[evlatests] Frequency-dependent sensitivity

[Rick Perley]

    I did a quick test today to check if the claims made in the past 
about frequency dependent, or independent, sensitivity were still true.  
The claims were:

    A)  At L-band, at the default frequencies of 1465 MHz (IF 1), and 
1385 MHz (IF2), the sensitivity at the latter is about 30 to 40% better 
than at the former.  This effect is dependent only upon the frequency, 
and not the IF, for the default settings of the L301 and L302.

    B)  At C-band, the IF 2 default frequency of 4835 MHz is somewhat 
(say, 20%) more sensitive than the IF 1 default of 4885 MHz.  But, 
extensive tests to see if there is IF dependency, or LO dependency, have 
failed to show any clear evidence. 

    C)  At X-band, there is no frequency or IF dependence of any kind, 
for the default frequencies of 8485 and 8435 MHz. 

    So, to check if these statements are still true, a short observation 
of a strong calibrator was made, as 12.5 MHz BW, with 16 channels in 
each parallel-hand IF pair combination.  I observed in both the standard 
setups, and with the IF 1 and IF 2 frequencies interchanged. 

    Results:

    All the statements made above remain true. 

    So -- I intend to focus efforts on L-band, where the effect is by 
far the strongest, is seen on all antennas, and is clearly not dependent 
on the IF pair.  I'll try various LO combinations for the default 
frequencies, to see if there are any dependencies on these.   Other 
ideas for testing will be gratefully considered!

    One special note:

     Antennas 14 and 18 have very poor sensitivity in LCP at X-band.  
The cause is immediately found in the Tsys tables -- antenna 14 LCP Tsys 
is about 55 K, for 18 LCP it is about 45K.  Way too high!