Joule Thief

[sarmasio]

Hi,

I have tried to replicate the two battery charger variant of the joule thief.
Could not make it run, then after 4 trial I gave up. Can somebody help me with
this, I would like to make sure I-m doing right the  toroid at least.
Which of the four variants is the good one?

best regards,
sarmasio

[kooler]

Hi,

I have tried to replicate the two battery charger variant of the joule thief.
Could not make it run, then after 4 trial I gave up. Can somebody help me with
this, I would like to make sure I-m doing right the  toroid at least.
Which of the four variants is the good one?

best regards,
sarmasio

it looks like your hook up on the switch is wrong.. you have both batteries on at once with feed back .. fix that i think you be ok..

robbie

[sarmasio]

Thanks for the fast response,

This is how it suppose to be the right side battery is the run battery
and the left one is the charge. By reversing the switch, the role of the two batteries
will change, the right one will be the charge battery, the left one will be the
run battery. I'm suspecting having a problem by winding the toroid, this is
where I'm getting lost.

best regards,
sarmasio

[TinselKoala]

Robbie the idea of the switch is to switch one battery from running to being charged, and the other from being charged to the running position.
I don't know about the switch being right or wrong... it looks like it at least switches the batteries around electrically.... I haven't built the circuit myself but  the problem with JTs not running is often incorrect coil phasing, as the sketches suggest.

@sarmasio
If you take the direction of the coil that is connected to the transistor collector as "fixed"... that is, hook it up and don't change it.... then you have four possible options to try with the other two coils. Right? Call the coils 1 and 2 and the ends a and b. You can reverse 1a and 1b, and you can reverse 2a and 2b. 4 possibles to try. Try them! If none work, then go back to the "fixed" coil hooked to the collector and reverse it. Then go back to coils 1 and 2 and try all 4 possible orientations again.
So in all, there are 8 different ways you can hook up 3 coils. If you are using breadboards and clipleads, you can try all eight in less time than it took me to type this, probably.You could also try applying a resistor of 220K to 1 megohm between the collector and base of the transistor to make the system start oscillating.... so now you have sixteen variations to try, maybe it will take you half an hour and you might need to make notes.

I'm all out of toroids or I'd try your setup, but by the time I wind a bunch of windings I lose track of which end is which anyway, so I _always_ try all possible permutations of coil hookups on any circuit like this, even if the first one I try works.

(ETA: don't forget to try both polarities of the LED too, and make sure your LED is actually working, this circuit can easily pop them and you'll be testing with a dead LED..... a real waste of time.)

Also, NTE109 is the expensive equivalent of the 1n60 germanium diode which might be a lot cheaper under that part number. But I'd use a more robust diode here until I got the circuit working if I were you. Subbing a 1n4007 high-voltage rectifier for a 100 volt germanium point-junction diode.... interesting. You might try 1n914 or 1n4148, low cost, relatively fast switching diodes. If a germanium diode isn't expected to pop, then I don't think a 1n914 would either.  Once you are running then optimize, by increasing the base -collector resistor if you needed one, and lowering the operating voltage by using the germanium diode.
Last night I tried MPSA18 to sub for 2n2222 and it works significantly better at very low voltages. Have not yet tried higher voltages with the MPSA18, though.... but they are very cheap.

Also again...... are you interpreting your toggle switch correctly? Usually on small toggle switches, I think the contacts "opposite" the toggle handle position are closed. The way you have it drawn is, in my experience, backwards. So you might be trying to run on your depleted battery and trying to charge your full one !! Check the action of your switch with the continuity or ohmmeter function of your multimeter.

[TinselKoala]

1ua . yes that is a nice one !  15 for 2.84 http://www.ebay.com/itm/15PCS-Transistor-FSC-MOT-ON-TO-92-MPSA18-MPSA18G-/260979077822?pt=LH_DefaultDomain_0&hash=item3cc390b2be

I'm finally down to the low microamp range!

In Ghost Light mode, with MPSA18 instead of 2n2222, and for the diode MUR1560 ultrafast high current diode, I am getting the light without touching at around 50 microAmps with 3 volts on the cap, and with touching a fully bright LED at around 2.5 volts and 1.2 milliAmps. (The current drops as the voltage drops. I started at 12 V and about 250 uA non-touching. I am amazed....)
And this is thru my 1.8 ohm series resistance LCheepo ammeter and also a 1 ohm series resistance in the negative lead at the 3055 emitter.