12 volt powered oscillator led light

[SkyWatcher]

Hi folks, I finally got around to making a 12 volt powered oscillator that is lighting a series string of 20 christmas leds.
As the circuit shows it's using a 2n3055 transistor and the 330 picofarad capacitor is important for getting extra brightness and efficiency from the setup.
It is drawing 100 milliamps at 12.44 volts for 1.244 watts and is at full brightness and actually may be beyond that a bit, by comparing to another string connected to wall outlet.
This led string is supposed to draw 2.4 watts and I'm using half that for the same job it seems.
Also, when i connect, say, 2 strands of pickup coil, it lights to just about full brightness, though it draws about 130 milliamps, then when I hook up the other 2 strands, the amp draw drops to 100 milliamps and is much brighter.
I'm using a 1-1/8" diameter ferrite toroid with 24 awg, 6 strands and 2 used for the bifilar bedini type oscillator.
Please share your thoughts and comments or questions if you wish. Here is the circuit.
edit: removed circuit and reposted update below
peace love light
tyson

[SkyWatcher]

hi folks, forgot to mention one more interesting thing about this setup.
When i place my finger on the case of the transistor, the amp draw drops to 20 milliamps and with a volt meter across the 12 volt battery, the voltage climbs back up much closer to its no load voltage, which seems to verify that it is indeed using much less amp input.
Also the leds maintain the same brightness while this is happening.
So if indeed it is in that case using only 20 milliamps, then that works out to
248 milliwatts to light 20 leds to full brightness and beyond, i say beyond because they seem to have a more whiter color with my setup that with the ac wall outlet, probably the much higher frequency.

Again, let me know what you folks think.

I made this setup for emergencies mainly, though if it's as efficient as it appears, maybe this could have use for house lighting with a few more strings in parallel or maybe series, more experiments will tell.
peace love light
tyson

[SkyWatcher]

Circuit fix

Hi folks, I made a mistake in the circuit, the capacitor is supposed to be in parallel with the base/trigger resistor. Here is an updated cad pic of circuit and a couple pics of setup.
Let me know what you folks think.



Uploaded with ImageShack.us


Uploaded with ImageShack.us


Uploaded with ImageShack.us
peace love light
tyson

[NickZ]

Sw:
Can you upload a picture of it?
This is what I found today. 150 led light solar system, runs on a single AA for 8 hours per night. Total cost is $35 or so.
82FT 150 RED LED 2-IN-1 SOLAR STRING FAIRY LIGHT-XMAS | eBay

Sounds like it is a pretty efficient solar powered system, and has to have some sort of Jt circuit to run 150 leds, off of single AA.

[NickZ]

Thanks I just saw the pictures.

[SkyWatcher]

Hi nick, thanks for reply.
Did you take note of how the input amps decrease when i add more pickup strands in series across led load and how placing my finger on transistor lowers input amps to 20 milliamps while maintaining same brightness.
Something different is happening here.
Also, main reason i made this is to replace using fluorescent lights for emergency or normal lighting and since we kept talking about powering these oscillators with 12 volts and I never did till now as far as lighting leds.
This circuit seems very efficient and the transistor has no heat noticeable.
If I can somehow get the circuit to draw that 20 milliamps or 250 milliwatts without having to touch the transistor casing, this could be a house lighting replacement and even now appears to be using half the power to do the same job.
peace love light
tyson

[NickZ]

Sw:
Might be that you are grounding it when you touch the transistor, I've heard of that happening before. Looks like a basic Jt circuit with several secondaries.
In the link is Kooler Backward Jt circuit, lighting 70 leds on just 27 mA.
‪hartley oscillator‬‏ - YouTube
Although you've seen it before, maybe the other guys haven't
It' not that it's not replicable, I haven't heard of anybody else trying it. I don't have the right toroid. I'm still working on it too though, and hopefully will be able to get it right. I may try the new pancake coil that Lidmotor just came up with for a Jtc. You don't need a ferrite core for that one, and it runs an led brightly on less than one mA.

[NickZ]

These solar panels are the cheapest way to get full 12 volt power to run your whole house using leds, you can also connect to Jtc, but I don't know if there would be an advantage when using 12volt source, might be better just connecting to 12volt 1-10 watt leds direct to the solar panels voltage controler.
‪Man makes Solar Power with Broken Panels‬‏ - YouTube

[SkyWatcher]

Hi nickz, thanks for the replies and information.
I looked at lidmotors steel coil joule thief and used a small ferrite bead toroid and was able to get what looks like the same brightness as he did with 1 milliamp at 1.7 volts.
The capacitor has to be matched correctly to get the best light output for lowest input.
Also, I thought about that 20 led christmas string I have and researched a little more and found out that that it uses the 3 resistors of 740 ohms each to drop the excess line voltage and that is why it draws 2.4 watts instead of around 1.2-1.4 watts.
At the least, this high frequency method of running the leds is more efficient and will probably give longer life to the leds than direct dc or ac house power and is again at the least a good emergency lighting option. I am going to box it up and work out a plug in setup for a separate lamp post type idea for the leds to mount to spread the light out better.
peace love light
tyson

[NickZ]

SW:
Yes, I found out the hard way about the Christmas Tree Light sets, and having to unsolder all the resistors, or it would not light at all. But I did it because I really like the flat top leds that my set came with, as they really spread the light evenly all around. But, playing with the captret project blew them all out. 50 of them, and as well as some other of my favorite leds. Until they all blew this amount of leds gave a nice amount of usable light, and light my whole living room as a night light. I see yours are different.
I'm surprised that you got a one mA draw on the little bead core. Good deal. Now, how many days does the battery last lighting those leds?

[SkyWatcher]

Hi nickz, If you mean how long does my 12 volt - 7ah sla last running my series 20 led christmas string, i haven't made a duration test, though at around 1.15 watts, it would last around 73 hours. Though i would assume getting dimmer as the battery voltage drops.
Yes the ferrite bead worked, though I think the capacitor is what does the trick, without it, it doesn't even light the led. Though the nickel/zinc AA running it, should run that led for around 2000 hours or 83 days, at least so one could see walking in a hallway.
I guess if we made many of these, similar to lidmotors circuit, with tiny ferrite beads or toroids or whatever works, it could make for more useful light.
I'd say 50 of these circuits in parallel running off the same AA might be like the light output of say 20 leds at normal output and yet would still be only drawing 50 milliamps.
Then again, if we could tune them like kooler did with his 27ma setup, that would be even better.
And as i write, i just came across another circuit somewhat similar to koolers, he is using a 20 turn secondary and has the leds in parallel. I think I will have to do something now to try and match these results and I will use the parallel off the secondary configuration.
I have a small ferrite e-core that I will use and just wrap 20 turns over the bifilar and see what I can do with it. Here's a link to another similar result as koolers.
‪Joule thief lighting up 20 ultra bright 10mm LEDS with 1,28Volt and about 60mA (22)‬‏ - YouTube
peace love light
tyson

[SkyWatcher]

Hi folks, hey nick, well the results so far are looking fairly good, not 70 leds mind you, though lol, i don't have 70 leds to even try.
I used that small ferrite e-core and wound 10 bifilar turns, then wrapped over that, 20 turns, all 24 awg.
Used a 2n3906 pnp and a capacitor 104k in parallel with base resistor of 500 ohms. I had to use 2 AA's to get the leds to what I would consider full brightness, too blinding to look at for more than a split second.

So at around 2.5 volts and 41 milliamps input, around 103 milliwatts.
Not a bad start for 5 milliamps per led since I'm running 8 in parallel connected to the 20 turn secondary.
I'm using the 5mm white/bluish leds.
I'm using joule thief circuit with the capacitor added and the secondary.
peace love light
tyson

[gyula]

hi folks, forgot to mention one more interesting thing about this setup.
When i place my finger on the case of the transistor, the amp draw drops to 20 milliamps and with a volt meter across the 12 volt battery, the voltage climbs back up much closer to its no load voltage, which seems to verify that it is indeed using much less amp input.
Also the leds maintain the same brightness while this is happening.
....

Hi Tyson,

In most cases our body is a capacitor with respect to a working oscillator's "hot point" like the collector point you touch. So it could be useful to attach some capacitors in parallel with the collector emitter while watching input current draw. I suggest starting from 100 pF and going up in value to some nF.
However, a tricky thing may happen when finding a cap value (which does reduce input current draw while LED brightness does not change but a little) and then you switch your setup off, then again you want to switch it on with this extra capacitor in place, you may be surprised the oscillator would not start... If this is the case then you have to reduce the cap value again to get a compromise. Oscillators are a beast, to say the least lol.

(The reason it might not start with a "good" cap value is that the starting current in the switch-on moment is less than without the cap, hence the voltage and current values induced are not enough to maintain the positive feedback from the initial switch-on 'kick'.)

Good job!

Gyula

[NickZ]

SW:
Nice results. I think that the secondary output is the way to go, but getting that low draw is a trick. I know what you mean about the 104 cap, I also use some of those, as well as the 103. I used a tiny ferrite core from a cfl on one of my Kooler replications, like the one he had going for 5 months on a button cell, but mine still discharges to AA battery.
They are neat little circuits with some interesting differences
compared to the regular Jtc. My aim was to get the BwJt going off of my cement cells, but for some reason they don't provide any more light than connecting the leds direct to the cement cells, I still don't know why.

It does take more juice to get the circuit to oscillate than it does to keep it going. I don't have too many loss leds left after the captret project, that are not mounted already, to be able to try the parallel connection only.
I just made another BwJt using the TIP31, 104 cap, 3 k resistor, and 3/4 inch toroid. So far it won't start with only one AA, but it will have to change the resistor to a pot. Kooler mentioned that the inductor on the battery negative side was the key to get all those leds to light up to 3/4 brightness off negative at only 27mA. He also did something similar to the guy in the video tht you posted, by putting another winding on top of an inductor.
I don't have a 12 volt battery source but do I do have a 4 volt lead acid battery with a pot on it that works just fine, I recharge it during the day off the grid, and runs all the leds I have, all night. The pot works like a dimmer that I lower down late at night.

[Guruji]

Chrcistmas leds

Hi Skywatcher very good setup and thanks for sharing. Why don't you try a pot instead of a resistor and try to tune for resonance to get more efficiency as Gagdetmall used to do?
Thanks