Conclusions

1. We demonstrate that with respect to the signal sideband, DSB observations are noisier than SSB observations by the term (Equation 11)
2. Using DSB systems, spectroscopists often analyze lines in both sidebands. Above certain levels of antenna and atmospheric noise, it may be more efficient to perform two SSB observations rather than one simultaneous DSB observation if SSB capability is available. In terms of receiver temperature, we show that the break-even point is given by (Equation 17) Below this receiver temperature, sky and antenna noise dominate and SSB mode is always more efficient.
3. Receiver noise temperatures are improving year-by-year and some devices are already within a factor of a few of the quantum noise limit. Because of the presence of atmospheric and antenna noise, we may soon reach the point of diminishing returns in further lowering receiver noise temperature. Parametrically, we show that the acceptable lower limit for receiver DSB noise temperature can be written as (Equation 22) where n is the degradation factor in observing speed that we are willing to tolerate over a quantum-limited receiver. An appropriate choice for n might be 2. In the 230 GHz band, the ``acceptable Trx'' is about twice the quantum limit; in the 650 GHz band, it is about 3.6 times the quantum limit.
4. We emphasize that keeping antenna spillover losses to a minimum pays large dividends for any system and that keeping the image termination temperature as low as possible is critical for SSB systems. We also emphasize that among ways to achieve lower noise, a dual polarization system usually leads to the most improvement.
5. We show that for the anticipated MMA antenna parameters and atmospheric transparencies appropriate to good weather on the Chilean or Mauna Kea sites, SSB systems offer observing speed improvements over DSB systems of 50% at 230 GHz, >85% at 345 GHz, and over a factor of 2 above 400 GHz.
6. We note that for single-dish mode, which the MMA is also intended to support, SSB systems are often important for improving calibration and reducing spectral confusion from the image sideband. The latter point is not an issue in interferometer mode as the sidebands can be separated by phase switching.
7. Below 300 GHz, the noise advantage of SSB systems is perhaps marginal, at least under the best skies and assuming that the specified antenna performance is achieved. However, the DSB-to-SSB integration time ratio is a steeply rising function at low opacity. Thus, if weather conditions deteriorate even slightly, the SSB advantage grows rapidly. Above 300 GHz, the noise advantage of SSB-mode observing is almost always significant.
8. Given the advantage to single dish observing and the noise improvement above 300 GHz, we conclude that development work to achieve a SSB capability for the MMA is well-justified.