Data reduction, page II

The complete series of recordings

The measurement of the spectrum of an object (actually the wavelength range corresponding to a defined grid position) consists normally of a series of recording series. Below is the entire procedures principles explained. In many cases you can save yourself some of the accompanying photographs (eg, when a library of scalable darkfield photographs and a Masterflat exists).

Biases: There are as many as possible (minimum 10) biases included. So short-time exposures <1 sec without light infall (the camera shutter is not open. To make sure that no light penetrates through the telescope, you can also cover the aperture of the telescope). These biases are then averaged for the data reduction procedure. The biases are made at the same CCD temperature as the target frames.

Darks: When the desired temperature of the CCD chip is reached many darks are shot (dark frames). They are carried out with closed shutter, and the same exposure time as in the object images. They will also later be averaged (to the exclusion of recordings with cosmic - or after reducing cosmics by the software).

Flats: The inclusion of good flats through the telescope and the spectrograph is a difficult undertaking. In my experience flats must be perfect or we should leave it. Poor flats worsen the result (compared with data reductions without regard to the flats).
"T-shirt-flats" are obtained as follows: one covers the opening of the telescope with a white cloth (no wrinkles!). And illuminates this "diffuser" with a ~150W halogen lamp (which is a bulb that has no lines in its spectrum (purely thermal radiator). Fluorescent tubes, energy saving light bulbs and gas discharge lamps are not suitable because they make no continuous and flat spectrum. Expose for several minutes close to the saturation of pixels (~80%). But beware: There may not be truly saturated pixels! All pixels must be within the linear range of their sensitivity. One should take many of these flats (later they will be averaged so that the flats introduce no relevant noise into the result). The flats show in general brightness gradients and dust effects. The flats are later freed from the bias and dark current. So sufficient darks must be made with the exposure time of the flats to correct later the flats (preparation of a Master Flats).
This correction images (biases, flats, darks) can be used quite some time. You need not be recorded each observation night. In particular, the flats and darks require considerable time. We shall, therefore, have a library that is constantly supplemented with recent recordings. One can also monitor the constancy of the camera and the apparatus. But the flats/darks of the data base may only be used when the CCD camera is not changed or anything has been changed on the apparatus and the same CCD temperature was used!

Caution: If you do not have full frames but receives image cutouts (clips), it must match the cutouts of all the images (object photographs, flats, darks ....)!

The images of the object are exposed as long as possible. It is better to make a few recordings in which the pixels are just not yet saturated (50 to 80%), then many short-term exposures. Large series of short-term exposures smuggle additional readout noise in the result. Long exposure times, however, require a stable and reliable tracking of equipment (autoguiding). With my ST-4, I have already 5 h tracked continuously for a single star on the slit. This is pure reliability.

Also important is keeping a diary. In this enter the weather conditions, assessment of seeings, the apparatus used (telescope, camera, spectrograph, slit width, grating ...), the boundary conditions (CCD temperature), the observed objects and the images (exposure times, darks, biases ... .) and the hours (Universal Time). I write a few phrases in a text editor and save this log file with the recordings. So I can always recapitulate what I have done what night.


Recording software (data acquisition software)

In the most cases I use Maxim DL. This program has the advantage that I can control the science camera (Sigma 1603ME), and the autoguider camera (St-4) on the same screen and watch. Of course you can also use the camera manufacturer's software, including the free IRIS or Astroart and many other acquisition programs. The images should generally be stored as .fit file. This format has an appropriate header, where the important data (object, camera, shutter speed, time, location ....) are stored. It is the scientific default format for spectra which is used worldwide. Of course, the header data at the beginning of a series of recordings to be updated.

next page

back to content