[evlatests] 26jul06 data: odd behavior of ant. 16/IF 1; no fringes in fs scans

[Michael Rupen]

This is a preliminary report of some odd behavior during the first
dynamic run I've had involving fast-switching with the EVLA antennas.

             -- Michael Rupen

-------------------------------
AR603: 26jul06, 10:25-12:55 IAT
-------------------------------
LCX bands
standard continuum
3.3s averaging
reasonable weather (overcast but API 3-5 degrees)

EVLA antennas included: 13 14 16

All I've done so far is looked at the raw data.

SUMMARY:
   * Fast-switching scans give NO FRINGES for EVLA-VLA and EVLA-EVLA
     baselines, *except* for faint fringes at L band, with relatively fast
     phase winds, on the _target_ source (not the calibrator).  This seems
     to be an effect of using fast-switching during a "dynamic" run, since
     fringes were nicely detected in identical scans during
     regularly-scheduled runs on (e.g.) 20 and 23jul06.

   * Ant. 16, IF 1 (AC), shows "double sinusoidal" amplitude, with
     a period (max to max) of about 1 minute.  Phases are stable.  This
     behavior was not present on 15jul06, but was clearly visible on 20 and
     23jul06.

More details:
* Data are there for all antennas, all bands

* NO FRINGES during any fast switching scan
   at C X bands

   At L band, EVLA antennas do show fringes during fast switching scans,
   on the target source (!), but not on the calibrator.  The target source
   fringes are MUCH noisier than the VLA-VLA baselines, and also show fast
   phase winds (a turn in 2 minutes, vs. a turn in 10-ish minutes for VLA-VLA
   baselines).

   The target is too weak at C/X bands to see any such fringes.

* "Standard" (non-fast-switching) scans look fine -- nice solid fringes.
   EXCEPT ant. 16, IF 1 (AC), at 6cm.
     This shows nice stable phases, but the amplitudes show an "oscilloscope"
     (eye/double-sinusoid) pattern, varying by a factor 3 on a timescale of
     ~1 minute, with the upper/lower envelope traced by alternate 3.3s
     records.  I.e.,
                  6   8
               4           10
            2                    12
          1
              3               11
                 5       9
                     7
     and then repeating, with max-to-max being ~60 seconds (yes, I know
     this should mean 20 rather than 12 samples, this figure is just
     illustrative).  The top of the upper envelope is the correct answer --
     here 0.6 in uncalibrated units, vs. the bottom of the lower envelope
     being about 0.2 .  ["Correct" answer taken from other baselines, and
     from ant. 16 IF 2 (BD).]

   Ant. 16 IF 1 (AC) looks OK at 4cm.  The single 20cm "sit still" scan is too
   short and too weak to be sure what's going on.

I looked back at the AR603 6cm data from 15, 20, and 23jul06, all regularly
scheduled (NOT dynamic) runs.

* fs scans show nice fringes and nice behavior on all dates (see previous
   reports for sensitivity issues)

* 15jul06 does NOT show the ant. 16, IF 1 (AC) problem
   20jul06 and 23jul06 DO show the ant. 16, IF 1 (AC) problem.

   So something happened between 15 and 20jul06.

================================================================================