[evlatests] Antenna 16 gain variations at C band

[Juan Uson]

Dear All,

I have analyzed the two data sets obtained on 3C84 (C band, ~23 Jy,  
July 22, 6 hours) and 3C147 (C band, ~8 Jy, July 23, 3 hours).  Each  
calibrator was observed on-axis every 12 minutes for a duration of 3  
minutes.  On July 22, antennas 14 and 16 were present, 13, 14 and 16  
were used on July 23.

In addition to other (minor) effects, antenna 16 showed a few  
distinct "gain" states on both days.  The enclosed plots show the  
results of an amplitude and phase self-calibration on the time scale  
of the 10 second integration that I used.  Previously, a 10 sec  
"phase only" solution was computed and applied.  The phase  
corrections were normal, mostly within +/- 5 degrees and mostly due  
to the atmosphere as evidenced by the changes in the phase difference  
being of order of 1 degree.

The subsequent amplitude self-calibration gave corrections of ~ 0.5 %  
on most antennas, with a few instances of ~ 2% corrections on some  
antennas.  Antennas 13 (second day only) and 14 (both days) behaved  
in this way.  However, antenna 16 showed a few "states" with gain  
differences up to +/- 10% from the average.  One can see between 2  
and 5 distinct such groups and the gain changes between them happen  
on the 10-second dump time but not in an apparently regular fashion.   
Moreover the ratio of the RCP and LCP gains does not vary by more  
than 1.5% so whatever is doing this seems to be affecting both  
polarizations simultaneously.  I have plotted the weights assigned by  
AIPS task FILLM for all antennas as a function of time and there is  
no noticeable difference between antennas, so that the gain changes  
are not reflected in the derived weights.  I have also plotted the  
"system temperatures" which are designated in AIPS as Tsys (with  
values usually about a fraction of unity that are used to derive the  
weights) and there are also no apparent effects.  AIPS also stores a  
second value related to system temperature (which it designates Tant  
and which is usually of ~40-50 "Kelvin" and is used by the task that  
assigns flags based on "antenna temperature" and the entries for the  
EVLA antennas are all zero.

It seems to me that the symptoms might correspond to a change of  
"scale factor."  Given that both polarizations change in synch, I do  
not see that this would be due to the data transmission (digitizing  
and recreating the signal for the VLA correlator) because it has to  
affect both polarizations the same way at the same time.  This could  
happen if the cal signal was fired at a slightly different frequency  
than it is sampled and my data are using a 10-second average to  
sample the beat (an offset would seem to only affect the scale but  
would not vary it).  There are probably a few other ways to cause it ...

I have posted a copy of this note with the plots in the testresult  
area as 20060727.ju

Cheers,

Juan


Encl.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: PLOTAMP7.PS
Type: application/postscript
Size: 122408 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20060727/cfa510dc/attachment.ai>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PLOTAMP8.PS
Type: application/postscript
Size: 126266 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20060727/cfa510dc/attachment-0001.ai>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PL3C147AMP7.PS
Type: application/postscript
Size: 58454 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20060727/cfa510dc/attachment-0002.ai>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PL3C147AMP8.PS
Type: application/postscript
Size: 69061 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20060727/cfa510dc/attachment-0003.ai>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PL3C84AMP-RoverL.ps
Type: application/postscript
Size: 87510 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20060727/cfa510dc/attachment.ps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PL3C147AMP-RoverL.ps
Type: application/postscript
Size: 63753 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20060727/cfa510dc/attachment-0001.ps>
-------------- next part --------------