|
The pros and cons of different frequencies.
The author has shown in a previous article how poor a compromise a fixed-selectivity receiver is in the matter of quality when both local-station and distant reception is wanted. In the present article it is explained how the choice of value of the intermediate frequency stages of a superheterodyne is linked with the question of providing variable selectivity.
If the number of sets incorporating it is any criterion, the superheterodyne principle has established itself as the only one capable of meeting the exacting requirements of modern broadcasting conditions. Although no one will deny that it is possible to build a straight set which possesses the requisite sensitivity and selectivity for long distance reception, the problems of obtaining accurate ganging and of maintaining stability are likely to prove serious, and the resulting set is by no means simple. This is especially the case when variable selectivity is needed. It is by no means difficult to include this in the fixed-frequency IF amplifier of a superheterodyne, but the difficulties attendant upon fitting it to a straight set would seem prohibitive.
If we decide, as we must, that the superheterodyne best meets the requirements of general reception, as providing the simplest way of obtaining high sensitivity and selectivity, our decision is confirmed by the comparative ease with which variable selectivity can be included, for this has such a far-reaching effect upon the quality of reproduction that it is soon likely to be deemed an essential. The superheterodyne, therefore, proves to be the best type of receiver not merely for distant reception but for general use where a very high standard of quality is demanded in local reception together with high selectivity when distant stations are required.
Now broadcast reception may be divided into three distinct categories: local, where the receiver is situated within the service area of the wanted transmitter; medium distance, where reception is wanted from high-power stations which are several hundred miles away, but which give a strong signal after dark; and long distance, where the signals are usually very weak and emanate from transmitters of low or high power many hundreds or thousands of miles away. The degrees of selectivity and sensitivity necessary rise as we go through these categories in order, and the quality of reproduction obtainable steadily deteriorates. The minimum degrees of selectivity and sensitivity needed are set by the demands of local reception and are quite low and easy to satisfy, while it is accordingly possible to obtain very high quality reproduction. The maximum degrees of selectivity and quality, however, are set by the weakest signal required and by the interference conditions, and the difficulties increase as the signal weakens and the interference increases.
Whatever the frequency adopted for the IF amplifier, it is not difficult to obtain all the amplification which can be used. A limit is set to the useful amplification by the noise generated in the receiver and this is of several kinds. Valves contribute their quota of hiss, but the limit is really set by thermal agitation in the conductors forming the first circuit, and this applies whether the set be straight or superheterodyne. For a maximum ratio of signal to internal noise the input circuit should be efficient and resonant to the incoming signal. Given this condition and a correct choice of the first valve, the frequency at which amplification is carried out in unimportant. In practice, of course, one would tend to choose a low intermediate frequency rather than a high, since fewer stages are then needed to achieve the amplification with stability.
Turning now to selectivity, provided that the efficiency of the tuned circuits is the same, the lower their operating frequency the greater is their power of discriminating against unwanted stations, so that here, again, a low frequency would seem advisable. For a given degree of selection before the frequency-changer, however, a low intermediate frequency far more likely to lead to trouble from second-channel interference than a high one. We cannot, therefore, arbitrarily choose any frequency, and the pros and cons must be carefully weighed bearing in mind the purpose for which the receiver is intended; that is whether it is to be used for local, medium, or long-distance reception.
The broadcasting bands can be taken accurately enough as 150-300 kHz for the long waves and 550-1,500 kHz for the medium, and for all ordinary purposes these must be taken as prohibited bands within which the intermediate frequency cannot lie. We have, therefore, three possibilities for the intermediate frequency; it can be a low frequency of the order of 110 kHz, it can be a medium frequency of about 465 kHz between the two wavebands, or it can be a. high frequency of some 1,600 kHz. We shall confine the discussion to these three frequencies, and the question of amplification need not be considered since it is quite possible to obtain any degree we may need at any of these frequencies.
The Demands of Selectivity
The question of selectivity very complicated. Although it is simple enough to compare the selectivity of single circuits of the same efficiency at the different frequencies, this tells us only a small part of the story. It is usually possible to build more efficient circuits at a high frequency than at a low, so that this offsets the natural tendency of low frequency circuits to be more selective, while the necessity for tightly coupled band-pass circuits at a low frequency still further militates against it. Dielectric losses, however, assume serious proportions at a high frequency and the measures necessary to combat them tend to increase cost. It is, therefore, impossible to lay down hard and fast rules, and experience indicates that there is not a great deal to choose on the score of selectivity between 110 kHz and 465 kHz. Assuming that the same number of tuned circuits be used and the couplings adjusted for the same degree of modulation-frequency response at about 5,000 Hz, the selectivity can be of the same order, but the higher frequency demands rather better coils. The difference, however, is by no means great enough to rule out the use of a frequency of 465 kHz even when very high selectivity is demanded.
The case is different, however, when we consider 1,600 kHz, for it then proves impossible to increase the efficiency of the individual circuits sufficiently without the use of regeneration, and to obtain the same degree of selectivity many more circuits are necessary. It is generally accepted that for long-distance reception the response at 10 kHz off-tune must not be more than one-thousandth of that at resonance. It is easy to achieve this with six tuned circuits at 110 kHz, and not unduly difficult at 465 kHz, but quite out of the question at 1,600 kHz.
Turning now to second-channel interference, experience indicates that two signal-frequency tuned circuits give adequate protection with an intermediate frequency of 110 kHz, except in cases where the interference is caused by a local transmitter; three circuits are then needed or a special rejection system. Over the band of 150-300 kHz, second-channel interference can occur only from stations in the range 370-520 kHz, and this is outside the broadcast bands. On the medium waveband, however, stations between 770 kHz and 1,720 kHz may cause interference, and the risk is serious because this is the major portion of the medium waveband.
If we turn to an intermediate frequency of 465 kHz we gain considerably, for each signal frequency circuit gives roughly the protection of two with the lower frequency. In addition, the number of broadcasting stations capable of causing second-channel interference is much less. On the long waveband only stations between 1,080 kHz and 1,230 kHz can cause trouble, and on the medium the second-channel range is 1,480 kHz to 2,430 kHz, of which only the 1,480-1,500 kHz band coincides with a broad-cast band.
When we turn to 1,600 kHz, the second-channel band for the whole range of 150-1,500 kHz is 3,350-4,700 kHz, and interference of this kind from broadcasting stations is impossible. With such an intermediate frequency it is, in fact, possible to dispense with signal-frequency tuned circuits and to employ in their place a fixed filter. Ganging is then abolished, and as the full range can be covered in one sweep of the tuning capacitor wave-band switching is unnecessary. The single-span system of tuning thus offers considerable advantages over the more conventional superheterodyne.
The Final Choice
The chief characteristics of the different intermediate frequencies have been briefly outlined, and we are now in a position to make a choice, and this choice will naturally depend upon the use to which we wish to put the receiver. For a purely local receiver one frequency is as good as another, but the high frequency has the advantage of permitting single-span tuning, with all its simplification, to be used. More often than not, however, the straight set will be used, for neither selectivity nor sensitivity is important.
For a very long-distance receiver used in a district remote from any powerful transmitter, a low intermediate frequency is best, for there is hardly any limit to the degree of selectivity which it is possible to achieve, and no more than two preselector circuits will be needed. If the set is to be used within at few miles of at broadcasting station, three, or even four, signal-frequency circuits will be needed to avoid all trouble from second-channel interference.
In general, a receiver is not used exclusively for one purpose, however. The local receiver must give occasional reception of the stronger Continental transmissions, and the long-distance receiver is sometimes tuned to the local. We have, therefore, to consider more than a single requirement, and we must ever bear in mind the important points of convenience and cost. With an intermediate frequency of 1,600 kHz it is easy to secure more than adequate selectivity for local reception, and by means of regeneration it can be increased sufficiently to permit very good results being obtained from the stronger Continental transmissions. The degree of sensitivity required for such stations is not so. high that background hiss is likely to prove important, so that there is no objection to the aperiodic aerial system of a single-span receiver. When account is taken of the simplicity of its tuning arrangements, moreover, this type of set will be seen to be ideal for these particular receiving conditions.
The next class of listener requires a set which will still be used largely for local reception, but which will be used to a much greater degree than before for moderate distance reception, and occasionally for long-distance reception. Here the single-span receiver is again suitable, but its adjacent channel selectivity is lower than that of a receiver with a lower intermediate frequency. The latter, however, must have a more complex tuning system if second-channel and kindred forms of interference are to be avoided. A definite ruling in this category is impossible, for it is on the border line, on either side of which one system seems definitely superior to the other. Too many factors are involved to be discussed here, and they are, moreover, factors which will vary greatly according to circumstances.
Extreme Sensitivity
When we turn to the more sensitive type of set, however, which, although it must be capable of high-quality local reception, is intended not only for medium but for long-distance reception, we shall find that signal-frequency tuning and a moderate or low intermediate frequency become necessary. The sensitivity of such a receiver must be so high that internally generated noise becomes of importance, and in order to make the limiting factor the circuit noise rather than the valve hiss, the signal applied to the first valve must be as large as possible, and this calls for a tuned signal-frequency circuit rather than the aperiodic circuit of a single-span receiver. If we use signal-frequency tuned circuits, waveband switching becomes necessary, so that there is no reason, save that of second-channel interference against using a moderate or low intermediate frequency. Adequate selectivity for this type of receiver cannot be obtained at 1,600 kHz at the present time, so that this is an additional reason for reverting to the standard type of superheterodyne in cases where both sensitivity and selectivity must be very high. We have, therefore, to choose only between 110 kHz and 465 kHzfor the intermediate frequency. A receiver of this nature should be fitted with variable selectivity if it is to be capable of giving the best performance as regards both distant and local transmissions, and it has already been pointed out that this is more readily achieved at 465 kHz than at 110 kHz. When we take into account the smaller degree of second-channel interference With the higher intermediate frequency, it is clear that there is ample justification for its use.
|
