[Aavso-photometry] CCD testing - charge/image shift?

[Tom Krajci]

I've dome some more analysis of flats in my CCD test series from 10 July. 

When I divide one flat by another...and both flats are in the linear 
range of the CCD.  No surprises here - random/patternless result.

But when I divide a linear flat by a non-linear flat...I get strange 
results.  The resulting entire image is clearly not featureless, and the 
dust shadows stand out very clearly (as well as large scale structure).

But I noticed something else about the dust shadows.  In the case of the 
non-linear flat...the image seems to have shifted vertically by about 
three pixels.

Did the CCD/light source move during this test run (all night, almost 
1000 images)?  No.  I examined other flats after this...linear ones, 
non-linear ones.  No sign of physical movement of any parts of the test 
setup.

If this 'vertical image shift' takes place for pixels above the linear 
range (three pixels shift!)...that could complicate the calibration of 
images for precise photometry.

Has anyone seen this before?

PS.  I'll be happy to send my screen shot/graphic of this effect to 
anyone that's interested, just ask.

Thanks in advance,
Tom Krajci

-- 
-------------------------------------------
Tom Krajci
Cloudcroft, New Mexico
http://picasaweb.google.com/tom.krajci

Center for Backyard Astrophysics (CBA)
http://cbastro.org/ CBA New Mexico

American Association of Variable Star
Observers (AAVSO): KTC http://www.aavso.org/
-------------------------------------------