[evlatests] Modcomp free tests

[Ken Sowinski]

Were very confusing.

Walter or Rich may need to correct my memory or false statements.

1.  The executor or its control interface seems to have memory.
When we begain Modcomp free work using Xsysstart the correaltor
and idcaf thought there were two subarrays, but idcaf never geot
information for sub2.  The correlator thought there were 17
VLA antennas in sub1 and 10 EVLA antennas in sub2.  After much
trial and error we got one 27 antenna subarray by running
Csysstart, a file with a different name.  A later attempt to
kill and restart the same file resulted in a subarray with only
17 antennas.

2.  With one subarray and all antennas things still did not look
too good.  In general EVLA antennas worked well and VLA antennas
did not.  But the effects were not closing; antsol and D10 gave
quite different ideas of some antennas.  It looked like there
were large random delay errors however there was no evidence that
there were mistakes.  Trying to measure delays in line mode did
not produce any consistent results.  Changing bands made no
difference.  All IFs for an antenna behaved about the same way.

Evenutally we gave up and went back to Modcmops.  The executor
was never restarted; it will be interesting to see what happens
after the power comes back on Thursday and we start with a
clean slate.