[Aavso-photometry] Dark Skies / Thin Air

[arne]

ben at davies.net wrote:
> Arne, thank  you,
> 
> I'm very relieved to hear that the negative values are not a problem.
> 
> The pixel fov was oversampled at .46 arcseconds.  I didn't get around 
> to using any binning this time, but was defocused.
> 
> The weather was in rare fine form.  Winds that had been off the desert 
> died at sunset to a dead calm.  The temp didn't get below 45 degrees at 
> night.  Can you believe it? no moon, no wind at the peak and warm, and 
> all this in November yet.  The skies were just magical.  Probably never 
> happen again.
> 
Ah, oversampled.  That helps in understanding the situation.
At SRO, we get about 50electrons/pixel in a 60-second exposure
with 1.25arcsec/pixel on a C14.  If you scale things roughly,
that would result in about 6 electrons/pixel with your setup.
SRO is in a pretty dark site in southern Arizona, so I think
the numbers will be about equivalent.

If you have the typical 2e-/ADU gain, then this equates to 2-3 ADU
per pixel in your 60-second exposure.  Since this is less than the
readnoise, and less than the dark current noise, you are dominated
by noise sources other than the sky.  Seeing a couple of counts of
sky would then depend on averaging over large areas, and negative
pixel values will obviously occur since there will be positive and
negative errors from the mean dark level.

I wouldn't worry about it, except in how your software handles the
negative values.  The incorrect way is to require all pixel values
to be positive, and to replace the negative values with zero.  If
your software does this, I suggest adding a constant value to your
image (say, 100 to each pixel) before any step that would truncate.
Arne