Gosh!

[Tom Droege <tdroege2@earthlink.net>]

Climbing up onto SoapBox:

Gosh everyone!  These chips are much better than I paid for.  The best are 
Class 3 (1 is best), but I only paid Class 4 prices.  This means that they 
have defects.  One should not expect flat fielding to cure real 
defects.  You have to cut them out of the image.  I assume that Michael has 
been doing this for his work.  What say Michael?  I don't see a streak in 
the Photometric Residuals plot.  I would expect one if he did not cut out 
this bad area.  On the other hand, I don't see a line with no data, so 
possibly he did not cut out the defect.  It would be hard to see at the 
scale of the plot.

I think that you have to map the whole chip if you are going to do good 
work.  Besides this dark streak that Andrew has found there is also a 
bright streak that is easy to see.  There will also be bad areas and bad 
pixels.  One just has to map them, I think.  Some will be of a type that 
flat fielding can fix.  One can imagine a small mask defect that affects 
the gain.  There will be hot pixels that turn on at some temperature.  So 
one has to pay attention to the chip temperature when creating the map.

Possibly Arne, who has been there when chips were much worst than these, 
can give us a tutorial on how to do it and what data sets need to be taken.

These are small problems on the scale of things.  For this I chip we might 
lose 40,000 out of 4,000,000 pixels.  This is only 1% of the chip, and 
takes into account expected edge effects.  For this we get a 95% reduction 
in price.  A bargain, I think.

Down from the SoapBox.

Tom Droege