Next: Traps and Bad Columns
Due to "under thinning" , good resolution has not been achieved with the some recent Ford chips such as the GCAM device. The FWHM of arc lines seems to be limited to about 2.7 pixels.
Despite the fact that the chip is 45mm long, the usable spectral coverage is on the order of 36mm. The spectrograph optics cause appreciable vignetting near the edge of the field. In terms of intensity, the illumination falls off fairly sharply in the first and last 200 pixels. In addition, inside of this the chip is not being illuminated by the entire beam, and the line shapes become irregular. Specifically (and this varies from grating to grating), below pixel 400 and above pixel 2600 the lines have shoulders. Depending on how critical the line shapes (and centroids) are to your program, you may want to trim the spectra accordingly. Also, one can't cover a wavelength range greater than a factor of two in first order and a factor of 1 1/2 in second order without order overlap. The wavelength range will depend on the order separation filters available in some cases.