On 2/5/16 8:02 AM, Grant Saviers wrote:
Roger,
From the link Dan AC6LA posted there are some long standing different
views of near and far fields from vertical antennas. A discussion above
my pay grade as to whether NEC 4.2 analysis is correct for these models,
but it is validated in my experience. I can offer an intuitive
explanation to part of your question.
So why does a vertical at the edge of the sea radiate more energy
seaward than landward? The relative conductivity is different by a
factor of 1000, 4 S/m for salt water vs 0.005 S/m for "average" earth.
So in that situation the return currents flow in the low resistance side
to a much higher value than the high resistance side. Further the
losses from a radiated field over salt water ground resistance
approaches that of copper. I think that accounts for the directivity
gain.
That's a very small effect. You can model it by doing a vertical in
free space with a variety of counter poise configurations. Start with a
90 degree bend dipole (e.g. 1 vertical, 1 radial) and then start adding
more radials.
Just not much change.. the direction of the main lobe changes a bit, but
the azimuthal variation is probably less than 1 dB. After all, an ideal
dipole has a gain of 2.15dB compared to an isotrope. An infinitesimally
small dipole has a gain of 1.75.
Perhaps the more important factor is that the pattern starts to
look like a vertical over "perfect" ground which shows the elevation
lobe at a maximum value at the horizon, which is great for long distance
DX propagation if you look at the HFTA statistics re arrival angles.
This is exactly what's going on and what's important. You shouldn't be
using NEC to model this kind of thing: you need a code that deals with
reflections from partial conductors. Jim Breakall did a model decades
ago for terrain that modeled the surface as a series of flat plates.
HFTA uses similar analysis, except it can't handle changing the soil
properties over the profile. Nor does HFTA do verticals, it's h-pol only.
You need a different modeling code for this problem. Something more like
used in the microwave fields, and you're going to need a very big grid,
and lots of computational horsepower.
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|