[evlatests] modcomp-free, Jun 26.

[Barry Clark]

> 2. At 18:46 at the start of a scan the phases went loopy on VLA
>     antennas: high rates for one scan, then fringes gone for the
>     rest of the time. EVLA-EVLA baselines still good.

At that time the internal time in the CMP vxWorks side went totally loopy,
leaping forward for about 40 minutes.  This was under investigation, and 
the problem sudenly cleared itself at about 20:05.  In the process of 
cleaning up the investigation, the vxWorks side was accidentally rebooted.  
The vxWorks reboot, coupled with the continued insistence of the Executor
to send commands left the system in a confused state for many minutes.
It finally sorted itself out, without help (though the script had to be 
restarted).  However, it has sorted itself out into a state where the 
CMP time is wrong by a few waveguide cycles, resulting in the behavior 
that Rick refers to as pulsing.  Currently under investigation.