[evlatests] S-Band Performance

[Rick Perley]

    A short report on S-band performance during the 'flux densities' run. 

    Executive Summary:  With a single exception, amplitude stability 
(after gain correction through application of 'switched power') is 
better than 1% over the 30 hour period. 

    Some details:

    Observing setup same as at C-band. 
    Eleven antennas fringed well:  6 7 8 9 10 12 14 15 24 27 28

    All antennas show the same 24-hour diurnal gain periodicity as seen 
at other bands -- 4% amplitude (power).  As at the other bands, this is 
removed to better than 1% by application of PDif power monitor. 
    Three antennas show poor stability in their switched power monitor:  
6RCP (both IF 'A' and 'B'), 7RCP (A and B) and 27LCP (C and D).  Of 
these, 6RCP is by far the worst -- with 'noise' or variations of nearly 
50% in the PDif value.  This is so large that these data had to be 
discarded.  Even after careful editing, the post-correction gains were 
unstable to 5%. 
    The reported system temperature for antenna 7 is far higher than the 
others, especially in the lower half of the band (80 K).   However, this 
is almost certainly an error in the recorded Tcals, as the range in 
reported Tsys from source to source is also a factor of 2 to 3 too high. 
    Two subbands are heavily affected by interference, centered near 
2180 and 2325 MHz.  (The latter of these is digital radio).  Despite the 
strong RFI, channels outside the broadcast frequency span, but within 
the subband, appear to be fine, and calibrated nicely. 
    The strong RFI in these two subbands does however destroy the 
switched power calibration system (as expected).  Eric has written a new 
task 'SYCOP' that permits the  user to copy the SY tables from an 
RFI-free subband to any subband with comprised switched power.  This 
should be a valid procedure for monitoring and correcting electronic 
gain variations, but will introduce a scaling which should be removed by 
regular  calibration on a source of know flux density. 

    Elevation gains were fitted -- as expected, the coefficients are 
very small, and virtually certain to be due to small elevation 
dependencies in the electronics. 

    Polarization calibration was as spectacularly successful in this 
band as in the others done so far (Ku, X, and C).  The effect of 
satellite downlinks is seen in 3C138 (dec = 16) at frequencies above 
3.70 GHz.  (The other three polarization calibrators are much further 
north, and were not affected). 

    Bandpass stability was about as good as can be expected at these 
frequencies (pk-pk of a couple percent). 

    Phase stability was only modest -- ten or twenty degrees from source 
to source.  The patterns are the same as seen at higher frequencies, and 
it appears the scatter is proportional to frequency.  This might suggest 
atmosphere, but I'm dubious, given the good weather and short 
baselines.  A close analysis will be needed (and is being deferred until 
I can finish reduction of the remaining four frequency bands).