Next: 5 Data Reductions
Previous: 4 Observing Philosophy

4.1 Observing Procedures

Once the detector has been aligned and focused (§3.5.2; §3.6.2), the required calibration exposures should be made. An initial set of calibrations can be made just before dark, and additional frames secured at convenient times during the night. It is important when trying to achieve high signal to noise ratios to secure flat fields at least three times during the night as the chip position does change slowly. The data can then be reduced with the appropriate set of flat-fields according to the time of exposure. Be sure to repeat all of the calibrations each night as small changes of dewar position may occur when the dewar is filled with LN2. If accurate radial velocity measurements are being made, comparison exposures should be obtained at least once an hour to map small pixel shifts throughout the night.

Before the calibration exposures are made the observer must decide whether to use on-chip summing or not (§3.2). On-chip summing will be more advantageous on camera 6 than on camera 5. This is due to the larger spread of light perpendicular to the dispersion on camera 6. If on-chip summing is selected, all calibration exposures must be made in that mode as well as the object exposures.

Comments are given below on the various types of pictures:

zeros

A DC offset is added to the overall signal from the CCD to insure no negative data. This ``pedestal'' level (a few hundred ADUs), formerly known as ``bias'' has no particular significance other than it should be stable. Obtain 10 - 20 per night. They can be spaced throughout the night or split between beginning and end of night. Zero exposures are important for the TI CCDs where preflash is generally used, but may not be necessary for some of the TEK CCDs where the bias level is very uniform. In this case one could use the overscan or the region on each side of the spectrum to subtract the bias level.

darks

If relatively short exposures are planned (1 hour or less), there is probably no need to obtain dark exposures. Obtain a couple 10 or 20 minute darks just to check on dark current and look for any hot pixels. If a hot pixel is seen, it may be possible to move the spectrum vertically to higher or lower column numbers to avoid it.

flats

Obtain a total of at least 10 times (and preferably more) the number of ADU's compared to the count rate of your most intense spectrum, using a decker large enough to be wider than object spectra. To avoid possible non-linearities, no spectrum (object or FLAT) should exceed 8000 ADUs with a TI CCD. See §3.2 for exceptions. For example, if your spectrum has an intensity of 5000 ADUs, obtain at least 10 FLATS with 5000 ADUs each. These exposures may be split between the beginning and end of the night or grating setting. If on-chip summing is in use, the spectrum intensity may exceed the above limits, but since the amount of summation varies, caution is advised.

comps

Obtain at least one per data set. More may be useful if accurate radial velocity measurements are to be made. Wavelength shifts on the order of a few tenths of a pixel have been measured over a period of hours.

observe

Main observing word. Again, be careful not to exceed ADU limits discussed above per pixel unless on-chip summing is in use. Any of the above data types can also be obtained using this command.

Carefully filling out the logbook on the logsheets provided will make subseqent data reductions easier. Please leave the originals in the notebook; copies can be made in the administration building for taking home.