Pulsar Scripts
On the right screen on naiad, start a couple of xterms as gbtops.
In one xterm, "cd include". There are a number of scripts here that will set up typical LO/IF configurations and/or run typical observations for you. Scripts that you may want to run are:
- setup_1A17_L.g : L-band, 1375 MHz centre frequency, 1.3-1.45 GHz RF filter in place, DCR controls cal switching signal.
- setup_2A39_800.g : PF1, 820 MHz centre frequency, DCR controls cal switching signal.
- pulsar_obs.g : allows choice of several different centre frequencies, sets up Spectral Processor as switching signal controller (not great if using only BCPM; follow instructions in Miscellaneous to change back to DCR). Run this script only to do the general and IF/LO setup; when asked to set up BCPM, hit "t" for "terminate". Remember to take the Spectral Processor out of the Scan Coordinator when taking data.
To run a script, type, eg, "glish -l setup_1A17_L.g". In all cases, it is wise to run the script twice in case the receiver manager wasn't turned on the first time through.
Cleo Windows
On a different desktop on the naiad screen, start the Cleo Launcher.
You will need to launch the following screens:
- The receiver you are using
- LO1
- IF Rack
- Converter Rack
- Device Explorer
Unlock each of the screens.
In Device Explorer, double-click on the BCPM device. Then go to the "Managers" menu and select "On". This should start the "init" sequence on the BCPM: lots of info should scroll by in the blue command-line window. If this does not happen, see Troubleshooting.
Check to see that the hybrid/gain settings, LO1 and Converter Rack LO settings are as expected for your observing setup.
If using the PF1 receiver, adjust the attenuators so that the IF powers are about 1.0.
GO
I usually start GO from another xterm on the same desktop where I ran the pulsar setup script.
Type "GO T" and deselect the spectometer before starting the main GUI.
You will probably need the "Observing Table" and "BCPM" panels, both of which can be started from the "Backends, etc." button.
Reminder: Never swipe text that you want to delete from the GO windows!
Type in the full path to your observing table in the "Observing Table" window and hit "Load". The message window at the bottom should say that the table is OK; if not, edit the file for mistakes.
In the BCPM panel, make sure the correct sub-manager is selected (currently only bcpm2 works) and select the desired bandwidth, polarization summing and sample rate. A table converting the mysterious "X2", etc. sample times to microseconds can be found here.
Select observing mode: you must run a scan in Monitor mode with the correct BW, pol summation and sample time before taking any data!
If in Search mode, select:
- Cal on/off. This is purely informational to the BCPM and has no effect on the presence or absence of pulsed cal signal in the IF.
- Disk/tape. "Disk" means the (small) local disk on the BCPM computer. "Tape" actually means to write to the /home/vraid2/scratch/your_login/bcpm2/ area on the vortex disk. For long data scans, always choose "Tape".
- Target and base filename: again, purely informational. The base filename has no effect if you are writing to "tape".
Timing and Stokes modes do not work yet. Sorry!
Observing
Choose the right IF input! In the blue command-line window, type "setif 1" for 600/800/Lband, "setif 2" for 340/450/S-band and "setif 3" for 85-3.
A typical sequence is: Monitor scan, cal scan, data scan. Here are procedures for each.
Monitor Scan:
- Select Monitor mode, BW, pol summation and Tsamp on BCPM GO panel.
- Set (long, about 10 minutes) scan length in main GO window.
- Single-click on the desired source in the Observing Table window.
- Hit "Run" in the Observing Table window, not "Start" in the Main window (this is because you are changing to a new source).
- When on source or very close to it, hit "balance" on the IF Rack, making sure you are talking to the correct set of Optical Drivers (1/2, 3/4 for 600/800L-band; 5/6 and 7/8 for 340/450/S-band).
- Check the powers and attenuators in the Convertor Rack (Rack A for PF, L-band; Rack B for S-band; always Convertor Modules 4/8 for the BCPM). The power levels should read about 0.1 (upper) and 0.15--0.2 (lower); I find attenator setting of 20 (upper) and 10 (lower) are about right for 600/800/Lband.
- By now the BCPM should be displaying changing levels. If not (ie the I/O rate is 0 MB/sec), see Troubleshooting. In the blue command-line interface window, type "setpower". For many BW/Tsamp combinations, the power levels won't be quite right and you will have to manually edit scripts such as "set1072" and "set14" to add or subtract dBs from the current settings. Make sure the bandpass settings (eg, 8e6) are correct for the given channel bandwidths!
- Once happy with the power levels on the BCPM, abort the scan.
Cal Scan:
- Select Search mode on BCPM GO panel; set cal status and source/file names. Set the file length to something sensible (e.g., 30 seconds).
- Set scan length in main GO window to about 35 seconds (leave a little extra time beyond a single file limit).
- Turn the cals on by going to the receiver Cleo window and using the pull-down buttons to change "Off" to "External" on both the X and Y cals. This will inject the 25Hz cal signal into the IF.
- Hit "Start" in the main GO window (assuming you have just done a monitor scan and it knows the right coordinates). No attenuator adjustments should be needed.
- After the scan finishes, wait for the BCPM to say "Idle" before proceeding to data scan.
Data Scan:
- Turn the cals off by going to the receiver Cleo window and using the pull-down buttons to change "External" to "MCB" on both the X and Y cals.
- Set the output device to "TAPE" in the BCPM GO panel.
- Select Search mode on BCPM GO panel; set cal status and source/file names. Set the file length to something like 900 or 1000 seconds. More than this is not recommended, as files will only be transferred to the vortex disk after first being written to the local disk. Of course, shorter file lengths are fine.
- Set scan length in main GO window to the total desired length plus 5 or 10 seconds (eg 14410 seconds; leave a little extra time).
- Hit "Start" in the main GO window (assuming it knows the right coordinates). No attenuator adjustments should be needed.
- After the end of the first file (900 or 100 seconds or whatever), it's a good idea to check that a file called "scratch" is being written to the /home/vraid2/scratch/your_login/bcpm2/ area. This file will be renamed when it is fully written. If no file is being written, see Troubleshooting.
- When the end of the scan is reached, the BCPM will continue to move data to the vortex disk. Eventually (30 seconds) the Scan Coordinator will give up on waiting for the BCPM to say it's finished and will issue a warning about the BCPM. This is normal and does not require action; the next scan will start just fine. However, wait for the BCPM to say "Idle" (probably 10 minutes) before doing anything that uses the BCPM manager. This does not preclude going to access to change receivers.
After observing:
Turn off the BCPM Manager in DevExplorer, then quit the BCPM gui.
Miscellaneous
Troubleshooting
- The Antenna doesn't go to the right place. If selecting a new source, make sure you hit "Run" in the Observing Table window instead of "Start" in the main GO window. Otherwise, you may not have control of the antenna; get help from the Operator.
- BCPM does not go through the "init" sequence when its Manager is turned on in Device Explorer. One thing to be aware of is that the BCPM gui must first be running in order for the manager to be happy. Try turning the manager off, starting the BCPM gui, and then turning the manager on again in the DevExplorer Manager menu. If this doesn't work, the manager is probably dead. Try the following recipe:
- Ssh to vortex as monctrl (get password from Operator)
- "Taskmaster vortex status" (BCPM manager is task 9.)
- "Taskmaster vortex stop 9"
- "Taskmaster vortex status" (Wait till task 9 is stopped, max 1 minute.)
- "Taskmaster vortex start 9"
- "Taskmaster vortex status" (Task 9 should have started again.)
- Turn BCPM manager on (or off, then on) in DevExplorer
If Task 9 doesn't stop or restart, or there are continuing Cleo error messages about the BCPM manager, the problem may be a dead process. Try doing "ps" (as monctrl on vortex) looking for bcpm and killing the process. Then go through the Taskmaster process again. If none of this works, call Mark Clark or Amy Shelton!
- Levels are always 0 and setpower does nothing. Make sure you have chosen the correct input with the "setif" command. Remember: 1 for 600/800/Lband, 2 for 340/450/Sband, 3 for 85-3.
- BCPM I/O rate is 0 MB/sec and the level monitor display is frozen. The BCPM scan is dead. This type of thing can happen for various reasons:
- The IF Rack was balanced (and levels changed a lot) while the BCPM was running.
- That BW/Tsamp mode simply doesn't work.
- The BCPM has decided it is unhappy and doesn't want to take data any more. Yes, this happens.
Suggestions:
- Kill the scan, balance the IF Rack, and restart the scan. If this is a normally functioning mode, it will probably come back to life.
- If the mode often works but is sometimes flaky (eg, 1.4 MHz BW, 50 usec sampling), try switching to another BW/Tsamp and running a monitor scan, then coming back to the mode you want. Alternatively, try cycling between summing and not summing the polarizations.
- Occasionally a normally functioning mode will simply stop working for a while. The only thing I can suggest under those circumstances is to run in a different mode!
- No file named "scratch" is being written to the /home/vraid2/scratch/your_login/bcpm2/ area even though the BCPM has been taking data for more than one file length. Things to check:
- Make sure you have selected "Tape" instead of "Disk" in the GO BCPM panel. If it's "Disk", the BCPM will stop taking data after one file length. Select "Tape" and start again.
- Make sure you edited hardware.conf before starting up the BCPM gui. Otherwise, it will be writing to the area that was previously indicated. Check the other person's directory!
- Make sure you did not create a subdirectory named "scratch". If you did, remove it and everything should eventually be all right.
- The levels drift very low or very high soon after the start of your scan. Your source is probably rising or setting, so the spillover is changing. Attenuators need to be changed somewhere. The IF rack is probably the best place; don't hit "Balance" but instead add or subtract dBs manually.
Writing backup tapes with the DLT drives on Vortex
For Example:
cd /home/vraid2/scratch/istairs/bcpm2
ls S22[78]*.bcpm2 | xargs -n 1 tar cvf /dev/rmt/2hn >
do.fill.GC_020815_tape19a.out &
ls S22[78]*.bcpm2 | xargs -n 1 tar cvf /dev/rmt/3hn >
do.fill.GC_020815_tape19b.out &
NOTE: Those last two commands should be on a single
line each (and 2 DLTs are required in the DLT drives obviously).
Modify the above for your situation.
Also note advice on
Using the DLT drives on vortex
Comments, inaccuracies and suggestions for improvement to Ingrid Stairs