[evlatests] X-band scan-change effects

[Rick Perley]

    Michael also ran the same test at X-band as at C-band:  30 minutes 
on a strong calibrator, with scan changes once every minute. 
    The results on this test are very similar to those at C-band.

    General Issues:

    1) The amplitude stability is not as good as at C-band -- some 
'rattiness' (meaning, amplitudes that deviate from the noise 
distribution) are noted for many antennas.  However, the wind was rising 
throughout this test, reaching 15 -- 20 mph, and I think most of the 
'rattiness' in the amplitudes will be due to this.  The effect is quite 
small -- typically 1%. 
    2) Delays are all less than 5 nsec, except for 5D = -139 nsec, and 
6D = 38 nsec.
    3) All antennas fringed, except for 5C, which was all pure zeros, as 
noted at C-band.

    Specific Issues and Oddities:

    A) Antenna 6B has the same high and low states in amplitude (but no 
effect in phase) as seen at C-band.  As before the 'high' state is quite 
stable, but the 'low' state (low by 7% in amplitude) is variable -- with 
a sinusoidal shape of period 7.5 minutes!  (I see now that the C-band 
amplitudes follow the same sinusoidal variability in this 'low' state).  
As at C-band, the  X-band gain changes occur once every 57 seconds -- 
and are not commensurate with the scan boundaries. 

    B)  Antenna 20D has a faintly sinuoisdal gain change -- as seen at 
C-band.  The period is about 5 minutes, the pk-pk amplitude  about 0.3%. 

    C)  Antenna 23C displayed a remarkable, drop in amplitude, beginning 
about 12 minutes into the test, and steadily losing amplitude for 10 
minutes, when it stabilized at a new level, 3% below the original 
level.  No such effect is seen in any other IF.

    D) Antenna 14A did something quite opposite:  Starting at a low 
amplitude, and steadily gained amplitude for about 3 minutes, 
stabilizing at a level 4% above the initial level. 

    E) Antenna 3 shows the same sporadically bad behavior as at C-band.  
All four IFs behave identically.   When behaving badly, the amplitudes 
are off by 10% or so, the phases deviate by up to 100 degrees.  It's in 
this state for only a small fraction of the time. 

    F) Antenna 17 showed an 'event' 9 minutes into the run, on all four 
IFs, wherein it seemed to go offsource (amplitudes dropped rapidly to 
zero), then returned (amplitudes rose at the same rate they dropped, 
returning to the same level as before). 

    G) Bandpass stability (1-minute solutions, based on the 1-minute 
scans) was generally excellent -- fluctuations << 1%.  Exceptions are: 
    2C:  Edge effects of a few percent, lower edge only.
    5D:  Generally very ratty.  Looks like low SNR.  This IF was bad at 
C-band also.
    6B:  Very large (10% or more) deviations at both edges.  Same effect 
as at C-band.
    23C:  Significant change in BP shape (1% or so) over time.  Not 
sinusoidal. 
    25B:  Significant edge effects, same as at C-band.  (I didn't report 
this for C-band, as this antenna has an old style polarizer, so the SNR 
is very poor at C-band).
   
    H)  An odd pattern of 'pure zeros' was noted.  AIPS reported a 
significant number -- about 5%.  The pattern is curious:  IFs B and D 
had large quantitities of zeros, apparently antenna-based, during the 
first 4.5 minutes.  In A and C, the high fraction of zeros lasted a 
shorter time -- the first 40 seconds or so.  Not obvious 'zeros' were 
seen beyond these times (except for 5 C, which was zero throughout). 

    I)  Baseline 6 x 11, on IF B only, shows a sinusoidal effect, the 
same as the 'wobble' reported at C-band, but of much lower amplitude.