[vsnet-chat 2085] Re: Plotting light curves

[aah@nofs.navy.mil]

Kato-san mentioned that they normally use 30sec
exposures and can take >1000 per night.  I can't,
and it is due to the deadtime processing that is
involved with our TK1024 chip.  It takes 5sec to
prepare the CCD for an exposure, 25sec to read it
out, 5sec to create a fits file and add header info
from the telescope, and 5 sec for display and
miscellaneous operations.  In other words, even
for a zero-length exposure, it takes us 40 seconds
before we are ready for the next exposure.  So
taking 1second exposures of some bright object still
means 41 seconds before we can take another frame.
This is a very low operating efficiency, so my
rule of thumb is to avoid projects where exposure
times are less than the dead time.  Of course, there
are several ways to get around this and improve your
efficiency, but that is another topic.
  We store our CCD frames in <name>.<extension> format,
where <name> reflects the UT date and <extension> is a
running counter that starts at 101 (don't ask me why!).
So for the observing software here at NOFS, you can only
take 900 frames per night before getting
an error.  This is just tradition and no one has taken
the time to modify the software to handle more frames
since we have never come close to the limit.  Our new
infrared camera can store frames at a 1Hz rate, meaning
that it can potentially take thousands of frames per night.
Again, the normal number is more like 400 since we coadd
fast exposures before storing to disk.
  So for the case of a telescope with tracking problems,
where your exposure times are limited by hardware rather
than by your science, then binning or coadding the frames
before analysis makes sense.  However, I'd still stay
clear of smoothing unless you have a specific goal in
mind and understand the interactions that take place.
Arne

VSNET Home Page

Return to Daisaku Nogami

---

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