The reset anomaly has been described numerous times at other conferences [1,2]. The chief component of the reset anomaly is a large intensity ramp at one edge of each quadrant. The edge where the reset anomaly is largest is also where the array readout electronics are located. A typical Correlated Double Sample (CDS) bias frame from one of our detectors is given in Figure 3, showing the typical profile from the edge inwards. The ramp drops from approximately 1000 ADU from the edge to 0 ADU near the centre. This is typical for most devices; however we have also seen devices where the maximum value of the profile is at least a factor four to five improved. More of a concern is the instability of this profile across the detectors for consecutive reads. In standard observing we subtract a dark and bias frame flattening the profile. However we find that when we subtract consecutive CDS images the profile does not flatten out, but is unstable. Figure 4 shows the images and profiles for three consecutive CDS frames, subtracted from each other as described.
Solutions for this have been reported elsewhere . These usually consist of running read or reset frames that have the same execution time as real reads and are then be dumped. However, for our wide field imager we typically only do simple correlated double samples, an extra frame time would reduce the readout efficiency by 33%. Our solution is, therefore, to perform extremely fast reads of all the pixels in the array using standard read routines. However we do not settle or digitize for each pixel. This typically takes approximately 0.1 s to complete. We do this twice only after the resetting of the array, which adds 0.2 s to our CDS time of 1.4 s. This process is therefore more than twice as efficient as other methods for reducing the reset anomaly.
Was this article helpful?