[Aavso-photometry] Question concerning linearity of CCD chip

[arne]

actuary at mchsi.com wrote:
> To All:
> 
> I am reading "Lightcurve Photometry and Analysis" by Brian D. Warner.  On page 
> 91, a procedure for testing the response linearity of a CCD chip is 
> described.  I am confused as to the meaning of the sentence "Find the average 
> of the maximum ADU values for each star at a given exposure...".  Say 5 
> imagess at 1, 2, 4, 8, and 16 second exposuress are taken. For a specific star 
> there are 5 images at 1 a second exposure.  Takling the maximum ADU for the 5 
> images for the specific star provides a singlr value.  However, the procedure 
> says to find the average of the maximum ADU values ... The average of the 
> single value is the same value. Since the test involves plotting the values 
> obtained through the application of the "averaging.. maxumum" algorithm by 
> exposure time, one does not average over exposure time nor over all stars.  
> Clearly I am misinterpreting the description of the procedure.
> 
> I would appreciate some guidance as to the correct interpretation of the 
> procedure.
> 
This must be from the second edition, as the page numbers and text
do not agree with my original copy.  I'll try to interpret the statements
above, with the understanding that they are taken out of context and
I'd need the correct edition to fully consider the problem, and I
certainly won't approve any linarity method without seeing its entire
description.

I think what is meant is that each star profile has a range of pixel
intensities - highest in the center and then falling off like a Gaussian.
For the 5 frames with the same exposure, you would pick the maximum
pixel value found for a given star - that gives you 5 "maxima".  Then
average those.  This is basically what Wolfgang suggested.  If true,
this is a very crude method of determining linearity, sufficient to
look for large deviations (such as saturation or ABG turnover) but
not much more.

Arne