ULTRASPEC remote observing checklist
The purpose of this is to provide a set of reminders for remote observing
with ULTRASPEC on the TNT.
- Check you can connect to the drpc. If you can't, you will need
to check with the TO that "observer" is logged into the drpc (see
next section for skype info).
- Make sure that the USB disks are mounted.
- Make sure no one else is doing anything over the network that
will slow you down (e.g. check with 'top' and 'ps aux' for ssh
processes). If you need to send a message to everyone of the system,
"wall" as root is a possibility. "su -" in a terminal, type "wall",
then your message ("Dear observer, please log out of the ultraspec drpc
as we are about to start a remote observing session") then ctrl-D.
- Check the sunrise / sunset times.
- Create an 'eplanner' plot for the night to work what you will
observe. Remember greyed out sections indicate mast obscuration which
affect the NE direction. You will normally want to avoid unless the
programme can tolerate dodgy photometry (e.g. "does this star have
eclipses?"). Note the filters that will be needed during the night.
- Get some food and drink in place. It might be a long session!
Start of night
- Open skype to "ControlRoom TNO". ("TNO Observer" old alternative
which might be out of date.)
- Before the dome is opened, and the dome lights are off, take
biases. Always full-frame 1x1 and 2x2, and specific formats from the
preceding night. Make median biases from these with "makebias".
(Ideal bias setup: dome lights off, air conditioning off,
primary mirror covers closed, M4 mirror moved to the 4k camera
position, lowest throughput filter in, focal plane slide to "block").
- Check weather at the TNT
- Remove FP slide, point M4 at ultraspec, open mirror. If the sky is
clear, tell the TO to point at a blank field near the zenith (they
should have a list). If more than one filter is in use, start with the
lowest throughput filters (e.g. u and z), end with the highest
(e.g. KG5). Check with the TO that the telescope is spiralling to
allow stars to be removed when taking medians.
Once you are done, make the mean flat with "makeflat".
- After the sky flat, you might want to take a flux standard, but
only if the sky is absolutely clear.
- If you can't take biases and/or flats, you might want to try to find
some old ones to use during the night.
- Move to the first science target.
- For each target, give the TO the telescope RA, Dec and PA from the
usfinder-generates chart. If there isn't one, you will likely have to
move straight to the target RA, Dec with PA=0, unless you can
quickly run 'usfinder'.
- Once you are on the target, get an exposure going, set it to have type
"acquisition" in usdriver. Check positions with "rtplot". You might
want to set the exposure time to be less than you will use during the
science run just to get frequent updates.
- You need to run rtplot with a file of defects to avoid the more
obvious ones. Vik created one during the Dec 2019 run. Also, if
possible, use a bias and flat in rtplot because it helps the
contrast. Set the xlo, xhi, etc region to avoid the extreme edges to
cut off any bias strip as it makes percentile image scaling
better. Check also that the obvious comparison stars are free of
defects. If you do need to move, there is a useful utility on the drpc
(should be an icon) which allows you to type in the X,Y offset you
want along with the PA and returns the shift you need to ask the TO to
make (which will be something like "3 arcseconds West" ... pause while
they do it ... "2 arcseconds North".
- Check exposure levels; make necessary adjustments to usdriver settings.
- If using drift mode, put the focal plane mask in; if you previously used
drift mode, take the mask out!
- Switch the data type in usdriver from acquisition to data (?? check),
start a science run, but don't relax just yet because you want to get
reducing asap to enable focussing.
- Generate a reduce script using "genred". The very first time through
you might want to run it with "prompt" on the command line to get the full
set of prompts.
- Get reduce going. First thing to do is look at the FWHM seeing to
see if you can improve the focus. Note the current focus (should
appear on usdriver window). Ask the TO to change. Typical useful
increment is +/-0.03. I normally like to change it enough that it is
obviously getting worse an to do so on both sides of the focus. Always
remember to let it settle so you are getting a good idea of the true FWHM
at a given focus position. Always difficult in very variable seeing.
- Now you can relax, sort of. Things to be alert for are focus and
position drifts. The latter seem to occur particularly though
meridian. You may need to ask the TO to make adjustments. Try not to
let the target drift more than 4 pixels in X and Y before correcting
- Also stay awake for: rotator limit and the need to move the
shutter. Both of these will cause reduce to fail and you are best
advised to stop and re-start.
- Remember to keep a blog of the night: problems, weather etc. These
are a good way to communicate what has happened to everyone. However
don't use them as an observing log. It's always best to write comments
about seeing and clouds into the run-by-run comments file. Remember
you can write comments lines as long as you want; just don't insert
any line breaks. Good run comments are specific and factual,
e.g. "Science. Primary eclipse. Seeing 1.3-2 arcsec, clear." or "Potential pulsator. No obvious pulsations. Seeing 2.3 arcsec, clouds affected 25 mins of the run".
End of night
- Possibly another chance for sky flats, although you and the TOs may be wanting to go to bed.
- Tell the TO that you are finished and thank them.
- Finish the night blog.
- Get Paul Kerry's archiving script going.
Tom Marsh, Warwick