[evlatests] 180 degree phase jumps, and more

[Rick Perley]

    Ken, Walter, and I continued to explore the phenomena of the sudden 
phase jumps which are so prominently revealed through the holography 
modes. 

    To review:  past tests have shown two different types of phase jumps:

    180 degree phase jumps which affect the all the EVLA antennas at 
once.  (Note, however, that the AC and BD IFs don't behave identially -- 
we've seen the 180 degree jumps on the AC side at 1275 MHz, and be 
absent on the BD side).    These jumps are generally long-lived, lasting 
many minutes.   It was earlier thought that they were uniquely 
associated with individual 'cuts' in holography mode -- more recent 
tests show this is not a perfect rule. 

    Smaller phase jumps, lasting exactly 10 seconds, which show up on 
the individual 10-second stopping points in a holographic raster.   The 
size of the jump varies.  Not all antennas jump at the same time. 

    These phenomena  are identified with holography, since no other 
tests, reported to date, have triggered these jumps. 

    Nearly all prior tests were done at L-band.  There is little data at 
any other band. 

    The new test was designed to see if the Walsh function phase 
switching is somehow responsible for either of these behaviors.   We ran 
6 in-main-beam short rasters with switching on, then 6 morre with phase 
switching off. 
This was done at L-band (with AC IFs at 1485 MHz, BD IFs at 1275 MHz), 
and at X-band (AC at 8435 MHz, BD at 8735 MHz).

    Short Answer:

    The phase jumps were seen at all frequencies both with, and without 
Walsh function switching. 

    The Details, plus other assorted phenomena:

    A)  L-Band. 

    i) 180 degree phenomenon

    As reported last week, the 180 degree phase phenomenom is very much 
worse at 1275 MHz (in the BD IFs) than at 1485 MHz (AC IFs).  I can't 
yet say whether this is a real frequency effect, or an IF pair effect. 
    At 1275 MHz, at least 180-degree phase jumps occurred within the 
30-minute duration of this test.   All four antennas jumped at the same 
time in the B and D Ifs.   The jumps are *exactly* 180 degrees (accuracy 
of perhaps 1 degree).   There are no exceptions.  Durations are variable 
-- from 7 minutes to less than one minute.
There was no change in overall behavior between Walsh switching on and 
off. 
    At 1485 MHz, no jumps occurred during the holography scans, but we 
got a clear, single, 180 degree jump between the two calibrator scans 
which separated the Walsh-on, and Walsh-off sequences. 
   
    ii) 10-second jump phenomenon

    Nine of these jumps were seen during the test -- five with Walsh on, 
and four with Walsh off.  The size of the phase jump is variable -- 
values from 20 to 150 degrees are seen.  All are 10-second long, 
exactly.   All antennas do it, and when they do, all jump together.  But 
some antennas jump less often -- for example,  18 (IFs A and C) jumped 
only 6 times, not 9 like the others.  A and C are always matched, as are 
B and D.  'Matched' means that the two opposite polarization jump at the 
same time, by the same amount.  Jump sizes, however, vary between the 
two IF pairs (i.e., between frequencies), and between antennas.  So, for 
example, 13A's phase jumps are not of the same value as 14A's, although 
the number and times of the jumps are identical. 

    iii) Special jumps phenomenon from 18 B and D. 

    Antenna 18B and 18D are jump-crazy!  Besides all the jumps described 
above, 18B and 18D, at both L-band and X-band, had an almost uncountable 
number (dozens) of extra `10-second' jumps.  All of these are of the 
same size (distinguishing them from the more generic variety described 
above) -- very close to 60 degrees.   In fact, some are 20 seconds 
long.  And, unlike the `true' 10-second jumps, these are seen within a 
calibrator scan.

    B)  X-Band. 

    i)  180-degree phenomenon:  Only a single 180-degree jump was seen, 
and it occurred between the two calibrator observations which separated 
the Walsh-on from the Walsh-off tests.  All EVLA antennas jumped 
(antenna 24 died between the two calibrator scans, so no determination 
could be made).  No change in frequency occured between these scans. 

    ii)  10-second phenomenon:  This is identical to that seen at L-band 
-- same occurance rate, same duration, same size, same relationships. 

    iii)  18B and D:  The same as at L-band. 

    C)  Amplitudes:  Other than the 400 msec duration integration 
broaching a phase jump, all amplitudes from all EVLA antennas were very 
good.  There are some minor curiosities -- I'll defer discussion until a 
later time. 

    D) Other oddities.  All EVLA -VLA antenna amplitudes went dead at 
the end of the calibrator scan, for 20 seconds, prior to the switch from 
Walsh-on to Walsh-off.  I presume this time is when the VLA was Walsh 
switching was on for the VLA and off for the EVLA, or vice-versa.