Big Joule Theif

[synchro]

Finished job video.

I have a new Flash mod video of the finished product with a push button on off switch.

YouTube - zebok3's Channel

[Groundloop]

@NickZ,

The circuit is not OU (yet) if that is what you are asking about.
I need approx. 35 Watt input to fully lit the 25 Watt 230VAC bulb.

I have some heating in the transistors and small heating in the
resistor and potmeter. But I think it should be possible to use Mosfet
switching or maybe other trasistor types to get the heating loss down.

Alex.

[NickZ]

Groundloop:
You have given us a lot of great information.
When using multiple secondaries, an additional inductor toroid for each secondary output can help to keeps the voltage more stable without dimming the other secondaries. So multiple secondaries around one big main toroid core should work fine. I've seen a diagram of the idea somewhere, if I can find it I'll attach a pic of it.
How many volts minimum input does it take to light your 25 watt, 220v bulb, full on?

[Groundloop]

@NickZ,

I have not tried to use more than one secondary toroid. But I think your
idea is good. One thing though, the center coil (L3) needs to be tuned.
Too few turns and no output at all. To many turns and poor output.
I'm using the positive voltage "spike" every time the transistor is trigged
on, as you can see in the drawing, via a diode to trigger the coupling
coil (L3). This way the input enery can ONLY be transfeered to the output
at each power coil triggering. The polarity of the L3 also needs to be
correct. Connect the L3 the wrong way and no or poor output is the result.

The two toroid with a coil in the middle is an AC switch. AC (or pulsed DC)
can ONLY transfer to the second toroid if the coil is shorted out. No work
need to be done on the coil since you can actually use a on/off switch
for this. So my idea was to use the forward voltage spike when the transistor
turns on for this switching since we do not need to provide any power to
do so. I hope my english is good enough to explain what is happening.

The toroids do not need to be of the same size. You can actually use a
larger toroid for the output (the right one). Like I said before, the importaint
thing is the center coil (L3). It MUST have the correct LOW Ohm =
thick wire, but still HIGH enough inductance for the voltage pulse.
Best way is by trial end error. Begin with enough wire to make many turns
and then remove turns until the voltage get really high at the output.
If the voltage starts to go down then you have removed too many turns.

I do not make videos but here is a picture where I lit up a 25 Watt 230VAC
normal light bulb.



Here is a 15 Watt 230VAC Osram Tube lit up to almost full.
(Input usage was 15 Volt @ 0,86A.)



My circuit enclousere is an old data transfer switch metal box. LOL
If you have more questions then do not hesitate to ask.

Alex.

[NickZ]

@ Groundloop:
Your double toroid circuit looks very interesting.
I have a bunch of questions, but what comes to mind is: what would happen if we were to use one main toroid with several secondaries. like you've done, but all the way around the main one. Like a 3 inch core with several other smaller toroids connected all around it? Possibly one of the secondaries could be used to feed-back to the source battery.
I'd be very interested to see more pictures or a video of your set up lighting a regular 25 volt bulb.
I'll be trying it with two one inch toroids to see what happens. Kooler mentioned that on his twin core Jtc using two cores made the voltage output a bit lower, but lowered the drain on the battery by a lot. If I understood him properly.
Sort of reminds me of the magamp. Could there be some similarities?
Thank you for sharing your results.
I'm still working on the Jt-Caprets tests, but will try your idea soon.
NZ

[Groundloop]

My big "Joule Thief"

@All,

Here is my big "Joule Thief". I have built and tested this circuit and it runs
well. You can get a very high voltage output with just a "few" turns of
enameled copper wire at the output toroid. This circuit can light a 230VAC
25 Watt normal light bulb. The circuit can also light up CFLs and tubes. It
can also desulphate and charge lead acid batteries.



Final note, do not touch the output when you run the circuit at high
input voltage and current. The output can be several hundreds of volt
with real amperage, and can harm you.



Groundloop.

[Tectalabyss]

Some very useful info

Hi all.
Been pondering just how many turns I have on my single layer wrapped 2 3/8". I was wrapping toroid #2 so I can do some comparison tests and I lost track, interrupted and what ever. Ah....30ga wire is .010, looked it up , measured it, and even with RS's red coating it measures .010. Ok, that 100 turns per inch. I do wrap tightly and make sure I dont over wrap. Can't measure the outside as it always will be "loose" with small gaps due to the larger dia than the inside, but the inner ring doesn't. I actually counted the outer ring wraps and was right at 350. I left a gap for the primary wind. Hmmm..ok that's abt 100-150 less turns than most schematics show for a good cfl. About 78% of what I need or most successful projects have posted. Actually, it's quite a bit since the voltage is multiplied across the # of turns and its 22% short! I took a piece of tape cut it to 1/4" wide and put it inside the core from where the wire started to where I ended, looked closely and nipped it with small scissors. The tape
Took it out and carefully laid it on a ruler. 3.5" dead on.

Ok, now I can do what I want without counting and be very close! A correct length can be cut for the 1st layer and put in the toroid before winding and just cover it. It could be marked and removed. If a second layer is needed, due to to space limitations, cut the length you need and after you've wrapped layer #1, put the correct length of tape on the inside again covering it or marking. I prefer to cover the 1st layer with stretched thin electrical tape. I split the tape by width as it leaves less over lap and keep the windings closer to the toroid.

Not all insulation will be the same. The clear seems to be a little thicker, still, you should be close without all that deep concentration. If another gauge wire is used for the long secondary wrap, now you can look it up and it's an easy calculation. If it were a really fine wire like the Fuji with 45ga, I tried it by marking a bamboo skewer stick 1/4" with an Xacto knife and wound from cut to cut. Works and you have a new multiplier.

Hope this helps some out as much as it has me.

Hello Vaporizer and All.
That is some very useful Info and will make it a lot easier to wind our coil's Thank you.

@NickZ.
I agree with you on the captret,There is more to it it. Just be careful has Vaporizer had stated.I use a fuse and put it in an open topped container.Good luck.Thanks to all. Tec

[synchro]

Twin Funsaver video.

Got a new video on the twin package:

YouTube - zebok3's Channel

[Vaporizer]

Calculating your number of turns without counting!

Hi all.
Been pondering just how many turns I have on my single layer wrapped 2 3/8". I was wrapping toroid #2 so I can do some comparison tests and I lost track, interrupted and what ever. Ah....30ga wire is .010, looked it up , measured it, and even with RS's red coating it measures .010. Ok, that 100 turns per inch. I do wrap tightly and make sure I dont over wrap. Can't measure the outside as it always will be "loose" with small gaps due to the larger dia than the inside, but the inner ring doesn't. I actually counted the outer ring wraps and was right at 350. I left a gap for the primary wind. Hmmm..ok that's abt 100-150 less turns than most schematics show for a good cfl. About 78% of what I need or most successful projects have posted. Actually, it's quite a bit since the voltage is multiplied across the # of turns and its 22% short! I took a piece of tape cut it to 1/4" wide and put it inside the core from where the wire started to where I ended, looked closely and nipped it with small scissors. The tape
Took it out and carefully laid it on a ruler. 3.5" dead on.

Ok, now I can do what I want without counting and be very close! A correct length can be cut for the 1st layer and put in the toroid before winding and just cover it. It could be marked and removed. If a second layer is needed, due to to space limitations, cut the length you need and after you've wrapped layer #1, put the correct length of tape on the inside again covering it or marking. I prefer to cover the 1st layer with stretched thin electrical tape. I split the tape by width as it leaves less over lap and keep the windings closer to the toroid.

Not all insulation will be the same. The clear seems to be a little thicker, still, you should be close without all that deep concentration. If another gauge wire is used for the long secondary wrap, now you can look it up and it's an easy calculation. If it were a really fine wire like the Fuji with 45ga, I tried it by marking a bamboo skewer stick 1/4" with an Xacto knife and wound from cut to cut. Works and you have a new multiplier.

Hope this helps some out as much as it has me.

[NickZ]

Vapor, Tec, Pen, and All:
I have been playing with the Captret circuit tonight, and I connected the Jtc to the Captret.
I still have a 20 led bulb on the Jt, and the Captret was able to blink all of the leds for a second, every couple of seconds between each time, Good start...
Next I changed the capacitor to a higher value one, (10 volt 4700 mf), and was able to light a single led or two leds steady and brighter, but still only blinking the 20 led Jt., although a lot brighter also. I will connect a 4 capacitor bank that I just made like in Lidmotor's Captret video, to see what happens. But, it does look very interesting, so far, and does seam like this captret thing is going to work.
Vapor: the duty cycle I mentioned is the time is takes the Jt to drain a single AA. If a Jt can last all night on a single AA, with a bunch of Leds on it, I would think that it's working ok.
I can't get any readings on my analog volt meter, but the leds are bright so.. I can't really trust the meter anyways.
I was also surprised to feel a tingle from the secondary on a one inch toroid, first time for me too. The coil is singing so loud that my wife can hear it from outside the house. If I put a magnet on it it gets louder yet, but I haven't really seen any improvement in the light output from the magnet.
If you leave the Jt circuit running all night (like I do) it can overheat and fry the transistor.
I cooked two trim pots already. The TIP3055 type transistor do have a lot of amplification power, even when using them on a one inch coil, and a single AA. Not so with the 2n2222, although they are hard to beat as far as being the most economical on the battery's charge by almost twice the duration.
My feeling is that the Captret circuit is worth upgrading and experimenting with, and might be just the thing the Jtc needs to run on its own. But still alot more work is needed to find all the magic.
NZ

[synchro]

Two mods,

Here's a cool white toroid lit up with two Funsaver flash circuit chips and two AA batteries at maximum brightness.

YouTube - zebok3's Channel

[Vaporizer]

Hi all,
Time for a followup. I swapped out the transistor for a TIP35. The circuit "sang" as it did with the TIP31C and it measured 10v higher on the output. No other changes were made. I did notice it tingled me a bit and the other transistor never would. Put the 9w cfl on it and after abt 5-10 seconds, it lit right up!
Woot! Ok, this proves that even with a working load, you cannot trust the output reading without a scope as 67v will not light a cfl.
I decided to try the 13w spiral and the 8" halo cfl. All lit just the same on 1.25v or 2.5v the same brightness as the Fuji/Kodak transformer. I did find and good oddity, the 13w spiral will run full bright at abt 100ma less than the the other 2.
It's taking abt 455ma right now on the 13w spiral @ 2.5v.

I looked at some saved pics in my folder and ppl that lit cfls with 1.5v , 2-2.25" core used 10, 10, 450 for windings. Next is to add the 10:10 and see if it gains or loses. Once I know that, I can stretch a thin wrap of electrical tape over the 1st layer and add a second layer if needed.
Also I can now start looking at adding the 555 to see if I can get the duty cycle down. Lidmotor & Slayer were down to abt 15% on time. If I can get it non-flashing at 20% on, that should extend my 2.5hrs to to well over 10hrs. I can live with that.

@ Nick,
Sounds good. What is the duty cycle you refer too? Are you using a 555 timer or is that the battery swap out time? Also how many total LEDs, mcd ea voltage & current draw if you don't mind?

Addon: I tried the 10:10 wind on it and wasn't as bright no matter where I adjusted it. Time to try a lower primary wind. I expected the ratio to go that way, but as odd as these work, one really needs to keep and open mind and try to find out.

[Tectalabyss]

Sounds great !

Guys:
I was given a 20 led warm white bulb. It is perfect for my Jt circuit, and it gives the right color. I am so happy that there are leds now that give off an acceptable light color. I'll be buying more of them soon. The Jt will light all the bulbs that I put on it. I have connected about 40 or so.
I am having shorter duty cycles on a one inch toroid with secondary on it, using TIP3055, probably due to the draw from the 20 led bulb. But it is very bright, and still last all night.
I bought a new 9 volt battery to use to light a Captret. I can light two leds with it, somewhat dimly, but as time goes by, the leds start to get brighter. At first the leds would only brink. But now they are both on steady at about 50% brightness. I will connect the captret to one of my Jtc tonight and see what happens. In anycase it's pretty neat. I'll need to find which capacitor(s) will work best. I used a 10mf 200 volt, but will try different ones when I get a chance.
Nz

Hi All.
@NickZ Sound like your doing great Getting it to run all night long now that usable light. How's Pentiger coming along? Thanks. Tec

[NickZ]

Guys:
I was given a 20 led warm white bulb. It is perfect for my Jt circuit, and it gives the right color. I am so happy that there are leds now that give off an acceptable light color. I'll be buying more of them soon. The Jt will light all the bulbs that I put on it. I have connected about 40 or so.
I am having shorter duty cycles on a one inch toroid with secondary on it, using TIP3055, probably due to the draw from the 20 led bulb. But it is very bright, and still last all night.
I bought a new 9 volt battery to use to light a Captret. I can light two leds with it, somewhat dimly, but as time goes by, the leds start to get brighter. At first the leds would only brink. But now they are both on steady at about 50% brightness. I will connect the captret to one of my Jtc tonight and see what happens. In anycase it's pretty neat. I'll need to find which capacitor(s) will work best. I used a 10mf 200 volt, but will try different ones when I get a chance.
Nz

[Vaporizer]

Hi Tec & all,
Had an interesting experience. Finished winding one of the 2 3/8" cores, same as the one in the previous post, full 1 layer with 30ga RS red. I didnt count the turns....naturally(to many distractions here), but there has to be 200-300 maybe more. I left enough room for a 3 & 5, trigger & primary of the RS 26ga green. Set it up with 1.25v, a 22ohm and 1K base resistor. I tried a floro on the output after adjusting the LED for max. Well, nothing...ok it takes quite a few hundred volts for a 9w with low current. I exchanged the cfl for a neon expecting major brightness. Hmmmm...not so. It lights both prongs indicating AC so I measured it with my meter. I get 57.8v ac. Actually abt what the neon should look like at that voltage. Disappointingly low after the other test. I upped it to 2.25v, 3v increase on the secondary. TIP31C on this one too.

Kinda puzzling. I'm wondering if the low number primary windings are limiting it? Best guess I have other than the transistor being flaky. I'd hate to think the core have that low a saturation, but I did contact the supplier and he has no data on them. I was trying to keep the ratio of primary to secondary high for a higher output. I was pretty sure I saw Jenna post a 3, 5, 80, but that may have been on a Goldmine core. This is my 1st big jump up in size. Guess I'll swap the transistor and if no improvement try a 10:10 primary and see where it goes. Neither will take much time. I'd like to find out before someone else might buy these.