[evlatests] High Resolution L-band Antenna Holography Images

[rperley]

There was considerable unused time during the recent A -> D 
configuration move.  I utilized these empty blocks to generate 
high-resolution images of the E-field illumination of the VLA antennas 
at L-band.  In all, 14 separate observations were required to piece 
together the 101 x 101 grid.  Half of the sample (l,m) plane was done in 
a single 16-hour block, the outer areas required the additional 13 
smaller observations.

There was no specific goal in doing this -- the results are of no use 
for antenna panel settings, as these are done with high frequency 
holography.  It was anticipated the diffractive effects will dominate 
the images -- and they do, as a quick peruse of the attached images will 
show.  One useful result of this effort will perhaps be an understanding 
of the origin of the complex structures seen in these images.

To keep data volumes with reason, I elected to correlate only two 
spectral windows.  (I regret this, in hindsight).  Scheduling was 
tricky, as the target source (3C147, conveniently unpolarized) transits 
at a high elevation (76 degrees), which when combined with the large 
throw angles (up to 18 degrees) can make for long travel times in 
azimuth.  As many of the antennas were in the D configuration, shadowing 
is a major concern.  And, as the pointing position of the 'moving' 
antennas is often very different (18 degrees different) than that of the 
reference antennas, the standing shadowing algorithm is not valid.  Eric 
Greisen wrote a special new task (SHAHO) to generate the appropriate 
shadow flags for both moving and reference antennas for holography.

Six reference antennas were employed.  The resulting images are from the 
phased sum of the reference antennas correlated against the phased sum 
of the moving antennas.  SNR is good, even for the furthest-out pointing 
positions (where the RR and LL voltage correlations are down about -30 
dB w.r.t. the center).  I have not yet attempted individual antenna 
images.

The on-axis cross-polarization has been removed, and the displayed 
images are properly in the antenna frame of reference.

Attached are 4 PDF files, showing the antenna complex electric field 
weighting, for the two frequencies (1445 and 1820 MHz).  Both sets are 
made with only 18 MHz bandwidth, to prevent spectral blurring of the 
outer regions of the beams.

The files ending in 'CORR' show the four correlations:  RR, LL, RL and 
LR from left to right.  The amplitudes are in the top row, the phases in 
the bottom.  Greyscales are not the same -- the parallel hands use an 
amplitude wedge from 2 to 120 units, phases from -100 to +100 degrees.  
The cross-hand patterns have amplitudes from 2 to 60, and phases -180 to 
180.

The files ending in 'IQUV' show the Stokes images, I, V, Q, U from left 
to right.  Amplitudes on top, phases below.  Greyscales are the same as 
for the correlator images -- 2 to 120 for 'I', and 2 to 60 for 'Q, U, 
and V'.  Phases -100 to 100 for I, -180 to 180 for Q, U, V.

The physical resolution of these images is about 30cm -- the highest yet 
done for VLA antennas at any band.  The Fourier transforms were done 
without weighting, so the 'dirty beam' is a perfect Fraunhoffer 
diffraction pattern, since the sampling region is a perfect square (with 
a few holes due to missing data points).  I'll later try imaging with a 
taper, to reduce the ringing (at the cost of resolution).

The obvious first conclusion is that diffraction is responsible for the 
fine-scale structure seen in the images.  Beyond that, I'll cheerfully 
admit that I don't have explanations for any of the curious features 
seen in these images.

It would be very interest to match these results against simulations 
produced by, for example, GRASP8.

I have the calibrated data files (amplitude phase as a function of 
offset direction cosines) carefully (painfully, actually) extracted from 
the 14 databases, available to all.

I proper memo is under construction.

At a later time, I"ll distribute what the beam images look like.


Rick
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 1445-IQUV.pdf
Type: application/pdf
Size: 103832 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20210402/b4be33af/attachment-0004.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 1445-CORR.pdf
Type: application/pdf
Size: 98325 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20210402/b4be33af/attachment-0005.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 1820-IQUV.pdf
Type: application/pdf
Size: 120114 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20210402/b4be33af/attachment-0006.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 1820-CORR.pdf
Type: application/pdf
Size: 111128 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20210402/b4be33af/attachment-0007.pdf>