I. File Nomenclature
====================
To avoid confusion in the data reduction process (and to provide
some redundancy of information), we will standardize the file
naming convention for the different types of observations as so
(where YYYYMMDD is the UT date of the night and NNNN=a running
id number that is automatically generated):
Type of Observation File Name
------------------- ---------
biases bias.YYYYMMDD.NNNN.fits
darks dark.YYYYMMDD.NNNN.fits
dome flats dflat.YYYYMMDD.NNNN.fits
twilight sky flats sflat.YYYYMMDD.NNNN.fits
pointing standard point.YYYYMMDD.NNNN.fits (or possibly just "junk")
focus exposure focus.YYYYMMDD.NNNN.fits (or possibly just "junk")
linearity tests lintest.YYYYMMDD.NNNN.fits
exposure map tests expmap.YYYYMMDD.NNNN.fits
grid tests grid.YYYYMMDD.NNNN.fits
standard star fields std-"stdFieldName".YYYYMMDD.NNNN.fits (e.g., std-CDFS.0003.fits)
science targets sci-"targetName".YYYYMMDD.NNNN.fits (e.g., sci-M104.0010.fits)
For afternoon biases, darks, domes, etc., use the UT date of the coming night
(even if the actual UT date these afternoon exposures are the "day before".)
(Admittedly, trying to keep to this naming convention during the
heat of battle can be difficult, but it is good to try...)
II. Afternoon Tasks
====================
(Based heavily on text from J. Allyn Smith)
* EVERY AFTERNOON WE WILL OBTAIN A FULL COMPLEMENT OF BIASES AND DOME FLATS. *
INITIAL SETUP:
Set for 1X1 binning - this more closely mimics the pixel scale of the
DECam on the Blanco, and we shouldn't be rushed for time for the readout.
EACH AFTERNOON:
1. Obtain a set of 25 bias frames
2. Obtain a set of dome flats in each filter (g,r,i,z) - 10 each.
Getting counts of ~40K-50K electrons (??K-??K ADU), or about
1/3rd of the way to saturation, is good to aim for.
The following exposure times and Variac lamp voltage settings
are working well for the Nov 2009 run (might need to change
for the current run...):
filter exp time Variac setting Mean Counts
------ -------- -------------- -----------------
g 30 sec 60V ~93Ke- (bias=40Ke-)
r 30 sec 45V ~84Ke- (bias=40Ke-)
i 20 sec 40V ~80Ke- (bias=40Ke-)
z 15 sec 40V ~110Ke- (bias=40Ke-)
3. Obtain a set of 10 dark frames flats for each of the following
exposure times:
a. 10 sec
b. 30 sec
c. 60 sec
d. 100 sec
If you run out of time on a given afternoon, the dark frames are the
most expendable and can be skipped on that afternoon if necessary.
3. Copy the afternoon's data to Fermilab via rsync:
ssh -Y dtucker@des06.fnal.gov
cd /data/des06.a/data/DES_CTIO1m_Runs/Jul2010/
rsync -avz v1@ctios1.ctio.noao.edu:/ua82/data/\*.fits n1/
(here I am guessing that we have the visitor account "v1"
and that the data are on /ua82/data on ctios1; change as
appropriate).
You should do the rsync after the flats and also throughout the
night -- it will save time in the morning.
4. Copy the afternoon's data to the staging area on the Brazil Portal via FDT.
See http://twiki.on.br/bin/view/Documentation/FdtTransfer
and consult Marcio Maia (maia@on.br), Bruno Rossetto (rossetto@on.br),
and Luiz da Costa (ldacosta@on.br) for help.
=====
FIRST EVENING:
Set a time to meet the Ops Support person to go through setup
proceedures, first night help, whatever... This will probably be
after the twilight flats are done.
EACH EVENING (this assumes we will try twilight flats, esp. if
we want to try u-band twilights):
As a general note, this is the most hectic time of each night.
You'll be rushing to get through dinner, grab the night lunch and
head up the hill to be ready to go. Getting the twilight flats
going on time and fast enough to get as many as possible takes lots
of practice. It will come as time goes on.
1. When you get in the dome (even a few minute before sunset), store
your night lunch, put on soothing music (if you like), make sure any
data transfers are completed (if you started any in the afternoon).
2. Go up to the observing floor and do a walk around of the telescope.
Make sure TelOps has set it up correctly (they should, they're good).
(If you have to open the telescope, plan an extra 10-15 minutes to
do this).
3. Verify the clocks are correct and the telescope is pointed to
the twilight flat location (usually about -2 HA at -30 dec).
4. Set up the first twilight flat (Don't execute it yet).
5. Go outside and enjoy the sunset! Camera for the green flash if the
horizon is clear. Watch the viscachas play (or sun).
For the twilight sky flats in u, start out about 3-4 minutes after
sunset with a 5-6 second exposure, aiming for mean counts of ~20K-30K
ADU (test the first few for photon level). The timing will take some
getting used to. Keep taking twilight flats, gradually increasing the
exposure time to keep the mean counts at ~20K-30K ADU, until you are
taking 30-40 sec exposures. Because of the shutter issues, we may
only get u twilight flats. But, after 5 or so good (30+ sec) u flats,
you could change to the z,g,r,i filter and try some of those. If you
can get 5-6 good u flats each day and 1-2 good g, r, i, z flats each
day, we can build a good master frame.
Different chips respond in different ways, but on the CTIO-0.9m and
the USNO-1m I shoot the twilight flats in the order: u,z,g,r,i
(z response is poor, so it takes a bit more light than you might think).
Also, you want any stars that appear to median out after combining the
twilight frames, so you should either:
1. turn off the telescope tracking while taking twilights, or
2. use the handpaddle to dither the telescope pointing between
twilight flat exposures. (I generally prefer this method, but
either will work if you have enough frames.)
=====
SUNSET+2 minutes:
1. Back in dome, turn telescope tracking on (if you dither the flats).
At the 3-4 minute point, start the first twilight exposure. After
readout, examine it for level and adjust exposure time accordingly.
Remember the sky is getting darker while you are doing all this.
General rule: Once you're at the exposure level that you want, each
successive exposure will need about a 25% increase in exposure time.
(You'll probably have help for these next few steps on the first night.)
2. Late evening twilight:
a. use a bright star (mag < 5) to refine the pointing
of the telescope
b. focus the telescope
c. take an "art shot" of a bright (NGC or Messier) galaxy
to verify pointing.
======
DURING NIGHT:
1. At the beginning of the night, observe a set of 3-4 or so different
standard star fields, one near airmass ~ 1, one near airmass ~ 2,
and one or two at intermediate airmass (~1.5) (or 1.3, 1.7) (all 5 filters).
2. Observe target fields throughout the night, interspersing
them with the occasional standard star field. The target
fields should be observed at low airmass (preferably < 1.3;
definitely < 2.0); the standard star fields should be
observed a variety of airmasses between 1 and 2.
3. Sometime during the night (preferably relatively early), do a
Star Flat/Stellar Linearity Test.
It is best if the field is observed at low airmass (< 1.3).
This set of observations can be treated as one of your standard
star fields.
4. In the middle of night, take 5-10 biases as a check on bias stability.
5. Periodically (every hour or so), go outside the dome and check sky
conditions by eye. (Note, it can take up to 15min to dark adapt.)
Note your findings in the manual observing log.
====
END OF NIGHT
1. At the end of the night, again observe a set of 3 or so
different standard star fields (all 5 filters) and 5-10 more biases.
2. Shut down and stow telescope, fill the dewar, fill out night logs.
3. At the end of the night, save the following data, which can later
be used to help determine the quality of the night:
a. The CTIO Environmental data from
http://www.ctio.noao.edu/environ/environ.html b. The 10-micron GOES East satellites animated GIF movie for the night from http://www.ctio.noao.edu/sitetests/GOES/movies/ c. The TASCA all-sky animated GIFS for the night from http://ctmt02.ctio.noao.edu/allsky.htm 4. Start an rsync of the night's data (including the ancillary data from Step 9) to des06.fnal.gov:/data/des06.a/data/DES_CTIO1m_Runs/Jul2009/Raw 5. Start an FDT transfer of the night's data to the staging area on the Brazil Portal. 6. GET SOME SLEEP! The madness will start again all too soon.