EFOSC2  ESO Faint Object Spectrograph and Camera        HOME INDEX SEARCH HELP NEWS

The four EFOSC2 imaging modes (Simple, Jittered , Coronographic and Polarimetric) require the following calibration frames :

Biases and Darks   are common to all EFOSC2 observing modes and are described on a separate page
Flat fields
Mask images (Only for  Coronographic and Polarimetric  modes and dealt with under those subsections)

Flat-fields

• types of flat fields
• flat field intensity level

Flat-fields come in several flavours :

• Internal flats
• Dome flats
• Twilight sky flats
• Super flats

• highly crowded science frames
• un-jittered images of a single field
• when targets in different frames are positioned on the same area of the CCD (this is often the case in polarimetric/coronographic imaging)
• bands in which fringing is significant (R, i and z  bands)
• bands in which the sky background is low (very blue bands, e.g. U band, and narrow bands)
• only a few science images are available for constructing a super flat.

Twilight sky flats provide the next best alternatives. However they do not always provide very good flat-fielding. In the case of EFOSC 2 we have noticed that morning and evening sky flats are consistently different. The difference is at a level of 2-3% peak-to-peak  across the CCD and so one should factor in similar photometric errors (due to flat-fielding - there may be other sources of errors). However this variation is smooth and in principle can be eliminated using observations of a standard star on many regions of the CCD.

With the red filters (Bessel R, Gunn-i, Gunn-z) CCD #40 shows considerable fringing. In these cases it is essential that twilight sky flats are taken since the science images will have to be used to constuct a fringe frame.

Dome flats are the insurance frames, to be used when all else failed. We do not expect flat-fielding to better than 5% and occasionally the variation is not smooth. Internal flats are quite useless for flat-fielding - the problem is that its illumination pattern is considerably different from that of external external photons.

However, it may be noted that CCD observations often have several other sources of photometric errors (including central light concentration in focal reducing instruments like EFOSC 2) and one should not expect absolute photometric errors of 5-10% unless one is prepared to put in a lot of observing/analysis  time and effort into reducing it.

Flat Frame Intensity Level :  a good rule of thumb is that the photon noise in each pixel should be considerably lower than the above errors (10000 counts => 1% photon noise). So 3-5 frames with a total of 100,000 ADU will suffice in all cases. In any case, taking more than 5 frames of 40,000 each is unnecessary. This is valid for all flat fields. Note : one needs flat fields for each combination of  binning  ,  readout speed  and  filter

back to top

Bias and Darks

Imaging Dome/Sky Flat Fields

Bias and Darks

Coronographic Flat Fields
While this is similar to simple imaging with the addition of a coronographic mask, in practice we use the spectroscopic dome flat template ( EFOSC_spec_cal_Flats ) to obtain flat fields. This is because we make use of the slit and grism wheels as well (in addition to the filter wheel). The coronograph goes into the aperure wheel while the Lyot stop goes into the grism wheel - choose the elements accordingly while making the OBs in jP2PP.

The Lyot stop is very efficient in avoiding scattered light from very bright stars. The installation and alignment of the element will be done by the telescope support staff either on the set-up night prior to a coronographic run or as the first exercise during the run.

Mask Position
Note the centre of the coronographic mask on the flat field images. Use this pixel value in the  Move-To-Pixel acquisition template as the position to which the masked star has to be moved.

back to top

Bias and Darks

Polarimetric Dome/Sky Flat Fields

Apart from a few additional features, Polarimetric flat fields are similar to Simple Imaging flat fields - observers should read those pages first before following the link in this section.

Image of the Wollaston Mask

An image of the mask is required to identify the reference pixel for use in the target acquisition image.

back to top

Send comments to : ls-spectro Last modified: