[evlatests] X-Band Fast Wobbles Found

[Rick Perley]

    The X-band dataset from yesterday's tests was taken with 0.1 second 
integrations, allowing a search for the infamous fast wobbles.  These 
were not hard to find ...

    The data were calibrated for delay (a single global solution), and 
for amplitude and phase (a 1 minute smoothing applied to both).  VPLOT 
was used to view the time-ordered data, using a vector average over the 
central 56 channels (100 to 156).  This was done so the delay 'clunks' 
would not be visible.  (And they weren't.  Similar plots of channels 
selected at the edges showed the 'clunks' quite nicely). 

    Wobbles were found in both amplitude and phase -- but with 
strikingly different characteristics. 

    A) Amplitude.  A 0.25 second periodicity, with pk-pk amplitude of 10 
-- 15 %, is seen on all baselines connected to antenna 27, in LCP only.  
The same phenomenon is seen on all four sub-bands.  There is clear 
evidence on some of the baselines that there are two periodicities 
present:  0.25 and 0.5 seconds.  (A beautiful double beat is seen, most 
clearly on 15 x 27, but also visible on 8 x 27).  This periodicity does 
not close (meaning, a CALIB solution will not remove it). 

    B) Phase.  A 0.25 second periodicity, with pk-pk amplitude of 4 
degrees, is seen *only* on baseline 27 x 28, in *both* RCP and LCP.   
The periodicity is seen on all sub-bands.