synopsis: (1) statistical uncertainty reduction by removing outliers (2) (positive) systematic uncertainty reduction due to removal of positive glitches.
limitations and applicability:
A minimum number of 25 read-outs per ramp are required for application
of the algorithm. Glitches occuring in measurements with less than 25
read-outs per ramp should be removed during signal deglitching
(sect. 3.8).
description:
Replacement of an anomalous increase in voltage between two adjacent
read-outs by the mean increase in the ramp. This is done in an iterative
way using a kappa-sigma criterion. The method assumes that an
integration ramp is linear.
purpose correction:
Glitches cause a voltage increase between adjacent read-outs which is
higher than expected from a constant photocurrent. Such glitches can
be isolated and corrected. Removal of these glitches lowers the scatter
on the individual signals and removes a positive bias from the signal.
uncertainty/noise introduced:
The glitch spectrum is such that the glitches with the
lowest energies have the highest frequency of occurence. The low
energy glitches cause a small steady voltage increase which cannot
be removed by this method. This causes a small positive signal bias,
which is considered to be part of a systematic increase in dark signal
(see sect 3.7).
auxillary data:
Default deglitching parameters:
| Nit | = | 4 | number of iterations |
|
= | 4.5 | minimum number of standard deviations for glitch detection |
| Nmin | = | 25 | minimum number of read-outs/ramp for application of algorithm |
keyword(s):
| RDITER: | number of iterations | SPD header |
| RDFSIG: | number of standard deviations for clipping | SPD header |
| RDMINP: | min. number of read-outs per ramp | SPD header |
| RAMPDEGL: | number of ramps deglitched | SPD header |