Ladybird Transistor Radio

Overview

This is my take on the MW radio from the marvellous Ladybird book making a Transistor Radio, also here, by G.C. Dobbs, with terrific illustrations by B.H. Robinson, first published in 1972.

The book shows you how to build a radio in stages:

• first, a crystal radio (of classic, but unfortunately, really poor design) on a 'breadboard' of DIY construction using screws and cups on wood

• add one, then two transistors (both OC71s) for amplification

• add a loudspeaker

• turn it into a  three transistor radio needing a shorter aerial and no earth lead

This final radio does away with that crystal stage of poor design and uses a third transistor (an OC45)  to amplify both radio and audio signals, making the radio a reflex receiver, although the book does not refer to it as such.  Some of the amplified radio signal is then fed back into the third transistor using a technique known as regeneration. 

(An aside: both techniques had been used originally in some (mainly DIY?) valve radios where cost was a major factor, presumably also true when transistors were a new thing and hence rather expensive. Other reasons, not applicable here, for keeping the number of transistors down, mainly (?) by the use of reflexing, were to save space in miniature radios such as the Sinclair Micromatic and for tax reasons, in the so-called Boys' radios imported into the USA from Japan. As to cost: regeneration as implemented here needs a trimmer capacitor and reflexing requires an inductor, another diode and a fixed capacitor. I suspect net savings, if any, using these techniques would have been small by the time the book was written and I wonder if the design was already a vintage one by then, especially as it uses germanium transistors from the 1950s? (One advantage of these old-school transistors is that they have nice long leads, making them ideal for the construction technique used here!) I think this would be consistent with information Dobbs gave in an interview with Practical Wireless magazine in 2009 (available via the link on his name above), where he tells how the radio had originated in the late 1960s as a project at an after-school radio club that he was running, his having just returned to the hobby after an absence of a few years, before which, as a youth, he'd been an enthusiastic amateur).

Although I bought the book in 1972/73,  I didn't make the radio at the time. I did make a crystal radio (of the same poor design), but using an air-cored coil, suggesting I'd followed instructions from elsewhere.  Reception was loud and clear, but was unfortunately dominated by just one or two stations across the entire tuning range. I remember the signals were so strong I could still pick up a station with a straight wire in place of the coil! I realize now that the lack of selective tuning was most likely because of that poor design. And I did end up with the equivalent of the penultimate version of the radio in the book when an enterprising friend hooked the radio up, rather ingeniously I thought, to the amplifier of my old Dansette-type record player via its (high impedance?) pickup. But what I didn't have was an actual Ladybird Radio, and certainly nothing resembling the final version, which is a shame, as the selectivity of that one would have been just fine. Anyway, after several decades I decided it really was high time to do what I had originally intended and finally build that radio: first, just the amplifier stage with loudspeaker hooked up to a 'toy' crystal radio and then, after some 45 years in all, the complete radio.

About the construction:

• my radio has all components on one side of a single piece of wood

•  the speaker is held in place magnetically by a steel bracket

• there is no separate volume control, as output is rather quiet (although as can be seen, I have drilled the board to take a potentiometer) & I increase the volume as much as possible using the regeneration trimmer

• the battery connector and coil are secured with double-sided sticky pads

About the components used:

• it has the vintage transistors specified in the book  and 'military grade' (so I was informed!) germanium diodes

• tuning capacitor: the book calls for an old-fashioned 500 pF (0.0005 uF) solid dielectric one, which I couldn't find at reasonable cost, so I opted instead for a modern 4 gang variable capacitor (polyvaricon) wired for 0 - 560 (280 + 280 + 0 + 0) pF.

• a 500 pF capacitor is specified in the book for use with the third transistor, but all I had to hand was a 471 pF one and used that first. After some experimentation (see section below on regeneration), I'm now using a 120 pF one.

• regeneration trimmer: I used a vintage 25 pF postage stamp type, not 10 pF as specified. Adjustment was originally by a plastic cap from a milk bottle connected to the trimmer screwhead with a short length of tubing from a soap dispenser, as shown above. It worked fine but I have now replaced the bottle top with a small tuning knob.

• Radio Frequency Choke (RFC) used in regeneration stage: not being able to get one at a sensible price, I first used a 4.7mH inductor shown in the picture, bought from an ebay seller who advised that it was a suitable substitute for this exact radio. It was later replaced with a higher value inductor using a common mode choke (see section on regeneration).

• speaker: a 3 or 5 ohm one is specified, but I could only find an 8 ohm one

Additional features:

• an on-off switch secured with a double-sided sticky pad

• a nod to more recent times in the form of a  red LED,  to show when power is on. It's in series with a 10K ohm resistor.

• a fine-tuning control (just squeezed in) using another 25 pF trimmer. It's a tight fit there and I've since removed it and have instead replaced the knob on the tuning capacitor with a larger one for better control.

Stations received:

I live near Southampton and can pick up BBC Radio 5 Live (909 kHz) easily. I can also receive a little less easily Talk Sport (1053 kHz) and  London-based Lycaradio (1458 kHz). 

At night, I've also been able to receive BBC Radio Scotland (810 kHz), Manx Radio (1368 kHz), a few Spanish stations and an East European station (Mar 24: Radio Baltic Waves International  - 1386 kHz) .  

Feb 24: also at night,  Radio Caroline  (648 kHz), just about, and a station with male and female voices, possibly in Arabic, with some music, strong reception, but on air for short periods only - TWR Europe (1467 kHz).

Sadly, MW is in terminal decline in the UK , as it is in the rest of Europe (with the possible exception for now of Spanish stations). BBC Radio Solent, which I was able to pick up on MW a few years' back is no longer there; BBC Radio Ulster (1341 kHz), has been gone since May 2021 and Absolute Radio (1215 kHz) as of Jan 2023. Smooth Radio (1557 kHz) went silent in Hampshire in Apr 2024. (The above-mentioned Radio Caroline bucked the trend however with the re-introduction of MW-broadcasting in 2017! ) I'm not aware of any campaign to keep MW going in the UK, but there is the Medium Wave Circle, which is dedicated to DX-ing and is still going strong after many decades, so perhaps with an upgraded coil and aerial, I could carry on tuning in for a while longer...

Changes being considered:

• longer coil: the ferrite rod is likely too short at around 2 inches - the book says to use half a 6 inch rod.

• add coils for SW and LW

• make new coils without ferrite rods. These will be much larger. Will the sound be better I wonder?

• make a new set with a longer board to allow more space for experimentation

Some observations:

• I tried a manufactured MW coil on a long rod and got much better results - the volume was much louder with little regeneration needed. You can also adjust the volume by sliding the coil along the rod.

• LED brightness can vary during tuning...

• occasional interference listening to stations beyond about 1000 kHz - presumed to be external

• volume (or signal) can be increased by touching a +ve terminal or by positioning hand near certain parts of the radio e.g. (very useful for distant stations) placing a fingertip close to the connection to the right of trimmer capacitor (with white bottle-top knob in the picture), i.e. opposite to the aerial connection.

• the original aerial was on a coil which was not fully extended. Not ideal, as this adds some inductance to the circuit. If I put my hand around the coiled aerial wire then the volume is increased significantly. 

• replacing the aerial with a telescopic one (intended I understand for FM reception!) salvaged from an old radio has made a big difference to the volume, even though it is much shorter. All domestic stations can now easily be heard from a short distance away from the radio and there's no more need for hand waving over the receiver to increase volume to an audible level!

Some notes on the new first stage

The diagram above is taken from the Ladybird book (and specifically from here) and shows the replacement of the crystal stage. First, note the additional few turns of the coil. The reason for these is that there needs to be a potential difference between emitter and base; those few extra turns provide a voltage drop determined by the r.f. signal received at the aerial that can be amplified by the transistor; they are not there for regeneration as such. Another way of looking at this - these extra turns provide a second coil which is used to provide the signal to the rest of the radio.

I've added impedance values in green for the components involved in regeneration. For the capacitors (impedance=1/wC, so low frequencies are blocked) and inductor (impedance=wL, so high frequencies are blocked), there are two numbers: the first is for w=10 kHz, approximately in the middle of the audio frequency range and the second is for w=1000 kHz, approximately in the middle of the medium wave radio frequency range. There is no specification for the RFC inductance in the book and here I make it 5 mH, approximately equal to the one I have used. C3, the trimmer capacitor is specified as being  10 pF (maximum) and the impedances given are for that value, although  I'm actually using a 25 pF trimmer and the impedances are actually 4M and 40K.

From this, we can see:

• radio frequency signals from the tuned circuit pass easily through C1 (impedance 100 ohms), are effectively blocked at a 390K ohm resistor at the top (out of view in the above image), blocked at the diode also, and so go through the transistor, via the base (low impedance assumed for now), for amplification;

• audio frequency signals from the diode are effectively blocked at C1 (impedance 10K ohms) and also at the 390K ohm resistor and hence go through the transistor, via the base (again low impedance assumed), for amplification;

• amplified audio frequency signals coming out of the transistor at the collector, at top, can travel one of three ways: through trimmer capacitor C3 (minimum of 4M ohm impedance in my radio), through the RFC (50 ohm impedance) or through capacitor C2 (200K ohm impedance). Clearly, the signal is effectively blocked at C3 and C2 and will go through the RFC and out, where it is needed, to the audio amplifier and then to the loudspeaker, having been amplified once (as an audio signal) by the transistor;

• amplified radio frequency signals coming out of the transistor at the collector, at top, can travel one of three ways: through trimmer capacitor C3 (minimum of 40K ohm impedance in my radio), through the RFC (5K ohm impedance) or through capacitor C2 (2K ohm impedance). Most of the signal will go through C2, after which it will pass through the diode and be demodulated to audio frequency. A tiny amount, sufficient for regeneration, will go through C3, then through the coil* and then through C1 after which it will go back into the transistor to be amplified again. However, a large amount will be wasted through the RFC, before demodulation, as the impedance here is only 2.5 times that at C2. At the lowest MW frequency, 500 kHz say, this loss is more pronounced with the impedance at the RFC (2.5 K ohms) being less than that at C2 (4K ohms). At low frequencies, the radio will be quieter, which is indeed the case. 

Finally, there are some problems with the section on regeneration in the book. Firstly, the diagram is at best confusing - in a commendable attempt to explain how the transistor is used "three times" to amplify the signal, the distinction between the two types of signal is unfortunately unclear and in addition, there is no output shown to the RFC. Also, the text is, I think, wrong to say of the amplified audio signal: The signal, now marked red, passes through C3 back to the coil. This signal will instead go through the RFC as described above; it is some of the amplified r.f. signal that is fed back into the tuned circuit, and just like the unamplified r.f. signal, will again go through the tuning capacitor, and not the coil, because of the high impedance of the coil. I've attempted to show all this in the diagram below, which now has just two colours for the flow of the signals: orange for radio frequency and blue for audio.

Experimenting with the first stage

I tried to remedy the loss of radio frequency signals through the RFC by first reducing the impedance at C2, but in error actually increased it by reducing the capacitance from 470 pF to 150 pF! I found that, unexpectedly, I was now able to receive lower frequency MW stations more readily, albeit at night, including  - just - BBC Radio 5L on 693 kHz. It seems that while it was now wasting more radio frequency signal through the RFC, perhaps more was also available for regeneration?  After some experimentation with a variable capacitor, the optimum value for C2 would seem to be around this number (150 pF) and needed to be greater than about 65 pF to work at all.

(It has occurred to me that perhaps another trimmer capacitor - in this case ~500 pF - could be of use here. I do have one of these and would need to find some space to add it. A bit bonkers perhaps, given one of the aims with this stage was presumably to reduce cost, but could be interesting and fun to try out!)

Next I replaced the inductor used for the RFC with a ~40 mH inductor from one side of a small (blue) transformer to give an impedance of ~ 40K ohms at a frequency of 1000 kHz and ~ 20K ohms at a frequency of 500 kHz. Result: stations quieter (e.g. Absolute Radio) or silent (e.g. BBC Radio 5L). Again, this is not what I expected, but then I realized that the impedance for audio frequency signals would also be increased 8-fold (e.g.  to 400 ohms at 10 kHz) and perhaps this was reducing their amplitude as a result?

I then tried a salvaged toroidal choke of around 10.5 mH inductance and various inductances obtained from common mode chokes of which I had 2, each comprising 2 coils of ~ 6.5 mH. Using these, I found the highest volume (when tuned to one of the stronger stations) occurs when I used ~13 mH and this is what I have replaced the original inductor with. Impedances at the RFC are now:

• 10 kHz (audio): 130 ohms

• 500 kHz (low MW radio): 6.5K ohms (cf. 4K at C2)

• 1000 kHz: (mid-range MW radio): 13K ohms (cf. 2K at C2)

Interestingly, I would say the difference between using the 5 mH and 13 mH inductors is not very great and perhaps that is why the author didn't specify a value for the RFC. Although it seems inefficient to lose some radio signal through it, using regeneration would seem to counter much of this.

If I remove the trimmer (or turn it fully anticlockwise) and hence remove the means for regeneration, the sound is now, not surprisingly,  quieter, but audible for major stations. The OC45 transistor is still amplifying radio frequency and audio frequency signals, and as such the radio is now operating as a reflex-only receiver.  The radio works fine in this mode with values of C2 greater than about 65 pF, as was the case with the regenerative receiver, but this time there seems to be no optimum value, which I also find surprising.

There is no mention of reflex receivers in the Ladybird book, but logically they fit in just before the section on regeneration. It's a terrific little book, but I think it would have been improved by the addition of a page on this.

BTW, it's interesting to compare this in reflex mode with the reflex circuit considered in the transistorized crystal radio.

A single transistor variant

It's possible to simplify the construction of this radio and do away with the amplifier stage and loudspeaker, using the first stage only with just the one transistor. To listen,  a crystal earpiece is used, with one lead connected to the output of the 10 uF capacitor and the other to the +ve terminal (or possibly to ground instead)  see relevant page in the book, here. The sound is more tinny, as is to be expected, but in all other respects the radio performs as well as the full version and with reasonable volume, making it a very good single transistor radio.

Modification to transistor biasing

There's an interesting critique of the radio here. It seems the specification of transistors to be used was relaxed in a later edition of the book, to the extent that the radio was very quiet for the author, who was using high-beta (hFE) transistors. He advises a change to the biasing of the first transistor in the amplifier stage to remedy this, making it, I believe, an example of a base-bias voltage divider and emitter resistor with AC bypass. Taking this to be good practice to follow (the gain approaches the beta of the transistor), I tried it on my radio, at first, with the original transistor, which turned out to have a very high beta of 141, compared with its neighbour's 86 and the datasheet's minimum of 30. On the plus side, volume was now increased greatly, with no need for regeneration at all for R5L. In fact, I had to turn it right down to avoid distortion, whereas previously I had had to turn it right up in order to listen. On the down side, there was now noticeable background static, tuning was imprecise, there was interference from SW stations at night (discussed elsewhere) and from bodily proximity and there were echoes when regeneration needed to be turned up (e.g. for TalkSport). Overall, the change made it unusable. I then replaced the transistor with one having a much lower beta (62). The disadvantages were now less extreme and outweighed by the increase in volume IMO. R5L was now decently loud without any regeneration and TalkSport was fine with just a little needed. For now, I'm keeping the mod with the replaced transistor. I'm puzzled though that it wasn't suitable for my high beta transistor whereas the author had implemented it to accommodate his...

I did try an emitter resistor in parallel with a capacitor for the second OC71, but the radio was quieter.  I also tried it with the OC45, but heard nothing, even with a 47R resistor alone. Perhaps a base-bias voltage divider would be needed for this to work? 

Below is a close-up of the radio with the biasing change to the first transistor, i.e. a 1K ohm resistor and 22 uF capacitor in parallel between emitter and ground. As can be seen, in order to accommodate the change, I've had to remove the LED for now.

In addition, there is now a 10K ohm resistor between base and ground of the second OC71. This was put there in an attempt to properly eliminate the potentiometer, but its presence seems to make little difference in practice...