low-cost LED-based "one sun" light source

You might be better off aiming for sun/10 or sun/20 as the target intensity and using a mixture of incandescent to get the IR and boosting it with some of the units made for aquarium lighting.

At least for silicon output varies nicely with intensity up to around sun*3 where heating becomes problematic. If you are on organic devices then they might be more tetchy at solar intensities and high uv content.

No idea if this lot are any good or not but on paper they seem to offer enough LEDs at the short wavelength end to give you a sporting chance of getting something like the right profile as a custom assembly.

They might even be helpful if there was a market for them.

BTW anyone know of a yellow/amber LED with the narrowest possible FWHM centred on 590nm. I'd like to have a solid state sodium lamp... (and a yellow laser is a bit on the expensive side for my requirements)

If ~ 1ms pulses would do, then I suggest a xenon photographic flash. You can also buy fairly cheap long (~ 300mm) xenon flash tubes meant for disco strobes.

The xenon arc lamps used in "proper" solar simulators are about $500 or so, and building a solar simulator with one of those might be worthwhile if you have some cheap labour available. You can buy the power supplies ready made or you might already have a suitable power supply (~ 30V, 1 -

You can also get small, very powerful halogen lamps sold for photographic or projection purposes, but the spectrum will be fairly wrong.

I think you would need a lot of different LED materials to get anywhere near the spectrum of sunlight, and some of the wavelengths would probably be hard to obtain or just unavailable.

ergy)

-color-of-the-sun/

all of the near UV. That's harder to simulate (especially with a single OTS LED) and will affect materials selection.

to generate the Earth-surface daylight spectrum) aren't marketed for peopl e with SAD or for general natural-equivalent indoor lighting. Maybe after t he Feds finally take marijuana off Schedule 1?

Well my point of the picture is that there is a lot of energy at ~1 um (300 THz) (where photodiodes tend to peak.) There are not that many photons in the UV (each photon makes an e-h pair) so missing out at the UV end isn't that ba d.

George H.

Skip the LEDs and go to a 175 Watt or 300 Watt Cermax lamp or one of the many Cermax clones... As the patents wore out, the clones appeared, and the cost went down.

Driver design is not trivial, but I can give you some pointers.

Steve

These are claimed to approximate 5760K but probably not very well:

Only $5 though, and they fit an E26 base.

The same 250W bulb without the blue filter (which cuts the lumens in half) has a color temperature of 3400K and the same 3 hour lifetime, so that's probably the actual temperature of the filament.

I guess anything much over 3400K in conventional incandescent would have too short a lifetime to be practical.

Carbon arc should be close, but they are a bit dirty ...

--sp

Paying ~$1/hr for light is a bit step, IMHO :^) (At the high end photography can get expensive.)

I'm not sure what Win is doing, but it seems to me a lower temp incandescent can provide the same photon flux.. (For PV) Unless he's looking at UV damage or something?

A bank of 100 Watt-ers would be bright. I've got this 250 watt bulb I bought at a garage sale, big base, bare bulb, I bought it to look at the filament. (It's dang bright when you turn it on..)

George H.

You won't get much blue light or UV compared to sunlight unless you run the filament at a temperature similar to the surface of the sun, which is about 5800K. The sun approximates black-body radiation in its spectrum.

So if you're testing heterojunction solar cells or something like that, the results may not be very representative.

Re-reading the question I guess he's looking for a one sun flux so maybe you're right. The 400nm lower limit caught my attention because it's pretty blue - about at the limit of relatively cheap so-called UV LEDs.

--sp

Blue LEDs are sufficient for treatment of S.A.D., non-visible wavelengths give no benefit,

Hi,

Collimation of the light wasn't mentioned, I think that may at some point come into play, as solar cells can be more efficient when they have larger areas of uniform crystallization, so it stands to reason that maybe the collimation of the light could also have an impact on the efficiency of the solar conversion.

cheers, Jamie

It's possible but of what practical use would it be?

Hi,

It's just one possible thing to watch out for, ie an uncollimated LED based "one sun" light source illuminating a solar cell may actually produce less electrical output than the sun if the solar cell is more sensitive to collimated vs uncollimated light, I'm not sure if any solar cells are more sensitive to collimated light, but it seems like highly efficient ones may at some point do better with collimated light.

cheers, Jamie

Sunlight is obviously not phase-coherent, being composed of the entire sp ectrum, but it is pretty nearly perfectly spatially coherent* i. e. collima ted by dint of the sun being 93 million miles distant from your typical sol ar panel.

AFAIK it shouldn't make any difference to test solar cells with uncollima ted light, or the mentioned other chemical and electrochemical "schemes" ei ther.

• That's the origin of those tiny colored speckles on your fingernails that appear in direct sunlight- interference in the irregular layers of keratin reinforces different colors of light). The nearsighted can see them naked- eye but 20-20 types will need a magnifier to see them.

P. S. DAMMIT would somebody PLEASE make an engineering-oriented spielchucke r that doesn't need to be told that words like "collimated" are real words? ??

Mark L. Fergerson

all that well collimated? That may be good compared to a normal light source, but I wouldn't call it nearly perfect.

I can't imagine light being highly collimated would make any real difference.

I think you are just going to have to keep adding words to your dictionary.

lly

"Nearly perfect" was based on some older references.

Nevertheless, raw sunlight is good enough to do the two-slit experiment w ith:

There are papers that take it into account when discussing rectennas:

y.

Yeah, I know.

I don't know why I bother maintaining a lawn at all.

Mark L. Fergerson

Yes. That's another in the reasons I've thrown in the towel and am purchasing an honest one-sun zenon-powered light source + atmospheric filter. But, sheesh, that's gonna be one bright light!

Awesome!!

way too hot though, those LEDs are burning paper. Standby for a life expectancy of 0.0000000000000000000001 suns.

NT

Sunlight is already collimated enough for anything a solar cell will notice, and you can't collimate it any better without either increasing the detection area or losing most of the light.

Seen from the Earth, the Sun subtends an angle of a bit less than 10 milliradians, i.e. a projected solid angle Omega' = pi*0.005**2 = 7E-6 steradians. (*)

Thermal light obeys the law of conservation of radiance: there's a 1:1 tradeoff between area and projected solid angle.

Concentrator solar cells receive light from wide angular aperture, e.g. an IBM project I was peripherally connected with was making 2400 sun concentrators, i.e. illuminating from Omega' = 0.2 sr, which corresponds to a cone of 28 degrees full angle.

Cheers

Phil Hobbs