AM, FM, SSB and CW

Everything up to here has been about producing a clean carrier and receiving it. A carrier on its own carries nothing — it is a pure tone at a radio frequency, and it tells a listener only that you are there. Modulation is the act of putting information onto it, and the way you choose to do that decides how much spectrum you occupy, how far you get on a given power, and what your emission is called on your licence. Section A tests the arithmetic; Section B tests the designators. They are the same subject seen from two ends.

Why modulate at all

Two reasons, and the exam wants the second one.

First, an unmodulated carrier conveys no information — and the Indian station-conduct conditions forbid transmitting one anyway.

Second, and more fundamentally, audio frequencies cannot be radiated efficiently from any practical aerial. An aerial has to be a useful fraction of a wavelength to radiate, and a 1 kHz wave is 300 kilometres long. A quarter-wave aerial for it would be 75 km tall. Shift the same information onto a 14 MHz carrier and the quarter-wave aerial is 5 metres. Modulation buys you an aerial of sensible size — and, just as importantly, it lets thousands of stations share the spectrum at once by putting each on a different carrier.

Compare the four modes on one screen

The panel below takes one audio tone and one carrier and modulates them four different ways, showing the waveform above and the spectrum below. Start on AM and push the depth past 100% to watch the envelope fold through zero and the sidebands spread out — that spreading is the splatter your licence forbids. Then switch to SSB and see the carrier and one sideband vanish, and to FM and watch the envelope stay perfectly constant while the spacing of the cycles changes.

CW — the simplest modulation there is

Continuous wave telegraphy is amplitude modulation reduced to two states: full carrier with the key down, nothing with it up, in the pattern of the Morse code. The emission designator is A1A.

Its virtues follow from its simplicity. Its bandwidth is only a few hundred hertz, so the receiver can use a 500 Hz filter and shut out almost all the noise and all the neighbours; and all the transmitted power goes into the one thing being sent. That is why CW has the best weak-signal performance of any of these modes and gets through when speech cannot.

Its one characteristic fault is key clicks. Switching a carrier on and off instantaneously is a broadband event — a step change contains energy at every frequency — and it splashes clicks across neighbouring frequencies. Shaping the keying so the envelope rises and falls over a few milliseconds cures it without making the signal any harder to read.

Amplitude modulation

In AM the amplitude of the carrier is varied in step with the modulating audio, while its frequency stays fixed. Draw a line joining the peaks of the resulting waveform and you get the envelope, which is a copy of the audio waveform. That is what an envelope detector recovers.

Sidebands and bandwidth

Multiplying a carrier by an audio tone is a mixing process, and mixing produces sum and difference frequencies. So a carrier at f c modulated by a tone at f m produces three components: the carrier, an upper sideband at f c + f m and a lower sideband at f c − f m. The information is in the sidebands; the carrier merely sits in the middle carrying none of it.

Worked example 1 — AM bandwidth

An AM transmitter is modulated by speech up to 3 kHz. What bandwidth does it occupy?

BW = 2 × 3 = 6 kHz — 3 kHz above the carrier and 3 kHz below it.

Percentage modulation

How deeply the carrier is modulated is expressed as a percentage. At 100%, the envelope just reaches twice the carrier amplitude on peaks and just touches zero in the troughs.

Drive past 100% and the envelope tries to go negative; the transmitter cannot produce a negative amplitude, so the peaks are clipped flat. Clipping generates harmonics of the audio, which appear as extra sidebands far outside the normal channel. That is splatter, and it is what a neighbouring station hears as your speech smeared across his frequency.

Where the power goes

This is the arithmetic that justifies the existence of SSB, and it appears in the paper as a plain multiple-choice question.

At 100% modulation, m = 1, so P total = P carrier × (1 + 0.5) = 1.5 × P carrier.

Take the carrier power as 1 unit. The total is then 1.5 units, and the extra 0.5 unit is split equally between the two sidebands, 0.25 each.

• Carrier: 1 ÷ 1.5 = two thirds of the total power
• Each sideband: 0.25 ÷ 1.5 = one sixth

So at full modulation — the best case — two thirds of everything your transmitter produces goes into a carrier that carries no information at all, and the two sidebands duplicate each other. Only one sixth of your power is doing useful, non-redundant work.

Single sideband

If the carrier carries nothing and the second sideband is a duplicate, remove them both. SSB suppresses the carrier and one sideband and transmits the other sideband alone. The designator is J3E.

Two gains follow. In principle the bandwidth halves, because you send one sideband instead of two. In practice the saving is larger: an SSB transmitter filters its audio to about 2.4 kHz, against roughly 6 kHz for a comparable AM signal, so real SSB occupies close to a third of the channel. And with the carrier gone, every watt goes into information rather than two thirds of it going into a tone that tells the listener nothing. Compared with AM at the same peak power, SSB gives an improvement of several S-units at the far end.

Generating it

Two standard methods, and the exam asks for both names together:

• The filter method: a balanced modulator produces double sideband with the carrier suppressed, and a very sharp crystal or mechanical filter then removes one of the two sidebands.
• The phasing method: two balanced modulators are fed with audio and carrier signals 90 degrees apart, and their outputs are combined so that one sideband adds and the other cancels.

Which sideband, and why it must be amplified linearly

By long-standing amateur convention: lower sideband below 10 MHz — so on 80 m and 40 m — and upper sideband above 10 MHz, on 20 m and up, and on VHF and UHF SSB. It is a convention rather than a rule, but choose the wrong one and nobody can read you: the audio comes out inverted and unintelligible.

Because an SSB signal carries its information entirely in the envelope, the power amplifier must be linear — class AB or class A. Put SSB through a class C stage and the envelope is destroyed and the result is splatter. This is the practical difference between an SSB rig and an FM one, and the reason an SSB transmitter's efficiency is lower.

Frequency modulation

In FM the frequency of the carrier is varied in step with the audio while its amplitude stays constant. Louder audio does not make the signal bigger; it makes the frequency swing further.

Deviation is how far the carrier is swung from its resting frequency by the loudest audio — a peak figure, quoted in kilohertz. The modulation index relates it to the audio that caused it:

Worked example 2 — modulation index

An FM signal has a peak deviation of 5 kHz and is modulated by a 2.5 kHz tone.

Index = 5 ÷ 2.5 = 2

Note that the index falls as the modulating frequency rises, even though the deviation has not changed. That is why an index is always quoted for a stated tone, and why it is not a property of the transmitter alone.

Carson's rule

Unlike AM, an FM signal has an infinite number of sidebands. Carson's rule is the practical approximation that captures about 98% of the power:

Worked example 3 — Carson's rule

Deviation 5 kHz, highest modulating frequency 3 kHz.

BW = 2 × (5 + 3) = 2 × 8 = 16 kHz

The tempting wrong answer is 10 kHz — simply doubling the deviation. That ignores the extra sidebands the audio itself contributes and understates the occupancy badly.

Narrow-band FM on 2 metres

Broadcast FM uses 75 kHz deviation, which is far too wide to fit an amateur channel. Amateur narrow-band FM uses a peak deviation of about ±5 kHz with audio to roughly 3 kHz, giving the 16 kHz above — comfortably inside the 25 kHz channel spacing used on the Indian 2 m band (144–146 MHz).

Capture effect, pre-emphasis, and the class C bonus

An FM detector responds to frequency, not amplitude, and a limiter ahead of it strips amplitude variation away entirely. So when two FM stations transmit on the same channel, the stronger one does not merely dominate — once it is a few decibels up it completely suppresses the weaker, which disappears. That is the capture effect. On AM the two would simply be heard together, heterodyning against each other. It is why an FM repeater is either perfectly readable or gone, with little in between.

Noise in an FM system rises with audio frequency, so the higher audio tones suffer most. The cure is applied at both ends: pre-emphasis in the transmitter lifts the higher audio frequencies before modulation, and matching de-emphasis in the receiver pushes them back down again — carrying the high-frequency noise down with them and leaving the speech unchanged.

Finally, because FM carries nothing in its amplitude, its envelope is constant. That means an efficient, non-linear class C power amplifier can be used, which SSB cannot tolerate. A cheap FM handheld gets more watts out of its battery than an SSB rig of the same size for exactly this reason.

Phase modulation

Phase modulation varies the phase of the carrier with the audio instead of its frequency. Since a change of phase over time is a change of frequency, PM and FM are closely related and a receiver cannot easily distinguish them: an FM discriminator demodulates a PM signal perfectly well. The practical difference is that PM is straightforward to apply to a crystal oscillator without pulling it, so many "FM" transmitters are in fact phase modulators with a compensating audio network in front. Both are constant-envelope angle modulation, and both are covered by the F and G symbols in the ITU emission classification.

The four modes side by side

Getting it back: demodulation

Each mode needs its own detector, and matching them is a reliable one-mark question.

The examiner will also ask you to name these modes in the ITU's notation — A1A, A3E, J3E, F3E and the rest — and the questions frequently pair the designator with the mode's properties in a single option. Read Reading an Emission Designator alongside this lesson rather than long after it; the two are examined together.