Some notes for observing at the Swope 1-m telescope
Kevin Krisciunas, 9 January 2012
How can one know the right exposures for a particular supernova?
You cannot know ahead of time. But during the course of a
night I display EVERY image I get of a field that is supposed to
have a supernova in it. Of course the simple way is just to type
display 1
in the iraf window. This will return two numbers (z1 and z2) which
"display" has used to scale the image. And the image is displayed
in the ds9 window.
But, often a supernova is buried in a host galaxy. So try this
once you have "z1" from one iteration of display. Say z1 was 500. Try
display zs- zr- z1=500 z2=1500
then hit and iraf will prompt you for a window (1:16).
Then you can use the right hand mouse button to adjust the scaling
on the ds9 display of the image, and you will probably see the supernova.
Make frequent use of imexam and the "r" command. This will give you
a radial plot of a star located where you have the cursor in the
ds9 window. If your standard star (or supernova) gives more than
22,000 counts in such a radial plot, then you are getting to the
non-linear realm of the chip. If you get more than 32,000 counts,
you are saturating, and the exposure time was definitely too long.
________
If you take twilight skyflats, you have to offset the telescope between
exposures. Why? Because as it gets darker after sunset, the stars
appear in the images and by offsetting the telescope you can get rid
of these stars when the filter by filter median skyflats are made. But you
need 3-5 images, not just 1 or 2.
At the telescope control system keyboard type
RBB 30
DBB 30
Then shortly before a skyflat exposure you type "NN" for example on
the TCS control keyboard.
Within 1-3 minutes after sunset one can usually start taking u-band
skyflats and not get more than 22,000 counts. When an exposure is
done you can type
imstat
in the iraf window and you will be given the mean counts of the image.
These should be 9000 to 20000 for skyflats and domeflats.
Though I've been taking domeflats too, I prefer skyflats for data
reduction. After sunset but before it gets too dark I can usually do
5 skyflats in u (exposure times 3-5 sec)
5 in B (exposure times 3-8 sec?)
then 3 in V (exposure times 10 to ??? sec)
(3 in g) or (3 in r) or (3 in i)
There isn't time to get skyflats in all 6 filters before it gets too
dark.
Of course, as the sky gets darker, you have to increase the exposure
times of the skyflats.
I get the mean counts of the skyflat and domeflat frames using
the "imstat" command. One can do them all by typing
imstat ccd*
____________________
For the proper calibration of photometry one also has to observe
standard star fields on clear nights. (If the sky has obvious clouds,
then the only real use of a standard star field would be for doing
a pointing test or a focus test.)
Here's a good plan for observing on a photometric night:
skyflats (exposure time 3 seconds at start, maybe 180 seconds at end)
pointing check
If you use coordinates from the Astronomical Alamanac they might be
2012.5 coordinates, so you have to type "MP 2012.5" on the TCS keyboard.
When done with pointing check don't forget to type "MP 2000.0" for
regular observing.
Go to a standards field. Do 7 exposure focus image with subraster off.
Filter is V. Don't forget to turn on subraster when done determining
focus.
If it is a clear night, observe the standards field in all 6 filters
(BVugri).
A photometric night might then have a sequence of pointings like this:
standards field
1 or more supernovae
standards field 90-120 minutes after previous standards field
1 or more SNE
standards field 90-120 minutes after previous standards field
1 or more SNE
standards field 90-120 minutes after previous standards field
1 or more SNE
standards field 90-120 minutes after previous standards field
Ending the night with a standards field.
Why start and end the night with standards and do them every
90-120 minutes? So that one can demonstrate from those images
how constant was the transparency of the sky.
_______________
Other things to consider:
If the supernovae were observed over a wide range of airmass (1.0 to 2.2),
then for best calibration of the SN images you should observe standards
over the same range of airmass. Why? So that the atmospheric extinction
coefficients can be determined in all the filters. Even under clear sky conditions
these coefficients can range by +/- 40 percent.
Make liberal use of "imstat" and "display". When first setting on a new
location make sure you're pointing at the right place by comparing a
displayed image with the finder at the CSP2 website. No sense using
up 45 minutes of telescope time making 6 images of the wrong place on
the sky.
When you display an image, look for a cosmic ray hit close to the location
of the supernova. These are often spikes that take up only 1 pixel, while
a star takes up many pixels. A surface plot of an image can be obtained
with imexam and its "s" command. If there is a cosmic ray hit very close
to a supernova, it's often best just to take another image. How close
is "close"? 15 pixels is not a problem. But 5 pixels could make for
problems.
Another reason to display and inspect every image is that maybe you forget
to change the focus, change the filter, or turn the guider back on. Of
course, you don't want to waste a lot of time doing this. One can inspect
an image while the next image is being taken. That uses up no telescope
time.
-------------------
I wasn't quite explicit in the notes I sent you earlier today.
1) If there are some clouds, then your only option is domeflats
and SN images.
2) If it's nice and clear one can also take twilight skyflats.
And if it's clear, you should really observe standards too,
not just supernovae.