Performing a test scan across a source (preferably a flux density
calibrator) is the best check we have that the receiver is in the
correct focus position and that it is operating correctly (the system
temperature is as expected). It is important this check be performed
each time a different receiver is used, either prior to the start of
the experiment, or at a suitable time soon after the start. It
will take about 10–20 minutes to do a test scan once the system is
setup ready for the VLBI experiment. Don’t worry about missing part
of the experiment, if the receiver is in the wrong position we get no
useful data at all.
The instructions below will work for all LBA/VSOP style experiments
with the S2 system. For IVS or Syowa experiments it will be necessary
to cable up the signal from the X-band receiver as if for an S2
experiment (its not necessary to do both frequencies). In general the
observatory manager will have checked the receiver position for an
IVS/Syowa experiment during the day prior to the experiment. However,
if you have to do this take the signal from the X-band system through
an amplifier and attenuator and then to the frequency translator.
Select a DAS profile of MP16S_F.PRO and set the SMX to 807.0 MHz.
Don’t worry about feeding a signal into IF#2 on the DAS.
How to do a calibration/test scan with bruce
- Make sure that the VLBI backend is setup to the right frequencies, you have coherence etc.
- Make sure that the power level for the system temperature measurements is OK (adjust the programmable attenuators until the level is −4, but do this with the antenna at the park position, not in service).
- Make sure that the cal is in remote and controllable by sam26m/sam30m (if in doubt issue the following command on smerd or mills as observer
/obs/bin/linux/calu -m sam26m -r rakbus). At Ceduna the cal will always be controllable by bruce.
- Login to newsmerd as observer
- Create two xterms, then stretch one of them so that it is at least
120×50 (otherwise bruce will crash when you start it). Using the command
xterm -geometry 120×50 & is an easy way to ensure this is correct.
- If there is a VLBI experiment schedule already running we need to pause it to stop bruce and it interfering with each other. Before starting bruce, go to the pcfs and enter the command
halt in the oprin window.
- To start bruce type the command
bruce -f 4 in the large xterm. When prompted for the names of the rakbus, sampler and antenna, the defaults are all correct for Hobart, so you can just hit enter, at Ceduna the machines are rakbus-ced, sam30m and sys30m respectively.
- In the other xterm, type the command
flux_cal.pl —freq=<freq> where
<freq> is the frequency of the observations in GHz. This will give you a list of flux density calibrator sources, pick the strongest one that is at a reasonable elevation (closest RA to the current sidereal time). At 22 GHz there are few good sources and you may be better to look in the file
/u/observer/cats/cal_22ghz. Planets such as Venus or Jupiter are generally the best targets - current planet positions and flux densities can be estimated through
MIRIAD by the tasks
MIRIAD is installed on ares.
- Enter the name and the coordinates of the source you have selected into bruce using the commands:
source <source name>
left <right ascension>
coordinate mode ‹b1950|j2000|date›
- If necessary switch the antenna drives on with the command
drvon and then slew to the source position using the command
- While the antenna is slewing you can set up the other bruce parameters :
- If you are observing dual polarizations, or two frequencies (S/X receiver) then enter the command
number channels 2.
- Set the length of the scan using the command
length <degrees>. The default of 1 degree is OK for most frequencies, however, for L-band observations you probably want 2 degrees and for 22 GHz you probably only need 0.5 degrees.
- Set the rate of the scan using the command
rate <degrees/min>, I normally set this to be the same as the scan length (i.e. the scan will nominally take a minute).
- If you are observing at 22 GHz issue the command
sample spacing 0.005, the default is OK for all other frequencies.
- It’s a good idea to set the frequency of the observations etc, although these parameters are only for information purposes and don’t effect the observations at all. Set the frequency to the center frequency of the VLBI observations
frequency 1 <freq in MHz>, set the bandwidth to 16 MHz
bandwidth 1 16.
- Once these parameters are set type the command
tsys. This should return sensible numbers which are similar if you repeat the command 2 or 3 times. If you get NaNs then the most likely problem is the signal levels into the VFCs, or the cal control. Go back and check the first few steps in this list again.
- Often the information in the bruce display disappears before you’ve had a chance to read it (this will be important when scanning). In the xterm where you issued the
cd /data/scratch/bruce and then
tail −100f bruce_yyddd.log where
yy is the last two digits of the current year and
ddd is the day of year (universal time). This will display the bruce log in the xterm where you can scroll back and look at any output that you missed.
- Once the tsys command is returning reasonable number and the antenna has slewed to the source you are ready to scan. Issue the command
- When the scan finishes you should get an average plot of the scan on the PGPLOT display with a Gaussian fitted to it and a scale in degrees on the x-axis. Bruce will report the offset of the centre of the Gaussian fit in its little display window. If you have a good quality scan and the Gaussian has been fitted correctly then the offset should be less than about 0.015 degrees (1 arcminute). If it isn’t then repeat the scan command to check, and check the focus
position of the receiver. In generally the offsets will be larger for sources at lower elevations.
- If the first scan (a
direction left scan) is OK, then do a scan in the orthogonal direction by issuing the commands
direction right and then
scan. Again the offset of the position should be less than 0.015 degrees.
- Once you have checked that the position offsets are OK then
quit from bruce.
- If the experiment is an IVS one using the MKIII or MKV, set the cal to be controllable by the pcfs by using the command
/obs/bin/linux/calu -m pcfs -r rakbus).
- If the VLBI schedule is halted then issue the command
cont in the oprin window on the pcfs. If the schedule is running then it may take it a few minutes to sort itself out after the pause.
- Make a comment in the pcfs log reporting on your check of the receiver position and stating why the schedule was halted (if it was).
Checking the system temperature
If the source you checked the position with was one from the
flux_cal.pl list then you can (and should) use the observations you just made to roughly check the system temperature from the information recorded in the bruce log, as follows :
- When bruce fits the Gaussian it lists the amplitude in VFC units. In the bruce log, find the amplitude of the left scan and divide it by the cal height in VFC units for IF#1 (this will generally give a number much less than one).
- Take the result of the previous step and divide the system temperature of IF#1 in units of the cal by it (this will generally give a number of the order of 10s or 100s).
- Multiply the result of the previous step by the flux density of the source you observed, at the frequency you observed as given by
flux_cal.pl (use the Ott et al. numbers in preference to Baars et al.)
- Finally look at the scan in the orthogonal direction and divide by the correction factor it gives. This is an estimate of the system temperature in Janskys.
- Repeat the process with the right scan and take an average of the two. Compare these numbers with those listed on the focus position sheet. They should agree to within 10% of each other (except perhaps at 22 GHz), unless there are other extenuating circumstances, such as the receiver being warm rather than cooled etc.
- Make a comment in the pcfs log reporting your estimate of the system temperature.