Because of the additional complexity of VHF/FM circuits, I would recommend that newcomers to transistor radio repair should repair a couple of MW/LW sets first. The information in this section assumes you have already read and understood the preceding section on AM (MW/LW) RF and IF circuits.
There are two main differences between MW/LW (AM) broadcasts and VHF (FM) broadcasts - both of which are indicated by the names. VHF radio signals are transmitted at a much higher frequency - around 100MHz for VHF compared to 1.6MHz (1600kHz) at the top end of MW. VHF radio uses Frequency Modulation (FM) whereas MW/LW uses Amplitude Modulation (AM). It therefore follows that the two main areas of the set that are changed are the mixer-oscillator and the detector.
This block diagram shows the makeup of a typical AM/FM receiver. The component functions on AM are shown along the bottom, and follow the arrangement we have already discussed. On VHF/FM, two additional transistors are included in the circuit, acting as an RF amplifier and mixer-oscillator. The transistor that was used as the mixer-oscillator on AM becomes the first IF amplifier on FM, giving a total of three IF stages on FM. The IF on VHF is a higher frequency - 10.7MHz compared to 470kHz on AM. Because of the different type of modulation, a different detector circuit is used on FM. The remaining audio and output stages are the same.
The RF amplifier and mixer-oscillator are normally assembled in a screened can, due to the high frequencies involved. A typical (simplified) circuit is shown here.
The first transistor is the RF amplifier and is configured in common-base mode to achieve the highest gain. The RF amplifier serves two purposes. Firstly it amplifies the signal before it reaches the mixer-oscillator, as its name suggests. Secondly it provides some isolation between the mixer-oscillator and the aerial, preventing the oscillator frequency being radiated by the aerial and causing interference to other equipment.
The second transistor is the mixer-oscillator and is also working in common-base mode. Because the transistor is running near the limits of its frequency range, the phase shift between emitter and collector is about 90 degrees. Because of this, the oscillator tuned circuit only needs to provide another 90 degrees of phase shift to achieve oscillation.
L3 and L5 are the tuning control. Variable inductors are used on most early VHF sets, whereas later sets used variable capacitors and fixed inductors. L3 and its parallel capacitor adjust the tuning of the RF amplifier to suit the received signal. L5 and its capacitor set the frequency of the local oscillator to 10.7 MHz above the received signal. A damping diode may be connected across L3 to prevent the amplified RF signals from more powerful transmitters from overloading the mixer-oscillator stage.
The received and oscillator signals are mixed as described for the AM circuit and appear as a 10.7MHz If signal on the first IF transformer (L6/L7).
The VHF RF amplifier and mixer-oscillator assembly is often referred to as the "VHF front-end". The entire circuit is only powered when the set is switched to VHF, the power being switched by a section of the waveband switch. In this circuit, OC171 transistors are used in both positions. In many sets the RF amplifier will be an AF114 and the mixer-oscillator will be an AF115.
VHF IF amplifier
This circuit shows the typical IF amplifier arrangement (Ekco/Pye/Invicta diagram). VT4 and VT5 are the AM IF amplifiers. The circuit is very similar to that used for AM only sets, however there are two IF transformers, one for 470kHz and the other for 10.7MHz, with their primaries and secondaries connected in series.
In order to use the AM mixer-oscillator (VT3) as an IF amplifier on VHF, the oscillator needs to be disabled. This is usually achieved by bypassing the emitter resistor with a capacitor (C21), so that the stage cannot work in common base mode, using contacts on the waveband switch (SW1F). The input to the stage is switched to the output of the VHF mixer-oscillator instead of the MW/LW aerial circuit (SW1E). In addition the VHF IF primary must be bypassed when the set is working on AM otherwise the oscillator will not run reliably. In this circuit this is done by SW2A (MW) and SW3A (LW).
On AM the IF bandwidth needs to be quite tightly peaked to give good selectivity and avoid interference from adjacent channels. However, since FM broadcasting relies on frequency variations, the IF performance for this band must be flat over the range of frequency variations expected for normal broadcasts (normally +/-75KHz from the nominal frequency) the gain of each IF amplifier stage will be lower. Three stages of IF amplification are therefore needed for FM, compared to two for AM.
On AM the gain of the first IF amplifier is controlled by AGC as usual. On FM, no AGC is applied since it does not matter if the signal is clipped.
The transistors used are all AF116 types. These have a better response at 10.7MHz than the AF117 types used in AM only sets. VHF sets cannot use the earlier OC45 transistors. These only operate reliably at 470kHz with some fiddling (neutralisation), and will not work at all at 10.7MHz. This is why transistor sets with the VHF band were not produced until a few years after MW/LW models appeared - manufacturers were waiting for suitable transistors and circuits to be developed.
Frequency Modulation Detector
This diagram shows the difference between Amplitude Modulation (a) and Frequency Modulation (b).
Since the amplitude of an FM signal is constant, the single diode type detector used for AM would not work for FM. For FM detection we need to convert the frequency variations into a voltage variation. This is a little more complicated than AM detection.
There are a few different types of FM detector circuit. The most commonly used type is the "ratio detector" which is shown here.
Whereas the IF transformers between the IF amplifier stages needed to have a flat response across the frequency variation of the modulated signal, the final IF transformer has a tightly tuned peak at the unmodulated IF. Therefore when an unmodulated carrier is received (silence is broadcast), this is exactly at the peak or resonant frequency of the IF transformer. When the carrier is modulated by the transmitted audio, its frequency varies either side of the resonant frequency of the transformer.
The actual operation of the ratio detector depends on the phase relationships between voltage and current in a tuned circuit at and near resonance. To explain this fully would involve covering a significant amount of AC electrical theory, which is not appropriate here.
When an unmodulated carrier is received, the IF transformer is at resonance. At resonance the voltage and current are in phase, so the output power is at its maximum. The output is rectified by the diodes and C1 (typically 8uF) charges to the peak voltage. These act rather like an AM detector, but the value of C1 is such that rapid variations are filtered and only slow changes in level are responded to.
When the carrier is modulated, the frequency varies away from the resonant frequency of the IF transformer. Then the frequency is above resonance the phase of current lags (is behind) the voltage and below resonance the current phase leads (is before) the voltage. The amount of lead or lag depends on how far away from the resonance the signal is.
These current/voltage phase variations cause current variations through the diodes and C1. Since C1 is fairly large, the charge stored in it does not vary with the current/voltage variations caused by the modulation. The variations appear as a varying voltage on the centre tap of the secondary. This is combined with the voltage/current variations induced in winding "L" (an untuned winding close to the primary), to form the demodulated signal at the bottom of "L".
The remaining IF signal is removed from the audio by C2.
One problem with FM broadcasting is that noise and hiss tends to be more noticeable, especially when receiving weaker stations. To reduce this effect, the treble response of the audio signal is artificially boosted prior to transmission. This is known as pre-emphasis.
In the set a corresponding filter or "de-emphasis" circuit is required to reduce the treble response to the correct level. Since most noise and hiss tends to be at the higher frequencies, the de-emphasis removes a lot of this. Pre-emphasis and de-emphasis thus allow an improved signal-to-noise ration to be achieved while maintaining the frequency response of the original audio signal.
In the circuit above, R2 and C3 are the de-emphasis components.
The main advantage of the ratio detector described above is that it is largely immune to amplitude variations, due to the effect of C1. Since most interference tends to affect the amplitude of an RF signal rather than the frequency, such interference is rejected as an inherent part of the detector operation.
The disadvantage of this circuit is that quite accurate alignment is needed to give good results. If the alignment is slightly out, such that the resonant frequency of the transformer is not exactly the unmodulated IF, then distortion can result.
In a few sets an alternative detector circuit, known as the "Foster-Seeley discriminator" (shown here), is used. At first glance the circuit appears very similar to the ratio detector, except the two diodes point in the same direction and an additional inductor is used. This circuit also relies on voltage/current phase shifts in the transformer, however the alignment is less critical.
The drawback is that it is not immune to amplitude variations so an additional circuit, known as a limiter, is needed. A limiter is basically an IF amplifier driven into clipping, so the amplitude of the signal is fixed. The distortion caused to the IF waveform is irrelevant since it does not affect the frequency variations that carry the audio information we want.
The additional cost of a limiter stage cannot normally be justified in a portable transistor radio. However the Foster-Seeley circuit can give better performance and audio quality than the ratio detector, so it is sometimes used together with a limiter in high quality equipment where the additional cost can be justified.
The most basic form of FM detector is the slope detector. The circuit is the same as an AM detector, and relies on the IF transformer being detuned.
This graph shows the frequency response of the final IF transformer. It is adjusted so that its resonant frequency (4) is higher than the IF frequency. The IF frequency (2) is therefore part way up the response slope, and the modulation variations move it closer to (3) and further away (1) from the resonant frequency. The output voltage from the transformer thus varies depending on the frequency input, and a simple single-diode AM detector is sufficient to resolve the audio.
This circuit has a couple of disadvantages. It has no immunity to amplitude variations, so a limiter would be needed to clamp the IF level. It also requires fairly careful alignment to get the frequency variations in an area where the slope is reasonably linear. In practice this circuit is rarely used in radio receivers. However it is important to be aware of the effect because you can have a set with a detector working this way, unintentionally.
If one diode or the electrolytic capacitor in a ratio detector circuit failed, the output would be distorted. If a previous repairer has tried adjusting the final IF transformer, this would cause an apparent improvement, and the repairer might think the problem was solved. However, rather than fixing the fault, the repairer has simply changed the operation to a slope detector. The set might work fairly well on strong signals, but on weaker signals it would be very prone to noise and interference because there is no limiter.
If an FM set is distorted, resist the temptation to twiddle with the IF transformers unless they have clearly already been disturbed. It is more likely that there is an electronic fault causing the distortion. By fiddling with the IF alignment you will make the diagnosis of the real fault more difficult because the set will sound worse when the fault is fixed, until the alignment is corrected.