[evlatests] 'late on source' problem

[Gustaaf van Moorsel]

It was unclear whether the problem, seen particularly in L-band,
that EVLA antennas would be on source on 1 min 20 seconds instead
of the more usual 20 seconds, was still with us.  It looks like
it isn't.  I quickly went through the following band changes, all
on one strong source:

X - L - X - P - X - C - K - L - K - P - K - C

In other words, switches to the low frequency bands (L, P, C) from
the higher bands (first X, then K).

Only for the first scan is the first 1 min 40 sec flagged by the on-
line system (except for antenna 7, which only comes online halfway
through the first scan, and antenna 20, which does not come online
at all during that first scan).  As an aside, we don't seem to be
hit by the infamous 'first scan' syndrome.  Maybe Ken is right and
it has been solved.

For all subsequent scans, it takes between 11 and 41 seconds for
all antennas to come online.  There does not seem to be a correl-
ation between band and time it takes to come online.  There were
only two scans where it took 41 seconds, and in those cases half
of the antennas were online in 31 seconds.  For what it's worth,
these were the third X-band scan and the third K-band scan, which
have in common that they are preceded by a P-band scan.

In 9 out of the 11 scans (not including the first scan here) all
antennas were online in 20 seconds.  In both L-band scans (which
were the main reason of the test), all antennas were online in 11
seconds!

Unless there is a strong objection, I am inclined to take the all-
eged 1 min 20 sec it takes to get to L-band out of the list of
'known problems'.  It must have been solved somewhere along the
way.

Gustaaf