|
Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://xmm.vilspa.esa.es/user/socpub/ccd_review/node13.html
Дата изменения: Fri Aug 30 15:54:50 1996 Дата индексирования: Tue Oct 2 12:14:07 2012 Кодировка: |
The full frame mode of operation is widely used for optical astronomy in order to obtain the greatest focal plane coverage. In X-ray astronomy this mode has a disadvantage in that for the relatively shorter image integrations required to maintain a photon counting mode, smearing of the images occurs during readout when the photons continue to fall on the moving charge pattern. The use of mechanical shutters to circumvent this problem would add a significant dead time to short exposures, and in any case their use in space is generally precluded on the grounds of cost and technical risk. In this mode with an image:readout ratio of, say 10:1, a trade-off between readout rates, noise and count rate capability is required.
The framestore mode may utilize permanently shielded store sections,
from which the X-ray data can be read out without smearing. With a frame shift time
of 
10msecs, any smearing which may occur while the data is shifted to the
store section may be at the level of 
1%. However this mode is subject to
uneven dark current generated when charge packets occupy varying amounts of time in
the store section.
Up to 50% of the chip area is occupied by the store section, which limits the focal plane coverage attainable. Complex packaging has been considered to use the image sections of one layer of chips to cover the stores of a lower layer. This is facilitated by the curved nature of X-ray focal planes, but the problems expected in qualifying multi-layer packages for flight have prevented the adoption of such a scheme to date.
It should be noted that, unlike previous generations of X-ray detectors, the format of CCD readout is not fixed, but may be altered at will by changing the sequence of clocks supplied the CCD. For example a windowing mode can be obtained by discarding unwanted pixels by rapid clocking without sampling, and only reading out pre-selected pixels. This allows a faster access to the desired pixels (higher time resolution) and higher count rate capability for the limited case of a small region of the image.
Pixels may also be added together by inhibiting row or pixel readouts until the required numbers of rows or pixels are collected in the output register or node. This allows some means of trading off resolution against the fraction of split events obtained. This flexibility may be important for coarse spatial resolution application such as XMM.
A final example of user defined formats is the timing mode. For point objects such as binary sources, where imaging information can be traded against improved time response, the CCD can be read out continually. The apparent pixel address now relates to time information rather than spatial information, and time resolution is determined by the pseudo-pixel cycle times and the point response of the optics [Lumb].