How can I build a class AB voltage amplifier?

[Ben321]

I know about the class AB current amplifier (output voltage is same as input, but maximum possible output current is increased). You just use a circuit like shown here:


I also already know about the class A voltage amplifier (output voltage is higher than the input, but the maximum possible current is less) like shown in the picture below:


The advantage of class A is that it increases the voltage of the signal.
The disadvantage of class A is that is that it uses 0 volts (ground) as the lower voltage rail. This means that the input and the output must be capacitively coupled, and thus you lose any DC component in the input signal that you want to keep in the output.

The advantage of a class AB amplifier is that it can use a negative voltage for the lower voltage rail, thus keeping the voltage at the input port and output port at 0 volts. This means that the capacitive coupling on the input and output is unnecessary.
The disadvantage is that it is not a voltage amplifier. At most, the output voltage is equal to the input voltage.

What I need is an amplifier that has the positive and negative voltage rails like in a class AB amplifier, but that performs voltage amplification like a class A amplifier.

An op-amp might seem like the solution, but if you look at the specs of any op-amp, you will find its amplifying properties suffer at higher frequencies, and they go away completely at about 20khz (open loop gain of 1x at 20khz, and continues to decrease even further above 20khz). This is completely useless for an RF amplifier.

For an RF amplifier, you need an open loop gain of at least 100x at frequencies of at least 1 MHz (and that just gets you to the middle of the AM radio broadcast band). I've heard that to get this kind of gain, at these frequencies, you need to build your own amplifier with discrete transistors. op-amps just won't work. And, if you want to keep the DC component (for example in an oscilloscope with DC coupling), it will need to be based on the class AB amplifier (power supply has +volts to -volts, not +volts to 0volts), so that the input and output ports with nothing connected will be at 0 volts without the use of an AC coupling capacitor.

Can you post a schematic for a simple 2-transistor amplifier powered with + and - voltage supplies, who's output signal has a voltage (not current) gain?

[Wimberleytech]

Can you tell us what you are building?

Are you really limited to ONLY two transistors?

[Ben321]

Can you tell us what you are building?

Are you really limited to ONLY two transistors?

Considering that you need 2 transistors for an amplifier running from a +and- voltage power supply, and I'm trying to keep my component count down, it would be nice to do this with only 2 transistors. If a voltage amplifier with 1 supply only needs 1 transistor, a voltage amplifier with 2 supplies should need only 2 transistors. I just need to figure out how to arrange them for voltage (instead of current) amplification.

[Benta]

This has nothing to do with class A, B, AB or even class C amplifiers.

The output stage is the same in all cases (your "class A" amplifier is just a common-emitter stage), the difference is how you set quiescent current - from cut off to full output current.

Voltage amplifiers always have common-collector outputs, which do not give voltage gain, but rather current gain. This is to get the output impedance as close to zero as possible (otherwise it wouldn't be a voltage amplifier).
The voltage gain is in the input stage and the driver stage. You could probably get away with using just one of those. But limiting your circuit to just two transistors will not give good results.

[Ben321]

This has nothing to do with class A, B, AB or even class C amplifiers.

The output stage is the same in all cases (your "class A" amplifier is just a common-emitter stage), the difference is how you set quiescent current - from cut off to full output current.

Voltage amplifiers always have common-collector outputs, which do not give voltage gain, but rather current gain. This is to get the output impedance as close to zero as possible (otherwise it wouldn't be a voltage amplifier).
The voltage gain is in the input stage and the driver stage. You could probably get away with using just one of those. But limiting your circuit to just two transistors will not give good results.

I anticipate my load will have a very high impedance (the load will be the input pin of an analog-to-digital converter IC). The common collector output stage is not required.

My plan is like this:
Weak input signal ---------> 2transistor-2powersupply-voltage-amplifier ------------> 8bit ADC ----------> parallel port in PC -----------> signal analysis software

[Benta]

This has nothing to do with class A, B, AB or even class C amplifiers.

The output stage is the same in all cases (your "class A" amplifier is just a common-emitter stage), the difference is how you set quiescent current - from cut off to full output current.

Voltage amplifiers always have common-collector outputs, which do not give voltage gain, but rather current gain. This is to get the output impedance as close to zero as possible (otherwise it wouldn't be a voltage amplifier).
The voltage gain is in the input stage and the driver stage. You could probably get away with using just one of those. But limiting your circuit to just two transistors will not give good results.

I anticipate my load will have a very high impedance (the load will be the input pin of an analog-to-digital converter IC). The common collector output stage is not required.

My plan is like this:
Weak input signal ---------> 2transistor-2powersupply-voltage-amplifier ------------> 8bit ADC ----------> parallel port in PC -----------> signal analysis software

Yes... Well... That wasn't your question, was it? Perhaps you should define your requirements a bit better, then you'll get better answers.

[dzseki]

Perhaps you should look harder into opamps, there are a great variety that have 40dB or more gain at 1MHz, let alone current feedback opamps...
Also be aware that ADCs often act as a rather difficult load for even properly built amplifiers!

[Ben321]

Perhaps you should look harder into opamps, there are a great variety that have 40dB or more gain at 1MHz, let alone current feedback opamps...
Also be aware that ADCs often act as a rather difficult load for even properly built amplifiers!

From my understanding, the input pins on ICs have a high impedance (infinite, for all practical purposes), so assuming that an ADC has a high input impedance is a good assumption. Right? Not worrying about impedance matching between a voltage amplifier and an ADC will cut down on component count.

What's a "current feedback opamp"? Is that the same as a regular opamp, with the feedback loop? Or is it a special type of opamp?

[Wimberleytech]

Perhaps you should look harder into opamps, there are a great variety that have 40dB or more gain at 1MHz, let alone current feedback opamps...
Also be aware that ADCs often act as a rather difficult load for even properly built amplifiers!

From my understanding, the input pins on ICs have a high impedance (infinite, for all practical purposes), so assuming that an ADC has a high input impedance is a good assumption. Right? Not worrying about impedance matching between a voltage amplifier and an ADC will cut down on component count.

What's a "current feedback opamp"? Is that the same as a regular opamp, with the feedback loop? Or is it a special type of opamp?

No, ADC inputs are not infinite impedance.  For many ADC types, the input is a sampling network (switched capacitor)...you have to pay attention to such things.  Here is a random Analog Devices 8-bit ADC input spec.

[Doctorandus_P]

A bunch of years ago Tuxgraphics(.com) designed a lineair power supply where the 0-5V voltage from a uC pin was amplified with a few transistors to 30V. The Hiland supply with uC uses a very similar topology.

VT3 and VT4 form is an unipolair DC amplifer and the amplification is adjusted by the ratio of resistors R2, R4, R6.
VT1 and VT2 is a Shizalky pair to amplify the current.

The circuit is not very good, but it does work and is very simple.

[Siwastaja]

so assuming that an ADC has a high input impedance is a good assumption. Right?

No. A completely wrong assumption. Depending on case, ADCs sometimes require even less than 50 ohm drive impedance, and very often less than around 200 ohms. The assumption of high impedance is a classical trap for the young players.

For low frequencies and low sample rates, you can supply it using a small capacitor (like a 100nF ceramic), since the average consumption is low, and low impedance is only needed for short peaks (while sampling).

For higher sample rates, you typically use an opamp, which does both voltage and current amplification, and yes, gain bandwidth products go up to hundreds of megahertz if needed.

[MiDi]

For SAR ADCs Analog Devices has a series of videos how to drive them.