Some known CCD3-controller features
- Oscillation in x-direction seen especially in bias frames.
Left:The vertical line (x=50) indicates the transition between the prescan and
imaging area.
Right: Overscan subtracted and trimmed masterbias frame
- Vertical strips seen in xbinned images:
Left:gray scale image of a bin2x2 ALFOSC frame,
Right: horizontal cut of bin2x2 ALFOSC frame, with different
read speeds.
The vertical strips seen in xbinned images
with ALFOSC is due to the lose of synchronisation of the controller power
supply with the pixel readout timing. When you do binning in X the pixel
time increases so giving a none integer period for the PSU to sync. For xbin
2 reducing the 'tsam' from 172 clocks to 136 clock returns the
synchronisation, and no strips are seen. The shorter sample time does
increase the noise by ~1ADN (0.3 e) compared with no binning, but
significantly reduced the xbin 2 noise from ~25 ADU.
So if xbin 2 is used the 'tsam' should also be reduced to 136 clocks
(readspeed 200).
- Over-exposing in binned mode
Basically what is happening is, due to binning and a probably saturated
image, the summed signal is giving a very large signal to the
pre-amplifier of the controller causing it to saturate. This is resulting
in the reset level been larger or equal to the actual signal, so when the
reset voltage is subtracted from the signal voltage the result is a
negative or zero (black) level in the image.
Normally for non-binned integration this should not be a problem since a
single pixel full well will not cause the pre-amplifier to saturate. To
prevent it don't bin with very bright sources in the field.
This feature is not confined to the new Generation 3 CCD controller, it
has also been reported with the SDSU controller and is seen in images from
the old ALFOSC controller.
|
