Need help classifying this circuit

[aspiringeng123]

Hi everyone, I am trying to understand amplifiers better and had one here that I would like to build and test:
http://www.ke3ij.com/amp.htm

How would this amplifier be classified as? I was thinking class AB, as it has two common emitter gain stages in the beginning which serve as a pre-amp like the guide says, but I am unsure as to what having the emitter for the pmos and nmos of the final stage does, but I've seen it for a few of these topologies across the net.

Essentially, I see the circuit is similar to the class AB shown here http://www.circuitstoday.com/few-transistor-amplifier-circuits, but I would like to know which of its differences make it either not an AB, or an AB with different input/output behavior.

Sorry if my question is a bit vague as I am a newbie to analog design.

 

[davenn]

Hi there
welcome to PF :)

yes, its class AB variation

but I am unsure as to what having the emitter for the pmos and nmos of the final stage

don't know what you mean by that tho ??
there are no MOSFETS in this circuit

did you read the Class AB description in your second link ?
it describes the operation of the output transistors

Dave

 

[aspiringeng123]

Oh I am sorry of course, I meant npn and pnp (it was late at night and I was quite tired).

Ah ok, so the second link is close enough in design to provide analysis as to what is going on in the first, even though there are additional caps in parallel at the collector end in the beginning pre-amp stage?

 

[sophiecentaur]

It's a while since I did any audio amp design but it surprises me that there are no emitter feedback resistors in the two driver stages, in both those circuits. In any amp I designed, I would always include emitter resistors in a basically common-emitter stage.
Has something changed in the last few decades?

 

[skeptic2]

Why would you always use emitter resistors? Do you see some advantage to their use?

 

[sophiecentaur]

If the emitter is held at 0V, what sort of voltage excursion can you have on the base and still keep the stage linear? An emitter resistor means you can bias the stage to operate at a sanitary voltage by fixing its operating conditions. You define the stage gain (Rc/Re) and the operating current (Ve/Re) because of the feedback. That sounds like a good idea to me. Most transistors, these days, have enormous but unspecified current gain so how can you predict how the stage will behave if you don't control the gain?
Are you saying that you haven't come across that technique? Google "emitter feedback" images and see just how commonly it's used in circuit design.

 

[skeptic2]

If the emitter is held at 0V, what sort of voltage excursion can you have on the base and still keep the stage linear?

The 390K resistor isn't just for biasing, it also provides a negative feedback path. The negative feedback allows just as much voltage excursion as an amplifier with an emitter resistor.

An emitter resistor means you can bias the stage to operate at a sanitary voltage by fixing its operating conditions. You define the stage gain (Rc/Re) and the operating current (Ve/Re) because of the feedback. That sounds like a good idea to me. Most transistors, these days, have enormous but unspecified current gain so how can you predict how the stage will behave if you don't control the gain?
Are you saying that you haven't come across that technique? Google "emitter feedback" images and see just how commonly it's used in circuit design.

Of course I've used emitter resistors in common emitter amplifiers but the use of emitter resistors may not be the best choice for all applications.

The gain for this this type of shunt negative feedback is also easily calculated. It is Av ~ Rfeedback / Rsource. You can have just as much control over the gain with this type of emitter follower as with the other type.

For instance if you want to design RF stages for anything above the lowest frequencies, you'd never use an emitter resistor. An emitter resistor not only may raise the input impedance of the stage so high it becomes difficult to match but may also make the stage unstable.

The various types of amplifier stages each have their own advantages and disadvantages and the one that is the best fit for the application is the one that should be used.

 

[sophiecentaur]

The 390K resistor isn't just for biasing, it also provides a negative feedback path. The negative feedback allows just as much voltage excursion as an amplifier with an emitter resistor.Of course I've used emitter resistors in common emitter amplifiers but the use of emitter resistors may not be the best choice for all applications.

The gain for this this type of shunt negative feedback is also easily calculated. It is Av ~ Rfeedback / Rsource. You can have just as much control over the gain with this type of emitter follower as with the other type.

For instance if you want to design RF stages for anything above the lowest frequencies, you'd never use an emitter resistor. An emitter resistor not only may raise the input impedance of the stage so high it becomes difficult to match but may also make the stage unstable.

The various types of amplifier stages each have their own advantages and disadvantages and the one that is the best fit for the application is the one that should be used.

Absolutely true. The application was an audio amp and not an RF Amp.
I realize that there are many configurations for operating a transistor but there are an awful lot of circuits around that make use of an emitter resistor - even if only for setting the current bias value. I was only reporting what I see.

 

[Baluncore]

There are fundamental problems with that amplifier design. The DC operating point of the output stage is decided by the current gain of Q3. Only the high value of the 390k bias resistor protects the output stage from thermal runaway since, as pointed out earlier, the emitter resistor(s) are missing. The two input stages are identical non-linear circuits, which will magnify the harmonic generation. That the author could put his name on the circuit demonstrates both design naivete and the rarity of such component sensitive designs in the real world. All in all, the circuit does not seem worthy of any deeper analysis.

 

[zoki85]

I realize that there are many configurations for operating a transistor but there are an awful lot of circuits around that make use of an emitter resistor - even if only for setting the current bias value. I was only reporting what I see.

In A-class ac amp basic topology bias voltage is stabilizied by resistor (R_E) in emitter circuit. If the supply voltage (V_cc) increases, the collector current (I_cc) increases, and this causes voltage drop across R_E to increase which,in turn, causes a base voltage (V_B) to increase. But V_B is held constant by divider resistors R₁,R₂ and the base-emitter voltage (V_BE) is lowered.This reduces the base current and keeps the collector current from increasing. The same stabilization principle happens if V_cc and I_cc try to decrease. Negative feedback at work.
Class B ac amp topology uses two complimentary transistors, each for a half of waveform. The output stage is configured in a push-pull fashion. There's no DC base bias current (quinescent current 0). No need for emitter resistor.
Class AB are crossbreads between A and B class toplogies and, in many cases, there's no need for emitter resistors.

 

[skeptic2]

The two input stages are identical non-linear circuits, which will magnify the harmonic generation. That the author could put his name on the circuit demonstrates both design naivete and the rarity of such component sensitive designs in the real world. All in all, the circuit does not seem worthy of any deeper analysis.

The way I read your post is that you are claiming that the non-linearity and harmonic generation in the shunt feedback amp, is significantly more than that which would occur with an amp with an emitter resistor of the same signal level, the same operating point and the same gain? Is this correct?

 

[Baluncore]

I was referring to the Q1 and Q2 stages as harmonic generators. Q3 has an emitter resistor.
The 390k "shunt" feedback resistor(s) are there to set the DC operating points of the class A stages, not the gain of the circuit.
To prevent thermal runaway, the complementary output stage needs one diode removed or an emitter resistor added.

I'm saying the design is a liability, and not worth serious analysis.

 

[skeptic2]

It is worth a comparison to an amp with an emitter resistor. The blue trace is taken at the collector of the shunt feedback amp and the red trace was taken at the collector of the series feedback amp. As we can see both amps are operating nearly at maximum gain, in fact the voltage gain in both circuits is about 430. At such a gain, with minimal negative feedback, the distortions of one circuit configuration with respect to the other should be apparent but they aren't. The Fourier Transforms of both waveforms are indistinguishable from each other. One may debate whether the gain is too high but it is obvious there is no issue that one is more non-linear or a magnifier of harmonic generation than the other.

 

[sophiecentaur]

Those pictures look like simulations so they depend upon the precise modelling of the components. How accurate are the characteristics of the transistors used?

 

[Baluncore]

Skeptic2, I cannot see the details on Common Emitter Comparison.jpg diagram.
Why is 390k feedback on one from collector, the bias on other is from Vcc?
What is Vpp amplitude of input simulation voltage?
What does FFT of output look like?
What are first four harmonic amplitudes for different input signal amplitude?

When two amps, (Q1 and Q2), are in series, Q2 will have higher signal input so net distortion of positive and negative cycles will be different, hence even harmonic generation.