[evlatests] Results of solar 20-db test

[Rick Perley]

    A script designed to test the ability of the EVLA to observe the sun 
was run last Thursday. 

    The array observed a strong calibrator (3C84), and the sun both in 
'normal' mode, and with the 20-dB attenuators located in the LSC 
converter (T302 modeule) turned on.  Observations were made in L, S, and 
C bands.  Calculations suggest that the electronics has sufficient 
dynamic range to handle the quiet sun with the attenuators off, and that 
the T304 downconverter module has sufficient attenuator range to adjust 
the overall gain so the samplers are kept happy for the quiet sun with, 
and without, the 20 dB attenuators.   (The 20 dB attenuators' major 
purpose is to permit observations of moderate solar flares.  Major 
flares, which will saturate the front ends require the 'Lilly' scheme). 

    All expectations were met.  Nice stable fringes were obtained both 
with and without the 20 dB attenuators on both the calibrator and on the 
Sun.  (Minor problem:  due to an error in the script, the attenuator-off 
observations of the sun at L-band were made with the attenuators set to 
cold sky values).   Delay differences of a few nsec were found between 
the attenuator-on and attenuator-off observations.    

    Not all is rosy, however.  The algorithm which decides how the T304 
attenuators are adjusted to maintain proper signal levels to the 
samplers (and avoids saturation within the module) is clearly 
inefficient, and makes very strange choices for many antennas.   The 
switched power (PDif) values while on the sun are biased negative -- 
they should be noisy with a zero mean, or at most a very slight positive 
mean.