[evlatests] Early First Scan Issue

[Rick Perley]

    At the end of software time, we ran the standard 'stress test'.  The 
first scan started quite early, (defined as before the stated start 
time), and lasted 15 minutes.  No fringes were seen on any baseline 
during this time.  On the next scan, good fringes showed on all 
baselines immediately.

    This is I think strong evidence in favor of Ken's hypothesis that 
the first scan always fails, in total, when it starts before the 
declared start time.