2004 August 23


When the VLA is in its A configuration, it is possible that the Pie
Town (PT) antenna may be substitued for a single VLA antenna (Y1) in
order to schedule a concurrent VLA observation using the VLA-PT link.
In the coming months, the antenna substituted is likely to be one of
the two EVLA prototype antennas.  While the VLBA and (e)VLA antennas
are similar in many respects, there are some differences that could
compromise a VLBA dynamic observation made with just a blind
substitution.  In order to help ensure that your scientific goals can
be achieved by an array with Y1 substitued for PT, we would like to
bring your attention to the following:

                      Top 13 VLBA+Y1 Gotchas

1) The VLA antenna moves at about 1/2 the slew rates of the VLBA.
Some care may be necessary to ensure that sufficient time is allowed
for the VLA to get on important sources.  On the other hand, this will
cause dwell time scheduling to be rather inefficient for the rest of
the VLBA.  In the event that the full VLBA is allocated to a dynamic
project, then it might be preferable to have a seperate file without
Y1 in it.

2) VLA scans start on even 10 second intervals in IAT.  Therefore they
may start at a time several seconds offset from the VLBA antennas,
which go on the even second UT.  The VLA also takes longer to set up
on source.  These effects, combined with the slower slew rates, make
fast phase referencing (cycle times less than 2 minutes) problematic
with the VLA in the VLBI array.

3) The VLA is less flexible in how it is tuned.  Whereas the VLBA
allows for 8 BBCs to be spread across 500 MHz, the VLA constrains the
8 BBCs to fit inside 2 X 50 MHz windows spaced no more than ~350 MHz
apart.  Some instances of Doppler tracking of high velocity sources
can shift the observed frequency of a spectral line outside the 
VLA's windows.  The latest version of SCHED will warn about 
frequencies being outside tolerances.  If an EVLA antenna is being
used then xxx.

4) At 43 GHz, the poorer pointing of the Y1 antenna may prove
detrimental.  With a single VLA antenna it is not possible to perform
reference pointing corrections.

5) No pulse calibration system is available at the VLA.  If you plan
to use more than one BBC, then you should observe a strong and compact
source to line up the IFs.

6) The bandpass of the VLA is not as stable as for the VLBA,
especially at L-band (1.4 GHz).  If you are planning a spectral line
experiment you may want to observe a bandpass calibrator more
frequently than you would for the VLBA alone.

7) The Modcomp computers may start to misbehave if the observing file
provided is more than about 1000 lines long.  For example a phase
referencing experiment with a switching time of 5 minutes over 24
hours would generate 1728 lines and be in trouble.

8) If the observing run is scheduled with the full VLBA+Y1 (including
PT) then the data rate computations made by sched for these 11
antennas will be larger by 55/45 or 22% than the true data rate and
data volume for 10 stations.

9) The VLA has no W-band (86 GHz), S-band (2.4 GHz), or 50cm (0.61 GHz) 
receiver systems.  The EVLA prototype antenna may be missing xxx.

10) Some of the VLA receivers do not cover as wide a bandwidth as the
VLBA receivers.  This is especially true of the 15 GHz band where the
VLBA can reach as low as 12 GHz but the VLA is limited to the range
14.2 to 15.7 GHz.

11) It is dangerous to schedule a single VLA observing run for longer
than 24 sidereal hours.  This is because scans contain no LST day
identifier, so if the file must be restarted at any time after the
first LST day, it will start reobservations of the first day's scans.
The remedy is to supply seperate files for each day.

12) Automatic flagging and calibration transfer have not yet been
implemented for Y1 observations.  This is not a scheduling 'gotcha',
but observers should be aware of it.  Details of how to calibrate Y1
can be found in Memorandum 25 at

13) The calibration interval is set by the first (main) subarray.  In
rare instances during rapid source switching (e.g., phase referencing)
the calibration sampling may not be optimal.  Also the system
temperature measurements are time tagged with the end of the
integration, not the middle as is the case for the VLBA calibration
data.  If an EVLA antenna is used then Tsys values will be 
copied from the Los Alamos antenna.

Further information about VLBI observing with the VLA can be found
at www.aoc.nrao.edu/vla/vlbivla/current/html
You may also contact the undersigned.

                                              - Greg Taylor
                                                Joan Wrobel
                                                Craig Walker