From the Executive Summary of the ASAC report of the April 2003 meeting, presented to the ALMA Board at their 26-7 May meeting and noted in those minutes: ... 6. Inhomoegeneous Array: The ASAC strongly recommends that a single antenna design be adopted for ALMA. Having two different antennas designs seems certain to impact the science capabilities of ALMA for wide field mosaics and polarization observations, while in a worst-case scenario, imaging of any significantly extended source could be affected. The ASAC recommends that the project consider whether additional specifications will [be] required to enforce consistency between two different designs. If two different antenna designs must be adopted, the ASAC recommends that an identical quadrupod design be used for both antennas, which should reduce any adverse effects on the science. Later in the report... The ASAC reviewed two written documents on the impact of an inhomogeneous array that had been prepared by A. Wootten and by the ANATAC. We also heard a presentation by S. Guilloteau. The science implications of having two different antenna designs arise primarily from "common mode errors", which would cancel if the antennas were identical. Common mode errors identified include pointing errors, phase/pathlength/focus errors, phase effects due to changes in the fiber length, and polarization matching and primary beam shape. For common mode pointing errors, errors due to wind are likely to be common in the compact configuration, while solar heating in this configuration may vary from one antenna to the next due to shadowing. In contrast, in more extended configurations, common pointing errors are likely to arise from solar heating, while the wind and its associated pointing error may vary across the (large) site. For errors in phase due to pathlength and focus changes, all mechanical deformations except that due to the non-intersection of the axes (likely the dominant effect) would benefit from having identical antennas. Phase effects due to changes in the fiber length are dominated by the run to the antenna; this normally common mode error could probably be monitored and compensated for in software. Polarization and primary beam shape are determined by the quadrupod leg design; having two different antennas with very similar quadrupod designs could mitigate the problems here. However, it is worth noting that the Vertex and Alcatel prototypes do not have identical quadrupod designs. Inhomogeneous array designs also have cost implications during the construction, commissioning and operations phase. In the construction phase, the cost effect could be either positive or negative, depending on the details of the antenna contracts. For commissioning and operations, it is clear that having an inhomogeneous array implies extra costs due to the extra work involved with commissioning and maintaining two different antennas, maintaining two software interfaces (for example, different pointing models), etc. The bottom line is that anything that increases the cost ultimately affects the science return from ALMA in a negative way. The ASAC reached the following conclusions concerning the inhomogeneous array: 1. The ASAC stronly recommends that a single antenna design be adopted for ALMA. Having a single antenna design will factilitate several key observing modes with ALMA, in particular polarization observations and wide-field mosaics. It will also reduce the effort and cost required to commission and operate ALMA. 2. If two different antenna designs are adopted, the ASAC recommends that the identical quadrupod design be used for both antennas. Having an identical quadrupod design should help to minimize science impact, again particularly for polarization and mosaic observations. Minimizing the problems introduced by having two different antenna designs implies that there should be additional specifications placed on the designs, for example on the lack of axis intersection, the thermal coefficient for the expansion of the quadrupod legs, the profile for the quadrupod legs, etc. It might be possible to minimize common mode errors with appropriate specifications on the change of the antenna with tempreature and gravity and on the wind response. However, placing a number of additional specifications on the antenna designs could drive the costs up. 3. If ALMA consists of an inhomogeneous array without stringent specifications on the quadrupod and other aspects of the 12m antennas, the ASAC believes the biggest potential impact on the science capabilities of ALMA will be in the areas of polarization observations and wide field mosaics. Polarization mosaics are probably the most demanding use of ALMA and would likely be extremely difficult with an inhomogeneous array. In a worst case scenario, imaging of any sources larger than roughly 1/4 of the ALMA primary beam could be adversely affected. For any type of inhomogeneous array, the potential extra costs involved will take money and effort away from other ALMA tasks and the end result will be a less powerful instrument. Having two types of antennas will have a negative impact on commissioning and operations, with extra training, software, spare parts, etc. required. In this context the ASAC wishes to highlight the impact on the software effort, as many of the corrections required to operate ALMA with different antennas will fall to software. Summary The ASAC strongly recommends that a single antenna design be adopted for ALMA. The ASAC recommends that the project review the antenna specifications to see whether additional specifications would be required to enforce consistency between two different antenna designs. If two different designs are adopted, the ASAC recommends that the identical quadrupod design by used for both antennas, which should help to minimize the impact of the different designs on science. If two substantially different antenna designs are adopted, the biggest potential impact on the science capabilities of ALMA will be in the areas of polarization observations and wide field mosaics. In a worst case scenario, imaging of any sources larger than roughly 1/4 of the ALMA primary beam could be adversely affected.