Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Laying in ground radials?

48 views
Skip to first unread message

Dan Richardson

unread,
Aug 10, 1998, 3:00:00 AM8/10/98
to
On Sun, 09 Aug 1998 14:28:25 GMT, tempo...@dim.com (Bill, KD0HG)
wrote:

>
>Anyone care to post their own special technique for shallow burial of ground
>radials? There must be a better method that I've missed for this baked-hard
>eastern Colorado foothills terrain other than waiting for rain with a shovel
>in hand.
>
>73,
>
>-BK

Hi Bill,

Well here is my story. My soil is very hard clay like and in the
non-rainy season it is almost like concrete. What I did, and it worked
well, was get out my gasoline powered lawn edger. Next I installed two
blades on it so the width of cut would be wider. Then I proceeded to
cut groves in the ground using the edger at full depth about 2-3
inches. The only problem is that you'll get dirty as the edger really
throws the dirt. Be sure to ware safety glasses and good luck.

73
Danny, K6MHE


George T. Baker

unread,
Aug 11, 1998, 3:00:00 AM8/11/98
to
Main thing that I would stress is that there is NO good electrical
reason for placing radials below the ground surface. Shallow (an inch or
so) burial is useful only to get the wires out of the way and prevent
them from being disturbed or causing problems, such as tripping people.
Deeper burial is an exercise in futility since that merely places the
higher conductivity wire further removed and surrounded by lower
conductivity soil.

So, if you don't have surface conflicts, forget about burying them. They
will work better lying on the surface and are out there for inspection
and repair if needed. Just lay them out and secure to the ground with
long nails or wires. Sure saves a lot of hard labor!

72/73, George didit dit
Amateur Radio W5YR, 52 years and counting!
QRP-L #1373 QRP ARCI #9583 FISTS #4930 ARS #403 ICQ #16819496
AutoPOWER Systems, Fairview, TX (30 Mi. NE of Dallas)

Gary Coffman

unread,
Aug 11, 1998, 3:00:00 AM8/11/98
to
On Tue, 11 Aug 1998 00:31:09 -0500, "George T. Baker" <w5...@swbell.net> wrote:
>Main thing that I would stress is that there is NO good electrical
>reason for placing radials below the ground surface. Shallow (an inch or
>so) burial is useful only to get the wires out of the way and prevent
>them from being disturbed or causing problems, such as tripping people.
>Deeper burial is an exercise in futility since that merely places the
>higher conductivity wire further removed and surrounded by lower
>conductivity soil.

Actually, there are a couple of good electrical reasons for burying
radials. Getting them down below the frost line will ensure that
they can serve as a lightning ground. And shallowly burying them
will negate any concern about making the radials exactly equal
length (though just laying them on the ground goes far toward
achieving this frequency insensitivity as well).

Deep burial is, however, counterproductive because RF can't
propagate below the skin depth of the Earth.

Gary
Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
534 Shannon Way | We break it |
Lawrenceville, GA | Guaranteed |

x...@x.com

unread,
Aug 11, 1998, 3:00:00 AM8/11/98
to
Bill,
Ref: CQ Mag., July '98, pg 20.
"How To Build an Effective All-Band Counterpoise"
By Richard Brunner, AA1P

Reg Edwards

unread,
Aug 12, 1998, 3:00:00 AM8/12/98
to
The depth of buried ground electrodes is frequently mentioned on these
walls.

I have been trying to discover for ages what happens to moist soil
resistivity and permittivity when it feezes.

Does freezing affect the effectiveness of lightning conductors ?

Does it affect the impedance looking into RF ground radial systems. If so
- to what numerical extent ?

Or, to ask the question in practical terms, does anyone know the dielectric
constant (permittivity) of ice.

Please and thank you.

Reg.

Roy Lewallen

unread,
Aug 12, 1998, 3:00:00 AM8/12/98
to
A while back I got a paper written by a professor in Alaska regarding
the RF properties of ice, but can't put my hand on it just now. As I
recall, the conductivity of ice is very low. Don't recall the
dielectric constant, but think it was fairly high. So it's a pretty
good dielectric, and very unlike liquid water. When I get some time
I'll see if I can find the paper. There was some discussion on this
newsgroup on the topic in October 1996.

Its effect on radial ground systems would of course depend strongly on
how deeply the ground is frozen.

Roy Lewallen, W7EL

Richard Clark

unread,
Aug 12, 1998, 3:00:00 AM8/12/98
to
On Wed, 12 Aug 1998 16:23:33 GMT, Roy Lewallen
<w7...@teleport.com> wrote:

>A while back I got a paper written by a professor in Alaska regarding
>the RF properties of ice, but can't put my hand on it just now. As I
>recall, the conductivity of ice is very low. Don't recall the
>dielectric constant, but think it was fairly high. So it's a pretty
>good dielectric, and very unlike liquid water. When I get some time
>I'll see if I can find the paper. There was some discussion on this
>newsgroup on the topic in October 1996.
>
>Its effect on radial ground systems would of course depend strongly on
>how deeply the ground is frozen.
>
>Roy Lewallen, W7EL
>

Hi Roy,

To your statements about the properties of water.

From "Reference Data For Radio Engineers," Table 19,
Properties of Materials:"

Water (distilled)
Dielectric Constant = 78.2
Dissipation Factor = 0.04 @ 1 Mhz = 0.005 @ 100 MHz
Volume Resistivity = 1,000,000 Ohms
Fresh Fallen Snow
Dielectric Constant = 1.2
Dissipation Factor = 0.0215 @ 1 MHz
Snow (hard packed, followed by light rain)
Dielectric Constant = 1.55
Dissipation Factor = 0.29 @ 1 MHz
Ice (from distilled water)
Dielectric Constant = 4.15
Dissipation Factor = 0.12 @ 1 MHz = 0.035 @ 100 MHz
Soil , sandy dry
Dielectric Constant = 2.53
Dissipation Factor = 0.018 @ 1 MHz

The quality of a dielectric is found in its Dissipation
Factor which directly relates to Q. For water, its
characteristics vary widely as a function of the content of
impurities. Because water is such an efficient de-ionizer,
when mixed with salts its resistivity plunges, but it never
approaches the conductivity of even the poorest metal. The
loss in ground systems is not from conduction but from the
dielectric loss in the heightened Dissipation Factor. (It
may be argued that the dielectric loss is a conduction loss
of the equivalent series resistance however; but this
argument would be a contortion.)

From "Electronic And Radio Engineering," Terman, 1955,
footnote 1 page 24:
"...The reciprocal of the dissipation factor is termed the
capacitor Q and is the ratio of the capacitor reactance to
the equivalent series resistance..."

So we find from above

Water (distilled)
Dissipation Factor = 0.04 @ 1 MHz = 0.005 @ 100 MHz
Q = 25 @ 1 MHz = 200 @ 100 MHz
Fresh Fallen Snow
Dissipation Factor = 0.0215 @ 1 MHz
Q = 46.5 @ 1 MHz
Snow (hard packed, followed by light rain)
Dissipation Factor = 0.29 @ 1 MHz
Q = 3.4 @ 1 MHz
Ice (from distilled water)
Dissipation Factor = 0.12 @ 1 MHz = 0.035 @ 100 MHz
Q = 8.3 @ 1 MHz = 28.6 @ 100 MHz
Soil , sandy dry
Dissipation Factor = 0.018 @ 1 MHz
Q = 55.6 @ 1 MHz

This again points out the age old saw of Ground (in nature)
being worse than Water (in nature) for losses as something
of a myth. In fact the opposite is true and it is the high
SWR presented by the air/surface interface that presents the
boon of reflected power in wet surface conditions, not the
imagined lowered loss of boosted conductivity. Of course
the deliberate pollution of the soil with salts can boost
conductivity, but this still pales in comparison to a simple
radial system for actual conduction.

73's
Richard Clark, KB7QHC


at...@imap1.asu.edu

unread,
Aug 13, 1998, 3:00:00 AM8/13/98
to
Roy Lewallen (w7...@teleport.com) wrote:
: A while back I got a paper written by a professor in Alaska regarding
: the RF properties of ice, but can't put my hand on it just now. As I
: recall, the conductivity of ice is very low. Don't recall the
: dielectric constant, but think it was fairly high. So it's a pretty
: good dielectric, and very unlike liquid water. When I get some time
: I'll see if I can find the paper. There was some discussion on this
: newsgroup on the topic in October 1996.

: Its effect on radial ground systems would of course depend strongly on
: how deeply the ground is frozen.

: Roy Lewallen, W7EL


I see that Richard answered most of the numerical questions.
I recently have been calculating the properties of small water
clusters, and I thought I would add some "cartoon" physics.

The high dielectric constant and loss of liquid water is because the
water molecule has a permanent dipole moment. That is, when the
hydrogen atoms bind to Oxygen, the Oxygen tends to pull the electrons
off the protons of the hydrogen atoms. This produces a little electric
dipole which has an electric field that looks like the magnetic field
of a bar magnetic. In water, the molecules are randomly oriented, but
when you apply an external electric field, the molecules rotate around
and the fields from the molecules tend to cancel the applied field
leading to a large dielectric constant. This is true up to the
frequency where water can't easily rotate in response to the field
around 10 GHz. There is also an electronic part of the polarizability
that is smaller caused by the field displacing the electrons relative
to the nuclei.

When you freeze water the molecules form a lattice where the net field
again is small. The molecules are locked together and can't rotate to
align with the applied field and the dielectric constant drops down to
just the electronic part. As you might expect, collisions while the
water molecules rotate leads to loss; there is much less loss from the
electronic polarization.

Since water conducts primarily by ionic transport, freezing it
reduces the diffusion of the ions and the conductivity plummets.

An amusing fact is that the absorption coefficient of water is quite
low in the optical range. It increases by about 8 orders of magnitude
on either side of the optical spectrum. Presumably that's why our eyes
see the frequencies they do.

73 Kevin w9...@ptolemy.la.asu.edu

Reg Edwards

unread,
Aug 13, 1998, 3:00:00 AM8/13/98
to
Richard, much obliged to you for supplying the info in such detail.

By far the greatest change of consequence is the permittivity. It must
happen rather suddenly around 0 degrees C. Sea water has the same
permittivity as fresh water. Presumably sea ice has the same permittivity
as pure ice.

Incidentally, soil and water impedance can better be represented for
propagation effects along buried radials as a conductance in parallel with
a capacitative susceptance.

Seems there's plenty of scope for experimenting with jugs of mud, the
domestic refridgerator, alligator clips and an Autec, so called antenna
analyser.

Reg G4FGQ
http://www.btinternet.com/~g4fgq.regp

Richard Clark <rwc...@rwclark.seanet.com> wrote in article
<35e9d14f....@news.seanet.com>...


> On Wed, 12 Aug 1998 16:23:33 GMT, Roy Lewallen

> <w7...@teleport.com> wrote:
>
> >A while back I got a paper written by a professor in Alaska regarding
> >the RF properties of ice, but can't put my hand on it just now. As I
> >recall, the conductivity of ice is very low. Don't recall the
> >dielectric constant, but think it was fairly high. So it's a pretty
> >good dielectric, and very unlike liquid water. When I get some time
> >I'll see if I can find the paper. There was some discussion on this
> >newsgroup on the topic in October 1996.
> >
> >Its effect on radial ground systems would of course depend strongly on
> >how deeply the ground is frozen.
> >
> >Roy Lewallen, W7EL
> >
>

Richard Clark

unread,
Aug 13, 1998, 3:00:00 AM8/13/98
to
On 13 Aug 1998 04:36:10 GMT, at...@imap1.asu.edu wrote:

>Roy Lewallen (w7...@teleport.com) wrote:
>: A while back I got a paper written by a professor in Alaska regarding
>: the RF properties of ice, but can't put my hand on it just now. As I
>: recall, the conductivity of ice is very low. Don't recall the
>: dielectric constant, but think it was fairly high. So it's a pretty
>: good dielectric, and very unlike liquid water. When I get some time
>: I'll see if I can find the paper. There was some discussion on this
>: newsgroup on the topic in October 1996.
>
>: Its effect on radial ground systems would of course depend strongly on
>: how deeply the ground is frozen.
>
>: Roy Lewallen, W7EL
>
>

Hi Kevin,

To amplify on your observations of water, perhaps a quote
from "Physical Electronics," by Mssr.s Hemmenway, Henry, &
Caulton:
"Electrolytic conductivities increase nonlinearly with
temperature by amounts of the order of 2% per degree
centigrade in aqueous solutions. Water solutions of
electrolytes show an interesting shrinkage in volume
("electrostriction") as the ionization of an electrolyte is
increased. Using the inverse of this effect, it is possible
to make devices whose conductivity is pressure sensitive."

This same source also discusses the relationships of virtual
crystal structure imposed at the freezing point which turns
the analysis of a liquid into an analysis of a solid. This
appears to conform to your discussion above about lattice
formation.

I will also offer other sources (unquoted) for your research
into these matters: Debye, Huckel (umlaut over the u),
Arrhenius and of course Van der Waals.

Rich Griffiths

unread,
Aug 20, 1998, 3:00:00 AM8/20/98
to
Gary Coffman wrote:
>
<snip comments on burying radials>

>
> Deep burial is, however, counterproductive because RF can't
> propagate below the skin depth of the Earth.
>
> Gary
> Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
> 534 Shannon Way | We break it |
> Lawrenceville, GA | Guaranteed |

That strikes me as an interesting, unusual comment.

Can a mass with heterogeneous, spacially distributed properties like
earth (ground/soil/whatever) display a skin effect? I've always
been suspicious of antenna modeling software that doesn't account
for this heterogeneity.

Ground-penetrating radars operate in the HF range. Some penetrate
strongly up to 1/2 wavelength.

Is it meaningful to talk about a skin depth for surface soil?


73


Rich W2RG

Reg Edwards

unread,
Aug 20, 1998, 3:00:00 AM8/20/98
to
If any material is capable of conducting an electric current then it has an
intrinsic skin depth which decreases with the squareroot of frequency, its
permeability and its conductivity.

The current is attenuated exponentially with depth. There is no sharply
defined "depth". The nominal skin depth is that at which the current has
decayed to 1/e (37%) of that at the surface.

Skin depth = 1/squareroot(Pi * F * Mu * C),

where F = Frequency, Mu = permeability, C = conductivity.

Its roughly a few feet at 10 MHz in your back garden soil.

Soil also has a propagation velocity. Of the order of VF = 0.15 at HF.
Makes a terrible mess of the carefully calculated lengths of buried
radials. But it is more like a diffusion process than propagation of waves
because of the high attenuation. Fascinating eh ?

Reg G4FGQ
http://www.btinternet.com/~g4fgq.regp

Rich Griffiths <ri...@one.net> wrote in article
<35DC8944...@one.net>...

K4SQR

unread,
Aug 20, 1998, 3:00:00 AM8/20/98
to
Why would any amateur station bury radials any deeper than a couple of inches?
Any deeper does not enhnace performance
on 1.8------3.5-------or 7.0

73,
Jim, K4SQR
A 4-Square Station>> Deep burial is, however, counterproductive because RF


can't
>> propagate below the skin depth of the Earth.
>>

> Gary Coffman KE4ZV

Deacon Dave Shrader

unread,
Aug 20, 1998, 3:00:00 AM8/20/98
to
Rich Griffiths wrote:
>
> Gary Coffman wrote:
> >
> <snip comments on burying radials>
> >
> > Deep burial is, however, counterproductive because RF can't
> > propagate below the skin depth of the Earth.
> >
> > Gary
> > Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
> > 534 Shannon Way | We break it |
> > Lawrenceville, GA | Guaranteed |
>
> That strikes me as an interesting, unusual comment.
>
> Can a mass with heterogeneous, spacially distributed properties like
> earth (ground/soil/whatever) display a skin effect? I've always
> been suspicious of antenna modeling software that doesn't account
> for this heterogeneity.
>
> Ground-penetrating radars operate in the HF range. Some penetrate
> strongly up to 1/2 wavelength.
>
> Is it meaningful to talk about a skin depth for surface soil?
>
> 73
>
> Rich W2RG

Well, it is different from dry sandy soil and wet loom. It is different
for humus and stoney or gravelly soil. So skin depth is real. How it
effects antenna patterns is more art than science. The skin effect in my
yard varies from winter to spring to summer etc. But for an 80 meter
dipole at 30 to 40 feet I'm still dominated by high angle radiation
caused by the effective image which is caused by skinn effect.

On a vertical, radials should be close to the surface or the ground
losses will be dominated by skin effect (losses in the image plane)

Dave W1MCE,
--
In the Love and Mercy of Jesus,

Deacon Dave

"Proclaim the Gospel Always!!! Use words if necessary"...St Francis of
Assisi

<http://www.StRichardDanvers.org>
<http://www.networknews.org>

Rich Griffiths

unread,
Aug 21, 1998, 3:00:00 AM8/21/98
to
Reg Edwards wrote:
>
> If any material is capable of conducting an electric current then it has an
> intrinsic skin depth which decreases with the squareroot of frequency, its
> permeability and its conductivity.
>
> The current is attenuated exponentially with depth. There is no sharply
> defined "depth". The nominal skin depth is that at which the current has
> decayed to 1/e (37%) of that at the surface.
>
> Skin depth = 1/squareroot(Pi * F * Mu * C),
>
> where F = Frequency, Mu = permeability, C = conductivity.
>
> Its roughly a few feet at 10 MHz in your back garden soil.
>
> Soil also has a propagation velocity. Of the order of VF = 0.15 at HF.
> Makes a terrible mess of the carefully calculated lengths of buried
> radials. But it is more like a diffusion process than propagation of waves
> because of the high attenuation. Fascinating eh ?
>
> Reg G4FGQ

I'm familiar with the formulas (although a figure for the velocity
factor of soil is a new one for me...interesting value).

But over the years I've become suspicious of mathematical
representations we have developed for reality. Sometimes we
over-generalize from them. Sometimes we forget that they're
representations and not the reality (physical/chemical/whatever
phenomenon) itself.

Skin effect is the result of interaction between the current and the
fields that the current itself creates. This effect is nicely
approximated by the above formula when the medium is homogeneous and
compact (e.g., a wire). What is it like when

M = f( x, y, z ) and C = g( x, y, z ) ????

I find this especially problematic when we're talking about the
near-field propagation from an antenna. On the higher HF bands, I
expect the "reflection" from the ground looks more like a refraction. I
further expect that the far-field pattern of an antenna will be
considerably different if the ground behaves like a refraction zone than
a surface.

Perhaps some of the antenna modeling software takes this into account.
I just don't know.


73


Rich W2RG

Roy Lewallen

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
Rich Griffiths wrote:
> . . .

> I find this especially problematic when we're talking about the
> near-field propagation from an antenna. On the higher HF bands, I
> expect the "reflection" from the ground looks more like a refraction. I
> further expect that the far-field pattern of an antenna will be
> considerably different if the ground behaves like a refraction zone than
> a surface.
>
> Perhaps some of the antenna modeling software takes this into account.
> I just don't know.

Modern antenna modeling software deals with the near field ground
interaction in a much more rigorous way than simple reflection.
However, NEC, MININEC, and derivatives all treat far-field interaction
as simple reflection. This is probably adequate for the flat-ground
model they incorporate. (But of course flat ground is unrealistic in
many situations.) Brian Beezley's TA (Terrain Analysis) software is
the only readily-available program I know of which does calculate more
complex refractive effects from terrain of arbitrary slope and shape.
It takes as its input a free-space antenna pattern from EZNEC or one
of his antenna modeling programs and gives the resulting pattern.

However, all the readily-available modeling programs, including TA,
assume that the ground is homogeneous to an infinite depth which, as
you point out, doesn't represent reality. (However, if a program were
devised which does take this into account, almost no one would be able
to measure the required values for the program to use. Even
measurement of simple RF conductivity at the surface is a tricky and
not very exact business.) As it turns out, the quality of the ground
generally has very little effect on horizontally polarized signals
except at high radiation angles, so the assumption of its being
homogeneous probably makes little difference for horizontally
polarized antennas unless being used for NVIS operation. Ground
quality is, however, very important for vertically polarized waves, so
the assumption may result in an unknown level of error.

Roy Lewallen, W7EL

Reg Edwards

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
Rich, I think you are taking an unduly pessimistic view of mathematical
modelling.

Don't let the wood get in the way of the trees.

Models allow what we can be reasonably sure of to made use of, and what we
have no idea of to be put on on one side for the time being. Manufacturing
or using a model sharpens our wits and on completion of the first crude
approximation we shall know much more about the real world than before the
model was constructed. Caricatures are very useful.

There's no need to overcomplicate by distinguishing near from far fields.
That's merely a human intervention. There is only one field and that field
changes with distance. KISS.

If you think propagation velocity on buried radials an interesting subject,
then make use of my program RADIOETH.exe, downloadable from my website in a
few seconds, not zipped up, usable immediately. What is of importance on
buried radials, as on any other sort of radials, is the resistive component
of input impedance. The program will tell you much about a single buried
wire at any frequency from DC up to HF, under a variety of soil conditions.

Program ENDFEED, as a sideline, will tell you the input impedance of a
bunch of any number of radials and the radiating efficiency of an antenna
which makes use of such a radial system.

If you are suspicious of "mathematical representations" as you call them
then don't use them. You won't learn anything from the refusal. But ask
yourself are there any other sort of logical representations? If you use
one you may learn a lot.
--
Click below my signature -

Reg G4FGQ
http://www.btinternet.com/~g4fgq.regp


Gary Coffman

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
On 20 Aug 1998 21:23:29 GMT, "Reg Edwards" <g4fgq...@btinternet.com> wrote:

>If any material is capable of conducting an electric current then it has an
>intrinsic skin depth which decreases with the squareroot of frequency, its
>permeability and its conductivity.
>
>The current is attenuated exponentially with depth. There is no sharply
>defined "depth". The nominal skin depth is that at which the current has
>decayed to 1/e (37%) of that at the surface.
>
>Skin depth = 1/squareroot(Pi * F * Mu * C),
>
>where F = Frequency, Mu = permeability, C = conductivity.
>
>Its roughly a few feet at 10 MHz in your back garden soil.

I think it is important to point out here that the attenuation due to
skin depth isn't attenuation due to dissipation (though there is
dissipation too in any lossy medium). The primary mechanism
here is due to the ratio of propagation velocity between waves
at the surface and at depth. So the wave is reflected or refracted
back out of the medium more and more strongly with increasing
depth, and less and less current can be found as you go deeper.
This is boundary condition behavior.

Gary Coffman

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
On Sat, 22 Aug 1998 01:41:51 GMT, Roy Lewallen <w7...@teleport.com> wrote:
>However, all the readily-available modeling programs, including TA,
>assume that the ground is homogeneous to an infinite depth which, as
>you point out, doesn't represent reality. (However, if a program were
>devised which does take this into account, almost no one would be able
>to measure the required values for the program to use. Even
>measurement of simple RF conductivity at the surface is a tricky and
>not very exact business.) As it turns out, the quality of the ground
>generally has very little effect on horizontally polarized signals
>except at high radiation angles, so the assumption of its being
>homogeneous probably makes little difference for horizontally
>polarized antennas unless being used for NVIS operation. Ground
>quality is, however, very important for vertically polarized waves, so
>the assumption may result in an unknown level of error.

An error bound can be assigned to the contribution at extreme depths.
Since the currents fall off expnentially with depth, it is a simple matter
to estimate the contribution of currents below a few skin depths. And
that contribution is virtually nil. So assuming that the soil is uniform to
infinite depth doesn't measurably upset the calculation.

What does upset the calculation is the assumption that the soil is
homogeneous at and very near the surface. This is often not the
case, and can result in a "lumpiness" to the near field which the
model calculations won't show. This is one of the reasons for
installing a dense and shallowly buried radial field. It will tend to
homogenize near surface soil conditions. (It also reduces the
dissipation due to the lossy soil medium, of course, but that's a
separate issue more germane to antenna system efficiency than
to azimuth pattern shape.)

Gary Coffman

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
On Thu, 20 Aug 1998 16:38:28 -0400, Rich Griffiths <ri...@one.net> wrote:

>Gary Coffman wrote:
>>
><snip comments on burying radials>
>>
>> Deep burial is, however, counterproductive because RF can't
>> propagate below the skin depth of the Earth.
>>
>> Gary
>

>That strikes me as an interesting, unusual comment.
>
>Can a mass with heterogeneous, spacially distributed properties like
>earth (ground/soil/whatever) display a skin effect? I've always
>been suspicious of antenna modeling software that doesn't account
>for this heterogeneity.

Yes, absolutely there is a skin effect for any conductive medium.
The main effect of hetrogeneity of surface soil conditions is a bit
of lumpiness in the azimuth pattern.

>Ground-penetrating radars operate in the HF range. Some penetrate
>strongly up to 1/2 wavelength.

Define "strongly". Field strength drops off exponentially with depth.
The return from these systems is *very* much weaker than a direct
ray. So the effect on the azimuth pattern of an antenna operating at
the surface is very slight for these deep reflections.

>Is it meaningful to talk about a skin depth for surface soil?

That's the only sort of soil for which it is meaningful to talk about
skin depth. Deep soil is out of the picture since it is so far below
the boundary that the currents in it are insignificant. Skin depth
is a *boundary* condition. Only the properties of the mediums near
the boundary matter.

Rich Griffiths

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
Reg Edwards wrote:
>
> Rich, I think you are taking an unduly pessimistic view of mathematical
> modelling.
>
> Don't let the wood get in the way of the trees.
>
> Models allow what we can be reasonably sure of to made use of, and what we
> have no idea of to be put on on one side for the time being. Manufacturing
> or using a model sharpens our wits and on completion of the first crude
> approximation we shall know much more about the real world than before the
> model was constructed. Caricatures are very useful.

I pretty much agree with all this and have made similar arguments to my
own colleagues.


>
> There's no need to overcomplicate by distinguishing near from far fields.
> That's merely a human intervention. There is only one field and that field
> changes with distance. KISS.

Perhaps, but I think making that distinction is inherent in the models.
Roy Lewallen responded with some good stuff on that, which I need to go
back to and re-read and ask questions.

> If you think propagation velocity on buried radials an interesting subject,
> then make use of my program RADIOETH.exe, downloadable from my website in a
> few seconds, not zipped up, usable immediately. What is of importance on
> buried radials, as on any other sort of radials, is the resistive component
> of input impedance. The program will tell you much about a single buried
> wire at any frequency from DC up to HF, under a variety of soil conditions.
>
> Program ENDFEED, as a sideline, will tell you the input impedance of a
> bunch of any number of radials and the radiating efficiency of an antenna
> which makes use of such a radial system.

I just downloaded both and will play with them later. Thanks. (BTW,
56K modems are great!)

> If you are suspicious of "mathematical representations" as you call them
> then don't use them. You won't learn anything from the refusal. But ask
> yourself are there any other sort of logical representations? If you use
> one you may learn a lot.
> --
> Click below my signature -
>
> Reg G4FGQ
> http://www.btinternet.com/~g4fgq.regp

I am indeed suspicious of models, but I can't avoid using them. My
suspicion comes from years of reading about them, doing my own modeling
of turbulent fluid flow, and now doing my own modeling of the
interactions of molecules and solids (adsorption/desorption of
pollutants on soils and sediments is the official justification).

I think modeling makes good science, for the reasons you give above.
And part of modeling needs to be continually questioning whether your
representation is any good.

A problem I have now is that we've bundled the issues of the behavior of
the ground + radials as a conductor with the behavior of the ground as a
reflector or refractor. They're connected, but the two are a lot to
deal with in a single posting.


73 Rich W2RG

Rich Griffiths

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
Roy Lewallen wrote:
>
> Rich Griffiths wrote:
> > . . .
> > I find this especially problematic when we're talking about the
> > near-field propagation from an antenna. On the higher HF bands, I
> > expect the "reflection" from the ground looks more like a refraction. I
> > further expect that the far-field pattern of an antenna will be
> > considerably different if the ground behaves like a refraction zone than
> > a surface.
> >
> > Perhaps some of the antenna modeling software takes this into account.
> > I just don't know.
>
> Modern antenna modeling software deals with the near field ground
> interaction in a much more rigorous way than simple reflection.
> However, NEC, MININEC, and derivatives all treat far-field interaction
> as simple reflection. This is probably adequate for the flat-ground
> model they incorporate. (But of course flat ground is unrealistic in
> many situations.) Brian Beezley's TA (Terrain Analysis) software is
> the only readily-available program I know of which does calculate more
> complex refractive effects from terrain of arbitrary slope and shape.
> It takes as its input a free-space antenna pattern from EZNEC or one
> of his antenna modeling programs and gives the resulting pattern.
>
> However, all the readily-available modeling programs, including TA,
> assume that the ground is homogeneous to an infinite depth which, as
> you point out, doesn't represent reality. (However, if a program were
> devised which does take this into account, almost no one would be able
> to measure the required values for the program to use. Even
> measurement of simple RF conductivity at the surface is a tricky and
> not very exact business.)

That addresses what would have been my next question: how do we get the
information/data we need to feed a comprehensive model? Conditions at a
site will vary with position and time. We overcome that somewhat with a
good radial system. But a system of 1/8-wave or even 1/4-wave radials
seems insufficient.

> As it turns out, the quality of the ground
> generally has very little effect on horizontally polarized signals
> except at high radiation angles, so the assumption of its being
> homogeneous probably makes little difference for horizontally
> polarized antennas unless being used for NVIS operation. Ground
> quality is, however, very important for vertically polarized waves, so
> the assumption may result in an unknown level of error.
>

> Roy Lewallen, W7EL

Do you have anything on how great the effect is? Or how the results of
modeling a reflective ground versus a refractive ground compare? So
often, Hams haggle over the alleged "take-off" angles of alternative
antennas. Their discussions are usually based on model results. Can
the available models correctly predict whether a given installation will
achieve 15 degrees or 20 degrees or ...?

Personally, I don't need that predictive capability. But some Hams
place great stock in their model results.


Rich W2RG

Rich Griffiths

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
Gary Coffman wrote:
>
> On Thu, 20 Aug 1998 16:38:28 -0400, Rich Griffiths <ri...@one.net> wrote:
>
> >Gary Coffman wrote:
> >>
> ><snip comments on burying radials>
> >>
> >> Deep burial is, however, counterproductive because RF can't
> >> propagate below the skin depth of the Earth.
> >>
> >> Gary
> >
> >That strikes me as an interesting, unusual comment.
> >
> >Can a mass with heterogeneous, spacially distributed properties like
> >earth (ground/soil/whatever) display a skin effect? I've always
> >been suspicious of antenna modeling software that doesn't account
> >for this heterogeneity.
>
> Yes, absolutely there is a skin effect for any conductive medium.
> The main effect of hetrogeneity of surface soil conditions is a bit
> of lumpiness in the azimuth pattern.
>
> >Ground-penetrating radars operate in the HF range. Some penetrate
> >strongly up to 1/2 wavelength.
>
> Define "strongly". Field strength drops off exponentially with depth.
> The return from these systems is *very* much weaker than a direct
> ray. So the effect on the azimuth pattern of an antenna operating at
> the surface is very slight for these deep reflections.

Good point. Strong enough to receive clearly back at the surface and to
construct an image of the subsurface. But probably not strong enough to
contribute significantly to fields any distance from the surface.

> >Is it meaningful to talk about a skin depth for surface soil?
>
> That's the only sort of soil for which it is meaningful to talk about
> skin depth. Deep soil is out of the picture since it is so far below
> the boundary that the currents in it are insignificant. Skin depth
> is a *boundary* condition. Only the properties of the mediums near
> the boundary matter.
>
> Gary
> Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
> 534 Shannon Way | We break it |
> Lawrenceville, GA | Guaranteed |

I'm still left with the question I just posed to Roy Lewallen in another
posting. Is the ground's behavior as a refractor vs. a reflector
sufficiently different that model predictions of the "take-off" angle
and other behavior that many Hams hold near and dear need to account
for this? Roy indicates that even advanced models like NEC do not
account for the difference and that a separate correction, such as by
TA, is needed.


Rich W2RG

Gary Coffman

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
On Sat, 22 Aug 1998 09:05:10 -0400, Rich Griffiths <ri...@one.net> wrote:
>I'm still left with the question I just posed to Roy Lewallen in another
>posting. Is the ground's behavior as a refractor vs. a reflector
>sufficiently different that model predictions of the "take-off" angle
>and other behavior that many Hams hold near and dear need to account
>for this? Roy indicates that even advanced models like NEC do not
>account for the difference and that a separate correction, such as by
>TA, is needed.

Questions of whether we are dealing with reflection or refraction
can best be answered by looking at the ratio of refractive indexes
at the boundary. For even poor soil this ratio is very high. We can
ignore refraction for take-off angles of interest to skywave propagation.

For horizontally polarized waves in the far field, there are essentially
no rays at very shallow angles of incidence because ground conductivity
shorts out the E field of the horizontally polarized wave traveling along
the ground. So we only have to consider the higher angle of incidence
waves and can treat them as purely reflected waves.

For vertically polarized waves, very shallow incidence rays couple into the
ground/air boundary and are trapped there as the guided groundwave. They
don't play any part in the far field skywave pattern. At slightly higher angles,
reflection of vertically polarized rays is complicated by the perspective
produced by the geometry of the situation. This spreads the mirror
currents over a longer distance, so these reflections are more strongly
influenced by soil conductivity than with horizontally polarized waves.
(I don't know if this description makes sense, you need to draw it out
to see the effect of perspective.)

In summary, the far field skywave pattern is solely a result of direct and
reflected rays. For horizontally polarized signals, this is pretty simple to
model. For vertically polarized signals, it is somewhat more difficult due
to geometric perspective. Terrain roughness also complicates the picture
due to the presence of terrain blockage and knife edge diffraction effects
as well as varying incidence angles in the far field. For relatively flat
ground, and the take-off angles of interest to us, we can usually ignore
terrain roughness, and the NEC based models do. But for very rough
terrain, or close in obstacles, use of a program like TA takes care of that.

Richard Clark

unread,
Aug 22, 1998, 3:00:00 AM8/22/98
to
On Sat, 22 Aug 1998 09:05:10 -0400, Rich Griffiths
<ri...@one.net> wrote:

>
>I'm still left with the question I just posed to Roy Lewallen in another
>posting. Is the ground's behavior as a refractor vs. a reflector
>sufficiently different that model predictions of the "take-off" angle
>and other behavior that many Hams hold near and dear need to account
>for this? Roy indicates that even advanced models like NEC do not
>account for the difference and that a separate correction, such as by
>TA, is needed.
>
>

>Rich W2RG

Hi Rich,

I was very amused by your response to one poster about the
faithfulness of computer modeling (especially since that
poster decries that same concern against others - vanitas).
Oh well.

As for the homogeneity of ground. Consider that at HF
wavelengths that heterogenous features need to be physically
significant against the dimensions of wavelength. Soil
maps available from city or county planners would resolve
such large body subsurface issues. Further, there are
distinctions of the soil/air interface that present
different concerns in the near field and far field. The
near field (within the first wavelength at least) concerns
are mostly matching issues as they do not contribute to
issues of launch angle (unless you are interested in launch
angles in excess of 25 degrees). Out beyond 10 wavelengths
(the far field by then), it is the reflective
characteristics of ground that contribute to low launch
angles; unless we are speaking of ground wave propagation,
and those extremely low angles suffer mightily.

There is also the debate between horizontal and vertical
that differentiates some of these issues. The comments here
throughout pertain to vertical more so than horizontal;
however, looking at ground from the vertical perspective
includes all features of ground as both an absorber
(refractor) and reflector.

For issues of reflection, Roy's comments as to the necessity
of TA is one of performing "ray-tracing" to far field
features of complex geometries. TA is a far field program
filling in the niche not covered by NEC's assumptions of a
flat featureless earth. I would go one step further and say
that even TA falls short in not addressing the extremely far
field geometries of the ionosphere and its interactions with
solar activity. This type of modeler is found in CapMan, a
propagation modeler (I also use the VOA's propagation
modeler for forecasting which accepts NEC-like or TA-like
launch angle inputs).

Refraction issues are loss issues. This term appears to
have entered the discussion with some laxity in definition.
If we are to understand refraction to mean turning the
incident wave applied against an interface towards the norm
of the surface (as it would in the real world); then this
power is lost. However, this loss is only found with high
angles of incidence (very near the antenna). For any
appreciable distance from an antenna (and thus low angle of
incidence), the ground is so reflective as to return nearly
all power in reflection.

The sky wave is built out of reflections and direct
transmission (at least for a vertical) and the angles
presented and the high SWR mismatch at the far-field
interface presents very little loss and a lot of reflection.
This loss (ground-loss) only accumulates over the distance
of 10's of miles for near 0 degrees launch angle, whereas
the near low launch angles (5 degrees and above) are
developed and on their way within 10-30 wavelengths.

Jake Brodsky

unread,
Aug 24, 1998, 3:00:00 AM8/24/98
to
On Sat, 22 Aug 1998 09:05:10 -0400, Rich Griffiths <ri...@one.net>
wrote:
<snip>

>I'm still left with the question I just posed to Roy Lewallen in another
>posting. Is the ground's behavior as a refractor vs. a reflector
>sufficiently different that model predictions of the "take-off" angle
>and other behavior that many Hams hold near and dear need to account
>for this? Roy indicates that even advanced models like NEC do not
>account for the difference and that a separate correction, such as by
>TA, is needed.

I wonder if it might be feasible to map ground conductivity by using a
look-down radar from an aircraft. What I envision is something like
this:

Given a good estimate of the altitude, the frequency used by the
radar, the strength and the composition of the echo; one should be
able to figure out the path loss and thus the ground reflection loss.
The existence of non-homogeneous soil could be accounted for in this
manner too.

I figure that if one could map the ground in this fashion, and store
the whole thing in a map file of some sort, couldn't it become part of
a model and get useful real-world estimates?

Of course, it would be necessary to do several surveys so that one can
bracket the degree of ground system conductivity changes...

Jake Brodsky, AB3A mailto:fru...@erols.com
"Beware of the massive impossible!"

Reg Edwards

unread,
Aug 24, 1998, 3:00:00 AM8/24/98
to
Yes, that would be OK. But only at the frequency at which the
investigations are made.

Translation of the data to widely different frequencies of interest is next
to impossible.
--
Reg G4FGQ
http://www.btinternet.com/~g4fgq.regp

Jake Brodsky

unread,
Aug 24, 1998, 3:00:00 AM8/24/98
to
On 24 Aug 1998 18:49:04 GMT, "Reg Edwards" <g4fgq...@btinternet.com>
wrote:

>Yes, that would be OK. But only at the frequency at which the


>investigations are made.
>
>Translation of the data to widely different frequencies of interest is next
>to impossible.

Why?

If the frequency (and thus the wavelength) is known, and the
attenuation from ground reflection is known, couldn't one normalize
this result to a DC resistance per cubic meter? I seem to remember
some formulas from my physics classes describing electric and magnetic
potentials of a wave in an attenuating medium. We even figured the
attenuation of a radio wave through sea-water.

These were the very equations which were used to describe the
"skin effect," if I remember correctly...

Roy Lewallen

unread,
Aug 24, 1998, 3:00:00 AM8/24/98
to
From measurements which have been made using the method currently
regarded as best (involving a buried transmission line; I don't know
the details), the ground conductivity varies considerably with
frequency -- 2:1 or more over the HF range. I don't know about the
dielectric constant. I don't believe that the mechanism is known, and
unless it is, the properties couldn't be extrapolated from VHF or
higher down to HF with any accuracy.

The skin depth at 3.5 MHz in average soil is around 12 feet, so you'd
really want to know the conductivity and permittivity profile to a
depth of 30 feet or more. The skin depth at any reasonable radar
frequency would be much, much less, so you wouldn't be able to see to
nearly enough depth. I think that with radar, the best you'd do is get
the surface properties, and at the wrong frequency. It's currently
possible to measure the surface conductivity with some accuracy at the
frequency of interest, so radar seems like a step backward. Like a lot
of problems, the devil's in the details.

Roy Lewallen, W7EL

Reg Edwards

unread,
Aug 24, 1998, 3:00:00 AM8/24/98
to
"Next to impossible" was perhaps an exaggeration. Let's say it's difficult.
It is necessary to have a mathematical model of some kind to translate
ground characteristics from microwaves down to LF and HF.

You may be interested in my small program RADIOETH available by downloading
from my website. This begins with soil characteristics at DC and works its
way up to VHF. It models the input impedance, R + jX, of a single,
shallow-buried radial wire versus frequency. Other programs available, such
as ENDFEED and TANT136, make use of this model to predict performance of
various antenna+ground systems.

Download in a few seconds. Not zipped-up. Can be run immediately. Click
under my signature.

Rich Griffiths

unread,
Aug 25, 1998, 3:00:00 AM8/25/98
to
I went away for a few days and unfortunately missed a lot. I'll try to
keep this follow-up simple.

Gary Coffman wrote:
>
> Questions of whether we are dealing with reflection or refraction
> can best be answered by looking at the ratio of refractive indexes
> at the boundary. For even poor soil this ratio is very high. We can
> ignore refraction for take-off angles of interest to skywave propagation.

I don't understand this, for a few reasons.

Soil is layered. The upper horizons are thin compared with the figures
that have gone by here for skin depth in soil. I would expect each
horizon to have distinctly different properties from the others. This
makes for a messy reflection/refraction situation.

In any case, if the ratio of refractive indexes is high, shouldn't we
expect a pronounced bending toward the vertical (normal), resulting in
even less similarity to reflection?


Rich W2RG

Rich Griffiths

unread,
Sep 2, 1998, 3:00:00 AM9/2/98
to
Gary Coffman wrote:

>
> On Tue, 25 Aug 1998 19:01:58 -0400, Rich Griffiths <ri...@one.net> wrote:
>
> >Soil is layered. The upper horizons are thin compared with the figures
> >that have gone by here for skin depth in soil. I would expect each
> >horizon to have distinctly different properties from the others. This
> >makes for a messy reflection/refraction situation.
> >
> >In any case, if the ratio of refractive indexes is high, shouldn't we
> >expect a pronounced bending toward the vertical (normal), resulting in
> >even less similarity to reflection?
>
> Well, do realize that HF skin depth of soil is a small fraction of a
> wavelength. Even on 80m the skin depth is less than 1/20th of
> a wavelength. So we have to be a bit careful applying optical
> ideas of ray tracing to it.
>
> The portion of the wave which penetrates the soil will move slower
> than the portion remaining in air. So there is a sharp bending toward
> nadir of the wavefront at the soil/air interface. This ironically has
> the effect of raising the effective skin depth closer to the surface
> as far as the ray geometry is concerned because the wave reaches
> skin depth closer to the source than it would if this effect weren't
> acting.

>
> Gary
> Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
> 534 Shannon Way | We break it |
> Lawrenceville, GA | Guaranteed |

If the wave bends sharply toward the normal to the surface (nadir),
isn't skin depth no longer a relevant phenomenon? I understand skin
effect to be one of those odd self-defeating phenomena in which the
fields produce currents that produce fields that tend to drive the
current to the surface. Here, the fields are no longer aligned so as to
produce current (and resulting fields) in the directions that would
produce skin effect.

If the GPR stuff is an indication, what we have is a layered medium that
reflects a portion of the energy at each interface and absorbs some
amount within each layer. The energy reflected by each layer diminishes
with depth, is further attenuated on the way back out, and suffers
multiple redirection due to the varying refractive indexes. In many
locales, it's common to have 3 horizons within the upper meter and 4 or
5 within 5 meters, so it sounds like a tough mess to model.


Rich W2RG

Gary Coffman

unread,
Sep 8, 1998, 3:00:00 AM9/8/98
to
On Wed, 02 Sep 1998 20:53:58 -0400, Rich Griffiths <ri...@one.net> wrote:
>If the GPR stuff is an indication, what we have is a layered medium that
>reflects a portion of the energy at each interface and absorbs some
>amount within each layer. The energy reflected by each layer diminishes
>with depth, is further attenuated on the way back out, and suffers
>multiple redirection due to the varying refractive indexes. In many
>locales, it's common to have 3 horizons within the upper meter and 4 or
>5 within 5 meters, so it sounds like a tough mess to model.

It would be at frequencies where those depths are a large fraction of
a wavelength. But at HF frequencies, we can ignore those layers as
mere fringing effects without seriously affecting our models.

Rich Griffiths

unread,
Sep 8, 1998, 3:00:00 AM9/8/98
to
On Tue, 08 Sep 1998 15:03:57 GMT, ke...@bellsouth.net (Gary Coffman)
wrote:

Ground Penetrating Radars generally operate in the range of 7 MHz to
30 MHz. So we're talking about the same range of frequencies, same
range of effects.


Rich W2RG


0 new messages