[evlatests] 180 degree phase jumps, and more

[Barry Clark]

The "ten second phenomenon" is due to an Executor software bug.  Improper
rounding was occasionally causing the phase command for the first ten seconds 
of a scan to go out too late.  Fix implemented, code committed, will be put 
in standard Executor next update.  Thanks, Ken, for finding the smoking gun.

"18 B & D phenomenon" appears to be the "needs another foot of cable"
fix already implemented on Antenna 13 BD.  Has 13 BD been behaving itself 
since we did that about a month ago?  (Hmm.  If it happens again on a BD 
L302, I think we should re-route that cable a little.)