[Aavso-photometry] Negative CCD observations

[Michael Newberry]

Robert,

Yes, you have the basic idea. As you say, using the 3-sigma value gives the 
pojnt of about 0.3 magnitude uncertainty, which is classically the point we 
call the detection limit of the image. But getting that value is simple if 
you want a rough number, or tough if you want a fairly accurate number.

Even if you carefully place the measuring aperture and everything appears 
"claen", extrapolating a "bright" star measurement down to the image limit 
is a risky proposition. There are a number of reasons: One is that the noise 
does not scale in a straightforward way as a function of brightness---not 
just because of the theoretical noise equations are not linear, but also 
because other nonlinearities arise in of the way noise is added to the image 
by processing (bias, dark, flat, and combining). Another problem is possible 
nonlinearities in the CCD response. A third issue is that the noise measured 
for one star is only one estimate of the noise at that magnitude and it is 
subject to some variation. Yet another problem is that most software does 
not estimate the noise very correctly and how incorrect the value is depends 
upon many factors. For getting a decent value for the limiting magnitude, 
you are best to measure a bunch of stars down to the image limit and plot 
sigma(m) versus m. You will see the curve rise abruptly at the faint end. 
Where it crosses sigma(m) = 0.3 is what you would call the plate limit.

On the issues of noise added by processing and the correct magnitude error 
equation for sky subtracted CCD photometry, you should see my paper in PASP, 
Jan 1991, p 122. It is available from the SAO preprint server.

Michael Newberry

----- Original Message ----- 
From: "Robert J. Modic" <rjmodic at n2net.net>
To: <aavso-photometry at mira.aavso.org>
Sent: Wednesday, January 25, 2006 4:30 AM
Subject: [Aavso-photometry] Negative CCD observations


>I was wondering what the best method is for making negative
> CCD observations.  The eyeball method would be the easiest,
> but somewhat subjective.  I'm looking for a quantitative
> method to find the limiting magnitude for a particular frame.
> I read that the rule-of-thumb is that a SNR of 3 is roughly
> the threshold brightness for a star.  Can I just measure a
> comp star with a SNR of ~20 to 50, divide the comp star SNR
> into 3, convert this result into a differential magnitude
> and then add it to the comp star magnitude?
>
> Something like this:
>
> STD+(-2.5*Log(SNR3/SNR20))
>
> I could then average the results for a few comp stars to get
> a better result.
>
> Does this make sense or is there a better way to do this?
>
> Bob Modic
>
>
> _______________________________________________
>
> Aavso-photometry mailing list
> Aavso-photometry at mira.aavso.org
> http://www.aavso.org/mailman/listinfo/aavso-photometry
>
>