Electronics Inductance charging

[Confused Fishcake]

I have made some glowing juggling balls, just a few leds inside a translucent ball, but they were never very bright, and I was going to make some new ones. Then I noticed that if you hold an led near a high voltage source (it happened to be a ccfl invertor) the led would light. I am trying to charge the 2 X 2000mah nimhs wirelessly. So far I have:

12V->CCFL invertor

thin wire wrapped once around the high voltage end.

one end of wire left open

one wire connected to + terminal of LED

Other leg of led connected to bit of foil.

Using this circuit, I can get the LED to glow. Considering how simple the circuit I made was, I feel sure that I can get the batteries to charge somehow. I will probably not end up using the CCFL invertor, and will wind proper inductors, but I would really appreciate it if anyone could say if this will be feasible, and what parts I should use. I am fairly adept at elctronics, and can solder etc.

[ConKbot of Doom]

The thing that lets that work isn't the high voltage so much as the high frequency. Make an inductor and hook up a transistor to it, and use a 555 to drive it at a few Khz (preferably over 20Khz, or 30 khz if you have animals) with 5 or 12v (depending on how much current each draws) and have a matching inductor (or possibly one with more windings) in your ball, lay the one inductor neat the other, and you basically have a small transformer with an air-gap in the core. Rectify and regulate the voltage, and use it as you see fit.

[cpemma]

This is Thomas Edison lighting his laboratory with fluorescent tubes, no wires attached. Just a big induction coil in the middle producing HF radiation.

I didn't know it also worked with LEDs, but I do know LEDs will also act as a light-sensitive cell. Bet you need a few hundred to charge a battery though.

[Moriquendi]

wasnt that tesla?

[Confused Fishcake]

conkbot, how should I make the inductors, and do i connect them across the output of the transistor, and treat the reciever inductor as a "battery"? Do you think this idea will be practical, I will need to draw about 50ma minimum at about 5V. Also, for the juggling balls, can you buy reed switches that are normally close, but open when near a magnet?

[ElThomsono]

This may be of some use to you, if not you might at least find it funny.

Hope it helps

[Confused Fishcake]

Thanks Elthomsono, I am going to buy the stuff he used, and see if I can get it to work. Only problem is, I will need a 40A 12V power supply to charge them at the same time. A bit like solid-state central heating. Anyway, I will post my progress when the stuff arrives, would anyone be interested in a how-to charge 2AAs wirelessly? If I am making it I can easily take notes/pictures. Anyway, I shall probably post my results in about 4 days.

[ConKbot of Doom]

Originally Posted by Confused Fishcake

Thanks Elthomsono, I am going to buy the stuff he used, and see if I can get it to work. Only problem is, I will need a 40A 12V power supply to charge them at the same time. A bit like solid-state central heating. Anyway, I will post my progress when the stuff arrives, would anyone be interested in a how-to charge 2AAs wirelessly? If I am making it I can easily take notes/pictures. Anyway, I shall probably post my results in about 4 days.

You can accomplish better results than afroman if you use an inductor with a higher inductance, a simple pancake coil like he used isnt going to have too much inducance, so a lot of current will flow. Add an iron core and more windings, using proper enameled wire, and your current usage should go down a lot, and your effeciency should go up a lot. You could either buy an inductor, or find a piece of steel to use as the core (though steel would have more eddy current losses than a powdered ferrite core)
Also, watch out for RFI, a simple RC oscilator like he has there can get into rather high frequencies if you have too small of a capacitance. (of course if you want a really high frequencey output, get an inverting mosfet driver, and hook its input up to its output, along with the mosfet, then the pretty small gate capacitance will act as the only cap in the circuit)

If you can, go salvaging, finding inductors might be possible.

[Confused Fishcake]

Thanks for the tip conkbot, i took apart loads of old PSUs and old electrical equipment and found loads. The trouble is, for some reason it just doesnt want to work. If I wire an led to one inductor, and bring it near a pulsing one, nothing happens UNLESS i switch power to the circuit on and off quickly. Any idea why this is? I thought it might be the mosfet overheatin, but im not sure.

EDIT: I was using a transistor with a massive voltage drop, I changed it to a TIP3055 and it will reliably light an led at a distance. I just need to find some better inductors now.

[fat-tony]

What were you using for a transistor previously? I'm just curious

Edit: I registered just to ask you this I've lurked for a long time, but now I think it's time to post.

[cpemma]

Originally Posted by Confused Fishcake

EDIT: I was using a transistor with a massive voltage drop, I changed it to a TIP3055 and it will reliably light an led at a distance.

A couple of things with the Afro circuit, 1k and 100n won't be all that high a frequency, although go too high with that circuit and you'll interfere with radio & TV reception.

5V is a bit borderline for MOSFET switching unless it's a logic-level type. A 12V circuit might be better (but use a fuse).

Main problem with a 3055 is its low gain, it draws a high base current with heavy loads, a MOSFET wouldn't. And a diode across the inductor coil would be a good idea.

[Confused Fishcake]

Fat tony, I was using some sort of FET, that got really really hot - I still have th omprint of the back of its heatsink singed onto my finger. cpemma, i am powering the NOT gate with 12V, it can take up to 15.

[cpemma]

Originally Posted by Confused Fishcake

Fat tony, I was using some sort of FET, that got really really hot - I still have th omprint of the back of its heatsink singed onto my finger. cpemma, i am powering the NOT gate with 12V, it can take up to 15.

Higher frequency would help, then. Try 10k & 2nF. And maybe a bigger heatsink?

[Confused Fishcake]

I used a 10cmX10cm heatsink, and it got that hot after less than a second, and I find that when I adjust the pot, there is a tiny spot in the middle which changed for each inductor which worked much much better, so I think I am getting the optimum frequency. So far I can light an led brightly enough to burn it out at 2mm air gap between the inductors (2mm is acceptable), and I can drive a 555 monostable. TBH, the circuit I have at the moment is good enough, but it will be a bit slow to charge the batteries. I am planning on having 3 or 4 inductors on the inside of the ball, lining up with 3 or 4 on the outside wired in parallel. Trouble is, I only have 1 really effective pair of inductors from an old printer power supply, and I will need at least 9 pairs. What I really need is some heatsink compound, although the 3055 is still cool enough to touch after 4-5mins. Do you think a higher frequency would really help? since the output voltage peaks in the middle of the range for my current circuit.

[cpemma]

Originally Posted by Confused Fishcake

Do you think a higher frequency would really help? since the output voltage peaks in the middle of the range for my current circuit.

You will need the two inductors able to resonate with each other, so maybe you've hit the sweet spot for your coils. I was thinking a higher frequency would give higher impedance with the coil, so less current & heat.

[Confused Fishcake]

I have recieved the parts in the post, and have made a prototype (bigger than it will need to be) and it works perfectly. I estimate the charger to be at least 25% efficient, and after I come back from holiday (1 week) will order the leds.

EDIT: I have run into a spot of bother, I am using changeover reed switches, so that when I put a magnet near the juggling balls, the lights turn off. The trouble is, the switch has something like a 100 ohm resistance. Also, I am using 2 2600mah nimh batteries (Only £1 each!) to power the leds, and for many of the leds on the internet, 2.4V is not enough. I would like to use red, green and blue leds, and need some help to find ones which will work at such a low voltage. I hope to use 12-14 leds per ball, and am still struggling with the reed switch problem. The only thing I can come up with is using a very low leakage transistor, but this will introduce a voltage loss. I would be really grateful if someone could help me with this, I will be away for a few days and will need to buy the parts as soon as I get back.

[fat-tony]

Originally Posted by Confused Fishcake

The only thing I can come up with is using a very low leakage transistor, but this will introduce a voltage loss. I would be really grateful if someone could help me with this, I will be away for a few days and will need to buy the parts as soon as I get back.

Could you use a MOSFET or something like that? It will incur very little extra current, and should be able to switch using a 100ohm input.

As far as increasing the voltage goes... you might want to investigate Maxim's line of step-up converters. I hope this link works: http://para.maxim-ic.com/cache/en/results/5074.html

[gideon]

this reminds me of tesla's free energy plans, basiscaly you are building his idea with the invertor as the tesla coil and the led as the lamp whci receives the energy through the air.

[Emerson]

would there be a way to have this same type of idea, but have just LEDs in something, and when you set in on a pad, they light up, and when taken away they don't

[Confused Fishcake]

Yes there would, I can get about 100ma at 3V wirelessly, are you planning on doing that? If so, I can give you the schematics.