Re: Saturation Improvements

[Andrew Bennett <andrew.bennett@ns.sympatico.ca>]

On Mon, 02 Oct 2000 23:02:40 -0500, Tom Droege 
<tdroege@veriomail.com> wrote:

>The thing that I have learned in going to CCD-World is that some of the 
>saturated charge can be lost.  There seem to be two full wells, Surface 
>Full Well (SFW) and Bloomed Full Well (BFW).  In SFW, some of the charge 
>can be lost due to recombination.  This is apparently what is happening 
>when saturated stars at the top and bottom of the frame have different peak 
>values.  I would be very cautious about using saturated measurements.
>
>This is something we can investigate in the future (or you can do it now - 
>Disk 15 will do it).

On Mon, 02 Oct 2000 20:43:37 -0700 Chris Albertson
<chris@216-178-250-236.lax1.phoenixdsl.net> wrote

>Yes you can do photometry on saturated stars.  You look at the shape
>and size of the plateau and fit a curve to the bottom part of the
>plateau then look to see where the peak would have been if it had
>not been truncated.  Likely a lot of messy details involved.

When I did this, 'way back when, it emerged that most of the
overflow charge ended up in adjacent pixels. A reasonable
estimate could be obtained by summing all connected pixels
above background regardless of how far the charge had wandered.
The scatter was a good deal worse than for non-saturated
bright stars and there was IIRC some evidence of charge loss for 
the very brightest. But the method did gain several magnitudes
in dynamic range.

Attempts at fitting proved worse than useless by comparison
with the simple sum.
As a life-long proponent of least-squares fitting, saying
this really hurts!

Andrew Bennett, Avondale Vineyard, Nova Scotia, Canada.