Date: Wed, 03 Sep 2008 13:00:22 -0400
From: Josh Crabtree <jcrabtre@nrao.edu>
Subject: [Alma-feic] NSI "beam averaging" issue
Hi All, this is the latest of my correspondence with Karl from NSI,
regarding the average beam issue. He said the NSI software assumes that
the Z distance of second beam has been increased/ /by 1/4 wavelength,
resulting in a -90 degree phase shift. We have typically decreased the
Z-distance of the second beam by 1/4 wavelength. Unfortunately the
software doesn't look at the actual Z distances for each beam, it just
assumes the beam 2 has a -90 degree phase shift. We intend to use the
longer Z-distance for beam 2 from this point on. As soon as we are able
to do another scan, I'll check to see if the NSI-calculated average beam
looks the way it should.
-Josh
Karl Wrote:
Josh,
The position phase correction is doing exactly what it is supposed to do.
The key to why it is not working for your case is the fact that in your
measurements increasing the Z (probe to AUT) distance results in a positive
phase change. NSI2000 expects a negative phase change. All the Agilent/HP
analyzers produce a negative phase change as the distance between probe and
AUT increases. See section 5.2.4.1 of the manual that I sent you.
Since I assume you are down converting you may want to change the LO/RF
relationship.
Karl
Date: Wed, 03 Sep 2008 14:46:49 -0400
From: Josh Crabtree <jcrabtre@nrao.edu>
Subject: [Alma-feic] More information regarding position correction
This is a quick correction to the last email I sent. Geoff informed me
that we are actually pulling Beam 2 back 1/4 wavelength further than
beam 1, but that the resulting phase change is +90 degrees because of
the sideband we are looking at. The nsi software assumes that beam 2 has
a -90 degree shift, as I mentioned earlier. It turns out that we have
done a few scans looking at the other sideband, for which beam 2 is
shifted -90 degrees from beam 1. I chose one scan (Band 3, 2008 Scan
17), and analyzed it. Curiously, the nsi software appears to apply a -20
degree shift to beam 2 when correction is turned on, shifting beam2 even
further away from beam1. I show the results in the link below. I don't
know it this feature of the software doesn't work like it should, or if
maybe there is just something else we are overlooking. I'm still
corresponding with NSI to get this figured out. Thanks.
-Josh
--
ToddHunter - 13 Sep 2008