What's with this Joule Thief thing?

[sophiecentaur]

It's amazing that a name like that can grab the attention of so many people and trigger such a flurry of constructing.

Is looks to me as if it's a way of driving a non linear device like an LED is such a way that it produces more light output than it would from a battery. Is it not just, effectively, a switch mode inverter to produce pulses of more current than the little battery, on its own, would provide?
The Joule Thief doesn't claim to be an over unity device - it just seems to make LEDs look brighter. Why is that a "Thief"? Surely it's just proving to be more efficient (or is it?) than an approximate voltage source (a battery). Novel, but anything more?

Has this been done on a large scale? Domestic lighting is so much more efficient using LEDs that it's hardly worth the bother but there would be many situations where PV / battery combinations could be cheaper for a given light output.

 

[jedishrfu]

wikipedia has a reference that explains how it works:

http://en.wikipedia.org/wiki/Joule_thief

 

[berkeman]

Yeah, it's just a boost DC-DC circuit to let you use up almost all of the energy in a battery cell:

http://www.instructables.com/id/Make-a-Joule-Thief/

In the general case, you would use a buck/boost or SEPIC converter to let you get the LED current from a voltage that can vary from over the LED drop to less than an LED drop.

 

[sophiecentaur]

But nearly all the energy in a battery is available at its fully charged voltage. Most modern cells are pretty well empty of energy once the volts drop. Is it not more to do with the LED characteristic?

 

[cjameshuff]

But nearly all the energy in a battery is available at its fully charged voltage. Most modern cells are pretty well empty of energy once the volts drop. Is it not more to do with the LED characteristic?

It has very little to do with the LED characteristic, apart from the LED requiring a higher voltage than a near-dead battery can deliver...it could be used to drive other circuitry as well. The Joule thief is just an extremely simple switching converter that can produce a few volts from very low voltages. That little energy remains in a mostly dead battery compared to its fully charged condition is rather irrelevant...a few joules do remain, and the Joule Thief makes use of them.

If you must have a practical, non-educational use for the circuit, you can drive LEDs or LED strings with batteries that don't produce sufficient voltage even when fully charged, or with supercaps that have an exponential discharge curve.

 

[vk6kro]

Rechargeable cells mostly give a constant voltage until they run out of charge and then the voltage collapses quite quickly.

However this is not the case for dry cells like alkalines.

See the blue curve for Zn/MnO₂.

If you wanted to run a 1.2 volt LED off this, you would not be able to do this after the cell had used up 40 % of its capacity. This means that the battery would be thrown out with 60% of the charge you paid good money for still in it.
All that lovely power going to waste.

As James suggested, if you have a single solar cell, you can't get more than 0.6 volts out of it, but garden lights use white LEDs that need 3.5 volts or so.
So, these have a very efficient DC=DC converters to convert from this voltage, stored in a NiCd cell, up to about 3.5 volts to run the LED.

 

[sophiecentaur]

But there must be more to it than just saying you're producing a DC-DC boost. If it were as simple as that, you would be putting more current than you needed through the LED when the battery volts were high - which would just be wasted as heat in the LED or series components. If there really is something in it then you would actually get More total light energy out but you'd have to measure it over the life of the battery charge. Merely showing brighter LEDs doesn't demonstrate any advantage.

I'm sure there must be something about the LED V,I,light-output characteristic that self regulates this oscillator to produce a more optimal current supply to the junction. If yoiu look at the data sheets for typical low power LEDs, they specify typical and max current for a nominal voltage and people have more or less left it at that. It isn't too surprising that a slightly different way of supplying an LED (with current pulses) produces more light.
I haven't the facilities (and can't be arsed) to investigate it but I'd bet an enthusiastic and slightly more technical than usual constructor could get some interesting conclusions from an in depth study.

 

[AlephZero]

I'm sure there must be something about the LED V,I,light-output characteristic that self regulates this oscillator to produce a more optimal current supply to the junction.

If you look at the circuit in http://en.wikipedia.org/wiki/Joule_thief it is fairly clear that taking more current through the LED reduces the Q factor of the inductors. The wiki page explains two other nonlinear effects that tend to limit the current.

This is a nice example of nonlinearity being a good thing. The circut will work pretty well for any reasonable component values. The only critical feature is that there must always be a load connected.