Comments for "Emitter Follower Regenerative Receiver"

18th June 2013 22:21

W. Thomas wrote ...

When you stated correctly 'Naturally the receiver lacks AGC and rapid QSB on long-path signals can be rather annoying to listen to', you also described very well a common problem of both direct conversion and regenerative receivers.

Did you consider adding an audio AGC?

In 1975, the following circuit principle of a then so-called 'Dynamik-Kompressor' ('dynamic(s) compressor') became famous, when Herwig Feichtinger, DC1YB, described it in the German 'Funkschau' magazine. The device could be built so small, that it could be integrated into a typical ham or CB radio microphone case.

In order to optimize the overall frequency response for speech transmission, coupling condensers where deliberately dimensioned in a way that they doubled as high-pass filters.

Field-tested values ranged from 22nF to 100 nF, depending on the voice of the main operator and the frequency response of the microphone and the transceiver.

In order to use the device also for music, e.g. for tape recording, coupling condensers with considerably more than 100 nF could improve the bass performance, but they also made the AGC function instable. Especially values of about 10 microF for the coupling condenser between the first and the second transistor resulted in 'blackouts' following fast 'attack time' reactions of the device.

Both attack and release time could be varied widely, but bass performance was never really satisfactory.

But for shortwave reception, this concept should do.


Amplified Ear


This is a description of the original 'classical' circuit


Speech compressor with three transistors

3rd May 2012 07:14

Mast wrote ...

I have built the emitter follower regenerative receiver. But in stead of the Colpitts oscillator topology described here, I use a Hartley oscillator in which the feedback cap is connected onto a tap at the coil. I did this for a reliable regeneration control over a wider range.

But this circuit has a tendency for producing nasty sqegging oscillations due to the loading of the oscillator by the RC lowpass filter and the relatively low impedance of the succeeding AF stage. It especially happens when I use a Medium Wave loopstick coil. A solution to this problem is by using an emitter follower buffer stage between the oscillator detector and AF stage.

This receiver can also be used for receiving FM broadcast stations. Use an RF transistor in stead of a BCxx when you want to receive FM broadcast stations.

Schematic of my emitter follower regenerative receiver:

1st January 2012 08:00

Simeon wrote ...

Hello Alan!First of all great page!Tons of fun and interesting stuff!I build that Emitter Follower Regenerative Receiver and it's the first transistor regen that worked for me.I bought bunch of lm384 and used one of them right after the first transistor,which gives great audio output to walkman headphones!I listened to some CW and many am stations,but there is no permanent coil yet.I want to ask you how did you connect the band spread capacitor?is it paralel to the tunning capacitor?Also the receiver is a bit tricky-sometimes I turn it on and it wont work,and seconds later it starts working normally!Happy new year Alan!

8th August 2011 22:10

Alan Yates wrote...


Sure, with enough signal it will injection lock broadcast stations. Adding a preselector in the amplifier or front-end would be a good idea though.



8th August 2011 21:05

9W2HFR wrote ...

Hi Alan,

Adding a RF antenna amplifier and operating this regen in oscillating mode, will this become an autodyne rx?

31st July 2011 16:56

Alan Yates wrote...


Quite often it is the carrier of an adjacent transmission. This is a particular problem on AM broadcast with 9 or 10 kHz heterodynes near the top of the passband caused by band crowding, especially at night when distant station signals get stronger.



25th July 2011 15:13

Bert wrote ...

From a lot of SW regen receivers that I have built, I heard a constant 5kHz sine wave. This happens with some but not all of the SW stations that I tune in. Actually I purposely design my low pass filter to have a corner frequency of 8kHz, hoping to improve the fidelity, reducing audio signal loss.

My question is, Is it typical to have a constant 5kHz sound coming out from a strong station? Why does it produce that kind of sound? Is it some kind of "pilot signal" as in FM radio?

10th January 2010 01:19

Pedro A wrote ...

A couple of weeks ago I was in Germany and bought the 'emitter follower' as a kit from It was fun to build and it is fun to operate. Slightly above oscilation, I'm able to use DReaM to listen to weaker AM and SSB stations :-) DRM should be possible for some stations which are near a strong AM station. The kit doesn't include a smaller variable capacitor for fine tuning and that is desperately needed fpr this. But what do you want for EUR 10.00 ?

23rd January 2009 00:57

Alan Yates wrote...


Injection locking is the tendency for an oscillator to capture an applied signal of similar frequency. Most oscillators will do this because of non-ideal behaviour of the circuit causing the applied reference signal to modulate the phase response of the circuit.

Injection locking can be useful, or bloody annoying, depending on the context. For example I once built a Theremin using a pair of RF oscillators, one with an exposed capacitive plate. The both oscillators ran at about 500 kHz using Hartley topologies and commercial IFTs from the junkbox. The first-try design injection locked through the power supply shared by the oscillators, making it nearly useless as a theremin; once the oscillators got within about 2 kHz of each other they'd capture each other and give no beat note. Once I fixed the supply decoupling they would injection lock through the mixer circuit, but the capture window was only a few 10s of Hz (ie subsonic) - this acted as a "zeroing" feature removing the need for an external zeroing knob.

The autodyne receiver I use as a poor-man's frequency reference uses injection locking to capture shortwave broadcast transmitter carriers. In the case of Radio Australia I know the carrier is derived from a Rubidium clock so it is a fairly good reference for checking the calibration of other devices. The video included with this article shows the "emitter-follower regen" also injection locked to RA's signal. As I tune with the bandspread knob I slide the window outside the detector oscillator's lock window and the detector oscillations break free of the RA signal, causing the audible hetrodyne. Tuning back towards the carrier frequency the detector once again locks the oscillator and the hetrodyne vanishes abruptly (not just like tuning through zero-beat). This can be done in either direction across the carrier frequency, there is a small window where the oscillator will lock the signal rather than just mixing with it. If you back off the regeneration control the window changes width until the detector stops oscillating. Advancing the regeneration control actually displays non-linear effects with the window size changing non-monotonically with regeneration level. It is most narrow near the point just past where oscillation begins. This is probably near peak-Q of the detector, further regeneration just broadens the response and eventually distortion/intermodulation rises rapidly - I'm not completely sure why, perhaps because the impedance of the active device drops and its parametric variations with each cycle increase. Less strong coupling to the transistor might help improve the overload margin of the receiver.

The other part of that sentence is also about using the radio in oscillating mode. Normally for listening to AM broadcast signals you'd run with the regeneration just below the point of oscillation, where the detector is most sensitive and its Q adjusted to give a bandwidth that matches the signal bandwidth. An alternative is to let the detector oscillate, so it is acting as a product detector. The stronger stations will injection lock as discussed above, but the weaker ones won't and you can either zero-beat them (which in practice is just about impossible - to demodulate DSB/AM with a product detector you need to inject the carrier not only at the right frequency but also in-phase so you achieve coherent detection, folding both sidebands over each other perfectly at baseband - in reality you can get pretty close but the distortion and phasing effects as the LO drifts across zero-beat is annoying to listen to). Another oscillating-mode alternative is to off-tune a bit so the LO is mixing the desired AM signal down to a low IF in the audio range, say about 11-12 kHz. Using a computer you can then filter and mix down one of the sidebands. Basically you use the computer as a SSB detector tuning from 0-22 kHz and the radio as a front-end targeting the low IF detector in software. The big drawback of doing this however is that both images are aliased onto the IF frequency, so 11 kHz each side of the RF local oscillator frequency get aliased to 11 kHz and if there is an adjacent station you will hear it. 10 kHz as an IF can be more useful in this case as spacings are often a multiple of 10 kHz so at least the adjacent carrier will be zero-beat. Noise from the other sideband also comes through and there is nothing you can do about that, you pay a 6 dB SNR penalty even if there is no adjacent signal. It works however and can be fun to try. This also lets you do other software-defined radio experiments. While you don't have the I/Q quadrature signals that let you phase-out the other sideband and avoid the aliasing problem you can still try decoding DRM signals or anything else your software supports without the Q channel.

Software wise I was actually using this program which was written by its author for VLF listening, of SAQ in particular - but the general scheme suited what I was doing. One of these days I'll probably have a go at SDR programming, starting with a mix-down and filter program like this. A general purpose low-IF mix-down utility would be quite handy for many things, not just RF related. Baudline lets you do it as part of its input decimation engine which is kinda handy, but isn't as versatile as what I have in mind.



22nd January 2009 22:18

Ben wrote ...

Hi Alan,

I do not quite understand what the following means. Could you explain it a bit?

"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."



Leave a comment on this article.