Calibrations
Flatfields and wavelength calibrations are part of almost every program's calibration set. These are discussed below. Darks, fixed pattern, and bias calibrations are less commonly made but are also touched on here.
The Hamilton is quite stable and it is not usually necessary to take flats or wavelength calibrations between targets; many observers make them at the beginning and end of the night.
The flat-field source is a continuum quartz lamp mounted above the entrance slit. It is turned on by selecting "Polar Quartz" (selection 8) from the Hamilton controller's "Lamps" submenu. (Note that if left alone, the lamp times out after about thirty minutes.)
Flats are made using a decker or aperture plate at least twice as tall as that used for observing (typically 5 or 6 arcseconds). This produces flat-field orders wider than the stellar orders, preventing edge effects when performing the flat-field division. An additional continuum exposure, through the same decker or aperture used for observing, is also usually made as a template for tracing the curvature of the orders.
Because of the Hamilton's very large wavelength coverage, combined with the lamp's temperature and detector's response, uniform illumination of all the orders in a single exposure is impossible. Depending on one's wavelength range and program requirements, various filters are used to temper the lamp, but several exposures, through several filters, are usually necessary to achieve adequate signal at all wavelengths. These are later coadded to create a master flat-field. (The most common formula for broad band coverage is a combination of bg12 and bg13 filters at a ratio of about 10-secs of the former to 1-sec of the latter.)
A Thorium-Argon hollow-cathode lamp, mounted near the entrance slit, is used for wavelength calibration. It is turned on by selecting "Th-Ar" (selection 5) from the Hamilton controller's "Lamps" submenu. (Like the quartz lamp, the Th-Ar times out after about thirty minutes.)
The Th-Ar lamp provides numerous lines in every order. Unfortunately, several very bright Argon features at the red end of the spectrum invariably saturate and bleed into adjacent orders. Some observers take two arcs--one unfiltered, and a second, longer one, using the BG38 filter to cut down the Argon lines.
Dark current and fixed pattern are very low for the CCD's used with the Hamilton. However, for very low signal-to-noise observations, you may wish to take these into account.
The data-taking program makes provision for dark exposures by providing a "dark" option (selection F) which inhibits the shutter. Dark exposures need not be as long as the stellar exposures to which they will be applied, but should be a substantial fraction, as darks may not scale linearly.
Fixed pattern results from the influence of the 60-cycle line frequency. Lick controllers, when operated in the "slow readout" mode, are synchonized to the line frequency, so that, rather than appearing as noise, it shows up as a low-level "fixed pattern", evident as vertical bars in a one-second dark exposure. The bars always appear at the same locations, and, using a one-second dark, can be removed from the data.
Note that in "fast readout" mode, the readout is not synched to the line frequency. The latter will instead appear as low-level randonm noise in the images. Nevertheless, some observers use "fast readout" for all their observations and calibrations, particularly if the program calls for numerous short exposures which would entail a substantial price in duty cycle due to the longer readout time in the slow mode (twice as long--or about 4.5-minutes for the full chip--as compared with fast).
Bias frames are not usually taken with Lick CCD's as the raw data are already bias-subtracted, with the bias levels stored in the highest number column. The data-taking program does, however, allow the option of disabling bias subtraction.
Finally, please note that all Lick CCD images are given a "fat-zero" to avoid negative-going pixels. This appears as an offset of 64 DN in the raw data. (New images, when initially displayed by the data-taker show this offset. However, images redisplayed using the data-taker or Vista will have their offsets subtracted.)