Solar Joule Thief Circuits

[SkyWatcher]

Hi NickZ, thanks for the reply and glad to hear your safe.

The feedback circuit I'm testing right now going on 4 hours so far is lighting the typical bluish/white 3 volt led, it is connected to the collector of transistor and the small regular diode is connected to the high side or other side of coil, of course these diodes are going back to the input battery.

Without the regular diode on the high side, it creates a short circuit, that's why it's used for feeding back to input.

I'm just testing this to see if we can raise the duration with the feedback, if we can then it's a worthy effort i think, without all that battery swapping stuff.

And i bet with the right combination of toroid and coils, i bet it could go a long time if tuned, tweaked, etc., etc.
We will see.
peace love light
Tyson

[NickZ]

SW:
Hope you have some luck with that direct feed-back through the transistor +, to diode, to battery.
Feed back to a capacitor instead of battery would probably work best. Or using both a cap and batteries in parallel.
Switching or swapping batteries in my set up, can be done away with, once the right circuit is found, just by using with a toggle switch.
I had a 1.5 volt voltage drop across the diode that I'm using off of the secondary coil. That's a lot of juice to lose. So I took it out today and ran the positive from the secondary coil through a 20 led light bulb (for now since it seam to like it), and then to the charge batteries (2 in series), no extra diode in there this time. Just to see what it will do like that. Output after the 20 leds going to the charge batteries is about 15 spiky volts. Input source is one AA or two in series. I know that the 20 leds won't light for too long with that much draw, but the idea is to increase the storage mA amount by using C size batteries instead, and keep the same low voltage of 1.5 volts or so. I still think there is an ideal voltage level to try to maintain, depending on the Jtc it might be like 1, 1.5 or 2 volts, to be most efficient on battery duration, or tuned for the highest light output, or a compromise of both, without frying the transistors. Even just two AAs will cook my transistor (TIP3055). I haven't found the right resistor yet, or another pot. Last pot burnt out on me, so I'm a bit handicapped.


Anyways. Time to make my new beach sand cement batteries...

[SkyWatcher]

Hi NickZ, thanks for the reply and information.
I finished the feedback test yesterday, discharging to a predetermined loaded voltage, since it's taking forever to discharge the AA.
Recharging the AA now and will test with just the led and no feedback in a little while.
Though yes, I will also try a toroid with a Jt bifilar and use a secondary to feedback for my next test, so as to eliminate that extra regular diode. Though I'll make it a 1 to 1 winding ratio just to power the one led still.
peace love light
Tyson

[NickZ]

Guys:
There is something I still don't understand about lighting leds off of the secondary coil.
I seams to me that the more leds are connected, the faster the battery's charge is still being used up, similar to what happens when connected onto the primary circuit. Since I'm not using high perm toroids, that may be what is holding me up, plus I don't have a lot of single leds to use, instead I have ready made multi led bulbs.
Has anyone using a one inch toroid Jtc running on a single AA been able to keep 100 individual leds (or more) lit Brightly, not barely lit, all night long, or 24/7 for a day or two. Yes, No?

[SkyWatcher]

Hi folks, Well i didn't have much luck with the feedback experiments, so i decided to try another circuit I've been wanting to try.

A joule ringer variant without the additional bifilar, just using a JT type winding on a toroid.
I'm using a 1" diameter ferrite toroid using 24 gauge magnet wire.
I used 27 turns on each side of toroid, so that means 27 turns on one half and 27 turns on other half as circuit shows.
I used 2 megaohms as shown.
I used 165 nano farad capacitor in parallel with resistor.
I used a 2n2222a npn transistor.
I used a 6800 uf, 25 volt capacitor

It lights up to probably 20% brightness at 12 volts in cap and when capacitor is charged to 24 volts, the led is at full brightness for about 8 seconds, then lowers and flickers and stays steady at about 1/2 brightness for awhile and very, very slowly tapers off for well over 20 minutes so far.

Will have to make better time and drain tests.

Though when at half brightness slowly draining, it's drawing about 400 microamps, then slowly tapers down from there.
It's a very interesting circuit and I'm going to be making many more tests on this circuit.



Uploaded with ImageShack.us

Let me know you folks think.
peace love light
Tyson

[SkyWatcher]

couldn't edit last post, that circuit was drawn by xee2

[dragon]

couldn't edit last post, that circuit was drawn by xee2

Looks just like a variant of one I posted on the Joule Ringer thread some time back.

http://www.energeticforum.com/attach...running-jt.jpg

They work extreemly well. The one shown will run for several hours at about 1/2 bright.

[SkyWatcher]

Hi dragon, yes sorry bout that, you did come up with the base idea, thanks.
I got an hour total run time on my setup, though it stayed half bright for only about 6 minutes and slowly tapered off from there, though i noticed xee2 had better results with air-core.

Could you please share the details of what you used as far as coils, core or not, etc. to get that 1/2 brightness for an hour or more, thanks.
Maybe the additional bifilar is needed.
peace love light
Tyson

[dragon]

The core was a basic 1/2" diameter ferrite toroid with the standard 8-15 turns, bifilar - as a normal JT circuit. The bifilar wound coil on the base circuit was 400 turns of 32 wire with an iron core. Basically a combination of Laser's Ringer and a JT that worked out quite well... actually has the best longivity of all the JT's I've put together. The ones with the capacitor/pot work well but for some reason don't seem to have the longivity. The bifliar base coil is simply a capacitor with a leaky diode ( resistance ), I just haven't found the right combination that performs like that coil yet.

[SkyWatcher]

Hi dragon, thanks for the information, appreciate it.
I'll have to make some experiments with the additional bifilar then as well.

I'm getting fairly good results with my setup.
I changed to 15kohm base resistor in parallel with 1.5picofarad 15kvolt capacitor.
Also blew up my 2n2222a npn transistor and i have no idea why, maybe i had a wired crossed.
So i replaced with a mpsa92 high voltage pnp rated at 300 volts that works very well.

At 1.75 volts input at 8.7 milliamps, the cool white led is probably 50-60% of full brightness, for 15.2 milliwatts.
Same toroid and coils used as before.

My question is, I'm using a 1 ohm resistor in line with negative rail and measuring with volt meter to get current measurement. Is this voltage reading accurate over the 1 ohm resistor, i mean is it averaging the values or since it's pulsed, could it be in error.
peace love light
Tyson

[dragon]

Hi SkyWatcher, If you use a .01 uf cap and a 50k pot you can adjust the current usage and get it quite low. Since the LED is going back to the positive the radiant spike has to be a higher potential than the cap ( or battery ) so you can use higher voltages and reduce the current even more. This will also increase the run time as there is more energy stored at higher voltages.

Readings from the resistor should give you the actual current use, I would imagine this is averaged but should be the actual usage over time. This would continuously change as the capacitor is depleted.

You can also connect a battery to the cap then connect your meter between the battery and cap to read the current directly. Once the cap is charged it should read the actual losses or current draw of the circuit.

[kavinash]

jus a thot

hey skywatcher.... i have this dye solar cell i am goin to fabricate for a project... i was thinking to connect the cell to a joule thief ckt in order to charge a battery which will light lamps... could you please help me out here
.. should i connect the juole theif after battery inorder to drainm all the electrns or after the solar cell to charge the battery completely??????
. using a joule theif will it reduce the life of a battery????

Hi folks, since ou site is down, figured I'd post my promising results here. I recently converted a 12 led lantern that is normally powered by 4 D cells and is now running off 1 AA rechargeable cell in standard JT emitter/collector connection. you'll notice in the pic showing the JT circuit inside the lantern that it has a 1" diameter toroid with 18 gauge insulated solid wire. Now this 1" toroid yielded about 4 hours run time using the 1- AA 2000mah rechargeable cell and the JT had a 150 ohm base resistor, since any higher ohm base would not light the leds very brightly. Then I decided to try a smaller diameter toroid of 1/2" diameter, this using 24 gauge magnet wire and with the same 150 ohm base resistor, run time was 8 hours at same brightness level. Now as we speak, I am running a duration test with the same 1/2" toroid and 24 gauge, though with a 1.5kohm base resistor and run time so far is 14 hours and still going at same brightness level. Many more experiments to be run, until I achieve the longest run time at same brightness level possible. Here are the pics. In the pic showing the various toroids, the 1/2" toroids are marked (1) and the 1" diameter toroid is marked (2). I'd like to hear your thoughts.
peace love light
Tyson


Uploaded with ImageShack.us


Uploaded with ImageShack.us


Uploaded with ImageShack.us


Uploaded with ImageShack.us

[SkyWatcher]

Hi kavinash, thanks for reply.
Since solar cells are not cheap and to get the most out of your cells, i would put them all in parallel and use the joule thief to bump the voltage and charge your battery, depending on battery voltage.
You will have to experiment to see how well the battery is taking a charge and tweak accordingly.
That way you'll have plenty of current and avoid needing lots of solar cells in series/parallel.
I suggest driving your leds directly with correct voltage to try and avoid using resistors.
Though with that said, one can build a very efficient joule thief if tuned properly for resonance, which is more difficult without instruments to see what changes are happening in system to tweaks. hope that helps.
peace love light
Tyson

[shubhamforme]

joukle thief with cop?>1

hi
i've been working on this design of joule thief,and i am able to get a cop of >>>1..
here the circuit--

Free File Hosting & Video Downloads, Free File Sharing, Online Friends Network - Ziddu


coule anyone verify it??
thanks

[SkyWatcher]

Hi folks, for fun and since I only have a couple of AAA rechargeable cells, here is my led ghost light conversion from dollar store.
joule thief uses 1/2" diameter ferrite toroid, 24 gauge bifilar, one AAA NiMh cell, PNP 2n3906, 1kohm base resistor all soldered together.
for your enjoyment


Uploaded with ImageShack.us


Uploaded with ImageShack.us


Uploaded with ImageShack.us
peace love light
tyson