[asac] minutes-2nd-corr-informal-meeting

[Al Wootten]

Folks,

Sachiko K. Okumura and Alan Baudry forward the appended minutes, bounced owing
to length (versions in plain text and in html) by our mailer.

Clear skies,
Al

Bounced message:
------------------------------------------------------------------
Dear ALG and ASAC members,

We send you the minute of the second
informal meeting of enhanced/future correlator at
Florence 2001 according to ALG request on early April.
Japanese and European teams plan to
talk further progress about the following Action items
with US team on 10 and 11 May in Paris.

Best wishes,

Sachiko K. Okumura and Alan Baudry
------------------------------------------------------------------
Second Informal Meeting
on Europe/Japan Cooperation on Correlators and
High Speed Samplers

Minutes of 24 February 2001
Florence, Italy

Approved by: European and Japanese participants
Distributed by: A. Baudry (baudry at observ.u-bordeaux.fr)


Participants:
Europe: A. Baudry, A. Bos, M. de Vos, S. Guilloteau
Japan: Y. Chikada-san, M. Ishiguro-san,M. Momose-san, S. Okumura-san

Agenda:
- Status of the FX Correlator Project
- Status of the Digital Hybrid XF Correlator/European Future Correlator
- Areas Open to Possible Joint Cooperation
- Status of High Speed Sampler Developments in Japan & Europe
- 3-way Partnership

0. Introduction

The second informal meeting on European and Japanese plans for the Enhanced
Correlator and High Speed Samplers was held on 24 February 2001 in
conjunction with the ASAC face-to-face meeting in Florence (23/24 February
2001).
It follows a first meeting held in Berkeley in September 2000; see:

http://www.eso.org/projects/alma/committees/backwg/minutes10sept_correl.doc

1. Financial Context and Enhanced Correlator

Ishiguro-san presented the Japanese proposal for their contribution to ALMA
in which the enhanced correlator is ranked as of major value. We discuss the
correlator issues in the context of the Enhanced ALMA. The total cost for
the correlator would include the Baseline correlator and the Enhanced
correlator. The first quadrant of the Baseline correlator can accomodate up
to 32 antennas (and thus can be used for the ACA) with a cost of about 8 M$.
The Enhanced correlator cost based on the Japanese FX architeture is
estimated to be around 43 M$. The present Japanese financial model is at the
35 M$ level and could evolve in the future. Prototyping the European Future
Correlator design requires more than 3 M$ with personnel costs included;
however, the total cost estimate for 64 antennas or more is not firmly known
at the moment (ongoing task).

Several financial scenarios are possible at this stage in the frame of
Europe-Japan cooperation and of the 3-way partnership for a second
generation correlator. Baudry and De Vos mention that they cannot make any
statement on financial issues since this depends on the ASAC
recommendations, ACC decisions and involvement of the European Science
organizations. However, they feel certainly committed, on behalf of the
European correlator team, to further
collaboration regardless of the approach chosen for the final architecture
of the Enhanced correlator.
We thus desire and agree to go ahead with joint cooperation and exchange of
information (see Section 4).

2. Report on FX Correlator and Japanese Work on Digitizers

Okumura-san and Chikada-san report on the main characteristics of the
Japanese FX design: high number of channels, flexible channel smoothing in
the X-part, etc. See details in:

http://www.nro.nao.ac.jp/~sokumura/eALMAcorr/FeasibilityofHPFXforE-ALMA3.pdf

System and board architectures (8 LSIs on one FFT board, 0.18-micron LSI
technology, 128 MHz frequency clock, etc.) have been defined and prototype
development is well underway. A test system (2GHz, 2 digitizers, 128
kilochannels, 1 baseline) will be delivered to Nobeyama at the end of
September 2001.

Japanese work on digitizers is summarized in a note distributed during our
informal meeting. Okumura-san presents the Allan variance results obtained
for their 1-bit, 8 GHz digitizer in the range 1-2 GHz. The note can be found
in:

http://www.nro.nao.ac.jp/~sokumura/eALMAad/8G-AD-iguchi.pdf

The Japanese 2-bit design obtained by combining two 1-bit systems is in
progress.

3. Report on DHXF Correlator (Future Correlator) and European Work on
Digitizers

Baudry uses his transparencies presented at the ASAC meeting to discuss some
characteristics of the European design and to outline the system
architecture. Various issues or options related to system architecture,
delay tracking, fringe stopping, system synchronization, etc. are briefly
commented by Bos.

Okumura had a comment on the cooling of the DHXF correlator board whose
power consumption might be more than 100 W per board.

Chikada asks about the effects of computational errors in the FIR filters
due
to internal quantization effects. De Vos mentions this has to be handled
in the quantization studies at ASTRON (see Andre Gunst's contribution in the
February meeting on the European Future Correlator). Other work has been
made in this field by the European team and Baudry will send some
references, including Gunst's contribution, to Chikada.

Baudry presents the major milestones of the European digitizer plan
including the Test Bench to be used for Allan variance and spectral analysis
of the fabricated multi-bit samplers. The first European 2-bit, 4 GHz,
bandpass sampling ASIC design has been accepted by the foundry and further
fabrications will be submitted in May and August this year. The European
team expects that qualification tests of the Test Bench will start in the
Summer.

4. Conclusions and Actions

Developments in both Japan en Europe have similar goals: wide band
correlation, high resolution and sensitivity with large number of channels,
flexibility.

The schedules and milestones until mid-2003 or the end of 2003 could
possibly be brought together more closely to ease the selection of a final
concept for the Enhanced Correlator.

It was agreed that we need to define some common test criteria to evaluate
the European and Japanese concepts.

The following Actions are to be completed before the next face-to-face ASAC
meeting and made public as soon as possible.

Action 1: Propose an initial set of common test procedures and performance
indicators [initiated by Bos for circulation to participants]
Action 2: Propose a joint timeline based on the current schedules to ease
synchronization and comparison of tasks in both teams [initiated by Okumura
for circulation to participants]
____________________________________________________________________________
____