When you get down to it, solar cells aren’t much different from the diodes and transistors in your parts drawers or inside your beloved electronics. They’re both made of silicon or some other semiconductor, and surprisingly can produce electricity in the presence of light. Here’s two semiconductors-as-solar panel projects that rolled into the tip line over the past few days.
[Steven Dufresne] cut open a 2N3055 power transistor to expose the semiconductor material to light. In full sunlight, he was able to produce 500 millivolts and 5.5 milliamps. In other words, he’d need around 5000 of these transistors wired up to turn on a compact fluorescent light bulb. A small calculator has a much lower power requirement, so after opening up five transistors he was able to make a solar-powered calculator with a handful of transistors.
[Sarang] was studying solar cells and realized a standard silicon diode is very similar; both are p-n junctions and the only real difference is the surface area. He connected a 1N4148 to a multimeter and to his surprise it worked. [Sarang] is able to get about 150 millivolts out of his diode with the help of a magnifying glass. While he doubts his diode is more efficient than a normal solar cell, he thinks it could be useful in low-cost, low power applications. We’re thinking this might be useful as a high-intensity light detector for a solar cooker or similar.
After the break, you can check out the videos [Steven] and [Sarang] put up demonstrating their solar cells.
33 thoughts on “Using Diodes And Transistors As Solar Cells”
Sounds like a waste of some nice transistors to me.
Naw, I can still desolder and then use them – probably will someday. Just have to be extra careful of the delicate wires. At least I stopped myself from digging up five more to connect in parallel for more current.
Interesting configuration for the transistors. I’m not in a position at the moment to think this through completely, so I’ll just put it out there to discuss.
It appears the energy is produced at the PN junction. Couldn’t you tie both E and C to the previous transistor’s B and effectively have your power sources in series/parallel? Wouldn’t that increase your current output?
Jeff, I measured C/E with respect to B and got the very same voltage and current as with just C with respect to B. This was with cool white compact fluorescents as light source. Strange, I know.
Hey have you tried to illuminate leds? it’s very sipler i think…
few years back the bbc/pbs had a show called “Rough Science” http://www.imdb.com/title/tt1292463/ each season they would grab a group of sciency people and put them somewhere and task them with objectives (kinda like Survivor, without the lameness) In the Ep linked above they sand off the top of a transistor and use it as a light detector to make a light beam communicator.
led’s will also generate a small voltage.
Well, an LED is a diode too. :)
LEDs are much more efficient because is packing is clear. They are often used in BEAM robots as light sensors because they don’t need to be powered!
Even better, the LEDs are spectrally sensitive, so you can use them as rough spectrophotometers. Forest Mims has written extensively about it.
correction: Forrest Mims. And here’s one of a million links: http://www.forrestmims.org/scientificresearch.html
Reminds me a bit of this show too: (from 7:15 they saw the top off a 2n3055 and use it to make a simple fax machine.)
The Secret Life of Machines is the most awesome show ever. I always recommend it to everyone, techy or not.
quick way to get the top open is to shoot the top with a BB gun
The 1N4148 converts HEAT to current. The temperature-voltage curves are in the datasheet.
Pretty old trick, the first time I saw this was in the mid 70s. Not worth the waste of an old though still very useable transistor if you need a real solar cell, but informative.
Few pages from late ’70s magazines with projects using transistors and diodes as solar cells: http://commonemitter.blogspot.com/2012/04/solar-battery-made-from-transistors.html
These surely are no match for real cells and I don’t think that cutting working transistors is a good idea, but damaged (with one of the junction damaged) transistors can be used also – that what is recommended in first article. Back in those years it was also possible to buy at reduced price factory new transistors that had worse that specified parameters but it was not a problem when using them as solar cells.
I tried this back in the early eighties when i was about 12, was a very cool experiment!
in my opingion, logically, there theoretically would be 5 or 6 possibilities of failure mode in a transistor;
e-c open, e-c short,
e-b open, e-b short,
all short, all open(very unlikely)
so is there a 1 in ~5 chance of using a blown transistor as a solarcell?
hmmm next time i have a to3 thats blown, ill try it and see! :)
ill post results if it works,,, one day…
Good to know for the coming apocalypse.
Connect the power pins on an EEPROM, point the clear window at a light source and you have a little solar cell.
I was surprised years ago, when I opened up some old “tin can” type transistors, to see a faint light as they operated.
It had to be dark, because there was only a little bit of light. But they were acting like LED’s.
Come to think of it, this could be a pretty cool project, some digital circuit, with a bunch of cut open transistors, emitting light everywhere…
The trouble with diodes and LEDs is that they don’t have enough area to give you much current. I’ve measured LEDs putting out 1-2v, but only about 10nA.
And while this is a neat hack, I see 2n3055s listed at Mouser for about $1.25 each in lots of 100. That’s kind of pricey for low-yield solar.
If you want cheap-and-cheerful solar, flip the circuit around and buy a $1 disposable calculator from Walmart. The solar cell in one of those will put out about 100mA, IIRC. This is also the season when solar-charged LED lawn spikes go on sale, and those only cost a couple bucks each. You get a fairly nice solar cell, a detector of some kind, a rechargeable battery, an LED, and possibly a few other components worth using.
I read a radio shack book about electronics back in the 70s, was it by Forrest Mims? Anyways, in the discussion of how holes work, I got the distinct impression that it was obvious that all semiconductor devices would emit light while operating, and if exposed to light, would produce electricity. Which is why they’re all (unless this is the desired function) in either sealed cans or black ceramic/plastic cases…
I would think an “exposed” VLSI chip would have all kinds of light induced error problems…
Power transistors make good alpha particle detectors.
If you saw off the top and place the emitter from an old smoke detector near the die, you get a series of extremely short pulses (less than 1us, shorter than I can easily measure) which can be amplified and logged.
The would also probably detect muons and other cosmic ray species, but the capture aperture is not very big.
I strongly suspect (but cannot prove) that the height of the pulse reflects the total energy of the particle and the angle of incidence of the particle to the die. If this is correct, it would be a way to make inexpensive alpha (and other particle) spectrometers as well.
Also, an array of these could detect sequential incident particles (for energetic particles, such as cosmic rays), and thus give an idea of the angle of incidence – telling you roughly which direction the particle came from.
Just a thought…
This I gotta try! I have the transistors already cut open, I have dies from old smoke detector tucked safely away, now I just need to make an amplifier circuit. Thanks!
A recent development, not yet escaped from the lab, is an LED panel that works equally well as a PV panel.
Only took 72 years to get there since Robert A, Heinlein’s short story “Let There be Light” in which he wrote about light emitting flat panels that worked just as efficiently at converting light to electricity, years before anything solid state in electronics was invented.
Read the story here. http://archive.org/details/LetThereBeLight_393
pls how can I construct the one that will power my inverter ?
You can’t. Or you can but you would need far too many of them and it would be hugely expensive. It’s much, much cheaper just to buy a commercially made solar panel.
I connect this three photo doide in series and when I expose them to sunlight I used to get 2.8v but to my greatest surprise it doesn’t even light just 1 LED . Why pls some body help me .
How much voltage does your diode need? How much current? If you have enough voltage, it’s possible you don’t have enough current, in which case you’d need to get three more in series and then connect those three in parallel with the first three in order to get more current, and then keep doing that until you have enough. But based on your previous comment, it sounds like you have an unrealistic goal of powering an inverter.
Can you please differentiate solar cell ,diode operations.
Right up there with scraping the black paint off an OC71 (“What’s an OC71, Grandpa?”).
[Note:- According to Andrew Wylie, the OC71 trick is (techie) urban legend anyway …but I still remember trying it, many, many moons ago; just can’t remember whether it worked or not.]
Please be kind and respectful to help make the comments section excellent. (Comment Policy)