>My input on the model would be this:
>Models amateurs use generally assume FS at some very large
>distance from the antenna. You can tell if the model is
>doing this by looking at FS at zero degrees. If you see FS
>at zero degrees is very low or nearly zero, you know the
>model assumes you are interested in FS at a distance of
>hundreds of wavelengths or more.
Most antenna modeling programs do not model the propagation medium (i.e.
the atmosphere). And, they make the "flat earth" and "uniform earth"
The pattern data is typically at infinite distance, referred to an
isotropic radiator. Field strengths (the V/m numbers) are typically
normalized to something like 1 meter distance (and then you can calculate
field strength by dividing by the radius to your observation point).
For low angles, where the "reflection point" would be many wavelengths
away, most modeling programs (e.g. NEC) use the simple reflection
coefficient calculation. Furthermore, since they don't model the curvature
of the earth, for an antenna that is high above the ground, that can make a
difference. Put that antenna up at 300 ft (100m)and the radio horizon is
around 25 miles (40 km) away. What this means is that, to the antenna, it
looks like you're on a gently sloping down hill (1:400 grade) (granted,
this is fractions of a degree..)
If you want to really model the interaction, you'd need to run a free space
pattern, then plug that into something that models the interactions with
>That really isn't how the antenna works in the real world.
>The real world isn't flat and on 160 meters the height of
>the ionosphere often compares to the distance the model uses
>to calculate pattern. The result of this is low angle FS is
Since the vast majority of modeling programs don't model the medium, the
ionosphere (or atmospheric refractivity, for that matter) don't feature
into it. The model just ignores it. What you're really talking about here
is a need for a better propagation model. HFTA does part of it, but only
does horizontal polarization where you can assume almost total reflection
and vastly simplify your calculations. Doing the same calculations for
arbitrary polarization is much more complex (Breakall, et al., did some
work using GTD to model terrain effects, and compared it with actual field
measurements: "It is necessary to use lossy dielectric plates to obtain
accurate predictions for vertical polarization; whereas PEC (Perfect
Electrical Conductor) plates provide estiamtes for horizontal polarization
that are quite similar to the predictions made using the lossy dielectric
plates." "The enhancements in gain (e.g., 20dB) at low elevation angles
(e.g. 3-5 deg) can be very significant....")
VOACAP does statistical predictions on a much larger scale. There ARE
models around that deal with the minutia of this stuff, but probably not
Title: Modeling and measurement of HF antenna skywave radiation patterns in
Authors: Breakall, J. K.; Young, J. S.; Hagn, G. H.; Adler, R. W.; Faust,
D. L.; Werner, D. H.
Journal: IEEE Transactions on Antennas and Propagation (ISSN 0018-926X),
vol. 42, no. 7, p. 936-945
Publication Date: 07/1994
The paper presents an evaluation of the perturbations of elevation plane
patterns of HF vertical monopoles and horizontal dipoles when the antennas
are sited in irregular terrain. The Method of Moments was used in
conjunction with the Geometric theory of diffraction for predicting the
elevation plane radiation patterns. The 3D terrain was approximated by
seven connected flat plates that are very wide relative to the largest
wavelength of interest.
>Models are excellent for antennas some distance above earth,
>especially horizontally polarized antennas. They really are
>just very rough approximations for vertical antennas near
>earth (or even dipoles at very low height).
Kind of depends on the model. NEC isn't all that hot for real terrain,
since the flat earth approximation isn't realistic, although (particularly
for NEC4), the actual antenna modeling is quite good, even very, very close
(as in touching) the ground.
>I wouldn't predict one way or another what would happen in
>other locations where a marginal ground system is used, but
>I'd never rest 100% on any model when earth is
>involved...especially when the models we use don't model the
>earth in any accurate detail.
See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless Weather
Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any questions
and ask for Sherman, W2FLA.
TowerTalk mailing list