Instrument Calibration Data

 

Once the CCD has been set up to obtain images of the correct format, the observer can begin taking calibration exposures. Normally the following calibration exposures are required:

• A number of ``Bias'' frames, which are dark exposures with zero exposure time. These can be taken with the BIAS command. These are often not required because the bias is structureless, and its absolute level can be determined from the overscan and underscan regions associated with each exposure (see Table 5). It is still good practice, however, to take a few BIAS frames at the beginning and/or end of each night.
• A single long dark exposure, which can also be taken with the DARK command. The dark count is very small, and a dark frame is usually used to measure the cosmic ray rate.
• Flatfields in each passband to be used. Flatfields can be obtained off the inside of the dome with the mirror covers open using a filament lamp which is kept in the WHT building (never use the fluorescent lamps, which have strong lines in them which can cause fringing with thinned chips), or off the twilight sky between about 10 and 40 minutes after sunset. U band flatfields, and flatfields with interference filters can be obtained at the bright part of this period, and other broad bands (in the order B,V,I,R) in the latter part. It is a good idea to shift the telescope by 20 arcsec-30 arcsec between exposures so that stars may be removed later. As a rule of thumb, try to get 30,000-40,000 counts per pixel, so as to have a high S/N flatfield. ``Dark Sky Flats" can be created by the ``shift and add" technique. Here the telescope is shifted (`dithered') in a raster pattern between program exposures and the Dark Sky Flat is obtained by taking the median of the stack of frames (each image is rescaled so that the sky level in each is the same); see article by D. Carter in Gemini 39. Dark Sky Flats are also very useful for removing any fringes (which are due to interference of night-sky lines, and commonly seen in R and I exposures).
• Shutter Corrections. The CCD shutter takes a finite amount of time to open and close, which results in a difference between the nominal (ie. as recorded in the FITS header) and real exposure times. Furthermore, this difference can vary with position on the chip, depending on the shutter design. The shutter correction (i.e. the length of time it takes to open *and* close the shutter) can be determined by taking several (dome) flatfield frames of equal illumination but with different exposure times; in practice it is best to take a series of short exposure frames and one longer exposure frame. We have found that the shutter correction depends on Zenith Distance, and only slightly on position within the CCD frame. Here are some preliminary results:

  20 ZD: S.C.= 0.005  -  0.02 sec

  45 ZD: S.C.= 0.055  - 0.065 sec

  90 ZD: S.C.= 0.08  - 0.09 sec

  However, these numbers are based on only a small number of frames, and users are encouraged to take their own data (the dome flats can be done in the afternoon and are usually taken anyway), especially if they need high-precision photometry. In any case, exposure times of less than 2 seconds should be avoided (for a shutter correction of 0.05 sec, a 2 second exposure is  0.05 mag fainter than would be calculated from the nominal exposure time!).