Recent Changes - Search:
MtPleasant26m /

LBA Observing

Setting up for LBA observations:

Make sure you are part of the LBA chat on Mattermost. You can access that by pointing a browser at https://chat.atnf.csiro.au/. If you don't have a personal account you can login as hbobserver@gmail.com (usual observer password). Keep an eye on messages here. The other LBA observers at different stations are a good first port of call for problems at Hobart/Ceduna. Chris Phillips is the main LBA coordinator at the ATNF. Jamie Stevens also knows a lot about the systems here. There are also Mattermost apps for smartphones etc if you want to use them.

Preparing schedule files

Instructions on preparing LBA schedules for use at Hobart and Ceduna is contained here.

Hobart Setup Checklist:

Hobart is usually run through a VNC server on newsmerd (newsmerd:1). If this isn't up and running, setup and login instructions can be found on this page.

  1. Check the information email for the session should say which receiver is required for each experiment. If you need to change receiver, put the antenna in the park position and then select the appropriate receiver in Otter (for further instructions see here). Once it is on-axis, you can use the .prc file for the experiment to set the oscillators etc. If in any doubt you can Cross-check these values against the information for the experiment at ATNF LBA wiki. For details on calculating oscillator settings see here.
  2. Is the quadrature hybrid OUT (except for L-band where quad is IN), and is the receiver in "Direct" mode? Instructions on checking this are here .
  3. Start the field system (FS) if it's not running. From a terminal in the newsmerd VNC session, login to oper@hobart and type fs-mk5ho.
  4. Load the experiment procedure file with proc=[expnameho] (e.g. proc=vc149ho) in the FS input command line in the oprin window. Then set the oscillators and other critical configuration information with sched_initi. Check that the Agilent is set to the correct frequency and set it if not. Instructions are here. The scripts still set SML01 and SML02 for the DAS, but these aren't used for the DBBC. Then make sure that the DBBC configuration etc is correct by running setup01.
  5. Use the remote controlled backend ("Palfreyman's Delight") to check the polarisation selection & adjust the attenuation so that iread returns levels close to the target. Adjust the attenuation in the "DAS Attenuation" section of the folly display to change the power level into the DBBC if the automatic gain (See here if you're unsure of what this is doing). Startup and operating instructions for the configurable backend are given here.
  6. If all is well, start the schedule with the schedule=[expnameho],#1 (e.g. schedule=vc149ho,#1) command to send the telescope to the first source.
  7. We are using fs-mk5ho (which means dbbcho and flexbuffhb) for all the experiments. The flexbuff is pretending to be a Mark5B (very much like the Ceduna Mark5C, but fortunately a lot less touchy). However, it does mean that many of the normal mark5 commands don't work/return errors (e.g. disk_pos, checkmk5). We have tried to remove most of the offending commands from the snap and proc files, but there are still a few present, particularly when running sched_initi. You aren't likely to crash the flexbuffhb with an unsupported mark5 command (unlike Ceduna), but you will get an error. The other important thing to note about flexbuffhb compared to mk5ho is that when we change mark5 modes (normally change of experiment), we have to reconfigure the Fila10G. This is done with a python script which resides in the home directory for oper on hobart. This script takes two arguments, the bitmask and the mark5B mode. For example for the experiment v560b, if we look in /usr2/proc/v560bho.prc we find a line in the setup01 proc which reads mk5b_mode=ext,0x00f0000f,,32.000. The bitmask we need is the 8 digit hex number (0x00f0000f). For each of the f's in the butmask we have 2 BBCS and 128 Mbps data rate, so here the mode is Mark5B-256-4-2 (while 0xffff0000 would be Mark5B-512-8-2). So to configure the Fila10G for this experiment we run the command ./configure_fila10g.py --bitmask=0x00f0000f --mode=Mark5B-256-4-2 in an xterm on hobart. NOTE: make sure you copy the bitmask correctly, as if that is wrong we send the wrong bits from the DBBC to the recorder and likely get no useful data from the experiment.
  8. The fringe test data is written when there is a disk2file command in the snap file. Use list to look for these (or look through the snap file in a text editor). If the schedule has been drudged properly the disktofile entries will look something like disk2file=,/mnt/flexbuf0/fringtest/v560b_ho_no0001.m5a,02h02m49.00s,02h02m50.00s, where /mnt/flexbuf0/fringtest tells you the directory on flexbuffhb where the files are being dumped to (NOTE: this directory has to also be mounted on hovsi) It also tells you the time of the fringe check data, which will probably need to be reported on the LBA chat to the person making the fringe checks.
  9. If there is someone present at the observatory you can have them adjust the programmable attenuators (switches at the top of rack 2 with green LEDS) to set the Square Law Detector (a.k.a SLD, power meter) reading to -4 dBm in both channels. The SLD is in rack 3, underneath the rotary attenuator dials. The metal switch under the meter (labelled with L and R) controls which channel is monitored (this is nice, but not critical, it just means we don't get independent tsys measurements).
  10. Check pmSTALM is running and if it isn't, start it in a newsmerd terminal window. (pmSTALM is the script that manages the station alarms. Details here)
  11. Record experiment details to the appropriate ATNF wiki page. (Go here, click the experiment number link, scroll down to "observing comments for each antenna" section and click Ho to update the Hobart log and Cd to update the Ceduna log.) The login username and password can be found here.

Ceduna Remote Setup Checklist:

Ceduna is run through a VNC server on pcfscd (pcfscd:0)

  1. The field system (FS) should be running. If it isn't, start it by typing fs in a pcfscd terminal window.
  2. Load the experiment procedure file with proc=[expnamecd] (e.g. proc=vc149cd) in the FS input command line in the "operator input" window. Then set the oscillators and other critical configuration information with sched_initi. Check that the Agilent is set to the correct frequency and set it if not using oscillator_gui.sh on pcfscd The scripts still set SML01 and SML02 for the DAS, but these aren't used for the DBBC. Then make sure that the DBBC configuration etc is correct by running setup01.
  3. Make sure that iread and bread give reasonable values and if not, use Lovell's Folly to adjust the attenuation.
  4. We are using dbbccd and mk5cd for all the LBA experiments. mk5cd is a Mark5C pretending to be a Mark5B - frankly its a horrible recorder and we're going to get rid of it soon, but its what is there at present. Many of the normal mark5 commands don't work/return errors and also potentially crash the Mark5C (it really is a hunk of junk). We have removed the offending mark5 commands from the snap and proc files, but basically apart from basic things like disk_pos and checking the bank in use, don't issue any additional mark5 commands unless you know what you're doing, or are desparate. When we change mark5 modes (normally change of experiment), we often have to reconfigure the Fila10G. This is done with a perl script which resides in the home directory for oper on pcfscd (NOTE: This is similar to, but different from the python script at Hobart). This script takes one argument, the name of the experiment. For example for the experiment v560b, to configure the Fila10 we run the command ./configure_fila10g.p --experiment=v560b in an xterm on hobart. It takes about 30-60 seconds for this script to run, the times reported at the end should be within a few hundredths of a second of each other. If not, run it again. If they are still very different seek help.
  5. If all is well, start the schedule with the schedule=[expnamecd],#1 (e.g. schedule=vc149cd,#1) command to send the telescope to the first source.
  6. The fringe test data has to be grabbed manually. To do this there is another perl script which needs to be run from the home directory of oper on pcfscd. This script takes the experiment name as an argument, for example ./fringe_recorder.pl --experiment=v560d. This script replaces the previous bash script and does a few additional steps, including halting and then restarting the schedule. NOTE: after halting the schedule the script issues a disk_record=off command to mk5cd which will give an error if there isn't a recording happening. This error is benign though and can be ignored. The perl script will also return a time - this is the time of the data for the fringe check and should be reported in the LBA chat so that whoever is doing the fringe checking knows what time to grab the data for from the other antennas.
  7. The fringe test data is written to /data/internal/fringtest on cdvsi. If mk5cd crashes or has to be rebooted, then it is necessary to manually remount /data/internal on mk5cd so that it can write the fringe test files out to the right place. To check if the disk is mounted ssh oper@mk5cd and then df -k . If you don't see a line which starts with cdvsi:/data/internal, then you need to remount that disk. To do that su on mk5cd (usual root password for observatory machines), then mount -t nfs cdvsi:/data/internal /mnt/cdvsi once that is done cd /mnt/cdvsi/fringtest and check that you can see files there. You can now logout from the root account.
  8. Record experiment details to the appropriate ATNF wiki page. (Go here, click the experiment number link, scroll down to "observing comments for each antenna" section and click Ho to update the Hobart log and Cd to update the Ceduna log.) The login username and password can be found here.

Monitoring LBA Experiments:

These web pages should always be open:
ATNF VLBI monitoring page
ATNF recorder monitoring page
Local Mt Pleasant live page
Local Ceduna live page

Hobart Monitoring Checklist

Check every 2-3 hours:

  1. Is the dbbc running OK? dbbc=pps_delay should return 61971, if it doesn't then the dbbc probably needs a restart
  2. Are the levels OK (iread and bread)
  3. Is the telescope on source? (type "onsource" in the oprin window of FS)
  4. Is the weather being logged? (type "wth" in the oprin window of FS)
  5. Is the clock offset being measured? (type "clkoff" in the oprin window of FS)
  6. Is the Tsys sensible? (scroll through the FS log)

Check every 2-3 hours:

  1. Is the dbbc running OK? dbbc=pps_delay should return 61959, if it doesn't then the dbbc probably needs a restart
  2. Are the levels OK (iread and bread)
  3. Is the telescope on source? (type "onsource" in the oprin window of FS)
  4. Is the clock offset being measured? (type "clkoff" in the oprin window of FS)
  5. Is the Tsys sensible? (scroll through the FS log)

Alarms:

If you lose the Ceduna VNC window, restart in a newsmerd terminal window with the following command: vncviewer pcfscd
The password can be found here.

Rebooting Ceduna drive PC

If you're getting "Error: Cannot get monitor info from antenna (8020002)" or the like, you'll need to reboot the drive pc (sys30m).

In a terminal run 'telnet lerips', the password is the usual admin.

Then '/Boot 1', follow by 'y' when prompted. Sys30m will take several minutes to come back.

After sys30m starts responding (eg. to 'ping') enter 'source=disable' in the operator input. If this doesn't work try reopening the field system.