Memo Review

Memo: 352 - Design and Development of 183 GHz Water Vapour Radiometers
      Hills et al., 2001Mar

Reviewer: Chris Wilson

Date Received: 2002Aug19


Memo 352 outlines two complementary designs for 183 GHz water vapor
radiometers for ALMA as well as a program of work to deliver two prototypes
by March 2003. There are several places in the memo where it would be useful
to provide updated information, given the 18 months that have passed between
the writing of the memo and this review. I have listed those items separately
at the end, in case it is not intended to update old memos. 

(1) The introduction lists 3 key aspects of the phase correction problem that
need to be pursued in parallel. Item 1, the design and testing of the
prototype, is the main concern of the memo, but item 2, on experimental
testing of the 183 GHz technique, and item 3, on atmospheric modeling, need
to proceed in parallel. To the best of my knowledge, the atmospheric modeling
has advanced quite well but the experimental testing of the 183 GHz
radiometers in true phase correction with an interferometer has hardly
progressed at all. Although two radiometers were installed on the SMA by a
different team in August 2001, a combination of bad weather, computer
crashes, and the SMA testing schedule has resulted in very little data being

In my opinion, the lack of experimental testing of the 183 GHz phase
correction technique on any system other than the JCMT-CSO interferometer is
a serious limitation and source of concern. I don't know what the solution
is, but I believe one needs to be found. 

(2) The memo discusses testing the prototypes on an interferometer. I agree
this is an essential part of proving the final design. Will this be possible
with the ATF or will these tests have to rely on some other interferometer,
possibly with completely different electronics and so on?  I think the
instruments need to be tested in the field as well as in the lab. 

(3) I am not up to speed on whether the project is considering using the WVR
to correct for tip-tilt errors. Has a decision been made on that, and if so,
could it be incorporated into the memo text? 

(4) I agree that leaky and unlocked LO could be a problem for the
astronomical receivers. When will it be possible to test this? Will the
prototype antennas at the ATF have electronics etc. that are sufficiently
close to the final production version for such tests to be useful? Is it easy
to lock the LO using frequencies available from the project and how
accurately does the frequency of the LO need to be known? 

(5) I didn't at first see where the decision to use filters in the backend
came from. The previous memo (303) recommended an analog filter bank. Under
other technical issues, the reason for chosing filter banks is briefly
discussed. What is the current status of the analog correlator systems i.e.
are they still limited to 4 GHz bandwidth? And so is this still the correct
choice, to use filters? 

(6) It may be my ignorance, but I don't follow the numbers in the discussion
of gain fluctuations in note 1. If the current systems achieve 2 parts in
10^4 over 1 second with a 2 Hz switch rate. The proposed systems have a 40 Hz
switch rate, so presumably we gain by a factor of sqrt(20). But I don't see
how that gets us to below 10% of the total noise. What is the total noise in
the current systems? And what fraction of the rms achieved with current
systems is due to gain fluctuations? 

(7) In note 6., if the filters don't match well from one telescope to
another, is it likely to be possible to work out a calibration method to
compensate for this with the necessary precision? 

(8) How useful will tests of the two prototypes be on an interferometer 
if the prototypes are of different designs?

(9) In the current optics design, is there room for M1 and M2 to grow 
twice as large to accomodate tip-tilt correction and not interfere with 
the optics of the astronomy receivers?

(10) I didn't understand the discussion of what happens if the 183 GHz 
signal is seen by the astronomical receivers. The memo says that it would 
be at a power level no greater than that form a 290 K black-body entering 
over the RF band of the receiver. Does this mean it would be no brighter 
than the sky? But wouldn't that increase Tsys substantially? Or would it 
only be a problem for observations around the 183 GHz frequency?

(11) In discussing the test program, the memo says that it is still too 
early for the SMA to commit to helping with ALMA-related testing. Sadly, 
I would say even 18 months later it is STILL too early for that. This is 
unfortunate, because I think such testing outside of the laboratory is a 
critical step. We are currently relying on a few hours of real data with 
a single-baseline interferometer that could not measure absolute phase as 
our justification for how well this technique will work for ALMA. This is 
rapidly becoming a critical piece of information that is missing from the 
ALMA project.

Possible updates to the memo:

(1) Can we refine the number for change in zenith angle that is to be 
allowed (first paragraph under specification)?

(2) Four bands are needed for the filter bank but more study of this is 
needed; any progress we can use to update the memo?

(3) Any progress on how good the matching of the filter banks need to be 
on the different antennas?

(4) Was the PDR held as scheduled in July 2001? Results?

(5) There is some discussion about various ways to provide a cold load, 
but I thought we have settled now on providing a cold load in the main 

(6) The memo says the authors intend to use existing atmospheric codes to 
investigate various topics. Have the authors carried this out, or has 
someone like Juan Pardo done it (or is planning to do it?), in which case 
he should be given the credit.

(7) Is the team still on schedule? If not, could the timeline be revised 
to reflect the current best estimate?