Very Long Baseline Interferometry is radio interferometry where the radio telescopes involved are not physically connected. Instead of this the signals are converted to video signals and recorded to high bandwidth tape. These tapes are brought to a Correlator centre (1) where they are played back and correlated. VLBI is used where the separations between the telescopes which make up the array are too large for cable, waveguide or radio links to be practical. The WSRT normally takes part in VLBI sessions as a member of the European VLBI Network (EVN) although ad-hoc experiments also occur. More information on the operations of the EVN can be found in The EVN hanbook. One way to get this handbook is via the VLBINFO account in Bologna. See See section The VLBINFO account.
At each observatory in the EVN there is a Friend of VLBI whose responsibility it is to make sure that the VLBI observations at that telescope are done according to the schedule. The `Friend' is usually assisted by a Technical Friend who takes care of the technical aspects. The names of the `Friends' can be found on the VLBINFO account. See See section The VLBINFO account
There are various operational requirements intrinsic to VLBI observations. These are basically the same as those needed for connected element interferometry.
Observing the same source is handled at the scheduling phase. Schedule production is a responsibility of the Principle Invesigator (P.I.) and the `Friend of VLBI' must check that the Schedules are put into the telescope control computer accurately.
Our maser's UT timescale must agree with that of the other observatories. The Global Positioning System (GPS) timing receiver acts as a check on the WSRT maser's offset and drift. Larger bandwidths demand increased accuracy with which we must know the offset (accuracy must be better than reciprocal bandwidth). Other Westerbork clocks are derived from the maser clock signals.
The frequency depends on all the LOs in our mixing chain being correct and all being in lock with the maser. These must all be checked in advance. This is most easily done by broadcasting into the band (also known as the `Instraal test'). See See section Instraal test
Normally VLBI is observed in circular polarization. Most telescopes have altazimuth mounts and so the plane of linear polarization appears to rotate as the parallactic angle of the source changes. For circular polarization the effect is only a phase shift. For most wavelength bands VLBI uses Left Circular Polarization (LCP) but for 3.6/13cm Right Circular Polarization (RCP) is used because of a pre-existing geodesy standard. The WSRT telescopes have linearly polarized feeds, called `X' and `Y' (see the WSRT User Documentation ). Current VLBI procedures assume all telescopes to have circular feeds, so the linear feeds have to be combined with the appropriate levels and 90degree phase shift to make circular polarization. This is done in the `IF/Pol select box'.
The telescope limits are the same as for normal observing: from -90 to +90deg in Hour Angle and from 90deg to 5 deg in elevation. See the WSRT User Documentation. Just as in normal observing shadowing can cause problems for sources at low declinations when observed at extreme hour angles.
The current agreement is that the WSRT will provide 3 cooled receivers at 6,18 or 21cm and 9 uncooled receivers at 49 and 92cm. If the WSRT happens to be observing at the VLBI wavelength anyway then more will be provided. Receiver changes cost a lot of time, with the exception of the 6/49cm uncooled dual frequency frontends, so operationally it is not possible to change the array wavelength in less than 1 week.
Some P.I.s want the local WSRT correlator output for further analysis; this can be useful for combined WSRT-MERLIN-VLBI mapping. This can be had by warning the `Friend of VLBI' in advance of the observations.
About 8 weeks before an observing session the EVN and NRAO schedulers agree on a draft block schedule for an observing session. This shows the dates and (UT) times allocated to observations and allows for errors to be found and corrected. There is also a possibility for small gaps to be allocated on a filler basis if tapes and correlator time are available. The block schedule is distributed to observatories and a final version is frozen about 4 weeks before the session begins.
It is the P.I's responsibility to prepare and distribute VLBI schedules to all observatories. Usually schedules are made using the program PC-SCHED. This is available on ftp from Haystack. This program comes with considerable online help and is documented. The data files (*.SCH) contain information on standard receiver setups, recorder setups, telescope positions, slew rates, and horizons, and source positions. There is an alternative program (SCHED) available from the NRAO. This can be picked up directly from the NRAO. It runs under various operating systems, but is not as user-friendly as PC-SCHED. Geodesists use yet another program (SKED) for schedule preparation. In future there will also be a program called GNOMES, produced by the NRAO, which will also prepare and check VLBI schedules. The standard format for the output of all of these is called SNAP, and all EVN observatories can read this, and convert it into their own telescope control format. See See section SNAP commands.
This SNAP is an ASCII file which must be converted by the WSRT local
program called STFMA for observation (using a default setup) and is also
used by the field system computer to operate the Mk3 rack and tape
unit. As well as the SNAP schedule the P.I should provide a
human-readable listing, a cover letter, a tape barcode label file,
and a procedure file.
The procedure file is a series of observing subroutines in SNAP, but it
is not needed for standard setups,
because station defaults are available. Inexpert users are advised
NOT to send procedure files that they do not understand.
The program PC-SCHED is an interactive VLBI scheduling program that runs on a DOS PC and was created by Alan Rogers at Haystack. It can create Mk2 and Mk3 schedules and there is a handbook available, as well as good online help. The program can be picked up from Haystack by anonymous ftp.
To run PC-SCHED on the Novell network in Dwingeloo, use the following dialogue:
C:\> F:
F:\login> login guest
F:\home\guest> cd pc-sched
F:\home\guest\pc-sched> pc-sched
F:\home\guest\pc-sched> logout
All schedules for EVN experiments are deposited in an account in Bologna. They must be placed there 2 weeks before the session begins and are the responsibility of the P.I.. It is possible for P.I.s to apply to JIVE (the Joint Institute for VLBI in Europe) for assistance in preparing their schedules. The schedules are put into a subdirectory named after the first 3 letters of the month and last 2 digits of the year (e.g. `nov94' ). Stations information is identified with a 2 letter code, so Westerbork is "wb". File transfer is by the program ftp, and the dialogue should go as follows:
ftp astbo1.bo.cnr.it
login: evn
password:
cd monthyear
prompt
mget *.wb
mget *.cvl
mget *.drg
quit
The name convention is
When the schedules are picked up they need to be checked. If there is anything wrong, or even suspicious the VLBI Friend must contact the P.I.. It is a P.I.'s responsibility to get a new schedule made and transfer it. Procedure files for standard setups are a local responsibility.
One common problem is that PC-SCHED automatically inserts a
STOP command if there is a long wait. At the WSRT these STOPs
must be removed in order to avoid problems with the field
system
The schedule files are transferred to a directory `/disk3/obs/vlbi/' in
the workstation `wsrt00' and put into subdirectories per experiment.
The SNAP format
is converted by the Westerbork observation specification program
STFMA into a schedule that can be observed by
Westerbork. This process needs a default setup, part of which is
wavelength dependent. The files are renamed to remove the `.wb'
common to them and given extensions by their contents ( so `gl15bsna.wb'
becomes `gl15b.snp' ) and the procedure files are edited to reflect
actual settings that will be used. When this is done (and checked)
the files are transferred to the field system computer (`wsrtfs') into the
directories `/usr2/sched' (for schedules) and `/usr2/proc'
(for procedures). If necessary they can be edited by any
normal text editor.
Procedure files can also be edited by the program pfmed during
operations. See the Mark IV Field System Documentation for more
details. With the Mk4 upgrade the field system software will be
upgraded also.
The field system is started by logging into the field system as
operator and typing fs. It will take a few seconds to initialize
all VLBI devices.
Then when it is running give the schedule name with:
schedule= name of schedule
For the observing system the observations should be started up with the normal starting procedure. When operating the field system will
It does not however know what the telescopes are actually doing.
The schedule operates using SNAP commands. See See section SNAP commands. These commands can also be grouped together into SNAP procedures in which case they are put into a procedure file and the procedure is given a name. This name can then be called in the schedule as needed.
In addition to the total powers for the backends needing to be correct, we also have total power levels in the Mk3 equipment: ASTI, IF/Pol select, Mk3 IFs, Video Converters (VCs). These all have levels marked with a red dot, except for ASTI and VCs, where "1V" is the nominal value. To get accurate circular polarization there is a ASTI setup procedure produced by Jaap Bregman. See See section ASTI Total Power setup.
The way that observations are prepared is by way of default observations. See section Default settings for tied-array VLBI. These are already provided for the Mk3 standard setups. The setups agreed by the EVN as 'standard' are summarized in the VLBINFO account in Bologna. See See section The VLBINFO account. Any nonstandard observations must be have their control codes and mixing schemes provided by the Tony Foley, in consultation with Albert Jan Boonstra. See See section WSRT Mixing schemes.
When VLBI observing is done there is a pre-session checklist provided to avoid some obvious mistakes. This can also be compared with the EVN checklist. During the session there is also a checklist for Waarnemers. See See section Checklists for a VLBI session.
See the interface of the tied-array to the Mark 3 rack