A "Free-Power", Batteryless, One-Transistor AM Radio that works off of AC hum and 'spherics

6/1/2007 by Rick Andersen, KE3IJ

Here's something that might intrigue you Crystal Radio builders out there: A one-transistor (Silicon, yet!) AM broadcast receiver that derives its power from ambient AC hum and 'spherics [static crashes]!

Gotta give credit when I can: The germ of the schematic was borrowed from Darryl Boyd's "Stay Tuned" website which features over 170 articles and plans on crystal radio construction. The URL for the site is http://www.crystalradio.net/crystalplans/ and my No-Power -- or, more accurately, Free-Power -- circuit was derived from #2 of the 3 circuits found in the site's project #153, "How to Build Free-Power Radios." That website has three jpeg pages which are copies of the original article, by Terry L. Lyon, from an old Popular Electronics magazine issue. The 2nd page is the page containing the 3 schematics of single-transistor, free-power receivers. My version has some modifications of the original circuit, which I will describe below.

Terry Lyons' original circuit was designed around a common-emitter amplifier, with emitter grounded; mine has a 100K ohm potentiometer inserted between the emitter and ground. The wiper of the pot connects to a tap on the antenna coil, 5 turns up from ground, with the pot determining how much of the emitter current is permitted to flow through the coil winding in a Hartley Oscillator feedback configuration, so that a fraction of the incoming RF signal is fed back into itself after having been boosted by the transistor. With enough battery voltage present, advancing the pot to full clockwise would connect the transistor's emitter directly to the coil tap, causing the circuit gain to increase through positive feedback (regeneration) until the circuit "plops" into oscillation at the tuned frequency. In our "free-power" receiver, there isn't enough voltage or current to achieve stable regeneration to the point of self-oscillation; instead, we get "close": the volume comes up on our radio reception, to a point where we begin to hear a lot of clipping distortion that sounds like "clicks" on voice peaks. If we back the regen ("Sensitivity") control a little, we clarify the sound but lose a little volume. This is the optimum setting of the Sensitivity control. Your mileage may vary.

Lyons' circuits were predicated upon the fact that using a 10 Megohm resistor as the collector load kept the small power supply voltage -- the charge accumulated, rectified and stored on the filter cap -- from being bled away too fast. He also used a very high value resistor between collector and base. I found out, by trial and error, that using a 1N914 silicon diode in the resistor's place, gave a louder signal. Also, I found that reducing the collector resistor from the original 10M (!) down to 470K, gave me quite a bit more sound volume--although it did hasten the bleeding away of accumulated charge. But I found I could live with it. (By the way, I did most of my experimenting with a Radio Shack Amplified Speaker hanging off the audio output. The crystal (ceramic) earphone I had lying on the bench also worked, as did a pair of stinky old 2000 ohm magnetic headphones... just make sure that the .47uF blocking cap is in circuit. If you try to connect the magnetic phones directly to the transistors's collector and ground, the phones bleed the charge away very quickly and you end up with a dead receiver!

I tried using Germanium diodes as rectifiers-- and, contrary to what I expected [they have a lower turn-on threshold than Silicon diodes -- .2v vs. .6v-- so I thought they would be more effective than the Silicons. Wrong. They only seem to generate tens of millivolts, whereas Silicon diodes in the same circuit provide a fairly constant 450 - 600 mV.

The circuit can actually be brought close to the edge of oscillation, but this attempts to pull more current than the "atmospheric supply" can deliver, so you will hear a sharp, clicking or rattling distortion rather than plopping into oscillation, when the Regen (Sensitivity) is at max.

One thing that enhances performance a lot is a series L-C in the antenna lead (another ferrite bar coil of around the same inductance [240uH or so] in series with another 365 pF tuning cap, which acts as a resonant antenna tuner, and boosts signals very nicely at the low end of the tuning range.

Very interesting (if theoretically dangerous) how the +600 mV rectified DC that powers the receiver, fluctuates during lightning storms, if you monitor the 10uF filter cap with a DVM connected across it!

I won't claim that this receiver is perceptibly 'louder' than a good single-diode crystal receiver with a Germanium 1N34, but it's amazing to me that this circuit works at all! I'd like to try it with a Germanium transistor but I don't have any yet. Meanwhile, it's neat to just leave the receiver "on" all the time, with no battery, and it sounds pretty nice through a Radio Shack Amplified Speaker.

It is essential that the antenna be nice and long and that an EARTH ground be used (or a cold water pipe). We're trying deliberately to pick up as much AC hum as possible, since it is the power source!

Let me know if you build this, and how it works for you.