On Fri, 4 Sep 2009 16:44:48 +0100, "christofire"
wrote:

"Jeff Liebermann" wrote in message
.. .
However, you can still get
fairly close if you make the antenna sufficiently small relative to
the operating wavelength. As the physical antenna size approaches a
point radiator, the pattern starts to look rather spherical.

That doesn't sound right. The directivity gain of an infinitesimal electric
doublet (i.e. a dipole with infinitesimal length) is about 0.4 dB less than
that of a half-wave dipole.

I'll plug a series of shortened dipoles, possibly with loading coils,
into 4NEC2 and see what happens. You may be right.

As I recall, the big holes in the pattern, that are inline with the
elements gets smaller is diameter as the antenna gets electrically
smaller. The rounded circular donut pattern tends to flatten. I
wanna play with the models to be sure. This still begs the question
of how close to spherical does the pattern need to be in order to call
it isotropic? Dunno.

I was once told a true isotropic radiator would have to be circularly
polarised

"Near isotropic circularly polarized antenna"
http://www.google.com/patents?id=saMgAAAAEBAJ
CP satellite antenna used on Intelsat V. I've been looking at the
patent for a while trying to figure out how it works.

Yeah, it should be CP because that would correctly fit the definition
of the field being identical along the sphere, in all possible
measurement antenna orientations. Note that the isotropic simulator I
posted is *NOT* circularly polarized. If you plug the deck into 4NEC2
and instead of looking at the total gain in the 3D window, look at the
vertical and horizontal gains individually, you'll see something
really ummm.... interesting.
http://802.11junk.com/jeffl/antennas/isotropic/isotrop2-vert.jpg
http://802.11junk.com/jeffl/antennas/isotropic/isotrop2-horiz.jpg
Needless to say, that the polarization is not even close to being
uniform over the sphere. (I'll add these to the menu as soon as I can
figure out what the latest JAlbum update broke in my photo
collection).

Drivel: Just got 4NEC2 setup on my new computah (Dell Optiplex 755
E8500 with 4GB). A messy tower and antenna simulation, that took over
an hour on my old PIII/1GHz clunker, now takes about 4 minutes. I'm
happy (for now).




--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558