Return to WHIRC Home Page
Several in-house reports were written during the commissioning process. These
cover aspects of the instrument performance in more detail than presented
on this webpage.
The mean-variance (photon transfer) measurement is used to
determine both the read noise and gain (e/ADU) of the detector.
This analysis assumes the only sources of
noise to be the signal-independent read noise and the shot noise from the
signal itself. When the variance of the
signal is plotted against the mean value on a log-log plot, the result is a
line of unity slope in the photon-noise-limited regime which intercepts the
X-axis at a value equal to the gain. The
Y-intercept of the actual data is the variance of the read noise.
The data from September 2008 and more recent data from November 2009
were analyzed using the sum/difference technique described in Schubnell
et al. (2006, SPIE, 6276), in which the variance of the difference
of two images is plotted against mean of the sum. For both of
the data sets, this yielded a conversion gain significantly smaller
than that determined using a temporal variance on a pixel-by-pixel
basis. Analysis of model data with known read noise and gain using the
sum/difference technique returned the correct values, whereas the
temporal analysis returned gain values about 10% larger, so we have
accepted the 3.3 e/ADU gain obtained by the sum/difference analysis.
The complete report (WHIRC Report VI) may be downloaded from the
29 October 2010