Reconfiguring joule thief for different input/output

Fallingwater

Flashlight Enthusiast
Joined
Jul 11, 2005
Messages
3,323
Location
Trieste, Italy
I have a small 2xAAA, 3-5mm-LED light (this one) that has no driving circuitry whatsoever - it just hooks the LEDs straight to the two cells. This, of course, gives abysmal output.
I'm looking into building a small circuit to drive the three LEDs at about 30 or 40 mA each from fresh cells. Now, I know how to build a joule thief, but the stock design is only good for 1.2/1.5V input and output for one LED.
How do I modify it so that it works for my application?
Thanks.
 

Mr Happy

Flashlight Enthusiast
Joined
Nov 21, 2007
Messages
5,390
Location
Southern California
The circuit is very malleable. I'd suggest trying it with 3 V input and put the three LEDs in series on the output. Experiment a bit and see how it goes. You might have to try different numbers of turns and different cores for the windings to see what works best.
 

Cemoi

Enlightened
Joined
Jan 23, 2008
Messages
527
Location
France
You might try this circuit which is more efficient than the plain Joule thief. I used it to modify a bike light, see this thread.

Although the above page is in Czech, the figure captions are in English and you should have no trouble building the circuit. You could try to decrease the value of R1 to set the output current to your liking (I used 8 ohms to get 18 mA which makes my two Nichia GS in series very bright).

An English-speaking Czech colleague kindly translated the page for me (but I don't have the result here at hand). The most interesting information is that you can make the circuit even more efficient by replacing the BD433 transistor by a more bulky BD243C.
 

Fallingwater

Flashlight Enthusiast
Joined
Jul 11, 2005
Messages
3,323
Location
Trieste, Italy
The circuit is very malleable. I'd suggest trying it with 3 V input and put the three LEDs in series on the output. Experiment a bit and see how it goes. You might have to try different numbers of turns and different cores for the windings to see what works best.
Will do, thanks :)

Cemoi: I'm not sure whether your circuit can fit in the small space available inside the light. Pretty sure a simple joule thief can though...
 

Cemoi

Enlightened
Joined
Jan 23, 2008
Messages
527
Location
France
I'm not sure whether your circuit can fit in the small space available inside the light.

I managed to make it fit on a 0.6"x1.4" (1.5cm x 3.6 cm) PCB but it was tricky, and this is maybe still too big to fit in your light.
 

Fallingwater

Flashlight Enthusiast
Joined
Jul 11, 2005
Messages
3,323
Location
Trieste, Italy
You might have to try different numbers of turns and different cores for the windings to see what works best.
Could you specify a bit? Do I just wrap more/less wire (and what effects does it have)? Do I have to wrap a different number of windings for the two wires?
 

Mr Happy

Flashlight Enthusiast
Joined
Nov 21, 2007
Messages
5,390
Location
Southern California
No, I can't really tell you much more, I don't have enough knowledge of the subject. For more information you would have to read up on inductor design and winding calculations.

What I do know is that there is a bit of an art to tuning such circuits. You can't always tell how they are going to work until you try them out, and the number of turns on a coil is something you may have to experiment with to get the best results. For a typical ferrite core or bead, I would guess that a number of turns between 20 and 60 would be a reasonable range to experiment with. To reduce the number of variables, I would keep the number of turns equal on both windings.
 

FRANKVZ

Enlightened
Joined
Jul 11, 2004
Messages
361
Location
Western PA
You might try this circuit which is more efficient than the plain Joule thief. I used it to modify a bike light, see this thread.

Although the above page is in Czech, the figure captions are in English and you should have no trouble building the circuit. You could try to decrease the value of R1 to set the output current to your liking (I used 8 ohms to get 18 mA which makes my two Nichia GS in series very bright).

An English-speaking Czech colleague kindly translated the page for me (but I don't have the result here at hand). The most interesting information is that you can make the circuit even more efficient by replacing the BD433 transistor by a more bulky BD243C.

Translated: http://translate.google.com/transla...z/ledlight/m2d.htm&sl=cs&tl=en&hl=en&ie=UTF-8
 
Top