Last Updated on February 3, 2024 by John Berry
At the transmitter, information (speech or data) is impressed by various means on a carrier generated by an oscillator. The result is a modulated carrier. The means by which this happens depends on which of several of transmission modes is selected. This resultant signal can be characterised simply by its bandwidth, and its power. This modulated carrier propagates from the transmitter via the antenna to the receiver antenna, and then to the corresponding receiver. There, it’s demodulated by a reverse process, and the speech or data recovered.
The signals or transmission modes used in ham radio are based on three basic schemes – on-off carrier (CW), amplitude modulation (AM), and phase or frequency modulation (FM). Phase and frequency modulation are very similar when it gets to the antenna and are often considered as one.
There are many good texts on the scope and complexity of transmission modes, so this page does not aim to cover all options. It concentrates on a discussion to aid understanding and give characteristics that might be used elsewhere on this site.
Amplitude modulation
Under amplitude modulation, voice or data information is impressed on the carrier such that its amplitude, or size, changes.
Viewed in the time domain, we see the carrier vector, with a varying information vector (blue). The resultant vector changes both in phase and in amplitude in line with the modulation. Viewed in the frequency domain we see the carrier and two side-bands, one either side. These representations are shown below.
There are then many forms of this core concept, and each option is selected for its efficacy in the required transmission system.
First, once the composite ‘carrier-plus-modulation’ has been generated, the carrier can be suppressed, and one sub-carrier filtered out, giving the ‘single sideband, suppressed carrier’ mode, ‘SSBSC’, or just ‘SSB’. The remaining sideband then gives its name to the mode – upper sideband, USB, or lower sideband, LSB. Other options are possible.
The phase of the carrier-plus-modulation vector depends only on the propagation distance from the transmitter and hence is of no interest.
Second, the modulation is not restricted to voice. Data can be sent by representing symbols from a digital bitstream as audio tones. The most common digital sub-modulation used today is FT8 where eight tones are used to modulate the sub-carrier under the SSB amplitude modulation mode. In this case, typically USB is employed.
Demodulation simply requires the detection of the amplitude change. In simple systems, this is done by re-introducing the carrier and detecting the change with a diode.
Frequency (or phase) modulation
Under frequency modulation, voice or data information is impressed on the carrier such that its frequency changes.
Viewed in the time domain, we see the carrier vector, with a varying information vector (blue) at varying phase. The information vector is at fixed amplitude. The modulation changes the resultant carrier-plus-modulation size, but, more significantly, it changes its phase. Viewed in the frequency domain we see the carrier and multiple sidebands, either side. These representations are shown below.
Like amplitude modulation, a data bitstream can used as modulation, with the modulator forcing a specific phase to represent each bit. If there are two data bits or symbols, a 1 and a 0, this would be referred to as binary phase shift keying, or BPSK. If the bitstream is coded, there could be multiple phases per symbol giving complex PSK schemes like 4-PSK, or quadrature phase shift keying (QPSK).
Many ham radio VHF digital transmission modes use PSK.
Demodulation simply requires the detection of the phase change. In simple systems, this is done by detecting the frequency change as a varying current in a tuned circuit.
As an aside, because FM has an amplitude varying component, it can be detected with an AM receiver. It’s not very efficient, but it works and is referred to as slope detection – detecting the amplitude change in the slope of the sidebands.
Complex modulation
There are all manner of complex modulation schemes that mix AM and FM, and, indeed, CW (continuous wave, keyed on and off) too.
An example might be 64-QAM – 64 level quadrature amplitude modulation, where the modulation vector is best viewed as one of 64 states in an 8 by 8 matrix.
A visual representation of 64-QAM is shown below.
Detection of QAM requires detection of amplitude and phase of the resultant, positioning it as a symbol on the matrix. Typically, detection of symbols is easy in simple transmission modes like BPSK, because there are only four. In complex modes, it becomes progressively more difficult to discern one symbol from another. In higher order transmission modes, progressively greater received signal to noise ratio is needed. As a result, they are only used in professional systems.
Characteristics of transmission modes
There are three core characteristics of transmission modes that matter to the radio amateur: necessary bandwidth, required signal to noise ratio (or bit error rate), and typical threshold of reception.
Necessary bandwidth
Typically, an amplitude modulated speech transmission requires about 7kHz of RF bandwidth. Such a mode would have both sidebands with full carrier and a maximum modulating frequency of 2.5kHz. Any less than 7kHz and information will be lost. To allow for frequency drift in equipment and for protection against adjacent channel noise, the channel bandwidth is generally a bit more. For 7kHz necessary bandwidth, the channel bandwidth would likely be 12.5kHz. FT8 and other minimalist systems only need, for example, 50Hz necessary bandwidth. This is achieved by employing two-stage filtering. First circuits filter for the recovered SSB audio signal, then microprocessors filter using digital signal processing (DSP) for the data.
Signal to noise
Typically, for speech, the recipient will need to benefit from about 10dB signal to noise (S/N) ratio at the receiver demodulator. Less, and information will be lost. More and it will be easier to listen to but much more will give no benefit. When transmitting data, one can still talk about S/N, but it is more appropriate to talk of bit error rate. In one of his presentations, Joe Taylor, K1JT, talks of an “arbitrarily low” BER. In essence he’s talking of a BER of better than about 10-3, one error in a thousand. Less than this and errors start to get annoying, but above, there’s little to be gained.
Threshold
Finally, we can define the necessary signal power into the front end of the receiver as the threshold of reception. This is quoted for either a given signal to noise ratio or bit error rate. In analogue systems this is measured in -dBW (decibels below one Watt). Strictly, in digital systems, engineers measure the threshold in Eb/N0 – energy per bit of information in a normalised bandwidth (of 1Hz).
There are three good reasons why the signal to noise threshold is more relevant for radio amateurs:
- No ham radio equipment is specified in terms of Eb/N0,
- Transmission systems like FT8 are really AM (SSB) systems pretending to be digital, and
- Developers like Joe Taylor quote his DSP system threshold in dB below the analogue (SSB) threshold.
I’ve used the threshold in -dBW universally on this site.