Introduction: Joule Thief Charger

Picture of Joule Thief Charger

Let your dead battery give life to another! An open circuit Joule Thief can put out 50 or more volts. Enough to charge a AA or AAA Nicad or NiMH rechargeable battery.

Step 1: Make the Charger

Picture of Make the Charger

Use this schematic to build a standard Joule Thief circuit with the added diode.

My joule thief uses twisted network wire passed through a small ferrite core. I use 6 turns of wire. You can find a core from a burned out compact fluorescent bulb. You can see how others have wound the coil and built the Joule Thief, since so many have done so. Just add a diode and LED in series with the charging battery. The LED is useful as a charge indicator.

A high-speed schottky diode would be the most efficient. The 1N4005 was handy at the time and works.

Step 2: Open Circuit Voltage

Picture of Open Circuit Voltage

I got a voltage of 52.6 volts coming out of the joule thief circuit without any connection to a load.

More than enough voltage to charge a rechargeable battery.

Step 3: Charge Current

Picture of Charge Current

I measured 9.33 milliamps into a shorted load. This is the charging current to the cell.

Step 4: Charger in Action

Picture of Charger in Action

The donor cell on the left has a voltage of 1.057 volts. This battery is charging the battery on the right.

Step 5: Charging Up

Picture of Charging Up

The receiving cell voltage is 1.375 and is getting a steady charge.

Step 6: Plotting the Results

Picture of Plotting the Results

I connected My two channel data logger and monitored the voltages of both batteries overnight. The life force from one cell goes to the other. Its like the Alien Healing Device from Babylon 5.

Steps 7 and 8 show the results using Excel.

Step 7: First 8 Hours

Picture of First 8 Hours

Top red trace is the cell being charged. The cell voltage has stabilized and is accepting a charge.

Bottom blue trace is the donor cell. Notice how the voltage is slowly dropping off. The life force of the donor cell is slowly slipping away.

The Joule thief output jumped between two voltages and is a bit steppy. Nothing is perfect.

Step 8: Last Chart

Picture of Last Chart

The final 5.7 hours before I stopped recording data.

The charged cell is still receiving a charge and the donor cell now has dropped to about .62 volts. The Joule Thief is still running.


Nakagwa Labs (author)2016-05-09

what if i connect a 9v alkaline battery instead of the 1.5v battery

elhnad (author)2015-03-19

this is really awesome....but just to make sure i've understood this right. So even tho this outputs 50 or so volts, when it actually is closed and there's a load receiving the charge, it will self adjust to the recipients voltage, just enuf so current (a few dozen milliamps) can flow from one to the other.

does the current change by a lot or is it always a small number and so will require overnight charging.

my practical use of this is to rejuvenate/restore almost dead lithium ion batteries.

botronics (author)elhnad2015-03-20

The voltage will adjust itself to the charging of the battery. The current is so small that is more of a trickle charger. Because there is no voltage limit, may not be safe on a lithium cell. Although many lithium cells have built-in overvoltage, undervoltage and overcurrent limiting devices inside. Monitor the voltage as its charging and establish a cut-off point.

Yerboogieman (author)botronics2016-03-14

IMR lithium batteries have this protection while similar ICR batteries do not.

zorcy (author)2015-09-02

So, what would keep this from working with, say 6 circuits, running parallel, with electrodes in the lake, to charge a 12v battery? Since the water is not alkaline, it would have a very small charge, very. But a small charge would boost up?

botronics (author)zorcy2015-09-18

I guess parallel operation is possible. The diode at the output would keep the other circuits from shorting each other out. You can eliminate the led and just keep the one diode. Interesting idea!

gmccusker (author)2014-08-17

Hmm, replace donar cell with a small 1.5v solar cell and the load with 3 1.2v Ni-MH batteries in series. Workable? I don't care much for how long it would take to charge the batteries, more interested to know if I can get a smaller cell to charge a higher voltage considering that's what the joule thief does in the first place.

PhilKE3FL (author)2014-07-21

What was the starting voltage and ending voltage of the NiCd battery?

blinkyblinky (author)2011-12-06

How many turns go BEFORE the 6 turns?

botronics (author)blinkyblinky2011-12-07

You wind the turns through the core as a pair. So both are 6 turns.

blinkyblinky (author)botronics2011-12-07

So technically If I wond 30 turns would it make any difference?

botronics (author)blinkyblinky2011-12-08

The turns is not critical for it to work. The frequency will usually go down with more turns. Experimentation is in order.

blinkyblinky (author)botronics2011-12-08

Thanks. On my digital meter, when I measured and checked for voltage I got 0.1 volts less than the battery had. What happened?

PhilKE3FL (author)blinkyblinky2014-07-21

You can go to my website to see how to optimize the Joule Thief circuit. I did a boatload of tests to see what was "best." I consider it to be the "best" when a circuit got the dead battery to last as long as possible, in other words it used the least amount of current. Brightness was not as important but the LED still had to be about as bright as a 3V source using the same LED. This is NOT the only "best" there is, it is simply the one I decided to base my test results around. I figured that as long as the circuit could be used as either a flashlight or a night light it was good, bright enough.

Increasing the number of turns, up to a certain number, which differed for different types of wire and inductor cores, increased the efficiency and then if you put more turns on things started to get worse again. I noticed little difference in peak pulse voltages but they were there and with the circuit running they never seemed higher than ~ 2.8 Volts as I measured with an O-scope. The lowest voltage measured was ~ 1.95Volts, for the white LEDs. Other LEDs, as the blue and amber LEDs showed, had different turn on voltages.

I tested three different transistors, three different colored LEDs, four different inductor toroid cores from Digi-Key (all under $0.50 each for single pieces) at least three different wire thicknesses, and many different number of non-easy type of loops. (The easy Joule Thief uses the same number of loops for both sides of the circuit, I tested different number of loops, found which part of the circuit should use the higher and lower number of loops and found the optimum number of loops for each to achieve the lowest current usage while maintaining a sufficient brightness.)

Digi-Key Toroid Part #s I used for my tests:

495-3851-ND, 495-3849-ND, 495-3848-ND, & 495-3874-ND

captainchaoss (author)2014-05-31

Total Noon here. this is probably a silly question, but could I use "c" cells as the donor(s) ? They are the same voltage, with larger capacitance, so to my understanding, they will still work, right ?
Through my job, I have a constant supply of partially discharged batteries, mostly "c" cells. It stinks to see most devoted to disposal. I would rather build a few of these and vampire the energy to some rechargeable !

PhilKE3FL (author)captainchaoss2014-07-21

Yes, in fact the circuit will work with any size battery and with voltages probably a good deal higher than 1.6Volts even, as long as the transistor, and other parts, can stand the voltage. Understand that the thing that keeps the battery, the NiCd under charge, from blowing up or having a meltdown due to the excessive voltage is the fact that these are very fast pulses. NiCds WILL meltdown if over charged even at a trickle rate, considered to be ~ 1/20 - 1/10 the AHr rating of the battery. They start to heat up, get extremely hot eventually and self destruct. I have never known them to start a fire but I have seem them melt plastic cases such that the case with dead batteries had to be thrown out and replaced batteries and all. I saw this happen to a pair of 5W CB walkie talkies and I had to replace the entire case of the walkie talkies as well as the NiCds.

NiCds are generally considered to be fully charged when they reach a charge voltage of 1.35 - 1.4 Volts and should NOT be charged further than that to avoid the overheating or meltdown condition.

kyooby (author)2014-03-27

Hi.Can I light up 12 leds(bright light not low) with this circuit? I sow somebody on youtube hooking up 53 leds lightning for about 15 hours.Can this be true or fake?

argha halder (author)2013-08-12

its impossible because leds will burn to ashes at 50 volts and even the best joule thief with 6 turns on coil wont give 50 volts.may be you put the wrong dial on the voltmetre to measure.but except it is a great instructable.

botronics (author)argha halder2013-08-13

No, No you are so wrong. You can place an led in series on high voltage circuits as long as the current is limited and not reverse biased. Its done all the time in led light bulbs that run on 120 vac. The diode in series protects the led from any reverse voltage and the joule thief current is limited. The meter clearly shows 50 volts. I kid you not.

PingoBags (author)botronics2013-12-27

The high volts if from the zener diode you're using right? I have some different zeners I've tested and a few of them output about 50 volts too.

argha halder (author)botronics2013-08-13

oh then sorry .i thought it to be mistake .but now i understood and thanks for both the comment and instructable.

rodan1956imod (author)2013-11-22

The network cable used for networking has 4 pairs. Does it make any difference if I use one of the 2 pairs used in networking (serial data comm) as I observed in your Joule Thief you use the blue / white blue pair which is for telephone connections. This pair is loosely twisted as compared to the 2 pairs used for networking.

harmhero (author)2012-07-16

Strange, my Joule Thief works fine, white LED is burning, but I get the same output voltage (without load) as my battery, 1.5 volts. Doesn't make sense really because the LED is burning with the JT but not with the battery.

namzer0 (author)2009-12-01

how much is the current draw? when not charging/open circuit and when charging?

mitchiko (author)namzer02011-12-06

You cannot measure the current draw in the open circuit only the voltage.

freenergyfuture (author)2011-11-07

I am very interested in Joule Thief's and i am curious how to place a variable ohm resistor into the circuit? please reply asap.

agatornz (author)2011-10-31

i like the idea of this for hiking - using a combination of nimh rechargables and alkalines for my electronics - it would be great to be able to charge the former - is there some way you can include a simple "charge level indicator" in the circuit so one would know how well charge the rechargeable is?

botronics (author)agatornz2011-11-01

Your better off using a solar cell for charging. You can attach one to your backpack. A picaxe mcu could be used as part of a circuit to measure amp-hours when charging.

twighahn (author)2011-04-21

can u put this in layman's terms?

ab3t (author)2011-03-17

What if the batteries have a voltage of 3.7 volts used? What should be changed?

botronics (author)ab3t2011-03-18

Nothing, the battery voltage will adjust itself. Just feed it electrons.

mikedoth (author)2010-02-18

Stupid question but why doesn't the LED blow from the high voltage alone?

SwingNoob (author)mikedoth2010-03-04

because your using a 1k resistor as well

mikedoth (author)SwingNoob2010-03-04

I'm still learning but will a 1k resistor drop 52.6 volts down to 1.5 to charge the battery? How about a voltage regulator?

botronics (author)mikedoth2010-03-04

No regulator is needed if you are charging a battery.  The voltage is high because it is an open circuit, but the current is low.  When the battery is connected, the voltage will drop to the voltage of the battery. Same thing with a LED connected to the output.  The voltage drop of the led will keep the voltage down. Think of the LED as a zener diode.

leodahsan (author)botronics2011-02-18

more like a constant current source? ;)
btw.. what are better options than the 2n2222? I need something that generates 40v @ 20uA (the more the better ;P) from anything around 1 to 3.4v. Anything in this range, but I don't know how to start..

knektek (author)2011-01-03

is this like a campain for charging up batteries. 'All dead batteries please save a ni-cd's life today and donate 500 milli amps each month'.

dasimpson1981 (author)2010-12-23

so hwne the battery is charge dose the led go out ?

dasimpson1981 (author)2010-12-19

i have tried this and well the charge is false the voltage is right but there is never any ampage to run anything i have tried charging then running down over and over and dose not seem to improve or work

CyborgGold (author)2010-09-13

you should have done this logging with the rechargeable battery starting out dead... this one just hovers at full charge... actually it appears to drop slightly in charge.

botronics (author)CyborgGold2010-09-13

I discharged the battery into a load till its voltage was .9 volts. That is the recommended lowest you can go without damage to the cell. When a battery just sits there open circuit, the voltage will rapidly rise to a higher level, which it did. During charge the voltage will only rise slowly. The battery was no way fully charged. It was only a test to see if the circuit would raise the voltage and allow current to flow in a charging direction. Given enough time and dead batteries, the battery will charge.

flying pie (author)2010-06-16

circuit is too small and can u add a solar panel to the "charge donor " so u can charge it without using up a battery to charge a diffrent battery

A good name (author)flying pie2010-09-12

Are you serious? The circuit is too small? Adding a solar panel would defeat the purpose. You're using a next-to-dead battery to charge another battery.

botronics (author)A good name2010-09-12

Next-to-dead batteries still have a lot of joules left. I have nite lites that run for months using next-to-dead batteries. We are just transferring whats left to something more useful.

tinstructable (author)2010-09-06

You know you don't need that diode... the led is acting as a diode. Light Emitting Diode. Just a tip.

botronics (author)tinstructable2010-09-12

There might be a negative going spike as the circuit oscillates that can harm the led. I could just use a diode and forget the led, but the led acts as a indicator.

electrofreak (author)2010-04-23

What is the frequency range for these types of circuits?I'm curious because if the frequency is high enough a fast type diode could be a better choice. Also it will help in determining the size of the capacitor if one decided to filter the output.

botronics (author)electrofreak2010-04-24

The frequency is in hundreds of kilocycles. Watson's e blog has a lot of information on joule thief circuits.

lolzertank (author)2009-05-19

Wouldn't the led eat up a lot of power? Assuming that its a red led (2v), then the total load is 0.7v (diode) + 2v (led) + ~1.3v (battery) = 4v * (your charge current). Without it, it would be 2v * (your charge current) total, increasing the efficiency dramatically. Of course, since you're using dead batteries, this might not be important at all.

webmasterpdx (author)lolzertank2010-04-03

 Shottky diode shouldn't be .7V. Probably could use a germanium too. Either gives you capability of working with lower input voltage. personally, I'd get rid of the LED too (maybe have a pushbutton to see if it's working. You could also pass it through a capacitor to remove the DC component and the try a full wave rectifier. Might get better results....I'm not sure.

About This Instructable




Bio: I like to tinker and experiment with electronics, robotics, programming, and photography. Along with my latest interest in Steampunk.
More by botronics:Convert 3-AAA Flashlite to LithiumPole Grenade PropSteampunk Vacuum Tube Night Light
Add instructable to: