[asac] DRAFT DSB/2SB recommendation

[ewine at strw.leidenuniv.nl]

Dear ASAC,

Below please find a very ROUGH DRAFT of our DSB/2SB recommendation to the
project.  Thanks to Geoff for the preparation. Please read carefully
and send comments to me and Geoff. We'll discuss it briefly tomorrow
at our telecon.

Talk to you all soon,

Ewine

****************************************************************************


Issues Associated With DSB versus SSB Receivers for the Initial ALMA Complement

 A number of ALMA memos (numbers 168, 170, 301, and 304) and reports by the
project scientists to the ASAC have recently considered the potential
sensitivities of double sideband (DSB) versus a number of single sideband
(SSB, which here includes variants such as sideband separating, or 2SB,
approaches) receiver designs.  While the conclusions differ to some extent,
the numbers and general trends driving the potential decisions are similar and
are worth summarizing:

-The potential sensitivity gains with various SSB options are greatest for
 observations of transitions in a single IF sideband or for observations of
 lines in separate sidebands where the needed correlator capacity is less
 than that available.

-The potential sensitivity gains with SSB receivers increase as the receiver
 noise contribution to the total system temperature decreases. That is,
 SSB receivers provide improved performance as the atmosphere begins to
 dominate the overall noise.

-For continuum or wideband spectral line observations that "fill up" a 
 correlator bandwidth matched to that available from the DSB receiver IF,
 SSB receivers with the same IF bandwidth are sometimes less sensitive
 because the DSB bandwidth is effectively sqrt(2) larger.

 With the current estimates of achievable receiver noise temperatures, the
estimated improvements in system temperatures with SSB receivers range from
1.4-1.2 (low frequencies to high) for observations in a single sideband (Memo
304, Figure 2).  As receiver noise temperatures drop, the improvement
attainable with SSB receivers gets larger.  Under the same conditions,
continuum observations with DSB receivers are more sensitive, particularly at
high frequencies. It is worth stressing, however, that with better receivers
SSB approaches will be equal to or superior to DSB receivers for all observing
modes, and that the potential improvement corresponds to a very large number of
additional antennas.  Recent work at submillimeter frequencies has
demonstrated that the SIS mixers themselves can operate near the quantum
limit, and that in the future it will be possible to build receivers that are
much more sensitive than those likely to be initially installed on ALMA as our
understanding of materials at THz frequencies improves.

 At that point, SSB receivers will clearly be superior, especially if their
IF bandwidths can be made sufficiently large to occupy most of an atmospheric
window and fill a very large correlator with a single sideband (and with 
dual polarization receivers).  In the meantime, the overall gains (or losses)
in sensitivity with SSB versus DSB receivers are a complex function of the
assumed receiver temperatures, the atmospheric conditions under which
observations are performed, the correlator capabilities, and the temporal
mix of observing modes used by the array.  It is largely differences in
these parameters that drive the differences in the various ALMA memos and
reports. SSB receivers also are more complex to design, build, and maintain,
and so if the gains are small or negligible then DSB receivers provide
better value from a total project perspective in terms of cost and risk,
especially early in the project lifetime.

 Given the likely pace of design and development after ALMA construction,
it seems unavoidable that both DSB and SSB receivers will be implemented
on the array at some point.  It is therefore important for the project not
to preclude either option at this time, at least in terms of making decisions
now that make it extremely expensive to implement new receiver layouts in
the future.  Some specific recommendations, by no means exhaustive, might
include:

-Dual polarization, DSB receivers provide the best alternatives for bands
 8-10 at present, and should be the baseline design. The correlator(s) must
 therefore provide for phase switching demodulation of the upper and lower
 receiver sidebands.

-Design and development of SSB receivers is critical for ALMA and should
 continue.  Decisions on when it is appropriate to implement SSB designs,
 especially for the lower frequency bands which are likely to have SSB
 implementations ready first, are best made by the Receiver and System
 IPTs. The ASAC requests regular updates on the progress in this area,
 especially as regards the first light receiver bands.

-The IF distribution and correlator downconverter systems should not
 preclude the introduction of SSB (read 2SB) receivers, or at least should
 not make the conversion to SSB/2SB approaches prohibitively expensive.

-The cryostat design, cryogenic systems, and interfaces should be compatible
 with a gradual migration from DSB to SSB receiver cartridges.