[evlatests] Sensitivity Variations -- Night vs. Day, and Summer vs. Winter

[Rick Perley]

    Analysis of the switched power data has shown an interesting -- and 
unsurprising -- dependency on system temperature (and hence, 
sensitivity) between night and day. 

    The observations were of 64 objects, each observed 2 to 7 times, at 
each of the eight bands, over a 30 hour period.  During this period, the 
ambient temperature changed by about 21C.  Plotting system temperature 
vs elevation shows the expected dependency.  Plotting this for the 
night-time and day-time period individually shows a clear time 
dependency -- with system temperatures lower at night.    This is seen 
at all bands. 

    The weather was clear throughout, so the effect must be due to 
variation in ground spillover power.    To roughly quantify this, I 
plotted the elevation dependency for one well-behaved and calibrated 
antenna at each band, first for the period 1AM to 7AM, and again for the 
period 1PM to 7PM.    The elevation dependencies are the same, with the 
only difference being a simple offset. 

    The rough (error about 0.5 K) are given below:

    Freq          Diff
------------------------
   1465            1K
   2465           1.5
   4885           2.0
   8435           2.5
   14965         2.0
   22435         3.5
   37400         2.0
   43340         3.0
   48400         3.0
-----------------------

    The deviations from a smooth function are surely due to different 
spillover efficiencies -- a more careful analysis could be done, if 
there is interest in this.  With a little thought, we could probably 
deduce the spillover contribution as a function of elevation -- this 
would be a useful parameter.   
    The lower values at L and S bands are I think reasonable -- the 
ground radiation originates from much lower down, and the diurnal 
temperature change will be less. 

    The variation is significant -- about 10% of total system 
temperature at the middle bands.  Extending a bit, we can predict that 
winter night observations will be considerably more sensitive than 
summer daytime -- the ground temperature difference could easily be more 
than 40 C, for which the variation in system temperatures would be  up 
to 7K, perhaps a bit more...