A Single Transistor Regenerative Receiver
with Big Loop Antenna, for AM Broadcast Reception

December 2, 2006 by Rick Andersen, KE3IJ

In my file on the Universal AM/FM/SW Regen , I presented a Colpitts-like regenerative detector that required a choke coil in the emitter circuit, which was basically a scaled-(up in value, down in frequency) version of the little 6-turn choke needed in the VHF aircraft Superregenerative receiver. In that same Universal article, I showed photos of the AM broadcast band version that I built into a balsa-wood box. Its 3 x 7 inch, 17 turn antenna loop coil gives it exceptional sensitivity... Unlike most homebrew Regens that are expected to be connected to an outside antenna, this one picks up AM broadcast stations as well as any inexpensive commercial transistor radio having a ferrite bar or rod antenna. The only "catch" is that you must set the Regen control just on the brink of oscillation, for exceptional sensitivity and selectivity-- but that's true of all Regens.

Well, I decided that if a 3 x 7 inch loop antenna (which is the "L" of the detector's L-C tank circuit) can perform so well, then a much bigger version ought to blow my socks off (think of Marty McFly and the gigantic loudspeaker in Doc Brown's lab at the beginning of the first Back to the Future movie)... and without the sensitivity to electrical noise that vertical whip antennas are known for. Loops respond to the magnetic component of the incoming radio wave, and not so much to the electric component. That makes them quieter (less prone to electrical interference).

So I went down to the basement "shack" and proceeded to get ridiculous: I strung up about 3 turns' worth of #22 enamelled copper wire from some shelving and ceiling struts, as well as from the legs on my workbench, and ended up with a rectangular loop that hangs vertically, with the vertical sides being about 4 feet tall, and the horizontal length being about 7 feet. This loop antenna is actually part of the circuit shown in the schematic below, with DC flowing through it from the 9v battery that powers the transistor Q1, the usual 2N3904 NPN silicon. The radio that springs from this brainstorm has been christened, the Big Loop Regen.

Unlike the Universal circuit, this one is the variant used in my AGC-80 article-- it's a Colpitts oscillator with not-quite enough loop gain to sustain oscillation, but enough to be able to set the gain right on the edge-- just what we want in a Regenerative receiver. But the point is that this circuit does not need any 10 mH inductor in its emitter circuit like the Universal did. And whereas the Universal tapped the audio off of the emitter circuit, this one takes it off of the collector side of the circuit, between the top of the antenna coil and the bottom of the 10K collector resistor. Both circuits work well; this one is simpler, in my opinion, because we literally have only one "L" in the whole circuit -- the Big Loop -- which is also sucking in any passing electromagnetic wave rippling by, with a large cross-sectional area.

And boy does this detector suck up signals!

Very respectable during the day, at night the AM band is packed from end to end with signals! I know this receiver is capable of hearing some serious DX (far distant stations), because I have been able to hear the time ticks (or "pips") of Cuban radio station "Radio Reloj" at several spots on the dial, which station I just learned about recently after doing a search on the Internet.

I had been puzzled about why I occasionally heard the Morse Code letters "RR" just below our local AM radio station, WHP 580 KHz, Harrisburg PA, on certain regen receivers I had built. With this Big Loop regen, I have heard these "RR"s not only at 570 KHz, but also around 820 KHz and one other frequency which slips my mind at the moment. With this regen I was able to hear, for the first time, that the "RR's" occur every minute, and that one-second time ticks are present continuously. Then when I heard the time being announced in Spanish -- even though the human voice sounded weak and distant -- I realized I was listening to a Time Station like our own WWV or Canada's CHU, but this one most likely being in some tropical area south of my QTH in Pennsylvania. The internet educated me about "Radio Reloj" and the fact that it broadcasts on more than one frequency in the AM band. I only wonder what this Big Loop could hear if all USA stations were to go off the air for a few moments....

I should mention here that, with the particular L and C values that I'm using, my receiver is unable to tune the whole AM band from 530-1700 KHz. Mine tunes from about 640 to 1300 KHz. This doesn't bother me too much, but it may bother you, in which case you are going to have to do some experimenting with L-C ratios to come up with something better than I did. But that's what we Radio-Heads like to do, don't we?

The schematic is shown below, and it is simpler than several of the other Regens shown on my site. I connect the output to my old standby, the Radio Shack Amplifed speaker, and fill the shack with loud late-night AM radio sound.

As I've described in my other articles, the circuit works as follows:

Q1 is configured as a variation on the Colpitts oscillator theme, but the inductor "L" has been expanded into a ridiculously large coil of wire which happens to be very sensitive to any passing electromagnetic ripples from radio stations.

Capacitors C1 and C2 form an AC voltage divider whose ratio sets up the right voltage and phase differences between emitter and collector, to provide positive feedback (regeneration) and potential oscillation. The tuning cap and the big loop coil comprise a parallel LC tuned circuit; the combination of L and all three C's set the resonant frequency of the circuit.

Any sine wave oscillator with insufficient loop gain to sustain oscillation, becomes a sharply-peaked bandpass filter instead. So we are essentially tuning an active LC bandpass filter so that its center frequency falls on the carrier frequency of our favorite AM radio station. The more gain (amplification) we give the circuit (by increasing the positive bias on the base via the Regeneration potentiometer), the more sensitive the circuit becomes, and the sharper the bandpass filter becomes (the selectivity increases, the bandwidth decreases, until at the critical point of breaking into oscillation, the bandwidth has decreased to just the center frequency itself).

Since Amplitude Modulation distributes the audio intelligence in mirror images on both sides of the carrier frequency (we call these sidebands), and we need at least 3 KHz' worth of bandwidth in order to preserve the intelligence, we can see how we must carefully adjust the Regen pot to the spot where the signal is as loud and sharp as we can make it, before it starts sounding like "seashell" filtering and very soon afterward begins to whistle/squeal-- this is when the bandpass filter (AM detector) has become an oscillator (Product detector) and the "beat frequency" (difference) between the station's carrier and your oscillator is heard as a squeal.

The resistors above and below the 5K REGEN pot act to scale the pot so that the critical oscillation trip point falls within its range of pot values; one minor annoyance of Regen detectors is that they require that you re-tweak the Regen control as you change the tuning frequency. In this design, the low-frequency end of the dial requires more positive bias (Regen knob clockwise, assuming you've wired it normally!), but will have to be backed off (counterclockwise) as you tune up higher in the band. You'll know when you've crossed into oscillation by the sudden appearance of whistling carriers sweeping by as you tune. Don't turn it counterclockwise sharply-- just a little; you want to be as close to oscillation as possible, without being in oscillation. For AM broadcast (which is oriented toward music and not simply voice communications), you will want to back off from the narrow-passband "seashell" sound, as noted above. You want the loudest signal you can get before the filter starts to restrict the fidelity. Sounds complicated, but it's more like teaching a kid to ride a bicycle: you pick it up very quickly-- don't let the length of my sentences fool you. You will eventually get used to the "feel" of your particular regen.

The 10uF cap hanging off the Regen pot's wiper bypasses noise to ground-- some pots make a lot of noise in the audio amp when you turn them.

The 365 pF Main Tuning capacitor is the old-fashioned meshed metal plate type. The little plastic tuning caps from transistor radios will work in this circuit, but you will have to re-scale the turns on the Big Loop antenna and possibly futz around with the feedback caps C1 and C2. The plastic tuning caps tend to be of a value around 170 pF or so; the older metal plate caps from old tube radios were 365 pF. In case you're wondering, No, we can't use a varactor diode here-- it has way too little range in capacitance and would not allow you to tune much of the AM broadcast band. Diode tuning works better at higher frequencies.

You may ask, What about the Big Loop itself? Is there anything "magical" about the 4 x 7 foot rectangular dimensions? Not at all. It's just what I decided to try in my basement, that's all. Some people wind about 15-20 turns of #22 or #20 enamelled copper wire around a cardboard box (called a Box-Loop antenna); some people scotch tape a loop around the perimeter of a wall in a room of the house; some have used a wood frame with 3-foot cross arms to support a loop antenna. I don't know how large you can make this loop and still have the circuit work (I imagine that different sizes present different load impedances and require modifications to the values of C1 and C2; see more comments about these caps below). But in general, the more cross-section of area encompassed by the loop, the more sensitive it is and the more DX it can hear.

Now, some comments on those voltage divider (feedback) caps, C1 and C2.

I have specified 220 pF ceramic capacitors for both of them, in the schematic. I have also varied both their values (keeping C1 equal to C2) and the ratio of their values to each other. In general, I have found the following:

* C1=C2=100 pF -- Circuit works, but there won't be enough feedback at the low end of the band to reach the oscillation threshold, and overall volume/sensitivity doesn't seem all that great.

* C1=C2=470 pF -- Nice and loud compared to 100 pF performance, but the upper end of the dial tunes more broadly (less selectivity) and is a jumble of loud, overlapping stations at night. Also, less of the AM band range is covered (because at 470 pF, C1 and C2 are already equal to or exceeding the maximum value of the Tuning cap. Since the overall capacitance is the combined value of C1, C2 and C-Tuning, this means that the variable component of the overall capacitance is only a small percentage, compared to the fixed part. Thus, less range of tuning).

* C1 (220 pF) is larger than C2 (150 pF) -- Well-behaved Regen characteristic at upper end, but lower end of band lacks sensitivity.

* C2 (220 pF) is larger than C1 (150 pF) -- Lower end of the band now works well, but upper end feedback is so large that the circuit "chops" or "motorboats" and then pops out of this chopping mode with too-abrupt a change; Regen pot is very touchy now.

* Changing the Big Loop antenna from 2 to 3 turns affected all of the above combinations of C1 and C2's performance.

I finally opted for the schematic's 220 pF for each cap. So start with those values and modify what you must to get the best performance. In other words, the price you pay for this ultra-simple radio is that YOU must decide how to optimize its performance for the particular band and physical loop dimensions that you use.

Note that if you were to adapt this circuit to Shortwave (World Band, 3-30 MHz) use, these caps would certainly need to be lowered in value, and the Big Loop would need to be reduced to 1 turn (or more for a small box loop). Then again, you probably DON'T want to try this Big Loop idea on the Amateur (Ham) bands, because any big loop is very prone to mechanical vibration, and this will cause the oscillating detector's frequency to wobble unacceptably for CW and SSB modes. But at AM broadcast frequencies, these small vibrations can't be heard, given that the detector is not oscillating in the AM reception mode.

The receiver will work without the 100 uF electrolytic cap, by the way. It's in there to bypass noise and keep the battery looking like a low impedance source, as is the case with all decoupling circuits. You can probably leave it out if you want to. On the other hand, if you end up building your own audio amp into this radio, leave the electrolytic in. Multiple stages invite interstage coupling (unwanted oscillations, motorboating, squealing).

As mentioned earlier, audio is taken off the junction between the Big Loop and the 10K collector resistor. A .022 uF ceramic or mylar cap filters out the RF component, bypassing it to ground and passing only the lower audio frequencies to the output coupling cap, which I show as a .22 uF, but anything from about .1 uF to 10 uF will work here. I send mine out an audio jack and into the Radio Shack Amplified Speaker, with comfortable listening levels at about 1/2 to 3/4 of the way up on the Volume control.

Now it's YOUR turn. Go slap one of these simple receivers together and string up a big loop antenna somewhere in the house, preferably where the kids won't trip over it and the wife won't banish you to sleeping on the couch for the next several nights.

If you build this Big Loop Regen and have good results with it, please feel free to email me at rick@ke3ij.com.
I enjoy getting "feedback" [regen pun] from readers!