[evlatests] crossed-hand delays

[Ken Sowinski]

I made a survey of crossed hand delays today and found significant
errors.  At all bands, the AC error was about 10 ns, the BD error
about 20 ns.  This has been corrected.

The error arises when delays are repeatedly adjusted after station
board reboots and power outages without giving sufficient care to
changes peculiar to the reference antenna.  Crossed-hand delays
were last set around March, so this represents the accumulation of
such errors over eight months.

The question arises of how important it is to measure these delays
and keep them small.  If routine calibration includes delay fitting,
for crossed-hands as well as parallel hands, when necessary, is there
a need to invent a mechanism to routinely measure and correct the
crossed-hand delays?  The alternative is to measure it only rarely
and make corrections when it gets "too" large.