• Please review our updated Terms and Rules here

transistor to work with a joule theif

L

linuxlove

Veteran Member
Joined
Jan 11, 2009
Messages
1,018
Location
Auburn, AL
what kind of transistor will work to make a joule thief? i have a lot of transistors on old boards, but i don't know what type of transistor will work.
 
right, well, i read somewhere that any NPN transistor should work. I tried a C945P34E transistor. The last transistor got very hot, so it was broken, and i used this which seems to be an NPN transistor. now when i hook it up to my AA battery, the transistor doesn't get hot, but the LED doesn't light. ideas? also i'm using aluminum wire on the ferrite bead instead of copper. if i should use copper, i'll pull some out from an old cable i have (a dog chewed on it so it's no use anymore).
 
A general-purpose switching transistor, such as a 2N2222 should work.

I wasn't aware that you can get insulated aluminum wire, however. If you're going to wind your own inductor, you need to use what's known as "magnet wire"--basically copper wire with a thin enamel insulation. It looks like bare wire, but it's not.

You might even get better performance if you can get your hands on some litz wire.
 
hmm. so any transistor in the 2N series should work? i'll have to look out for those types. as for the wire, i might have run across some of that type, but it got recycled cause i didn't know about joule thiefs then.
 
linuxlove said:
hmm. so any transistor in the 2N series should work? i'll have to look out for those types. as for the wire, i might have run across some of that type, but it got recycled cause i didn't know about joule thiefs then.
Click to expand...

No--the 2N series of transistors basically covers all 3-terminal transistors, from the simplest germanium point-contact 2N105 to the hugest stud-mount power transistor.

You want a general-purpose switching NPN transistor with sufficient gain to make sure that the thief operates at low voltages and sufficient high-frequency gain to make sure that your little doodad operates efficiently (i.e. don't use audio transistors).
 
ok, i'll have to experiment with my transistors. also, for it to work, does it have to be a blue or white LED or can it be any LED?
 
i'll have to try that once i get the thing working. I also made a new coil using that cable wire -- it was actually aluminum, not copper. I gout about 10-12 turns out of it and... no go.
 
Is the wire insulated (i.e. enameled)? If not, the windings will just short out against one another. There are plenty of places to scavenge insulated magnet wire. Old transformers, CRT deflection yokes, etc.
 
the wire is insulated. as for magnet wire, i have an old speaker here i could pull it out of, but would un-coiling the wire from the speaker damage the insulation? pulling the wire out of a CRT is not feasible because

a) we don't have any broken monitors to pull it from
b) i really don't know what i'm doing, so even if i did discharge the CRT, i risk breaking something and hurting myself bad.
 
You get the wire not from the CRT, but from the yoke on the CRT neck. Surely there must be a junked TV or two in your area.

Another source is to crack open an old transformer-type "wall wart" and use the wire from the transformer. I don't recommend using speakers, as they've usually been abused and the insulation is toasted and brittle. Besides, you can't solder to aluminum.
 
Last edited:
yeah, i have a few old transformers around here. i could probably get the magnet wire from a TV, but i'm not comfortable poking around inside a CRT, even if it is discharged.
 
Hi
I looked at the circuit on the web. I thought I'd
list a few of the possible problems:
1. Transistor pins not connected right ( base emitter and collector are in
different orientations for different transistors ).
2. Transformer winding not in correct phase ( it makes a difference which
wires are tied together. The dots in the diagram have specific meaning ).
3. The resistor may be to small or too large a value for the particular
transistor you are using.
4. The LED has its leads reversed ( the long lead should go to the negative
of the cell ).

I should also note that connecting a capacitor my damage the LED and
or blow the transistor. The current setup uses the coil as a current limiter.
With a capacitor added, it is no longer a current source but changes
into a voltage source. LEDs don't like this. Also, you'd need a diode
to block the current flowing back into the transistor when it turns
on.
Dwight
 
Dwight Elvey said:
Hi
I looked at the circuit on the web. I thought I'd
list a few of the possible problems:
1. Transistor pins not connected right ( base emitter and collector are in
different orientations for different transistors ).
Click to expand...
if i had the pins wrong, wouldn't the transistor get hot when voltage was applied?

2. Transformer winding not in correct phase ( it makes a difference which
wires are tied together. The dots in the diagram have specific meaning ).
Click to expand...
That may be why it isn't working. however, i am no electronics engineer and can't figure out waht anything is on an electrics diagram like that.

3. The resistor may be to small or too large a value for the particular
transistor you are using.
Click to expand...
maybe. i have a bunch of other size resitors here i can use; i was just using a 1K resistor.

4. The LED has its leads reversed ( the long lead should go to the negative
of the cell ).
Click to expand...
nope, i've tried the LED both ways and it still didn't come on.
 
You must log in or register to reply here.
Back
Top