The very first thing you should do now is to locate the limits of your slitlets because this information is needed for all further commands. Therefore you type
Midas
...> LOCATE/MOS
This should produce the table
.tbl (mos) with the columns
:SLIT | sequential number of slitlet |
:YSTART | first row of slitlet (world coordinates) |
:YEND | last row of slitlet (world coordinates) |
:XOFFSET | offset of slitlet from center of CCD |
and write the total number of slitlets to (0). It may be that the threshold defined by (0) is either too low (e.g. below bias value) or too high for your data. Also, the width ( ) (0) may be chosen to high or too small. If you detect too many slitlets, where only noise is visible, you should increase ( (0)). You also can change the scan position and width ( ).
Then you should have a look at your flat-field and try again with
Midas
...> SET/MOS flatlim=threshold,width,limit
Midas ...> SET/MOS scan_pos=xpos xbin=width
Midas ...> LOCATE/MOS
You may also try to identify the slitlets interactively with
Midas
...> LOAD {flat}
Midas ...> DEFINE/SLIT init
Midas ...> DEFINE/SLIT add ##
where ## stands for the number of slitlets you want to identify.
You will have to determine the offsets between the slitlets with OFFSET/MOS (see below).
If you do not have FORS data the column :xoffset will be set to zero. This is due to the fact that for FORS data the slitlet positions given in the header of the frame are transformed to offsets from the center of the CCD. This transformations is obviously not valid for other instruments. As you will need the offsets for the wavelength calibration frame you can determine the offsets relative to the first one (which is not necessarily identical with the center of the CCD) with
Midas
...> OFFSET/MOS
As this command does a line search in the wavelength calibration frame
.bdf (wlc) and correlates only the detected arc lines
the resulting offsets may be wrong if there are not
enough lines to get unambiguous correlation results.
Setting the parameter
(200,5) to the values successfully
used for
SEARCH/MOS will help to yield reasonable results.
Normally spectroscopic flat fields show the spectral signature of the lamp with which they were taken. You can take out this spectral intensity distribution with
Midas
...> NORMALIZE/MOS
By default this command will fit a polynomial
of
(3)
order to the averaged
(along the slitlet) spectral intensity of the flat field (
(flat)) (separately for each slitlet) and divide it
by these fits. The results are stored in the frame
(normflat).
Alternatively, you can normalize the flat field by dividing through
an average smoothed with a median filter.
To perform the actual flatfield correction together with the
normalization type
Midas
...> FLAT/MOS
This command will do the normalization and divide the frame
.bdf by
.bdf.
If you have not given any name for the result frame
it will derive the name of the
flat field corrected object frame by adding an `F'
to the name of the input
frame (e.g. Ffors0001).