set the current in a push pull stage - 1.06Amps!

Lets assume you use 0.47 ohm emitter resistors, you'll need to set the voltage accross both resistors to exactly 1V. If you used 0.33 ohm then you'd set the voltage to be 0.7V

The total dissipation of the amplifier will be 1.06Amps x 17V x 2 or 36.04W



In answer to your question it'd depend entirely on how many transistors are paralleled as to how much current each one passes, it also depends on how powerful the amp is. You can already see that just a 10W amp dissipates a lot of power - a big amp will need huge heatsinks & loads of transistors 😀

The reference is the quiescent current, IE the voltage accross the emitter resistors. You don't need to worry about what the base emitter voltage or gate voltage is - just the voltage (which gives you current) accross the emitter resistors.

 

[megasat16]

Hi, I'm still confused how to identify the circuit of Power Amp is pure Class A or just Class AB (A if driven with lower wattage).

I know that SE may become class A, but how about regular PP (complementary BJT or MOSFET) and Quasi Complement?

Could we just increase the bias with lower voltage supply to reach Pure Class A from any Power Amp Circuit (so that BJT or MOSFET will be "on" in idle/without input, or at its peak dissipation)?

Can Quasi Complement work into pure Class A?
Can Complement BJT work in pure Class A (how much Iq in basis and/or output)?
Can Complement MOSFET work in pure Class A?

Thanks,

Ervin L

It eludes me why any other topology other than NPN / N-Channel based designs can be called PURE Class A! To me, PURE Class is always been Single Ended in nature (ok, Balanced input can be used but not fully balanced from input to output for PURE CLASS A) and the entire positive and negative waveform conduction happens in the NPN or N-Channel Type Device.

There are many other Class A variations that can use all topology of what you asked but I think it's fitting to call them Push-Pull Class A, Class AA, Super A, Quasi-Class A, Sliding Bias Class A (Krell and Levinson), Hyperbolic Class A (HCA from Yamaha). While I don't have experience with all of these, I believe many of these Class A variation can be used in Full Balanced (from input to output).

 

[event horizon]

It eludes me why any other topology other than NPN / N-Channel based designs can be called PURE Class A! To me, PURE Class is always been Single Ended in nature (ok, Balanced input can be used but not fully balanced from input to output for PURE CLASS A) and the entire positive and negative waveform conduction happens in the NPN or N-Channel Type Device.

I must admit i prefer the sound of single ended class A power amplifiers, but they are just sooo damn in-efficient. As long as none of the output transistors turns off then a push pull amp will be operating in class A & the second previso is that half the peak output current is accross the emitter resistors.

As i mentioned, set the quiescent for a slightly lower speaker impedance & as you can see with peak current flowing through one transistor the other still won't turn off 😉 Class A is simply about getting rid of crossover distortion & hopefully biasing the output transistors into there linear regions to further reduce distortion. As i mentioned in another thread on here there is a difference in the way push pull & single ended amps produce distortion products. Push pull produce mainly odd order distortion, this may explain why i prefer single ended as they tend to produce mostly even order which is far more acceptable 😀