Joule Thief

[Pirate88179]

@ Abbarue:

Crap!  I forgot to mention your contributions to this topic also.  For that I am sorry.  This is a good question you ask of MK1.  Could it be that the brass wire has a higher resistance and this might account for his results?  Brass is a mixture of copper and zinc right?  This would have to have a higher resistance than copper or say, aluminum right?  This is why I proposed using the VR so we can adjust this variable.  I know Jeanna tried this but I am not sure she used it where I was thinking of in the circuit.

The "ping pong effect".

What I am hoping to try to do with the supercaps, among other things, is to get a ping pong effect going such that the current flows through the toroid and windings and transistor and lights the led. (diode)  Where does the excess juice go?  what I want to do is to set up the caps so that any unused juice flows back into the 2nd cap.  If I can put a secondary on the bifilar toroid and feed that back to the 1st cap, then maybe the efficiency would be improved by an order of magnitude?  This is to say nothing about combining this with the Bedini SG motor or earth battery.  I have many ideas and so little "expert" electronics knowledge.  The main reason I am so hooked on supercaps is the way they "imitate" batteries.  All of the other caps I have played with "dump" everything they have in like .00002 seconds.  All at once.  The supercaps seem to provide energy at a rate demanded by the circuit until their energy is gone.  In the case of my earth battery circuit, this was for up to 2 hours after disconnecting it from the earth battery.  And this was a very small supercap (my first one) salvaged from a "shake flash light." (5.5 volt .22F)  Here is a photo of that circuit.

Bill

[Mk1]

@all

I also tried aluminium 20 gauge 1 turn shows nothing on multimeter , but led worked anyway.

@ pirate

It dose go fast , i answered your question on last page 

By the way it was Jeanna that found that 1 turn is enough to light a led.

[Pirate88179]

@ MK1:

Well, I am the first to admit, I can't keep up with this topic.  I wish this was all I had to do every day.  Actually, this is my favorite part of my day, checking OU dot com for responses and posts.  I pulled a 1.25" OD ferrite toroid from an old computer power supply.  It was already wound with a bunch of turns of what appeared to be magnet wire.  But, it was not wound in a bifilar manner so I am going to remove that wire and use this one to make a JT also.  I now have 4 of these and that will make 5.  You would think I would know all about these by now, but I don't.

I was hoping those zaps from the Fuji circuit would make me more intelligent and activate parts of my brain I am not using.  This did not seem to work as of this writing.

Bill

[Pirate88179]

What bugs me here the most is nobody seems to be concerned with the input consumption. Everyone is rejoiced they can lit two or 300 LEDs, but at what expense?

I think the idea with JT is that you should be able to light LED(s) for a long time, implying a low current consumption from the battery. Further more with most of these designs there is a limit to which you can use the battery with the JT until you lose the functionality, LED will go off basically once voltage in the battery drops below certain level.

The point is if your JT uses 20-25mA of current then you might as well hook the LED directly to the battery source and not bother with anything because you are consuming the actual current LED needs to operate.

The best result seems to be around 45-50% duty cycle square wave oscillator. With shorter impulses the LED does not light up as bright, although the current consumption is < 5mA (depending on the circuit).

I think there should be some ground rules set here what the goals are. In my humble opinion some of those would be:

1. Usable light output.

2. Being able to connect more than one LED for greater light output.

3. Minimum current consumption/maximum battery source longevity.

and then the tall order one...

4. Feedback into the battery source for recovery of power.

Even though I am not a big fan of quantifying things, we need to start doing some measurements and see what the gains (if any are) of individual circuits. There should be a solid science behind this and a rational explanation of why. When I say solid science that means experimentation, repeatability, enforced with some theoretical background.

I did post a link to two ebooks earlier that most probably ignored. One was about LEDs, basically A-Z and the other about commercial power drive systems for LEDs. Both did appear to have interesting bits of info that could be used in our quest here.

I just wanted to correct your post....that was 400 leds.  Each led requires 3.0 vdc and 250 mA's to light.  Now I don't really know how many mA's are in a AA battery but when I put my meter to one, it is no where near enough to light 10 leds much less 400.  We are not creating energy here and this is not OU.  I have seen your posts on this board and I KNOW that you understand more about electronics than I do, and I respect that.  I just think that this JT circuit is already pretty great (in my mind anyway) and if we all work at it, we can improve it to be something even better.  With all of your knowledge and experience, I know you can help us.  This circuit is not the answer to everything and maybe not even to anything.  But, it just might raise enough questions to help find the answers we all need and want.  Thanks for posting here.

Bill

[Mk1]

@all

I made some experiment trying to make a small version of the Meg,I will call the Meg thief , So far i got some strange and interesting stuff,
I actually made my led sing a high pitch sound , I hooked it up to the bridge and got 66 volt dc per pickup coils so 2X66volts dc.(or a presume 2x120 volt ac)

This is a picture of the joule thief a modified to Meg spec , i saved the 2 windings already there and made 2 coils where the tie rap are
They are made 7 turn on each side of the toroid(regular 22 gauge covered single strand(like phone lines)) , and used a 2n3904 and 1 k resistor. And even added a mini neo magnet taken from a cd drive reading head (but the led cried mercy whit or without it). The turns that are on the picture are 24 on both half , 12 going up and 12 down so both ends are in the same location larger than 20 gauge.

Sorry i don't have camera , i wish i could show you , i got friend coming over this week-end i hope they got a camera.

So far this little circuit has showed incredible possibility.

Even a brand new 9 volt don't make that led sing, lol

The sun will rise soon , literally good night .