[evlatests] EVLA Holography Phase Jumps

[Rick Perley]

    With the antennas close together, and the weather still good, we 
have begun a program of holographic measurements of the EVLA antennas. 

    When I last checked holography mode (in October), there appeared to 
be only two issues left to resolve -- one merely annoying, the other 
more serious. 

    The latter is the curious issue of the EVLA antenna phases jumping 
for sub-mode 1.    I ran a test over noontime today to see if this 
problem is still with us.  The test involved observing 10 raster scans.  
I made the even-numbered antennas the references, the odd-numbered 
antennas are the 'movers'. 

    For the impatient, the bottom line is that:  The phase jump on 
sub-mode #1 is still with us.  How annoying. 

    The effect is very simple to describe:   A holographic scan consists 
of moving the antenna in a straight line (usual in azimuth or elevation, 
but any angle can be chosen), stopping N times along the way to measure 
the amplitude and phases of the antenna pattern.  The stepping distance, 
the number of stops (N), and the start and stop positions can all be 
specified.  The 'N' stopping points are sequentially labelled as 
'sub-modes'.  Each raster point has a duration of 10 seconds.  The phase 
jumps I see are always, and only, seen in sub-mode #1 -- the first 
position. 

    The characteristics are interesting:

    1) Only EVLA antennas do this.  They do it on *every* submode=1 
observation -- with a single exception:  Antenna 1, on the 5th scan, did 
not have a phase jump. 

    2) The phase jumps are always the same on opposite polarizations, 
(i.e. A = C, and B = D), but the size of the jump is different between 
the IFs. 

    3) The transition time is very fast -- much less than 0.4 seconds, 
and the transition itself takes place 0.5 seconds after the beginning of 
the scan.  In this test, the antennas did not

    4) The phase jumps (for any given EVLA antenna and IF pair) come in 
two different magnitudes, which are of opposite sign, but not of equal 
magnitude.  The typical magnitude is ~100 degrees.   For the 10 
observations I made, there were 4 jumps one way, 6 jumps the other 
way.    There is no pattern in the sequence -- a jump in one direction 
seems to have no influence on the sign of the following jump. 

    5) Each EVLA antenna's phase jumps are unique to that antenna!  
(This is how I can tell this is an EVLA phenomenon, not a VLA one). 

    6) The jumps for A/C are always of opposite sign to those in B/D:  
When an antenna's A and C phase jump up, B and D always jump 'down'. 

    I  hope these curious features can trigger an explanation -- and a 
solution!  Since the occurance of the jumps are perfectly predictable, 
we can fairly easily flag the offending data (which, fortunately, occur 
at the furthest edges of the raster scan in the normal way we do 
holography).