The Science IPT has received no written communication from the Front End IPT on its ability to meet the total power requirements. D'Addario has written a memo on this subject which unfortunately has a few confusing items whose impact we are assessing. What stability can be achieved? What measurements have been made to suggest these levels? Once we know this we can finish the simulations begun years ago by Mark which led to the 1/f stability requirements which were written into the project book. You will recall that these were based on a lesser-evil sort of analysis, where the atmosphere was considered, and switching/wobbling was considered, and the requirement was that the stability could not be worse than the sum of these. This was in 1998 I think. Subsequent memos by Wright, Welch and by Crutcher through the ASAC expanded on these requirements. We agreed on them at the November meeting here in Charlottesville, and the ASAC endorsed them as codified in version 2.2 of the (Draft) Front End Specifications (no further revision has been distributed to the Science IPT): 4.10. Gain stability FEND-80960-ZZZ This section applies only to the operational mode. Variations in gain, as defined in section 4.9, as a function of time should comply with the following: Application Power Gain Variation (RMS) Single dish observation 10-4 (TBC) over 1sec Interferometer 3x10-3 (TBC) over 1 sec Differential polarisation 5x10-4 (TBC) over 1 sec This was presented to the ASAC in April, and agreed to by them. There has been some confusion over the interpretation of the first point, which I have clarified as being applicable to continuum observations on four dedicated total power antennas; I believe that this was the context of the November discussions. In the area for the Groningen meeting, on ALMAEDM I have found a document called Front End Specs v2.5, which has the same wording as above. I deduce that discussion in the Front End IPT has led it through several further iterations without change to this section. Stability levels of this order have been achieved on existing interferometers and on total power instruments. The classic way to achieve this is through temperature stabilization; I can cite examples going back to the 60s of this. This was called for in the MMA SYSTEMS WORKING GROUP REPORT, Memo No. 142 of the ALMA Memo Series. Has this been considered in receiver design? If so, why was it not considered? That memo reports achieving gain variations of 1.3 x 10-4 in an hour, with then-existing receivers. This may not be possible with InP amplifiers and with wide bandwidth receivers; if so the FE Group should demonstrate that. This suggestion was endorsed by the Calibration Strategy PDR of June 2001, which occurred in Cambridge England. Currently at BIMA, in my understanding, stability is achieved by slight detuning of the receivers. This results in a loss of sensitivity. Observers at BIMA have found it worthwhile to suffer this loss of sensitivity in order to reap the gain stability benefits it provides. For that subset of experiments which require sensitive continuum measurements over large fields, ALMA observers may find a similar tradeoff acceptable. An indication from the FE Group evaluating the sensitivity loss would be useful. I'm not sure that observers would want to go so far as to suggest that more stable amplifiers be used, rather than InP devices, as that would affect all observing at all wavelengths. Yet a third possibility is that for the four antennas dedicated to total power continuum observing a special receiver be constructed, perhaps a correlation receiver. This seems to me to be a drastic solution, to be attempted only when the alternatives have proven unacceptable. The ASAC and the Science IPT endorse the goals as spelled out in the Draft Front End Specifications Documents, Versions 2.2 (which was provided to us) and Version 2.5 (which was on ALMAEDM). The Science IPT has, of course, lost manpower over the last few weeks but we feel that the highest priority task, that of defining the configuration, is well in hand. Current work is now concentrated in the calibration area, and in defining the Design Reference Science Plan. However, recognizing the importance of demonstrating the fundamental importance of gain stability to ALMA Science, we have a task of high priority to consider Science losses in continuum images owing to gain stability variations of differing levels. (All Science IPT documents are posted in the clear at http://www.cv.nrao.edu/~awootten/mmaimcal/ and have been for five years, including specific lists of action items). What are those differing levels which should be used in this activity? I have suggested to Mark Holdaway, who is undertaking to perform the simulations, that the most important levels are those in the Draft Front End Specification document. John Webber suggested to me that even in the total power receivers, he thought that only 7-8x10-4 over 1 sec might be achieved. We've waited for a long time for such a number; I hope that the FE IPT will provide that or a similar number so that we can illustrate the effects that will have on science. What science? The Design Reference Science Plan will give us a quantitative (but tentative) idea of what the relative standing of total power continuum needs are with ALMA. With these two elements, a quantitative measure of the deleterious effects of receiver instability, and a quantitative measure of the demand, we should be able to endorse whatever course of action it is that the Front End IPT recommends. But we need to know what you think you can achieve (which I thought was what the FE Spec gave us) and an inkling of how you think you can achieve it. I.e. what does the matrix: Method Stability Cost sensitivitity achievable implications implications Default plan Moderate? Low? Low? Temp stabilization High? Moderate? Low? Stabilization by mistuning Moderate? Low? High? Correlation receiver High? High? ?? look like? Numbers in the first column would, of course, be most helpful.