Re: [TASS] Flat fielding

[Michael Gutzwiller <deepsky@FUSE.NET>]

It is true that the current flat technique is oblivious to N-S gradients.
That's why the FlatComp program was born.  It takes the 0.050 noise floor
down to 0.035.

Mike G.

-----Original Message-----
From: Andrew Bennett <andrew.bennett@NS.SYMPATICO.CA>
To: TASS@LISTSERV.WWA.COM <TASS@LISTSERV.WWA.COM>
Date: Friday, September 10, 1999 11:52 AM
Subject: Re: Flat fielding


>On Thu, 9 Sep 1999 22:29:46 -0500, Tom Droege <droege@WWA.COM> wrote:
>
>>The flat fields were definitely not up to snuff.  I don't even have any
>>good idea about how to make a good flat field for the Mark IIIs.  You
>>cannot make twilight flats with a device that takes a 470 second long
>>exposure.  At twilight the conditions are changing much faster than that.
>>You can't point it at a "dome flat" spot because it does not point.
>
>If the camera won't turn to point at the dome, the dome must turn
>to point at the camera.
>
>>  You
>>just have to look at the data and use that.  This is what Mike G's program
>>does.  I think that this is all one can do.  I think Mike went to a lot of
>>effort to think of the best possible way.
>Yes indeed! Simple and elegant.
>
>But if the N-S illumination gradients are really putting in
>5% variations in sensitivity ...
>
>How big are the variations in average sensitivity from
>column to column, assuming a minimal population of
>dust bunnies? Is most of it lens vignetting rather
>than CCD?
>
>******** heretical matter follows *********************
>Could it be that one would be better off in a drift-scan
>instrument not doing flat correction at all?
>******** end of heresy. You may open your eyes now ****
>
>Andrew Bennett, Avondale Vineyard, Nova Scotia, Canada.
>