[vsnet-chat 1719] Re: CCD's Lumpy flat field?

[aah@nofs.navy.mil]

Stan wrote:
>How real is this 'few parts in a thousand' value? Or what is a few? Just at
>the moment I'm having trouble getting better than 2-4% with CCD BVRI
>measures but these are across 20-30 degrees between images. Unfortunately
>the ST6 with the filter wheel vignettes badly at f6.3 which may be a part
>of the problem. But the values quoted by Lew seem to be very good for CCDs.
>How are these determined?
  Differential CCD measurements, where the variable and its comparisons
are in the same field, can reach a few millimag accuracy.  Gilliland and
his collaborators did this a few years back for M67, and Brian and I just
heard a talk from Ken Janes where he is reaching this accuracy as well.
Remember too, that in order to do millimag precision, you have to have
S/N=1000 or so, which is much more difficult than what most observers
are used to.
  All-sky work, like what Stan is doing, is a different beast.  You must
have a stable sky, so that variations between one measurement and the next
can be correlated; you must account for extinction variations between the
two regions of sky; seeing variations will affect the photometry if you
use the same size aperture for both fields; flatfielding problems will
rear their ugly head since the objects in both fields will not fall on
the same pixels.  I usually consider 2 percent as a good number for random
fields and placement within a field for a CCD system, and 1 percent as
being excellent.  Lew's quote:
>Anyone doing 0.05 mag all-sky CCD photometry is doing really well.
Should be taken with care, since I would personally discard any night
with photometry this poor.  For a beginner, 5 percent is a good goal,
but you should work hard to get beyond that limit.
Arne

VSNET Home Page

Return to Daisaku Nogami

---

vsnet-adm@kusastro.kyoto-u.ac.jp