[evlatests] Feb 14 tests and confusion

[Ken Sowinski]

Barry worked on evla reference pointing at the beginning and
fought problems in the evla script he was using as well as
problems in the system.  

When he gave up I tried to track down what was happening.
EVLA antennas were weaker than they ought to be, but some
EVLA to EVLA baselines showed much larger amplitudes than 
EVLA to VLA baselines.  After a little while I saw that
the system would occasionally snap into proper behavior
for a few records and then go away again.  All the EVLA
antennas did the about the same thing at the same time.
I tried disabling phase switching for antenna 16, the worst
offender, and it magically began to work much more solidly
after the Modcomp skip to get phase switching turned off at
the correlator.  From then on only 16 misbehaved at times;
sometimes both IF pairs at the same time and sometimes not.
When the amplitudes came and went the phases would jump; 
sometimes the phases would jump in anticiaption of an 
amplitude change.  In its most broken state most of the
phase jumps were 120 degrees.  The problem certainly must 
be in the L302(s) or its timing signals.  Do we try to do
something about this now or what for the new L302 firmware
and MIB software?

There is a long track of data on 3c84 at X band that the
brave might want to find and look at.  The synoptic view
provided by AIPS can be of more use that my limited observations.
IAT time
2300   Start X band integration
~2330  Turn of phase switching for antenna 16.  This
       will be obvious because 16AC data will disappear.
2335   Do another Modcomp skip
2337   Turn phase swithcing on again for antenna 16.

Once the system started working more solidly I made a survey.
X band
13A has a lower response than expected and the SD voltage in
the T5 seems abnormally large.

C band
16 CD looks weaker than all the others
13 does not work at all. No fringes, no TP, no SD.

K band
16 is rather weak; most likely it is not well pointed.

L band
13A has a lower response than expected and the SD voltage in
the T5 seems abnormally large, just as at X band.


VLA pointing was curious, but that is a story that belongs in
a different forum.  It may only be a combination of wind and
daytime tilt from differential solar heating.