[evlatests] Results from Friday night tests

[Rick Perley]

    We had 3 hours of time Friday evening.  Ken used the first hour or 
so to tweak up
some things.  He will presumably issue a report. 

    With the rest, I did the usual 'fast continuum' observations of 
strong sources
and blank fields, at all the bands.  X-band was selected to test the 
phase connectivity
issue -- half a dozen sources were used. 

    A) L-band.  The calibrator was missed, so no conclusions can be 
reached.  The
off-source data looked normal.

    B) C-band.  16B has the multiple-level problem (the same as that 
which used to be in 14B).
There are two flavors of multiple-level problems.  This one is that 
which has 6 different amplitudes,
separated by about 1 dB in power equivalent. 
                   No amplitude drops or phase spins are seen. 
                   We have normal rms on antennas 14 and 16 (other than 
16B).  While reassuring, I
remind readers that the rms should be reduced by a factor of 1.4 or so. 
                   There was very high closure between 14 and 16, 
visible both has reported closure
errors (up to 20%), and and high rms *between 14 and 16 alone*.  High by 
a factor of 22! 
No phase effect is seen. 

    C) X-band.  16B problem the same as above.
                   - No closure errors seen between 14 and 16. 
                   - No phase spins or amplitude drops. 
                   -  The 'end of scan' problem with 14 and 16 is still 
with us -- but in almost all cases,
it affects only the last record (or 0.416 seconds) only.  In one case, 
it took out 11 seconds of data. 
                   - Phase connectivity is good for antenna 16, but 
there is an odd variation for antenna
14 -- identical on all IFs -- where the phase changes by 10 or 20 
degrees between sources, but
in a clearly organized way.  For 30 minutes, all the changes were -10 
degrees.  For the next
30 min, each change was +10 degrees.  The next ~1 hours saw -10, etc.  
As I was moving north
and south on a similar timescale, there might be a relation.  I'll check 
this later. 
                   -- Short-term phase behavior is better than the VLAs 
(no 'delay clunks'). 

    D) K-band and Q-band.  The data look good, but I'll analyze them 
later.  Outta time!