The Plateau de Bure Interferometer (PdBI) is currently undergoing two major upgrades. A 6th antenna is being built and is scheduled to be fully operationnal by the end of 1999, while the Northern track will be extended from 232 meters (station N29) to 416 meters (N46) within the same timescale. This will provide higher sensitivity, higher observing efficiency, higher angular resolution, and more circular beams. To take full advantage of those improvements, new configurations of the instrument are needed, and are currently under investigations. This document provides informations about the improvements that will result from the upgrades and describes likely new configurations of the interferometer. In the last part, we compare the relative merits of extending the Northern track even further (up to N52) or of doubling the length of the Eastern track (from E24 to E48).
Currently, the antennas can be arranged in 6 primary configurations (D, C1, C2, B1, B2 and A1 - http: iram.fr/PDBI/bure/bure.html) that can be combined together to yield four angular resolutions ranging from 6'' (at 100 GHz for configuration D) to 1.3'' (at 100 GHz for configuration A). With the upgrades, a similar set of four angular resolutions will be kept, but will probably require less primary configurations. Furthermore, the 6th antenna will provide a better sensitivity to the most compact configuration, making this configuration even more adequate for detection experiments.
Configuration D: With 6 antennas, the most compact configuration
will become ``W05 W08 E03 N02 N07 N11'', which provides 5'' resolution
at 100 GHz instead of 6'' for the previous D configuration. This slightly
extended configuration was chosen because for more compact situations,
heavy shadowing occurs. The main improvement will be a 22 
increase in
sensitivity.
Figure 2: Long baseline configuration with North track extension and 6 antennas.
Contours are -20, -10, -5, 5, 10 to 50 % by 10%.
Configuration C: For the next most compact configuration (C) we suggest ``W09 W12 E03 E10 N03 N13'', which provides about 3.5'' resolution at 100 GHz, or 1.5'' at 230 GHz. This configuration contains a few short baselines, and provides a fairly complete coverage of the UV plane. It is therefore likely that - thanks to the 6th antenna - the C configuration will require only one primary configuration, instead of two previously.
In selecting this configuration however, the balance between the number of short baselines with respect to longer ones, and the possible complementarity between the C and D configurations argue in opposite directions. The community is invited to present arguments on the relative interests of ``fast'' imaging (single configuration) versus low sidelobes (D + a more extended C' configuration).
Configuration B: The B type configurations provide a resolution of about 2'' at 100 GHz and is intermediate between the compact (D and C) configurations discussed above and the extended (A) configuration below. We are searching for a single configuration which would provide proper UV coverage when combined with the C configuration, and which could still be combined with A type configurations for high resolution mapping of moderately extended objects.
Configuration A: This most extended configuration will make use both of the increase in the number of antennas and of the extension of the Northern track. While the 6th antenna will obviously improve the sensitivity, the extension of the Northern track will have two effects.
First, it will improve the ``absolute'' angular resolution since the longest baseline (E24-N46) will now be 460 meters long instead of 400 meters previously. Second, the extension towards the North will greatly improve the shape of the beam, especially for low declination sources.
There are two possible classes of A configurations, both yielding roughly
equivalent angular resolutions (1.1'' at 100 GHz; 0.50'' at 230 GHz).
The first type corresponds to the antennas put on the two farthest stations
of each track (in our case: ``W23 W27 E18 E24 N29 N46''). This will provide
the highest number of long baselines, a few short ones, but no intermediate
ones. Such a configuration would then have to be combined with a B type
configuration to provide proper mapping, but when this is done, a large
range of spacings will be recovered and high resolution imaging of even
fairly extended objects could be done. Another possibility is to put one
antenna at the end of each track and one roughly at the middle. One
example is W12 W27 E12 E24 N29 N46 (Fig.
).
This configuration alone
is well suited for mapping or size measurements of very compact objects.
In addition, because it contains both long and intermediate baselines,
it could still be used during marginal weather conditions - combined with
some tapering - to yield resolutions equivalent to those of a B-type
configuration. This could subtantially improve the observing efficiency,
particularly during the winter season. However, such a configuration is
certainly not ideal for the high resolution mapping of moderately extended
objects, and even when combined with a B type configuration may be less
adequate than the other class discussed above. The possibility of having
two A configurations (one of each type) is worth considering. Note
furthermore that switching between the two would require the moving of only
three antennas, an important aspect since shuffling the snow along the
tracks is a time consuming task necessary prior to the moving of the antennas.
An additionnal question is that of a future extension of the array. During the preparation of the Northern track extension, it appeared technically possible to go up to N51 instead of N46, thereby gaining an aditional 10% on the longest baselines. However, the cost is high (most likely above 2 MFF). For a very similar cost (2.5 MFF), owing to the topography of the Plateau de Bure, we could probably extend the array to the East, up to E48 (instead of E24 currently). The longest baseline would then be 590 meters. As an example, we show the result of configuration W12 W27 E24 E48 N29 N46 (Fig.3), which yields a resolution of 0.95'' at 100 GHz (and 0.4'' at 230 GHz).
Figure 3: Long baseline configuration with North and East track extension
Users of the IRAM PdBI are invited to comment on the choice of these future extensions and configurations (contact L.Loinard - loinard@iram.fr).