Need help with understanding the biasing of a class AB output stage

[Plecto]

Hi. I've been messing around with class AB output stages and their biasing. The first schematic I looked at was this one:

**broken link removed**

I thought it was a great way to bias a class AB stage, but when looking at the output current of the op-amp (which I use as a input stage) as well as the output voltage capabilities, I wasn't as happy anymore. To increase the bias current, R1 and R2 have to be lower in value, but lowering them too much will quickly exceed the current capabilities of the op-amp. By having a low bias current and high R1 and R2 values, the output peak voltage will be lowered. I made this formula to calculate output peak voltage vs. R1 and R2 values: VRL=(Hfe*RL(10Vs-7))/(10(Hfe*RL+Rb)), where VRL is the load voltage, RL is the load resistance and Rb is the biasing resistor values. Using darlington pair won't help at all if my calculations are correct.

I thought about using mosfets instead (despite the negative opinions people have towards fets in audio applications) and this will remove the issue as no current flows into the gate. This means that the bias current can be set with high value resistors without limiting the output voltage, but I would instead have to deal with the threshold voltage (which is often around 4V perhaps?) which will limit the output voltage instead.

I then thought of replacing the diodes with resistors, doing this seems to be better solution, but I'm not sure if there's any drawbacks of using resistors instead of diodes with biasing a class AB stage that I'm not seeing?

I'm wondering if there are better ways of doing this. I would really like to have as high efficiency as possible without having to resort to more powerful input stages

 

[erikl]

If you want to increase the output cross current, try inserting a resistor (potentiometer) in series (between) the 2 diodes. The diodes intend to stabilize the class AB stage's cross current against temperature changes, if they are in temperature contact with their respective transistors.

 

[FvM]

I don't think that the circuit is a good candidate for a practical class AB output stage. You e.g. want a DC coupling to control the output bias point. But some of the discussed problems also apply to a standard class AB design, e.g. the trade-off between R1/R2 value, driver current and transistor current gain. The usual solution for higher output current is a complementary or pseudo-complementary darlington output stage to achieve higher current gain.

My general suggestion is to study circuits of real amplifiers, calculate their parameters and try to understand why they have been designed as is.

To refer to a few of the addressed detail questions:

MOSFETs are an option. They lead to partly different designs. Most amplifiers that have been designed in the last 20 or 30 years have probably MOSFET output stages, the high time of bipolar output stages was in the decades before. But you can make good audio amplifiers in both technologies and now we are talking about bipolar.

The biasing diodes are a standard element found in most bipolar class AB, including some variant like an "adjustable" bias diode formed by a BJT and a potentiometer. The diodes offer bias voltage stabilization and temperature compensation. Usually they are combined with output transistor emitter resistors as feedback means.

The current delivered by the biasing resistors R1/R2 varies with amplifier output signal, in many designs they are replaced by active current sources, or a resistor is combined with a capacitive bootstrap circuit to keep the voltage across the resistor constant.

 

[E-design]

Another option you may consider is to take an op-amp with a suitable type output stage and add a current booster. This works well for low to medium output power applications. The op-amp must have a fast enough slew rate to accommodate the required voltage swing and needed frequency response.

 

[BradtheRad]

I thought about using mosfets instead (despite the negative opinions people have towards fets in audio applications)

I saw an expert state that mosfets are okay in audio amplifiers, as long as the power supply is tightly regulated. It may have been Rod Elliot's website. (Take a look at the mosfet amplifiers, a screen or two down the list.)

**broken link removed**

I then thought of replacing the diodes with resistors, doing this seems to be better solution, but I'm not sure if there's any drawbacks of using resistors instead of diodes with biasing a class AB stage that I'm not seeing?

In essence, you add some amount of DC bias to your AC signal. The idea is to apply just sufficient DC current to eliminate crossover distortion.

To make things simple I used a bipolar power supply.



Your input signal needs to overcome 0.65 V threshold in the base, thus it needs to be about 1.3 V greater amplitude than your desired output amplitude.

With careful adjustment of the pots, you should be able to obtain satisfactory operation, where the transistors are turned on long enough to minimize crossover distortion, yet without too much wasted power. At that point the transistor bases differ by 1.3 V at all times in the cycle. Then the diodes appear to be unnecessary.

Variations might include:

* a single polarity supply

* putting the NPN transistor at the lower side