Contents
Quick Guide for Dummies - The novice 30m observer (DRAFT)
by Carsten Kramer
This guide is meant to help the novice observer at the 30m telescope getting started. At the end of each section, and in the Appendix, we give links and references to more detailed documents. THIS IS ONLY A YET VERY INCOMPLETE DRAFT!
Howto observe ?
Well before your observations are scheduled to start, it is a good idea to prepare observing scripts, i.e. small PaKo routines, to setup the receivers and spectrometers, to select your sources and lines, etc.. The Astronomer-of-Duty is there to help you doing this. Scripts should reside on the subdirectory ~/pako of your project account.
When time has come to start real observing, take a seat and let the Astronomer-of-Duty and the Operator guide you through the first steps. It is really easy. You'll need to log into your project account onto mrt-lx1 and start the observing software called PaKo, run the receiver and backend setup scripts, select a source, and then e.g. start a pointing. This will send the commanded frequencies to the synthesizers, and it is now on the operator to "tune" the receivers, to make them ready for observations to start. This may take 10 minutes or more. Once tuning is finished, you may want to check the pointing and focus of the telescope first, before going to your science source. See below.
Getting started
- Log into your project account at the PC on the observers desk
- Start a new X-window/ KDE session under your project account
Click on icon "paKo". this will connect to your project account on mrt-lx1
In the X-window, type: goPako, pakoDisplay, pako
Run your setup scripts a the pako prompt, e.g. @ setup
And then continue with set doSubmit yes, start
A typical observing session: PaKo
Observations on the science source are mixed with calibration, pointing, and focus scans. Observations of a line calibrator are recommended at least once after a new receiver tuning. Observations on the science source always switch between a reference and the on-observations. The time for the various cycles (on/off, hot/cold/sky calibration, pointing, focus) depends on the stability of the sky and the instrumentation, among other things. Typical cycle times are 2minutes for the on/off cycle, 10minutes for the calibration cycle. Pointings should be done every 1-2 hours, depending on the stability of the pointing. Focus scans should be done about every 2 hours, and more often during sun set and rise.
CHECK: typical session to be added
Computer System
What are all the screens for ? Network topology or entering the unknown
mrt-lx1 is the name of the workstation to run the observations. This is the central machine, the observer is using for running PaKo. The online data processing software Odp is running on mrt-lx3. During observations IMBfits raw data are continuously transferred from mrt-lx1 to mrt-lx3. The observer should use mrt-lx3 (or mrt-lx2) for offline data processing.
While the observer logs into mrt-lx1, the observers account lie on the vis directory on disks connected to mrt-lx3.
The main machines of importance to the observer run under Linux, Debian or Ubuntu.
CHECK: A sketch of the main computers and disks relevant for the observer, and a sketch of the entire 30m-Granada network system.
My machine is called mrt-lx18. Now what ? Software
PCs with numbers larger than 4, are small PCs used as X-terminals. Not much software is installed on these PCs and the user may want to quickly change to one of the main workstations described above, e.g. via ssh -X mrt-lx3.
CHECK: Say which software is installed on lx3: mira, class, greg, xemacs, xephem, IDL, etc.
Printing
A typical print command is lp -Pmrt-pr1 file.ps. At the 30m, we have printers mrt-pr1, a fast b/w printer, and mrt-pr2, a color printer.
Data
Online data processing (Odp)
The Odp creates calibrated spectra online during the observations, stored in the so-called 30m files, and readable by class. In most cases, the observer works offline only with these files. However, in case of problems with the data, e.g. calibration issues, the observer needs to know to some extent what Odp is doing.
During observations, the backends take data and save them in streams in FITS format. These streams are merged with streams from other parts of the system, e.g. antenna mount drive, and also with messages, to be included in IMBFits raw data files. These are then combined by mira/odp to create calibrated spectra, usually on the antenna temperature scale TA*, and saved as 30m files. These in turn can then be read offline with class by the observer. Mira, class, greg are part of the GILDAS software package. Astronomical data therefore is created in several levels: level-1 are the streams, level-2 are the IMBfits files, level-3 are the 30m files. In case of calibration or other problems, it is possible to recalibrate data using MIRA, or even to recreate data from the streams.
CHECK: Give small sketch of data-flow
Where are my data ?
- On the observers project account,
30m data are found here: ~/observationData/mira. One file is created per backend and per day, e.g. FTSOdp20120105.30m
IMBfits raw data are found here: ~/observationData/imbfits. Their file name gives date, scan-number, backend, e.g. iram30m-fts-20120107s167-imb.fits.
What happens to my data ? Backup, Storage
- At the end of an observing run, the operators make sure that the observer receives a copy of the project account. Upon request, IMBfits raw data are included. Depending on the project size, the data may be provided on an external USB disk (provided by the observer, or by IRAM) or DVDs.
In addition, we retain at least one copy of the project account (including all 30m files and IMBfits raw data files) at the observatory. Upon request to the computer team, we will retrieve data. Depending on when the data were taken, data are stored on different media. Most recent projects are stored on a local disk array, the project archive prArch and/or on Ultrium tapes. Older data are still available via DVDs, and CDs.
- Data streams are not regularly backuped.
How many data to expect ?
This depends on the spectrometers used, and on the observing mode and data rate, i.e. how many spectra are taken per second. Don't worry if you are not using the FTS backends. In this case, your data will fit on DVD(s). If you are using the FTS, you may well need a hard disk.
The 30m has 2SB dual polarization receivers covering 32 GHz of instantaneous bandwidth. The FTS backends cover the full bandwidth at a resolution of 200 kHz. This gives 160,000 channels. The typical dump rate is 1 spectra per second (e.g. for position switched observations). This gives a typical data rate of 2.3 GB/hour. It has been tested that the 30m system is able to cope with a dump rate of 5 Hz (100 msec phase time), for instance when conducting frequency-switched on-the-fly mapping observations.
The amount of required disk space is doubled if you plan to take as well the IMBFits raw data back home.
Copying data back home
This is an option when the amount of data does not exceed a couple of GB.
CHECK: give some typical commands, and give typical data rates
Saving data on external disks
Especially, for projects creating large amounts of data, it is a good idea if the observer copies data over to his/her own laptop or external hard disk. This should be done in regular intervals during the run, and not only at its end.
Assuming your project number is 100-11, change directory to the subdirectory on your local disk where project data shall be stores, and then copy the data over using rsync -rtvH --safe-links --bwlimit 2000 100-11@mrt-lx3:/vis/100-11/* ./.
Detailed information is given here (internal).
Next steps: user dp
Spectral line data are calibrated automatically with MIRA during the online-dataprocessing (odp). Next, the observer takes a look at the spectra, fits and subtracts spectral baselines, fits Gaussians, creates data cubes, etc., using CLASS which is part of the GILDAS software package.
Howto work with spectra: CLASS/GILDAS
The basic CLASS commands are explained in the online CLASS documentation, e.g. in this short example session.
More advanced, recently introduced new commands are explained in a lecture given at the 30m Summerschool 2011. It gives examples on how best to visualize data and goes through a state-of-the-art data reduction session.
Details like telescope and instruments
Whom to complain to ? Help needed
The Astronomers-of-Duty and the operators are your prime contact at the telescope. Don't be afraid of asking "silly" questions, please!
Appendix
Displays
CHECK: Show screen shots of main displays, and explain what is most important for the observer. Where to find the scan number, e.g..
More information, references
See the lectures on telescope, frontends, backends, and software given at the IRAM 30m Summerschool 2011.
This page is maintained by CK. Any comments are welcome.