The picture on the left shows the underside of the radio with its power supply and audio output stage on the chassis extension.
The set uses eight miniature B7G all-glass valves as follows:-
RF Stage 6BA6, Frequency Changer 6BE6, IF Amplifier 6BA6, Detector, AVC and Audio Amplifier 6AV6, Audio Phase Inverter 6C4, Push-Pull Audio Output two 6AQ5s. A 6X4 is used for rectifying the rough HT from the vibrator unit.
The set, unlike modern transistor types, can be used with either a positive or negative earth without any changes, but that is where the advantage of this old valve set finishes. Power consumption from the 6-volt (not 12-volt) battery is a huge 4.5-amps. Not fluctuating with audio output but a solid 4.5-amps whether giving a whisper or a roar. The valve heaters of course are largely responsible for the current drain. A total waste of energy compared with modern semiconductors.
The first job with any old valve receiver is to replace the audio coupling capacitors and most of the decoupling capacitors. As I removed each one I measured its leakage for interest and found values of around a megohm to several megohms. These values may not seem particularly significant to those familiar with transistor circuitry, but the input impedance of a valve can be more than many tens of megohms, and one sees circuit voltages, up to several hundred volts. With this combination, a leakage of 5 megohms begins to get important. For example 300-volts across 5Mohms will result in current of 60microamps. With a grid leak of say half a megohm, connected to such a leaky capacitor, one could see around 30-volts across it. Take it from me, that sort of grid source potential will make an output valve pretty hot, hence the importance of changing coupling capacitors.
Once I'd changed the vital capacitors, and having checked that all the resistors seemed OK, I connected the set to a 6-volt DC power supply and found complete silence from the loudspeaker.
Fortunately, pasted to the lid of the receiver was a circuit diagram. A very useful practice, I must say. After a few minutes I was able to discover that there was a resistor missing. This was the main feed from the rectified HT supply at the reservoir capacitor following the 6X4, to the smoothing capacitor and the remainder of the set's HT circuit.
I fitted a suitable new component of adequate rating and was rewarded by a slight hum from the speaker and sensitivity to touching the grids of the output valves. So far so good, but no stations and very little background noise.
As the owner had supplied a complete set of spare valves, even though all the heaters were lit, I started swapping the valves for new ones starting with the RF Amplifier.
I was not surprised to hear the set come to life when I inserted a new 6BE6, as a superhet without its local oscillator will not work, and it seems to me that the most common "dead set" fault of this nature with a valve set is failure of that function.
The set burst into life and after 30 minutes, I'd re-aligned the RF stage trimmers and peaked the IF stages, which are set unusually low at a mere 265KHz.
So far the repair had been straightforward, but what else remained to be done?
The set has five push-buttons for station selection. To adjust these, a button is pulled outwards, the set tuned to the desired station, then the button is pressed firmly back into position. Whether by design, or stiffness from the age of the set, this is no mean feat. The big problem is that the mechanism, which is provided for this feature, wasn't working. A clutch is provided, which connects the tuning knob spindle to the tuning capacitor via a gear train, and the lining of the clutch had long since perished. This prevented the set from being tuned.
To remedy the fault was not easy. The thickness of the clutch plate and its coefficient of friction were both critical. This was so when the set was first manufactured, but was now even more critical, because of wear.
When a push-button is pressed a mechanical gubbins opens the clutch. Over the years however, wear has taken place so that the original movement has been reduced to perhaps half. The result is that the clutch now hardly moves when a button is pressed.
A spring applies pressure, in the radio's normal condition, so that the clutch plate is allowed to trannsfer movement from the tuning knob. General wear and stiffness has taken place so that more friction is now required than was previously needed.
The use of a friction plate, such as one I made from neoprene rubber I initially tried, was no use. The plate certainly had the right amount of friction, but buckled and distorted from drag when a push button was operated. Not enough clearance from the gubbins.
A thinner tougher plate was needed, and eventually after a lot of trial and error I found a circle of alumium oxide-backed material from a sanding disk with a mating circle cut from a cycle inner tube, suitably roughened with emery paper served the purpose. The combination is strong enough to not distort from drag when a button is pressed and supplies almost enough friction for tuning. I say almost enough because, at the leftmost end of the scale one experiences slippage. To fix this I added a second spring to the mechanism so that the clutch pressure is increased slightly.
The radio now works properly, that is if one has sufficient strength in one's thumb to operate the push-buttons!