[evlatests] P-band bandpass issues, and something about ea13 ...

[George Moellenbrock]

Rick-

What about the cross-hand phase (_NOT_ == pos ang)?  I gather you are not 
calibrating this because if you were, you'd be able to assert orientation 
offsets directly by quoting just the real parts of the D-terms, rather 
than depending on orientation offsets dominating the total instrumental 
polarization.  No doubt there is some degree of non-zero ellipticity in 
each dipole, too...  Maybe ea13 has a uniquely large ellipticity 
dominating it?

Since the cross-hand phase is not calibrated, it is not clear to me what 
the Q and U you estimate really are.  If Q and U are derived formally from 
the correlations (all _four_, not just XY and YX, right?), a non-zero 
cross-hand phase residual will distort the result.  Even though you are 
expecting Q=U=0, I don't think you are actually measuring them ideally....

-George


On Thu, 3 Aug 2017, Rick Perley wrote:

>    We are preparing for another 'moon-polarization' observation, to occur 
> Sunday night.  For this, a very short observation was taken last night, for 
> the purpose of establishing suitable nearby phase calibrators.  Key results:
>
>    1) All 27 antennas in the array (15 is in the barn) worked well, with 
> similar sensitivities.
>
>    2) However, 3 antennas have very bad bandpasses, with very high 3.2 MHz 
> ripple.  These are:
>
>    7X, 22X, 26X, 26Y.
>
>    3) I did a polarization calibration.  As usual, there is a wide range of 
> leakage values.  Since our (AIPS) software does a 'relative' solution, the 
> values obtained reflect the reference antenna's cross-polarization.   If (as 
> I believe) the major contributor to the cross-polarization is the antenna 
> orientation, then the solution amplitudes reflect the mis-orientation of the 
> antennas w.r.t. the reference.  Frank tells me that he believes ea20 is the 
> antenna with the guaranteed correctly oriented dipoles.  So, I list below the 
> cross-polarization amplitudes for all the antennas.  I note that the 'X' and 
> 'Y' solutions are in all cases nearly identical -- this is strong evidence to 
> support the contention that nearly all the cross-polarization is due to 
> dipole misorientation.
>
>    Antenna        Cross-Pol Ampl
>
> ----------------------------------------------
>
>    1                         2%
>
>    2                       15
>
>    3                       13
>
>    4                       14
>
>    5                         3
>
>    6                       11
>
>    7                         3
>
>    8                         2
>
>    9                         4
>
>    10                     14
>
>    11                       8
>
>    12                       2
>
>    13                     21
>
>    14                     12
>
>    15                      out
>
>    16                      11
>
>    17                       7
>
>    18                      12
>
>    19                      14
>
>    20                      ref
>
>    21                      14
>
>    22                      13
>
>    23                      12
>
>    24                      20
>
>    25                       4
>
>    26                       4
>
>    27                      17
>
>    28                        2
>
> ----------------------------------------
>
>    There is a remarkable grouping here -- these values are not evenly 
> distributed:
>
>    There are nine antennas between 0 and 4%:  1, 5, 7, 8, 9, 12, 25, 26 and 
> 28.
>
>    There eleven antennas between 10 and 14%: 3, 4, 6, 10, 14, 16, 18 19 21, 
> 22, and 23.
>
>    There are four antennas at 15% or greater:  2 (15%), 13 (21%), 24 (20%) 
> and 27 (17%)
>
>    The two remaining antennas (11 and 17) are 8 and 7%, respectively.
>
> -------------------------------------------------
>
>    4) The most interesting result comes from looking at the results of 
> applying the polarization calibration to the data.  As always noted before:
>
>    a) The values of the cross-hand channels (XY and YX) are up to 20% of the 
> parallel hand channels (XX and YY) without applying the polarization 
> calibration.  This is consistent with the results cited above.
>
>    b) The values of the cross-hand channels after polarization calibration 
> are all below 2% (in the noise) of the parallel-hand channels.  Looking good 
> .. but ...
>
>    c) The values of Q and U (which are formed by combining XY and YX), for 
> all antennas but one, are also in the noise -- this is the correct result. 
> But for one antenna -- *ea13*, the result of forming Q and U is very 
> different -- large values are obtained.  The only way I can think of that we 
> have have proper XY and YX amplitudes, but improper Q and U values, is if the 
> phase relationship between the dipoles is incorrect.  The orientation or 
> phasing of the dipoles in ea13 is different than that of the other antennas.
>
>    *It would be very useful for somebody to climb up ea13, and check if its 
> P-band dipole connections are different than the other antennas*.
>
>    Rick
>
>
> _______________________________________________
> evlatests mailing list
> evlatests at listmgr.nrao.edu
> https://listmgr.nrao.edu/mailman/listinfo/evlatests
>
>

--