Episode: 3781
Title: HPR3781: The Joule Thief
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr3781/hpr3781.mp3
Transcribed: 2025-10-25 05:16:34

---

This is Hacker Public Radio Episode 3,781 from Monday the 30th of January 2023.
Today's show is entitled The Jewel Thief.
It is the 20th show of Andrew Conway and is about 13 minutes long.
It carries a clean flag.
The summary is, using the Jewel Thief to suck energy out of flat batteries.
Hello Hacker Public Radio people, this is McNally, also known as Andrew, and today I'm
going to tell you about a delightful little circuit called the Jewel Thief.
Now the Jewel Thief allows you to extract energy from an otherwise flat battery.
And more remarkably than that, using a flat battery, so like I say I'm using a AA battery,
it was 1.5 volts when it started out its life, it's now down at around about 1 volt.
And even at 1 volt, the Jewel Thief circuit is more than capable of allowing you to light
up a 2 volt LED, or indeed a string of LEDs of even higher voltage than that.
So the magic of this circuit is that it's so simple as well as being able to use a flat
battery to power an LED that requires a higher voltage.
And the secret of the circuit is very fast switching.
So the LED isn't in fact on all of the time, but it's on enough of the time that you
can't really tell that it's flickering on and off.
In fact, the thicker I've built, as far as I can tell, it's flickering on and off 100
kHz, which is 100,000 times a second, and not going to pick that up with my eye certainly.
So it only uses four components, the battery and four components.
So I have a flat battery, and it's worth noting that when I say a battery is flat, you
will not be able to use it to power most things, but of course, flat is relative.
When a battery is exhausted, it's exhausted for the particular task that you put it to.
For example, lighting up an LED torch, running a radio, running a Bluetooth mouth, Bluetooth
mouse, I've got that right in the end.
Once the back voltage goes too low, without some jiggery procury trickery, you can't use
it anymore.
But if what you really just want to do is lighten LED, there is way to coax remaining
energy at the battery.
Indeed, something like 50% of the energy in the battery is still there.
It's just coming out to too low a voltage.
So if you can, put it to another use, that would be great.
So let me just describe the circuit.
So the four components are a resistor.
The one I've got in front of me is a 10 kilo ohm resistor, but I used a 1 kilo ohm resistor
to start with, and that would work too.
I've got a transistor, an NPN transistor.
It's a BC548, which is a totally bog standard transistor, nothing fancy about that at
all.
A yellow LED, I guess, is to work with almost any colour of LED.
And the fourth and most exciting component is myself, one, the inductor.
Now, this is not an ordinary inductor.
An ordinary inductor is the search, which is just a coil of wire around a ferragmagnetic
core usually.
This has a ferragmagnetic core.
It's in the shape of a torus, and most inductors will look like that.
But the clever thing about this inductor is it's got two windings.
One that goes one way around the torus, and another that goes the opposite way around
the torus, and that's the key to how this thing works.
It's actually quite easy, a little fiddly, to widen one of these yourself.
So what I did, I had a dimmer switch from my kitchen light.
It was a main voltage dimmer switch that eventually liked in through some LED driver circuits
to all the lights in my kitchen.
The switch on it became soft, so I had to replace it.
But I don't like to throw things out, so I kept it thinking that might come in handy
one day, and indeed it did, because when I opened it up, I found it had an inductor.
So I dismantled the inductor.
Now, a normal inductor just has, as I say, one winding of wire around the ferragmagnetic
core.
It's a ferrite bead, so it's quite large, actually, it's about a centimeter across,
and maybe half a centimeter deep, and it's like a donut of some ferragmagnetic material.
So I removed just unwound by hand, well, sat watching some television programme,
looked at me a few minutes, to unwind, about a meter worth of the wire that was wound onto the
inductor. Then I took this wire and folded it over. So I have two free ends,
and the other end is now a fold, where the wire is folded in half.
I anchored the two free ends by wrapping it around the ferrite bead once, and then proceeded
to thread through the looped end, the folded end, repeatedly until I had windings that went
all the way around the torus. What's it done that? I cut the wire at the fold, so I now have
four ends, and using a multimeter, the next thing I did was I used a bit sandpaper to
scrape off the enamel, because it's all wire-insulated, it would work in the inductor,
if it was just bare copper wire, it's not, it's enameled wire. So at the ends,
at the four ends at hand, I used sandpaper to scrape off that enamel, so I could make a
electrical contact, and then what I did was I just joined two bits of wire that weren't connected
to each other, and I used a multimeter to figure out which two ends didn't really matter, it doesn't
matter which two ends, as long as they're not showing continuity in the multimeter. So I took
those two ends and joined them together, and that will be the north, north, the positive end of
my inductor I'll use in the circuit. Now it's not an ordinary inductor, so it's got a common positive,
and it's got two negative ends, if you like, and so what you do is you take the positive end,
and you connect that to your flat battery, so the positive terminal, your flat battery, one
of the two negative ends, one of them gets connected to a resistor, and then that resistor gets
connected to the base of the transistor, so the base of the transistor, if you imagine it is like
the tap that you can turn, so you can turn the transistor on and off, which will be crucial in
a moment, so the base of the transistor controls whether it current can flow through the rest of the
transistor. The other negative wire that comes in the inductor gets connected to the collector
of the transistor, and the third and final leg of the transistor, the emitter gets connected
to the negative terminal of the battery, and you also then take your LED and connect the positive
leg of the LED to the collector, which is also connected to one of the end of the negative end of
the inductor, and that's the positive end of the LED, the negative leg of the LED gets connected
to the emitter, and to zero volts, or the negative terminal of the battery. Look at a circuit
diagram, it'll be easier than that, but this is the crucial bit, so what happens is when you connect
all this up, current will flow from this flat battery, and it only needs a small amount, one
volt is plenty, through a resistor, one kilo ohm, ten kilo ohm, doesn't matter, a tiny little
current will flow into the base of the transistor, and then the transistor, after a short delay,
will turn on. When the transistor turns on, it will allow current to flow through the other winding,
so one winding of the inductor remember goes to the base of the transistor to open it up,
the other winding comes down through the to the collector, and at first the current can't
flow that path because the transistor is off, but once some current flows into the base through
the other winding in the inductor, then the current can flow through the transistor, and it
will have a very low resistance, so very quickly a much larger current will start to flow through
the transistor, and a much larger current will start to flow through the other winding of the inductor,
and because the two winding inductors share the same core, and they're winding opposite directions,
what this does is stops the current flowing into the base of the transistor, so what happens,
well when that current stops the transistor shuts off the tap, shuts, and then that main current
path suddenly closes off, and this is the key property of an inductor, if you rapidly change the
current flowing through an inductor you get a spike in voltage, so the voltage the inductor
equation is voltage equals inductance times rate of change of current with time, so if you very
suddenly switch off the current you'll get a spike in voltage, and so even though you've only
got one say a one volt battery, you can easily produce two volts, three volts, four volts are
potentially very large voltage is actually for a brief time, and that allows you to light
the LED, which is connected across the collector and the meter of the transistor,
and as I said at the beginning, this happens extremely quickly, all of that takes place
say a hundred thousand times a second, depends on the value of the inductor, I've got quite a high
value of inductor, just because of the what came out of the dimmer switch, and to a lesser extent
it depends on the resistor, but really the switching, as far as I understand it, is mostly dependent
on the transistor, and the timings inside the transistor, now this transistor that I'm using is
not really designed for switching, and if you look at its data sheet, it doesn't tell you anything
about the timings, other transistors, MOSFETs, they do tell you stuff about timings, but this BC
548, now you don't get any information because it's not really intended for that, it doesn't matter,
it just works, now I have tried to understand the equations that govern this, and it's for
such a simple circuit with four components, it's it's deceptively simple, because the equations
are actually quite complex, so I actually went right back to Maxwell's equations, if you
have done any physics you might have encountered Maxwell's equations, and I was able to derive
the inductor equation in this case, and for those of you that are interested, it looks very much like
the usual inductor equation, so the voltage drop across the inductor is equal to the inductance
times the rate of change of the difference of the two currents that are flowing in the two
windings of the inductor, but actually to try and model what's going on in the circuit,
using that equation is difficult, because as I said earlier, it isn't the inductor and the
resistor that is where the magic is happening, well some of the magic is happening in the doctor,
a lot of the magic is happening in the transistor, and transistors are nonlinear devices, and if you
don't have documentation and the time delays involved, I don't see really how you can model
the dual thief in any simple way, you can certainly model it, but it's not simple, but it doesn't
really matter, because if you want to just light up an LED with a flat battery, well a dual thief is
certainly the way to go, now if you would like to build your own, I highly recommend about half
and half our video by Clive Mitchell, or Big Clive, as he's known in YouTube, he is the person who
coined the term dual thief, which is a lovely little pun, but the circuit itself is much older,
I think maybe the first patterns are almost 90 years ago, but those components are very different
from the ones we would use today, and the ones I've got in front of me, and I think Clive references
a letter in a magazine that was published in 1999, but anyway, for those details and for a
lovely clear description on how to make it and how it works as well, do refer to Big Clive's
every wish as a link I'll put in the short notes, I hope you found that interesting, bye bye
you have been listening to Hacker Public Radio, at Hacker Public Radio does work,
today's show was contributed by a HBR listener like yourself, if you ever thought of
recording podcasts, you click on our contribute link to find out how easy it leads, hosting
for HBR has been kindly provided by an honesthost.com, the internet archive and our sync.net,
on the Saldois status, today's show is released under Creative Commons,
Attribution 4.0 International License.