Joule Thief

[stprue]

Hello Conrad,
I will hand draw a diagram and post it. 

@All
Someone here recently gave a link to a great little free circuit diagram program.  It also converted to breadboard views.  Does anyone know what the name of the program?

[TinselKoala]

I like the use of the magnets to make the battery connections, that's a nice touch and one that I will shamelessly start using in my lab.
Thanks!!

What happens if you remove the Schottky diode that is in series with the LED? I'm not grasping why this diode needs to be there.

[kooler]

@All
Someone here recently gave a link to a great little free circuit diagram program.  It also converted to breadboard views.  Does anyone know what the name of the program?

someone posted this one which is real nice..
http://fritzing.org/

[conradelektro]

I like the use of the magnets to make the battery connections, that's a nice touch and one that I will shamelessly start using in my lab.
Thanks!!

What happens if you remove the Schottky diode that is in series with the LED? I'm not grasping why this diode needs to be there.

I also use two little magnets, one has a red litz wire soldered to it and the other one a black litz wire. The other end of the wires has endings which plug nicely into the plus and minus rail of a breed board (solid wire ends soldered to flexible litz-wire). This gives an easy connection of a breed board to a battery. I saw this on a web site which discussed various circuits. Most things I know about electronics are gleened from some forums or web sites.

The Schottky diode in series with the LED turns the LED into a "good diode" which really blocks current in the reverse direction. In the forward direction it does no harm because its forward Voltage is lower than the forward Voltage of a LED. The power consumption of the 1 V 50 µA circuit I posted was lower when using the Schottky diode in series with the LED. I attributed this to the better reverse current blocking, but I could be wrong. The effect is not hight, may be some 20%.

I measured power consumption with a scope over a 1 Ohm shunt (in the positive power supply line) doing an approximate averageing of the Voltage curve by hand. The true RMS measurement of my scope over the shunt was pretty accurate as well (compared to the averageing of the Voltage curve by hand). The power comsumption of a Joule Thief happens in bursts. But my digital multimeter measured the power comsumption also very well (the average Voltage over a 1 Ohm shunt and the direct current measurement in µA when the multimeter was put in the positive power supply line).

I still believe (after many experiments) that a Joule Thief with an air coil can have less power demands in comparison to using a Ferrite toroid. But the air coil will be bigger than a Ferrite toroid. 100 turns for the base coil and 200 turns for the collector coil over a 2 cm plastic tube seem to be well suited for the circuit I posted. To keep the power supply voltage down to 0.5 Volt, the number of turns can be doubled. To start the Joule Thief I needed 0.7 Volt, but once started the Voltage could be dropped to 0.5 Volt. Air cores also allow for a higher frequency of oscillation which can reduce power demand when the currents become more like spikes.

A high electrolytic capacitor over the power supply (e.g. 100 µF) helps with weak power sources but increases the average power consumption. The reason seems to be that an electrolytic capacitor can give the peak power but a battery is to slow. So, when the peak power demand happens the electrolytic capaciter will fulfill it and then is slowly filled by the battery till the peak power demand happens again in the next cycle.

I am not so sure that a Joule Thief with two transistors uses less power in comparison to a one transistor JouleThief which is carefully designed (if one wants a LED that is continuously lit). The key features are frequency (should be 100 KHz or better more) and a very well designed coil, best air core. A 1 nF capacitor over a 100K to 500K resistor to the base helps to reduce power consumption by reducing the "switch on time" of the transistor and by assuring the necessary threshold voltage at the base to switch the transitor on.

I believe that power comsumption can be brought down to below 1 V and 50µA by having the LED blink, but this is not pretty according to my humble opinion.

White LEDs which I bought recently seem to need less power than red LEDs in order to be dimly visible. There was some progress concerning the light output of white LEDs.

I still try (without success so far) to have a LED continuously dimly lit by the electro smog in our environment (mobile phone transmissions, wireless LANs and 220 V 50 Hz mains grid). The energy density seems to be too low. A long antenna and a good ground seem to help, but still too little energy.

Greetings, Conrad

[Pirate88179]

Jim:

Great to see you back here Mate.  As you know, I too am very sorry for your loss.  Here's to better days ahead for all of us.  There is some great work being done here by many folks.  I have been unable to keep up with it all lately.  Lighting leds on such low power...it is amazing to me.

Happy New Year to you Jim, and to everyone.

Bill