(Revision iv, 2000-10-19)
NRAO Tucson, Tuesday October 10 2000
DTE, JGM, LRD.
Written material was prepared in advance of the meeting and made
available on the WWW; this material can be found at:
http://www.tuc.nrao.edu/~demerson/holocdr/
The agenda for the meeting is given below.
15:00 Coffee
Since the material presented is covered in detail in the written material (see the above link), no attempt is made to summarize the presentations. However, comments from the panel and others present are included below.
Introduction
Darrel Emerson reminded the group of the requirements and purpose of a CDR. Quoting from the ALMA guidelines (see http://www.mma.nrao.edu/administration/cdrpdrgl.pdf):
Critical Design Review (CDR)
To be held before expenditure of significant funds on the
construction of field equipment which will be incorporated into
the test interferometer.
The purpose of the CDR of an ALMA (MMA) Subsystem is principally to
review 4 questions:
General Overview
Peter Napier pointed out that the spatial resolution requirement may be driven more by the need to allow for the high phase gradient left from the near field correction, rather than just the ability to resolve individual antenna panels.
Robert Lucas stated that, if the antenna is not refocussed
to allow for the near-field siting of the transmitter, then the
needed scan range for holographic maps is the sum of the
scanning distance that would be calculated for a transmitter at
infinity, and the near-field spreading of the field response
from the antenna. The latter approximates a 12-meter diameter
tube. At the distance of our near-field transmitter, and if
0.1m resolution across the surface is required, the necessary
map scanning distance is approximately doubled. However, this is
less of an issue in practice since the intention is to put the holography
receiver main dish feed to the near field focus position for
the distance to the transmitter tower.
This point was covered in more detail during the Data
Analysis presentation later in the meeting.
Bill Peters pointed out that it may not be correct to consider effects due to multi-path propagation just a sum of random, uncorrelated vectors. In practice the reflections may not be random, but rather may produce (for example) rapid fringes across the surface.
Ukita (Nobeyama, Japan) thought that the level of cross-talk between signal and reference channels may be more important than we had considered; this had been found to be the case at Nobeyama with holography on the 45m telescope. He had specified 80 dB as the required isolation, rather than the 50-60 dB we contemplated. More investigation of the effects of cross-talk and the precision of antenna measurement is required.
Transmitter Design
Peter Napier noted that the current planning (i.e. the use of one of the Evaluation Receiver photonic phase reference assemblies as the photonic transmitter) did not permit interferometric holography using the terrestrial transmitter. However, Bill Shillue and Larry D'Addario responded that an easy workaround for this was possible - for example by splitting the fiber photonic reference to use one photonic synthesizer as a reference simultaneously for both antennas. Although not included in the design presented today, this would be a simple and cheap modification. There are other options as well.
Peter Napier said that it may be useful to rotate the transmitter box about the axis of radiation, so that the signal polarization of transmitter can be varied.
Dick Sramek asked about how the transmitter antenna would be rotated to illuminate different ALMA antennas. Bill Shillue responded is that the current design requires this adjustment to be made manually, at the top of the tower.
Richard Simon asked if truly simultaneous single dish holography using both antennas would be possible. Given that only one holography receiver is being built, this is not possible, but nor is it a requirement. The decision was made some time ago to limit the holography implementation to a single receiver.
Receiver Design (front end)
Peter Napier asked if cross-polarization holographic measurements might be required: for example, to investigate the effects of the grooving of the panels for solar scattering in the Vertex antenna. The response is that the holography receiver is NOT designed to have different polarizations; the concensus at the meeting was that this is acceptable.
Larry D'Addario remarked that it might be possible to use the holography transmitter, with little or no modification, to produce sufficient power at higher frequencies - say in the 211-275 GHz band - for the astronomical receivers to measure in interferometric mode. The astronomical receivers are much more sensitive, and the total collecting area is much greater than for the holography system with its small reference horn.
Ukita suggested that all exposed surfaces at the receiver should be covered with absorber to cut down on reflections. We should be sure that this is permitted by our receiver design.
Receiver (digital section)
Brian Glendenning asked why FPGA had not been chosen for the real time digital processing, rather than a DSP chip. Larry D'Addario responded that the FPGA might be harder to implement, and the DSP approach had the advantage of being easy to modify and reprogram. (The interesting circumstance of Larry apparently defending a software approach, with Brian suggesting a seemingly more hardware oriented approach, was noted.)
Infrastructure (tower, cabling etc.)
The question was discussed of whether a 50m or 60m tower is required. No-one had any compelling arguments in favor of a 60m tower, so the decision is to keep to the plan of a 50m power.
Real Time Software
Ukita asked how well encoder reading is synchronized to data. The response is that both the encoder readout and the data integration periods are latched to 48 ms timing pulses, so that encoder readings and data are in synchronism to within 10 microseconds or better.
Analysis Software
Peter Napier suggested that we need a full-scale simulation of the near-field focus correction and its effects on phase error. Rather than performing a fast Fourier Transform and then applying corrections, might we be better off with a direct Fresnel transform? The power of modern PCs might make this feasible today.
Test Schedule This needs to be integrated in with the schedule plan for the so-called zero-baseline interferometer electronics test (the Laboratory Interferometer or LI).
The LO system installation milestone needs to be added to the TI plan.
Panel Discussion
A number of relative minor issues were discussed, and a list of action items was generated. In no particular order:
It was considered that none of the above action items were "show stoppers," and all could easily be resolved in a timely manner.
The panel considered each question to be reviewed in an ALMA CDR (see the list at the head of this document.) It was felt that all questions had been answered satisfactorily.
After the CDR, it was discovered that
a significant discrepancy exists between the Vertex and EIE antenna
ICDs. Although the holography receiver design as presented at the CDR
is completely compatible with the Vertex antenna design, there
is not currently enough space at the apex in the EIE design. It
is essential for this issue to be resolved without delay.