[evlatests] Failures to Fringe -- Part 2

[Rick Perley]

    A second test to measure the variation of antenna gain with 
elevation (and, it turns out, other variations as well!) was run last 
Friday night.  Because it is now clear that temperature variations are 
influencing the electronic gains (and, more ominously, the power from 
the calibration diode which is used to remove these gain variations), 
this 2nd test was crafted to minimize these influences by running 
between midnight and dawn. 

    The object selected was J0741+3112, which was tracked from its rise 
to its transit -- an elevation range from 8 through 87 degrees.  The 
weather was clear, dry, and calm.  Temperatures declined on average 
through the run from +4C at midnight to -2C at dawn -- there were 
significant periodic temperature oscillations of approximately 1 hour 
period, with an 8C peak-peak, throughout most of the night. 
   
    The band sequence was:  X(Ref) - X - C - S - Ku - K - Q - Ka - L , 
with about 1 minute spent at each band.  The X(Ref) frequency pair was 
different than than the following X observation -- the BD side was 
changed from 8.7 to 11.3 GHz when I went from the referenced pointing 
calibration to the on-source gain observation.  This sequence is 
slightly different than the first test -- for that one, I used C-band as 
the referenced pointing band (because one antenna had a non-functioning 
X-band receiver), which was followed by a C-band gain observation, again 
with a different frequency selection. 

    Recall that the first test (the previous week) showed a high rate of 
fringing failures in C and Ka bands, but only on the BD side.  For that 
test, the C-band observation followed the C-band referenced pointing 
(with a different frequency selection), and the Ka-band observation 
followed Q-band.  For this Ka-band observation, the BD frequency 
selected was 29.5 GHz.  (Note -- none of these 'fringing failures' was 
flagged by the system). 

    For this second test, the only significant unflagged 'failures to 
fringe' occurred at Ka-band -- which as before followed the Q-band 
observation.  For reference, the overall statistics are given below.  
There were 39 separate observations for each band -- a total of 27*40 = 1080
 tuning changes. 

    Band        Number of failures
------------------------------------------
    X                  2 (ea14 on all IFs, and ea24 on AC only)
    C                  3 (ea11 twice on all IFs, 24A on AC only -- not 
the same time as the failure above).
    S                  2 (ea23 and ea15, all IFs)
    Ku                0
    K                  1 (ea24 on AC only -- different time than those 
at X and C bands)
    Q                  0
    Ka                40 -- 38 of which are only on the BD side. (ea24AC 
and ea18AC also failed once each).
    L                   0
---------------------------------------
   
    For the Ka-band fringe failures:  As before, the distribution 
appears to be flat amongst the antennas.