Make a Solar Panel using Diodes!

So, solar panels are made of silicon...diodes are made of silicon.  Have you ever wondered if diodes conduct current induced by light?  Sure you have, and so have I, so let's experiment and make a solar panel using 1N4148 (run-of-the-mill) diodes.   This is just a brief instructable, as the full construction is left up to you, intrepid reader, as a fun project. 

Check out the video below to see how much voltage you can get from just four diodes.

Step 1: Silicon Diode

A silicon diode is a two-lead semiconductor that gates current flow in one direction.  The symbol below shows how a diode is lined up with the schematic symbol.  The image was taken from http://www.gadgetjq.com/single_fire_tach_adapter_diode.jpg for copyright purposes.  Diodes are used in circuits that convert AC voltages to DC voltages, and also as voltage regulators, clamps, and multipliers.

Current flows in the direction of the arrow.  A few other terms of technical use are:

Forward-biased is when the anode is more positive than the cathode, and reverse-bias is the opposite:  the anode is made more negative in voltage than the cathode.

There are different kinds of diodes, too.  Pin diodes, germanium diodes, schottky diodes, rectifier diodes (p-n junction diodes) name most of them.  In this instructable we're going to work with silicon-based pin diodes, although if you're really curious I'd encourage you trying the different kinds of diodes to see how it all shakes out.

Diodes are pretty cheap.  You can pick up a pack of 50 from Radio Shack for around $3.

 

Comments

[DIVYA GARG]

I think you have used pn-juction zener diode .............Can I use a simple pn-juction diode . I mean as looks in this picture............

[nevdull]

Yes, I believe you should be able to.
Good luck!

[kdorji]

I am not able to go beyond 4 to 5 diode connected in series and i was able to get 100 to 150mV. But going beyond that numbers of series connections i saw voltage drops to below 100mV which is confusing. Please let me know how to increase it to usable Voltage level..or give me website to refer.

Regards

[dpeek]

lost of power converted to heat.

[robot1398]

by connecting diodes in series or parallel is there any difference in voltage

[dpeek]

connecting them in series will increase the voltage, while connecting them in parallel will increase the current.

[zis3485]

Hello sir..It's an awesome project....I am very interesting about this project and want to make one for me. You write "You could add in capacitors....throw in an op amp or even a charge pump and turn mV into V. " but I am very new here and don't know how do I do it. Can you please give me a diagram? Thank you

[macman808]

what you need to do is add a joule thief to the circuit. because then you have 3 volts >:-)

[kdorji]

Good Idea .... May i have circuit diagram please or the web site where i can go though for it?

[nevdull]

That's an interesting thought. :) There are some charge pumps that operate on a very low voltage and if a joule thief circuit could boost to 3V, the charge pump could easily make that 6V or 9V, although I think the current would severely suffer. Thanks for the comment!

[macman808]

yes the current would suffer but if it didn't, some charge pumps can turn 3v into 12v and then you just need an inverter to make that 240v.

[nevdull]

I bought five 3", 5", and 7" LCDs at dirtcheap prices the other week from my local surplus electronics shop and put together a few CFL inverters for the backlights, but much to my chagrin it still required a -24V on VLED and all the charge pumps I was willing to buy (ie not over $5) couldn't invert more than 10-12V *and* perform a boost (although I saw a few more expensive ones that could and found a nice voltage tripler pump), but figured how to string them together to get

V_out = 2D - (2 * V_in)

I wonder, though, how much of an AC  voltage swing you'd get if you're pushing, say, 10 uA.  Sadly, I'm not sure it could even meet the charge pumps I_in requirements.  It would be interesting to see if you could get 120 or 240 VAC from it, though. 

[dsandds2003]

It is kinda funny....in the 80's i used to scrap out old tobe radios for copper ans other metals. I found that those old selenium rectifers also produced electricity when exposed to light. I used to 4 or 5 volts from 3 or 4 of them hooked togather. But back then their wasn't much use for these electricity producers.
Foe those of you who do not know what these are, they were square pices of metal. Usually about 3" by 3" and about 1/32" thick. They were stacked with spacers between the plates and anywhere between 6 and 10 plates per rectifer. Now you can replace then with 1n4004 diodes.
Dave

[beehard44]

how many volts did you get in the PCB panel?

[nevdull]

with my configuration I was only able to raise a few hundred millivolts. :-\

[ljsggh2n77]

I believe that it is so far out and crazie that is just beautiful man!!!!!!! Cheers!

[nevdull]

hey thanks!
Cheers!

[drresearch]

This is great,
I used only 2 zener diodes (I think 5v, but not sure) and got up to 2 V in parallel and even 3 V in series, though I did not manage to light up an LED. But I am planning to try buying more of these diodes and making a whole PCB full of them or something and see how much power I can get. I'll let you know.

[nevdull]

That's awesome! I'll be curious to know if you ever manage to light up that LED. Keep us posted on your progress and feel free to post any schematics/pics/etc to this instructable.
Cheers,
/nev/dull

[drresearch]

Well, I am sorry that I took so long, but finally I tried and bad news, with about 20 diodes in series or parallel the voltage is not even as much as I had got before (I am afraid they might sold me slightly different diodes in the shop, there they make confusion sometimes). I might try again, but not very soon, perhaps in a few weeks.

[nevdull]

No problem. Too bad it didn't source more voltage. Maybe a new design might help if you're tinkering out with it again! It's always fun to try new setups, I think. Thanks for the post!

[drresearch]

Well, I forgot to mention that the diodes had to be almost touching the light source (I used a 15watt CFL bulb), so perhaps it's not that great, but I will still try. I'll also post a picture most probably.

[onesuperdon]

lol, it neva crossed my mind that a zena diode could be used to generate current :P but it makes perfect sense since it does contail n-type & p-typ semiconductors. I have so many of them at home, so gonna start as soon as I get there! Yay, excitement ^_^

[8v92]

This is a good use for clear glass diodes found on discarded circut boards in older computers and other devices. The semicondutor junction will only convert a small part of the visable spectrum to electrical current because of the band gap potential of the PN junction. So a diode or infared LED will have a peak output in red light , a green LED will only output in blue light , a blue LED output in ultraviolet light . It takes red, green and blue light to produce white light so a white LED has an output voltage of about 3 volts in full sunlight because it can convert most of the visable spectrum. The output current of a single diode or LED is too small to power anything however if an electrolytic capacitor (2000 mfd 10vdc) is connected for several hours in full sunlight a small amount of energy can be collected. For maximum output the diode or LED must be aimed at the light source and any size large capacitor will work.

[E_MAN]

how many volts did you generate with the pcb shown? I would like to charge a 12v lead acid battery, I would like to know if it is remotely possible.

[nevdull]

i generated a couple hundred mV but there are other configurations that could probably get you more. however, i dont think it would have the current to push as a recharger unless you used some intermediary caps or something, or again, changed the configuration of parallel and serial connections.

[E_MAN]

ok, thanks. I did some calculations, and it was not very efficient, but a good idea anyways. :D

[chriskarr]

May I know what your calculations were? =)

[E_MAN]

sorry, but i do not remember, i did it awhile ago, try for your self, i remember that they were kind of easy. I found that they were not worth the price. Buying a 12v solar panel would be cheaper and smaller.

[chriskarr]

So, I take it you're measuring the voltage output vs. size, correct?

[E_MAN]

Yep :D

[chriskarr]

Well, if you're going to be 'doing calculations', you'll need to take into account the amount of amps which are 'generated'. For more details on the subject, I recommend joining 'www.thegeekgroup.org', a forum dedicated to advancing the knowledge of others and giving others resources with which they might do something amazing.

[chriskarr]

Another thing which shows the very-cool aspects of this project - parallel voltages remain the same, but parallel currents add. You should see how many µ-Amps you're producing. =)

Also, you'll get a higher voltage across the system if you place the diodes in series...

...then again, there's also the fact that a 1N4148 diode has a forward voltage-drop of ~0.5V, so I'm not sure if that would affect the operation, or not..

[dnlgreenwood]

How about using a step up transformer? I need about 10v to run my Sega Game Gear and it would take a while to get the thing started, but its worth a try!

[dnlgreenwood]

Also: I did some calculations, and it would take about 400 diodes to get 10v if 4 diodes produce 0.1V. ratio of diodes to Voltage: 2 to 0.05 2*10/0.05=400 This may be inacureate because of unconstant conditions. For example, if there are clouds obstructing the sun, or if you are not in direct sunlight, a lower voltage will be produced. this would be better for charging batteries than running a device such as a Game Gear or a GameBoy. you can try hooking up some NiCd rechargeable batteries to your diode panel, and see what happens, and how long it takes to charge, then test them on your device. It is a good idea to use a voltmeter to make sure you won't destroy the batteries!

[nevdull]

that's pretty cool....it would then be more like a real solar panel (in the sense that it's a panel. i don't think it'd ever come square inch for square inch to rivaling the output of a real solar panel, tho). I tend to think of this as a technique for generating power when the zombie invasion begins... :)

[markjoseph]

i suggest using an LED connected in series..and use it in reverse bias condition..this would act just like a solar cell..(white LED is more prefereable)

[hamid manzoor]

Mr, nevdell. How much maximum power you can get from this idea? How many maximum no of diodes you can link togather to get 12 Volt+1A? Thanks.

[thebair]

Will any other diodes work?

[nevdull]

Yeah, but they need to be clear (ie not opaque like germanium diodes and rectifiers). See the comments for a more detailed discussion of which diodes, configuration, and other components (ie LEDs) that can be used in this way.