NEC-LIST:Second ground medium

I've just been made aware of a large difference between the far field
pattern when a second, highly conductive medium is under the antenna as
opposed to a large group of radials not connected to the antenna. For
tests, I've been modeling a vertical dipole (roughly a quarter
wavelength at the test frequency of 1.8 MHz), with bottom end 20 feet
(6.1 meters) above ground and the top 150 feet (45.7 meters) above
ground, fed in the center. The one medium model has a single ground
medium with conductivity and relative permittivity of 0.002 and 13
respectively. The two medium model has an inner medium with salt water
characteristics (5, 81) extending to 150 feet with a radial boundary,
and a second medium of 0.002, 13 beyond that. The two medium ground
model shows an increase in field strength up to about 3 dB over the
single medium model, with the increase beginning at elevation angles
above about 10 degrees. Using the NEC radial model with a large number
of radials in place of the highly conductive inner medium gives similar
results.

In contrast, the single medium model with a field of 120 radial wires of
150 foot radius placed either just above the ground under the antenna
and modeled with either NEC-2 or NEC-4, or just under the ground and
modeled with NEC-4, shows no significant field strength difference from
the model without the radial field. Likewise, a wire grid of 300 X 300
feet just above or below the ground alters the far field strength very
little.

As I recall, a reflection coefficient is calculated at the point that a
straight ray from each segment strikes the ground. When two media are
defined, the reflection coefficient is calculated for whichever medium
is struck by the ray. When the NEC radial model is used, the ground
conductivity is modified by the assumed conductivity of the radial
field, decreasing away from the center. This is, of course, a highly and
overly simplified view of what really happens with regard to the
interaction of the field with the ground close to the antenna. So my
thinking is that the model with a single medium and a wire radial or
wire grid field is likely the more accurate, and the model with a second
medium or NEC radial model isn't accurate due to being overly simplified
in the interaction of the antenna's field with the ground close to the
antenna. If this is so, then a caution is in order to not use two very
different media or the NEC radial model when the boundary is relatively
close to the antenna.

I'd appreciate any comments or other points of view on this.

Roy Lewallen

-- 
The NEC-List mailing list <nec-list_at_gweep.ca>
http://www.gweep.ca/mailman/listinfo.cgi/nec-list