DIY Digital Spectrometer

Here's a spectrometer/spectrograph I put together with a digital camera, a transmission grating made from a CD-ROM, a cardboard tube, some aluminum foil, tape, and hot glue:

The camera is a Digital Concepts 4.1 megapixel unit, with a fixed-focus lens and equipped with a 512 MB SD memory card. It was modified for remote triggering for a model airplane application, for which it turned out to be unsuitable (the camera is very sensitive to movement as it takes a picture, easily bluring the image). The transmission grating was made from a CD-R disc by cutting out a section with an aviation snips, then removing the reflective foil backing by adhering a piece of packing tape to it and pulling it off. The resulting grating was hot glued to the front of the camera lens with the tracks oriented vertically. This orientation causes the spectra to appear to the side of the light source. The angle to see second-order spectra was determined experimentally, and the cardboard tube from the center of a roll of plotter paper was cut appropriatly. The camera and grating were then hot glued to the angled tube and some aluminum foil added to block stray light:

The other end of the tube was equipped with an optical slit made from two pieces of folded aluminum foil, approximatly 1 mm apart.

On August 12, 2007, I tested the completed spectrometer on the sun:

Some Fraunhofer lines are visible in this specrum, though the instrument was not aligned exactly with the sun. Proper alignment resulted in the camera being saturated, so I settled for an off-axis spectrum. Another spectrum was shot of the sky even farther from the sun, in which the lines are not visible. The spectrum also shifted towards the blue end of the spectrum, which may be due to internal reflections within the tube:

To test the unit with a less-bright source, I went indoors and aimed the instrument at a compact flourescent light bulb. The results look good:

Using an image processing program, the brightness and contrast can be boosted to make more emission lines easily visible:

On August 14, 2007, I cut down the tube length from over 3 feet to about 1 foot. This made the instrument far easier to point and handle. I also put a proper switch on the ends of the shutter control wires to make the unit easier to use. Lacking a slit on the input end, I none the less shot some test images, using my fingers over the end of the tube to make a crude slit.
Here's a standard flourescent light:

A white LED flashlight:

Another shot of a compact flourescent light, and the same image enhanced:

A red laser pointer has a nice tight spectra, though the intensity overloaded the camera a bit and washed out the color:

A blue LED, without a slit. You can see both the second order and third order spectra:

Spectrometer links