[mmaimcal] THE NECESSITY OF N > 4 SUBARRAYS FOR ALMA

[Steven T. Myers]

  ON THE NECESSITY OF A MODERATE NUMBER OF SUBARRAYS FOR ALMA

                        S. T. Myers (NRAO)

                        29 February 2000

  -----------------------------------------------------------

ABSTRACT:

For some time now, there has been debate about the actual number of
simultaneous subarrays required for ALMA.  From the system design
point of view, arguments have been presented that 4 is sufficient.
There is even a baseline design for 2 subarrays that has been floated.
However, for some time now, scientific arguments have been given that
compelling cases exist for having more than 4 simultaneous subarrays.
I propose that a baseline of 8 subarrays would be a moderate design
goal, even a reasonable design specification, given the probable uses
of the ALMA array.  Below I present a concrete example of a case using
8 subarrays, and discuss why this should be taken seriously in the
current debates about the design.

  -----------------------------------------------------------

The following is a strawman subarray allocation table given as a
concrete example of an Nsub = 8 case:

#1 Antennas out for reconfiguration and baseline determination
#2 Antenna(s) used for mm-wave VLBI
#3 Antenna(s) undergoing receiver testing or warm-up/cool down

#4 Main interferometric subarray
#5 Main single-dish subarray

#6 Secondary interferometric subarray at second frequency band
#7 Secondary single-dish subarray at second frequency band

#8 Secondary interferometric subarray for transient event monitoring


Discussion:

First, the engineering subarrays -

#1 Antennas out for reconfiguration and baseline determination

   It is highly likely that at any given time there will be several
   antennas being moved or having just moved, and some number of
   antennas being used to anchor the new baselines into the array.

#2 Antenna(s) used for mm-wave VLBI

   Several time a year (at least) one or more ALMA antennas will be 
   used to perform mm-wave VLBI with the VLBA and other VLBI antennas.

#3 Antenna(s) undergoing receiver testing or warm-up/cool down

   It is almost certain that at any given time there will be at least
   one 4K dewar out of commission at any given time, and probably more
   than one, especially during the early years of operation as we gain
   experience.  Furthermore, we will likely be testing new receiver
   components and bands, and therefore a third engineering subarray
   is probably fairly common.

Now, the observing subarrays -

#4 Main interferometric subarray

   This is the primary observing array

#5 Main single-dish subarray

   Many (most?) observations will require or desire simultaneous
   single-dish OTF scans to allow high-fidelity mosaicing.  There
   are numerous documents out there on this.  'Nuff said.

#6 Secondary interferometric subarray at second frequency band
#7 Secondary single-dish subarray at second frequency band

   It is also possible (likely) that not all antennas will have
   receivers at all bands (especially the submm bands and especially
   during the early years of the array).  For example, an inner ring
   might be outfitted with the highest-frequency receivers doing the
   primary observations (subarrays 4-5) and the outer antennas in the
   array be observing at some lower frequency band.  THIS WAS DE
   RIGEUR DURING THE EARLY DAYS OF THE VLA Q-BAND INSTALLMENT!

Finally,

#8 Secondary interferometric subarray for transient event monitoring

   At any given time, there is a GRB to observe. Furthermore, there
   are a number of X-ray transients requiring observation.  It is
   impossible to extrapolate from the current VLA situation, but it
   is absolutely clear that there will be the desire to commandeer a
   subarray for catching, following-up, and monitoring these
   interesting objects or ones like them.

   Furthermore, there are more mundane cases where the ability to pull
   out a small subarray for monitoring is called for.  An example is
   time-delay monitoring of gravitational lenses.  (It might be
   advantageous to monitor at 85GHz where variability is high, and
   there is the whole southern sky.)

Although this may be an extreme case, it is absolutely clear that 2 
subarrays is incapable of supporting even normal operations, and that
even 4 subarrays is insufficient.  This is particularly true in the
early stages of the array construction, where there is most likely
going to be a heterogenous mixture of receiver suites among the
existing elements.  Therefore, even a design where 2 initial subarrays
can be expanded later on will seriously hamper operations and testing.

I would like to reiterate that even 4 subarrays is insufficient.  For
example, a common mode might be 

#1 Antennas out for reconfiguration and baseline determination
#2 Main interferometric subarray
#3 Main single-dish subarray
#4 Second interferometric subarray at second frequency band

which uses a full 4 subarrays.  There is no headroom in a 4-subarray
design for a transient response subarray or VLBI.

Conclusions:

I propose that the design goal for ALMA be 8 simultaneous subarrays.
There may be reasons why the actual specification be "at least 4
subarrays", though this will be insufficient for all but the most
mundane operational modes and thus I strongly urge Nsub = 8 be adopted
as the spec from the beginning.  It is also greatly desirable that no undue
limitation be imposed on the number of antennas allocatable to any
given subarray, as this may be driven by either the scientific
requirements of a given program, technical requirements of an
engineering task, or the chance availability of receivers or
configuration status.  Although these requirements may have
substantial budgetary impact, unnecessary frugality or conservatism at
this point will likely hamper the full use of ALMA as the flagship
radio-astronomical observatory of the following decade.

There are compelling scientific and operational arguments in favor of
having the capability of 8 simultaneous subarrays, and in strong
opposition to having fewer than 4 subarrays.  Although 4 subarrays may
be seen as fulfilling the baseline requirements for "normal"
operations, there are clearly cases of interest where more than 4
would be desired, if not necessary.  Furthermore, this capability must
be included in the current design, and cannot be left as a later
upgrade, as it is most likely the early testing stages where a number
of engineering subarrays will be used which will detract from the
number available for scientific observations.

Note that none of the arguments presented here are new, and have been 
expressed by other members of ALMA/MMA working groups, particularly
Michael Rupen and Bryan Butler.  We (the Imaging and Calibration group
in particular) would like to see these concerns more clearly addressed
in the discussions currently underway before crippling limitations are 
built into the rapidly evolving designs.

The reader is referred to the scientific requirements most recently
presented in the white paper "Astronomical Requirements for the
Millimeter Array Correlator" by Rupen, Shepherd and Wright (1998).

-------------------------------------------------------------------------
|:| Steven T. Myers                      |:|  Associate Scientist      |:|
|:| National Radio Astronomy Observatory |:|                           |:|
|:| P.O. Box O                           |:|  1003 Lopezville Rd.      |:|
|:| Socorro, NM 87801                    |:|  Ph:  (505) 835-7294      |:|
|:| smyers at nrao.edu                      |:|  FAX: (505) 835-7027      |:|
|:| http://www.aoc.nrao.edu/~smyers      |:|                           |:|
-------------------------------------------------------------------------