[evlatests] VLA antenna gain-elevation determination

[rperley]

As we have completed the determination of the subreflector rotation 
coefficients (except for ea01) for improving the antenna gains at high 
frequencies, it seemed appropriate to redetermine the gain-elevation 
dependencies.  The data were taken the night before last.  The data 
quality was outstanding, and the post-fit residuals, for nearly all 
antennas at the high frequencies, is of order 4 arcseconds.  The 
coefficient tables will be separately sent to the appropriate people 
(but if you want them, let me know, and I'll forward them).  I've 
attached the plots of the gains vs. elevation, along with the best-fit 
curves.  These are produced by the AIPS program 'ELINT'.

The observing procedure was described yesterday.  Antennas 14 and 19 
were out of the array.  Data were taken from X through Q bands, at 
elevations between 11 and 82 degrees.  At each band, two spectral 
windows were utilized, whose frequencies were chosen to span the entire 
bandwidth, using the 8-bit samplers.  The solutions were determined 
separately for each polarization, for each of these two frequencies.  At 
X-band, these four solutions can be rather different, indicating effects 
not directly associated with elevation.  (In all such cases, the gain 
variations are very small).  At all other bands, the four solutions are 
very similar.  There are no solutions for ea01 at Q-band -- due to 
repeated failures in referenced pointing (no surprise, given the current 
situation with that antenna).

Some other points of general interest:

1) Referenced pointing was done for every antenna, at every band, for 
each of the 23 elevation observations.  Discounting the ea09 shadowing 
issue, and ea01 at Q-band, the referenced pointing algorithm gave a good 
result for 2928 out of 2947 determinations.  Nearly all these failures 
were at Q-band, at elevations below 20 degrees, where we know the 
primary beam is severely distorted.  Indeed, for X, Ku, and K bands, 
there was only a single failure, and at Ka band, only 5 failures, 4 of 
which were on ea01.

2) The observations were made between midnight at 6AM, during which time 
the ambient temperature dropped considerably.  This had a strong effect 
of the antenna gains, as the post-amps are directly attached to the 
feeds, which extend out into the cold night air.   Examining the 
switched power, it was noted that the system gains changed monotonically 
by 10 to 20% at all bands -- an effect larger in  most cases than the 
elevation gain  dependency!  Thus, in order to determine the actual 
elevation gains, and not the temperature coefficients of the post-amps, 
the switched power values were applied to the data.  (There remains one 
problem -- the temperature coefficients of the noise diodes.  
Fortunately, these seem to be much smaller than those of the post-amps). 
  Comparison of the gain curves with, and without, the application of the 
switched power shows a dramatic difference.  A conclusion which should 
be made from this is that, for proper correction of the elevation gains 
through use of these coefficients, the temperature sensitivity of the 
electronics should be accounted for through use of the switched power.  
Unfortunately, at the present time, this cannot be properly done when 
observations are made with the 3-bit quantizers.

3) Besides the universal temperature-induced gain changes, there were a 
small number of antennas whose gains 'did weird things', during the 
observations.  'Weird' here means jumps of tens of percent.  All of 
these were perfectly corrected by the application of switched power.  
Likely, these are issues associated with the band switches.  In no case 
were these 'weirdnesses' of a 'shredded' nature -- the gain change was 
the same throughout the observations, thus permitting effective 
corrections through application of the switched power.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ElGain-Q-May2020
Type: application/pdf
Size: 426390 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20200527/1f6ce9d8/attachment-0005.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ElGain-Ka-May2020
Type: application/pdf
Size: 424027 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20200527/1f6ce9d8/attachment-0006.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ElGain-K-May2020
Type: application/pdf
Size: 444645 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20200527/1f6ce9d8/attachment-0007.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ElGain-Ku-May2020
Type: application/pdf
Size: 445780 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20200527/1f6ce9d8/attachment-0008.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ElGain-X-May2020
Type: application/pdf
Size: 445843 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20200527/1f6ce9d8/attachment-0009.pdf>