<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
<p>The suggestion made by Barry and others that some elevation- and
antenna- dependent effect (like droopy cables) is responsible for
the weird R-L phase plots was investigated this afternoon. <br>
</p>
<p> As expected, for those antennas showing 'even' symmetry,
plots of that phase difference against elevation showed very nice
alignment for all four sources. For these (few) antennas, the
dominant factor is elevation-dependent. <br>
</p>
<p> The question of the effects of the reference antenna is
important, so I spent a good part of the afternoon looking at this
as well. <br>
</p>
<p>Bottom line is: The choice of reference antenna is important,
but the overall message remains unchanged. Some antennas show an
'even' symmetry with HA, some antennas show an 'odd' symmetry. <br>
</p>
<p>As it turned out, my choice last week of ea10 as the reference
was probably the worst one I could have made -- ea10 has nearly
the largest 'odd' symmetry pattern, which made more antennas look
'odd' than there probably are. After trying a number of reference
antennas, I chose ea09, as this resulted in the largest number of
antennas with 'even' profiles. <br>
</p>
<p>I also simplified the processing. The plots made Friday made use
of both 'FRING" (to remove the delays) and 'CALIB' (to form the
antenna-based solutions). For the attached plots today, only
'CALIB' was used, on a narrow range of central channels. I'm
also using 1-minute averages in the solutions, as shorter time
intervals introduce effects due to the 'delay clunking'. (These
are C configuration data, so short-duration solutions show much
scatter due to this effect). <br>
</p>
<p>Attached are the R-L phase plots, as functions of (a) Hour Angle,
and (b) Elevation, for all antennas (except ea17, which had some
serious weirdness unrelated to the issues being discussed.) These
data are from C-band. There are four plots per attachment. <br>
</p>
<p>For all plots:</p>
<p>3C286 (dec = 30.3) is RED</p>
<p>OQ208 (dec = 28.5) is LIGHT GREEN</p>
<p>3C273 (dec = 2.0) is TEAL (darkish green)</p>
<p>3C287 (dec = 25.2) is BLUE.<br>
</p>
<p><br>
</p>
<p><b>Some observations:</b></p>
<p> The plots against elevation show the dominant effect for most
antennas -- a simple elevation dependence of the R-L phase. The
outstanding examples are ea01 and ea22. Although these are the
cleanest (meaning, all four sources follow the same curve), most
antennas have some (smaller) dependence on elevation as well.
Note that the phase range can be remarkably large -- 3 to 4
degrees for ea01, 3, 5, 6, 8, and 22. Note also that the sign of
the phase change is always the same -- positive -- meaning the
change in 'R' is larger than than in 'L'. <br>
</p>
<p> But not all antennas show this dependency. The use of ea09
as reference has removed many antennas from the 'odd' symmetry,
but not all. The plots against HA show that some antennas retain
significant 'odd' symmetry: ea10, ea16, and ea18 are the best
examples. Note here however that there is little sign of
elevation dependence in these antennas' phase difference. Also
note that in all cases, the magnitude of the variation is smaller
than the 'even' effects noted above. <br>
</p>
<p> Finally, some antennas are just different. ea15 is the best
example -- something else entirely going on here. <br>
</p>
<p><b>So -- What to Do?</b></p>
<p> There are at least three different things going on here. If we
are to pursue these issues, I advise starting with the simplest
and largest -- the elevation-dependent 'even' phase. It has the
cleanest signature, and the largest amplitude. The antennas with
the largest and cleanest (meaning, other dependencies are much
smaller) effects are ea01, ea03, ea05, ea06, and ea22. <br>
</p>
<p> I'm willing to conduct specialized tests, but note that,
being in the A configuration is not the most optimum for tests
involving phase... <br>
</p>
<p>Rick<br>
</p>
<p><br>
</p>
</body>
</html>