[evlatests] Odd flagging, and integration issues

[Rick Perley]

    We ran an extensive test of high frequency capabilities last night.  
Results will be distributed later (once we understand them!).  Here I 
report on two peculiar issues regarding flagging and integration:

    The observations were of three sources, each observed twice (once 
with subreflector rotation off, once with it on).  Each 'observation' 
consisted of a calibrator observation (without referenced pointing), a 
referenced pointing determination, then calibration (with referenced 
pointing applied), then a small holography raster. 

    A)  Each calibration observation which followed a referenced 
pointing observation started with ~15 seconds of  unflagged data which 
were clearly of an incorrect amplitude.  It looks as though the system 
is applying the referenced pointing offsets during this period, as the 
amplitudes are all changing upwards during the 15 seconds.  This 
phenomenon was seen on all six (three sources, observed twice each) 
occasions.    The flagger is apparently unaware that something is being 
modified. 

    B) On two occasions, the calibrator observation which *preceded* the 
referenced pointing observation was observed with 10 second integration, 
rather than the 1.66 seconds which was requested.   As  referenced 
pointing observations are made with 10-second integration, it seems 
clear the system, on occasion, is 'getting ahead of itself'.