[evlatests] Pointing Offset Results

[George Moellenbrock]

Rick-

    1) Many antennas show a strong elevation dependency in the X-band
> results.  This can only be due to a model error.  Ken, on the basis of
> last night's pointing run, has found the origin, and inserted corrections.
>

Does this change affect what the gaincurves actually are for bands subject
to the sub-reflector rotation trick?   (Should they be remeasured?  At least
at high-freq bands where the differences  could be meaningful....)

I guess this is another way of asking how orthogonal nominal
or referenced pointing and the sub-reflector rotation trick are.  The
nature of the rotation-trick=OFF gaincurves seem to imply (or not
rule out, at least)  that they are coupled to a considerable degree.
I think you allude to this sort of thing in your gaincurve email (trying
to optimize at el~50 deg, etc.).      Is the subreflector rotation trick
implemented  in an antenna-dependent manner?  (Different
amounts for different  antennas, etc.)

I fully appreciate that the gaincurves' principal mandate is
to merely remove elevation-dependence.   However, like the nominal
efficiency, the coupling details noted above amount to (at least) scaling
effects for which the precise nature of the factoring (and normalization)
is important if we want to tacitly assume (or even advertise) that the
_net_ nominal calibration (sw power, efficiency, gaincurve, opacity, etc.)
is within a few % of true Jy, and doesn't mysteriously change
discontinuously
(even at less than a few %) every maintenance day without explanation,
or every time new gaincurves are published.

I personally would discourage such adverts, frankly, and encourage
explicit external f.d. calibration, but I also recognize the convenience
from a characterization point of view---e.g., stress tests, pipelines, and
similar---of having a _stable_ net nominal calibration regime.
This means understanding (and keeping track of, if necessary)
how the various factors, including
implicit ones like subreflector rotations (and focus, etc.), care about
each other, and only update more than modestly coupled parameters
in parallel, for example, especially when (only) some corrections are
deferred to offline processing.   We can be as precise as we deem
reasonable and useful in all of this, of course, considering available
maintenance effort, etc. (about which I am agnostic), but we should
at least be clear about what the downstream expectations should be
(about which I care).    In particular, since some of the effects are
corrected
offline, there will be fewer surprises in using offline s/w if the net
nominal calibration model is well-defined and collectively
understood and coded-to, and consistently adhered-to,  and the
consequences for subtle but  deliberate and desirable changes in
telescope parameters recognized.

-George
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20121003/40188fcc/attachment.html>