[TenTec] NEC, ground, grounds, and radials.

[Dr. Gerald N. Johnson]

On 1/5/2011 3:44 PM, Richards wrote:
> Er... ah...   No joke.   Do you doubt the proposition that a
> good ground plane lowers radiation take off angle?  If so,
> I would be curious as to your reasons.   (Seriously, and with
> no intention to flame or cause an argument, as I have
> made quite an investment in time and resources in vertical
> developing a reasonably decent vertical antenna system
> for my small, suburban back yard.   Any info would be
> greatly appreciated. )
>
> FYI ---
>
> ------ Extra Class Exam Questions of interest -----
>
> Question E9A12  - and the answer is that the efficiency
> of a quarter wave grounded vertical antenna can be improved by
> installing a good radial system.
>
> Question E9A13 - answer is - soil conductivity is the most important
> factor in determining ground losses for a ground-mounted vertical
> antenna operating in the 3-30 MHz range.
>
> Question E9C13 - answer is - When a vertically polarized antenna
> is mounted over seawater versus rocky ground, the far-field
> elevation pattern low-angle radiation increases.
>
> Question E9C17 - answer - The main effect of placing a vertical
> antenna over an imperfect ground is that it reduces low-angle
> radiation.
>
> Also, on the audio study guide, Gordo makes some stray comments
> about using 3 inch copper strap for radials on his roof, and
> mentions improving the ground field increases low angle
> radiation  - he goes on to say it does not increase signal strength,
> of course, but only that it increases the amount of signal that
> has low take off angle.
>
> This information is consistent with all that I have read on verticals
> in the ARRL Antenna Handbook, and I did LOTS of research before
> installing a large vertical monopole in the back yard.
>
> A huge ground pane does lower take off angle (as NEC modeling
> shows) and also improves antenna efficiency -- I stopped at 65 radials
> but I wish had installed even more just to be sure.    Also, my back yard
> soil is very conductive and remains moist even through the summer, so
> the soil, itself, helps me considerably.   Rob Sherwood and I exchanged
> some nice email at the time I was doing this homework, and I believe
> he lives over a more dry, rocky soil, and that is much harder to work
> over.   I also corresponded with the infamous  Rudy Severns N6LF
> and his findings are consistent with this conclusion.    NEC modeling
> produces consistent results.
>
> N'est ce pas?       Happy trails OM.
>
> ===============  JHR  ============================
>
>
NEC modeling of a ground plane is not all that dependable. NEC has some 
significant foibles, it doesn't couple currents to that ground plane, it 
just reflects from that ground plane to the distant receiving site. It 
tosses in some loss for a real vs a perfect ground plane in that 
reflection. It neglects the induced ground currents. Its bad at wire 
junctions that aren't at exactly 90 degrees. Its results depend on the 
segmentation of the wires in the antenna. It can't hack wires in 
parallel with close spacings. It models wires for some of its 
computations as infinitesimally thin, using the diameter mostly for the 
resistive loss part of the computation. e.g. it models the current as 
being at the center of the conductor. And I maintain that its 3 dB high 
in antenna gain or field strength when a ground plan is involved. It 
models connections to earth as resistive, but doesn't work out the 
current distribution in the ground.

I've been using NEC and MININEC since the days of the Apple II+. Since 
the program for the Apple II+ came in source code, my copy drops the 
gain of an antenna over a ground plane by 3 dB.

Segmentation. The docs with MININEC and NEC suggest you start with a few 
segments and keep increasing them until small changes in the number of 
segments results in practically no change in results. Maybe I got it 
into round off error problems from the larger matrix, but on the Apple 
II+ with MININEC, I found that phenomena and then if I kept increasing 
segments there came a number where the solutions got wilder and wilder 
with more segments. That didn't give me super confidence in that concept 
for selecting segments. Segmenting the wires and emulation shapes other 
than wires with wire meshes are some of the arts required to get useful 
results from NEC, often limited by the number of total segments allowed 
by EZNEC or the computer available memory and acceptable computation time.

Lately I see some suggestions that parallel wires (but further apart 
than a feed line) should be segmented so that the segment joints are in 
line even though the wires are different lengths (think a driven element 
and a closely spaced first director in a long yagi). W7EL suggests and 
implements a tapering segment length near wire junctions hoping to solve 
some of those problems or limitations.

Some years ago I had an inspiration to do a directional antenna using 
wires close to 3/4 wave long and slanted to use the main lobes of such 
wires. I called it a "cats cradle" because from the top it looked a lot 
like a cat's cradle string without the straight side strings. I modeled 
and I optimized, shifting frequency sometimes to get more optimization. 
And I varied the corner angles which were never exactly 90. To make them 
model better I inserted a very short wire between the two angled ones, 
that wire being at right angles to each of the longer wires. I couldn't 
tell it made much difference, sometimes it would show 12 dBi gain, 
sometimes 9 dBi gain. I wasn't able to detect what foible made such a 
difference in free space gain. At the last minute before a Central 
States VHF Conference in Cedar Rapids, I plotted out the antenna on 
multiple sheets of paper, taped them together and grabbed a roll of #12 
wire and made one for 1296. I'd also computed its results at 902. It was 
distinctly better if stacked four high to cancel out the radiation broad 
side to the diamonds. At the conference, where half a day is dedicated 
to antenna measuring, it came in around 9 dBi on 1296 and showed the 
strong up and down lobes. Dashing my hopes of 12 dBi. When I got home 
and tried to reproduce the 12 dBi, the program didn't ever again.

As for the effects with ground planes and my claim of error. I base it 
on this: Model a quarter wave vertical on a perfect ground plane. It 
will show 3 dB more gain than a half wave dipole in free space. Yet the 
theory of images in the ground plane insists that the quarter wave 
vertical on the ground plane has a image of the other half making it the 
exact equivalent of a half wave dipole. I claim that while the program 
in free space is comparing the signal intensity from the antenna to that 
of a perfect isotropic radiator located at the 0,0,0 origin of the axes, 
that when the ground plane is present it cuts that isotropic radiator in 
half, shielding half of its radiated power and so the reference to a 
full isotropic radiator is 3 dB in error. 3 dB too much gain.

Anyway because the NEC family doesn't couple wire currents to the ground 
plane it ignores the effects of those ground currents radiating and the 
added losses to the wires near the ground from coupling to the resistive 
ground. That makes some conclusions about low radials and low antennas 
somewhat in error based on NEC calculations. Just an area its a whole 
lot better than wild guesses, but no where near precise.

Even for free space calculations, those who hang their hats on antenna 
designs for VHF and up yagis don't believe any NEC results until 
confirmed on the antenna range.

While NEC has its foibles its a whole lot handier than the techniques 
used during WW2 where the antenna had to be described with a 
differential or integral equation that had to be solved. Pure analytical 
solutions might have taken a year of a good mathematician or numerical 
solutions a half year by a bunch of grunts with log tables and a room 
full of Friden calculators. And then short cuts introduced 
approximations that weren't always perfect. Assumptions were made on 
current distribution on the resonant antenna.

In the 1930s some broadcast verticals were made with lots of taper from 
ground to top and were sometimes 5/8 wave tall which for a wire is an 
optimum height for gain at the horizon. The upper half wave is separated 
more from the image, yet the radiation from the 1/8th wave isn't strong 
enough to get the null at the horizon typical of a 3/4 wave antenna. 
When they checked field strength, it didn't compute. Then they checked 
the current distribution. It didn't compute. There wasn't the phase 
change 1/8 wave from the ground. The varying cross section of the self 
supporting tall tower messed all that up. Another popular tower for a 
few years was made up of two self supporting tapered towers, the bottom 
one upside down and guyed and the top one not guyed. That fat in the 
middle construction didn't match the current distribution assumed for a 
simple wire. So those also fell out of favor though there was one north 
of Des Moines supporting TV and state communications VHF antennas after 
the radio station abandoned it. Its gone now.

Now to the details of this discussion about grounding and losses. For 
sure a vertical with just a ground rod or I think with very short 
radials and a gang of ground rods has a lot of ground resistance that 
the feed line sees in series with the radiator and its radiation 
resistance. That tends to raise the impedance and so lowers the RF 
current and the radiation from the antenna. Additional ground rods are 
ineffective unless spaced a ground rod apart at 60 Hz and DC. A lot of 
that is because the earth is such a lousy conductor it takes that 
spacing to get beyond the voltage drop of the dirt around the first rod. 
The only thing that makes an earth conductor good for telephone and 
power is that while resistive it can have a large cross section. 
Carson's uniform ground theory (from the 30s taught to power line 
engineers) shows that no matter how much copper you put into the aerial 
neutral of a power line more than half the return current will flow in 
the earth. It does that with significant voltage drop at the earth 
terminals, called ground rods. But not more significant than miles of a 
practically sized aerial conductor. One of the benefits of high voltage 
transmission is that the current is smaller. At 7200 volts 60 times 
smaller than the 120 volt load current. Carson's theory shows for power 
line computations the resistance of the earth conductor is practically 
consant whether the distance is a few hundred yards or many miles.

You have two phenomena with a vertical and earth. One is that conductive 
loss when the ground contact area isn't large, the other is the 
reflection from the earth that sets the vertical pattern. And a 
vertically polarized signal induces earth currents at a distance that 
cause resistive attenuation of the traveling/propagating wave. Ideally 
the ground plane should extend from the transmitting antenna to the 
receiving antenna and be solid metal with no resistance. That's hardly 
practical so we compromise at quarter wave radials because even in the 
air the open end high impedance transforms to a low impedance at the 
feed point and reduces the currents on the outside of the feed line. My 
NEC models say it takes more than three in the air to be effective at 
decoupling the feedline current, that four isn't bad, more is better. 
Broadcast standards like 160 and up because they measure actual field 
strength at a few miles distance and up to 160 they see a stronger 
signal with more radials. And some beyond that but the increased signal 
strength isn't commensurate with the added costs. Of course they aren't 
working for round the world DX, they are working for ground wave 
coverage signal strength.

The halfwave and 5/8 wave verticals whether bottom fed or center fed 
have small feed line currents and so the resistance of the earth 
connection is a much smaller fraction of the antenna feed impedance and 
the losses from that smaller with little or no conductive ground plane. 
Lifting the radiating portion of the 1/2 wave center fed vertical or the 
5/8 th wave ground mounted vertical should reduce coupling to the earth 
and those coupling losses, but NEC won't show that.

Raising the radiator above the ground increases the distance to the 
image and so is like spacing dipoles further apart, you get more gain 
from the separation, to a point.

When I was in HS, my dad and I put up our second vertical, 42 feet of 
pipe balanced (with guy wires) on top of a mast up from a 25 foot 
telephone pole. We probably had the feed point about 32 feet with four 
radials a quarter wave long for 40 meters. Fed with 75 ohm coax through 
a series capacitor it was a good match on 40m. That height gives a 
theoretical 75 ohm resistance when the reactance is tuned out with the 
series capacitor. We where using a Viking Ranger 160-10 meter 
transmitter at the time and I found the Ranger would tune it from the 
shack on all those bands. Contacts on 160 and ten were not too common or 
at great distances. 80 I don't remember. 40 was superb for DX, I could 
(with my 50 watts on CW) take mid Pacific stations from west coast 
stations. But it was poor for working within the US. I don't remember 
details of it on 20 and 15, but that was a half century ago. I do have 
my logs but I'm not going to dig through them now. After it fell down I 
don't recall we put up another vertical. I was given a vertical later, 
but I used it to support a short 80 meter inverted V that was better for 
the path I wanted to work then (Allen TX to St. Louis, MO).

It can be misleading to compare two antennas in the same space at the 
same time. The test antenna CAN couple to the reference antenna and take 
on some of the reference antenna's characteristics. Two verticals in the 
same yard would do that. The best example of that is the original 
exposition of the CCRR that was tested ON the ground plane of a quarter 
wave vertical. The high currents in the CCRR coupled to the vertical and 
made it radiate very nicely. When tested at microwave well isolated, its 
20 dB down. Another failed test was the crossed field antenna tested 
near a reference dipole. There the required feed lines also radiated to 
make it appear to work.

73, Jerry, K0CQ