[evlatests] Sensitivity and Dropouts with Phase Switching off

[Rick Perley]

    Ken and I ran tests this afternoon with the goal of comparing
performance with, and without, phase switching on the EVLA
antennas.   As before, we used 400 msec averaging/dumping, with
50 MHz BW.  We observed a cluster of calibrators, one of which
was chosen to be rather weak (0.35 Jy) so as to let the noise
unquestionably be visible.   About 20 minutes of good data were
obtained, at X-band. 

    For reasons unknown, the data from both 14 and 16, with the
phase switching on, was of very poor quality.  Amplitude dropouts --
total dropouts -- occurred regularly in both antennas.  These are
different than the 10-second interval, partial 'attenuation' that we've
reported before.  Thus, unhappily, we have no 'before' state.

    The no-phase switching state data are very good, with the
following exceptions:

    14B is giving multiple power levels, as reported yesterday. 
    14D is just plain broken.  No sensible data were recorded.
    16A has a variable correlated power, more or less sinusoidal
with a 40 second period, and a 5% peak-peak amplitude variation
(about 0.3 dB in power units). 

    From the 5 good IFs, we find:

    a) No amplitude drop-outs at all.  Not one.  Given that this
phenomenon has been seen in *every* 400-msec averaging
experiment taken with the phase switching on (prior to today),
the absence of any dropouts in today's phase-switching-off
data is a strong argument that the phenomenon is connected
to the phase switching. 

    b) Sensitivity:  Close to normal.  16C is indistinguishable
in rms (on the weak source) from the best VLA antennas.
The other 4 good IFs (14 A and C, 16 B and D) are perhaps
15% too noisy compared to the best VLA antennas. 

    Ken notes that the delays are a bit off, and pointing not
quite right, plus some other level settings perhaps off a bit.
The sum of these (and probably some other factors) might
well recover the last 15%. 

    c) Reported Tsys.  All IFs, except 14D, report normal
backend Tsys.  14D's value was in excess of 5000 K. 
Clearly rubbish. 

    d) Phase Behavior.   No problems.  Nicely coherent,
with all fluctuations convincingly explained by atmospheric
fluctuations. 

    Next action:  Do some delay, pointing, and amplitude
level tweaking, and repeat the experiment.