In principle we can use a simple radio receiver which receives the required channel, amplifies it, demodulates the signal, and outputs the required information pattern (the music, or whatever). This type of system is a Tuned Radio Frequency (TRF) receiver of the type illustrated in figure 16.5a.


In practice, TRF systems are rarely used. Nearly all radio receivers (and TV's radar receivers, etc, etc,) use a superheterodyne system similar to that illustrated in figure 16.5b. ‘Superhet’ receivers have a number of advantages over the TRF. One of the most important of these can explained by considering figure 16.6.


In order to select the channel we want (‘Classic Sound’ in this case) and reject other stations the receiver has to include filters. Ideally, these filters would pass the desired channel without altering the carrier+modulation but completely block all other frequencies. To do this we would need a filter with a ‘flat top’ and steep ‘edges’. The ideal filter shape would be like rectangular brick. In fact it is possible to prove that it is physically impossible to make filters which have this perfect behaviour. All real filters have an imperfect frequency response. The filter shown in 16.6 shows typical behaviour. The passband shape isn't flat, hence the filter tends to attenuate sideband components which are at relatively high modulation frequencies. (This ‘high frequency droop’ is why cheap AM receivers sound muffled and boomy.) The filter's response also isn't zero everywhere outside the wanted channel. As a result some signal ‘leaks through’ from transmissions at nearby frequencies.

The performance of a filter is normally specified in terms of a Rejection Ratio. This indicates the filter's relative gain (as usual with engineers, gain here really means ‘loss’!) at the center of the passband compared to just outside. It is conventional to specify this in either of two ways: the Adjacent Channel Rejection Ratio and the Alternate Channel Rejection Ratio. The adjacent channels are those immediately on either side of the wanted (tuned) channel. The alternate channels are those one channel further way. For example, in the medium wave band when tuned Virgin AM on a carrier frequency of 1242 kHz the adjacent channels would be those using carriers of 1251 kHz & 1233 kHz. The alternate channels would be at 1260 kHz & 1224 kHz. (Note that it is usual when quoting a transmitter's ‘frequency’ to give the carrier frequency. The actual modulated transmission is spread out over the channel centred on that frequency.)

A good radio receiver will provide an adjacent channel rejection ratio of 30-40 dB and an alternate channel rejection ratio of around 60dB. This means that signals received on the alternate channels are reduced by a factor of around 1,000,000 compared to the power of the desired signal. Whenever possible broadcasting engineers keep transmitters using adjacent frequencies well apart geographically. This means that under most circumstances the alternate channel rejection ratio figure is the best guide to how well a receiver will reject unwanted nearby signals. Note that, although at first sight a rejection ratio of 60dB seems a lot, a nearby powerful transmitter may couple much more power onto the receiver's antenna than a distant low-power transmitter. For example, a 100 kW transmitter just 5 km away will provide an antenna signal level which is 10,000 times greater (+40dB) than a 1 kW transmitter 50 km away. In these circumstances a 60dB rejection ratio would mean that, when we were trying to listen to the distant station, we would also hear the local one coming through in the background 20 dB quieter. This would be quite audible and likely to spoil music listening. For this reason very good receivers (professional communications receivers and high performance hi-fi FM tuners) should have alternate channel rejection ratios of at least 80dB and top-class receivers reach 120dB!

This need for very high filter selectivity is the downfall of the TRF system. To be able to listen to a different station we have to re-tune the receiver. On the medium wave band this may mean changing the received radio frequency from around 600 kHz to around 1600 kHz. However, no matter where we tune in this range we'd still like a ‘flat topped’ filter response which has a rejection of 60 - 80 dB away from the required channel. Given the money we could make an RF filter which fits the requirement at a specific channel frequency. However the need to make the filter tunable over this range makes it impossibly difficult (i.e. expensive!) to maintain this performance over a wide tuning range.



Content and pages maintained by: Jim Lesurf (jcgl@st-and.ac.uk)
using HTMLEdit3 on a StrongARM powered RISCOS machine.
University of St. Andrews, St Andrews, Fife KY16 9SS, Scotland.