Re: [TASS] Flat fielding

[Tom Droege <droege@WWA.COM>]

We had many discussions about light boxes and flat fielding.  Show and Tell
Note #1 has pictures of the design that I ended up building into "Dead
Barney".  TN42 has drawings of the flat field box and pictures of a model I
made to explore the desigh.  TN-39 shows an earlier version of the design.
 There was a design using a folded light path and mirrors that I was talked
out of by various people.

One of the reasons for having a temperature controlled TEC to cool the CCD
is to keep a uniform dark field from day to day.  This is a modest help for
the flat field problems.  In theory, the flat once known, will stay pretty
constant.  OK, we will collect dirt.  But it is not an every day
requirement to make continuous flat field experiments.  We shall see how
many of us feel the need to build flat field hardware.

My opinion at the moment is that there are so many problems with these wide
fields, that a median flat is apt to produce the best results.  We have
lots of exposures to median.  But I have an open mind on this.  I would be
delighted to find someone really work on the filat field and demonstrate
that lower errors can be achieved.

As for the Mark IIIs.  They are a special problem and are (from one fiew)
much easier.  I see more error in my data from changing conditions that
from real flat fielding problems.  At least in suburban locations like
mine, there is more problem from a neighbor turning on a porch light that
shines on my camera than any dark or flat field problem.

I don't think that we ever thought that "flat comp solves the problem".  It
is more that flat comp does good enough with what can be done that we are
not likely to do better.  OK, if we all lived on mountains and owned rifles
to shoot out neighbors porch lights we could do better.  Not that I want to
give up on doing better.  There is just a limit to what we can do in our
locations.  A catalog is still useful even though it does not have the
accuracy that could be obtained if we ran the project at professional
locations.  It is a trade off.  We will do what we can do well.  That is to
make frequent measurements at lesser accuracy.  The mountain locations will
(hopefully) supplement this data with a lessor number of measurements of
higher accuracy.  Both results will be useful, I think.

Tom Droege

At 11:41 PM 9/26/99 GMT, you wrote:
>I think it is worthwhile at this point to summarize the
>progress that has been made since this thread started
>Sep8 with my query as to whether Mk III flat fielding was
>as good as it could or should be. This question was
>intended as a lead-in to the more difficult problem of
>Mk IV flat fielding.
>
>The initial response that there is no problem
>(with the Mk III) was amplified (Mike G, Sep10) to
>>It is true that the current flat technique is oblivious to N-S gradients.
>>That's why the FlatComp program was born.
>There is general concensus that, with FlatComp, the
>Mk III problem is solved.
>
>The Mk IV flat fielding is a more serious problem.
>Tom Droege (Sep11) reported work on a light box and on the
>same date Chuck Shaw told us about his, described at
>http://www.ghg.net/cshaw/flat.htm
>On Sep23 Arne Henden (on another thread) wrote.
>>  Flatfielding a Mk4 should *not* be a major
>>problem as long as a light box is constructed.
>So we seem to have arived at a concensus here too.
>
>This, I think represents real progress. I, at least,
>was not aware before that the Mk IV problem was fully
>recognized and that plans included construction of a
>suitable light box.
>
>Andrew Bennett, Avondale Vineyard, Nova Scotia, Canada.
>
>