Found a circuit for driving 2 white LED's from a single C cell Nicad battery.

I've put it in spice and think it will be a good one. The output changes very little as I switch among the few 30 mA LED models offered in spice.

Any other opinions on this one. I THINK its a relaxation oscillator, but I'm not sure on that.

Kermit

 

Do your simulation again and get a graph of the collector voltage on Q1. A typical white LED needs 3V to 3.6V to work so you should see regular peak voltage in the region of 7V to light the LED.

Looks OK though. Build it!

 

The voltage changes slightly according to the LED model I choose. Ranging between 6.2 and 6.8.

It also works down to 1.0 volt without much loss of brightness. (according to the sim)

I'd really like to get this one going since I have a load of metal can 2222's and even more of the 3906's. The cap is simple, but finding a $free$ source of the 100uH inductor might be a problem. Still they are cheap enough at these current levels.

 

This circuit is classics multivibrator circuit plus "boost converter" (more or less) to light the LED from low AA cell

http://www.electro-tech-online.com/...or-amplify-current-voltage-37.html#post903396

 

The voltage waveform across the two LED's.

 

http://electronicdesign.com/article/components/single-alkaline-battery-drives-white-led5886.aspx

It would seem its a variation of the circuit found here.

 

That is the 2 transistor version of a "Joule Thief", there are 1 transistor versions out, but I'm unsure of their efficiency.

They were all the rage a couple years back, so some very efficient and simple circuits were created. I forgot which was "best", but the 2 transistor one did have a few advantages. The disadvantage I remember was not being able to go to as low of voltage as the single transistor version. If this isn't for running a weak battery to totally completely dead, then that limitation doesn't apply.

 

That's really a rather ingenious circuit. As far as 100 uH you can always "roll your own" so to speak, not too hard to do with a spare toroid coil and some new or salvaged magnet wire. Might not be hard to salvage something close enough from an old motherboard or PC power supply either.

 

The toroid inductors to be used as a "base" for winding your own are probably all around you and you've thrown away several. They are inside every Compact Flourescent (spiral type) bulb, in the wide plastic base.

You only need to saw the base open at the seam with something like an X-Acto saw, and the circuit board can be removed, usually one small and one large toroid are on the board, in addition to a nifty transistor or two.

That's what I used for a source of cores anyway when creating a bunch of "kits" for kids to make the single transistor version, which uses a tapped inductor.

There are about 200 or more instructables on instructables.com for these types of circuits, that was a "phase" they went through a while back.

 

A 10 MegOhm resistor, with 120 turns of 38 AWG magnet wire around it gives me about 28 microhenry. The LED glows about half brightness. (I put in on a bread board just before dinner) Pulses slightly shorter than 1 usec. Going up to 100 microhenry will probably be around 450 turns. I'll hook up the drill to wind that one. That should stretch the pulses out to 4-5 microseconds wide. Should bring the brightness up too.

 

The toroid inductor gave a lot better response with the single transistor version and only around 50 turns total, not sure about performance in the two transistor version.

It might have just been better transformer type coupling for the single transistor version. Might be a cool experiment, but the toroids are a pain to wind.

 

That ends up being essentially a coreless coil so you end up with a lot more wire (and DC resistance) than if you wind it around a true coil form and even though harder to wind a toroid gives you the most inductance per turn.

Really shouldn't be too hard to find something already wound that's close enough. PC repair shops toss out power supplies and older motherboards on a regular basis, tons of other potential sources as well since they're used in most everything nowadays.

 

I've got the guts of a small computer UPS in a box under the desk. Since I've already gotten the soldering iron out, I'll have a look and see what its got.

 

As you move on in experimentation you might even consider buying one of those $20 LC meters from eBay, sure beats pulling out my old heavy equipment and it seems to be rather accurate. Resolution isn't great when you get way up or down in values but 100 uH is no problem for it.

 

Your 100uH inductor needs a ferrite core to reduce the number of turns. It's important to keep the DC resistance as low as possible in this circuit.

If you can find a small ferrite ring about 1cm diameter and wind on about 20 turns of magnet wire as thick as possible. As others have said, an old PC motherboard should have some.

The actual value of inductance is not too important.

 

Your 100uH inductor needs a ferrite core to reduce the number of turns. It's important to keep the DC resistance as low as possible in this circuit.

If you can find a small ferrite ring about 1cm diameter and wind on about 20 turns of magnet wire as thick as possible. As others have said, an old PC motherboard should have some.

The actual value of inductance is not too important.

This is a true statement, oddly enough.

Inductors are usually in the 20% tolerance range.

In a microcontroller world where clocking and bit-banging require pinpoint accuracy, this is weird to hear.

Knowing that resistors set many things regarding IC's it is figured that the more accurate the resistor, the more accurate the circuit.

When dealing with inductance, you have the ability to be "more free" with your choosing as long as you deal with the result properly.

That is one good thing about "slow moving" inductors.. The "snails" of electronics. You have a "range" of voltages where a component will be happy.


Noe RF inductors are MUCH more picky. Above my head.

 

Agreed on the core, the 28 microhenry with 120 turns, has almost 3 ohms of resistance. If I wind a 100 uH one, the resistance would be about 15-20 ohms.

 

This is a true statement, oddly enough.

If you haven't read Jim Williams' treatise on power supplies, you can find it at Linear Technologies, AN35. Here is what he has to say about selecting the inductor:

John

 

HAHAHHA!!

Ill have to give that a read!

 

I've got a winner. Found a proper size ring inductor and 'knitted' myself a proper inductor.

I Wittled the winding stick from a popsicle stick.

nothing there for size ref. ??

smaller than a US dime, (7-8mm across)