Emitter Follower Regenerative Receiver

I find myself really inspired by Burkhard's work. He experiments with some unique topologies and always comes up with something interesting to play with. Google Translate is as always very useful for reading his work if like myself you can't read German.

This receiver he calls the "The emitter-follower Audion". As usual he built it on a tinplate lid and uses a PC as the AF amplifier system. The unusual oscillator/detector topology drew me to it, naturally I just had to try it, being so simple it was easy to throw together.

I built mine in a Altoids Tangerine Sours tin I picked up in Las Vegas. A 12 k switch pot was used for the regeneration and a pair of polyvaricons for bandset + bandspread tuning rather than a single trimmer capacitor, and a toroidal core inductor instead of a slug-tuned one. Otherwise I followed his design (except I used 2N390x devices, despite having BC-series devices in stock - I had the non-European transistors sitting on my bench, left overs from the dekatron emulator build - no other reason). The antenna connection was applied at a tap on the tank circuit (5th turn from the cold end through a 3p9 capacitor). The inductor value (about 5 uH; 32 turns on a T50-2) was selected using my LC VFO calculator to optimise the tuning range given the polyvaricon capacitance of about 210 pF with the gangs combined.

The Insides of the Emitter Follower Audion

The radio works quite well. It tunes from about 4.8 MHz to beyond 23 MHz. Break into oscillation is fairly controlled, but a multi-turn pot would make it a little easier to optimise the regeneration point for non-oscillation reception of AM broadcast stations. With strong signals it tends to injection-lock the carrier, becoming an autodyne-style receiver. Weaker signals can be resolved as hetrodynes in oscillating mode, and by using software on the PC I was able to mix-down either sideband of a weak AM signal using the receiver oscillating and off-tuned.

Stability is quite reasonable, but degrades as expected with increasing frequency. A more robust mechanical construction might help in this regard (flexing the tin shifts the frequency a bit), but long-term drift is actually very good, surprising considering a type-2 toroid was used. The regeneration control pulls the RX frequency quite a lot, which is not unexpected when the tank is so heavily coupled to the transistor. It might be interesting to tap-down on the tank the connection to the transistor base. The bandspread polyvaricon makes tuning in stations quite easy. It's "antenna" gang (160 pF) is coupled to the tank via a 10 pF, this might be reduced if you want to more easily resolve SSB signals. It would be very unpleasant to use the radio without the bandspread/bandset tuning, at least with weaker stations.

The Front-Panel, So Far Unlabelled

I haven't labelled the front yet. I'll probably calibrate the bandset capacitor (chicken-head knob) with the bandspread centred so I'll know roughly where it is tuned, at least highlighting the shortwave broadcast and HAM bands.

31 metres was very active during testing, and Radio Australia and Radio New Zealand dominated the band. Voice of America and China Radio International were easily heard. Weaker stations were a bit harder to tune in with Radio Australia overloading the receiver at times, a pre-selector would be helpful, but reducing coupling to the antenna is helpful and once the regen is tightened towards the onset of oscillation the detector Q becomes sufficient to reject adjacent signals, even ones within 10 kHz (like the pairs of RA transmissions). Many stations were not identified, carrying non-english content or failing to identify within my limit of patience for listening to rapid QSB.

This video is of Radio Australia in oscillating mode, showing the receiver injection locking to RA's enormous signal.

Some Music on Radio Australia
Some Music on Radio Australia
(5.974 Mbytes)

Naturally the receiver lacks AGC and rapid QSB on long-path signals can be rather annoying to listen to. However, for its simplicity you can't help but be rather amazed how well it actually works. As Burkhard suggests, it might be a great first-project for someone interested in building shortwave receivers, but direct drive of "walkman" headphones would be helpful, untying it from an external amplifier or PC. (I tried a crystal earpiece, which works but only with the very strongest stations). I might design a matching 1.5 Volt amplifier capable of driving 32 ohm headphones and build that into the unit, making it stand-alone.