Dear PI,

This email is to provide you with some more details of progress with
the CABB system.

Observing for the 2009 APR semester started on schedule on Wednesday
22 April. At this stage there is no reason to expect any changes in
the schedule for the rest of this semester. (Version 2 of the ATCA
schedule was released on 20 April, with some rescheduling of a small
number of observations in June.)

The overhauling of the ATCA User Guide is continuing, with the new
version, to be available in the next week or so, containing detailed
instructions on preparing schedule files and observing in the CABB
era. The MIRIAD manual will also be updated in the near future to
describe the changes made for CABB.

A new scheduling tool has been produced to make schedule files for
CABB. Pre-CABB schedule files will no longer work, however, old
schedule files can be read into the scheduling tool and written out in
the new format.  The new scheduler is web-based, and can be accessed
via a link from http://www.narrabri.atnf.csiro.au/observing/sched/. If
accessing from outside the ATNF network you will need to use your ATNF
unix account details to access it. Features of the new scheduler
include the ability to search for nearby calibrators and inspect the
calibrator catalogue information for potential calibrators of
interest, and the ability to use the velocity calculator for selecting
zoom bands (as they become available). Files are saved directly to the
sched file area on xbones. Not all functionality of the command line
ATCASCHED has been reproduced, but experience with the new tool has
shown it to be easy to learn.

The previous CACAL routine no longer exists. Calibrating delays,
phases and amplitudes is currently conducted as three separate steps,
which are issued as commands within CAOBS (or alternatively, CACOR).
VIS continues to be used to inspect delays, phases and amplitudes of
data. A new CABB version of SPD is used to monitor the delays and
amplitudes.  As CABB always produces autocorrelation spectra, these
can also be viewed with SPD and are a useful diagnostic tool. The 2GHz
CABB bandwidths show some amplitude variation across the band due to
the characteristics of the receivers, which can be dealt with by
bandpass calibration during data processing. The much wider bandwidths
of CABB are accompanied by an increased likelihood of RFI. The
harmonics of 128MHz produced by the old samplers are no longer seen in
spectra, but there are a number of other features, some
self-generated, other external, which now appear. In some cases, care
is required to select channel ranges (with tvchan) or to flag out bad
channels to enable delays to be calibrated correctly.

CABB provides 2048 channels across 2.048 GHz, though not all channels
are useful. As detailed on the CABB webpage, the top and bottom 32
channels are affected by analogue filter roll-off and should generally
not be used.  The delay calibration, by default, uses the central 50%
of the CABB band, though this can be adjusted by the user.  Channels
513, 1025 and 1537 are always flagged as the digitiser produces quite
large birdies in these channels due to the interleaved sampling
process.  At present, in the second IF band, we have two (of 16) CABB
signal processing boards removed for investigation, and so this band
is currently missing an additional 256 channels. These are not
contiguous, but are concentrated in four areas of the band. The small
loss in sensitivity will not seriously affect continuum observations:
spectral line observers planning to use the second IF should contact
me to discuss their particular observing plans. We expect to have the
boards repaired or replaced within a month or so.

CAMON is no longer used to monitor the status of observations, with
this functionality taken over by CACOR (for system temperatures and
CABB sampler statistics etc) and MONICA (for updates to pointing
parameters etc). ASSISTANCE is currently still being used to monitor
visibility data for calibrator scans and selected monitor points.

Experience is still being gained with CABB data reduction. It is
apparent that the much wider bandwidths at 6cm and 3cm in particular,
with significant change in the primary beam between top and bottom of
the band, will require careful consideration in data reduction for
extended sources. At present, the best approach within MIRIAD is to
divide the data into smaller chunks, e.g. 128 MHz, and reduce each
separately. This is less of an issue at shorter wavelengths as the
fractional bandwidth is appreciably smaller.
 
All in all, CABB is working very well, which is credit to the long
hours devoted to developing, installing, testing and commissioning the
CABB system by many ATNF staff.  We will continue to develop our
understanding of observing with greatly increased bandwidths and in
reducing the resulting data, and expect that our documentation (User
Guide, MIRIAD manual, webpages) will be updated regularly in coming
months. We recommend observers arrive a day or two earlier, and leave
a day or two later, to become familiar with the new system, and
encourage feedback on all aspects of observing and data reduction so
we can make improvements wherever possible, and make the most accurate
documentation of the system available for all users.

Regards,
Phil

P.S. The call for proposals for the ATCA 2009JUL semester (15 July to
30 Sep) is now available at
http://www.atnf.csiro.au/observers/apply/avail.html