Handbook, chapter 7. Modulation: AM, SSB, FM, CW, digital
A radio wave on its own carries nothing. A plain sine wave at fourteen megahertz is just a plain sine wave: switched on, switched off, and there is nothing more in it.
For it to carry anything it has to be spoiled according to an agreed rule. Vary the amplitude, or the frequency, or simply switch it on and off. The rule is called modulation, and the way of reading it back, demodulation.
One wave, four ways to spoil it
Three lines: the carrier in the middle and two sidebands. All the information is in the sidebands, while the carrier takes most of the power and carries nothing.
The upper picture is what an oscilloscope would see. The lower one is what the neighbour's receiver would see. Switch between modes and watch both at once: that is the whole chapter.
AM: the simplest and the most expensive
Amplitude modulation varies the height of the wave with the voice. The envelope follows the sound, and the demodulator is literally a diode and a capacitor. That is how the first receivers were made, and how medium-wave broadcasting still works.
Look at the spectrum: three lines. The carrier in the middle and two sidebands either side.
And here is the arithmetic that has all but driven AM out of the amateur bands. All the information is in the sidebands. The carrier carries nothing while taking at least two thirds of the power. The two sidebands are identical — one of them duplicates the other.
So of every hundred watts, about sixteen go to any useful purpose.
SSB: the same thing without the waste
Remove the carrier. Remove one sideband. One line is left — and in it everything needed to reconstruct the voice.
That is single-sideband modulation, and it is what almost all voice work on the short waves uses. The gain is twofold, and large both times:
- half the bandwidth — about 2.4 kHz instead of 6, and a narrower band means less noise in the receiver;
- all the power is in the signal, because there is no longer a carrier or a duplicate.
Together that is about four S-points over AM at the same power. No amplifier at a sensible price gives an advantage like that.
And there is a third advantage, seldom mentioned and used every day. When two single-sideband signals end up on the same frequency they do not merge into porridge: the ear hears both voices separately and can concentrate on the one it wants. In AM or FM the stronger signal would simply swamp the weaker. That is exactly why, in a pile-up where dozens of stations call at once, the operator of a DX expedition still picks somebody out.
There is a price, and it is an honest one: a single-sideband signal cannot be received with a simple detector. The receiver has to put the discarded carrier back itself, accurately and stably. A hundred hertz out and the voice becomes a mumble or a squeak. That is where all the complexity of the chains in chapter 8 comes from, and where the need to tune precisely onto every station comes from too.
Either sideband can be discarded, and the convention is simple: below 10 MHz people use the lower sideband (LSB), above it the upper (USB). The reason is purely historical and has no physical sense to it at all. Just do as everyone else does — otherwise you are heard as unintelligible burbling.
FM: wide, clean, for VHF
Frequency modulation does not touch the amplitude at all — it shifts the frequency up and down with the voice. So it is indifferent to crackles and to changes of level: the receiver looks at the frequency, not at the loudness.
The price is bandwidth. Switch the drawing to FM and pull the deviation: the lines multiply, and the greater the deviation the further they spread. Amateur FM occupies about 12 kHz — five times more than SSB.
On the short waves that luxury is out of the question, while on two metres and seventy centimetres there is plenty of room. So FM is the language of repeaters, mobile stations and handhelds, while long-distance work on those same bands is done with SSB and CW just as on HF.
CW: what is audible when nothing else is
CW is the oldest modulation and the simplest: the carrier is simply switched on and off. Formally it is amplitude modulation at a hundred per cent depth.
The bandwidth comes out tiny. A receiver can narrow the filter to 250 hertz — ten times narrower than a voice channel, which is ten decibels less noise, and all of it free. On top of that, the human ear latches onto a steady tone long after speech has fallen apart.
Together that is the difference for which people work CW when a band is all but closed.
It seems that a CW signal is one line and cannot be any wider. In fact the width is set by how sharply it switches on. A perfectly rectangular key would give an infinite bandwidth; a real one with sloping edges fits into a few hundred hertz.
That is why transceivers have a rise-time control, and why too «hard» a key is heard on neighbouring frequencies as clicks. You can work out the bandwidth for your own speed with the timing calculator.
Emission designators: three characters in the licence
In the Radio Regulations, in datasheets and in amateur licences the modes are written as three characters. It looks like a cipher and reads simply:
| Designator | What it means |
|---|---|
| A1A | CW: amplitude modulation, one channel of digital information, copied by ear |
| J3E | SSB: single sideband with suppressed carrier, an analogue telephony channel |
| A3E | ordinary AM |
| F3E | FM telephony |
| J2D | digital modes in a single-sideband channel — this is where FT8 falls |
The first character is the type of carrier modulation, the second the nature of the signal, the third what exactly is being sent. Only the first two are worth memorising, but recognising them is worth it: what you are allowed on each band is written in these designators.
Digital: the same modulation, different content
Digital modes are not a separate kind of modulation. FT8, RTTY and PSK are audio fed into the same single-sideband transmitter, and what goes on the air is an ordinary SSB signal. The only difference is that the audio is produced by a program rather than by a voice.
So the licence for them is the same and they occupy bandwidth within that same channel — and FT8 fifty times less of it.
A tree of the code, not a table
Finally, something about the Morse alphabet that a «letter — code» table does not show.
Morse code is built around the frequency of letters in English: E is one dot and T one dash because they are the commonest; Q, J and Z are four elements each. Alfred Vail counted letters in a printer's type case a good hundred years before Huffman proved that this was optimal.
In a flat table that optimisation is invisible. On a tree it is obvious at first glance — along with the fact that it was not optimised for our language: the Morse code tree, drawn as a printed board.
Next — chapter 8, the chain: why a receiver needs an oscillator of its own.