[TowerTalk] Fw: Why radials improve radiation!

[Jim Lux]

>
>> 2a. For vertical antennas, it is said that there is no reflection (radiation) from
>> the ground system.  It isn't very clear why, horizontal or vertical polarization or ?
>
> well, there can be radiation from the ground system.  using the term 'reflection' will
> lead to lots of heated discussion as that implies certain things that normally either
> aren't true or are hard to visualize.  the ground system normally can't be said to
> reflect anything, it can provide currents that may be visualized as images of the
> vertical element.  in a well balanced ground system the currents on opposite sides
> cancel out so there may be no far field contribution from the ground system, though
> there may be near field effects.
>


One has to be careful when using terms like "reflection" because it 
sometimes is just a nice conceptual model for a much more complex behavior.

For instance, if one is calculating the electric field between an object 
and a infinitely conductive plane, you can replace the plane by a copy 
of the object spaced appropriately, if that makes the calculation more 
tractable.

Particularly in the days when we used slide rules and pencil and paper, 
such approximations and simplifications made things feasible to analyze, 
but that doesn't mean that the conceptual model necessarily represents a 
real behavior.

These approximations persist even in modern programs like NEC (e.g. 
reflection coefficient approximation vs Somerfield-Norton ground), 
because until very recently, they were computationally faster, and work 
adequately for a lot of modeling purposes.

In the near field, magnetic fields are as important as electric fields 
(in fact, one of the three classical definitions for where the near 
field ends is where the ratio of E to H field becomes 377 ohms: free 
space).  The "stuff" under an antenna (of any polarization) is part of 
the system.

The buried ground radials have lots of different functions: They serve 
as a low impedance connection from base of antenna to soil currents; 
they reduce the apparent soil resistivity. and so on.  An odd 
observation from some modeling with NEC 4 is that if you model a buried 
radial field with lots of long radials that you can "break" the radials 
a short distance from the antenna, and the performance doesn't change 
all that much. That is, the "conductivity improving" aspect of the 
radials seems to be as important as the "low impedance connection to the 
soil"


In the far field (a long way from the antenna), a reflection coefficient 
model works fairly well.   The problem is that you really need the 
terrain shape to do it.  HFTA does a nice job assuming horizontal 
polarization and that the terrain is not tilted cross ways to the 
direction of propagation, but that's because waves polarized parallel to 
the surface have almost perfectly reflectivity, regardless of the soil 
properties, so it make the calculations simple: all you have to worry 
about is diffraction.

As soon as the polarization is tilted at all, it gets a LOT more 
complex.  And, in fact, so much more complex that it's probably not 
worth modeling: not only would you need a very detailed terrain model 
(on a scale better than 0.1 wavelength) but you'd need equivalently 
detailed information on the dielectric properties.  Unless you happen to 
live in the middle of a flat dry lake or salt flat, or maybe among a 
bunch of sand-dunes, that just isn't going to happen.

So, we make some approximations, make a qualitative assessment that 
something is better or worse in the general case (e.g. is higher 
better?) or identify a specific case where it's really good or bad 
(verticals on the beach = very good, regardless)