Choccy Block Transistor Radio Circuit Description

All the voltages assume a battery voltage of 9.0V.

RF Amplifier Stages

The ferrite rod antenna has a single high Q tuned circuit which picks up the magnetic component of radio signals in the conventional way.  The output from the rod is coupled into the radio input from a smaller, lower impedance coupling coil via the d.c. blocking capacitor C1.  Q1 and Q2 are identical cascaded RF amplification stages.  At low signal levels the d.c. bias point is initially set by R5 and R6 giving 1.1V at the base of PNP emitter follower Q3.  This results in Q3 giving 1.7V at the emitter.  The base bias currents through R14 and R15 drop about 0.5V each to give about 1.2V on each base of Q1 and Q2.  This directly results in 0.6V at Q1 and Q2 emitters.  The operational collector currents of Q1 and Q2 are then set by the values of R1 and R3, that being approximately 0.6V / 1K = 0.6mA.  With that collector current, the values of R2 and R4 are set to put the collectors of R2 and R4 somewhere close to or a little above the mid-rail voltage.  This typically works out at about 4.7V.  C2 and C4 bypass the emitter resistors to maximise the gain of the two stages.

Demodulator and AGC

As well as setting the initial d.c. operating point of the RF amplifier stages, Q3 acts as both demodulator and buffer for the amplified radio signal.  As the RF voltage hits minimum peaks at the base of Q3, the emitter tracks this voltage down strongly with a low impedance drive.  As the RF voltage at the base increases above its previous minimum, capacitor C6 holds the previous negative peak and the voltage only rises at a rate determined mainly by R7.  This produces the required demodulated audio signal on C6, and the loading on the previous RF stage is minimised by the current gain of Q3.  The audio appears across VR1 and the wiper picks off the desired level.

The longer term average level of the audio signal plus the 1.7V Q3 emitter voltage appears on C7, low pass filtered through the 10K track of VR1 and C7.  As it is the negative RF envelope peaks that are detected, the average d.c. value on C7 drops with large signals.  As this is the voltage used to bias the previous stages, the collector currents of Q1 and Q2 will drop and the RF gain of the circuit will drop.  This provides automatic gain reduction on strong signals and contributes to the radio behaving more like a traditional superhet with AGC.

Audio Amplifier

C8 provides d.c. blocking and couples the audio signal into the audio amplifier stage made up of Q4, Q5 and Q6.  C8 must must be a non-polar type as the bias voltages on either side of it are about the same.  The audio stage is a very conventional class B output made up of Q5 and Q6 with Q4 being a driver which also exhibits significant audio frequency voltage gain.  Q4 is set up for a collector current of about 2.1mA, an operating point at which it shows high gain and will have sufficient drive for the output pair without using too much current.  R9 and R10 set up the d.c. voltage at Q4 base for 1.8V and the resultant emitter voltage is 1.1V across R11 470R.  That sets the Q4 collector current.  So, inline with the 0.7V diode drops of D1 and D2, a 1.8K resistor is required for R12.  This conspires to result in 4.5V at Q5 base which sets the output at the emitters of Q5 and Q6 to about 3.9V.  That is almost mid-rail and allows the best output swing into the loudspeaker.  The top of the output swing is limited by the current drive for Q5 from R12 and by the Vcesat of Q4 at the bottom end.

D1 and D2 should ideally be mounted in physical contact with the transistors so that base bias voltage, and so the quiescent battery current tracks with the device temperature.  R13 is present to reduce the quiescent current to a minimum, saving battery power.  In a more complex design the quiescent bias current would typically be adjustable.  This has been avoided to reduce the tweaking required.

This output stage configuration with these devices can have a full power bandwidth of up to 400kHz into an 8 Ohm resistor if no steps are taken to limit this.  Such a high bandwidth is undesirable, as small amounts of RF signal will get through the demodulator stage, be amplified by the audio output stage and feed back into the antenna causing high frequency oscillation.  The bandwidth of the driver transistor Q4 is reduced by adding collector-base feedback capacitor C9 and by driving the stage through resistor R8.  Both components are necessary.  C9 provides a path for high frequency negative feedback from Q4 collector to the base, but if this was driven from the low impedance of Q3's emitter directly when at maximum volume, it would be ineffective because the output of Q3 would dominate.  The result would be a system where the bandwidth was reduced when the volume control was in the middle range, but which increased again with the wiper at either end of the track.  R8 ensures that there is always some resistance in line driving the audio amp from the volume control, so the variation is reduced.  The resulting -3dB audio bandwidth of the amplifier is about 6kHz.  This is about right for AM radio, and higher bandwidth would just reproduce more annoying noise and make most stations sound excessively bright.  (See note 1)

100uF capacitor C11 couples the audio signal from Q5 and Q6 emitters into the loudspeaker while blocking the d.c. voltage.  100uF is quite a low value to use here and this avoids driving the small speaker with lots of bass signal which requires a lot of power.  The -3dB roll-off will start at about 200Hz with 100uF.  This saves battery current in a small radio, but you might want to increase this value for more bass output especially if you are using a bigger speaker which can reproduce the lower frequency sounds.

R16 and C12 form a low pass filter for the power supply to the sensitive RF and demodulator stages.  Without this, signals induced on the battery voltage by the high current taken by the output stage will feed back along the power rail into the earlier stages.  The result is a low frequency oscillation at high volume settings independent of the layout of the ferrite rod or loudspeaker.

Possible Problems

There is a lot of overall gain in the design and only one tuned circuit.  It is possible to get feedback and oscillations if you put the ferrite rod near the loudspeaker and its windings.  Feedback is also possible into the wiring of the tuning coil and the variable capacitor.  It is best to keep the tuning capacitor and coil at one end of the radio, and to have the loudspeaker at the other.  Feedback might sound like a howling noise or it could be that the radio sounds very hissy or quiet, and moving the tuning control just producing whistling noises.  Q2 collector and Q3 base are the parts of the circuit which have the most radio frequency energy in them, so the component leads in that area need to be short and you need to keep the ferrite rod and coils away from there.  The layout shown in the cardboard enclosure works well.


A TRF design is never going to be quite as good as a superheterodyne design for station selectivity and rejection of strong off-station signals.  This radio is designed to work just off the ferrite rod and batteries and it will suffer from picking up strong short-wave and long-wave signals if you connect an earth, an untuned aerial wire or a mains power supply.  If you live in a particularly mountainous valley where a normal radio struggles, you will have problems here too.

Note 1:  AM band modulation bandwidth in Europe is specified to have -3dB at 6.3kHz. (and -40dB at 9kHz).  Some years ago the -3dB point allowed was substantially narrower, more like 4.5kHz.  This fitted well with the IF bandwidths of radios made at the time.  IF bandwidths of most AM radios haven't changed a great deal though, as many are still made in the same way.  The approximate 4.5kHz bandwidth that you get out of them is inherent in the simple design. Consequently many AM broadcasters reduce the level of their low and mid-range audio transmission somewhat and gradually increase the treble above 3kHz.  This compensates for the limitations of many receivers.  This is called pre-emphasis.  Our radio does not have a traditional IF strip where this bandwidth limitation occurs.  The mildly reduced bandwidth of the audio amplifier is deliberate to fit in with most broadcasts.  There is another reason for this.  The channel spacing on the AM bands is specified as 9kHz in Europe by international agreement.  The result of imperfections in any radio system mean that all the signals interact a little in each of the stages.  One consequence of this is that a tone at the channel separation frequency, and at harmonics of that frequency tends to be created.  This is especially prevalent at night when many strong stations appear on channels adjacent to each other.  Adult males above the age of 40 tend to ignore this tone and hear it as the "traditional AM sound," but in a more sophisticated radio you might choose to put in a 9kHz notch filter specifically to remove it.  Children and dogs may appreciate this especially, as they have better hearing at high frequencies.  In our radio, the gentle roll-off of the amplifier bandwidth above 4.5kHz provides some relief.

Navigate Up

All content Henry J. Walmsley 2011