LED Power Sources...Investigations using LED's as input source power

[Peanutbutter29]

A lot of people are aware that LED's will work (to some degree) in reverse.  E.G-  will output a small bit of voltage in the presence of light.   I thought it might be good to have a thread for people to post any findings using LED's for input power.  Not sure what all has been done...

I was able to do a couple tests of things and may have found some useful innovations.  I'm hoping that maybe these methods can help others improve upon this. 
   I'm sure a lot of you have played around with these LED's more than I and can probably make a better setup than I show easily.  Also I assume, some of the EE wiz's in the forum here could also expand this pretty easily. 

At any rate, I made a couple videos that should explain it enough.

First, I figured out a way to add and more evenly regulate the output from LED's receiving light.  I show assembling one in the first video and I'm sure if you've played with these you know they can't add in series.  With this setup you can add 2 led's and get MORE than either 2 separate (slightly).  This will be constant relative to intput.

Second I show a method to connect these LED's receiving power, into a charge pump.  In this varation, all of the energy the LED's emit is stored and allowed to build.  I believe this to be the most promising of the two circuits.  I also show this stored energy transfers from outer to intermediate terminals and may also represent a negative polarity.  This circuit will drain to 0v if left connected as opposed to the rectifier above that will output a steady amount, but not store. 

http://www.youtube.com/watch?v=h67kcyBdmHQ&feature=plcp&context=C47db8d5VDvjVQa1PpcFPFfx6ZklIUC_2d470drKTSY_48x9Tljtc=

In the second video, (I get a better shot of the light meter at end) I show another improvement / variation to the Charge pump design.  Here I replaced the Lv switching diode with another LED (still reverse biased like diode).  Also, I connected a mylar cap across the main terminals.  This allows for not only a greater storage (more efficient) but also a greater transfer for discharge (which can be done from multiple terminals in a single charge).

http://www.youtube.com/watch?v=sFhuayjkN9E&feature=youtu.be

Now, I just played with these, some and from stuff I had around.  I imagine there are better choices of LED's than I chose.  Also, capacitor sizing will be important and I've not played with that at all.  Also, to note, any diode can be used for the reverse bias;  and may in fact need specific types for various situations.  E.G.  a Zener will allow a controllable max V stored.  Also if you notice the way the energy and polarity changes with the charge pump;  it would seem to be screaming for an AC input (could almost represent C in a LCR circuit). 

Hope maybe these things can help someone with their work.  Please also post if you have findings as well.
Thanks

[Peanutbutter29]

Well I finally had some sunlight to test these two circuits in a more realistic situation.  You can see that I was using light of Appx. 1/10 power production, as from the sun.
  With the rectifier / regulator circuit = I get 220mv continual.
  With the Charge pump it quickly builds up = I show 2v!!!!!!!  (this can max at around 3v from the green led diode)

http://www.youtube.com/watch?v=l9XWtErJL1Y&feature=youtu.be

Again, just trivial stuff atm.  Can anyone think of a way to discharge the "charge pump" circuit or multiples?  I assume this would be the way to take power.  The whole thing would be essentially a modern, crude Tesla Radiant receiver I suppose.  Dopants, plates, caps, timing circuit = power out.

That's all for now,
Thanks

[Magluvin]

You know those little solar sidewalk yard lights? 
It would be hilarious if all you would need is just a few more led's instead of the solar panel.  This would be a good test for you...

Try to make one of those sidewalk lights with just led's(cheap cheap cheap from china on ebay) Even if you needed 10 for the charge side and 1 for light, it should be cheaper.

Have you tried all wavelengths(colors including UV and IR?  of led's?

Those yard lights are cheap enough to buy and compare/test your theory. Those yard lights put out a poor amount of light (1 led) from the ones I have seen.

If you can charge a small cap to high voltage, say 10 led's 2V each is 20v, then discharge that to a converter to charge the storage cap/battery, you might get better efficiency.
Or for quick easy tests, just use the leds in series to charge the storage cap/battery, then run one led for the night to see if it works.

Mags

[Magluvin]

I just did a quick test, with what I have on the bench project at the moment.

I have a 10mm white led from RadioShack, a laser tach(pretty bright) and some led desk lights that mount under the first shelf of my computer desk and are variable in color by mixing red blue and green and the control box.

The laser produced 0 volts. Zero.  The variable lights gave me something I didnt expect.

Red shows 0v and the most, blue gave 2mv. 2.19mv at peak.  But when all the leds were adjusted to white, I couldnt get 2mv only 1.6 mv.

Either having all the RBG led's on took away some current from the blue ones, or the other spectrums caused an opposing affect to the led chips ability to convert the light into voltage. Like the red laser had shown 0v. But without the laser the led reads .22mv in ambient light, blinds closed.  So the red may be a an issue using sunlight, with white leds as the solar receiver. If definitely so, then filtering the red could increase efficiency. Very very Interesting. ;]  Makes me wonder about solar panels in this aspect.  Could a filter help a conventional panel?

Lol, wouldnt that suck if chem trails reduce solar panel efficiency. Just a far fetched thought, but I still thunk it based on what I know. ;]

You should get a nice prism to experiment with the different wavelengths/colors(from the sun, being that is the main solar source) to make a record of the findings for different led's. The prism wil get you from IR to UV, separating all the colors produced by the sun.

These desk lights are inside a sealed fogged plastic lens, so much of the output is scattered, dispersed.

The sun is a couple hours from setting here. The 10mm white led in that sun was just over 500mv. Will try tomorrow at noon during my lunch break.

Not bad info for 5 min of tests  ;]

Mags

[Peanutbutter29]

Glad to see another experiment!  Try and be quick....

ya the clear or white LED's don't seem to output well for low light.  Some colors, interestingly will output in VERY low light (add charge pump, lol). 

**I assume this is related to the "trapping" of certain wavelengths related to the color of the LED coating.  (similar to greenhouse)

The prism idea, excellent and yup could work.  Would be a good way to show electrical powers of various spectrums (or alterations to said spectrums *coughs* chemtrails, etc).

For your multicolor LED.....I assume this is a 3 lead type?  I have one but, haven't tried yet.  I assume the white you mention is across both LED's?  if so then they normally won't add;  though they may behave different with only one cathode shared by 2 anodes.  Be interesting to see how they can connect.

thank you for the results!!