I would like to make cheap simple device to measure the intensity of
various wavelengths emitted by the sun (aka spectrometer). It would
only need to cover the visible spectrum. With this information, I
could use Wein's law to estimate the temperature of the sun by finding
wavelength of maximum intensity. Looking for direction or suggestion
on how to accomplish this. Thanks.

---

Clarification of Question by bra00424-ga on 16 Oct 2006 07:10 PDT

Also, this is just for fun ... a demonstration for class.

---

Clarification of Question by bra00424-ga on 16 Oct 2006 12:36 PDT

I have thought of something...Get a photocell and use color filters
and read induced voltage to get idea of different
intensities....Should get maximum around yellow light..

---

Request for Question Clarification by hedgie-ga on 18 Oct 2006 08:27 PDT

Hi, bra...

Interesting question.
The digital camera idea is not bad - and  I can find formula to convert 
RGB values to frequencies.

But for a class demostration I would recommed to replicate classical Newton's
experiment - and add a bit to get spectrum max.

Newton's  experiment:  

http://physics.bemidjistate.edu/archives/Col122502newton01.htm


and a diode at green (not yellow) will show max  (at noon) --> T= 6000K

Do you want few references on that (for $20) ?

Hedgie

One possible method is as follows:

Use a digital camera to take a photograph of the sun.  The resulting
image file will be millions of pixels, and each pixel will consist of
varying amounts of red, green, and blue.

You could probably write a simple algorithm to average these values
for each pixel and give you a mean effective color.  You can convert
that color to a wavelength, and voila, you can use that information to
find the temperature of the sun.

The only problem is that you don't want to completely saturate your
image, which would result in a bright white sun.  A bright white sun
will not really work for this method.  So, you'll have to access the
manual controls on your camera and underexpose the image as much as
possible.  A large f-stop value combined with a high shutter speed
would do the trick.

It sounds a bit tricky, I know, and unless you're used to photography,
some of what I just wrote might make no sense.  If it's unclear, ask
one of your photographer friends on advice in taking the photo.

The camera trick would not really work, since the rgb values are really
outputs of fixed filters. You would essentially get three different weighted 
averages, and not the complete spectrum. If you need to use a camera for 
recording you really must separate the different frequencies first, using 
a prism or something similar. Taking a picture would destroy this information.

However I don't think that you need the complete spectrum, just the maximum. 
As a first approximation I would propose the following:

Set up a dark room where you allow a single ray of the sun to enter. Let it 
pass through a prism, and use a moveable slot to block everything from the 
output other than a single frequency. Use a photometer at the other end to
measure the intensity of the light, and find the location of the maximum.

The one (big) problem though is how do you calibrate your measurements, that 
is how you find the wavelength corresponding to each slot position. One 
approach would be to use the ends of the visible output, and mark them with 
the (known) values for the frequencies at the ends of the visible spectrum.

I hope this gets you started, although it certainly needs more work to make
a practical experiment.