[evlatests] recovery

[Ken Sowinski]

Software time today was used primarily for correlator work but some
time was spent to get things working again after the various power
problems and the move of antenna 13.

Antennas 16 and 18 were not looked at at all.  When I left the VLA
the operator was complaining that the emergency stop was set on 16
and he was unable to point it for observations this evening.

Antenna 14 seemed to work fine at L, C and X bands modulo small
delay errors.

An attempt was made to get 13 back but there are many problems.
After much effort a reasonable fiber delay was found while observing
at L band to minimize the effect of any pointing errors.  Other
problems:
1.  The response was very poor in BD and even worse in AC.  An
improvement of about 50 percent was realized by changing the azimuth
offset by about 30 arc-min.  This seems unreasonable for a move to
the same arm, so:
2.  I wonder if the subreflector is working properly.  Looking at
the FRM with the device browser while changing bands leads me to
doubt it is doing the right thing, but no systematic investigation
was made.  Additional evidence is that even when the pointing was 
"optimized" at L band there was almost no response at X, or even C band. 
3.  AC is much worse than BD.  That could be delays; no attempt was
made to set delays for each IF.
4.  It is conceivable that the transition module is set to select the
wrong sideband.  This would just about account for the poor response
at L band if there were nothing else wrong.  The deformatter parameters
were set properly yesterday but were not looked at today.
4.  RCP for L band produces a very ratty signal as seen at the T5s.
Rob knows about this and it has little effect on the correlated signal
since B was about as good as D.

At various times synthesizers and L305 were checked and they professed
to be in sync.  L305 sync was verified by looking at the SD signal.
Sampler statistics and T5 parameters all seemed reasonable for the
bands we were looking at.

I have surely forgotten something; Rob may be able to fill in more
details.