In the average circuit diagram the frequency- changer (FC) stage appears to be the most complicated part of the radio receiver. This is due mainly to the waveband switching arrangements, which are only too often drawn in a spidery fashion with lines crisscrossing each other all over the place. It is much more helpful initially to regard the FC as a single waveband device, as in FIG. 1, with the items in the broken line boxes being those which are concerned purely with tuning. Note that the two coils in the local oscillator circuit (extreme right) are shown as having variable iron-dust cores, and that the padder C11, is shown as a semivariable capacitor. This is to illustrate both possible arrangements; an actual receiver would employ only one or the other (fixed padder with variable core, variable padder with fixed core).
In this circuit the popular triode-hexode type of frequency-changer tubes (valves) is shown, but it is applicable to most other types. When a heptode (pentagrid) or octode is employed the first grid corresponds to the triode grid and the second grid to the triode anode.
FIG. 1 A simplified frequency-changer stage
Each additional waveband would have its own pair of boxes connected into circuit by the wavechange switch. The latter may develop poor contacts over fifty, sixty or even more years of use, resulting in crackling or perhaps the complete loss of one or more bands. With any luck a squirt of cleaning fluid may cure these problems. Certainly, this should be the first thing to be tried if a set refuses to work on one or all bands.
Let’s suppose you are confronted with a receiver that will receive no stations on any band, but is ‘lively’ from the IF amplifier onwards, as demonstrated by injecting an IF signal into the grid of the IF amplifier. Now transfer the signal generator output lead to the control grid of the FC valve. Does it pass through the set at increased volume? If so, at least the mixer section of the tube (valve) is all right. If there is a much reduced sound or none at all, check the voltages on the anode and screen grid of the mixer. If low or absent, look at the associated HT feed resistors and decoupling capacitors.
If the voltages appear to be on the high side, check the cathode voltage (if cathode bias is used). Little or none suggests that the tube (valve) itself is at fault and it should be tried by substitution. A very high reading, say 100v or more, suggests that the cathode resistor may be 0/c. Note that certain Ekco sets of the mid-1930s used large value wire- wound cathode resistors in connection with ‘silent tuning’ systems, and these are known to be vulnerable.
Let’s suppose now that an IF signal will indeed pass through the mixer at good strength, yet no stations can be heard. With a good aerial connected tune along each band in turn, listening carefully. Does a ‘rushing’ noise come in at the low frequency end of the MW band or the high frequency end of the LW band? If so, this is an indication that the LO (local oscillator) is not working. Measure the voltage on the triode grid (or Gb of a heptode or octode), expecting to find it several volts negative with respect to chassis. If not, check the triode anode or G2 voltage. A zero reading points to an o/c feed resistor or shorted decoupler, but a low reading does not necessarily mean the same. When the LO fails to oscillate it draws a much higher than usual HT current and thus the voltage will fall considerably.
Test the resistance of each winding of the LO coils for the non-working band(s). You may find that one or more has gone o/c, in which case it must be removed for close inspection. If you’re lucky the break may be visible and repairable without too much trouble, but all too often a subsection of Murphy’s Law will cause the break to be in the middle or at the inner end of the winding. At this stage, working on the ‘why me?’ principle, you may be wondering what makes a coil that has worked satisfactorily for all those years suddenly go o/c. In most cases it is due to a phenomenon called ‘green spot’, the development of a small speck of corrosion due to the wire having been handled by damp hands back in the dim and distant past.
You are now faced with the job of rewinding the entire coil. Try to remove the old wire carefully and take it up on a suitable spool. You may be lucky enough to find a single break and to repair it without too much difficulty, but multiple breaks may necessitate all the wire being replaced. In this case saving as much as possible of the old will give you a good idea as to how much you need for the rewinding job. Large diameter coils are fairly easy to rewind tidily, but you are unlikely to be able to do the same with small coils unless you are lucky enough to have a winding machine. Don’t worry about this unduly as untidy coils seem to work pretty well, all things considered.
The ‘Q’ meter mentioned in the section dealing with test instruments comes into its own for getting the inductance of a coil right to tune over a particular range of frequencies. If you are not lucky enough to own one, take comfort from the fact that coils with iron-dust cores have a reason able amount of latitude in this respect and a certain amount of trial and error should eventually win the day.
If you have to remove any metal screens to gain access to coils for repairs, always replace them before you commence realignment.
A case in point
Whilst this guide was being prepared the writer was asked to look at a Murphy receiver which had baffled several previous experienced but non professional repairers. A wartime model, it had only SW and MW bands, of which neither worked. The presence of the ‘rushing’ noise at the low frequency end of the MW band indicated that the LO was not working, and indeed several components had been replaced in this area, including the triode anode and grid resistors and the small coupling capacitors. Since this work demonstrably had been fruitless, attention was turned to the local oscillator coil windings. There should have been resistance readings back to chassis on the grid windings and both appeared to be o/c when measurements were taken from the wiper of the wavechange switch. Now, to paraphrase Oscar Wilde, to lose one winding may be attributed to accident but to lose two points to carelessness; in any case, the SW winding was of such a robust nature as to make ‘green spot’ highly unlikely. Although the wavechange switch contacts looked perfectly good, just in case readings were taken directly across the coils, which then proved to be in order. Gentle pressure on the switch contacts with a thin bladed screwdriver restored the connections and normal reception was obtained on both bands. It was impossible, however, to ‘set up’ the contacts to make them work reliably and in normal circumstances it would have been necessary to replace the switch. However, when the set was made wartime shortages must have caused Murphy to use any switches that were to hand, and this example had a complete set of spare unused contacts right alongside the faulty set, and to which all the connections could be transferred without difficulty. The owner of the set was suitably impressed, but the fact is that a little thought on the part of the previous would-be repairers might have suggested to them the same answer, and the moral is: don’t make assumptions from visual examinations. The wavechange switch, as mentioned, looked so good as to be above suspicion, but as Sherlock Holmes famously remarked, when all the possible solutions have been investigated without result the answer must lie with the impossible.
Faults on the aerial coils
Now let’s consider how to proceed if the LO stage is working correctly but only the very strongest stations can be heard, probably with considerable background noise or whistling. See what happens when the aerial is taken directly to the control grid of the mixer. Do the signals become louder or remain unchanged? These reactions point to either the primary or secondary of the HF transformers in the aerial circuit being 0/c. Measure each as for LO coils and if necessary carry out repairs using the same methods.
To return briefly to the wavechange switch, occasionally you may find a case of persistent noisiness, although all bands are present, which does not respond to cleaning. This may be due to HT tracking across the wafers. The RF circuits are nearly always switched on the ‘dead’ side of coupling capacitors, but in sets having a ‘gram’ position the HT to the FC and IF stages may be switched out to silence them. It is frequently found that the section of the switch used for this job breaks down, causing deafening results. The cure is to remove the HT leads from the switch and join them permanently. Not having the ‘gram’ position is no great loss as in practice it is seldom used.
When realignment is necessary
The most probable source of poor performance is interference with the adjustments. Owners of radios have been known to take them in for repair after ‘tightening all the loose screws’, as they put it! Sometimes the trimmers and cores are sealed with paint, which not only reveals any subsequent adjustment, but also indicates the original settings. You could try returning to these in the hope that this will restore the performance of the set, but all too often a complete realignment job will be necessary.
Before starting ensure that the IF alignment is correct. Now make sure that the dial pointer is at minimum frequency when the tuning capacitor is fully closed, adjusting its position if necessary by sliding it along the drive cord or runner, etc.
It is always best to follow the set manufacturer’s instructions as given in service manuals, but if these are not available the standard procedure to be described should give satisfactory results.
With the simplest medium and long wave receivers, and in the absence of specific instructions, it is advisable to start on the medium wave band, because some cheaper sets do not have a separate long wave oscillator coil, but load the MW one with an extra capacitor.
Connect the output leads of the signal generator to the receiver via a ‘dummy aerial’. Many generators are furnished with one of these when new, but there is always the chance of its having been lost over the years. FIG. 2 shows a suitable substitute that may be made up quite simply.
Connect an output meter or the AVO on a low AC voltage range across the secondary of the output transformer.
FIG. 2
Tune to 550m (545kHz) and adjust the signal generator to this frequency, noting whether or not it comes in at the correct position on the dial. Adjust gently the core of the local oscillator coil (or the padder as the case may be) to obtain this, then adjust the core of the aerial coil (if fitted) for maximum reading on the meter, keeping the output of the generator as low as possible to prevent the AVG from coming into operation.
Now retune the set and the generator to 200 m (1.5mc/s) and adjust the oscillator trimmer for correct pointer position and the aerial trimmer for maximum signal. In all probability this will affect the tuning at the other end of the band so repeat the entire operation several times until overall accuracy is obtained.
Switch to long waves and tune the set and generator to 2000 m (150 kHz). Adjust the local oscillator coil or padder to set the pointer, then the core of the aerial coil (if fitted) for maximum output. Retune to 1000 m (300 kHz) and adjust the oscillator and aerial trimmers for pointer position and maximum output. As for MW, repeat several times for overall accuracy.
Short waves present a bit more of a problem. The reaction to adjustment of cores or trimmers is much sharper, particularly at the higher frequencies and very slight movements are called for. Non-metallic trimming tools are essential if hand capacity effects are to be avoided.
The standard short wave coverage for receivers was approximately 16m to 52m. Suitable alignment points are 50 m (5 mHz) at the low end of the dial and 20 m (15 mHz) at the high end.
For band-spread SW ranges choose alignment points towards the top and bottom of each. As an example, the ‘41 m Band’ in the Murphy A126 covers from 40 m 50 m (7.5 mHz 6 mHz), so use 7mHz and 5.5 mHz.
Seal cores and trimmers with wax (as from an old capacitor) or quick-drying paint when the alignment process is completed.
Optical bandspread
An alternative to the purely electrical band-spread described above was to retain continuous coverage of, say, 16m to 50m, but to expand the dial artificially so that it had enough effective length to make tuning simple. Ferranti pioneered the idea in 1936 with their Magnascope dial. It consisted of a separate 180 inch scale mounted inside the set, with a small optical system to project the markings on to a screen above the main tuning dial. By this method the scale was made to be equivalent to one 6 ft in length. Murphy produced an updated version some twelve years later and quite wrongly claimed it to be an original idea, which provoked a very angry reaction from Ferranti. In both systems maintenance is confined to keeping the lamp, lenses, mirror, screen, etc., clean, but don’t scrub the scale!
As the actual frequency coverage is still the same as with ordinary receivers, use the 5 mHz and 15 mHz alignment points.
RF amplifiers
Sometimes known as signal frequency amplifiers, there are a number of advantages to having a tuned RF amplifier preceding the frequency-changer stage. The extra gain makes the set more sensitive and improves the signal to noise ratio, but there is another useful consideration. Conventional superhets rely largely on their IF stages to provide selectivity, and there is often only one tuned circuit prior to the FC, which in itself may be relatively inefficient. A good RF amplifier immediately at least doubles the number of pre-FC tuned circuits, and receivers intended for serious short wave listening may have two or more RF amplifiers. The stage gain diminishes fairly rapidly as the frequency increases, but this may not be so serious a consideration as the maintenance of good selectivity
Sometimes the RF amplifier was called a ‘pre selector’, especially when it took the form of a separate add-on unit.
As far as fault finding is concerned an RIF stage resembles an IF amplifier, frequently employing the same tubes (valves) type and similar component values. There has to be an extra section or two on the wavechange switch to cope with the various tuning coils, and there will also be another set of trimming capacitors plus, of course, the required additions to the gang capacitor. Alignment again follows the usual rules of cores at low frequencies, trimmers at high.
Rejectors
We spoke earlier of aerial rejector circuits (‘wave traps’) intended to counteract swamping by powerful stations. With the various changes in the wavelengths of stations over the years it could well happen that the wavetrap might be reducing the strength of a weak but desired station, and slight readjustment may be necessary. Usually the adjustment core or trimmer needs to be altered only very slightly to achieve the desired result.
Another type of rejector is designed to reject signals at IF. A signal at this frequency should be injected into the aerial socket, strong enough to break through the tuned circuits and be heard in the loudspeaker. Tune the rejector for minimum sound.
On completion of adjustment, rejector cores or trimmers should be sealed.
Image rejector
The problem of the ‘image’ is largely confined to superhets with low IFs. The local oscillator of a superhet may be designed to operate either above or below the signal frequency, but in practice it is usually higher, i.e. the sum of the RF and IF. Suppose that a set with an IF of 125 kHz is tuned to a station on 1000 kHz. The oscillator has to run at 1125 kHz, and should there be a strong station on
1250 kHz, capable of breaking through the aerial tuning, it too will beat with the local oscillator to produce 125 kHz. The result will be either crosstalk between the two stations or an intrusive whistle of varying pitch. This is not all. If the receiver is tuned towards the low frequency end of the band, when the pointer reaches the 750 kHz mark the local oscillator will be running at 875 kHz, and thus capable of combining with the 1000 kHz signal once again to give the IF In this case the station will be heard again, although at lower volume, and the false point of reception is called an ‘image’. Many early superhets had image rejectors fitted in the aerial input circuits and adjusting them has to be carried out in accordance to the specific instructions of the set manufacturer. If these are not to hand, it is better to leave the rejector well and truly alone, particularly some of those used in EMI receivers which involved the physical movement of coils with relation to each other.
An image is always twice the IF away from the genuine station, and so raising the former to 450/470 kHz alleviated the problem greatly, at least as far as medium waves are concerned, by spacing the wanted and unwanted signals more widely apart.
Adjusting ferrite aerials
Receivers manufactured from the mid-1950s onwards were often fitted with ferrite rod aerials. The windings are not always adjustable, but where they are the rule is the same as for any tuning coil:
the core to be set at the low frequency end of the band. In fact, the accepted method is to slide the appropriate coil along the rod to the point of maximum volume and to seal it into position with wax. Broken rods need not be replaced if the pieces can be easily glued together, because magnetically, joints don’t matter.
Faults on the tuning capacitor
Crackling or intermittent tuning can be due to shorting of the main tuning capacitor, or occasion ally bad earthing of the moving vanes. The latter is sometimes caused by dirty or broken spring contacts on the moving shaft, resulting in earthing taking place through the bearings instead of directly. Even when the earthing arrangements are in order, lack of lubrication on the bearings can result in slight crackling as the set is tuned, particularly at the higher frequencies. Shorting of the vanes might be caused by conductive dust getting between them. It has been known for very enthusiastic amateur repairers to remove the capacitor from the set and to clean it in a dishwashing machine! I will refrain from describing the professional’s most effective method since it is a little hazardous and simply suggests a high pressure air blast.
Another explanation for shorting capacitor plates is that some enthusiastic but incautious person has adjusted them, either directly by bending them, or via the spacing screw which is to be found at the rear of the capacitor. Make sure that this is set properly, so that the fixed and moving vanes are equally spaced. There should be a locknut to ensure that the screw stays put. If it is found that reception ceases at certain parts of the tuning range, find the exact spot with the aid of an ohmmeter, as follows.
Switch to long waves, where coils connected across the capacitor will have the least effect on the meter readings, and place the test prods between earth and each set of moving vanes in turn. Rotate the tuning knob slowly until a sharp drop in resistance is shown on the meter. Knowing now which set of vanes is at fault, and where, it should be possible to clear the short circuit by gentle bending of the moving, not the fixed, vanes. The former, incidentally, were usually slotted on the rounded edge so that precise adjustments to alignment could be made at the factory by slight bending. Don’t confuse this intentional warping with out and out mishandling.
Tackling broken dial drive systems
The dial drive probably has been the one item in radio sets responsible for the greatest amount of sweat, tears and profanity on the part of service engineers, so be warned.
The simplest sets have but a knob fitted directly on the tuning capacitor shaft. The next up the scale have a small epicyclic reduction gear, and then come the complicated ones!
Drive cord setups can be divided into two broad classes: (a) Philips, (b) The Rest. There must have been someone at Philips from the 1930s to the 1950s who was a cross between a mechanical genius and a sadist. Not content with intricate cords for the capacitor, he also used fiendishly clever Bowden cable and steel wire drives for the tuning dial. To attempt to describe one of these mechanisms is only slightly more difficult than expounding nuclear theory on the back of a postcard, and to restring one calls for all the patience attributed to the prophet Job.
Approach a Philips dial drive warily. Unless the cord or the wire is actually broken, leave well alone, and merely lubricate the capacitor bearings and the runner which carries the pointer, to relieve the load on the drive as much as possible. If you cannot obtain steel wire you might like to explore the chances of using fine brass wire as used for picture hanging.
Whatever you do, before you start to restring a Philips (or any other) dial, make quite sure that you know which way the pointer should travel with respect to the dial markings. There is nothing more likely to make a strong man weep than to spend hours installing a cord drive only to find the pointer moves in the wrong direction.
It is to be hoped that the entire tuning drive system is mounted on the chassis of the receiver so that it may be attended to with the latter out of the cabinet. Unfortunately, all too often Philips made this impossible by mounting the dial itself and all the pulleys and rods, etc., inside the case which makes matters very difficult indeed. The very first drive cord ever tackled by the writer, at the age of fourteen, was just such a one as this and it took several evenings of solid hard work eventually to figure it out. However, it did give the satisfactory feeling that nothing ever could be as bad again.
Some common types of dial drive
It is impossible to enumerate all the types of dial drives, but at least there is a basic pattern from which to work, shown in FIG. 3. Some more expensive models featured a flywheel on the inner end of the tuning spindle to facilitate tuning from one end of the dial to the other.
Should the tuner knob turn without effect, and yet the cord drive appear to be intact, check the tightness of the grub screws in the centre of the drum, and for the presence of oil on the narrow part of the spindle where the cord runs. The spring itself may have lost tension over the years so either replace it or stretch it further to a convenient point of the drum. Incidentally, when withdrawing the chassis of a radio from the cabinet for any reason, have a look to see if the drive cord has to be disconnected from the dial pointer first. This simple check can save a broken cord and much hard work!
If the cord has broken but is still in position around the drum and various guides, do not disturb it until you have made a sketch of how it is lying and of the number and direction of turns round spindles, etc. If the worst has happened, and a tangle of ancient drive cord is found lying in the bottom of the cabinet, try to find all the broken pieces, both to reclaim the spring(s) and in order to get an idea of the length of new cord required, thus avoiding waste. Incidentally, good flax stranded fishing line is just right for restringing cords. It was, in fact, officially recommended by EMI back in the 1930s.
Before starting to restring the drive, ascertain the direction in which the pointer has to move. When the capacitor is fully closed the pointer needs to be at the low frequency end of the dial and this varies from set to set — it may be on the left or right according to the whim of the designer. Knot one end of the cord to the spring after threading it through the hole or slot in the drum. There may be more than one hole, so use the one which allows the greatest amount of cord to go round the periphery of the drum.
Where the cord passes over small pulleys, ensure that these are free to turn, oiling sparingly if necessary. Do not be tempted to put more than two turns around the tuning spindle as it will not increase the available grip, but merely cause piling up and binding of the cord. The loose end of the cord should be passed around the drum in the direction opposite to the initial wrapping, so that as it is drawn off in one direction, it is taken up in the other. Finally pass the end back through the slot and knot it to the spring as well. The latter should be stretched to hook into one of the holes usually provided.
The drive should now operate correctly, but note that the initial stretching of the cord will probably have to be taken up by altering the spring tension.
Permeability tuning
This was discussed earlier. In its practical form the tuning coil is made long and thin, with a movable core mounted on drive cord passing through it. Permeability tuning was used in a few table radios just after the war, but is more likely to be found in FM receivers and car radios. The only fault normally encountered is restriction of the tuning range caused by the jamming of one or more of the cores in its coil. A smear of silicone grease will usually effect a cure.
Dial pointers
Pointers are of two basic types, those that are actually carried by the cord, and those which ride on some kind of rail and are merely pulled along. In the latter case make sure that it can travel freely by applying a little light oil to the rail.
To set up the pointer correctly, examine the dial glass or the backing plate for a datum mark and slide the pointer to coincide with it. If such a position is not marked anywhere, set the pointer so that it travels equally just beyond the upper and lower limits of the dial markings. If its previous position has been incorrect, it may be necessary to slightly realign the set to obtain the correct readings for wavelengths.
A word of warning regarding the cleaning of glass dials. In some sets the paint used to mark them was such poor quality that even water would remove it, so it is best to experiment first with something of no consequences, e.g. the maker’s name, which nearly always appeared in an unobtrusive place. If this resists a good hard rub it should be all right to use a mild detergent on the rest, but be very careful. Far too many dials end up as plain glass thanks to overvigorous cleaning.
FIG. 3 Typical arrangements for cord drives