CHAPTER 3: Observing
LAST UPDATE: 25/02/2004
Introduction: Observing with CPSR2There are two different ways in which CPSR2 can be run. The method you choose will depend on whether or not you wish to simultaneously observe with the filterbanks. It is possible to observe simultaneously with the wideband correlator using the Parkes Telescope Control System, but the filterbanks are not yet integrated into that system. No matter which method you choose to follow, you will always need to run the CPSR2 cluster monitoring and configuration tool, "GUIMonitor", as described in the previous chapter.
It should be noted that unlike the filterbank "pmmon" system, the GUIMonitor does not attempt to drive the telescope. It is purely an interface to the FFD and cluster nodes. Our software has been designed with the new Telescope Control System in mind. TCS is a graphical user interface written by staff at the ATNF. It handles schedule files, configuration of the observatory hardware and control of the telescope itself. It can also connect to and control the various backend instruments (eg. the multibeam correlator). The latest version of TCS will talk via a network socket to the CPSR2 control system, sending it all the information required to start and stop data aquisition. This means that all you have to do is keep an eye on the cluster and step in if something crashes. If you observe in this mode, the CPSR2 GUIMonitor is used primarily as a means of monitoring load, disk space and memory buffer levels, all the things specific to CPSR2.
If you want to take simultaneous baseband and filterbank data,
one of the two instruments will have to be run in manual mode,
where you enter the observation parameters and start/stop the
data recording using an instrumant GUI while driving the dish
with different software.
Section 1: Observing with TCSFollow these instructions if you want to use TCS to control the CPSR2 instrument. This is usually the easiest way to do things as it is the only way to make CPSR2 operate under a schedule file. TCS runs most happily under Solaris. This means that you should log onto one of the Sun machines in the lower control room as "pulsar" and run it from there. Be sure to enter the command atcsver next, which will ensure you are running the version of TCS that can talk to CPSR2. Once you have done that, simply enter tcs and wait for the first window to come up. You will be asked to select an observing mode, use the "pulsar" setting and be sure to check the "expert mode" box at the bottom of the panel.
Once the TCS GUI has launched, you can load or construct your schedule file (see the TCS manual). Make sure you enable the CPSR2 instrument in the backend selection panel (lower left) You may have to disable the multibeam correlator (we do not use it, but TCS tries and fails to configure it anyway). If you intend to use the wideband correlator, make sure the appropriate symbolic link is in place and run the wbcorr control GUI before trying to start TCS. If you choose to observe without the correlator, disable it using the pull-down menu and leave the first orange warning message up, just move it to a blank spot on the screen. Otherwise the warning will reappear every time a new instruction goes through.
In theory, that should be all you have to do! When you command TCS to observe, it will slew the dish to your source and once there it will send a string of commands to cpsr2d which will construct a valid header for the observation and automatically start recording data. You should see the two terminals on pegasus that contain the "20cm_cpsr2_das" program acknowledge the "GO" signal and start counting the seconds. If the "GO" signal is not received, you must stop the track in TCS and figure out what went wrong (see chapter 5).
One point to note is that TCS doesn't send the bandwidth flag to CPSR2, so it just uses whatever was set last. You have to make sure you set the box in the manual control panel to match the down-conversion chain configuration and then do a short observation to set the headers correctly. Normally it will be either +64 or -64 but the difference is crucial.
From here on, every time TCS changes source it will stop the data recording and send the new header parameters to cpsr2d. All you have to do is keep an eye on the GUIMonitor to make sure that you have plenty of memory buffers and disk space free.
See chapter 4 for instructions on data reduction. Note that you must process the baseband data while observing if you do not want to fill the disks in a matter of hours.
Section 2: Observing in Manual ModeIn manual mode, it is up to you to configure the conversion chain and drive the telescope using your software of choice. When the telescope is on source you will have to manually type in the information requested in the first six fields of the GUIMonitor manual control panel and then hit the start button. When a track finishes, you must manually hit the CPSR2 stop button (before the source moves out of the beam!) and enter the information required for the next source. In most cases you can even get away with leaving the sky coordinates set to zero, provided you are sure you will not need to refer to them after the data has been recorded (eg. if the pulsar has a well known catalogue position).
When running with TCS, the LO chain is automatically configured based on the parameters you enter in the TCS CPSR2 control box. When using the filterbanks to control the telescope, you will have to manually configure the down-conversion system to provide the proper LOs and attenuation for the CPSR2 bands. This can usually be accomplished by loging into a sun machine as pulsar and running one of the losetup scripts. Thse live in:
and all have a ".cmd" extension. The syntax required to use one of these files is as follows:
Some of the most useful ones are described below:
You may have to manually adjust the attenuation settings in the conversion chain to provide the right levels for CPSR2. This can be done with the ATNF "logui" program. The levels themselves can be monitored using the "lomon" program. The CPSR2 FFD is connected at the four points called "c40 IF64M", and the levels should stay close to 3.0 volts. Both of these ATNF programs should be run on the sun machine called sagitta, using terminals on the quad screen linux box on the shelf in the middle of the room.
Back to the index