Jim and Rich had thought quite a bit about a scope that was needed over the previous three years. In 1988 we all discussed it at dinner one night having obtained input from a number of engineers and astronomers. Jim laid out the detailed characteristics of the survey telescope required. To observe a fair sample of galaxies, and to reach quasars at the edge of the Universe in a period of five years, required a 2.5 meter telescope. The telescope would need two instrument types: spectrographs and a camera that could take 5-color images. The field of view of the telescope had to be about 7 square degrees, an unheard of sky coverage for such a large telescope. For reference, the Moon is half of a degree in angular size. The spectrographs would have to take spectra of at least 400 objects at a time; we finally settled on 640. The camera required 30 large CCDs, a number never acquired up to that point.
SDSS Plates: Rich Kron placing fiber optic cables in each of the 640 holes drilled into the plate for observing. The entire process could take up to one hour.
Credit: Thomas Nash, Sloan Digital Sky Survey
Two spectrographs fed with fibers from the 640 simultaneously acquired objects were needed. Images from the telescope were used to determine relative positions on the sky of galaxies quasars and stars. Those relative positions were used to drill holes in large 30-inch, thin aluminum plates. Fibers were placed in the holes and the plate would be mounted on the telescope. When the telescope was pointed at the proper place on the sky, each of the fibers would be lit up by a different object, so that 640 individual galaxy spectra could be obtained simultaneously. Previously, most telescopes could only obtain one spectrum at a time, so this made the 2.5-meter telescope hundreds of times faster than typical telescopes used for this purpose. The 5-color images could be used to do a preliminary classification of millions of objects and a pre-selection of objects for which we needed spectra could be made. The first five years of the project acquired images of two hundred million objects and spectra of over one million objects. For various reasons it actually took us eight years to reach the five year goal.