A Superregen Receiver for FM broadcast or Aircraft band AM reception
Feb 2004; Revised October 2006 by Rick Andersen
The Superregenerative detector was popular, in its vacuum tube version, in the early days of VHF reception up through the late 1950's, early '60s. After that, it was found in its transistorized version in simple kits from Radio Shack, etc. It's also the reason for the noisy hissing sound made by the 27 MHz CB Walkie Talkies many of us had as kids. Though a lot of people haven't even heard of it, and others think it's obsolete, the fact is that the superregen is still around in specialized high-frequency applications. Our use of it here is a throwback to the old days when homebrew 'supergennys' were in vogue -- as an extremely simple, yet extremely sensitive, broadband VHF receiver that will demodulate FM broadcast stations as well as AM aircraft band transmissions.
In the circuit below, Q1 serves as the superregen detector. Now this configuration is a strange beast, if you're only casually acquainted with transistor circuit design. First of all, in the original 2004 embodiment of this circuit, the signal was coupled into the collector circuit of the 2N3904 NPN transistor; usually we connect the input to the base or emitter, and the collector is considered to be an "output". Well it turns out that the high impedance of the LC tuned circuit, at the collector, makes it susceptible to outside influence, if we connect an antenna at that point...... this circuit teaches us that you can impress a signal on any lead of a transistor and take the output off any other lead, and the damn thing will still work!
In this revised version of Oct. 2006, I've moved the antenna to the point between the top of the RF choke and the bottom of the emitter of Q1-- the radio picks up aircraft signals even better now than it did before. Again, there are several places where you could couple the signal in.
The tuning cap is a 15pF variable I happened to have lying around in my junque box; the inductor "L" is nothing but a 2" piece of #20 stiff copper wire, bent into a "U" shape. If you squeeze that U into a near-bobby-pin shape, you will have eliminated most of the inductance, raising the frequency of the tuned circuit. You will have to experiment with the amount of "squeezing" you do to the wire; FM broadcast stations (88-108 MHz) will need more inductance, and the lower half of the Aircraft band (approx. 109-130 MHz) will need less, since it's above the FM broadcast band.
The antenna has been changed from the February 2004 original of this article. Originally I had a 24" whip, made of #8 or #10 solid copper wire, soldered to the ground plane and coupled to the L1 / C-tuning circuit through a 1 pF cap. My prototype was plagued with a lot of "dead spots" and "flutter" as I tuned across the dial. Adding a buffered input stage helped, but I later discovered the present method of making an antenna that worked even better, and allowed me to get rid of that extra input stage, with little change in sensitivity. The antenna is a 24-inch piece of plastic-coated #20 or #22 hookup wire, bent in the middle and folded back on itself, then twisted so that we now have a 12-inch "whip" with two connections at the bottom, which are soldered between (in series with) the emitter of Q1 and the little 6-turn RF choke that was made by winding #22 enamelled wire on a BIC pen (the pen is used as a coil form and later removed). Since very little inductance is wanted at VHF, my folding the antenna wire back on itself and then twisting its two halves together, gives it a "non-inductive" characteristic that allows me to get 12" of length without so much inductance as to spoil the circuit's operation around 110-130 MHz. This unusual antenna seems to work very well (plenty of sensitivity) and most of the dead spots and flutter are gone.
Just make sure you build everything nice and tight -- long, floppy leads are a disaster at VHF. I usually use 2 pieces of Radio Shack's 5x7" copper clad board, connected together perpendicularly so as to make a bottom piece and a vertical front panel -- an L-shaped chassis. Then I construct the circuit above this ground plane, using 10 megohm resistors as "standoff insulators" or posts, on top of which I wire the circuit.
Positive feedback required for oscillation is provided by the small 7pF cap from collector to emitter. You may need to increase this to as much as 10 -15pF -- or, try another transistor, if your detector won't oscillate. The way to tell that your detector is oscillating is to listen for a loud white-noise "rushing" sound from your amplified speaker, connected to the audio output of this radio. If all you hear is "quiet", it ain't workin!
Q1's emitter has a 6 turn RF choke connected to it; below that choke are a 10k resistor to ground and a .001uf cap across that resistor. These 3 components form a subcircuit that makes the superregen unique: the Quench Oscillator. (The 10k resistor is also the emitter load across which we take audio which is sent to the second transistor, Q2, for amplification).
Here's how it works -- and I'm giving all this detail because so many people have no clue about the mysterious goings-on inside this circuit:
The 330K resistor provides + bias to turn Q1 on. The .001uF cap on the base provides an RF ground at the base. Like a normal regenerative detector, if there is enough positive feedback, the transistor will oscillate. However, in a normal regen, we find that the most sensitive reception occurs at the point just before oscillation sets in; after that, we get a heterodyne squeal if there's a carrier present in the signal being received. In a superregen, we interupt the buildup of feedback, between about 20,000 to 30,000 times per second, with a lower-frequency sawtooth waveform that is automatically produced in that 3-component subcircuit mentioned earlier. The .001uF cap across the 10K resistor forms an RC time-constant circuit that sets the "quenching frequency". It turns out that if we alternately "quench" (interrupt) and then buildup the feedback, at >20 KHz (above the range of human hearing), we can effectively drive the transistor much farther into the range where positive feedback would have resulted in self-oscillation. I.e., by delaying the onset of oscillation, or chopping it up, we effectively operate in a much more sensitive region of the transistor's bias. We hear practically down to the "noise floor", and we get fantastic sensitivity.
The drawbacks? The frequency response is much broader than a normal regen detector -- in fact, the superregen can only "hear' broad (wide-bandwidth) signals like FM radio stations with their 200 KHz -wide deviation. Narrow-band FM is not well-detected by this circuit. The detection method itself is known as Slope Detection, and you may notice that FM music stations are a bit hard to tune in clearly; there always seems to be a little bit of distortion. AM Aircraft transmissions, on the other hand, come in clear as a bell, and the broadness of tuning, which would be a serious disadvantage in a professional receiver, works to our advantage in this simple radio. The reason I say this is that aircraft transmissions tend to be short, choppy remarks, and the pilot may be on one frequency, while the tower may be on another, so that a selective radio will hear only one side of the conversation. This receiver will often pick up both sides, if they're near enough in frequency. My homebrew version of this receiver tunes from approximately 112 to 130-ish MHz; the band actually extends to 174 MHz.
This circuit has been around for a long time; I didn't invent it. I did, however, come up with an enhancement based on another radio design found in these pages -- the Reflex Receiver: I connected a 12k ohm resistor in series with a .47uF capacitor; this pair is connected from the bottom of the RF choke back to Q1's base. This is essentially feeding back some audio (with 25 KHz quench waveform) which I find boosts the audio gain and also gives it a fuller, slightly "bassier" sound, compared to the "tinny" sound I got before putting the RC pair in.
A .47uF cap blocks Q1's DC emitter voltage while passing the audio to an RC low pass filter (22k, .01uF) at Q2's base; the RC filter gets rid of most of the RF, leaving the audio intact. Q2 audio amplifies the signal and the output is coupled through a 2.2uF electrolytic to an output jack. I found that the addition of the .047uF cap to ground was necessary to kill the observed tendency for stray RF to interact with the Radio Shack Amplified Speaker through the connecting cable, which caused an instability which varied with my hand position near the radio. The cap also softens the "tinniness" some more.
Even though this unbuffered Superregen probably doesn't radiate very much noise, I would not bring this radio anywhere near an airport, especially in these paranoid days of terrorist threats! I wouldn't want to find out that I was causing any interference to aircraft or tower communications! I live about 30 miles from Harrisburg, PA, and have no trouble at all receiving pilots' transmissions as they enter and leave the area. So you don't have to be close to the airport to hear aircraft communications. Being way up in the air, the planes' transmissions are receivable for many miles around.
I hope you'll take the time to build one of these almost-forgotten Superregen circuits and see how well they can hear, for such a simple circuit. Just be ready to do a little futzing around with parts layout, etc., before it will work smoothly. Be ready to make slight substitutions, try another transistor if the circuit won't oscillate, and understand that you'll have to "calibrate" the receiver for the band you're interested in (FM or Aircraft) by custom-bending the U-shaped wire that is "L". Take my 2" length for the main inductor "L" as a starting point; you may need to modify its length to get your radio into the correct frequency band.