[evlatests] Curious Gain Variations at Ku and K bands

[Rick Perley]

     As noted in the previous message, two new phenomena have been 
uncovered by the band switch tests.

     This note details 'type A' behavior.

     At Ku band especially, but also at K band, the gains are seen to 
continuously change over the span of the 10-minute switching cycle.  To 
appreciate this better, refer to the attached plots.

     ea01-Ku-Switch.png shows the PDif and PSum values for ea01. RCP is 
shown in the top two, LCP in the bottom two.  The block on the extreme 
left are the data when switching back-and-forth with L-band, the next 
block with S band, the next with C band, the next with X, and the 
right-most block with K, Ka, and Q.

     ea02-Ku-Switch.png shows the same data for ea02.

     All show that, within each block the gains are continuously 
changing.  The effect is by far the largest on EA01 LCP -- the gain 
change is about 40% !!!.  There is no correlation between the IFs, so 
the effect cannot be caused by subreflector settings or pointing 
errors.  It is present on both AC and BD Ifs.  All antennas are affected 
at Ku-band, usually by small amounts, typically that shown by ea02.  
ea01 LCP is the only antenna-IF where the gain rises -- all others show 
a decline, typically by a few percent.

     The effect is present at K band, but less prevalent.  It seems to 
be completely absent at Ka band, while at Q-band, only one or two 
antennas show this.

     The effect is completely absent at L, S, C and X bands.

     These are true gain changes, as shown by the antenna gain plots:

     Ku-RawGain.png shows the solved antenna gains with no switched 
power -- they reflect perfectly (and inversely) the PSum and PDif values 
shown above.

     Ku-FixedGain.png shows the antenna gains following application of 
PDif.  The gain changes are completely removed, and the resulting gains 
are stable to much better than 1%.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: ea01-Ku-Switch.png
Type: image/png
Size: 33890 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20181115/094f6904/attachment-0004.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ea02-Ku-Switch.png
Type: image/png
Size: 34591 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20181115/094f6904/attachment-0005.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Ku-RawGain.png
Type: image/png
Size: 30440 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20181115/094f6904/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Ku-FixedGain.png
Type: image/png
Size: 40427 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20181115/094f6904/attachment-0007.png>