Joule thief general question on how they work

[antknee]

Joule thiefs look pretty good and I understand the principle, for example take a single 1.5V AAA battery and with minimal components get 5V out.

What are the limitations?

I mean I'd like to get 5V and 500mA out of the AAA battery but that seems a bit too good to be true.

Regards.

 

[colin55]

I have a circuit that produces 5v at 70mA:

The parts cost $3.00

 

[Grossel]

Colin: How do you make operating voltage (5v) to the microcontroller first time it start? Will it run?


antknee:
I found a schematic on the net involving several transistors and coils:
**broken link removed**
Source: **broken link removed**

The limitation in a regular JT is:
* Maximum current capacity for the battery.
* Resistance in coil. Lower is better.
* Voltage drop Vce in transistor.
* The duty cycle that is providing isn't optimal and also may providing long rise and fall time for Vbe - wich make transistor much less efficient. Microcontroller fix this issues better.

If the battery isn't the limiting factor, the schematic above might be a solution.

 

[colin55]

The micro works down to 0.8v

 

[antknee]

Thanks for the replies. It is interesting that both of you have said "I have a better way". I instinctively look for an IC because I think it will be easier when that isn't necessarily the case.

I found the NCP1400 joule thief. It will give 5V at 100mA from a 1.5V battery. I'm guessing 100mA is about tops.

 

[crutschow]

I mean I'd like to get 5V and 500mA out of the AAA battery but that seems a bit too good to be true.

The required battery current is multiplied by the ratio of the input and output voltages. Thus for 5V output and 1.5V battery voltage, the required battery current would be 5/1.5 * .5 = 1.7A or more depending upon the efficiency of the joule thief. Thus an AAA battery would not last long. You would need at least a D cell to provide that amount of current, which could last perhaps a couple hours.

 

[antknee]

I'd like 5 Watts for 5 minutes. That has given me something to think about. Regards.

 

[MrAl]

Thanks for the replies. It is interesting that both of you have said "I have a better way". I instinctively look for an IC because I think it will be easier when that isn't necessarily the case.

I found the NCP1400 joule thief. It will give 5V at 100mA from a 1.5V battery. I'm guessing 100mA is about tops.

Hi there,


If you are really serious about this project then you should look into the Zetex line of converter IC chips. These things are quite easy to use and can run off of 1.5v all the way down to 0.7v input.

A long while back a bunch of us from another forum got together and created a design for a high powered LED driver using one of the Zetex chips. Lots of people ended up building these things and everybody was pretty happy. A very nice gentleman from CA was good enough to fund the whole project and even board designs came out of it, all about the size of a quarter dollar (USD) or a tiny bit bigger. I probably even have those designs still around.

Part number ZXSC300 for example, but since then they came out with more advanced versions too.

I also posted a complete analysis of the Joule Thief circuit way back around year 2000 on How Stuff Works, but i dont know if it is still there or not.

 

[Sceadwian]

A micro controller that works down to .8 volts, what controller is that?!

 

[MrAl]

A micro controller that works down to .8 volts, what controller is that?!

Hi,

In a word, "Any one bootstrapped". Oh wait, that's three words

 

[antknee]

I have found a boost converter that can output 1A /60V, (maybe not together). It is complicated though with 10 pins and perhaps a dozen components.

 

[audioguru]

A Joule Thief and most other simple voltage-boosting circuits use an alkaline battery that starts as 1.5V or 1.6V when new, then its voltage drops as it is used. The output power of the circuit also drops as the battery voltage drops so an LED will be dimming slowly all the time.

 

[amando96]

I think one of the atmel attiny works with voltages that low.

 

[Sceadwian]

Well the bootstrapping makes sense, colin didn't use quiet the right words though =) The micro controller will continue to run while the battery goes as low as .8 volts. The MCU will need more than that. You must have mis-read amando, the picopower tiny's that I've seen work down to 1.8 volts. I don't know of any cmos based micro controller that can run on .8 volts, they require at least two diode drops for some reason.

 

[amando96]

Yeah, i did misread it, however, there is this feller:

"ATtiny43U
Ultra low voltage 8-bit AVR Microcontroller for portable and other battery operated applications.

Integrated boost converter automatically generates a stable 3V supply voltage from a low voltage battery input down to 0.7V."

So the micro itself cannot operate at 0.7...

 

[Sceadwian]

Never seen those before amando, they're new as of this year, nice though. I wonder at what niche they'll fill though, anything powered by a single alkaline or NiMh battery is going to be pretty cheap.
Personally I think it's easier to just use a single lithium cell. They'll run any conventional low voltage cmos IC's, they're not difficult to charge and are very safe and affordable now.

I've made joule thieves before, and I find little they're useful for that couldn't be made easier using the proper voltage cells other than units that can make an LED glow by taking the last bits of juice out of dead batteries. They're more novelties that useful.

 

[BrownOut]

A Joule Thief and most other simple voltage-boosting circuits use an alkaline battery that starts as 1.5V or 1.6V when new, then its voltage drops as it is used. The output power of the circuit also drops as the battery voltage drops so an LED will be dimming slowly all the time.

I once tried to figure out the "magic" of the joule thief, how the heck they get more life from a used-up battery. I thought I had it at one time and it was due to the saturable inductor, and the fact that it will always switch at the same voltage, even as the battery voltage goes down. I even posted my "discovery" But I realize I was on the wrong path, and there isn't really any way a JT can get any more from a used up battery. The problem you describe makes it pretty much unusable, unelss..... one can control the pulse rate when the battery is fully charged, so that he only makes the power his applications demands. I'm thinking an emitter coupled pair with the average output voltage fed into the base, and the coil feeding the other. That way the output voltage determines the pulse rate, and not the battery voltage ( using ECP transistors to keep it simple as originally conceived)

 

[Sceadwian]

What you're describing is a basic regulated boost converter Brownout. You would just have to make sure that the entire circuit worked on at no point more than a single diode drop, or you'd loose most of the usefulness of a typical Joule thief.

 

[BrownOut]

I'm talking about adding very simlpe PWM control to the basic joule thief, which would IMO, deliver the unsubstantiated promises of the same (thinking of a real simple LED blinky here). It would not change any drop in the convertor path. I don't know if it would work or not, and I probably won't take the time to fine out. But I like to think about it. The JT is the simplest convertor there is, and my proposed improvment wouldn't change that by very much.

Maybe if I feel energetic, I'll try to make a sim. I doubt anyone would be interested in seeing that though.

 

[acmefixer]

JT simplest converter? Here's a simpler one

The Joule Thief using a transistor, resistor, coil and LED is the second simplest. I replaced the transistor with a TN0702 low gate threshold MOSFET and eliminated the 1kΩ resistor. Instead of four parts (not counting the battery), there were only three parts. In addition, it would work at a much lower voltage than the silicon transistor. See my blog for more info.