[evlatests] EVLA Time error fixed.

[Vivek Dhawan]

The good news:  Its still fixed.

The other news: It did not break when we originally thought,
                 but later With No Apparent Provocation.

                 It broke twice on Feb 14th, by differing amounts,
                 again WNAP.

                 It fixed itself WNAP, before we 'fixed' it again
                 by re-synch. It has stayed fixed since Feb 18.

A revised timeline is below (times approx).

The only science project I can find that might be affected is
Urvashi's. Its hard to tell because of the frequency switching.
Since she can apply self-cal to M87, and the time offset is
smaller than the integration time of 5s, even a wide field
should be immune after self-cal.

UT date     start UT  end   project  ! comment

2009-02-11  07:14           POINT    ! Known to be good.

2009-02-11  10:18           AB1321   ! Probably good (not examined)

2009-02-11  15:43           MAINT    ! Originally accused to be start
                                      ! of problem

2009-02-12  03:30 - 06:00   POINT      ! expected BAD, found GOOD.

2009-02-12  05:42 - 15:40   AB1321     ! expected BAD, found GOOD.

2009-02-14  01:30 - 06:30   BASELINES  ! BAD, 0.59sec offset.

2009-02-14  06:30 - 11:30   POINTING   ! BAD, 0.85sec offset.

2009-02-14  02:30 - 07      POINTING   ! expected BAD, found GOOD.

2009-02-15  05    - 15      AR664      ! Urvashi Rao. GOOD??
2009-02-16  05    - 15      AR664      ! Urvashi Rao. GOOD??

2009-02-17  12:50           AW761      ! Expected bad, found GOOD
                                        ! hard to tell.

2009-02-17  23:20           SOFTW     ! Problem fixed by re-synch

2009-02-18  23:30 - 00:50   BASELINES !  GOOD.

2009-02-19  03:00 - 07:30   BASELINES !  GOOD.

------------------------------------------------------