Before we consider the rest of the set, it is worthwhile having a quick look at the overall makeup of a typical MW/LW transistor superhet radio.
The aerial signal is tuned by one section of the tuning control, and then passes to the mixer-oscillator. The oscillator runs at a set frequency (typically 470kHz) above the tuned frequency. The mixer combines the oscillator and tuned frequencies such that sum and difference frequencies are produced. The difference frequency will therefore be 470kHz regardless of the frequency the set is tuned to. This is known as the "Intermediate Frequency" or "IF".
The IF is amplified through two stages of tuned amplification (IF amplifiers). The tight tuning ensures that the wanted signal is highly amplified while all other signals are rejected.
The Detector stage extracts the modulated audio signal from the IF signal. This stage also produces a voltage, which varies depending on the strength of the received signal. This voltage is fed back to the first IF amplifier to adjust its gain. Thus, if a strong signal is received, the gain of the IF amplifier is reduced to prevent overload. This signal is called the "Automatic Gain Control" or "AGC".
The audio from the detector passes through the
volume control and then on to the Audio Frequency (AF) amplifier and output
stage, to drive the speaker. This section has already been discussed in detail.
The circuit of a typical mixer-oscillator is shown here. As can be seen, a single transistor performs both functions. The circuit is a bit complicated to understand because the transistor is simultaneously operating in common-base mode as the oscillator, and in common-emitter mode as the mixer.
The "aerial tuned circuit" consists of the ferrite rod aerial coil and one section of the tuning capacitor. This is tuned to the frequency of the station being received. A second winding on the ferrite rod coil connects to the base of the mixer-oscillator transistor. The 56k and 10k resistors are the base bias components, and are decoupled by the 0.1uF capacitor.
At the oscillator frequency (which is typically 470kHz higher than the received frequency) the "aerial tuned circuit" is not tuned at all, and this appears as a virtual short-circuit. The base of the transistor is therefore grounded at this frequency, so the transistor is working in common-base mode. In common-base mode, the input is the emitter and the output is the collector. The "oscillator tuned circuit" provides tuned positive feedback between the collector and emitter, causing the circuit to oscillate. The frequency of this tuned circuit is set by the other half of the tuning capacitor. Fine adjustments would be included in the set to ensure that the oscillator tracks accurately 470kHz above the received frequency.
Although the aerial tuned circuit is low impedance at the oscillator frequency, it is at much higher impedance at the received frequency because, of course, this is what it is tuned to. The received signals therefore pass freely to the base of the transistor. At this frequency, the transistor is working in common-emitter mode, so the signals are amplified and appear on the collector as usual.
So on the collector we have the oscillator signal and the received signal mixed together. When two signals of different frequencies are mixed in a non-linear circuit, there are actually four definable frequencies on the output. These are the oscillator and received frequencies as expected, and also the sum frequency (a frequency equal to the two added together) and the difference frequency (the lower frequency subtracted from the higher one). The sum and difference frequencies are both amplitude-modulated the same as the received frequency. Since the oscillator frequency is always 470kHz higher than the received frequency, the difference frequency will always be 470kHz. The primary winding of the first IF transformer is connected in the collector circuit, and is tuned to 470kHz.
The mixing process is actually a bit more complex than this description suggests. The transistor biasing is taken partly into the non-linear region (similar to a class B amplifier at the brink of conduction) due to the oscillations being rectified on the base-emitter junction, developed across the 3.9k resistor and smoothed by the 0.01uF capacitor. This non-linearity is needed for the mixer to produce the sum and difference frequencies. Don't worry if this doesn't make sense - it's not important!
The circuit on the right is the mixer-oscillator
stage from a typical MW/LW radio. The waveband switching selects the appropriate
coils on the ferrite rod aerial depending on waveband. In addition the unused
aerial section is short-circuited to prevent it interacting with the used section.
Additional capacitance is switched into the local oscillator on LW, to reduce
the oscillator frequency. Trimming capacitors are included in the aerial and
oscillator circuits for alignment on both wavebands.
A basic IF amplifier stage is shown here. The circuit is very similar to the basic audio amplifier stages we have discussed previously. The main difference is that the input and output are connected through transformers having tuned primaries. The transformers are tuned to the IF.
With the simple transformers used the capacitors required to tune the circuit to 470kHz would be around 3000pF. Values this large are inconvenient in high frequency circuits. Also the transformers would only have a few turns, which reduces the efficiency of the transformer and makes repeatable manufacture more difficult.
To solve this, transformers with tapped primaries are usually used. The transistor collector is connected to the tapping and the capacitor is connected in parallel with the whole winding. This allows capacitors of around 300pF to be used. This arrangement is shown in the circuit below.
Another advantage of this arrangement is that
the low impedance of the transistor does not damp the tuned circuit as much,
resulting in a sharper IF response.
Early transistors, such as the OC45, did not perform particularly well at these frequencies. Feedback within the transistor, caused by internal capacitance, can cause instability or distortion, which limits the gain that can be obtained from the stage. A technique known as "neutralisation" or "unilateralisation" (depending on which book you read) is used to cancel this effect and allow the full gain of the stage to be achieved. Neutralisation involves applying external feedback around the stage to cancel out the effect of the internal feedback. It normally consists of a capacitor of a few tens of picofarads and a resistor of a few kilohms in series, connected between the secondary of the output IF transformer and the base of the transistor.
This circuit shows a typical two stage IF amplifier. The neutralisation components are at the top of the diagram. They consist of a 1.2k resistor and 56pF capacitor for the first stage, and a 3.9k resistor and 18pF capacitor for the second stage. The actual values will vary depending on the design of the circuit and IF transformers.
When replacing the IF transistor it is preferable to use the same type. If a different type of transistor is used the set may be prone to whistles if the neutralisation is not appropriate for the different device. If it is necessary to modify the neutralisation, the instructions in the service data should be followed if possible. If the service data is not available or does not give this information, the following technique should work.
Tune the set to a powerful local transmission. Switch off and connect a temporary link between the emitter of the transistor and the base bias circuit. If we were doing the first transistor in the above circuit we would link the emitter to the junction of the 56k and 8.2k bias resistors. The transistor now has zero bias and is unable to amplify. The only signal that can get through is via the internal capacitance in the transistor. Switch the set on and turn the volume right up. The station should be heard quietly. Adjust the value of the neutralisation capacitor between 10pF and 100pF (switch off when changing components) and select the value that gives the minimum output. This will be the value that has most completely cancelled out the internal capacitance. With the capacitor fitted and the temporary link removed, the set should work properly.
Double Tuned IF Transformers
In the circuit above, only the primaries of the IF transformers are tuned. To achieve a sharp response it would be preferable to tune the secondaries too. This causes additional phase shifts, which means that the neutralisation would not work correctly when connected to the transformer secondary. Where double-tuned IF transformers are used, the neutralisation is taken from the top of the transformer primary instead. This is less effective than the secondary connection, but the additional gain and selectivity achieved by double-tuning more than compensates for the slight loss in gain realised from the transistor itself.
This arrangement is shown above. Cf and Rf are
the neutralisation components.
The OC45 and equivalent/similar devices from other manufacturers were the first types of transistor to have widespread use in IF amplifiers. Transistor development continued, and the next family of devices, the AF117 and equivalents, used a different manufacturing process that reduced the internal capacitance, eliminating the need for neutralisation in IF amplifier circuits. The circuit below shows a mixer-oscillator and IF amplifier using these devices.
There are a number of devices in this family, namely AF114, AF115, AF116, AF117, AF118, OC170 and OC171. These devices have metal cans and a fourth lead connected to the can. This lead is connected to the 0V line in the circuit, so that the can acts as a screen around the transistor. Over time (several years) these transistors have a tendency to develop internal short-circuits between the can and one of the other connections (normally the collector or emitter). The problem was already known about and mentioned in a transistor radio repair book published in the mid-1970s. I will give more details about this and possible solutions in the repair section.
The next family were the last germanium devices to be used regularly in radios. They are the AF121, AF124, AF125, AF126 and AF127. They are in smaller cans than the AF117 series and do not become short-circuited internally.
In the circuits above and below, the AGC only controls the gain of the first IF amplifier. This is the usual arrangement, however in a few sets both IF amplifiers are controlled. The AGC voltage from the detector is connected to the base bias circuit of the IF amplifier transistor, and therefore alters the bias current. This alters the gain of the stage.
The mechanism for controlling the gain of the stage by altering the bias is something that many textbooks seem to gloss over. The gain of a transistor does not vary significantly with changes to the bias current, so adjusting the bias current does not in itself account for the gain variation.
Because the base-emitter junction is a forward biased PN junction, the voltage across it remains fairly consistent with changes to the base-emitter current. Ohms law tells us that R=V/I, so if the base-emitter voltage remains constant and the base-emitter current varies, the resistance (or more accurately, impedance) of the base-emitter junction must be varying.
The stage is being driven by the preceding IF transformer, so the load presented to the secondary of this transformer therefore varies depending on the AGC voltage. If the base-emitter impedance is fairly high, the transformer cannot deliver as much current into the base of the transistor. If there is less current variation in the base, there will be less current variation in the collector, thus the output of the stage will be reduced.
If a very strong signal is received, the AGC voltage could increase so much that the IF transistor bias current is decreased to the point where virtually no collector current flows. The IF stage would be working partly or completely in class B mode, resulting in distortion to the IF signal and the resulting audio signal.
To overcome this, some sets have a damping diode arrangement (D1 in the circuit above). An additional resistor is connected into the collector circuit of the first IF stage (2.2 k-ohms in the circuit above), with a decoupling capacitor to remove IF signals. When the transistor is biased normally, current flows through this resistor, and it drops a couple of volts. D1 is therefore reverse-biased and has no effect on the preceding mixer-oscillator stage.
If the AGC voltage is such that the IF transistor bias is reduced to virtually zero, the voltage dropped across the resistor will be virtually zero. D1 will then be forward biased and will effectively connect the 680 ohm resistor across the primary of the first IF transformer. This applies a considerable amount of damping, which reduces the output from the transformer and thus the drive to the IF amplifier. The circuit will settle in a state where the diode is conducting just sufficiently to prevent distortion in the IF amplifier. Note that the damping diode only comes into play when the received signal is so large that the normal AGC circuit cannot bring the circuit under control. With such a large signal, the reduction in selectivity caused by damping the IF response is of no consequence.
The detector stage serves two purposes. Firstly it has to extract the audio signal from the IF signal with the minimum distortion. Secondly it has to produce the AGC voltage to control the gain of the IF amplifier as discussed above.
Waveform (a) shows the modulated IF signal. The detector removes the negative half cycles, leaving the waveform shown in (b). Filter components remove the IF, leaving the demodulated audio signal (c).
The most common form of AM detector is a diode as shown here (left). T5 is the final IF transformer. The OA70 diode is the detector and the 0.01uF capacitor is the filter.
The AGC voltage is taken from the same point, but passes through a resistor-capacitor filter circuit to remove the audio signal and leave just an average DC voltage. This DC voltage increases as the IF signal increases.
The diode is sometimes contained inside the IF transformer can, so some disassembly may be required if the diode needs to be replaced.
Although a diode is the most common form of detector, it is not the only option. A few sets use transistor detector circuits (right).
The secondary of the final IF transformer is connected directly to the base-emitter junction of the transistor. There is no other bias applied to the transistor. The IF signal is sufficient to bias the transistor on the positive half-cycles, so the transistor is operating in class B mode. The transistor also amplifies the detected signal and AGC voltage.
Although the addition of an amplifier appears to be a good idea, in practice transistor detectors were not often used because it is more difficult to design a circuit that works well with minimal distortion over the full range of signal levels.
The circuit shown is a straightforward
example. Often the circuit will appear more complicated, but the principle will
be similar. The clue that a set has a transistor detector is the lack of a detector
diode and the presence of a transistor in the circuit between the final IF transformer
and the volume control.
As mentioned before, some sets have an extra audio amplifier stage between the volume control and the driver stage. This improves sensitivity and allows some negative feedback to be used to improve the sound quality. However the additional transistor increased the cost of the set (transistors were expensive in the early days), so an alternative method of increasing the audio gain was needed.
The trick is to use the final IF amplifier transistor to perform the additional task of being the first audio (AF) amplifier. Although this may sound odd, since the two signals are at completely different frequency ranges, the one transistor can do both jobs as long as the signals are combined and separated carefully.
This circuit is a typical reflex circuit, as used in several Ekco, Pye and Invicta sets. VT2 is the first IF amplifier stage, and works as described previously. VT3 is the reflex IF/AF amplifier. The IF signal path through the stage is the same as described previously, and the diode detector is also completely standard. The AGC voltage across RV1 is fed back via R5 to the first IF amplifier as usual.
The audio signal from the wiper of RV1 is passed back around via C29 to the base bias circuit of VT3. R9 is decoupled by a 0.04uF capacitor (C16), which is sufficient to bypass the IF signal but not the audio signal. The IF transformer secondary (L12) is a low impedance at audio frequencies, so the audio signal reaches the base of the transistor for amplification.
An additional resistor (R24) is connected in the collector load circuit, again with a 0.04uF IF bypass capacitor. The amplified audio developed across this resistor is passed to the driver stage via R13 and C23. R24 has quite a low resistance, so the amount of audio developed across it is not vast. However a fairly small amount of audio amplification will give a worthwhile improvement in the performance of the set.
This improvement is obtained without the
cost of an additional transistor. The only cost is a handful of resistors and
capacitors, and some careful circuit design. This type of design is a compromise,
since it is necessary to obtain some audio gain without significantly impacting
on the performance as an IF amplifier. In practice the performance of these
sets is better than a six-transistor set without reflex, but not quite as good
as a seven-transistor circuit.