[TowerTalk] K8UR wire 4-SQR characteristics vs 1/4 vertical array design from N0AH

[L. B. Cebik]

On Thu, 10 Aug 2000, Guy Olinger wrote:

>How about in EZNEC, using a radius of near-perfect ground for the  extent
of the radials, and the actual ground after that, with a direct 
connection to ground in the model, versus the same with a single 
actual ground. This would seem to describe the degree of sensitivity, 
though it would not help to estimate less than maximal radial 
configurations. 
>
Actually, the use of good quality ground to approximate a radial system is
not an especially good approximation of the action of buried radials.  If
you look at Rudy Severns article in last months QST, you will see that the
calculations of ground effects involve some complexities such that the
interaction of ground radials modeled as elements do not come out the
same.

The local ground/distant ground line is useful for determining far field
gain with different distant soils without changing the conditions directly
under the antenna.  It is useful, for example, in setting up an antenna
near an ocean line or near an abrupt change of terrain type of
(lower) height.

Interestingly, there are some simple antennas, such a quarter-wavelength
monopole, where models using 32 radials above and below ground (within
0.001 wl of the surface) tend to converge for good or better soils,
although there are differences in reported impedances.  However, not all
arrays show similar convergences, so this is mostly an artifact, rather
than a usable convergence.

In the end, shortcut models are unreliable until and unless compared to
full and complete models meeting all known software constraints.  Only
then can their reliability be assessed as models.  At that point, of
course, one no longer needs the shortcut model--just the patience to run
the large model.  And it is the large model and its physical correlate
that need comparison to assess the limits of the software.

-73-

LB, W4RNL



--
FAQ on WWW:               http://www.contesting.com/FAQ/towertalk
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com