Make a Joule Thief

In this week's Weekend Projects I show you how to make a Joule Thief. The PDF file that goes with the podcast is here(450 kB PDF file). So whats a Joule Thief? It's a little wisp of a circuit that allows you to drive a blue or white LED from a low voltage. Normally, if you want to light up a blue or white LED you need to provide it with 3 - 3.5 V, like from a 3 V lithium coin cell. But a 1.5 V battery like a AA cell simply will not work. But using the Joule Thief, it works like a charm. Not only does it work with a brand new battery, but it works until the battery is nearly dead-- down to 0.3 V. That's well below the point where your other toys will tell you the battery is dead, so it can steal every last joule of energy from the battery (hence the name). To learn how to make one, watch the video, which is available in a variety of formats.

It's a little hard to see through my fingers in the video, so here's the detailed view of the coil and winding your own.
	First, here is a selection of ferrite toroids, inductors, and transformers that are suitable for using to make a Joule Thief. Depending on the type that you start with, you may be able to use the existing wires or need to take them off and wind it yourself.
	To wind your own coil, start with two colors of insulated wire and a bare ferrite toroid.
	Take the two strands of wire through the center of the toroid.
	Keeping the two strands together, wrap them around and through the toroid again.
	Keeping the two wires together, make a few more turns through the center.
	Keep winding until you fit as many turns as will fit in a single layer around the toroid, typically 7-10 turns with thin insulated wire.
	Clip the wire leads down. Note that we have two pairs of wires: one coming out the front, and one coming out the back.
	Strip the wire ends. Take one wire from each pair of different color and attach them together.
	Solder the cross-over pair together. This is the "common" point of the coil windings.
In the circuit diagram for the Joule Thief, the common point of the toroid is the connection at the top of the hand-wound ferrite toroid, in the upper right of the diagram. This goes to the positive end of the battery. The other two wires from the toroid go to the resistor and to the intersection of the transistor with the LED. One other detail that you may need to know is the symbol and pinout of the 2N3904 transistor. In the symbol, the part with the arrow is the "emitter", the "collector" is the end above it, that also connects to the LED, and the "base" is the wire leading off to the left, between the collector and emitter. (Also remember that the end of the LED with the flat side and short lead is the end that has the flat bar in the diagram.)
	An actual 2N3904 transistor looks like this.  The pins, holding it so that you can read the text on the flat side are (left to right) Emitter, Base, and Collector. I particularly like this one because it has that little EBC legend on the bottom.
So how does it work? Pretty well, actually. (The technical discussion has been removed.) As a side note, this is not the most efficient circuit around; its beauty is that it works with such a low voltage. So, this is a great circuit to use with a dead or dying battery, and less so for use with a brand new battery.

Electric Origami

This little LED-lit cube is much more than just a paper lantern: It's a translucent and flexible thin-film electronic circuit that hooks up a battery to an LED, limber enough to be folded into an origami box. And the coolest thing about circuits like these? You can make them at home.

In what follows, we combine basic electronics (an LED Throwie) and papercraft (a traditional origami balloon) to make what might be called an "LED Foldie." The circuitry consists of aluminum foil traces, ironed onto adhesive paper such as freezer paper, photo mounting paper, or even a laser printed pattern. Something constructed this way can then be folded so fit an LED and battery to complete the circuit. Besides the throwies themselves, this was inspired in part by the beautiful LED Floaties ,
The first step in designing a three-dimensional circuit like this is to see where the parts go. After that we will unfold the model, draw circuit paths between the points that we want to connect, and go from there. To get started, we first folded an origami balloon, and then inserted the components where we wanted them. The balloon has a convenient pocket on the side for a lithium coin cell, and a single hole that allows you to point an LED into the interior of the balloon. (And you can follow along with balloon folding in this flickr photo set.)
	We knew we wanted the LED to go into the hole in the bottom of the balloon, so we tucked the leads into the folds on the side opposite the battery.
We marked up the locations of the battery and LED terminals on the origami balloon-- while still folded-- and then unfolded our "circuit board." At this point, we have the component locations marked, but no lines drawn between them. The next step is to add those circuitry lines (circuit board wires, or traces) between the battery and LED. One thing to keep in mind for interfacing papercraft to electronics: it's helpful if the circuit traces fold over the leads for the LED in order to maintain good contact.
After connecting the dots (so to speak) we have the resulting layout of our circuit. (See PDF below as well.) Pretty simple here-- only two wires! The two round pads contact the two sides of the battery, and the two angled pads contact the two leads of the LED. The next step is to actually fabricate our circuit board. We've actually found two slightly different techniques that work well, so we'll show you both. First is the "Freezer paper" method (which also works with sheets of dry mount adhesive), where you laminate foil traces to the plastic-coated paper. Second is the "Direct Toner" method, where you print out a circuit diagram on a laser printer and laminate the foil to the printed toner. (Both of these methods of fabricating paper circuitry can be applied in all kinds of other arenas besides origami. Our origami balloon example provides a good demonstration of the techniques!)
METHOD I: The "Freezer paper" method
Lay the circuit layout diagram over a piece of aluminum foil and trace over it with a stylus (a wooden skewer or blunt toothpick works well) to make an indented outline.
Next, cut out your traced pattern. Scissors work well, of course. Be careful not to tear the foil!
Prefolding your paper and comparing to your circuit layout will show you where to lay the aluminum foil pieces out on your paper. Then, use an iron to laminate the foil to the paper. What kind of paper? The easiest (but slightly obscure) choice is "dry mount adhesive," which is tissue paper infused with high-quality hot-melt glue. You can get sheets or rolls of it from art supply places for use in mounting artwork and photography. Much more common and equally workable is freezer paper. Freezer paper is a common plastic-coated paper that you can get on rolls at the grocery store-- look in the section with the aluminum foil. (Place foil on the shiny side of the freezer paper). We used a small hobby iron to fuse the foil to our different papers, but a regular iron works just as well. The dry mount adhesive did not require much heat, while the freezer paper needed the iron to be on high-- that plastic has to melt. We folded a larger sheet of parchment paper over the whole circuit during ironing in order to keep the adhesives from sticking to the iron and other surfaces. We also experimented with waxed paper, which was not sticky enough for the aluminum foil. We even tried ironing copper leaf onto waxed paper, and though it adhered well, it was too fragile and the traces broke upon folding. It would probably work reasonably well in an application where folding isn't required: It was absolutely beautiful and completely unreliable for origami.
Once the foil is adhered to the paper, it is time to refold it.
	Insert the components, and it lights up. If it doesn't light up, try turning your battery around. If it still doesn't light up, make sure your LED leads are contacting the traces. Hint for this circuit: You won't hurt the LED by plugging it in backwards to that little battery, so this is a better method than actually trying to keep track of the polarity.
The LED Foldie naturally wants to sit on the heaviest part, the battery, with the LED projecting into the side of the balloon. The weight of the battery helps keep the circuit connected.
METHOD II: The "Direct Toner" method
Our last breakthrough came when we created a pdf pattern to print out. We realized that you could fuse the foil directly to the toner from a laser printer. You can print out the pattern (laser printers only: no inkjet!) and iron your foil pieces directly to the paper. Caveat: while the foil sticks well to the toner, it isn't quite strong enough that you can just iron on a giant sheet of foil and have it only stick where there's toner, so you still need to cut out the foil shapes, at least roughly.
After printing out the pattern, place it over your aluminum foil, trace the outlines, and cut out the foil pieces.
Place your foil carefully over the pattern, and iron very well, very hot. Be sure to cover your work with parchment paper or you will get toner on your iron.
	When your foil is stuck to the toner, cut out the square and get ready to fold.
	Fold gently so as to disturb the traces as little as possible. They may come loose in areas with multiple folds, but should stay on enough for assembly. Inflate, add battery and LED, and admire the glow. As before, if you have trouble, try turning your battery around and making sure that the leads of the LED are making contact with the foil.
And there it is: a bridge between papercraft and electronics, or perhaps between etch-at-home printed circuit boards and high-end flex PCBs

A Simple and Cheap night -detecting LED Lamp

Here's a simple problem: "How do you make an LED turn on when it gets dark?" You might call it the "nightlight problem," but the same sort of question comes up in a lot of familiar situations-- emergency lights, street lights, silly computer keyboard backlights, and the list goes on. Solutions? Lots. The time-honored tradition is to use a circuit with a CdS photoresistor, sometimes called a photocell or LDR, for "light-dependent resistor."  Photoresistors are reliable and cost about $1 each, but are going away because they contain cadmium, a toxic heavy metal whose use is increasingly regulated. There are many other solutions as well. Look here for some op-amp based photodetector circuits with LED output, and check out some of thetricks used in well-designed solar garden lights, which include gems like using the solar cell itself as the sensor. (Our own solar circuit collection is here.) In this article we show how to build a very simple-- perhaps even the simplest-- darkness-activated LED circuit. To our LED and battery we add just three components, which cost less than thirty cents altogether (and much less if you buy in bulk). You can build it in less than five minutes or less (much less with practice). What can you do with such an inexpensive light-controlled LED circuit? Almost anything really. But, one fun application is to make LED throwies that turn themselves off in the daytime to save power. Throwies normally can last up to two weeks. Adding a light-level switch like this can significantly extend their lifetime.

Here are our components: On top: a CR2032 lithium coin cell (3 V). On the bottom (L-R): the LED, an LTR-4206E phototransistor, a 2N3904 transistor, and a 1 k resistor. This LED is red, blindingly bright at 60 candela, in a 10 mm package. It casts a visible beam, visible for about twenty feet in a well-lit room. We got the LEDs and batteries on eBay, and the other parts are from Digi-Key, but Mouser has them as well. As we mentioned, the last three cost about $0.30 all together, and much less in bulk. The LTR-4206E is a phototransistor in a 3mm black package. The black package blocks visible light, so it is only sensitive to infrared light-- it sees sunlight and incandescent lights, but not fluorescent or (most) discharge lamps-- it really will come on at night. Our starting point is the simplest LED circuit: that of the LED throwie, which has an LED driven directly from a 3V lithium coin cell. (Funny looking example here.) From this, we add on the phototransistor, which senses the presence of light, and we use its output to control the transistor, which turns the LED on.  The circuit diagram looks like this; please ignore the messy handwriting. ;) When light falls on the phototransistor, it begins to conduct up to about 1.5 mA, which pulls down the voltage at the lower side of the resistor by 1.5 V, turning off the transistor, which turns off the LED. When it's dark, the transistor is able to conduct about 15 mA through the LED. So, the circuit uses only about 1/10 as much current while the LED is off. One thing to note about this circuit: We're using a red LED. That's because the voltage drop across the transistor allows less than the full 3 V across the LED. The full three volts is really only marginal for driving blue LEDs anyway, so two-point-something really doesn't cut it. (Might be able to work around that with a cheap FET-- haven't tried yet.) And now, let's build it. You can certainly put this together on a breadboard, but there's something more satisfying about the compact and deployable build that we walk through here.
First get the transistor and the resistor. The pins of the 2N3904 are called (left-to-right) Emitter, Base, Collector, when viewing it from the front such that you can read the writing. We're going to solder the resistor between the leads of the Base and Collector of the transistor. Unusual part: hold the resistor with its leads at 90 degrees to those of the transistor while you solder. Stay safe when you do this: Use Mr. Hands. After soldering, clip off the excess resistor lead that is attached to the transistor base (middle pin), as well as the excess length of the collector pin.
Next, we add the phototransistor. Note that it has a flatted side, much like an LED does. This pin on that side is the collector of the phototransistor. Solder the collector (flatted side) to the middle pin (the base) of the transistor, again at 90 degrees. The other pin of the phototransistor, the emitter, is left unconnected for the moment. (Here is an alternate view of what that should look like when you're done.) Finally, we need to add the LED. To do so, we need to know which side is the "positive," or anode side of the device. Regrettably markings of LEDs are not consistent, so the best way to be sure is to test it with the lithium coin cell-- put the LED across the terminals of the cell and, when it lights up, note which side is touching the (+) terminal. (Usually, it's the one with the longer lead.) Solder the "positive" lead of the LED to the emitter pin of the transistor-- it's the one on the left, which doesn't have anything soldered to it. Trim away the excess lead of the LED that goes past the solder joint. Solder the other pin of the LED (the "negative" pin, or cathode) to the emitter of the phototransistor, the pin on the non-flatted side, which does not have anything connected to it yet.
By this point, there are only two pins sticking down below the components: One that goes to the resistor and collector (rightmost pin) of the transistor, and one that goes to the emitter of the phototransistor and to the cathode of the LED. To test the circuit, squeeze the coin cell between these two terminals, positive side goes to the lead touching the resistor. You can't see the LED on here because these photos were taken with incandescent lighting-- it wouldn't turn on.
Bending the leads to contact the lithium cell a little more reliably, you can try it out a little more easily. In the photo on the right, I cupped my hand over the circuit-- so the LED turned on.
To make this into an actual "throwie," you still need to add some tape and a magnet, but that's quite easily done. This one makes a pretty good nightlight attached to the top of a doorframe-- when the room lights are off, it shines a bright, bright spot on the ceiling. Where to go from here? While this little circuit can do something on its own, it would probably also be happy as part of a larger circuit. At a minimum, note that if you work with batteries that have lower internal resistance than the lithium coin cells, you should place an appropriate resistor in series with the battery before trying to operate this circuit-- or else you may put too much current through the LED. Certainly, this is one of the easiest and least expensive ways to control an LED with a photosensor.  You could also consider crossing it with some more extreme mods, like the Talkie Throwies that know Morse code, or for more extreme hackers, bagel throwies.