[evlatests] Some Very Interesting Data ...

[Rick Perley]

    Michael took some C-band data yesterday evening, which has revealed 
some problems, some old, some new. 

    The dataset consists of a single strong point source, observed 
continuously for 30 minutes, but with a new scan every minute.  The goal 
was to see what (if anything) these new can commands did.   The 
frequencies used were 4800 (AC) and 7000 MHz (BD).  The usual 128 MHz 
BW, with 64 channels in all four polarization products for each subband.

    Some general observations:

    *  Antenna 13 has no receiver. 
    *  Antenna 28 gave no fringes on either subband at any time.
    *  Antenna 23 was tardy getting on source.   If this is recurring 
behavior, servo should be notified. 
   
    Overall, data quality was simply outstanding!  But, (as always, eh?) 
there are some notable exceptions:

    A)  Antenna 5 gave perfectly zero amplitudes in IF C, and a very 
strange spectral shape (rapid oscillations in frequency) in IF D.  The 
latter showed a delay of 136 nsec as well.  MPR says these were due to a 
formatter problem, which has been repaired.
    B) Antenna 6D has a 40 nsec delay.  All other subbands show delays 
less than 5 nsec.
    C) I generated bandpass functions for each antenna/IF/polarization 
for each 1 minute scan.  Perusal of these solutions should fabulous 
bandpass stability (changes << 1% from scan to scan), *except* for
             2C:  turnups or turndowns of ~1% at lower edge,
             6B: very large -- (10s of percent) -- turnups or turndowns 
at both edges edges, with poor stability in the central channels;
             11B:  Same as 6B, but not as bad -- few percent.
             17 has poor sensitivity in B and D -- this would be 
expected if there was an old-style polarizer in it, but according to the 
EVLA retrofit sheet, it's supposed to have been retrofitted.

    D)  Antenna 5 drops amplitude by 15% on every 1-minute scan change, 
on all IFs.  No other antenna showed this. 
    E)  Antenna 6Bchanges amplitude by ~ 6% every 57.6 seconds.  
NB!!!!:  This behavior is *not* connected to the 1-minute scan changes.  
The nature of the two amplitude states is curious:  The higher state is 
relatively
stable -- every other transition comes to the same level.  The other 
(lower) state is variable in amplitude -- and within the 57 seconds 
spent in that state, the amplitude usually (but not always) is changing. 
    F)  Major-League Wobbles are found!!!  Baseline 6 x 11, IF 'B' only, 
has a 25% sinusoidal amplitude oscillation, with two periods clearly 
visible to the eye:  116 seconds and 12.5 minutes.  Baseline 11 x 25, 
also IF 'B', has a much weaker wobble, 7% amplitude (pk-pk), with 129 
second period.  There is *no phase* oscillations seen on either 
baseline.  There could be more wobbles out there -- I haven't looked at 
all baselines on all IFs.  If there are, they can't be as large as the 
ones I found (which were large enough to show up in CALIB solutions). 
    G) Antenna 3, all IFs shows some odd amplitude/phase drops and 
changes, of small (few percent) amplitude, of which there is no easy 
description. 
    H) Antenna 20D shows a faint amplitude oscillation 0.5%) of ~4 
minute period, which does not appear to be due to baseline-dependent 
wobblese.
    I) Antenna 5A shows two phase states, separated by 60 degrees.  
There were 13 phase jumps, *each of which* exactly coincided with a scan 
boundary. 

   
    E) An