Sensitivity, Selectivity, AGC and Squelch

The syllabus lists receiver characteristics as a shopping list — sensitivity, selectivity, fidelity, adjacent channel and image interference, A.V.C. and squelch, signal-to-noise ratio — and it is tempting to treat them as six things to memorise. They are not independent. Sensitivity is limited by noise, selectivity fights fidelity, and the AGC exists because sensitivity without it would be unusable. This lesson takes each in turn and says three things about it: what it means, what actually limits it, and what it looks like on the front panel of a real radio.

Sensitivity — and why gain is not the answer

Sensitivity is the ability to receive the weakest possible signal. It is quoted as the input in microvolts needed to produce a stated audio output at a stated signal-to-noise ratio — for example "0.2 µV for 10 dB S/N".

The tempting assumption is that sensitivity is a matter of gain: add another RF stage and you hear weaker signals. You do not. Every amplifier generates noise of its own — thermal noise in its resistances, shot noise in its devices — and amplifies that noise along with the signal. Turn the gain up and the signal and the noise rise together; the hiss gets louder, but nothing becomes readable that was not readable before. Sensitivity is limited by noise, not by gain.

Which stage's noise matters? The first one. Noise added at the input is amplified by every stage after it, while noise added late in the chain is amplified by nothing. A quiet first RF stage followed by ordinary ones gives a quiet receiver; a noisy first stage followed by superb ones cannot be rescued.

How much noise a stage adds is quoted as its noise figure, in decibels: the amount by which it degrades the signal-to-noise ratio passing through it. A noise figure of 0 dB would be a perfect, noiseless amplifier. Below about 20 MHz it barely matters, because atmospheric and man-made noise arriving on the aerial swamps anything the receiver contributes — which is why an HF preamplifier usually makes the band noisier without making a single signal more readable. On VHF and above, where the band is quiet, noise figure is the whole game.

Signal-to-noise ratio

Signal-to-noise ratio is the honest measure: the ratio of wanted signal power to unwanted noise power at the output, almost always expressed in decibels.

A speech signal needs roughly 10 dB of S/N to be comfortable; a CW signal can be copied at 0 dB or below. And where is it decided? In the front end, by the first stage — not in the audio amplifier.

Selectivity, shape factor and fidelity

Selectivity is the ability to separate the wanted signal from those on nearby frequencies. In a superheterodyne receiver it is set almost entirely by the bandwidth of the IF filter. The tuned circuit ahead of the mixer is deliberately broad — its job is image rejection, not channel selection — and contributes very little to the shape of the passband.

Bandwidth alone is not the whole story, because a filter has skirts. The standard way to describe them is to measure the width at two depths and take the ratio.

Worked example — shape factor

The filter drawn above is 2.4 kHz wide at −6 dB and 7.2 kHz wide at −60 dB.

Shape factor = 7.2 ÷ 2.4 = 3.0

A perfect filter would have vertical skirts and a shape factor of 1.0. A cheap ceramic filter might be 4 or 5; a good crystal or mechanical filter reaches 1.5 to 2. Lower is better, and that is the one thing worth remembering, because it is the opposite of the intuition that a bigger number means a better component.

Fidelity, and the fight with selectivity

Fidelity is the ability to reproduce the modulation without distortion. It needs a passband wide enough and flat enough to pass all the sidebands, plus a clean audio stage. And there is the conflict: every kilohertz you cut off the filter to reject the neighbouring station is a kilohertz of your own signal's sidebands thrown away. This is why a broadcast receiver, with a 9 kHz passband, sounds rich and a communications receiver squeezed to 2.4 kHz sounds thin and nasal. It is not a fault. It is the deliberate trade, and on a crowded 40 m evening you will take selectivity every time.

Adjacent-channel and image interference

Adjacent-channel interference is a strong station a few kilohertz away breaking through. The cure is a narrower IF filter, because the offending signal is arriving inside the IF passband.

Image interference is entirely different: a signal 2 × IF away from the wanted one, converted by the mixer onto the IF, where no filter can touch it. It is cured only ahead of the mixer — see TRF and Superheterodyne Receivers for the arithmetic and the three remedies. Confusing the two costs a mark almost every paper.

Blocking and cross-modulation

Blocking (or desensitisation) is a very strong signal — often well outside the filter — driving the front end towards saturation so that everything else goes quiet as it appears. Cross-modulation is the modulation of a strong unwanted station transferring onto a weak wanted one, so you hear a broadcast station's speech riding on a signal that is not carrying it. Both are front-end overload, and both are improved by turning the RF gain down or switching in an attenuator — the counter-intuitive fix that marks out an experienced operator.

AVC / AGC — automatic gain control

Signals fade. Over a minute on 20 m the same station can vary by 40 dB, and no one wants to ride the volume control. Automatic volume control — the older name, still used in the syllabus — or automatic gain control keeps the audio output roughly constant as the input varies.

The mechanism is a feedback loop. A sample of the detected signal is rectified and smoothed to give a DC voltage proportional to signal strength, and that voltage is fed back as bias to the RF and IF amplifiers. A strong signal produces a large bias, which reduces the gain; a weak signal produces little bias and the gain rises.

Fast and slow

Two time constants matter. Attack is how quickly the gain is pulled down when a signal appears; it is always fast, so a sudden loud station does not blast your ears. Decay is how quickly the gain comes back up when the signal stops, and that is the one you choose:

• Slow decay for SSB and CW. Neither has a carrier during a pause, so a fast decay would send the gain shooting up in the gap between every syllable and every dot, and the band noise would surge in and out. Slow AGC holds the gain steady across the gaps.
• Fast decay for FM and for rapidly fading signals, such as a mobile station driving through Bengaluru traffic, where the level genuinely changes several times a second and the AGC must follow it.

Squelch (muting)

An FM receiver at full gain with no signal present produces a loud, continuous hiss — the limiter has nothing to limit, so it amplifies noise to full output. Nobody can sit beside that all evening waiting for a call.

Squelch, also called muting, silences the audio stage until a received signal of sufficient strength arrives. You set the threshold with the squelch control: turn it up until the hiss just stops, and no further. Turn it too far and weak stations will not open it; leave it fully open and you get the hiss back. It changes nothing about the passband and nothing about the transmitter — it is a gate on the audio, and it is a VHF FM feature.

The S-meter, and RIT

The AGC line already carries a voltage proportional to signal strength, so it costs almost nothing to put a meter on it. That is the S-meter: it reads the strength of the received signal, and nothing whatever about your own transmission. By the convention IARU Region 1 recommends, one S-unit is 6 dB and S9 corresponds to 50 microvolts at the receiver input on HF; above S9 the scale is marked in decibels, so a report of "S9 plus 20" means 20 dB above S9. Real S-meters are rarely calibrated anywhere near that faithfully, which is why experienced operators treat the readable/strength part of an RST report as judgement rather than measurement.

RIT — receiver incremental tuning, called the clarifier on older sets — shifts the receive frequency by a few hundred hertz without moving your transmit frequency. You use it when the station you are working is slightly off your frequency: pull him into your filter with RIT and he still hears you exactly where he called.

Receiving CW and SSB

Both modes present the same problem: an ordinary AM detector cannot demodulate them.

Why a CW signal is silent

A CW transmission is a bare carrier switched on and off. Feed a steady unmodulated carrier into an envelope detector and the output is a steady DC level — no variation, so nothing for the loudspeaker to reproduce. The key going down would produce a faint click and then silence.

The cure is the beat frequency oscillator. The BFO injects a locally generated signal a few hundred hertz away from the IF; the two beat together in the detector and the difference is an audible tone. Set the BFO 700 Hz from the IF centre and a carrier at the centre of the passband produces a 700 Hz note. This is also why the pitch of a CW signal changes as you tune through it — you are changing the difference frequency. Most operators pick a beat note somewhere between 500 and 800 Hz: low enough to be restful over an hour, high enough that the ear separates it from the rumble of band noise.

SSB and the product detector

An SSB signal arrives with its carrier suppressed, so there is no carrier for an envelope detector to work against — the output would be unintelligible. The receiver has to supply the missing carrier itself, which is what a product detector does: it multiplies the incoming sideband against a locally generated carrier from the BFO or carrier oscillator, and the difference product is the original audio. The same circuit, with the same oscillator, is what recovers CW, which is why one switch position on a transceiver usually reads "CW/SSB".

How much selectivity do you actually want?

The right filter is the narrowest one that passes all of your signal and none of anybody else's. That depends entirely on the mode.