[TowerTalk] K8UR wire 4-SQR characteristics vs 1/4 vertical
array design from N0AH
L. B. Cebik
cebik@utkux.utcc.utk.edu
Thu, 10 Aug 2000 08:44:44 -0400 (EDT)
On Wed, 9 Aug 2000, Tom Rauch wrote:
>It's been my experience that NEC tends, for some reason, to under-
>estimate these losses.
>
>73, Tom W8JI
Tom is correct that NEC-2 underestimates losses for most soil grades
when using either no radials in a model or using the limited facility
of placing radials very close to the ground--say about 0.001 wl above
ground. In fact, neither the use of the no-radial MININEC ground
(available in EZNEC's version of NEC-2) nor the close-proximity substitute
for a buried radial system is an adequate substitute for a model of a
buried radial system if one is modeling a buried radial system.
Of the available commercial software, only NEC-4 (GNEC or EZNEC
Pro/4) will model buried radial systems, and there are some modeling care
points necessary to get a model within the constraints of
NEC-4. Nevertheless, over very poor soil on 1.83 MHz with a 1/4 wl
monopole and various size radial fields ranging from 4 to 128 radials,
each 1/4 wl long and buried .001 wl deep (about 6.5"), the far-field gain
varied by about 4 dB from the smallest to the largest size of radial
field. Neither radial fields elevated 0.001 wl nor no-ground MININEC
models showed anything like this range of far-field gain variation
over very poor soil. The better the soil used in the model, the smaller
the variation, but be careful not to over-estimate your soil conductivity:
make measurements. As well, do not confuse increases in far-field gain
that result from have better soil quality at a distance in the reflection
zone with antenna proximity effects of better soil qualities. (You can
sort out this latter point by using multiple ground qualities with a
sensible radius for the inner zone.)
The upshot is that--for modeling radials that are in fact buried--nothing
short of a model that also uses buried radials will give close to useful
results. NEC-2 is not up to this task, and attempts at approximations
become more seriously misleading as soil quality decreases. As well, the
use of the MININEC ground is also fatally misleading on any vertical array
using one or more sloping elements, whether driven or parasitic, with
errors increasing with the amount of departure of the element(s) from
vertical. My tentative conclusion so far from looking at dozens of models
in articles and other sources upon which we tend to rely is that the
modeling will all have to be done over again using more adequate software
to the task. The next generation of models--which will be more adequate,
but not necessarily close to perfect--will be very large if they properly
model all of the physical radials used by the actual antenna system being
modeled. Only when we use adequate models can we sort out the questions
of basic software limitations from those of simply using inadequate
models.
This does not mean that NEC-4 resolves all issues of modeling radial
systems. There are cases for which NEC-4 produces no usable
results. Other cases, such as the physical radial field placed on or very
close to ground, remain objects of study in the pursuit of correlations
and deviations between physical antenna systems and modeled ones. Nor
does NEC-4 overcome problems associated with stratified vs. homogenous
soils beneath antennas. Those are just starter problems in the entire
cluster.
I plan to make my stock of data gathered so far on comparisons of modeled
radial fields for various vertical array types accessible, but the venue
is not yet established. Eventually, it will be at the site.
-73-
LB, W4RNL
L. B. Cebik, W4RNL /\ /\ * / / / Tel: (865) 938-6335
1434 High Mesa Drive / \/ \/\ ----/\--- http://www.cebik.com
Knoxville, Tennessee /\ \ \ \ / / || / e-mail: cebik@utk.edu
37938-4443 USA / \ \ \ \ || e-mail: w4rnl@arrl.org
--
FAQ on WWW: http://www.contesting.com/FAQ/towertalk
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com