Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.mrao.cam.ac.uk/~jsy1001/dimmwit/calibration.txt Дата изменения: Fri May 7 19:44:46 2004 Дата индексирования: Tue Oct 2 06:10:40 2012 Кодировка: Поисковые слова: п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п р п р п р п

Why flatfielding has to be done 1by1 binning:

I have modified finder so that the X and Y offsets of subframes are
saved in FITS files taken in slow 2D readout mode (they're saved in both
fast modes too but there the values are always zero). The units are
unbinned pixels from the origin.

This allows you to do flat fielding on subframes if you want to, but
there is a subtle point to note: Suppose you have chosen 32x32 binning
and then select a subframe. The top left corner of the subframe starts
exactly where you told it to (to one unbinned pixel resolution), NOT
some multiple of 32 pixels from the top left corner of the field. So it
is inappropriate to take a full screen flat field at 32x32 because the
binned pixels probably don't match the ones in your subframe (they will
each group together different sets of pixels). When you are flat
fielding you should take a full screen with 1x1 binning and work out
32x32 chunks of it (which you can now do given the header file mods).


Assumptions when taking dark frames:

We want to be able to assume that the dark frames are uniform and only depend on
the readout noise and the dc offset applied to the camera head. If this is true
then a subframe of any size and position can be used as a dark frame. We then
take dark subframes in each binning mode (to check that binning isn't making
a difference) with the exposure time set to 1ms (to ensure that no
stray light is recorded). To average out the readout noise we want to record
50-100 frames. Also the dc offset could change from night to night so dark
frames should be recorded every night of observation, once the CCD has been
allowed to cool, to check that the offset hasn't changed.

Testing uniformity of dark frames (needs to be done only once):
Each camera needs to be tested to ensure that a dark frame is indeed uniform.
To do this, we need

50 full frame slow mode with exposure time 1ms and 1by1 binning
50 full frame slow mode with exposure time 1ms and 2by2 binning
50 full frame slow mode with exposure time 1ms and 4by4 binning
1000 frames fast 2d mode
1000 frames fast 1d mode.

Plotting a histogram for each of these 5 files will prove the uniformity of the
dark frames. This must be completed with new and old drivers.

What to do to calibrate your measurements:

1. Dark frames with old and new drivers in place (while observing):
1000 frames in t0 mode
1000 frames in d0 mode
400 frames in r0 mode - in three binning modes(1,2,4) and exposure
time 1ms, subframe of any size will do (position doesn't matter either).

2. Flatfielding (only needs to be done once with old drivers):
For r0 mode the binning should be set to 1by1 and the full frame read out.
The flux bar should read about 50% so choose an exposure time that gives
this. The t0 and d0 modes are more sensitive to light because the entire
ccd gets read out so light level used for r0 mode needs to be dimmed to
stop saturation in these flat fields. Take 50 frames in each mode.

3. Pixel Scale (while observing):
Focus telescope on bright binary or multiple star and take over 100 images.
Could do this several times a night to check the focal length hasn't changed.
Binary separation should be between 10 and 30 arcseconds.

4. Measure Sky Background:
Point telescope at dark patch of sky and record 50 1sec exposure files in r0
mode (slow 2D) with 1by1 binning and any size/position subframe. This is to
check that the background is insignificant.