Before we get to the actual minutes, Al has the following request:

We are starting to confront strawman configuration designs with sources. I would like the MAC to help me build a library of images they think relevant for this study. Tamara has done a good study of M51, using her BIMA SONG image as well as a mock M51 which she has 'observed' with BIMA, OVRO and IRAM to understand differences in the various maps... Our library also includes Mark's famous 60 Jy source, and an HII region in M31 which he has been using more recently. I have also added a fairly simple 450 micron image from SCUBA (since most of the flux is decidedly resolved out by our 7" primary beam). We are having a configuration meeting apres-con on 9 June in Tucson. - Clear skies, Al

The October meeting will start off with the demonstration and reception for members of Congress in the Rayburn building. Unfortunately the room only seats 50, so it may not be possible for everyone to attend. An alternate reception may be held elsewhere. It is still important for those of us who plan to attend to contact our Senators and Representatives and invite them to attend the reception, with a personal offer to explain the project, demonstrations, and equipment on display.

Al reported that the speakers list is beginning to fill up for the meeting, which runs 7-8 October.There are a number of invited speakers who cannot make the meeting and their spots need to be filled. Beckwith cannot: Koerner and Natta have been suggested as replacements; Calvet has been suggested as a replacement for Hartmann. Genzel is trying to come. The tentative schedule can be seen on the meeting webpage. SUGGESTIONS FOR SPEAKERS ARE ENCOURAGED. It is anticipated that alternate invitations will go out next week.

Email registration for the meeting is now available. Please register! Registration and general information on the meeting and the speaker status can be found at:

http://www.mma.nrao.edu/science/science99.html

Globe Travel in Charlottesville can handle your travel arrangements, including air travel and accommodations at the Omni Shorham Hotel. There is an email link to them on the meeting webpage under the "Washington and Accommodations". A limited number of rooms will be reserved, so early reservations are advised for the hotel. The deadline for the hotel to hold the room block is sometime in August.

Reminder: following the science meeting, a joint MAC/SAC meeting will take place on 9th Oct. to discuss issues relevant to both advisory committees.

AS THE TIME FOR THIS MEETING APPROACHES, MAC MEMBERS SHOULD BE THINKING OF ISSUES THAT WE WILL WANT TO HAVE DISCUSSED AT THIS MEETING.

The NSF oversight committee (MOC) met last week (they meet thrice yearly). The committee consists of six people who advise the NSF on the project management. They have been very helpful in the past with issues of management, procurement, and the evolution of the ALMA initiative. The issue facing the NSF at the moment is what dollar amount to put in the the FY2001 budget request for the U.S. portion of ALMA. The budget is being sketched out for FY2001 at the present time; it needs to go to the President by October. His budget is then put before Congress at the State of the Union address in early 2000, which must approve it sometime before the beginning of the fiscal year in October 2000.

The problem is that construction funds are now being requested for ALMA, which is now formally different from the entity which is presently being funded, the MMA. Reference to the MMA will probably be retained in the budget in words to the effect that the U.S. participation in the joint project ALMA will be based on the MMA. The U.S. contribution to ALMA will probably be presented to Congress in the form of an upper limit: "not to exceed XX dollars". NSF has therefore asked for the cost of the "U.S. reference project", i.e., the MMA (= 36 x 10m antennas) the cost of which NRAO has recently calculated. A percentage of that "U.S. reference project" cost will be used as the cap for the U.S. contribution for ALMA.

The NRAO cost estimate will be audited by a committee of scientists and auditors July 7-9 in Charlottesville. The committee is currently being put in place. Both the cost estimate and the audit will be presented to the MOC supporting the conclusion that it is a sound estimate, and they will then convey their findings to NSF. The MOC is giving strong support in this process.

The initial MOU with ESO was initialed on 30 March. The parties are in good agreement: "there are issues to be discussed but no profound issues". The NSF and the European organizations meet next on June 10; it is hoped that they will sign the MOU at this meeting. ESO has prepared a press release for the signing with a very pretty picture from Peter Schaefer (Bob Fosbury has worked on this as well), which can be seen on the MMA website. NRAO and AUI will coordinate with ESO in making the press announcement for the "planned" project.

From the date the MOU is signed, Brown and Kurz have 120 days to come up with a plan for phase 1 (design and development) of the project. This may be in place in time for the October press conference.

Took place on May 18 in Socorro. The RFP was released on March 30 and bids are due June 30. The prebid meeting allows the vendors to ask about specs, get clarification on the RFP and to see the construction site. 32 people and 20 companies showed up. 8 companies have declared their intention to bid as prime contractors, although the number that may actually submit a bid may be somewhat smaller, perhaps half a dozen. The European RFP went out April 30, with responses due July 30.

Following the list from the April 99 meeting.

THE NEXT MAC TELECON WILL BE 23 JUNE AT THE USUAL TIME.
Note that this is not the usual date, because of AAS and other meetings.

R.S., J. Turner, 25-MAY-99

AL'S SUMMARY OF THE FIR ISSUE RESOLUTION

The European/US meeting in February addressed the FIR filter issue:

* FIR Filter. The digital approach using a finite impulse response (FIR) filter to select the bandwidth and convert to baseband is the recommended approach. However simulations should be done to verify that the filter accurately preserves phase. There is also concern over whether the sampler running at 4 GHz will have sufficient stability. Action: Simulations and prototype tests needed to confirm the FIR and sampler performance.

'The digital finite impulse response filter (FIR) sounds like a clever way to filter out specified unwanted portions of a spectrum. However, the loss of 13% efficiency is not a good trade-off. We recommend keeping this option open until the correlator design has to be fixed. Perhaps the advance of technology will make it possible to avoid losses of more than a few percent. '

On the basis of this memo, Webber and Emerson proposed that the MMA project adopt the FIR Filter. In the MMA project, this process occurs via approval by the eleven division heads(DHs). John and Darrel made that proposal on 16 Feb. On 20 Feb, The DHs, sitting as a control board, agreed with their recommendation. The digital filter is now the MMA project baseline. In their proposal, John and Darrel made a summary of advantages and disadvantages of the FIR vs. analog filters; perhaps Darrel or John would send you a copy of that recommendation.

I agree with the decision of the 'control board'. It seems possible to minimize the losses through dual quantization and I think that the payoff in terms of stability and adaptability outweigh the losses. I'm looking forward to the development of the hardware prototype, now in its early stages.

Here is what John and Darrel presented to the control board:

MMA baseband filter recommendation

John Webber and Darrel Emerson
1999-Feb-16

We believe the following to be true about the differences
between analog and digital baseband filtering:

ANALOG:

Advantages:             The usual technique, well understood
                        No aliasing due to quantization--signals in the
                        rejected band are firmly rejected

Disadvantages:          Filters are all slightly different in amplitude
                        and phase response
                        Spurious response problems
                        Differential temperature effects lead to unstable
                        baselines in the single dish case and phase
                        closure errors in the interferometer case,
                        limiting the dynamic range of maps
                        Requires multiple oscillators & mixers, making
                        assembly labor intensive
                        Very expensive with packaged components, many
                        pitfalls in surface mount PCB packaging
                        Absolutely fixed bandpasses

DIGITAL:

Advantages:             Flat phase response
                        All filters are identical
                        Pushes temperature effects into sampler only
                        Fully tunable and adaptable
                        More of the band is usable than in the analog
                        case--compensates for loss of SNR due to
                        second quantization
                        Inexpensive as surface mount PCB analog filter
                        without signal leakage problems
                        Easy to build

Disadvantages:          Technique not widely used in radio astronomy
                        Extra loss of SNR due to quantizing twice--
                        once at sampler, once at filter output; effect
                        is 5% with a 3-bit sampler
                        Very strong signals in rejected band are aliased
                        into accepted band at a level dependent on the
                        coarseness of the quantization

Given all these factors, we find that the FIR filter will be able to provide significant advantages in performance and cost for the MMA, and we recommend that it be adopted. We further recommend that a hardware prototype be developed and tested, using a suitable FPGA as the processing chip.