[evlatests] Bad Power Setting

[Rick Perley]

    I ran a test on Saturday evening, utilizing Ku and K bands.   At 
Ku-band, the (center) frequencies utilized were 14.645 and 17.488 GHz, 
for A/C and B/D, respectively.  The 8-bit system was used. 

    The operator noted in the log that the power levels were far too 
high in certain antennas, and successfully reset the affected antennas 
in the middle of the run.  (Kudos to Operations for noticing, and doing 
this). 

    The settings were indeed *way* out of range.  Looking at the PSum 
and PDif values, I find that, for some unknown reason, the system set a 
power level about a factor *40* times too high!!!  That is -- the T304 
attenuators were about 16 dB below where they should have been.  The 
affected antennas were: ea02B, ea09B and D, ea13B, ea21B, ea22D, ea25D, 
ea26B, and ea28D. 

    The script was set so as to give more than adequate time at the 
beginning of the run for the proper power levels to be set.  The other 
three frequencies settings (two at K band, and the other Ku band 
setting) were fine. 

    Unsurprisingly, performance of the affected antennas during the 
period of excessive power was considerably degraded.  Claimed system 
temperatures were a factor of a few higher -- however, these cannot be 
trusted, as the 'PDif Compression' problem was notably worsened for this 
period.  This is strong evidence for saturation within the T304 and/or 
the digitizers.