Author Options:

Alright.. So can someone explain to me what each part of a joule thief does? I understand the LED, Resistors and battery Answered

So, in a recent Instructable about a Joule Thief, I was slapped in the face with a gap in my knowledge. I have no idea what the toroid DOES, I know it's an inductor... but is it suppose to transform low voltage to higher voltage? What happens If I wind it more? How about less?

What does the transistor in this setup do?

In a recent Instructable, as I was saying, they did it with only a regular inductor.

I'm in the middle of finals too, my brain feels over used...



Best Answer 8 years ago

The toroid acts like a transformer. The transistor part produces pulsed input, which is enough like AC that the thing works like an AC transformer.


so the toroid turns into a transformer, when the transistor is off, electricity charges the toroid with a magnetic field, when the field is strong enough, the transistor turns on and convertets the energy stored in a coil into electricity.

From the point of view of understanding what's happening energy-wise, it's better to regard the toroid simply as an inductor.

With the transistor switched on, the current steadily increases and energy is stored in the magnetic field in the ferrite.
When the transistor is switched off, the current continues... driven by the collapsing magnetic field. Because the transistor is off, the current can't flow that way and the voltage gets high enough to make the LED conduct.

The secondary winding on the toroid is used to provide the positive feedback that drives the transistor, it only takes a small percentage of the power in the system (so some people get upset if you call it a transformer).

The thing that puzzles many people is how it oscillates.
The base of the transistor has a small bias from the 1.5V rail (via the resistor and secondary winding), so the transistor starts to conduct. Current now starts to flow through the primary winding and the changing magnetic field induces a current in the secondary... it's in the same direction as the dc bias so turns the transistor on hard. The transistor is now in "saturation" i.e. has ample base current and so the collector voltage is very low.
The current continues to rise until the transistor is turned off.
That will happen either because the capabilities of the transistor are being reached and collector voltage starts to rise or resistance of the battery and primary winding starts to limit the current or because the toroid has reached magnetic saturation. The latter is the normal mode of operation when the tiny toroid specified in the circuit originally christened "joule thief" is used.
What ever causes the current to stop rising, once it happens, the drive to the base of the transistor is reduced because of the positive feedback and the transistor is turned off hard. As soon as the energy in the toroid has been dumped (mostly into the LED), the negative voltage induced in the secondary ceases and the cycle begins again.

Anyone have some 'scope traces of the joule-thief circuit, you know like, in time domain.  That'd be neat.  That would be something I'd like to see.

The basic form is that of a blocking oscillator.

I ran across a website a while back that treated it pretty well, but I'm afraid I didn't save the link and can't recall the exact search terms (I just did searches on "joule thief theory", ... simulation, ... circuit,  and so forth after reading your request with no success to a depth of 3 pages (sorry, I'm lazy)

The circuit is very basic, so I suppose you could input it into a spice simulator to view the waveforms. It should turn out to be a pair of square-ish waves approximately 180 degrees out of phase, one for the LED and one for the transistor base

I don't think I'll do that. If you don't have the courage to post what you know to this thread, then please just butt out.

I could simply copy and paste what I said in the blog. But then I would have to come back here to add/edit it if I decided I made a mistake or want to make changes. Putting what I said in the blog just simply makes sense, and you're welcome to read it. Sorry if you don't agree.

Writing "NO" and then referring to some offsite blog is pathetic and indicative of someone who's a trol. Get a life. I have no patience left for people like you.

My adenda is trying to help others here. So thanks for being so appreciative of my helpful answer.

BTW, It's troLL.

If that was your intent, you should have posted the content instead of a link to an off-site blog. This site isn't about diverting people to one's own personal site. I find your method of information dissemination suspect, especially since (thank you Nacho) it has been reported that your answer is very similar in substance to what I posted, and you responded to mine by a singular use of "No"

The word was trol when I learned it back in ~1985 or so on Usenet, although it seems to have changed to troll, probably after Al Gore "invented" the internet.

And in my estimation, it fits like a glove.

.  Ermmmmm. Your linked-to page is just a more verbose version of what seandogue said. So wot's, uh, the deal?

If you clicked on the links I gave in my blog, you would get a more technical explanation. One doesn't even get into the math behind the blocking oscillator. I'll even put in a link to my scope pictures.

.  But you, very succinctly, said he was wrong. While his answer may not be in-depth, it is, as far as I can tell, correct.

Careful what you click on. He's a trol and trols usually have an agenda.

Wow! That sure is a nice picture. I thank you for sharing, and/or bringing it to my attention.

We know so little about the person asking the question. If I give him or her a highly technical answer, will he just not understand, shake his head, and close the tab on his browser? If I give a simplified explanation, will he misunderstand? Here's a link to my blog with a simplified explanation with links to others that hopefully will help the reader understand.

I have a pic of a waveform of a conventional JT here in my blog.