One of the many strange facets of vintage radio collecting is that one of the most sought-after sets which commands prices of up to $700 should happen to be what was just about the cheapest set on the market when it appeared in 1936. This was the Philco Model 444 ‘People’s Set’ ( FIG. 1), and the resemblance of its Bakelite cabinet to the front of the German Volkswagen or ‘People’s Car’ may or may not have been fortuitous. It cost 6 guineas, then about two weeks’ wages for a manual worker and half the price of many con temporary four tubes (valves) mains sets. Its performance was adequate on the long aerials used in those days and the fact that large numbers of them are still in working order more than sixty years later attests to the fact that they certainly were not shoddily made. In fact, the circuitry of the 444 serves to illustrate many of the conventional features described in earlier sections, plus some unconventional ones including an odd AVG sys tem that reverses the normal rules.
The frequency-changer stage uses a 6A7 heptode (pentagrid) to produce an IF of 451 kHz. This is coupled to the IF amplifier (a 78E) by a normal type of transformer, but the second IFT was, for some unknown reason, untuned. This can scarcely have been on economy grounds since the cost of a pair of trimmers could not have been all that much. The lack of tuning reduces severely the gain of the stage which, as this is a ‘short’ superhet appears to have been a perverse piece of design.
A double-diode-pentode acts as detector and output valve, and since no suitable 6.3 V heater type was then available Philco had Mazda make a special one just for the 444. Known as a PenDD61, it was effectively an AC2PenDD with a modified heater which Mazda pretended never happened, for nowhere can it be found in that firm’s technical or sales data.
The power supply stage uses a standard type of mains transformer with a directly heated full-wave rectifier (an 80) and HT smoothing by the field winding of an energized loudspeaker. Note that a 0.015 uF capacitor is wired from the live side of the mains input to the chassis of the set to suppress modulation hum; as the capacitor is connected on the mains side of the on/off switch it is in circuit at all times. Even when new and in perfect condition this capacitor would have passed sufficient cur rent to give a nasty tingle to anyone touching the chassis, and after the passage of time it must be regarded as a definite hazard. Incidentally, many of the capacitors in the 444 (as in other Philco models) were molded into what appear to be four large terminal blocks with three soldering tags each. Three of these blocks contain pairs of capacitors (G2 and C6, the AVG decouplers, C3 and G7, the cathode decouplers for the 6A7 and 78E, and G9 and Gb, the IF by-pass capacitors on the detector load) and the other holds just G16, the mod. hum capacitor. They are notorious for developing leaks but only G2, G6 and C16 are likely to cause trouble.
FIG. 1 The circuit for the Philco Model 444 ‘People’s Set’
In earlier discussion of ‘short’ superhets it was mentioned that the AVG was usually delayed heavily to permit maximum sensitivity on weak signals. Following American practice, the AVG in the 444 was tapped off from the detector load without benefit of delay, which must again count as being deliberately perverse. We now need to look at the AVG in detail as its curious circuitry can easily baffle newcomers when things go wrong.
It will be seen from the circuit that the standing bias on the 6A7 is provided by a 700 ohm cathode resistor. This provides about 7 V bias which is over twice the usual amount. Likewise, the 800 ohm cathode resistor of the 78E provides about 6 V bias which is again far more than is usual. The reason for the high cathode bias figures is that the detector load resistor, from which the AVG is derived, is returned to the cathode of the PenDD6I, which itself is 5 V positive with respect to chassis. Thus the AVG line, and consequently the grids of the 6A7 and 78E, is itself 5 V positive in the absence of a signal. This offsets the cathode bias on the two tubes (valves) so that the actual grid bias they receive is, respectively, 2 V and 1 V. Just why Philco went to all this trouble when it would have been far easier and better to have used a proper delayed AVG system with a separate diode is something we probably shall never know, but what is certain is that the strange design introduces another perversity almost guaranteed to fool the unwary.
Many times in this guide have we spoken of the undesirability of permitting positive voltages to get onto the control grids of tubes (valves). In the 444 we have to turn this upside down because the set simply will not work well unless there is positive voltage on the grids of the 6A7 and 78E. Without that, the standing bias on them reduces their gain to the extent where reception is restricted to only very powerful stations. This condition arises when the two AVG decoupling capacitors G3 and G6 develop leaks; as the AVG line is fed via a 2 M–ohm resistor only a small amount of leakage will drop the positive bias sufficiently to spoil the performance of the set. After the two capacitors have been replaced it is as well to check the value of the 2 M-ohm resistor in case it has gone high.
Replacing the output valve (tube)
The present writer had been over forty years in the radio trade before he saw a brand new PenDD61 in a box, because Mazda discontinued making them as soon as the production run of the 444 and its wooden-cased derivative, the Model 269, ceased. To digress for a moment, in view of the current worshipping of Bakelite artifacts, it is interesting to note that in 1936 it was very properly regarded as a cheap substitute for wood, the price of the 269 being half as much again as that of the 444.
Back in the late 1940s there were many 444s still in use and replacing defunct PenDD6ls was a common problem. The easiest way of achieving this was to fit an AC2PenDD and to drop the 6.3V heater supply down to 4V to suit it. To do this job a resistor of just over 1-ohm was required, a value not then easily obtainable commercially, and it was made up by cutting a length of resistance wire from an ordinary electric fire element. To measure accurately 1 1ohm with an ordinary meter is difficult so it was calculated by length. A 1 kW fire element has an overall resistance of about 50 ohm, so if it is stretched out to a length of 20 inches each inch represents about 2.5 ohm. Thus half an inch of fire element would give the required resistance plus a short length at either end for connections. It may sound crude, but it worked.
The Ekco BV67 ( FIG. 2)
Again appearing in 1936, this receiver dispensed with an HT battery and employed a vibrator pack, thus providing an excellent example of how a synchronous vibrator operates. It also illustrates the wiring of filaments in series—parallel, the use of a separate local oscillator in the frequency-changer stage and how an indirectly heated double-diode is used to provide delayed AVG.
The power for the BV67 was provided by a pair of heavy-duty 2 V accumulators connected in series to give 4 V. This was used directly to supply the vibrator pack, which provided a smoothed HT output of 135 V. Five tubes (valves) were employed, two VP2 RF pentodes, a PM1HL triode, a 2D2 double-diode and a PM22D output pentode. The first three mentioned formed one group and the last two another group, each with the tube (valve) filaments wired in parallel. The two groups were then wired in series across the 4 V LT supply. The combined filament currents of the first group amounted to 460 mA and those of the second group to 390 mA, so a shunt resistor was wired across the latter to make up the difference.
The frequency-changer stage used one of the VP2 pentodes as a mixer with the PM1HL operating as local oscillator, its output being coupled into the suppressor grid of the pentode. The IF amplifier was conventional. The cathode of the 2D2 double-diode was returned to the positive side of the 4 V LT supply and the anode of the AVG diode was returned to chassis, thus providing a 4V delay bias. To prevent this bias from reaching the detector diode anode its load resistor also was returned to the 4 V positive line.
Bias for the output pentode was provided by treating the filament as a virtual cathode. Since the filament was between 2 V and 4 V positive with respect to chassis, by returning its control grid resistor to chassis the grid received an effective bias of about 3V.
The BV67 was a very interesting experiment, but it failed to set a popular trend, probably due to the expense of buying and maintaining four accumulators (two in use, two on charge) and to the unavoidable hum from the power pack, which must have been obtrusive at low volume levels. However, it is worthy of study for the insight it gives into the various techniques it employed.
The Philips V7A
This was perhaps the most ill-conceived design, if that is the correct appellation, of any radio receiver ever. Its circuit was that of a fairly conventional four- tubes (valves) plus rectifier superhet but it had no chassis, all the various components being fitted around the inside of a Bakelite cabinet. Some were bolted into position but others were held in by no more than pitch. The V7A must have been extraordinarily difficult and expensive in labor to produce and it presented service engineers with great problems when repairs were required. Fortunately the idea was very soon scrapped but even so, fair numbers of V7As remain in circulation to plague set repairers to this day.
The ‘monoknob’ receivers
Again a brainchild of Philips, a series of sets under that name and of its subsidiary Mullard appeared in the late 1930s. The outstanding feature of what again were reasonably conventional superhets was the use of a single control knob to change its wave bands, to rune it and to adjust its volume and tone. In fact it was a bit of a trick, because what appeared to be one large knob in the bottom centre of the cabinet was actually two, working concentrically. The front section worked the tuning and the rear section the wavechange, whilst the complete assembly could be moved up and down and side to side to control, respectively, the volume and tone. It all depended on a complicated system of Bowden control cables which again must have been costly to produce and install, and which are extremely difficult to repair. By about 1938 the idea had quietly been dropped in favor of a new Philips specialty.
Fig. 2 Circuit of an Ekco BV67 receiver
Philips and Mullard pushbutton receivers
Again ignoring the principle that conventional designs had become so because they had been proved in practice to be the best, Philips decided to produce a radical new tuning capacitor for its pushbutton sets. In this the plates were circular in section, graded in overall size and mounted con centrically. The spacing was arranged to permit the moving plates to slide in and out of the fixed plates, so that the tuning action depended on lateral and not rotary movement. This capacitor was used in conjunction with mechanical pushbuttons which simply pressed the moving plates into the fixed plates a sufficient distance to tune in the desired station. This in itself sounds to be a simple and sensible design but unfortunately it was allied to a desperately complicated manual tuning system which had to change the rotary movement of the knob into lateral movement of the capacitor. This in turn was linked to a dial pointer drive mechanism which was non-linear in action and evil to set up. Since the dial has to be disconnected for chassis removal be very wary about taking on a repair job on one of these sets!
The wartime civilian receiver
The shortage of domestic radio receivers during the Second World War led to the introduction, in 1944, of what officially was called the ‘Wartime Civilian Receiver’, the official title not revealing that there were to be two versions, one for mains, the other for batteries. The public soon dubbed these as ‘utility sets’, bracketing them with all the other various household articles that were produced under the term ‘utility’ to spartan Government standards. Although the sets were not successful from the sales angle, it is well worth examining them in detail, particularly the mains version. The latter has some ingenious design points and illustrates the use of a Westector as detector and provider of AVC. Both mains and battery versions are totally unique in being the only domestic sets ever to have had their full technical specification published, enabling its performance to be checked against known criteria rather than against advertisement hyperbole.
In 1943 the Government authorized, via the Board of Trade, the Radio Manufacturers Association to set up a committee to design and build a standard receiver that was simple, in order to use a minimum of components, but capable of an adequate performance in wartime conditions. What emerged was the ‘utility’ set in its two versions. It might easily be imagined that the combination of urgency and the need for economy could have resulted in a ‘cheap and cheerful’ approach but on the contrary the design specifications devised by the RMA were extremely detailed and stringent. They covered sensitivity, selectivity, overall response, AVC operation and IF rejection. They were as follows.
For the mains version:
• Sensitivity: to be not less than 325 mV @ 220 m and 625mV @ 500m for 50mW output measured at the loudspeaker terminals.
• Selectivity: bandwidth not to exceed 11 kHz @ 50% response and 21 kHz @ 10% response.
• Overall response: to be not more that 7dB down @ 100 Hz or 9dB down @ 4 kHz with respect to the level at 400 Hz, to be measured on a resistive output load and using an RF input of Y mV modulated at 30%, applied to the Al (direct aerial socket with the volume control adjusted to give 50mW output @ 400 Hz.
• AVC threshold: the AVC to be delayed so that its operation commences when the output is approximately 1 W on a signal with a modulation depth of 50%.
• IF rejection ratio: not to be worse than 5:1 at any point on the dial.
For the battery version:
• Sensitivity: not less than 300mV @ 200m and 600 mV @ 500 m for an output of 50mW across the speech coil terminals with an HT voltage of 120V (rather surprisingly, better than for the mains set).
• Selectivity: as for mains version.
• AVC threshold: not specified.
• Overall response: not more than 10 dB down @ 100 Hz or 14dB down @ 3kHz with respect to the level at 400 Hz, under the same input/output conditions as for the mains version.
• IF rejection: as for the mains version.
In addition, the oscillator section of the frequency changer to continue to operate with the set fed from a 60 V HT battery (i.e. 50% down on normal) via a 2.2 k-ohm series resistor.
The cabinets used by all manufacturers to be as nearly as possible of the same appearance. Standard drawings to be prepared by the British Radio Cabinet Makers’ Association and kept at the RMA offices. The tuning scales to be finished in the standard manner and to have the same appearance as the prototype.
General quality: ‘In view of the difficulty of producing a sufficiently detailed specification to cover such points as loudspeaker performance or the durability of workmanship of the receiver, the RMA wishes to draw the attention of manufacturers to the fact that the sets will be so coded that defective apparatus can be traced to its source, and it is therefore in the interests of each manufacturer to adhere to the spirit of the specifications.
In other words, you have been warned!
Meeting the specifications
The technical design committee appointed by the RMA came up with what was a near conventional four tubes (valves) superhet for the battery set and a highly individual three- tubes (valves) plus rectifier superhet for mains. It is doubtful if much argument was required over the battery set, which in fact bore more that a passing resemblance to the Murphy B89 of 1940, shorn of its long waveband and variable tone control. On the other hand, the mains set must have provoked a great deal of discussion, much of which must have been due to the restrictions imposed on what tubes (valves) types might or might not be used. The committee had concluded that the required performance could be obtained with a ‘short’ superhet, which in peace time would have meant the use of a frequency changer, an IF amplifier and a double-diode- output pentode. Unfortunately, whilst the first two types of tubes (valves) were in quantity production, the last, having no military application, had been abandoned ‘for the duration’ and the Board of Trade would not permit its manufacture solely for the Civilian Receiver. To get around this the committee opted for the ‘Westector’ miniature metal rectifier as detector, driving an ordinary high slope output pentode, which type was also in plentiful supply. It would be the first time that the Westector had been used on a large commercial scale since the early 1930s. Its adoption here, whilst solving one problem, would cause others to arise, as we shall see.
Valves (tubes) and secrecy
The tube (valve) used in both battery and mains versions were to be as anonymous as the sets themselves and carried arbitrary type numbers devised by the British Tube (valve) Makers Association to cover groups of equivalents made by individual members, with the final figure of each number indicating the actual manufacturer, as follows: 1, Cossor; 2, Mazda; 3 Ferranti; 4, Osram; 5, Marconi; 6, Mullard; 7, Brimar. It seems odd that even at that critical time the illusion that Marconi and Osram tubes (valves) were unrelated, instead of coming from the same factory, had still to be fostered. The types employed in the battery version were known as the BVA172, BVA142, BVA132 and BVA262; due to their unique bases they were patently the Mazda series TP25, VP23, HL23DD and Pen25 and as the numbers reveal were made only by that firm. The prototype mains set used the BVA273, BVA243, BVA264 and BVA211, or in other words the Ferranti 6K8G and 6K7G, the Osram KT61 and the Cossor 431U. Oddly enough, the BVA273 was not included in the list of tubes (valves) used in production models, which were (including equivalents):
BVA274/5/6 (BVA277 was listed but not produced); BVA243/6/7; BVA254/5/617; BVA211/ 4/5/6/6. A list of commercial equivalents is reproduced in Figures 4 and 5.
An opportunity missed
Other design aspects intended to minimize the use of components in short supply included the single medium wave band and the use of resistance capacity smoothing for the HT, the latter ‘to avoid the use of large quantities of copper wire in . . . a smoothing choke’. It seems rather curious why this aim was not pursued further and more profitably by making the set AC/DC and eliminating the much more copper-hungry mains transformer. The cabinet, like that of the battery version, was to be of plain wood with no identification of any particular manufacturer.
The design of the prototype in detail
The aerial input was to the primary of a simple iron-dust cored RF transformer via a 500 pfd capacitor, with another of that value in series with the bottom of the winding and chassis. The bottom of the winding was also strapped to that of the secondary, which provided a mixture of inductive and capacitive coupling, the first at the high end of the band and the second at the low end. The top of the secondary was connected directly to the control grid of the frequency changer (Vi, BVA273) and the bottom taken via a surprisingly low resistance (680 ) to the AVC line. The local oscillator also used an iron-dust cored RF trans former with the top of the tuned winding taken directly to the grid of the triode section of Vi, with another 500 pfd from the bottom end to chassis (shunted by a 47kI resistor) acting as both grid capacitor and padder. The anode winding was supplied from the HT via the same 6.8 k resistor which supplied the screen grids of the frequency changer and the IF amplifier (V2, BVA 283). The grid of the latter was coupled to the anode of the triode-hexode by an iron-dust cored transformer operating at 460 kHz.
A second iron-dust cored transformer coupled the anode of V2 to the detector (MR1, WX6). The Westector has different operating conditions to that of a thermionic diode, requiring a lower impedance input and a small standing current flowing through it to ensure conduction at low signal strengths. The first was achieved by making the IF transformer have a step-down ratio of 1.5:1, which had the unfortunate side-effect of lowering the ‘Q’ and thus the gain of the IF amplifier; to counteract this some close-wound coupling turns were added to the primary winding. The way in which the standing current through MR1 was achieved will be discussed later when the AVG system is described.
The AF from the detector was filtered by a 47 k-ohm resistor and 100 pfd capacitor before being coupled via a blocking capacitor to the top of a 1 M1 volume control. The tap was taken directly to the grid of V3 whilst the bottom end was returned to the junction of the two cathode resistors in order to establish a standing grid bias of 4.5 V. The output tubes (valves) was conventionally coupled to the loud speaker with a fixed tone control capacitor across the primary of the output transformer.
The power supply stage employed a mains transformer with only two tappings to cover 195/250 V inputs, and a full-wave rectifier (BVA211). As mentioned earlier, resistance-capacity smoothing was used with a resistor between the HT— and chassis providing a bias voltage of approximately 10 V. The mains on/off switch was not conventionally ganged to the volume control but took the form of a single-pole toggle switch mounted at the rear of the chassis.
The AVC system
We have already seen that in any short superhet where the output tubes (valves) grid is driven directly from the detector it is essential to delay the AVG to prevent its coming into operation until sufficient voltage is developed across the detector load to give full volume. In normal practice, when a double-diode-output pentode is employed one of the diodes is used as AVG rectifier. It would have been possible technically in the utility set to employ another Westector for the job but this was ruled out by the Board of Trade, possibly because of supply difficulties. This left as the only source of AVG bias the DG existing at the top of the detector load resistor, which, had it been used direct, would not have been delayed and would have restricted severely the performance of the set. This problem was solved by some rather complicated circuitry which included pressing the suppressor grid of V2 into service as an effective diode. The circuit needs to be studied carefully in conjunction with the following description (incidentally of three contemporary descriptions studied during research for this topic none appears entirely to be correct).
It will be seen that the negative end of the bias resistor R13 is taken to R8 (4.7M-ohm) and via this resistor to R7 (1.5 M-ohm), the other end of which is connected to the suppressor grid of V2. The latter is also taken via R6 (1 M to the top of the detector load resistor R9 (330 k-ohm) which, as mentioned earlier, returns to the cathode of V3. Thus the four resistors effectively form a potential divider between a point that is at —10V and one that is at +12.5 V — but with the important addition of the path to chassis provided when the AVG ‘diode’ is conducting, as occurs under no signal conditions. It may then be regarded as a fairly low resistance, so that the centre of the potential divider is ‘clamped’ to chassis and effectively split into two sections, the upper at between +12.5 V and chassis, the lower at between —10V and chassis. The voltage at the junction of R6 and R9 will be approximately +9v which is applied to the anode of MR1 via R5 (47k). Meanwhile, MR1 cathode returns to the junction of R11 and R15 where a voltage of 4V with respect to chassis exists. Since the anode is more positive than the chassis the required steady current will flow through MR1.
The lower half of the potential divider provides approximately —2.5 V at the junction of R7 and R8 and this is applied to V2 and Vi as standing bias. When a signal arrives at MR1 a negative voltage will be developed at its anode that is in opposition to the positive voltage applied from the junction of R6 and R9, but until it exceeds —9 V it will not have any effect on the AVG system. However, as soon as this critical delay voltage is exceeded the ‘diode’ anode will become negative and it will cease to conduct, removing the ‘clamp’ to chassis. The negative voltage is then passed on as AVG bias to Vi and V2, but it is important to bear in mind that it will not reach their grids at the same potential, due to the dividing effect of R7 and R8. In fact, the voltage on the actual AVG line will be approximately 0.4 of that at the ‘diode’ when the latter is at —9.5 V and approximately 0.5 when it is at 15 V, with the proportion rising as the ‘diode’ becomes more and more negative.
The battery version
As mentioned earlier, this bore a strong affinity to the Murphy B89. It is a fairly straightforward four- tubes (valves) superhet using a triode-pentode frequency changer, pentode IF amplifier, double-diode-triode detector and AF amplifier, and pentode output. One feature worth examining in detail, however, is the way the AVG bias is obtained. For some reason instead of using the second diode of the DDT as an AVG rectifier, bias was drawn via R7 (2.2 M( from the top of the volume control (R9, 1 Me), which also acts as the detector diode load. The frequency changer and IF amplifier also derive their standing grid bias from the AVG line, which is returned via R8 (3.3 M-ohm) to the HT— line, at a point approximately —3 V with respect to chassis. To prevent this voltage reaching the detector diode, the bottom of the volume control is returned to LT+, thus making R7, R8 and R9 into a virtual potential divider. Note that since the LT— is taken to chassis, the total voltage across the divider will be 5 V and not 3 V. The tube (valve) thus receive approximately —2.7 V standing bias from the junction of R7 and R8, whilst the diode anode receives about +0.7V, so no AVG bias will be developed until the signal strength exceeds this.
Breaking the mold
The conception of every manufacturer (140 in all) involved in the production of the Utility Receivers building absolutely identical sets was hardly likely to have been achieved in practice and in fact many of them modified the design to a certain extent.
Whilst some changes are unlikely to puzzle anyone working on the sets, others could well cause problems, especially to those with lesser experience of them. FIG. 3 to 27.6 show a complete list of all modifications for both mains and battery versions.
Likely problems in servicing
Mains receivers
Many years of experience in servicing these sets suggest that generally speaking trouble is confined to the detector and power supply stages. It cannot be emphasized too much that the voltages around the unusual combination of metal rectifier and suppressor grid for detection and AVG are critical and that discrepancies can ruin the sensitivity of the receiver. Particular attention needs to be paid to the 4.7Mfl resistor which feeds negative bias to the suppressor of V2 and which is susceptible to going very high. The steady negative voltage developed across the Westector appears on G24, the coupling capacitor to the volume control, and if this should leak there is a danger of the output tubes (valves) receiving unwanted negative bias, in spite of the bottom of the control being returned to the junction of the bias resistors R11 and R12. One consolation is that although this may cause V3 to become less sensitive, at least it will not make it draw excess anode current as in most sets. To offset this it has been found that certain output tubes (valves) have had a tendency to leak internally between cathode and grid and thus bring about overloading. Note that the voltage reading at the cathode (nominally 13 V) does not represent the actual grid bias; this should be measured between the junction of R11 and R12 and the cathode. It should be a shade over 4 V and any discrepancy should be investigated, commencing with trying a replacement valve. Another item worth checking is the bias capacitor G13, by means of shunting another across it.
Apart from the normal chance of one or more of the smoothing capacitors losing its capacity there is always a possibility that someone in the past may have replaced them incorrectly, for it is all too easy for the uninitiated to ignore negative bias and return the reservoir to chassis instead of to the centre tap of the HT winding on the mains transformer.
= ===
FIG. 3 The circuit, component values (with tag strip details) and tubes (valves)
voltages for the battery version of the utility set.
FIG. 4
FIG. 5
FIG. 6 The circuit, component values (with tag strip details) and tubes (valves)
voltages for the mains version of the utility set.
== ==
The mains transformers themselves, of whatever make, have a good record for reliability. This is probably due to the fact that standard types were employed of the same ratings as those used for much larger sets and consequently were only lightly loaded: the total HT current was under 45 mA and the heater current not much more than 1 A, varying slightly according to the tube (valve) types fitted. In general these figures were far lower than those applying to most pre-war table receivers.
In not a few instances the mains on/off toggle switch has been found to have become ‘sticky’ and not throwing over properly. If this should be the case it is worth trying some WD4O down the sides of the toggle.
Realignment of the RF and IF circuits is seldom called for unless it has previously been disturbed or if tests reveal that low sensitivity is due to gradual loss of alignment. Should the IFTs be only a little off their correct frequency only slight adjustment of their cores (or trimmers) ought to suffice, but if there is a large discrepancy consideration should be given to the possibility of one or more of the capacitors shunting the windings having changed in value. This has not proved to be a serious problem with the utility set but it is something worth bearing in mind. The RF alignment is simple, with only the single wave band to worry about, but do study the list of modifications for possible divergences from the standard procedure.
The battery version
It is likely that anyone operating one of these set nowadays will be using a battery eliminator to provide the HT voltage. Note that although the suggested HT battery for the set was the conventional 120V type, it was stated that any between 108V and 150V might be used as available.
The tube (valve) used in the utility set, which were exclusively made by Mazda, were some of the best battery types ever made in this country and will give excellent service for a very long time if treated considerately. Note that C14 (0.5 uF) is connected from the HT line to chassis, following the decoupling resistor R11 (10k). If it should happen to leak the HT to all tubes (valves) electrodes save those of the output tetrode will be reduced, and R11 may itself be damaged; therefore check both these components if either one needs to be replaced.
Credit must be paid to the designers for using as original a 1000VW capacitor for the coupling between V3 anode and V4 grid. As V3 anode works at only just over 50 V, the chances of the capacitor leaking ought to be remote but one never knows! Check it if the output tubes (valves) appears to be drawing excess current, bearing in mind that there should an actual negative voltage on the grid via the stopper (R14, 100 k) and the leak (R13, 330 k). The bias resistor (R15, 390 ohm) should also be checked at the same time.
Alignment of the IF and RF stages is straightforward as far as most of the makers chassis are concerned, but as for the AC version, check through the list of modifications especially those dealing with Pye-made sets.