[evlatests] More Results from Thursday data

[Rick Perley]

    Results from spectral line observations. 

    Data were taken on a calibrator, and on cold sky, in modes 1A, 1B, 
1C and 1D (that is
single polarization in each IF), for each band.  Averaging was set to 
the fastest possible --
1.67 seconds.  BW = 50 MHz. 
    This made a lot of data.  I reduced only mode 1B (with a single 
exception, noted below),
for L, C, X, and K bands. 

    1) L-Band. 
   
    16B is bad, as noted by Ken and me many times.  *This problem needs 
to be addressed*

    All data from IF A are wildly fluctuating.  As Ken and I have both 
speculated, we suspect
local RFI.  But there is no evidence of anything in the X-power or total 
power spectra --
they look entirely normal!  The fluctuations are in both amplitude and 
phase.  Data from
IFs B, C, and D look fine.  I don't have an easy explanation for this -- 
RFI is still the leading
candidate, but it has to be outside the 50 MHz bandpass, and must not 
affect IF B at 1385 MHz. 
It is *definitely* not the sun. 

    Otherwise:  calibration of IF B shows excellent stability, 
sensitivity, and closure.  Everything
looks exactly as it should be.  (But note that 16 was missing, so I 
can't see 14 x 16 closure issues). 
Antenna 14B is as sensitive as the best VLA antennas -- which is what we 
expect. 

    2) C-Band.

    Unfortunately, 16B chose to be dead in this run, although Ken saw it 
working normally
shortly before.  We don't know what happened.

    Bandpasses, closures, amplitude and phase stability all look 
great.    The rms of a single
channel (8) shows 14B's noise is as good as, or slightly better than, 
the best VLA antennas. 
This is good news, but it is still not as good as it should be -- about 
40% better than the
best VLA antennas. 

    3)  C-Band in IF A
   
    I reduced this one to see if 14 X 16 is giving high residuals (as I 
think we expect) due to
aliasing.   Indeed, all the conclusions from IF B (above) apply to IF A, 
with 14 x 16 showing
much higher rms than it should, when on the calibrator, but not when on 
blank sky.  This
is consistent with the aliased signal in this IF channel. 

    4) X-Band. 

    both 14 and 16 worked here.  16 B showed the 'asymptotic gain' 
effect noted yesterday. 
We need to fix this! 
    Bandpasses and amplitude and phase stability are excellent.  There 
is no unusual phase
closure between 14 and 16 after the bandpasses are applied.  This lends 
support to my
suggestion that the 8% closure seen in continuum data is due to 
mismatched bandpasses --
although the bandpass solutions look good.  (I can't resolve this 
conflict!) 
    Antenna 14 rms is a little higher than best VLA antennas (and its 
reported Tsys is also
high), while antenna 16 rms is as good as the best VLA antennas (and its 
reported Tsys
is reasonable). 

    5) K-Band

    The calibrator was a bit weak, but we plunge on nevertheless.  Both 
14 and 16
were working. 

    All the good news reported above applies here also.  STability is 
excellent. 
Closure levels are normal and low.  RMS for 14 and 16 are where they should
be -- as good as any VLA antenna (as expected).  Bandpasses are fine. 

    6) Q-band.  I did not reduce these data -- calibrator too weak. 

******************************************************************

    Overall observation:  A new phenomenon, not seen by me before, has 
arisen. 
The phases of *all* antennas (both EVLA and VLA) suddenly flip to new 
values (and
the deltas are not the same for the antennas) about 10 seconds into 
every scan.  This
was seen at all bands, all IFs, all calibrator sources.  It's no fluke 
(possibility of a bad
pun here, but I'll leave it for somebody else). 

    OVerall conclusion:  Other than the remaining alias at 6cm, 
everything looks quite
good indeed.