When rating MH lamps what does the Kelvin refer to? I am use to it refering to temperature (ie absolute kelvin). Does to refer to wavelenth of radiation? bandwidth of radiation?

Thanks Brien

Brien ~

Color temperature describes the color of the light by comparison with a theoretical radiating blackbody and is measured on the Kelvin temperature scale. Unfortunately, some of the light manufacturers are obviously color blind.

Anyway, if you're into astronomy, you know that a red giant is cooler than a yellow star (like our sun) and that a white star is hotter than a yellow star, and that a blue star is even hotter. If you remember this analogy when thinking of metal halides (which are a point source of light) you would know that a lamp with a lower Kelvin temperature rating (like 4300K) would appear yellow, one with a rating of 10,000K would appear white, one with a rating of 12,000K would appear bluish-white, and one with a rating of 20,000K would appear blue.

Kelvin temperature ratings are only marginally informative when refering to metal halide lamps with full spectrum light. So a 20,000K number is meaningless. The numbers assigned to fluorescent tubes are more a marketing ploy than anything else. The only useful information might be in comparing three different tubes produced by the same manufacturer... this might give you an idea of each tube relative to another tube by the same manufacturer. Just don't compare a tube manufactured this year with the same tube manufactured by the same manufacturer a few years back, because they decided to "move up" all their numbers... 10,000K sounded much better than 5,500K.

And forget about Kelvin numbers when it comes to actinic fluorescent tubes... 7100K is just an arbitrary number that was plucked out of thin air.

How's that? Clear as mud, right? [img]/ubb/wink.gif[/img]

N I N [img]/ubb/biggrin.gif[/img] N G

------------------
Irrational Exuberance!

[This message has been edited by Ninong (edited 01-18-2001).]

I knew it would be this question based on the title. Just to elaborate a bit on the above (which is entirely accurate), the theory of black-body radiation was developed about 100 years ago by Max Planck. It is in most physics books because it marked the beginning of quantum theory, and is one of the classic examples about how classical physics failed. It is also the first equation to use Planck's constant, h, which is a fundamental constant of atomic physics that first appeared when quantum mechanics was being born. None of this need worry you, but I mention it to give a bit of historical perspective. Anyway, physicists including Planck were interested in studying the peculiar properties of so-called black-body radiation. One ideal of what that is comes from having a hollow sphere that is perfectly reflecting. You heat up this sphere, and light is generated in the interior but keeps reflecting off the inside of the sphere. If you poke a tiny hole in it, some of the light can get out. Now you can measure the intensity of that light as a function of frequency and the temperature at which you heat the black-body, and what you find is that when you plot along the horizontal axis the frequency (or color) of the light and on the verticle axis its intensity, you will get a bell-shaped curve with one of the tails, that near the red end of the spectrum, tailing off more broadly than at the other (more blue) end of the spectrum. Planck developed an analytic expression for this curve, and it is a function of frequency and also of the temperature, in absolute units, or degrees K.
So for any particular value of K you plug into the equation, you get a different curve, and lower K has a peak in the curve that is more red-shifted, and higher K has a peak that is more blue-shifted.

Stars are good approximations of black-bodies, and fit the curve rather well.

Incandescent lamps are also reasonably good approximations to black-body irradiators, and MH lamps perhaps less so. The main point is that these lamps have so many emission peaks that their emission spectrum looks like an overall smooth curve. You can fit that curve to a theoretical black-body curve by adjusting the parameter K until you get a best-fit. So you could do this for your lamp and get a best fit of say 6500K, which is around what sunlight looks like as well at high noon. Although that is pretty close to the temperature of the surface of the sun, it doesn't imply that the K value is really the temperature of the lightbulb, but rather that a blackbody radiating at that temperature would produce a spectrum more similar to that bulb than temperatures that were higher or lower. So the intent, originally at least, was to quantify the spectral qualities of a lamp with a single number by making the approximation that the lamp's emission spectrum fit a theoretical black-body spectrum. But as Ninong indicated, it is now more of a marketing ploy and you have to take the whole thing with a huge grain of salt.

I thought this "dumb" question would have a simple 2 line answer, boy was I wrong. This is great though, I loved the answers. Heady stuff eh, astrophysics, quantum mechanics, Max Planck.
Thanks so much for the time you spent with the answers, just as long as I don't have to open up my old P-chem book
I'll be happy.
Too bad its mostly fluff marketing ploys though.

Brien

Originally posted by Brien:
just as long as I don't have to open up my old P-chem book
Brien

Uhmmm... That's how I make my living. It is probably in there.