At 07:44 AM 8/14/2006, Pete Smith wrote:
>Rather than looking just at resistance, why not try to more fully
>characterize ground parameters? I fuzzily recall that a VE2 working at
>the Canadian government's antenna lab suggested years ago that you erect a
>very low dipole for the frequency of interest and then tweak ground
>parameters in a NEC-2 model until the resonant frequency and feedpoint
>impedance closely match the observed values. Is there something wrong
>with that approach?
That was the approach that George Hagn started with. Turns out that it has
sensitivities to the wrong things.. small changes in distance from the soil
have a much larger effect than changes in soil properties, and, you have to
have a very clear area around the dipole.
A better approaches is the OWL (Open Wire Line) technique (developed by
Hagn at SRI, and fairly extensively validated against field tests):
Basically, you drive two rods into the ground to form a two wire
transmission line, measure the Z of the line, and figure out what the
ground parameters are from that. You really need to do several
measurements, using different length rods. The technique is described in
the ARRL Antenna Handbook (at least in editions since the 19th).
Another technique, fairly good for measuring loss properties, if not so hot
for epsilon, is to put a resonant small loop on the surface and measure the
Q. Then raise it a known distance, and measure the Q again. The advantage
of the loop, over the dipole, is that the fields are fairly well contained,
so the surroundings doesn't have as big an effect. You're basically
measuring the Q of an inductor with this approach.
A similar approach is to lay an open wire line on the surface and measure
the propagation coefficient (gamma), either directly, or by measurement of
Z at several frequencies. Ideally, you'd have the conductors precisely
half immersed in the the medium, etc.
The latter two have a fair amount of analysis to back them up, and the
dominant error sources are going to be:
1) measurement uncertainty of Z or gamma
2) geometric problems: the loop isn't perfectly on the surface, or the
surface isn't flat, etc.
I would imagine that you'd want a loop that's big enough that the thickness
of the loop is "small", and that it averages out a big enough area of soil
(and deep!).. The penetration of the field is roughly proportional to the
The big name in the literature for these sorts of things is J.R. Wait.
There's a lot of information around on these sorts of things in connection
with geophysical prospecting and with underground (i.e. mine) communications.
>73, Pete N4ZR
TowerTalk mailing list