[evlatests] X-band performance, from last night.

[Rick Perley]

    I did a 'full system check' yesterday evening with all 5 EVLA 
antennas, using
bands L, C, X, K, and Q. 

    For all of these, we ran at mode '4' with 6 MHz bandwidth, giving 32 
channels
in each of the AA, BB, CC, and DD pairs.  (No cross-polarization).  
Averaging
was set to 5 seconds (so short dropouts can't be seen). 

    Ken claims to have fixed referenced pointing (other than the Cos(El) 
application
problem), so this was done at X-band, and applied to X, K, and Q band 
observations. 
(No secondary referenced pointing was attempted -- Ken makes no claims 
yet that
this works correctly). 

    I used 1800+784 as the calibrator -- elevation about 35 degrees (low 
enough to
minimize any effect due to the Cos(el) problem).  We also spent time 
integrating
on blank sky to confirm sensitivity measures. 

    Each band was observed twice, separated by about an hour. 

    I give here the X-band results. 

    A)  Overall. 

All 5 EVLA antennas gave stable fringes at X-band in IFs
A and C.  Antennas 13, 14, 16 and 18 gave good stable fringes in IFs B 
and D -
Ant 24 has not been outfitted in B or D yet.  No dropouts in amplitude 
or phase
were noted (but none were expected, due to averaging).  Gain stability, 
on the
5-second timescale, looks as good for EVLA antennas as VLA antennas. 
    
    B) Returning to correct phase: 

It is hard to conclude antennas 13 and 24
are, or are not, returning to phase correctly.  For IFs A and C, only 
antennas
13 and 16 (of the EVLA antennas) returned to their initial phase on the 2nd
observation.  Antennas 14, 18, and 24 were greatly different.  However, the
change in phase was the same on IFs A and C, so a phase difference shows
no change.  Whatever caused the phases to change on the individual IFs 
worked
identically for both.  Weather is unlikely -- only small changes were 
noted over
the 1-hour gap for the VLA antennas. 
    For IFs B and D, the situation is more bizarre, as nearly all VLA 
antennas
'jumped' by differing amounts, always the same between IFs B and D.  The
changes are comparable to the EVLA's jumps, which were also identical 
between
B and D.

    C) Bandpasses. 

These looked superb for all EVLA antennas.  With two
observations, two bandpass solutions were obtained.  The two solutions were
indistinguishable, to a quick eye inspection.  I used an EVLA antenna as 
reference,
so the famous 'buttonhook' in phase should not be visible.  It wasn't. 

    D)  Sensitivity. 

AIPS weights were generated for all antennas and IFs. 
       - Antennas 13 and 16 are amongst the 3 best antennas in the 
array, for
all IFs.  (Antenna 13 is #1 for all!)   The reported Tsys for all IFs 
for these
two antennas (these are backend Tsys values, known to be a few K higher
than reported frontend values) was about 35K. 
       Antennas 14 and 18 are in the top 3rd of the sensitivity 
distribution on
IFs A and B (i.e. -- RCP).  But both show notably reduced sensitivity in
LCP (IFs C and D).  The likely cause was easy to find:  The reported Tsys
for both antennas in LCP is well above the median:  55K for antenna 14,
and 48K for antenna 18.  (The RCP Tsys values are also a little high --
about 40K). 
      Antenna 24 has a very low AIPS weight -- 2nd worst in the array --
whose cause is easily found to be in the Tsys:  reported at 100K.