The instructions below are the same as those written by Simon for remote VLBI observations, but information not relevant to CHI observations has been removed
Setup and instructions
The observing is done from a vnc session running from the vlbi account on cdvsi (password the same as the vlbi account). Its probably best to make it display 1 so that its easy for everyone to find. If not, let people know what it is, or you can always
ps -ef | grep vnc on cdvsi to find it. If you are connecting to the remote VLBI observing VNC session, make sure your client is setup so that it allows others to also connect.
This is a work in progress. If you have additional suggestions or improvements, please edit the page and add them.
Recommended VLBI setup for Ceduna
Current usage of cables from UER:
- L1 - has RCP signal from receiver.
- L6 - has LCP signal from receiver.
- L2 - is connected to receiver tone port.
- L5 - has the cal control signal.
- Line behind DSO - has 5 MHz from the maser.
Dual Polarization setup
- Signal from L1 goes to LHS 30 dB amplifier;
- LHS 30 dB amplifier connected to input of LHS dial attenuators;
- Output of LHS dial attenuators connected to RF connector on LHS of frequency translator;
- SMY01 (middle SMY?) goes to input of splitter and the two outputs go to the LHS and RHS LO connectors of frequency translator;
- IF connector on LHS frequency translator does to LHS programmable attenuator (doesn’t matter which end)
- other end of LHS programmable attenuator is connected to RF IN connector on LHS of square law detector.
- DC OUT on LHS of square law detector is connected to VFC #2 input
The setup for the other polarization is the same, except the signal comes from L6 and replace all occurrences of LHS with RHS and the square law detector DC OUT goes to VFC #3
Single Polarization setup (Not used for CHI)
This is very similar to the dual polarization setup, the only differences are:
- Take the polarization to be observed and put it in the input to a splitter. So if it is LCP we want to observer, then take the signal from L6 and put it into the input of a splitter.
- Take the two splitter outputs and put them into the inputs of the LHS and RHS 30 dB amplifiers
- Instead of using one SMY and splitting the signal you will need to use two SMYs and put the signal from one directly into the LO connector on the LHS of the frequency translator and the other into the LO connector on the RHS of the frequency translator.
NOTE: Don’t use the SMY connected to the tone port for the second LO signal. The SMY not used in the dual polarization setup (the bottom SMY?) is either not connected, or doesn’t work on the GPIB, so that will have to be set manually by Bev. I would recommend checking that that oscillator is working while Bev is present (do a coherence check), since we use it in-frequently and its possible it may be faulty.
Setting the levels
- With the DAS running and an appropriate profile loaded, make sure that there is no CLOCK error on the DAS (this happens when the Ceduna tick phase changes, which happens when the station clock is restarted). If there is a CLOCK error the DELAY pots on the DAS High-resolution samplers will need to be manually rotated to get the right setting.
- Adjust the dial attenuators until the level on both DAS channels is as close to the centre of the range as possible.
- Adjust the programmable attenuators so that the square law detector is within its operating range (marked on the dial meter).
Record the setup
For all sorts of reasons it is critically important that you accurately record the setup used for the experiment. This information should be put in two places :
- The telescope log information on the experiment wiki
- In the PCFS log as a comment.
The information which should be recorded as a minimum is:
- The agilent/front-end LO frequency and level.
- The SMY/back-end LO (or LOs) frequency and level.
- The tone used to check coherence (and where it was observed - bandsplitter port, fine tuner port etc)
- The DAS profile and the cable connecting the DAS to the vsib box (BG1/straight through or BG3/64 MHz etc).
- The disk the data is being recorded to (make sure to also update the disk summary table).
- The approximate clock offset (as reported by clkoff).
- Which polarization is being sent to which DAS processor.
Timed commands on the PCFS
Its a good idea to make sure that certain information (clock offsets, system temperature, weather etc.) are monitored regularly throughout an experiment. For some experiments some, or all of these checks may be scheduled in (as part of a preob or midob procedure say). Either before the experiment, or soon after it has started look at what checks are being performed regularly by the schedule and add any which are not using a timed command.
- To check what is part of the schedule go to the directory
/usr2/proc on ceduna and look in the .prc file for the experiment (e.g. for the experiment v275c this would be the file @@/usr2/proc/v275ccd.prc). Look at the procedures preob, midob and postob (if they are defined) and see what commands are being issued. These commands will be executed once per scan and so if the experiment has short scans (a few-10 minutes), then you don’t need to worry about those commands.
- For commands not in the pre/mid/post ob procedures, or for experiments with long (> 15 minute) scans, you can set the pcfs to issue timed commands using
This will execute clkoff now and every 15 minutes into the future (when you change schedules many of the timed commands get canceled). To see which timed commands are in the system
If you want to cancel a timed command
The following process should be running in the cdvsi vnc session during a VLBI session :
PCFS monitor (ceduna)
STALM (ceduna) - haven’t got this working remotely at present, and since you can’t hear alarms sound in Ceduna when you aren’t there it probably doesn’t matter.
pmSTALM.sh (your local machine)
This is also the appropriate place to run
bruce calibration observations or change oscillator frequencies/test coherence with
To start the vnc session on cdvsi (if necessary)
- Login to cdvsi as vlbi
vncserver -geometry 1024×768 :1
The window manager is pretty primitive, but it works. To get an xterm left click in the background and select xterm from the menu.
To access the vnc session
- On a Linux machine (e.g. newsmerd or cdvsi)
vncviewer -shared cdvsi:1 You will need to enter the vlbi@cdvsi password to see it.
- On a graphical VNC client (e.g. Chicken of the VNC on a Mac), just give the machine name as cdvsi.phys.utas.edu.au and the display as 1 (assuming that the vncserver has been started on display 1). Make sure that you allow other users to also connect.
To run the
ssh -l oper ceduna
To start the pcfs status monitor
From an xterm on ceduna (as oper)
xterm -name monit2 -e /usr2/fs/bin/monit2 &
To run the DAS
- From an xterm on ceduna (as oper)
# When the xdos window starts it will automatically run the DOS DAS program. When it starts, select vsop.pro as the profile initially (that has the right communications options). After the initial setup, then select the profile you want.
To run cdisko
From an xterm on sille (as observer)
NOTE: There is currently a bug in cdisko (version 2.5) where it does not correctly report disk sizes > 2 TB. So if you have a 3.5 TB disk pack mounted it will only show 2 TB available. It will remain showing this much available until the space drops below 2 TB. If in doubt about the remaining space use
df -k on cdvsi to look at available space. It also bases the amount of recording time available on this disk space reporting. A 3.5 TB disk lasts about 24 hours (slightly more) recording at 256 Mbps.
To run pmSTALM.sh
See the instructions here. pmSTALM.sh will play an audible alarm from your machine if significant errors occur in the fs.
To change oscillator frequencies etc
- From an xterm on ceduna (as oper)
Take care to exit this nicely, as its easy to get the GPIB bus in a knot which requires rebooting Ceduna to fix
To check coherence:
The coherence checking assumes that:
- You have the IF monitor port outputs connected to programmable attenuators, then to the square law detectors and then to VFCs 2 & 3. This is the normal setup for system temperature measurement, as is described in the setup section above.
- SMY02 is set to produce a tone 500 Hz (not 5 MHz) above the lower-band edge for the current DAS profile setup.
- SMY02 is connected to the tone port on the receiver.
- From an xterm on ceduna (as oper) type the command
coherence. Make sure that the system temperature isn’t being measure at the time, since only one user can access the VFCs at a time.
coherence is an alias which
- Switches the output of SMY02 on
- Collects 3 seconds worth of data from VFCs# 2 & 3, sampling at 4000 Hz
- Switches the output of SMY02 off (the tone is on for a total of about 5 seconds, so try and do this when the antenna is not collecting good data (e.g. during a slew).
- Runs an octave program which Fourier transforms the data and plots the amplitude of the FT against frequency. NOTE: The octave script is quite slow and to see the plots you need to have the display set properly (
ssh -X oper@ceduna will usually do the trick). Once the first plot appears, hit return to see the second which zooms in on the region around 500 Hz and also shows the difference between the two channels. Once you’ve looked at this, hit return again (perhaps twice) and the plot should disappear (eventually).
- Should you ever want to look at it again the zoomed version of the plot is also saved to /usr2/log/coherence.XXXXX.YYY.ps where XXXXX is the modified Julian Day and .YYY is the decimal fraction of a day.
When we check coherence we inject a tone, sample the output from the IF monitor ports and then Fourier transform that to find the tone. Normally we take the output of the DAS bandsplitter or fine tuner ports compare them with 5 MHz from the maser on an oscilloscope, however, that isn’t possible for remote observation. Also, the signals from the bandsplitter and fine tuner ports are at baseband and go through the DAS’s automatic gain control system. This means that they aren’t good for system temperature measurements. The IF monitor port outputs (which we normally use for system temperature measurement) are centred at 96 MHz (whatever sky frequency is picked by the LO input into the frequency translator ends up at 96 MHz on the IF output). This means that if you are observing with VSOP profile with 8425 MHz centre frequency you have 2 x 16 MHz bands 8409–8425 MHz and 8425–8441 MHz (the first is output on the bandsplitter 0–16 MHz) and the second on the fine tuner (0–16 MHz). A tone at 8409.0005 Hz will appear 500 Hz above the lower band edge (i.e. at 80.0005 MHz on the IF monitor port). When we sample this broadband signal at 4000 Hz, we still see at tone at 500 Hz, due to aliasing in our massively undersampled system. Although it is significantly weaker than the signal you get if you use the output from the bandsplitter ports, the advantage is that when the tone is switched off (which is nearly all the time), you can use it for system temperature measurements. If you want to demonstrate that there is no tone at 500 Hz with when its switched off run the
octave ~/prog/coherence.m programs by hand with the tone off (
alias | grep coherence will show you the sequence of commands and their syntax which is use for the coherence checking process). One problem discovered in July 09 is that high SMY02 output levels (10+ dBm) will not show coherence with this process, at least for the 8.4 GHz receiver. Levels of 8 dBm are ok.
To run bruce
- Start an xterm on ellis (as observer)
/obs/develop/sol/mapping/bruce -f 4 -r rakbus -s sam30m -a sys30m