Examples of Dark frames are given in Figure for the SW array and for the LW array.
Measuring a dark current frame is a time consuming operation, because of the long stabilization time it requires for the detector: going from a previous flux level, a priori unknown, to a dark level, requires 200 stabilization exposures. This can be shortened, if the previous flux was known and low, as shown in Figures to .
The SW array has an automatic procedure (FAST-RAZ) which efficiently shortens the stabilization time required to about one minute when switching from light to dark, by doing many exposures and readouts with the shortest integration time. It is automatically set up every time a dark frame is measured.
A dark frame measurement will be done at the beginning of each orbit, with an accuracy of 0.5 ADU rms (SW) and 0.3 ADU rms (LW), this is likely to be the best dark frame to be used for all the observations of this orbit.
The time penalty for a dark frame measurement being high, the use of library data will always be recommended. The decision to perform a dark measurement for a given observation should be supported by strong accuracy arguments. In the following cases, the dark correction will have negligible effect on the accuracy of the observations: