I decided to bite the bullet and purchase AutoEZ, which is a front end
to EZNEC which automates building models using formulas, generating
test cases, optimizing designs, etc.

http://www.ac6la.com

One of my first models was a common ground plane antenna with the
variables being the radiator length, radial length, radial droop
angle, and the height of the bottom of the radials over ground.

The first run was to optimize a ground plane for 10M built of 5/8"
aluminum tubing with a droop angle of zero degrees, i.e. horizontal
radials, in free space.

The next run was to optimize the same ground plane with a radial droop
angle of 45 degees, as is more common for elevated ground planes.

The results are shown below with the lengths being in wavelengths.


0 degree radials 45 degree radials

SWR gain radiator radial SWR gain radiator radial
1.01 1.7 0.322 0.086 1.00 2.2 0.231 0.244

Note that the antenna has slightly better gain with a 45 degree
radial droop.


The next step was to run test cases for the two antennas at various
heights above real, average ground, which yielded the following.

Height is in wavelengths above ground for the bottom of the radials
and @ elev is the elevation angle of the main lobe.

0 degrees 45 degrees

Height gain @ elev gain @ elev

0.01 -0.76 24 0.78 21
0.10 0.27 21 1.32 18
0.15 0.61 19 1.50 17
0.20 0.85 17 1.59 15
0.25 1.01 16 1.63 15
0.30 1.09 15 1.63 14
0.35 1.11 14 1.62 13
0.40 1.11 13 1.63 13
0.45 1.12 13 2.02 43
0.50 1.18 44 2.37 40
0.55 1.59 41 2.64 38
0.60 1.91 39 2.86 36
0.65 2.16 36 3.03 34
0.70 2.37 35 3.15 32
0.75 2.52 33 3.21 30
0.80 2.61 31 3.22 29
0.85 2.65 29 3.18 27
0.90 2.65 28 3.16 9
0.95 2.62 27 3.29 9
1.00 2.70 9 3.42 9
1.05 2.83 9 3.54 9
1.10 2.95 9 3.67 9
1.15 3.08 9 3.81 8
1.20 3.20 8 3.95 8
1.25 3.35 8 4.07 8

The very surprising result, at least to me, was that 1/2 wavelengths is
the worst height to mount a ground plane as the elevation angle of the
main lobe is at the maximum elevation of about 45 degrees around that
height for both versions of the antenna.



Special note:

Most people understand that the results of an antenna analysis program
reflect the material used to construct the antenna and the type of
ground, if any, used for the analysis, are an approximation, and are
not accurate to 27 decimal places.

Further, most people also understand that absent them being a part of
the model used for the analysis, objects in the near field of the antenna,
such as, but not limited to, 20 foot prision walls, blimp hangers,
skyscrapers, a deluge of biblical proportions, giant sequoia trees,
hovering 2 mile wide alien spacecraft, hords of locusts, large gold
deposits under the antenna, battles between Autobots and Decepticons,
beached aircraft carriers, and stadium domes may well effect the
actual antenna perfomance.

Your mileage may vary, void where prohibited.

Any spelling mistakes in this article are all entirly my fault. Any grammer
errors spotted in this article were put there because I could.

--
Jim Pennino

[...]

manual calculation of a lossless Lambda/4-Ground-Plane-Antenna (radials at
0°):

www.leobaumann.de/unbenannt.pdf

Izur Kockenhan

[...]

manual calculation of a lossless Lambda-5/8-Ground-Plane-Antenna (radials at
0°):

www.leobaumann.de/unbenannt1.pdf

lol - Izur Kockenhan

El 19-11-14 7:24, Izur Kockenhan escribió:
[...]

manual calculation of a lossless Lambda-5/8-Ground-Plane-Antenna (radials at
0°):

www.leobaumann.de/unbenannt1.pdf

lol - Izur Kockenhan



The document is nice for the antenna only, however we can't avoid
mother earth. To get the far field elevation pattern for the antenna
with respect to mother earth, you need to take the ground properties
into account.

When the quarter wave or 5/8 lambda antenna is say 0.25lambda above
ground, you can use the two ray model using the direct ray and the
reflected ray to get a good impression. The reflection (versus
elevation angle) you can calculate using the Fresnel Equations for
media interfaces (using p-polarization).

--
Wim
PA3DJS
Please remove abc first in case of PM

On 11/18/2014 2:23 PM, wrote:
I decided to bite the bullet and purchase AutoEZ, which is a front end
to EZNEC which automates building models using formulas, generating
test cases, optimizing designs, etc.

http://www.ac6la.com

I got it a few weeks ago. Once you get familiar with it, there is no
turning back. I love it.

His app is listed on Roy's EZNEC home page as well.

He also has a free demo available for try-out.

Cheers,
John

On Wednesday, November 19, 2014 5:48:31 AM UTC-6, Wimpie wrote:

When the quarter wave or 5/8 lambda antenna is say 0.25lambda above
ground, you can use the two ray model using the direct ray and the
reflected ray to get a good impression. The reflection (versus
elevation angle) you can calculate using the Fresnel Equations for
media interfaces (using p-polarization).

I remember when I was using a 40m GP at slightly higher than a quarter
wave up, I always had to increase the ground qualities when modeling, if I
wanted the modeled test results to match the real world results I saw when
comparing to other antennas. And even then, they often came up a tad short
in the models. If I used "average", the modeled vertical results would
be quite stunted and way below the results I saw in the real world.

Instead of "average" ground quality, I would have to use "very good"
or "excellent", or whatever the exact names they called those in the
programs.

This area has pretty decent ground conductivity according to the maps.
Not sure how that would apply to people with a lesser ground quality.
Their modeled results may be closer to their real world results.

I found one thing when comparing a full size 40m radiator. The ground
mount version with 32 radials, was no where near as good as the same
32 ft whip elevated at 36 ft, with only four sloping radials.
It was like a whole new antenna when elevated.

After using it a while, I rigged up a 24 volt relay at the feed point
so I could add a small matching coil for 17m 5/8 wave GP use.
I could take it in or out of the circuit from the shack by plugging or
unplugging a 24v wall wart. :/ Worked out very nice.

On 11/19/2014 1:37 PM, wrote:
On Wednesday, November 19, 2014 5:48:31 AM UTC-6, Wimpie wrote:

When the quarter wave or 5/8 lambda antenna is say 0.25lambda above
ground, you can use the two ray model using the direct ray and the
reflected ray to get a good impression. The reflection (versus
elevation angle) you can calculate using the Fresnel Equations for
media interfaces (using p-polarization).

I remember when I was using a 40m GP at slightly higher than a quarter
wave up, I always had to increase the ground qualities when modeling, if I
wanted the modeled test results to match the real world results I saw when
comparing to other antennas. And even then, they often came up a tad short
in the models. If I used "average", the modeled vertical results would
be quite stunted and way below the results I saw in the real world.

Instead of "average" ground quality, I would have to use "very good"
or "excellent", or whatever the exact names they called those in the
programs.

This area has pretty decent ground conductivity according to the maps.
Not sure how that would apply to people with a lesser ground quality.
Their modeled results may be closer to their real world results.

I found one thing when comparing a full size 40m radiator. The ground
mount version with 32 radials, was no where near as good as the same
32 ft whip elevated at 36 ft, with only four sloping radials.
It was like a whole new antenna when elevated.

After using it a while, I rigged up a 24 volt relay at the feed point
so I could add a small matching coil for 17m 5/8 wave GP use.
I could take it in or out of the circuit from the shack by plugging or
unplugging a 24v wall wart. :/ Worked out very nice.





You are not the first ham to discover that elevated radials work so much
better than those in the ground. I guess in theory, 120 radials in the
ground would work OK, but who wants to do that when elevating four
radials does better.

I was pretty shocked when I first read about this in the attached pdf