[mmaimcal] Stephane's comments on ALMA receivers, etc.

[Al Wootten]

Bryan, and others:

Stephane sent this to the ASAC but I'm not sure you all have seen it.
As you know, he thinks you and I were over optimistic with Trx in the 
ALMA Memo.  I think we should have a specification, worse than which a
receiver must not be, and a goal, near current achieved performance, which is
what we defined.  Here is a proposal from him.  I think I agree with most 
of it, as a spec.  He has a number of good points, of course. - Al

A)

We have been asked by the Receiver group leader to refine the performance
specifications of the receiver.
Given the goal of getting the best possible receivers, yet producing them in
due time and quantity, I propose
to specify the noise performance in two steps

1)  Receiver noise temperature should not exceed (NUMBERS GIVEN FOR Freq.
below 370 GHz)
        for Single Side Band receivers, 6 h.nu/k averaged over the best 80 %
of its nominal bandwidth, and
        10 h.nu/k at any place in the nominal bandwidth for Double Side Band
receivers, these numbers
        should be halfed

2) The receiver temperatures should be measured in a reference plane outside
the dewar.
        - there is a technical issue here. The ideal would be a measurement
at the secondary focus, including
        all receiver related optics, i.e. with the full final cryostat and
optics. However, since receiver cartridges
        will be developped and tested separately, this cannot be performed
on the development site.
        We should propose that the receiver group instruments the test
cryostats foreseen for the cartridge
        testing in such a way as to provide a comparable receiver
temperature reference plane.

3) Numbers have to be debated. I am unhappy with the current 10 h.nu/k in
the proposal at all frequencies,
        and suggest 6 h.nu/k in (1). What about sub-mm frequencies. Would 10
h.nu/k be acceptable, conservative
        or too difficult ?

4) The proposed receiver specification do not define coupling efficiency
with the telescope. Yet this is a major
    parameter in performance. A measurement scheme should be proposed by the
receiver WG.

5) The sideband rejection of 10 dB is probably more than we need for
sensitivity purpose only. Do you have
    any other requirement on that number (e.g. single-dish observing mode,
calibration) ?

B)
The ASAC has also asked the project scientists to work out recommended
sensitivities, since many different
documents have been using very different numbers. I propose a two step
approach, always using the
mean receiver temperature as quoted above (6 h.nu/k at mm,  10 or more
h.nu/k at submm)

1) Give a "rule of thumb" number for Tsys under normal conditions i.e.
        75 % quartile of water vapor content for mm wavelengths
        25 % quartile of water vapor content for submm wavelengths
        Frequency far from an atmospheric line

    This "rule of thumb" could take the form
        Tsys(nu) = 35 + 0.45 (nu/100 GHz) for nu < 400 GHz
        Tsys(nu) = 1000 K                         for nu > 400 GHz in the
atmospheric windows
    (numbers to be derived according to receiver specifications)

2) For frequencies near an atmospheric line, the full atmospheric model
should be used. I propose to use
    the same opacity reference as before. A simple atmospheric model may be
sufficient ?

I all cases, I suggest the forward efficiency used to be 0.95 at the lowest
frequencies, falling down to 0.90
    at 900 GHz (in a parabolic way). This is to account for telescope/optics
performance degradation which is
    usually found on all mm telescopes. Note that these numbers are already
better than any telescope currently
    in operation.



The noise performance influence the mixer design and optics design. If we
fail to set agressive enough specifications, engineers may be free to select
a simpler but more noisy design... Yet we must be reasonable to avoid
throwing out all receivers... This may require a faster input than the next
face to face meeting.

I can try to draft a recommendation on the lab measuring device to evaluate
"cartridge" performance with B.Lazareff.

    Stephane