Detector Characterisation Tests
-------------------------------

A list of tests that ideally should be accomplished in the laboratory
before shipping instrument to Chile.


Readout Noise

  robust measures of rms readout noise for each channel

Gain 

  gain in electrons per ADU for each channel, preferably in more than 
  one location

Full Well Capacity

  saturation limit for each channel for large scale illumination and for
  "point-like" sources

Systematic Pattern Noise

  quantify level of non-random noise for each channel (eg. pickup noise,
  odd-even row/column noise)

Histograms

  verify ADC quantisation only occurs at +/-1/2 ADU level for each channel
  by analysing illuminated frames with large range of levels

Non-linearity

  use a sequence of constantly illuminated flats to assess linearity of
  each channel by doing exposure time sequence, possibly also at individual
  pixel level

Dark Frames

  measure dark current and assess stability of dark frames as a function
  of no. of coadds, and exposure time.  Use dark frame sequence to assess
  spurion rate and hot pixel frequency.  How do dark frame properties 
  depend on previous frame(s) illumination level(s).  

Flat Frames

  use a temporal sequence to quantify no. and location of bad pixels
  and QE defect regions.  Repeat at intervals to assess bad pixel and QE 
  defect stability, particularly after thermal cycling.  Stability of
  relative pixel QE over whole detector.

Reset Anomaly

  assess stability of and quantify level of reset anomaly in dark frames
  and in dark-corrected flats

Cross-talk

  measure cross-talk between channels and possibly between detectors.
  Requires some sort of point-like source to be projected on the channnels
  or if present hot-enough pixels.  Are crosstalk properties stable.

Image persistance

  temporal and spatial behaviour of residuals from bright point-like sources
  dependency on no. of resets, reads, time since last exposure