Building a Speaker with a Class AB Amplifier

[abro]

Hello,
For a school project I need to build a simple but working speaker. I am currently stuck on explaining how the signal amplifier works.
Here is a schematic of an AB amplifier from Wikipedia. I have add a positive and negative sign indicating the connection of the battery. It contains 2 transistors, an NPN and PNP. These react individually on the polarity of the input signal (AC). The amplified semi waves then combine and form an audiosignal.
I can only confirm that it works when the input signal is positive, then the PNP transistor blocks the current and all the current flows through the speaker. If its negative then the NPN transistor completely blocks the DC power supply.
Could you guys please explain how the rest works, would be really grateful.

 

[LickMyEyeball]

Hello there!

As far as I understand, you are correct that on negative input signal swings, the 'upper' BJT will not conduct, while the 'lower' BJT will conduct. Assuming that the current driving the load is not negligible, where is it coming from? (We will ignore the currents into the transistor base, which is usually relatively small.)

If the upper BJT's collector is not conducting, then current must be flowing from the load (connected in between the two transistor's emitters) down to the lower voltage marked with a negative symbol in your diagram. During negative signal swings, the energy required to drive the load comes from the potential difference that exists between the load and that lower voltage.

 

[vk6kro]

That is not intended to be a working amplifier. There are several important components missing and this has been done to simplify the circuit and to demonstrate how the circuit works.

Mark the bottom (-) line as 0 volts and the top (+) line as 12 volts.

Let the load be a speaker connected from where the two transistors meet to 6 volts.

Now you can see what happens as you vary the drive.

If the top transistor is conducting, current flows from the 12 volt line to the 6 volt one through the speaker.

If the bottom transistor is conducting, then the current flows from the 6 volt line through the speaker to the 0 volt line.

But don't build it like this. The input circuit must be properly biased to get undistorted output.

 

[abro]

May I ask why the speaker is 6 volts? Is there a 6 volt battery which provides the current or...

That is not intended to be a working amplifier. There are several important components missing and this has been done to simplify the circuit and to demonstrate how the circuit works.

Mark the bottom (-) line as 0 volts and the top (+) line as 12 volts.

Let the load be a speaker connected from where the two transistors meet to 6 volts.

Now you can see what happens as you vary the drive.

If the top transistor is conducting, current flows from the 12 volt line to the 6 volt one through the speaker.

If the bottom transistor is conducting, then the current flows from the 6 volt line through the speaker to the 0 volt line.

But don't build it like this. The input circuit must be properly biased to get undistorted output.

 

[vk6kro]

Yes, you could use a 6 volt supply like that. Or a 6 volt battery.
It is an example to show how the amplifier works.

In a real amplifier, you might use a capacitor in series with the speaker which has its other terminal connected to the negative line.

Try Google in image mode. Search for "audio amplifier". There are dozens of practical schematics there.
If one looks interesting, you can read about it on the website the schematic comes from.

 

[skeptic2]

Though Vk6kro's circuit is correct let me suggest an equivalent circuit to Vk6kro's that might be easier to understand. Instead of one rail at +12 V and the other at 0 V, let's make the voltages +6 V and -6 V. The speaker is connected between the junction of the two emitters and ground. Now with a positive input the current flows from the positive supply through the NPN transistor and through the speaker to ground. When the input is negative, the current flows from ground though the speaker through the PNP transistor to the negative supply.

 

[sophiecentaur]

The original diagram has been drawn too simply for it to make proper sense. Current will only flow through the transistors when there is a load connected to ground (mid point) so it should be included. It is easier to consider a simple resistive load and the (other terminals of the) two batteries must be explicitly connected to that same point (taken to be zero potential).
How it works:
The npn transistor conducts and 'emitter follows' the input signal when it goes positive and the pnp transistor conducts and follows the input signal when it's negative. This gives you amplification for both polarities of input signal.

The 'half sinusoid' waveforms shown represent the currents from each emitter - so they are not to scale with the input waveform (there will be current gain involved). But the overall input and output (complete) wave forms can be of the voltage variation and, because a common base amplifier has unity voltage gain, they will be to scale with each other. The output current will be the V/R where I is the output voltage and R is the load resistance. Power will be V²/R.
When the input signal is between + and - 0.7V, neither transistor will conduct and you will get 'crossover distortion' - but this is easy to avoid by using a slightly more complicated circuit to drive the bases.

"In a real amplifier, you might use a capacitor in series with the speaker" (vk6kro)
Where you have a DC output level at ground potential, it is not necessary to use AC coupling because large value capacitors are needed and they can be expensive. It is better to use a bit of feedback to ensure that the mean voltage at the output terminal is zero - so no net DC current will flow. It is only when you use a 'single ended' power supply that AC coupling is strictly necessary. You pays yer money and you takes your pick, as they say.

 

[abro]

In my drawing the ground would be the negative sign, right? But how can current flow through the NPN transistor when the input signal is negative? It will allow current to flow through the PNP transistor, but the NPN transistor blocks the entire voltage supply. Or is there something I don't know about transistors yet?

The original diagram has been drawn too simply for it to make proper sense. Current will only flow through the transistors when there is a load connected to ground (mid point) so it should be included. It is easier to consider a simple resistive load and the (other terminals of the) two batteries must be explicitly connected to that same point (taken to be zero potential).
How it works:
The npn transistor conducts and 'emitter follows' the input signal when it goes positive and the pnp transistor conducts and follows the input signal when it's negative. This gives you amplification for both polarities of input signal.

The 'half sinusoid' waveforms shown represent the currents from each emitter - so they are not to scale with the input waveform (there will be current gain involved). But the overall input and output (complete) wave forms can be of the voltage variation and, because a common base amplifier has unity voltage gain, they will be to scale with each other. The output current will be the V/R where I is the output voltage and R is the load resistance. Power will be V²/R.
When the input signal is between + and - 0.7V, neither transistor will conduct and you will get 'crossover distortion' - but this is easy to avoid by using a slightly more complicated circuit to drive the bases.

"In a real amplifier, you might use a capacitor in series with the speaker" (vk6kro)
Where you have a DC output level at ground potential, it is not necessary to use AC coupling because large value capacitors are needed and they can be expensive. It is better to use a bit of feedback to ensure that the mean voltage at the output terminal is zero - so no net DC current will flow. It is only when you use a 'single ended' power supply that AC coupling is strictly necessary. You pays yer money and you takes your pick, as they say.

 

[vk6kro]

For a simple amplifier, a dual power supply probably adds too much complication, even if you need a large electro go with a single supply.

It is universal now to use a dual supply for high powered amplifiers.
This adds extra cost and a need to have perfect balance to get zero DC current in the speaker.

 

[Jony130]

Try to read this
http://forum.allaboutcircuits.com/threads/compensating-diodes-used-on-push-pull-amplifier.101749/

 

[vk6kro]

Current never flows through both transistors at the same time.

It flows through the speaker in opposite directions depending on which transistor is conducting.

 

[abro]

Back to your 2 power supply amplfier, is this what you mean't? It should work, I used Kirrchoffs rules and it should...

Though Vk6kro's circuit is correct let me suggest an equivalent circuit to Vk6kro's that might be easier to understand. Instead of one rail at +12 V and the other at 0 V, let's make the voltages +6 V and -6 V. The speaker is connected between the junction of the two emitters and ground. Now with a positive input the current flows from the positive supply through the NPN transistor and through the speaker to ground. When the input is negative, the current flows from ground though the speaker through the PNP transistor to the negative supply.

 

[sophiecentaur]

In my drawing the ground would be the negative sign, right? But how can current flow through the NPN transistor when the input signal is negative? It will allow current to flow through the PNP transistor, but the NPN transistor blocks the entire voltage supply. Or is there something I don't know about transistors yet?

Not at all. The push pull amplifier is a sophisticated circuit and it is easiest to describe, initially, by connecting it in a circuit with a positive supply and a negative supply, with a 'common' ground. (It's one step up from the simple transistor amplifier circuit that we all start with ) That allows you to have your load connected to ground and everything is symmetrical about that ground. If you choose not do do it that way, you need to have your input signal 'biased' so that its mean value is 'half way' between two connections of a single battery and you also need to connect your load to a half voltage point. This is very often done in practice but this is much harder to understand for the beginner (well - it confused you! ;))

It would be a lot of trouble for me to draw you a picture of the dual power supply arrangement but you will notice that most analogue circuits (ICs and discrete) actually use a +Vs and a -Vs connection.
Have a look at the circuits in this link and you will see what I mean about the extra complication of using a 'single ended' Power supply.

 

[abro]

Thanks I think I understood it, the speaker has to be connected to the positive and negative terminals?

Not at all. The push pull amplifier is a sophisticated circuit and it is easiest to describe, initially, by connecting it in a circuit with a positive supply and a negative supply, with a 'common' ground. (It's one step up from the simple transistor amplifier circuit that we all start with ) That allows you to have your load connected to ground and everything is symmetrical about that ground. If you choose not do do it that way, you need to have your input signal 'biased' so that its mean value is 'half way' between two connections of a single battery and you also need to connect your load to a half voltage point. This is very often done in practice but this is much harder to understand for the beginner (well - it confused you! ;))

It would be a lot of trouble for me to draw you a picture of the dual power supply arrangement but you will notice that most analogue circuits (ICs and discrete) actually use a +Vs and a -Vs connection.
Have a look at the circuits in this link and you will see what I mean about the extra complication of using a 'single ended' Power supply.

 

[sophiecentaur]

Thanks I think I understood it, the speaker has to be connected to the positive and negative terminals?

Just to clear this up:
The picture you drew is perfect and the simplest way to think of this thing working. It shows how each half only works for a particular polarity.
This phrasing is confusing though. It just shows how much better a diagram can be than a description.

 

[sophiecentaur]

For a simple amplifier, a dual power supply probably adds too much complication, even if you need a large electro go with a single supply.

It is universal now to use a dual supply for high powered amplifiers.
This adds extra cost and a need to have perfect balance to get zero DC current in the speaker.

Naah! Feedback takes care of that for pennies. It also takes care of crossover and other distortions. A dual power supply can often be cheaper because the supply hum can cancel out when fed to a balanced amplifier (saving again on big Capacitors). The amp can act as its own DC regulator, to some extent. Something in my distant memory tells me that balancing technique was used on some big valve amps, too.

 

[vk6kro]

Big valve amplifiers used coupling transformers. There was no need to balance them beyond selecting valves with similar gain.

They were big, heavy and needed lots of feedback to even sound reasonable.

This thread is about producing a very simple amp and explaining how a complimentary symmetry amplifier works.

If it was permitted, there are IC amplifier chips which inherently solve the balance and distortion problems we had to wrestle with.

The circuit discussed above consists of two emitter followers, so it is already low voltage gain and adding feedback will only make it worse.
So, it would be better to take care and get the bias and balance right without having to resort to feedback.

 

[sophiecentaur]

Big valve amplifiers used coupling transformers. There was no need to balance them beyond selecting valves with similar gain.

They were big, heavy and needed lots of feedback to even sound reasonable.

This thread is about producing a very simple amp and explaining how a complimentary symmetry amplifier works.

If it was permitted, there are IC amplifier chips which inherently solve the balance and distortion problems we had to wrestle with.

The circuit discussed above consists of two emitter followers, so it is already low voltage gain and adding feedback will only make it worse.
So, it would be better to take care and get the bias and balance right without having to resort to feedback.

Typically, of PF, this thread is operating at multiple levels. For the fewest components and for the easiest analysis, a dual voltage power supply is by far the best way to go. The OP has come up with a perfectly adequate diagram with bipolar supplies with which to answer the question and he has got what he needed from the thread, I think.. There is no way that it would be used, as is, for serious applications but that's the nature of all simple exemplar circuits and I'm sure that he appreciates that fact.

On the next level and introducing several more concepts - we both know that the design is not optimal and also that any properly designed power amplifier will involve feedback so why suggest it is something that one 'resorts to'? It's the essence of all good analogue design, surely and anyway this unity gain stage already involves 100% feedback by its very nature (i.e. it uses emitter followers which have inherent voltage feedback). DC coupled Audio output stages are very common and they will always use dual supplies - or the equivalent - with DC feedback to eliminate DC offset.

 

[rude man]

Thanks I think I understood it, the speaker has to be connected to the positive and negative terminals?

No. In your original circuit the speaker gets connected between the emitters and a 6.00V power supply.

Notice that for input voltages between about +5.40V and +6.60V there is essentially no output current. As someone may have said previously that can be taken care of by a large gain stage feeding the bases, plus feedback.

 

[sophiecentaur]

No. In your original circuit the speaker gets connected between the emitters and a 6.00V power supply.

Notice that for input voltages between about +5.40V and +6.60V there is essentially no output current. As someone may have said previously that can be taken care of by a large gain stage feeding the bases, plus feedback.

To be fair, he has drawn a correct diagram - it's just the way he has described it verbally. (A picture speaks a thousand words - as does an equation)