|
Last updated: 31 Aug 2009
A comparison of STA1 versus T2KA can be found here.
All of the data used in the analysis shown below were taken with
the CCD operating at 152K.
The Quantum Efficiency curves of STA1, T2KA and T2KC are shown above
for comparison. The STA1 data come from UA-ITL.
The plot above shows fringing near 7500 Å
The plot above shows the Dark current (expressed as DN) averaged
over 3 columns for a 1800 second dark. The binning was 2x2.
The histogram above shows the numbers of pixels as a function of dark
level for an 1800 second dark with noo binning to highlight the hot
pixels. The hand drawn line is meant to approximate the distribution
of normal pixels. All of the pixels above the x-intercept can be
considered as Hot Pixels. For 1x1 binning the percentage of Hot
Pixels is 0.03%.
The plot above shows that the "hottest" of the hot pixels probably
saturate, meaning that for a small fraction of Hot Pixels, dividing
a 3600 sec dark by 2 to represent an 1800 sec dark gives too low a
level for a few score pixels. Optimal extraction should minimize this
effect on science data.
The histogram above shows how we estimated the fraction of hot pixels
for 2x2 binning. The fraction is 0.10%.
The histogram above shows how we estimated the fraction of hot pixels
for 4x3 binning. The fraction is 0.25%.
The column plot above shows the "injected charge" apparent in a mean
bias image (Zero). The vertical scale is electrons. Whlle this
ramp-like effect is removed by subtracting the Zero image during
processing, the noise introduced is shown below.
This plot show column plots averaged over 40 columns, the lower
in the overscan region, and the upper points in the imaging area,
demonstrating the injected change relative to the base bias level.
This plot shows the rms noise in the imaging area of several
overscan-corrected mean bias frames, as a function of row number.
The effective read noise changes with binning factor and position
on the device due to the noise contribution of the "injected charge"
present.
This shows the decay of the bias level with time when the CCD has just
been powered on after having been powered off for long and short time
intervals. It is advisable to run a series of bias frames immediately
after poering the device on to help speed up the decay. Serious data
taking can resume after about a 30 minute wait.
|
