Gee, Richard, it's only been seven months this time since you last
posted essentially the same comments and questions. You were going a
year or so between.

Here are a couple of the previous ones, with my responses. Anyone
interested in more information can do a google groups search of my
postings on this newsgroup containing "surface wave".

4-20-2008:

Richard Fry wrote:
"Roy Lewallen"
However, you can't compensate for this factor when the ground is
poor by improving the ground system. The reason is that the reflection
takes place much farther from the antenna than nearly any ground system
extends. And low angle radiation, where the improvement is most needed,
reflects the greatest distance away.
___________

Roy, didn't the experiments of Brown, Lewis & Epstein of RCA in ~1937
show that the h-plane field measured 3/10 mile from a vertical monopole
of about 60 to 88 degrees in height, over a set of 113 buried radials
each 0.41 WL, was within several percent of the theoretical maximum for
the applied power as radiated by a perfect monopole over a perfect
ground plane? And conductivity at the NJ test site was poor -- 4 mS/m
or less.

That tends to show that the fields radiated at very low elevation
angles also will be close to their theoretical values when measured at
this radial distance, even though ground conductivity at the antenna
site is poor. The relative field (E/Emax) for radiators of these
heights and propagation paths approximately equals the cosine of the
elevation angle.

I believe we've discussed this before, so I'll be brief.

Their calculation of the field at the receiving site when the radial
system is perfect was adjusted for the effect of ground wave attenuation
caused by the imperfect ground conductivity. If the ground between the
antenna and receiving site were perfect, the field strength would have
been greater.

Also, I'm speaking of sky wave. Ground reflection isn't a factor in
determining surface wave, which is what they measured and which isn't of
interest to most amateurs.

The greatest radiated fields always will be directed in or near the
horizontal plane when measured/calculated for such conditions. This
also will be true for any monopole from infinitesimal to 5/8 wavelength
in height, although the elevation pattern of monopoles from /4- to
5/8-WL no longer are described by the cosine function (see
http://i62.photobucket.com/albums/h8...omparison.jpg).

Elevation patterns show maximum relative field centered at various
elevation angles above the horizon, when those fields are measured at
progressively longer radial distances from the monopole, due to the
propagation loss for the surface wave over other than a perfect, flat,
infinite ground for those ranges. Earth curvature and terrain
diffraction add to those losses for longer surface wave paths over real
earth, and for very great distances the h-plane relative fields falls to
~zero.

As I thought you were aware, the surface wave propagates considerably
differently than the sky wave.

But that pattern shape is not the pattern shape originally radiated
by the monopole, it also includes the effects of the propagation
environment at the range where it was measured (or calculated).

If this were not true then MW broadcast stations would have
essentially zero coverage area for their groundwave signals.

It would be a mistake to design HF antenna systems based on optimizing
surface wave propagation as AM broadcasters do, unless you desire
communication for distances not exceeding a few miles.

Roy Lewallen, W7EL

4-22-2006:

Richard Fry wrote:
"Roy Lewallen" wrote:
The maximum far field (sky wave) gain of a ground mounted
quarter wave vertical over average ground, with a completely
lossless ground system, is on the order of 0 dBi, and this occurs at
roughly 25 degrees above the horizon (both depending on
frequency as well as ground characteristics).
_____________

The above is an understandable conclusion using NEC analysis, however
it is not supported empirically. If it was, AM broadcast stations would
perform very much differently than they do.

NEC analysis has been supported many times by measurement and observation.

The measured data in Brown, Lewis & Epstein's 1937 benchmark paper
"Ground Systems as a Factor in Antenna Efficiency" proved that the
*radiated* groundwave field from a vertical monopole working against 113
buried radials each 0.41 lambda in length was within a few percent of
its calculated peak value for a radiation pattern with maximum gain in
the horizontal plane. The path length for the measurement was 0.3 miles,
which was in the far field of the vertical monopole configurations measured.

Yes. The question is what is the calculated value. B, L, and E
normalized their measurements to the unattenuated field strength at one
mile for 1000 watts radiated power. I couldn't find anywhere in their
paper where they explained how they determined the ground attenuation
between the antenna and their observation point.

BL&E's measurements, and the results of thousands of measurements
made of the groundwave fields of MW broadcast stations using such radial
ground systems ever since demonstrate that their peak gain always lies
in the horizontal plane.

No, the field strength is strongest at low elevation angles only close
to the antenna, as you further explain below.

It is true that, as a groundwave propagation path becomes longer, the
field measured at increasing elevations above the earth at distant
ranges might be higher than measured at ground level at those ranges.
But that is not because more field was launched by the original radiator
toward those higher elevations -- it is because the the groundwave path
has higher losses, which accumulate as that path lengthens. Therefore a
NEC plot showing the conditions reported in the quote above do not
accurately depict the elevation pattern as it is launched from the
radiator, and the groundwave field it will generate.

Of course the standard far field analysis doesn't accurately depict the
field close to the antenna -- it's a plot of the field at points very
distant from the antenna, as clearly explained in the manual. NEC allows
you to include the surface wave if you want, and it accurately shows the
total field including the surface wave at a distance of your choice.
(Accurate, that is, up to a hundred km or so, beyond which the deviation
of the flat ground model from the curved Earth begins affecting results.)

Don't feel bad -- Reg has a lot of trouble understanding this, too.

There are software programs designed for calculating MW groundwave
field strength given the FCC "efficiency" of the radiator and the
conductivity of the path. The radiator efficiency is the groundwave
field developed by the radiator with a given applied power at a given
distance (1 kW @ 1 km). These values must meet a certain minimum level
for the class of station. I think in all cases, they must be within
~0.5 dB of the theoretical value for a radiation pattern with its peak
gain in the horizontal plane. In the case of directional MW antennas,
this performance must be proven by field measurements.

Finally, standard equations show a peak field of ~137.6 mV/m at 1
mile from a 1/2-wave dipole radiating 1 kW in free space. The
calculated groundwave field at 1 mile radiated by 1 kW from a 1/4-wave
vertical MW monopole over a perfect ground plane is ~195 mV/m. This is
the same field as generated by the free space 1/2-wave dipole, when all
radiation is confined to one hemisphere (137.6 x 1.414).

The groundwave fields measured from thousands of installed MW
broadcast antenna systems confirm that their intrinsic radiation
patterns are within a fraction of a decibel of that perfect radiator
over a perfect ground plane, no matter what is the conductivity at the
antenna site (N.B. Reg).

No, the measured fields from quarter wave broadcast antennas are
considerably less than 195 mV/m for 1 kW at one mile, unless perhaps
there's only salt water between the antenna and measurement point. As
you explained above, the surface wave is attenuated with distance. What
you seem to be missing is that the attenuation is strongly dependent on
ground conductivity (between antenna and measurement point, not just at
the antenna site) and frequency, so the actual field strength at one
mile for 1 kW radiated will always be considerably less than the perfect
ground case. The 195 mV/m and associated values for various antenna
heights is the "unattenuated" or "inverse" field, which doesn't include
the surface wave attenuation beyond simple inverse distance field
strength reduction. It's the field strength you'd get if the ground
between antenna and measurement point were perfect, not what you get
over real ground. I'm not very conversant with FCC antenna measurement
methodology, but somewhere the measured field strength is normalized to
the unattenuated field strength by fitting to a ground attenuation
curve, which in turn depends on frequency and ground conductivity. (I've
been told that this is the way broadcasters determine ground
conductivity -- by seeing how far the measured field strength deviates
from the unattenuated value.) I believe that the surface wave
attenuation curves used by the FCC are from the 1937 I.R.E. paper by
K.A. Norton. That paper is also the basis for NEC's surface wave
calculations.

Roy Lewallen, W7EL

"Roy Lewallen" wrote
Of course the standard far field analysis doesn't accurately depict the
field close to the antenna -- it's a plot of the field at points very
distant from the antenna, as clearly explained in the manual. NEC allows
you to include the surface wave if you want, and it accurately shows the
total field including the surface wave at a distance of your choice.
_________

Not the versions of NEC used by most amateurs.

Those versions show zero or very low gain in/near the horizontal plane
for a
vertical monopole over real earth. This leads to the common (mis)
belief
that those are the gains of the radiation pattern _originally
generated_
by the monopole.

But that belief is untrue.

RF

On Wed, 26 Nov 2008 03:39:41 -0800 (PST), Richard Fry
wrote:

"Roy Lewallen" wrote
Of course the standard far field analysis doesn't accurately depict the
field close to the antenna -- it's a plot of the field at points very
distant from the antenna, as clearly explained in the manual. NEC allows
you to include the surface wave if you want, and it accurately shows the
total field including the surface wave at a distance of your choice.
_________

Not the versions of NEC used by most amateurs.

Those versions show zero or very low gain in/near the horizontal plane
for a
vertical monopole over real earth. This leads to the common (mis)
belief
that those are the gains of the radiation pattern _originally
generated_
by the monopole.

But that belief is untrue.

RF

Roy and others have answered this one in the past too. You employ the
near field table to observe the ground wave. It works approximately
well, even out to the edge of the implicit flat universe. If you
object to flat universes, you are no longer in the realm of ground
wave. If anything, modelers give MORE response in comparison to the
BL&T data.

73's
Richard Clark, KB7QHC

Richard Clark wrote:

Roy and others have answered this one in the past too. You employ the
near field table to observe the ground wave. It works approximately
well, even out to the edge of the implicit flat universe. If you
object to flat universes, you are no longer in the realm of ground
wave. If anything, modelers give MORE response in comparison to the
BL&T data.

With EZNEC, you have to use the near field analysis to include the
ground wave; direct ground wave analysis isn't included in EZNEC because
it's of very limited use to most amateurs. In fact, it's really of
interest only to AM broadcasters and a very few other very small and
specialized users, who just about universally use EZNEC Pro which does
include direct ground wave analysis.

But anybody who's interested in direct ground wave analysis can also use
one of the other modeling programs which includes it, or use NEC-2
itself, available free from http://www.si-list.net/swindex.html. The
manual is available there also. NEC-2 produces field strength data
including the ground wave if requested with the proper entry on the RP
"card".

Roy Lewallen, W7EL

"Roy Lewallen" wrote
With EZNEC, you have to use the near field analysis to include
the ground wave; direct ground wave analysis isn't included in
EZNEC because (etc).
_________

After the comments of Richard Clark and you, Roy, I attempted to use
EZNEC to determine the ground wave (see link below).

The near-field analysis of EZNEC for radiation in the horizontal plane
at a point 1 km from a 1/4-wave monopole having two ohms in series
with a Mininec r-f ground, while radiating 1 kW over an earth
conductivity of 8 mS/m is shown as 72 mV/m.

The same setup when analyzed using the FCC's radiation efficiency for
this monopole height, and their propagation charts for these
conditions shows about 295 mV/m as the result, which value is
supported by the measured performance of real-world AM broadcast
stations, and is also a value in a range that could be expected from
the BL&E data.

Hopefully you or Richard Clark can tell me the reason(s) for this
difference, which could easily be my own setup of the NEC model.

Roy, would you mind posting the ground wave value EZNEC Pro reports
for these conditions?

http://i62.photobucket.com/albums/h8...FldExample.gif

RF

On Wed, 26 Nov 2008 11:41:04 -0800 (PST), Richard Fry
wrote:

Hopefully you or Richard Clark can tell me the reason(s) for this
difference, which could easily be my own setup of the NEC model.

I modeled their structures as they built them explicitly (they had
many variations), at the frequency they used, took readings at the
distance they reported. For your 1 kilometer distance (not one they
used), I get 303 mV/m at 3 MHz for their 70 foot radiator over a field
of 113 x 135 foot radials with an average ground conductivity. When I
use their distance of a mile, I get 188 mV/m, all else identical.
Their paper reports by formula that I should see 194.5 mV/m.

It would appear that with the average of the two distances, my model
accords quite closely to BL&E. That average would suggest results are
within an unreasonable accuracy given my experience with making RF
power determinations at the bench. However, my model is repeatable,
the paper is chiseled into the granite of history and your original
complaint seems to be moot.

73's
Richard Clark, KB7QHC

Earlier posts in this thread:

From: "Richard
Date: Wednesday, November 26, 2008 12:00 PM

Roy and others have answered this one in the past too. You employ the
near field table to observe the ground wave. It works approximately
well, even out to the edge of the implicit flat universe. If you
object to flat universes, you are no longer in the realm of ground
wave. If anything, modelers give MORE response in comparison to the
BL&T data.

Then I posted my result of using the near-field analysis of EZNEC
showing a value much LESS than the "modeler" value for those
conditions when using the BL&E data and the FCC curves.

Following that is posted:

From: "Richard Clark"
Date: Wednesday, November 26, 2008 3:32 PM

I modeled their structures as they built them explicitly (they had
many variations), at the frequency they used, took readings at the
distance they reported. For your 1 kilometer distance (not one they
used), I get 303 mV/m at 3 MHz for their 70 foot radiator over a field
of 113 x 135 foot radials with an average ground conductivity. When I
use their distance of a mile, I get 188 mV/m, all else identical.
Their paper reports by formula that I should see 194.5 mV/m. ...

No, the BL&E paper (accurately) stated that 194.5 mV/m is the
theoretical maximum field possible at 1 mile for 1 kW radiated by a
perfect 1/4-wave monopole over a perfect ground plane. The peak
values they measured came very close, but never quite achieved that
value.

It would appear that with the average of the two distances, my model
accords quite closely to BL&E.

Mr. Clark - kindly note that in your first quote above you say that,
if anything, "modelers" show MORE response than BL&E Then when
pressed a bit you say that your model "accords quite closely" with
BL&E.

Yet the results of my EZNEC near-field model showed considerably LESS
ground wave field at 1 km than either the FCC approach or the BL&E
data.

Clarifications, please?

RF

Richard Fry wrote:
"Roy Lewallen" wrote
With EZNEC, you have to use the near field analysis to include
the ground wave; direct ground wave analysis isn't included in
EZNEC because (etc).
_________

After the comments of Richard Clark and you, Roy, I attempted to use
EZNEC to determine the ground wave (see link below).

The near-field analysis of EZNEC for radiation in the horizontal plane
at a point 1 km from a 1/4-wave monopole having two ohms in series
with a Mininec r-f ground, while radiating 1 kW over an earth
conductivity of 8 mS/m is shown as 72 mV/m.

The same setup when analyzed using the FCC's radiation efficiency for
this monopole height, and their propagation charts for these
conditions shows about 295 mV/m as the result, which value is
supported by the measured performance of real-world AM broadcast
stations, and is also a value in a range that could be expected from
the BL&E data.

Hopefully you or Richard Clark can tell me the reason(s) for this
difference, which could easily be my own setup of the NEC model.

Roy, would you mind posting the ground wave value EZNEC Pro reports
for these conditions?

http://i62.photobucket.com/albums/h8...FldExample.gif

RF

My model has 120 0.5 wavelength radials buried 1.2 feet deep (the
unusual depth due to rescaling another model). Ground conductivity 8
mS/m, dielectric constant 13. The antenna is 0.25 wavelength high. The
whole structure is made from #12 wire to eliminate any problems due to
dissimilar diameters. Field strength is Ez at 1000 meters with 1000
watts applied.

Using the NEC-4D calculating engine, EZNEC Pro/4 shows (NF = near field
analysis, GW = far field analysis with ground wave):

Z = 40.08 + j27.91
GW = 297.7 mV/m
NF = 297.7 mV/m

Same, but with 0.25 wavelength radials:

Z = 39.56 + j26.55
GW = 292.7 mV/m
NF = 292.7 mV/m

Note that the feedpoint R and field strength don't exactly correlate if
you make the assumption that the resistance difference is due to loss.
This would be due to a slightly different current distribution on the
radiator due to interaction with the different ground fields. Other
experiments have shown that the impedance will also vary some with
radial burial depth.

Following are the results using the NEC-2D engine with a 0.25 wavelength
vertical and 120 0.5 wavelength radials one foot above the ground, all
other conditions otherwise the same. This analysis can be run with
EZNEC+, but only the NF results will be available:

Z = 66.83 + j1.894
GW = 230.0
NF = 229.7

As above, but 0.25 wavelength radials:

Z = 32.42 + j18.87
GW = 311.4
NF = 311.4

Elevated radials, even when elevated only this amount, show distinct
resonance effects, and making them longer than about 0.25 wavelength
often results in reduced efficiency which I think is due to movement of
the radial current maxima away from the center. The above results
illustrate these phenomena. While slightly elevated radials can be used
to approximate buried ones, as you can see the substitution isn't perfect.

The same 0.25 wavelength vertical over perfectly conducting (or MININEC)
ground showed a Z of 37.95 + j21.49 ohms. However, the resistances of
the various examples above aren't just this resistance plus loss
resistance, since the current distribution isn't quite the same when
radials are present.

The results you got weren't valid due to use of MININEC ground with near
field analysis, as I explained in another posting. As you can see, you
can get reasonably good results using EZNEC+ and near field analysis,
although the vast majority of people this intensely interested in the
mechanisms of AM broadcasting aren't hobbyists but rather professional
engineers who are using EZNEC Pro/4.

Roy Lewallen, W7EL

On Nov 29, 1:10*pm, Roy Lewallen wrote:

As you can see, you can get reasonably good results using
EZNEC+ and near field analysis, although the vast majority
of people this intensely interested in the mechanisms of AM
broadcasting aren't hobbyists but rather professional
engineers who are using EZNEC Pro/4.
________

Thanks very much for your numbers and comments, Roy.

I would never have thought to try to use EZNEC near-field analysis to
compute the groundwave if I hadn't read the suggestion to do so in
this thread. That was my first, and will be my last attempt at that.

When I need to calculate the MW ground wave for a particular distance,
monopole height, frequency and ground conductivity I use the FCC
method of first determining the inverse distance field of the radiator
at 1 km for 1 kW of radiated power, and then using that value in a
program I have with the FCC's MW propagation curves in digitized form.

My point when starting this thread was to show that the elevation
pattern radiation actually launched by vertical monopoles on any
frequency does not have a zero/very low relative amplitude at/near the
horizontal plane, which from what I read on these NGs seems to be a
popular belief.

RF

Richard Fry wrote:
"Roy Lewallen" wrote
Of course the standard far field analysis doesn't accurately depict the
field close to the antenna -- it's a plot of the field at points very
distant from the antenna, as clearly explained in the manual. NEC allows
you to include the surface wave if you want, and it accurately shows the
total field including the surface wave at a distance of your choice.
_________

Not the versions of NEC used by most amateurs.

Those versions show zero or very low gain in/near the horizontal plane
for a
vertical monopole over real earth. This leads to the common (mis)
belief
that those are the gains of the radiation pattern _originally
generated_
by the monopole.

But that belief is untrue.

RF

I'm not aware of such a widely held misconception, but it is indeed untrue.

You must have conducted some sort of survey to find out what "version"
of NEC-2 is used by most amateurs, you know more than I do about it.
Although there are a few minor variations of NEC-2, there's not enough
difference to justify characterizing them as different "versions". And
every one I've seen has the capability of producing data which include
the surface wave. It's done with a simple entry on the RP "card".

A number of different compilations of NEC-2, every one of which as far
as I know which includes ground wave analysis, can be had at
http://www.si-list.net/swindex.html. You can also get the manual there,
where you can learn about how to get the ground wave data from it.
You're obviously much more interested in ground wave field strengths
than most amateurs, so I suggest that you download one of the free NEC-2
compilations, and the manual, and them for your investigations.
Remember, it's done with the RP "card".

Roy Lewallen, W7EL