Joule Thief

[drodenbe]

Me again!

If you noticed by the last graph it really looks rough!  With all the fluctuations and all, but as it runs and auto adjusts the scale this will smooth out and you won't be able to see the detail unless you zoom in on an area.  But the important thing is the upward trend.  The LED is not lit yet but will when the voltage gets higher.  The scope peak to peak reading is around 1.5 volts, so when it gets to 2 and above, it should light.  Never tried this before with this circuit so this is my thoughts at this time.  As we all know time will tell.  However I believe that if this voltage comes up to the point of lighting the LED and continues to charge I will have a very warm feeling that this is a success!!!! 

About the first graph I stopped it last night and forgot to save the graph.  My mistake!  It had leveled off to the 1.160 voltage and was running constant for while.  It may have gone up more or less but I feel it was a good test anyway.   Thanks again.

David

[drodenbe]

Well here is a new update on my latest run on an old junk crystal cell that was a failure compared to the ones I make now.  I am making cells with aluminum and copper that are producing over 1.volt. Usually around 1.25 volts at about 150ma.  This one was a test of a different mix and idea. Was around .46 to .5 volts at about 10ma.  But as you can see by the graph the cell is coming to life.  The LED just came on a little while ago.  Getting around 1.5 volts Peak to Peak on collector.  Also one other thing I wanted to say is that the .0047 uf cap across the collector JT coil is not necessary to get what you see here.  I am not using any cap across the coil.  I did notice when I was designing the circuit, before I used the capret,  that it did make a small difference.  At that time I wanted all I could get.  Also I found that the charging effect is enhanced if you do not have a resistor in series with LED.  Strange isn't it!  I put the variable resistor in series because it gave me a easy way to change things during design.  I usually make my circuits with small two terminal  pc mount connectors.  In the capacitor position and anywhere I would change something.  Keeps me from messing up the pc board, due to unsoldering, soldering in components during design.  So I have a experimenters board so to speak.  Terminals for the joule thief coil.  Terminals for the base capacitor, terminals for the base variable resistor,  terminals for the transistor,  terminals for input voltage and a possible capacitor across the input voltage.  So Basically I have a board with terminals!!!  Ha! Ha!  How many times have you popped a transistor or wanted to change the pot or cap.  Works really great for design. Then I make a final board when finished.  Well,  Ok.  Later


David

[NickZ]

  You might want to check out the Lasersaber site:
http://laserhacker.com/

  He's coming up with new stuff all the time.

  NickZ

[TinselKoala]

@Pirate:
Take a look at this. I like the way the little device is mounted right to the LED. Very very simple. Somebody should go ahead and manufacture LEDs with the circuit and inductor built right in.

http://www.muzique.com/news/not-a-joule-thief/

[Pirate88179]

@Pirate:
Take a look at this. I like the way the little device is mounted right to the LED. Very very simple. Somebody should go ahead and manufacture LEDs with the circuit and inductor built right in.

http://www.muzique.com/news/not-a-joule-thief/

Thanks TK.  I think I might even have some of those chips but....I am still working with the NCP 1402 which will switch down to .3 volts and output a low ripple 3.3 volts.  I went through a bunch of those newer style chips and my hat is off to that guy for being able to solder it like he did.

I like your idea and the led folks should pay attention, although the Chinese are probably already doing this.  It would be great to spend a few cents more and buy an led, or leds, that already have the modern day JT type circuit "built in".  I do wonder, however, what would the problems be if using say 3 of them in series?  Would each led "circuit" work independently or would there be a problem in the sync between all of the different occilations from the 3 different inductors?

To me, this is really cool stuff that is coming out now about the JT style circuits.  These new chips give us even more areas to explore.

Thank you for posting this and showing me.

Bill