Amplifiers and Classes of Operation

An amplifier is where a signal grows: a few microvolts from the aerial become loudspeaker audio, and a few milliwatts from an oscillator become the 50 watts your licence allows on HF. The exam asks two kinds of question about them — decibel arithmetic, which is pure calculation, and classes of operation, which is the one place in Section A where a technical answer is also a legal one. Use class C on SSB and you will splatter across the band.

What an amplifier actually does

An amplifier does not create energy. The extra power in the output comes from the power supply; the input signal merely controls how much of it is released, in the same way a small movement of a tap handle controls a large flow of water. This is why a transmitter with a 50 W output needs a supply that can deliver rather more than 50 W, and why an amplifier that is doing nothing at all can still be getting hot.

That framing also explains why bias matters so much. The previous lesson set the operating point; this lesson is about what happens to the signal depending on where that point was put.

Gain

Gain is simply output divided by input, and there are three of them:

Gain is a ratio, so it has no unit. An input of 2 mV producing an output of 2 V is a voltage gain of 2 ÷ 0.002 = 1000 — not 1000 volts, not 1000 anything, just 1000.

The decibel

Ratios in a radio get inconveniently large, and they multiply when stages are put in series. Taking logarithms turns multiplication into addition and squeezes a range of a million into a range of sixty. That is the decibel.

Why the two different multipliers? Because power goes as voltage squared (P = V²/R), and the log of a square is twice the log. It is the same decibel; only the quantity being compared has changed. The figures worth knowing by heart:

A decibel is a ratio, so on its own it never tells you an absolute level. Add a reference and it does: dBm means decibels relative to one milliwatt. 0 dBm is 1 mW, 30 dBm is 1 W, and 47 dBm is 50 W.

Worked example 1 — an output power in dBm

An amplifier delivers 100 mW. Expressed relative to 1 mW:

dB = 10 log (100 ÷ 1) = 10 log 100 = 10 × 2 = 20 dBm

Worked example 2 — a voltage gain in dB

2 mV in, 2 V out. The ratio is 1000, and this is a voltage ratio, so use 20 log:

dB = 20 log 1000 = 20 × 3 = 60 dB

Use 10 log by mistake and you get 30 dB, which is the answer the paper puts in the list to catch you.

Worked example 3 — gains and losses in cascade

An amplifier with a gain of 30 dB feeds a 10 dB attenuator. Decibels add:

+30 dB − 10 dB = +20 dB

Check it in ratios: 30 dB is a power ratio of 1000, 10 dB of attenuation divides by 10, and 1000 ÷ 10 = 100, which is 20 dB. The same answer, with more arithmetic.

Cascading stages

Real equipment reaches its gain in several steps, and the decibel makes the bookkeeping trivial: every gain is a plus, every loss a minus, and you add along the chain.

Classes of operation

The class of an amplifier is nothing more than a statement of how much of each input cycle the device conducts for — the conduction angle — and that is set entirely by the bias. Everything else about a class, its efficiency and its linearity, follows from that one number.

Step through the four classes in the panel below. Watch the bias line move down past cut-off, watch the collector-current waveform lose more and more of the cycle, and watch the efficiency figure climb as it does. The trade you are looking at — linearity against efficiency — is the whole content of this section.

Class A never switches off. Its transistor draws full quiescent current with no signal at all, which is exactly where the poor efficiency goes: the power is dissipated as heat whether you are transmitting or not.

Class B biases the device precisely at cut-off, so it conducts for half the cycle only. One device therefore cannot reproduce a whole waveform, and class B is used in push-pull: two devices, one handling each half. That arrangement has a characteristic fault. Near the zero crossing the incoming signal is still smaller than the 0.7 V needed to turn either device on, so for a moment neither conducts and a step appears in the output. That is crossover distortion.

Class AB is the cure: give each device a little forward bias so the two halves overlap slightly. The conduction angle rises above 180°, the step disappears, and you give up only a little efficiency. That is why the linear amplifier in an SSB transmitter runs class AB.

Class C and the tuned output

Class C biases the device well beyond cut-off, so it conducts only on the peaks — a conduction angle well under 180°. What comes out of the device is not a waveform at all but a train of short current pulses. Left alone that would be useless. What makes class C work is the tuned circuit in the output.

A parallel LC circuit resonant at the signal frequency behaves as a flywheel: each current pulse gives it a push, and between pulses the energy sloshes between the inductor and the capacitor, filling in the missing part of the cycle. The output is a clean sine wave again. A class C amplifier must have a tuned output circuit — it is not an optional refinement, and the tank also rejects the harmonics that those pulses are rich in. The resonance lesson explains the flywheel action in detail.

Distortion and harmonics

A perfect amplifier's output is a scaled copy of its input. Any departure from that straight input-output line is non-linear distortion, and it always shows up the same way: as new frequencies that were not in the original. Those new frequencies are harmonics — whole-number multiples of the fundamental. The second harmonic is twice the fundamental, the third is three times, and so on.

This is why a transmitter has a low-pass filter at its output. A 7 MHz transmitter with a second harmonic at 14 MHz is transmitting into an amateur band it may be entitled to use but is certainly not entitled to interfere with, and a third harmonic at 21 MHz is no better. Where two signals are present at once, non-linearity also produces intermodulation — sums and differences that land nowhere near a harmonic and are far harder to filter out.

The buffer amplifier

Not every amplifier is there to make a signal bigger. A buffer has a voltage gain of about one and exists purely to provide isolation: high input impedance so it barely loads the stage before it, low output impedance so it can drive whatever comes after. The emitter follower from the previous lesson is the standard buffer.

What a buffer protects is the stage in front of it, and the classic case is the oscillator in a transmitter. Without a buffer, every change of load in the later stages — keying the PA, a change in antenna loading — reaches back and pulls the oscillator frequency. On the air that is heard as chirp on CW and drift on phone. The next lesson takes this up properly.

Coupling between stages

A tuned LC or tuned-transformer coupling is the RF version of the same idea: it transfers energy efficiently at one frequency and rejects everything else, which is exactly what an IF strip needs.

Feedback

Feedback returns part of the output to the input. Which way round it arrives decides everything.

Negative feedback is a deliberate trade: you throw away gain, which is cheap, and buy predictability, which is not. Positive feedback is what an amplifier is doing when it howls, whistles or takes off on some frequency it was never meant to work at — and it is also, applied on purpose, exactly how an oscillator is built. That is where this module goes next.