Two Stage Amplifer

[Hero999]

Yes, Wiki maximum power transfer for more information.

Even for RF power amplifiers the impedances are only matched to the the characteristic impedance of the amplifier, not to the output of the amplifier which needs to have as low impedance as possible.

 

[Nigel Goodwin]

Yes, Wiki maximum power transfer for more information.

Even for RF power amplifiers the impedances are only matched to the the characteristic impedance of the amplifier, not to the output of the amplifier which needs to have as low impedance as possible.

No, the output stage is matched as well, and it's vitally important that it is - it's by far the most important matched stage in the transmitter. 50 ohm output impedance amplifier, matched to 50 ohm impedance coax, matched to 50 ohm impedance aerial - if any of the three isn't 50 ohms, extra circuitry is used to make them match.

 

[marcbarker]

drawn wrong with the base bias resistor
The 2nd stage buffer doesn't need capacitive coupling, or bias. DC coupled is fine. Unless it's a level-shitter as well.

Narrow band RF amplifiers are usually LC coupled, with an L or Pi matching network. About 28 years ago I was actually doing this! Except the the second stage transistor is class C biased. Or is even biased NEGATIVE (more than class C)

 

[Hero999]

No, the output stage is matched as well, and it's vitally important that it is - it's by far the most important matched stage in the transmitter. 50 ohm output impedance amplifier, matched to 50 ohm impedance coax, matched to 50 ohm impedance aerial - if any of the three isn't 50 ohms, extra circuitry is used to make them match.

A transmission line only has to be matched at one end to avoid reflections and distortion. If the aerial matches the cable and the transmitter is designed to drive the load then it doesn't matter if its output isn't the same impedance.

 

[marcbarker]

No, the output stage is matched as well, and it's vitally important that it is - it's by far the most important matched stage in the transmitter. 50 ohm output impedance amplifier, matched to 50 ohm impedance coax, matched to 50 ohm impedance aerial - if any of the three isn't 50 ohms, extra circuitry is used to make them match.

Ask a HF Radio Ham transmitting from a centre-fed dipole, with a parallel conductor feeder between antenna and his shack, ask him what his VSWR he runs on his transmission line feeder?

Getting 1:1 match VSWR is a misnomer for some setups. That set up could be something like 20:1 swr, most if that reactive impedance mismatch.

Granted, between TX and ATU, it will be reasonably close to 50 Ω resisitive, but the point I'm making is not all transmission lines always need to be perfectly matched for maximum power transfer.

 

[Nigel Goodwin]

Granted, between TX and ATU, it will be reasonably close to 50 Ω resisitive, but the point I'm making is not all transmission lines always need to be perfectly matched for maximum power transfer.

For MAXIMUM power transfer they do, but obviously in practice 'perfect' doesn't happen, and ACCEPTABLE power transfer doesn't require perfect matching.

 

[marcbarker]

For MAXIMUM power transfer they do, but obviously in practice 'perfect' doesn't happen, and ACCEPTABLE power transfer doesn't require perfect matching.

Ah I think we're cross-purposes, and I might had misunderstood the argument a bit. Point I was highlighting is it's known in amatuer radio there is no significant power transfer advantage from getting a transmission line (antenna feeder) down to unity vswr, and that the swr can be very high. Almost the same power with unity swr goes out the antenna even with swr very high on the dipole feeder.

A bit I can add about 'acceptable' power transfer: an RF amplifier needs as much inter-stage power transfer as possible to make the stage gain acceptable, i.e. a 50 input resistance needs matching to the base of an RF transistor (= single figure ohms). If there was no matching at all, the transistor acts more like an attenuator than an amplifier.

But with an audio amp it's acceptable not to need a lot of power transfer. Besides, more cheap transistor junctions can be added to make it up. At audio the ac gain of a transistor is quite high compared to at rf.

 

[andrebc]

The value of the output capacitor (1nF) is so small that it passes only radio frequencies to the 50 ohm load.
If its value is increased to 100uF and the value of the emitter bypass capacitor (10uF) also increased to 100uF then amplifier will have some gain.

The value of the base bias resistor (215k) for the output transistor is so high that the transistor is almost cutoff when it tries to have a current of 52mA into the 50 ohm emitter resistor. It should be about 6.8k ohms.

Hi Audioguru,

Attached are the simulation results of the circuit you modified. for a Vin of 10 mv and Frequency of 1K you see the signal clipped at the negative cycle, but yes for a certain range of Vin I should not see any distortion if I keep trying Vin values. But the main issue here: is that I am drawing current equal to 42 mA through the buffer and my power budget is 15 mW max for a VCC=5V. So I believe at this point comes the effect of impedance matching. I think if do not have your impedances matched then you would have to give more power to your second stage Amp if you want to get something out. I might be wrong on this point. But in any case the modifications you did are not the solution to my problem. IC for thr emmiter follower cant be more than 3 mA. and that's one reason why I had to put a higher R value there.

Thank for the contribution!!!

 

[Hero999]

The transmitter output impedance of the transmitter isn't critical, ideally it should be as low as possible to achieve maximum efficiency. Providing the aerial is matched to the cable it doesn't matter.

I've simulated it in LTSpice, see attached PDF.

The transmission line is 1/8 of the length of the fundamental so if it isn't matched there will be ringing and distortion as the harmonics will be reflected. The characteristic impedance of the transmission line is 50Ω.

1) RS = 1m, RL = 50
The ideal situation, nearly all the power is transferred to the load. No reflections or standing waves occur because all the power is absorbed in the load which matches the line perfectly.

2) Unmatched, in this case the load is high impedance so the voltage will be higher than that of the supply. Theis is why operating the transmitter without a load can damage it. Reflections of the harmonics cause distortion.

3) Unmatched, the load is low impedance so the current will be higher which can damage the transmitter. Again reflections are causing distortion.

4) Unmatched again - lots of distortion.

5) The source is matched but load is too low, no reflections or distortion because most of the energy is absorbed at the source the voltage is low - inefficient.

6) The source is matched so no reflections or distortion, inefficient but better than #7 (total match).

7) Both the load and source impedances are matched to the transmission line. 50% efficiency as half of the energy is lost at the source.

The point is that the transmission line only needs to be matched at one end. In the case of a transmitter we want the aerial matched and the transmitter's impedance to be low.

 

[marcbarker]

If this subthread is about transmission lines and matching, I think you need to 'think outside of the box' a bit more

In a practical tranmitter station setup, where Hams compete against each other to be the strongest with same input power limited by law, they have to find ways around problems, like obtaining a wide frequency range where there's narrow band components (i.e. antenna) that can't be adjusted easily.

One 'problem' is the keeping the transmitter output, feeder and the antenna matched for best power transfer. The problem is with a "matched antenna/line/transmitter", is that a large and fixed HF antenna can only be made to size for a narrow band of frequency, and that operating outside of this range there's so much mismatch, that only a fraction of the allocated band is available. So a station would have to pick a 'slot' and hope others chose the same slot.

The solution to the problem of covering the whole band, with all of the power, is don't try and match anything with the feeder/transmission line at all. Construct an antenna feeder which has some arbitrary Zo doesn't matter too much, but nevertheless relatively constant along length. It'll be two parallel conductors separated by air, with spacers holding the wires apart. No attempt to match the antenna to this feeder either. Then instead of the transmitter having 50 ohm output, it has a variable real and 'imaginary' impedance. The imaginary impedances are set to annilhlate/cancel eachother.

The transmitter real impedance is set for maximum RF field strength from the antenna, in practice it won't be the same as the antenna's real impedance, because the grossly-mismatched transmission line feeder will modify the impedance from one end to the other.

 

[marcbarker]

andrebc, a question about the circuit.

Why is the second stage AC coupled?

Why I ask is because the 2nd Q base voltage is nearly the same voltage as 1st Q collector. If that's intentional then you might as well just delete 10uf and 6k8!

If there's meant to be AC coupling for some reason, then you don't have to make the 2nd Q send 43 mA through the 50 ohm, you can set the base to a lower voltage and use the capacitor as a level-shifter. You could have a lot less current in the 50 ohm.

That is of course if you have enough signal headroom, i.e. only low output signal was required.

 

[audioguru]

Hi Audioguru,

Attached are the simulation results of the circuit you modified. for a Vin of 10 mv and Frequency of 1K you see the signal clipped at the negative cycle

The input level is so high that the output clips the top and bottom.
Maybe the gain of the amplifier is too high.

IC for the emitter follower cant be more than 3 mA.

Then the output power to a 50 ohm speaker is only 0.00011W which is nothing.
I used a current in the emitter-follower of 53mA so that the 50 ohm speaker will have an output power of 0.06W and will be heard faintly.

You should not use an emitter-follower to drive a speaker.
An audio power amplifier uses two complimentary emitter-followers at the output (NPN and PNP) as a class-AB pair with a low idle current but enough current-handling to drive a speaker.

 

[marcbarker]

An audio power amplifier uses two complimentary emitter-followers at the output (NPN and PNP) as a class-AB pair

What about a class A output audio amplifier?

Class A output (or single-ended as it was known before) was once so common that transformers (and speakers at first) were specially developed for it, being optmised for having a standing current.

Battery equipment lost single-ended output first because it was more competitive to have a longer lasting battery radio. Meanwhile much mains powered audio stayed single ended right up to when transistors replaced valves (tubes) in output stages.

 

[audioguru]

Andre in BC has a supply of only 5V and wants the output transistor to draw only 3mA. Then the amplifier power is only 15mW. A class-A amplifier wastes most of its power making heat so then the output to the speaker is nothing.

A class-AB amplifier is about 60% efficient so then the output to the speaker is 9mW which will barely be heard with the speaker to your ear.

 

[Hero999]

From the Word attachment containing the question:

https://www.electro-tech-online.com/threads/two-stage-amplifer.95250/#post766574

https://www.electro-tech-online.com/attachments/amplifier-doc.31526/
https://www.electro-tech-online.com/threads/two-stage-amplifer.95250/#post766574

Output impedance <50Ω
Output Swing ≤ 2V p-p
Max Supply Power = 15mW

That isn't possible using a class A amplfier.

2Vp-p into 50Ω is 10mW RMS and a class A amplifier is nowhere near 66.7% efficient.

 

[marcbarker]

Wow so defeatist!

Isn't the 50 Ω shown on the circuit the actual speaker itself? When there's a DC bias* on a speaker it still works, and in practice it'll sound louder one way round than the other. Does that change anything?

And does it have to be hi-fi quality? if it's for beeps or noises, like a continuity tester, then it could be class B or even C biased S/E transistor if the output sound level is constant.

*Sensible amount, not an amp.


ps. 9mW which will barely be heard with the speaker to your ear
Yuck! What kind of inefficient speaker is that?

 

[Hero999]

I don't see your point regarding DC in the speaker: DC current doesn't do anything, it just wastes power, as far as the amplifier is concerned it's only AC out that counts.

Class B or class C could be used to achieve higher efficiency at the cost of fidelity, indeed you could go for class D and get both. It's a moot point though because of the low power, the bias currents will be the dominating factor.

 

[marcbarker]

You double the output power by having the DC in the speaker instead of a resistor. Because a 50 resistor shunts the speaker 50. You can lose the blocking capacitor as well with direct drive speaker.

If it's lo-fi, use a CMOS gate as an amplifier and there'd be a high signal / bias ratio.

 

[Hero999]

You double the output power by having DC in the speaker instead of a resistor. Because a 50 resistor shunts the speaker 50. You can lose the blocking capacitor by direct drive speaker

How does the DC bias double the AC RMS power?

It doesn't.

 

[marcbarker]

How does the DC bias double the AC RMS power?

Who's saying "DC bias doubles AC RMS output power"? I'm not.

Maybe you missed my point. Why shunt the speaker impedance with a 50 resistor? All I'm saying is remove the resistor and drive the speaker direct. Is this difficult to grasp? I'm sorry I can't explain it better.