[evlatests] Tests June 30 part 2

[Barry Clark]

I was trying to drive a VLA antenna (Antenna 10) from the Executor.

Various improvements that Hichem has made makes this apparently work
at L band in IFs AC.  Encouraged, I tried firing it up at C band, also
IFs AC only.  It worked pretty well.  There are noticeable artifacts.
There is a phase glitch every 5 sec.  It seems to correspond to a phase
jump of a few degrees in the integration before the IAT 10 second, and 
the one five seconds later (a smaller jump there, I think).  A little
hard to explain - the java code is quite capable of producing glitches
on a 10 second interval; 5 seconds is hard.  There is a five second
periodicity in the CMP, where the 1.25 sec transmission of phases to
the L7 coincide with the 10/192 sec resyncing of the system heartbeat,
but this seems to have the wrong magnitude.  Might be capable of producing
a heartbeat loss, which would produce something like a 20d glitch, not
the 6 or 8 observed.  But syncing the commands with TAI, which we should
get around to shortly may help this.

When I went back to L band, the effect is visible there also, but is only
two or three degrees, so easy to overlook.

When I tried to drive both IF pairs, all hell broke loose.  IF pair AC
appeared to be getting wildly wrong frequencies; IF pair BD seemed to be
getting reasonable frequencies but wildly wrong phases.

Vis-a-vis the weird phase rate at 16 L band.  I was running at 1 sec 
integration at L band anyway, and had a look at it.  Its phase is constant
during the ten second intervals, then takes a leap to a totally new phase.
This could be explained as a problem with the heartbeat strobe to the
L302, except that the fact that it occurs at some frequencies and not others
is very difficult to explain.