[evlatests] Solar Tests Results (at L-band)

[Rick Perley]

    I grabbed a spare hour to review and calibrate the L-band solar tests. 

    The observational sequence was this:
   
    3C84 in normal mode.
    3C84 with 20 dB attenuator in T302 on.
    Sun, in normal mode.
    Sun, with 20 dB attenuator in T302 on.

    Ken arranged the T304 attenuator setting algorithm to wait until the 
antennas were on source before setting the gains.  (That's important for 
this test!).  Examination of visibilities in PSum showed this worked as 
expected. 

    Some results:

    1) As noted earlier, the transition from normal to 20dB attenuator 
on is very different between 3C84 and Sun.   The attenuators are in the 
T302 (upstream) and we hope the T304's input attenuator to quickly (few 
seconds) to 'undo' the attenuator done by the T302.  This happens as 
expected when on 3C84 (the adjustment is made with 6 seconds) , but does 
not when on the Sun (it can take 30 seconds or more, and often does not 
end up at the right level).  The experts are aware of this. 

    2) We expect a small delay change when the attenuators are switched 
in.  It is indeed small -- maximum change in delay is about 0.2 nsec.  
Most are much less. 

    3) Similarly, a phase change is expected when the attenuators are 
in.  And this is so -- but it's not small, with typical changes of a 
radian or more.  It appears to be stable, as we did two cycles through 
the sequence, and the phases were the same on both for nearly all 
antennas. 

    4) Because we don't yet have the solar cals installed, we cannot 
calibrate the gain scale (the normal cals are completely buried by the 
solar emission).  The earlier reported offset in PDif while on the sun 
is *not* visible in the data I analyzed.  The previous report came from 
C band data.  I'll check this tomorrow. 

    5) We can compare the rough visibilities after 'calibrating' with 
the 3C84 data.  For the 'normal mode' observing there are clear fringes 
on the innermost ~10 baselines (we're in B configuration), with maximum 
amplitude of ~16 Jy.  (To put this on a proper scale requires knowing 
what the differential attenuation was when we went onto the sun).  In 
'20 db' mode, the maximum amplitude is ~18 Jy.  I consider this 
agreement to be quite good, given the difficulties the system has in 
recovering the correct attenuation while on the sun. 

    6) When on the sun, and after 'calibrating' the data, some antennas 
give clearly wrong answers, likely due to incorrect attenuator settings. 

    Overall, I regard this as quite a successful test.