Joule Theif

[BrownOut]

Can someone explain this? I just don't see how a simple switcher can make a battery work past its normal useful life. Is it because a dying bettery acts sort of like a capacitor, and charge appears on the terminals during the swticher's "off" time???? Or does it have more to do with the coil reversing the voltage and "pulling" the charge out.

I wanted to ask on the last thread about the Joule Theif, but couldn't find it.

BTW, I've visited several sites, and have not found the information.

 

[ericgibbs]

Can someone explain this? I just don't see how a simple switcher can make a battery work past its normal useful life. Is it because a dying bettery acts sort of like a capacitor, and charge appears on the terminals during the swticher's "off" time???? Or does it have more to do with the coil reversing the voltage and "pulling" the charge out.

I wanted to ask on the last thread about the Joule Theif, but couldn't find it.

BTW, I've visited several sites, and have not found the information.

hi,
I know that you use LTS.

This is a JT sim I tried a while ago.

 

[Nigel Goodwin]

Can someone explain this? I just don't see how a simple switcher can make a battery work past its normal useful life. Is it because a dying bettery acts sort of like a capacitor, and charge appears on the terminals during the swticher's "off" time???? Or does it have more to do with the coil reversing the voltage and "pulling" the charge out.

It's simply that a 1.5V cell, on it's own, can't light an LED - it doesn't have enough voltage. The Joule thief increases the voltage using a crude transformer in a self oscillating circuit, enabling a LED to be lit from a single cell. Because this circuit operates down to a very low input voltage, it can work with batteries considered flat for other uses.

It only really works because of the low current required to light an LED, as there's not that much energy left in a battery considered 'flat' - it's rather like licking the last of the cakemix out of your mothers bowl

 

[BrownOut]

Thanks Eric. This is going to be a challenge to simulate, because the coils need to be coupled, and I think we need to specify a saturation flux density for the core. It's not impossible, I just need to ramp up my Spice tools.

 

[jta]

My impression is that it acts like a DC/DC converter, able to up the otherwise unusable low voltage of a dying battery (say .5V) to something usable (say 1.5V). Therefore, it extends the life of a battery.

If someone else knows better, please correct me.

 

[ericgibbs]

Thanks Eric. This is going to be a challenge to simulate, because the coils need to be coupled, and I think we need to specify a saturation flux density for the core. It's not impossible, I just need to ramp up my Spice tools.

hi,
The LT directive K1 L1 L2 1 , where 1, is the mutual coupling

 

[BrownOut]

hi,
The LT directive K1 L1 L2 1 , where 1, is the mutual coupling

Ah! A new wrinkle

 

[BrownOut]

My impression is that it acts like a DC/DC converter, able to up the otherwise unusable low voltage of a dying battery (say .5V) to something usable (say 1.5V). Therefore, it extends the life of a battery.

If someone else knows better, please correct me.

The problem I have with that explanation is this; since energy cannot be created, a DC/DC convertor can only trade current for voltage. It's my understanding that as a battery dies, the output resistance rises to a high value, making such a trade impractical.

However, I think I figured it out. Current in an inductor rises as the integral of the voltage. The inductor time constand is R/L. With a healthy battery, R is small and the inductor charges quickly. In a dying battery, R is big and the inductor charges slowly. However, since the inductor continues to charge until the core saturates, the current will be the same wether R is high or low. The output is a function of the inductor current, which has been constant due to the properties of the coil. Thus, the Joule Thief is not only a boost convertor, but self-regulates as well.

As soon as I figure out how to make a saturable core, I'll be able to simulate Eric's circuit, and prove or disprove my hypothesis.

EDIT: Is there a way to make K a function??? I'll try to look it up in the help files.

 

[Nigel Goodwin]

The problem I have with that explanation is this; since energy cannot be created, a DC/DC convertor can only trade current for voltage. It's my understanding that as a battery dies, the output resistance rises to a high value, making such a trade impractical.

Read my post #3 - you are perfectly correct that the output impedance increases, but wrong about it been impractical for the reasons I stated.

 

[mneary]

My impression is that it acts like a DC/DC converter, able to up the otherwise unusable low voltage of a dying battery (say .5V) to something usable (say 1.5V). Therefore, it extends the life of a battery.

If someone else knows better, please correct me.

The voltage of a brand new AA alkaline cell is about 1.5V, barely enough to light even red LEDs. After less than 10% of its energy is used, its voltage is about 1.2V, which isn't usually enough. So the almost new cell won't light an LED. Of course, a 1.25 NiMH or NiCd is hopeless even when fully charged.

Joule Thief doesn't create energy, it just takes that 90% of the battery which lies below 1.5V and makes it useful to an LED. By the time the cell is down to 0.8V most of the energy has been harvested.

Anything that causes a halt to flux increase will end the feedback and permit oscillation. Core saturation works, but some of the transformers I see won't saturate. In those cases, the tiny transistor runs out of beta at fairly low collector currents.

 

[BrownOut]

The voltage of a brand new AA alkaline cell is about 1.5V, barely enough to light even red LEDs. After less than 10% of its energy is used, its voltage is about 1.2V, which isn't usually enough. So the almost new cell won't light an LED. Of course, a 1.25 NiMH or NiCd is hopeless even when fully charged.

Joule Thief doesn't create energy, it just takes that 90% of the battery which lies below 1.5V and makes it useful to an LED. By the time the cell is down to 0.8V most of the energy has been harvested.

Anything that causes a halt to flux increase will end the feedback and permit oscillation. Core saturation works, but some of the transformers I see won't saturate. In those cases, the tiny transistor runs out of beta at fairly low collector currents.

True. If the core doesn't saturate, you don't get any oscillations, and thus no power conversion. I proved that by simulating Eric's circuit. It ddin't oscillate. In order for this to even work, the core must be saturable, and thus is responsible for the correct operation from a new battery to an almost finished one. Websites that explains that that it "works" down to discharged voltage aren't sufficiently detailed.

 

[mneary]

The collector current for a small transistor levels off as the beta goes away at higher currents.

This causes oscillation even if the core never saturates..

 

[BrownOut]

I can't seem to make that work very well.

 

[BrownOut]

Here is something that I can make work. Be adding the capacitor, the base can be forced into going negative enough to cut off the transistor. Coupling between the coils is constant. What changes is the transistor comes out of saturation, which makes a sharp transistion in the coil from a linearly rising current, to a constant one. That kills the induced voltage in the base coil, and allows the cap to cut off the transistor. The cap eventually charges and the process is repeated.

That still doesn't illustrate proper operation for the JT, but it does show that a changing B field determines the oscillation. I have to make a better model for the saturable core so I can properly analyze this circuit. Honestly, I don't think it operates properly otherwise.

 

[BrownOut]

Ok, I figured out why I couldn't simulate Eric's circuit. I didn't have the same LED model as he did, and so I substituted a diode, but it had he wrong turn on voltage. So, I strung enough together to equal his model. Now it finally works.

Stupid, Stupid, Stupid!!!

 

[ericgibbs]

Ok, I figured out why I couldn't simulate Eric's circuit. I didn't have the same LED model as he did, and so I substituted a diode, but it had he wrong turn on voltage. So, I strung enough together to equal his model. Now it finally works.

Stupid, Stupid, Stupid!!!

hi,
This is a copy of my LTS 'standard.dio' in Comps folder.
Changed the .dio to .txt so I can post it.

Either replace your 'standard.dio' with this one or copy/paste any diode/led etc to your existing 'standard.dio'.

 

[BrownOut]

I copied your file. Don't I need a symbol too? BTW, how did you come up with the values for your coils. I hate the articles say just say "wrap XXX times around a ferrte." Too many variables to know what the final value will be, and I like to be able to analyze circuits before I build them.

 

[ericgibbs]

I copied your file. Don't I need a symbol too? BTW, how did you come up with the values for your coils. I hate the articles say just say "wrap XXX times around a ferrte." Too many variables to know what the final value will be, and I like to be able to analyze circuits before I build them.

Hi,
The 'diode' is the standard diode [led] symbol in the components file of LTS.
Right click the diode/led symbol on your drawing and choose a different type.

Ref, the JF transformer, there are so many ring ferrites its difficult to give exact winding data, most web sites use an empirical approach using recycled ferrites from old gear.

I do have an article for LTS transformer design, I will dig it out and attach it later today.

EDIT:
Look at these sites, two use online calculators.
TRANSFORMERS

**broken link removed**

**broken link removed**

Attached LTS doc file , look at the transformer section.