SSG, "Joule Thief" and other

[amigo]

I was surfing the Web and stumbled upon this simple circuit called "Joule Thief" that lets you run a white LED on 1.5V battery for a long time, by creating high voltage pulses through PWM or back EMF or whatever.

The circuit layout looked familiar to JB's standard SSG oscillator, minus couple of parts, but never the less. I'm attaching a generic JB circuit, the "Joule Thief" and my modified version. Link to the "Joule Thief is here.

Since I did not have a whole bead but small bead rings so I got four of them together and wound the feedback transformer. Did not have the 35 AWG so I used 30 AWG. (talk about a hack, eh?)

In their version I've put together, the circuit uses ~22mA of current with pulses that are 45.40uS long or 21.55KHz.
With my test battery (1.2V NiMH 2300mAh) that would give me over 100 hours of work time, no?

After modifying the circuit to resemble JB's SSG, I got it down to ~14mA, with pulses that are 32uS long or 31.25KHz. Not bad I guess with a barely noticeable change in LED illumination (probably 5-8% less).

Being evil megalomaniac that I am, I hooked 14 LEDs not one to the circuit, without any change in current usage or light illumination.

So my questions are:

- how do I make this self-resonate so that it does not need a button shorted to begin the process?

- also, I always wondered what is the purpose of the Diode across the EB, in JB's SSG circuit?

- can this be optimized to get even less current consumption

- what affects the pulse duration/frequency; how do I make them even shorter/faster?

Thanks

[Shamus]

The diode is there to ensure 1) that no current will flow from the primary battery to the base of the transistor and 2), to ensure that current will only flow through the trigger/base/emitter sections on the current induced as the magnet moves away from the coil. There's a probably a 3), in that it forms a part of the environmental amplification of negative energy formed by the sharp gradient of the field reversal in the coil but I'm no expert on this technology.

[amigo]

Thanks for the reply Shamus.

I've tried adding a diode to the Bedinified "Joule Thief" circuit across the Base/Emitter but that wasn't working.

You said "that no current will flow from the primary battery to the base of the transistor" but I do not see how is that possible when the trigger circuit is only sharing negative pole?

[Shamus]

You are, of course, right. The main thing is to limit the current to fire at the right time (when the magnet moves away). Since the "Joule Thief" doesn't have this feature, the diode probably wouldn't add anything to the circuit. Dunno for sure, I haven't studied self-oscillation in depth as of yet.

[Aaron]

self oscillation

Hi Amigo,

Awesome that you're looking at other practical applications for these circuits.

Try putting a resistor directly off the emitter of the transistor. Start smaller and work to higher resistance and see if it gets into oscillation. Not sure if 20 turns will work very well but you can try it.

[amigo]

Hi Aaron,

I am not sure what you mean, off of Emitter towards...(Base)?

I used wrong AWG to begin with, should be using thinner 35 instead of 30 which I think influences the ability to oscillate. Also, my bead is not in one piece so that must have other magnetic effects as well.

Though do you know what determines the pulse duration/frequency?

Thanks.

[smoky]

Hi Amigo, I think the diode is there to protect the transistor from blowing on the negative swing from the feedback coil's back emf. ie when the transistor is not conducting. They usually can only withstand about 6 volts or so of reverse bias (BVbeO)

[amigo]

Hi Amigo, I think the diode is there to protect the transistor from blowing on the negative swing from the feedback coil's back emf. ie when the transistor is not conducting. They usually can only withstand about 6 volts or so of reverse bias (BVbeO)

Thanks for the post. I did try adding the diode in the Joule Thief circuit mod but then it was not working or not working as intended (don't recall exactly).

[CaptainScat]

thanks for posting this amigo!
i looked at this a couple weeks ago but had other things i was doing. tonight i was watching the energy from the vacuum and just listening to john talk as i putzed around putting together this circuit. anyways it was just at the part where john says "...of course under the conventional normal circuits you cant have more out than you put in" when i hooked up a single AAA with a voltage reading of 1.09 to the circuit and the LED lit up bright as can be. it struck me then how great this circuit is in demonstrating the basic principle of bedini's circuits. it is even simpler to build then an SG and lets you find a use for those dead 1.5V batteries other than tossing them in the garbage.

one of the things i noticed while playing about with different batteries of differing voltages was the weaker the battery i use for the input, the lower the amp draw is on the circuit. anyone have any ideas why this is? is it not causing the transistor to switch as fast?

oh and my circuit is a total hack as well, i used a pretty big toroid core with about 40 turns of an unknown gauge (probably 30 or smaller) scavenged from a neon light circuit. the transistor is a 2n5551 scrounged from an old broken PC power supply and a 680ohm 1/2 watt resistor cause i had it laying around. i used the "variant on the circuit which has been optimized for fudging together without a PCB" from the joule thief web page you linked to with the components listed above. i don't know how to measure the speed of the pulses so sorry i cant give you those. i'm guessing you need a scope to measure that? i'm now looking about for more LED's... i fear i might be a megalomaniac also

thanks again for the great post!

[amigo]

CaptainScat thanks for the report, sounds like you are having fun hacking this - just the way it should be.

I never thought to try some almost depleted battery and see what happens with the current, thanks for the idea. I usually use rechargable NiMH batteries and whenever they fall down a bit I recharge, while the normal batteries I toss away - figures I should keep them and give it a try.

This gives me an idea to try a variation of the circuit, perhaps using a 2N7000 MOSFET and see if anything happens thats better.

Will report here if anything is accomplished...

P.S. Scope is really a necessity for any kind of work and if you do not have one I suggest you procure one, or build one, here's a really cool and simple project if you are into PIC uC: semifluid.com » PIC18F2550 KS0108 Graphical LCD Oscilloscope

[adam ant]

Capitan... there are LED's hidden EVERYWHERE!!! the industry has been using them in VCRs since the early 80's. they are in old computers, TVs, some old microwaves, stereos, appliances, etc. the trick is getting them out without breaking them. sometimes the leads are so short that it is almost impossible.
i have hacked so many appliances this year it is crazy. i actually got a hold of a full size, professional office Xerox copier. it was at least 200+ pounds, stood four feet tall, five feet long, and two feet wide. i couldnt tell you how many DC mototrs, gear set-ups, transformers, and excess resistors/diodes/transistors/LED's that i got.

radioshack has a one handed solder sucker that costs about $11. it heats and has a bulb together. it is the fastest/easiest/cleanest way that i have yet to find.

[amigo]

radioshack has a one handed solder sucker that costs about $11. it heats and has a bulb together. it is the fastest/easiest/cleanest way that i have yet to find.

hehe, you truly sound like someone who has gutted his way through many devices, with lots of experience under his belt (or just the Radio Shack solder sucker on the belt? )

[CaptainScat]

amigo i did some poking about the internet and found out that more turns = lower freq. less turns = higher freq. and higher resistance has the effect of reducing the supply current (osc only)
i found this and and bunch more info here
1.5 VOLT WHITE AND UV LED DRIVE CIRCUIT
he even used a rusty nail and cat5 wire for an inductor... i love it!

on getting your circuit to self oscillate, i really don't know squat about circuit design but could you charge a cap similar to this to get the emitter-base junction to conduct?
Simplest_LED_Flasher_Circuit
it's probably not possible but i don't know any better so i gotta ask

adam ant, thanks for the tips. i try to scavenge stuff if i can, something about taking something and recycling it for more than it was worth appeals to me. i snagged a couple of LED's from an old unused keyboard

[Watson]

can this be optimized to get even less current consumption

I was surfing the Web and stumbled upon this simple circuit called "Joule Thief" that lets you run a white LED on 1.5V battery for a long time, by creating high voltage pulses through PWM or back EMF or whatever.

The circuit layout looked familiar to JB's standard SSG oscillator, minus couple of parts, but never the less. I'm attaching a generic JB circuit, the "Joule Thief" and my modified version. Link to the "Joule Thief is here.

Since I did not have a whole bead but small bead rings so I got four of them together and wound the feedback transformer. Did not have the 35 AWG so I used 30 AWG. (talk about a hack, eh?)

In their version I've put together, the circuit uses ~22mA of current with pulses that are 45.40uS long or 21.55KHz.
With my test battery (1.2V NiMH 2300mAh) that would give me over 100 hours of work time, no?

After modifying the circuit to resemble JB's SSG, I got it down to ~14mA, with pulses that are 32uS long or 31.25KHz. Not bad I guess with a barely noticeable change in LED illumination (probably 5-8% less).

Being evil megalomaniac that I am, I hooked 14 LEDs not one to the circuit, without any change in current usage or light illumination.

So my questions are:

- how do I make this self-resonate so that it does not need a button shorted to begin the process?

- also, I always wondered what is the purpose of the Diode across the EB, in JB's SSG circuit?

- can this be optimized to get even less current consumption

- what affects the pulse duration/frequency; how do I make them even shorter/faster?

Thanks

For my version, the Supercharged JT, I get double the efficiency compared to the conventional JT. You can find out more here.

[totoalas]

I remember lidmotor has made some auto oscillation