[evlatests] EVLA SW tests, May 24 2006

[Ken Sowinski]

Today was mostly spent getting things to work again and
working with pointing questions.

The start was slow becuase initially no antennas were avaialable.
Antenna 13 is out becauase major pieces are in the basement.
Antenna 14 took a while to get put together again, but does work.
Antenna 16 had an E-stop set and took a while to get reset.
Antenna 18 Had not yet been unlocked after the Q band Rx was secured.

Once all three antennas were available the only problem noted was
that gain settings for 16B were not right anymore because the
downconverter had been replaced.  Perhaps Walter can use his
program to set this aright tomorrow.

Pointing and subarrays
I was able to verefiy that pointing works as expected regardless
of which Modcomp subarray is doing pointing.  I first ran a version
of Barry's sysptgx.evla in mchost and the usual SYSPTGX in sub 1,
evla antennas wiggled and believable pointing offsets were delivered
by the Modcomp and telcal.  I then put most antennas in sub 2
running SYSPTGX while sub 1 looked at SYSSTART.  Again everything
worked as expected.  The implementation of the synchronization signal
however is not foolproof for multiple subarrays and it will not work
in general when multiple subarrays have source changes at arbitrary
times and one (or more) subarrays are doing pointing.

There were small problems with antenna 18; it may only be a coincidence
that only IFs A and C work in this antenna.  The Modcomp did not deliver
pointing results when antenna 18 was in sub 2; telcal did.  However
telcal refused to belive there was valid data for LCP and only delivered
RCP pointing offsets, regardless of subarray.


Referenced pointing
Two tests were done and will have to be looked at more carefully to
substantiate my preliminary impressions.
1.  I rerean Rick's file from last Friday.
2.  Vivek ran a file which alternated X and K band pointing scans
     with and without application of measured pointing.
The first was run by a freshly compiled executor in a shell on mchost.
The second was run by the operator using the standard methods.

The first test looked to do all that was expected.  All the pointing
scans worked.  Antsol gains were higher when pointing offsets were
applied than when they were not.  I watched the pointing solutions for
the second test from time to time and saw no failures to point.  I
think I saw cases where applying pointing solutions worked as expected
but a more careul look is needed.  vivek will look at this and make
a separate report.


Thursday
I will be at the VLA tomorrow to work on the system controller again.
It may be possible to fit some evla work in if the pressure is great
enough.