[Towertalk] choices

[Chuck Counselman]

At 8:45 PM +0100 10/9/02, Tom Osborne wrote:
>...I've always wondered how an antenna can be modeled in
>NEC to give a reading like "3.2db" or "5.1db" gain.  Especially
>on a vertical. It seems there are just too many variables to
>model.  How many radials?  How long?  Are they buried?  Are they
>laying on the ground?  Are they elevated?  Is it wire?  Tubing?
>If tubing, is it tapered or not.  How far apart are the ends of
>the radials?  What kind of support?  If metal, how far away is
>the wire or does it just stick up in the air by itself?  How is
>it fed?  Directly?  Through a matching device?  I think a better
>figure would be "around 3db" instead of "3.2" absolutely.

NEC-4 can model every detail that you've mentioned, and many more, 
e.g., insulation on the wire, both lumped and distributed loading, 
and traps.  If you know what you're doing, it works.  I've modeled 
antennas professionally as well as for hamming.  When you build them, 
they work as predicted, and the input impedance and resonant 
frequency are also as predicted.  At 
<http://force12inc.com/F12-ant-specs-r31.htm> in Note 2 it is stated, 
"Computations are performed using state of the art modeling software: 
NEC-4, EZNEC, AO, YO and are validated using consultants."

MININEC, NEC-2, and their derivatives cannot model conductors on or 
below the ground, and cannot accurately model conductors within a 
small fraction of a wavelength of ground.  NEC-4 can.

Regarding "around 3 dB" instead of "3.2" -- I share your disdain for 
false precision, but precision may be justified when you're comparing 
two antennas under a common set of assumptions, and you're focusing 
on differences.

73 de Chuck, W1HIS