The Joule Thief (Wikipedia) is a well-known “instant gratification” hobby circuit that uses just a handful of components to pull off a pretty impressive parlor trick — using a single 1.5V battery, the Joule Thief can light a high-voltage blue or white LED that normally requires 3.5V or greater to turn on. Even more impressive, it can do so using a battery that is so drained of energy as to be counted “dead” for almost all other purposes. I have not measured this value myself, but it is commonly claimed that a Joule Thief can light a white LED from a battery with an open-circuit voltage as low as 0.6.

Even a fresh AA will not light a white LED by itself. This battery had an open-circuit voltage of 1.62 right before I connected it. It takes at least 3V DC to light this white LED.
Even a new 1.5V battery will not light a white LED by itself. The fresh AA cell to left had an open-circuit voltage of 1.62 right before I connected it. At least 3 volts, for instance from the two AAA cells shown to right, are needed to turn the LED on.
Chester Winowiecki's "Zombie Flashlight" from MAKE 37.
Chester Winowiecki’s “Zombie Flashlight” from MAKE 37.

We published a miniaturized flashlight build based on the Joule Thief circuit as Zombie Flashlight in MAKE Vol 37. That build uses an empty lip balm tube to house a AAA battery and the Joule Thief electronics, which are assembled on a tiny slab of perfboard. We had a great time building our test prototype in the MAKE labs, and got interested in the idea of designing a free-form Joule Thief flashlight that could be soldered together from individual components with no prototyping board or PCB. We were playing around with the components and had a batch of those new(ish) “Piranha-pin” low-profile LEDs lying around at the time. The discovery that a TO-92 transistor package nestled perfectly back-to-back into the empty space between the four LED pins kind of sealed the deal — we needed to do a free-form build. The Vampire Flashlight is the result.

Resistor, transistor, and LED neatly folded into a package smaller than your little fingernail.
Resistor, transistor, and LED free-form into a package smaller than a little fingernail.

The design is supercompact and minimal, and the end product is a usable flashlight only just slightly bigger than the battery itself. While the Joule Thief is a very simple circuit to build, this form factor demands a bit of soldering skill and patience. If you don’t enjoy the process of meticulous electronics craftsmanship, it could easily be adapted for construction by mounting battery pack, switch, toroid, and everything else on a small rectangle of perfboard, then completing the circuit connections with 22AWG solid-core jumper wires.

How it Works

There is of course no magic here. The Joule Thief does not produce more energy than it consumes — rather, it uses the energy that is available to it in very clever ways. Just as it is possible to design a hydraulic system (like Heron’s Fountain) that uses a relatively large volume of water at lower pressure to transport a relatively small volume of water at higher pressure, it is possible to design an electrical circuit that “steps up” a relatively large current at lower voltage to a relatively small current at higher voltage. There are in fact several tricks for doing this, and the Joule Thief uses more than one of them.

Vampire Flashlight waveform with fresh alkaline AA, measured at the LED.
Vampire Flashlight waveform with fresh alkaline AA, measured at the LED.

The most straightforward to understand is that the circuit oscillates to produce short bursts of current at higher power than the battery could continuously sustain. These bursts of power come tens or even hundreds of thousands of times per second, which is much faster than our eyes can detect; the LED appears to be continuously lit in spite of the fact that it is actually powered less than half the time.

caption

Voltage from the battery (A) trickles through the transformer’s secondary winding (B) to the transistor’s base (C), opening the transistor’s collector path (D) partway. This energizes the primary winding (E) partway, storing up energy in a magnetic field which in turn induces current in the secondary, boosting the voltage back to the base, opening the transistor further … around and around. This feedback loop rapidly slams the transistor fully open. The transformer’s tiny ferrite core (F) rapidly saturates, at which point induction ceases, the base voltage drops again, the transistor slams shut, the magnetic field collapses and its stored energy is dumped to the LED (G). The cycle starts again, switching on and off about 50,000 times a second (50kHz) — more than fast enough to make the LED appear steadily lit.

in-use

What will the next generation of Make: look like? We’re inviting you to shape the future by investing in Make:. By becoming an investor, you help decide what’s next. The future of Make: is in your hands. Learn More.

Project Steps

Score a toroid

A toroid, in an electronics context, is a ring-shaped “bead” made of some material with a high magnetic permeability. Sometimes they are made of epoxy or another plastic filled with powdered iron, and sometimes they are made of ferrite, a dark gray or black ceramic material consisting mostly of iron oxide. They are typically wound with coils of wire to make toroidal inductors and/or toroidal transformers, which are very widely used in alternating-current (AC) power supplies. The “donut” shape is useful because it shapes the magnetic field in a way that reduces the inductive noise created in nearby components.

We’re going to start with a bare toroid and wind our own toroidal transformer. If you want to scavenge a toroid, find a junk PC motherboard, inline laptop or printer power supply, or compact fluorescent (CFL) light bulb and open up the case. (If you’re working with a CFL bulb, take care not to breach the fluorescent tube, as it contains trace amounts of mercury.) If you’d rather just buy a toroid than have to hunt one down, I recommend Fair-Rite #5943004901, which is the part shown in the title photo and through the rest of this build.

Once you have the case open, look for a little donut-shaped bit wrapped with coils of apparently bare copper wire. I found two inside the AC adapter (ViewSonic model HASU12FB60) for a long-dead LCD monitor. Either the big one or the small one would work with the Joule Thief circuit, but the small one is more practical for a compact build. You can try to desolder the toroid from the printed wiring board if you want to, but we’re going to be stripping the coils off, anyway, so the easiest thing to do is just snip the windings on the component side of the board.

Once you have the toroidal inductor loose, unwind and/or or cut away the wire until you’re left with just the bare magnetic core.

Hack the battery pack

We’re going to be soldering directly to the battery case’s built-in terminals. These are mounted in relatively thin plastic and it is easy to melt them loose and ruin the battery pack if you try soldering to them while they’re in place. It can be done with skill, speed, and luck, but you’ll likely ruin one or more cases getting the knack of it. A better way is to demount the terminals from the plastic, tin them, and then put them back in place. Start by prying up the folded-over metal flaps securing each terminal in place with a hobby knife and/or small flat screwdriver. Once you’ve got all four flaps on each terminal pried out straight, pop the terminals out of the ends of the battery case.

Clamp the terminals in a vise or other work-holder by their wire leads. Apply heat with your soldering iron and then “tin” the inner, rounded sides of the terminals with solder. If you accidentally solder the flaps together, don’t worry — just remelt the solder with your iron and shake or blow it away (in a safe direction), then try again. Remember that solder will flow towards the heat, so apply your iron to the rounded side (which is where you want the solder to go), and not to the flaps.

Once you’ve got the solder where you want it, clip the terminal leads off right at the ring, then carefully reinstall the terminals in the battery holder and secure them in place by first folding the metal flaps back down with the tip of a small screwdriver, then squeezing with a pair of small pliers.

Install der svitch

We use a normally open SPST momentary switch (aka “a button”) to turn the flashlight on. The switch is mounted directly to the negative terminal of the battery holder, and activates the circuit by completing the connection to ground while the button is depressed.

Position the switch so that two diagonally opposite leads fit into the negative battery terminal and the ready-made hole in the battery holder where the black lead used to come out. Mark the location where the other two leads contact the battery holder, and drill clearance holes for them using a small (about 0.04″) twist bit in a pin vise. Fit the switch on the battery case and make sure the leads protruding inside are not shorted against the negative contact spring. The bottom-center switch lead is the one to watch out for; if it does contact the spring and short the switch, just clip it off long enough to penetrate the case wall from the outside but short enough that it does not protrude inside.

Mount the switch in a vise or other workholder and generously tin the lead that will connect to the battery holder’s negative terminal. Then strip one end of the black wire you cut off the battery holder earlier and solder it to the lead diagonally opposite. Make sure your solder joint here doesn’t add too much width — you’ll need to be able to pass it through the hole in the corner of the case.

Thread the black wire through the corner hole and fit the switch in place on the end of the battery case. Touch the tip of your soldering iron to the switch lead at the battery negative terminal just long enough to reflow the solder on the lead and on the terminal and get a good joint, then immediately remove the iron and blow to cool. Again, the thing you want to avoid here is overheating the case terminal and melting it loose from the surrounding plastic. Fold the unconnected switch leads over, inside the case, using the tip of a small screwdriver. This will help hold the switch in place.

Temporarily install an AA battery in the case and connect a voltmeter across the end of the black wire and the positive battery case terminal. The reading should be 0V, or something very close to it, until you press the button, at which point you should see it climb to about 1.5V.

Wind the transformer

Reel out and cut 16″ each of green- and clear-enameled magnet wire, and twist the two wires together securely over a length of about 1″ at each end.

Wind the pair 10 times around the toroid. Take your time and make sure the coils are evenly spaced, and that the wires do not cross each other at any point on the core. When you’re done, one pair of wires should protrude from the top of the coil, and one from the bottom.

Clamp one end of the wire pair in your vise and seal the whole assembly with two or three coats of nail varnish.

After the last coat of varnish is dry (usually about an hour after application), release the toroid from the vise, untwist the two wires on each side, swap out the green wires between the pairs, and twist them lightly together again so you don’t get them mixed up. This step is critically important. The circuit will not operate unless current is flowing in opposite directions along the two wires wrapped around the toroid.

Free-form the shiny end

It is a very good idea to connect transformer, transistor, resistor, LED, and battery pack together on a breadboard and verify that the entire circuit is working before you start soldering. Use the actual components that will go into your finished flashlight and, once you’ve proved they’re all working, take them directly off the breadboard for assembly.

Wrap one leg of a 1/8W 1K resistor around a scrap of wire or cut-off component leg to form a small loop right next to the resistor body. Then bend the middle (base) leg of the NPN transistor out 90 degrees away from the flat side of the transistor case and solder it to the resistors’s other leg, as shown. You want the body of the resistor right up against the transistor’s rounded face.

Cut a 3/16″ length of 1/4″ diameter heat shrink tubing, slip it over the resistor/transistor pair, and apply heat to shrink and secure. Snip off the excess transistor base lead, and the excess resistor leads on both sides of the resistor body (leaving the loop attached).

Butt the flat side of the transistor against the back side of the LED — it should just fit between the LED pins. With the LED facing down and the legs of the transistor protruding toward you, make sure the “clipped” corner of the LED case is in the near-right position, as shown. Wrap the outer transistor legs (the collector and emitter) around the outside of the two LED pins on each side and solder them in place. Snip off any excess lead from the LED pins and the transistor legs.

Connect the transformer

Cut the wire pair on either side of the toroid to about 1″, then carefully and thoroughly clean the enamel off each wire with a piece of fine sandpaper. Twist the wires together firmly, and evenly, then solder them together. Be generous with the solder, but don’t add too much width.

Slip the soldered wires through the hole in the center of the positive battery case terminal. Make sure the toroid is centered and firmly seated against the end of the battery case, then briefly touch your soldering iron to the positive terminal, inside the case, to reflow the solder there and on the toroid windings and make the joint. As before, be careful not to overheat the terminal and possibly melt it loose from the surrounding plastic. When the joint is secure, trim the protruding wire pair flush with the battery terminal.

Cut the remaining wire pair protruding from the toroid to about 1″, then carefully and thoroughly clean the enamel off the entire length of the protruding clear-coated wire with fine sandpaper. Also clean the enamel off the green-coated wire over a length of about 10mm right at the end. Slip a 1/4″ length of 1/16″ diameter heat shrink tubing over the green wire, then hook the stripped section through the resistor loop and secure by soldering. Slide the tubing up over the solder joint and apply heat to shrink.

Test and finish up

Strip the end of the black wire that runs through the battery case to the power switch, and hook it around the bent transistor lead on the emitter/negative side of the LED case. Hook the bare copper wire from the toroid around the bent transistor lead on the collector/positive side of the case. Don’t solder anything yet. Insert a battery in the case and press the power switch to confirm that the LED illuminates before proceeding.

Carefully cinch the LED/transistor/resistor bundle down along the toroid and ground wires until it is seated against the toroid face with the transistor body nestled inside the toroid ring, folding the green wire in underneath as you go. Once everything’s in place, bend the toroid and ground leads out to either side as shown. Confirm that the circuit is still working properly, then solder the leads in place on either side of the LED.

Trim the excess ground and bare copper wires away on either side of the LED, reinstall the battery, and you’re done!

Conclusion

Vampire Flashlight built with scavenged powdered-iron toroid and "Super Flux" white LED.

Vampire Flashlight with scavenged powdered-iron toroid and “Super Flux” LED.

Troubleshooting

The Joule Thief is an extremely simple, reliable, and well-proven circuit. The most common construction error is wiring the coil incorrectly. Remember that conventional current must flow in opposite directions along the two toroid windings. Though the two wires are physically wrapped around the toroid as a pair, if you don't swap one of them end-for-end before connecting them to the battery pack and LED/transistor/resistor bundle the circuit will not work.

If you're sure you've got the toroid winding correct, the next likely culprit is the enameled wire. I chose green and clear because I liked the color combination aesthetically, but in fact the clear-enameled magnet wire can be a bit frustrating to work with because, unlike colored enameling, you cannot easily see whether the clear coating has been completely removed from the wire before soldering. You can substitute 30-AWG red-enameled wire for the clear if you prefer.

Finally, if your circuit is still not working, and you're confident of both your transformer winding and that your solder joints to the enameled wire are good, check the switch again. Then check the transistor and the LED, making sure both are wired in the correct orientations. If the switch is working correctly and the semiconductors are hooked up the right way, it is likely either the LED or the transistor is bad. In this (fairly unlikely) case, the quickest fix is likely to build a replacement LED/transistor/resistor assembly.