Preliminary numbers from Frank's NEC-4 run on Reg's model below:

Caveat: I have not been able to ask Frank if the segments are all the same
length along the radial wire. The info below is based on that assumption.

The radial is 10 meters long, buried about 1 inch. I'm reading the numbers
from the graph that Frank sent me. The radial wire is 40 segments long or
..25 meters per segment. The antenna is 9 feet long and modeled at 8.07 mhz.

If I'm reading it right, at 30 segments along the radial wire, the current
has dropped from a peak of 0.6 amps to 0.2 amps. 30 segments seems to be 7.5
meters out. If the current is still 0.2 amps at 7.5 meters out on a 10 meter
radial, then Reg's approach fails. He indicated 20 dB down at a short
distance out. 75% of the way out on the 10 meter radial, the current is down
0.2/0.6 = .33. 10log * 0.33 = 4.8 dB (if I did that right).

So...it seems that the current along the radial is down only 4.8 dB at 75%
of it's length. Reg indicated that it should be down 20 dB at about 1/3 of
its length.

At the 35th segment of the radial, the current is 1/6th or 7.8 dB down. This
is at 90% of the radial's length.

At the 39th segment of the radio the current is .025 amps. 0.025/6 = .0146.
10log * .0146 = 14 dB down. That is only 14 dB down at 100% of the radial
length.

I'm using 10 log * (I1/I2) for for the dB calcs...I think current ratios and
power ratios are 10log, and voltage is 20log.

It is possible I'm interpreting Frank's graph incorrectly or applying the
attenuation that Reg refers to incorrectly. I'm just so glad to see some
numbers for current distribution along a radial wire from NEC-4, that I had
to post what I see.

Eyeballing it looks like thisthe radial wire starts at segment 39 and runs
to segment 79)

Segment 39 0.60 amps, distance from source = 0, dB = 0
Segment 49 0.54 amps, distance from source = 2.5 meters, dB = 0.46 dB
Segment 59 0.42 amps, distance from source = 5.0 meters, dB = 1.5 dB
Segment 69 0.22 amps, distance from source = 7.5 meters, dB = 4.3 dB
Segment 79 0.025 amps, distance from source = 10 meters, dB = 14.8 dB

What does Reg's program predict for dB down on this sample antenna?

Using 25 and 25 for soil and the info Frank gave me:

Reg's program shows radial attenuation of 20 dB at 2.3 meters from the
source.

Side by side with the NEC-4 data

Distance Reg NEC-4 (dB down)

2.5 m 21.2 0.46

5.0 m 42.4 1.5

7.5 m 63.9 4.3

10 m 83.3 14.8


These numbers are so far apart, it looks like I did something terribly
wrong. Someone please correct me.
Keep in mind these are preliminary attempts to analyze the NEC-4 based graph
that Frank sent me. I really do hope I did something wrong.

....hasan, N0AN

"Reg Edwards" wrote in message
...
Frank,

Just to confirm we are both working on the same system, I have -

Number of radials = 36
Length of radials = 10 m
Diameter of radials = 2 mm
Frequency = 7 MHz
Antenna height = 9 m
Antenna diameter = 1.64 mm = 14 AWG
Ground resistivity = 150 ohm-metres
Ground permittivity = 16

IMPORTANT:

If NEC4 gives you the input impedance of the radial system I should be
very pleased to know what it is.

Otherwise we shall have no idea where the discrepancy arises - in the
radial system or in the antenna efficiency calculation.

Radiating efficiency is estimated by my program by the well-known
formula -

Efficiency = Rrad / ( Rrad + Rradials )

provided antenna and radials reactance are tuned out.

Whereas NEC4 calculates efficiency by integrating power flow over a
hemisphere WITHOUT tuning out antenna and radials reactance.
Altogether different.
----
Reg, G4FGQ

CAUTION CAUTION CAUTION:

The wire segments are NOT equal in this model. Frank is sending me a new one
with linear segments. I'll correct the errors below as soon as I get the new
values.

....hasan, N0AN

"hasan schiers" wrote in message
...
Preliminary numbers from Frank's NEC-4 run on Reg's model below:

Caveat: I have not been able to ask Frank if the segments are all the same
length along the radial wire. The info below is based on that assumption.

The radial is 10 meters long, buried about 1 inch. I'm reading the numbers
from the graph that Frank sent me. The radial wire is 40 segments long or
.25 meters per segment. The antenna is 9 feet long and modeled at 8.07
mhz.

If I'm reading it right, at 30 segments along the radial wire, the current
has dropped from a peak of 0.6 amps to 0.2 amps. 30 segments seems to be
7.5 meters out. If the current is still 0.2 amps at 7.5 meters out on a 10
meter radial, then Reg's approach fails. He indicated 20 dB down at a
short distance out. 75% of the way out on the 10 meter radial, the current
is down 0.2/0.6 = .33. 10log * 0.33 = 4.8 dB (if I did that right).

So...it seems that the current along the radial is down only 4.8 dB at 75%
of it's length. Reg indicated that it should be down 20 dB at about 1/3 of
its length.

At the 35th segment of the radial, the current is 1/6th or 7.8 dB down.
This is at 90% of the radial's length.

At the 39th segment of the radio the current is .025 amps. 0.025/6 =
.0146. 10log * .0146 = 14 dB down. That is only 14 dB down at 100% of the
radial length.

I'm using 10 log * (I1/I2) for for the dB calcs...I think current ratios
and power ratios are 10log, and voltage is 20log.

It is possible I'm interpreting Frank's graph incorrectly or applying the
attenuation that Reg refers to incorrectly. I'm just so glad to see some
numbers for current distribution along a radial wire from NEC-4, that I
had to post what I see.

Eyeballing it looks like thisthe radial wire starts at segment 39 and
runs to segment 79)

Segment 39 0.60 amps, distance from source = 0, dB = 0
Segment 49 0.54 amps, distance from source = 2.5 meters, dB = 0.46 dB
Segment 59 0.42 amps, distance from source = 5.0 meters, dB = 1.5 dB
Segment 69 0.22 amps, distance from source = 7.5 meters, dB = 4.3 dB
Segment 79 0.025 amps, distance from source = 10 meters, dB = 14.8 dB

What does Reg's program predict for dB down on this sample antenna?

Using 25 and 25 for soil and the info Frank gave me:

Reg's program shows radial attenuation of 20 dB at 2.3 meters from the
source.

Side by side with the NEC-4 data

Distance Reg NEC-4 (dB down)

2.5 m 21.2 0.46

5.0 m 42.4 1.5

7.5 m 63.9 4.3

10 m 83.3 14.8


These numbers are so far apart, it looks like I did something terribly
wrong. Someone please correct me.
Keep in mind these are preliminary attempts to analyze the NEC-4 based
graph that Frank sent me. I really do hope I did something wrong.

...hasan, N0AN

"Reg Edwards" wrote in message
...
Frank,

Just to confirm we are both working on the same system, I have -

Number of radials = 36
Length of radials = 10 m
Diameter of radials = 2 mm
Frequency = 7 MHz
Antenna height = 9 m
Antenna diameter = 1.64 mm = 14 AWG
Ground resistivity = 150 ohm-metres
Ground permittivity = 16

IMPORTANT:

If NEC4 gives you the input impedance of the radial system I should be
very pleased to know what it is.

Otherwise we shall have no idea where the discrepancy arises - in the
radial system or in the antenna efficiency calculation.

Radiating efficiency is estimated by my program by the well-known
formula -

Efficiency = Rrad / ( Rrad + Rradials )

provided antenna and radials reactance are tuned out.

Whereas NEC4 calculates efficiency by integrating power flow over a
hemisphere WITHOUT tuning out antenna and radials reactance.
Altogether different.
----
Reg, G4FGQ

Corrected numbers for linearly segemented radials from Frank's latest NEC-4
model of one buried radial wire, compared to Reg's program.

Side by side with the NEC-4 data


This is how many dB down the current is as you move outward from the origin
of the radial.

Distance Reg NEC-4 (dB down)

1.0 m 2.5 1.3

3.5 m 8.7 4.4

5.9 m 14.9 8.7

8.5 m 21.4 10.0

9.7 m 24.4 23.8

Conclusion: the current drop along the radial is no where near as fast as
Radials3 predicts, therefore shortening the radials as much as the program
shows will increase losses significantly.

I find it VERY interesting, that at the full length of 10m, there is good
agreement between Reg's program and NEC-4.

If I were going to base my conclusions on this preliminary small sample, I
would say that Reg's program does not hold up for short radials. BL&E, W8JI
and now NEC-4 all indicate that there is no where near 20 dB of attenuation
in short radials. To confirm this isn't an odd case, a lot more runs would
need to be done with varying lengths and radial numbers...but I have to say,
it ain't lookin' good for Radials3 in terms of fairly representing the
rapidity with which currents diminish on a radial wire over its length.

Bottom Line:

For the present, the articles in QST, ARRL Handbook, Low-Band DX'ing and
W8JI's findings are the ones I would follow. The first three are all the
same study and that formula is based on BL&E. The following data are from a
spreadsheet I used to calculate the optimum length and number of radials
based on the above sources. I put the BL&E data in the spreadsheet as a
reference. The numbers are how many dB down the field strength was for a
given number and length of radials.

Brn/Lw/Ep



# Rad 0.137 wl 0.274 wl 0.411 wl
2 -4.36 -4.36 -4.05
15 -2.40 -1.93 -1.65
30 -2.40 -1.44 -0.97
60 -2.00 -0.66 -0.42
113 -2.00 -0.51 0 (Ref)


Here are a few runs for 80 meters of various numbers and lengths of radials
that should be within a dB or so of optimum (BL&E).(Based on the references
noted above) 3.7 mhz, 1/4 wave vertical. The formula is based on tip
separation at the perimeter. Too much separation increases loss, too little
wastes wire. All based on wavelenthgs, of course. I believe the maximum tip
separation recommended was .015 wavelength.

Available Wire # of Radials Length of Radials
500' 25 19.7' (not within a dB, not
enough wire)
1000' 36 27.8' (not within a dB, not
enough wire)
1500' 44 34.0'
2000' 51 39.3'
3116' 63 49.0 (should be within 0.5 dB
of BL&E Optimum)

My final setup will be 46 radials 50' long. I have 26 right now. It looks
like for 50' long radials, I should really have 63 of them, otherwise, I
could have stayed at 51 radials only 39.3' long. All this says is that I'm
not making the "most" out of the available wire I had. (which makes sense,
given I've added radials over time, and didn't have a final plan).

At this point, it looks like when copper prices drop, I need to get another
850' of wire and put in 17 more radials and I will have met the criteria for
the formula. (Be within 0.5 dB of maximum field strength according to BL*E).

If anyone wants a copy of the Excel spreadsheet, just email me and I'll send
it to you as an attachment.
Only two variables should be entered: Total length of available wire and
Frequency in Mhz. Everything else is calcuated.( I did not protect any of
the fields, so if you enter data into a calculated field, you'll have to
reload your spreadsheet from a non-messed-up one...so save a virgin copy
somewhere until you protect the appropriate cells.)

73,

....hasan, N0AN

Fellow Experimenters, Frank and Hasan.

I havn't the foggiest idea what you are doing with NEC4 but you should
be aware that, according to Radial_3, there are 3 resonant frequencies
with a single radial at lengths shorter than 10 metres and at a
frequency of 7 MHz.

The propagation velocity is very low. VF = 0.225

Funny unexpected things happen on multi-resonant lines especially when
Zo has a relatively large positive angle. Before you draw any
conclusions about deducing attenuation from your output data you
should take into account the line is -

1/4-wave resonant at 2.4 metres.
1/2-wave resonant at 4.8 metres.
3/4-wave resonant at 7.4 metres.

and at 10 metres it is very near to full-wave resonance. It can be
assumed the far end is open-circuit. Actually it isn't. It behaves
as if it is slightly longer.

It is significant that at 10 metres and 7 MHz, you have concluded that
the radial is about 20dB long. Which approximately agrees with my
program as being the length beyond which there is not much point in
extending it.

But the best way of determining attenuation is to do what I have
suggested - increase radial length in short increments and observe
what happens to radial input impedance. Eventually, Zin will converge
on Zo if it hasn't already done so. I should very much like to know
what Zo is and at what length it occurs. I have to assume NEC4 knows
what it's doing! ;o)
----
Reg.

Fellow Experimenters, Frank and Hasan.

I havn't the foggiest idea what you are doing with NEC4 but you should
be aware that, according to Radial_3, there are 3 resonant frequencies
with a single radial at lengths shorter than 10 metres and at a
frequency of 7 MHz.

The propagation velocity is very low. VF = 0.225

Funny unexpected things happen on multi-resonant lines especially when
Zo has a relatively large positive angle. Before you draw any
conclusions about deducing attenuation from your output data you
should take into account the line is -

1/4-wave resonant at 2.4 metres.
1/2-wave resonant at 4.8 metres.
3/4-wave resonant at 7.4 metres.

and at 10 metres it is very near to full-wave resonance. It can be
assumed the far end is open-circuit. Actually it isn't. It behaves
as if it is slightly longer.

It is significant that at 10 metres and 7 MHz, you have concluded that
the radial is about 20dB long. Which approximately agrees with my
program as being the length beyond which there is not much point in
extending it.

But the best way of determining attenuation is to do what I have
suggested - increase radial length in short increments and observe
what happens to radial input impedance. Eventually, Zin will converge
on Zo if it hasn't already done so. I should very much like to know
what Zo is and at what length it occurs. I have to assume NEC4 knows
what it's doing! ;o)
----
Reg.

Reg, Note that I am making all my calculations at 8.07 MHz, where
the structure is very close to resonance. NEC does indicate
the resonant lengths of the radials as follows:

1/4 wave = 2 m;
1/2 wave = 4 m......etc., to 1.25 wavelengths at 10 m.

I will try your suggestion of gradually increasing radial length
until I see a convergence trend at the complex Zo.

Frank

Frank,

After correcting the resonant lengths for the change in frequency from
7 to 8.07 MHz there is almost exact agreement between NEC4 and
Radial_3.

Keep a record of length, Rin, jXin for each incremental increase in
length. They could be useful.

When using Radial_3 set the number of radials to 1. The input
impedance of the radial system will then be same as the input to the
single radial and will be displayed with a greater number of
significant figures.

When you compare results between the two programs set the frequency on
Radial_3 also to 8.07 MHz. The resonant frequency of the Radial_3
antenna is slightly higher - it's something to do with the end-effect
and the fact that a vertical antenna needs pruning by a few percent to
make it resonate at the theoretical value of 75/Height MHz. Since at
present we are concerned only with the radials it is better to use the
same frequency for both programs.

I think that will complete all I have in mind. It may involve you
with a tedious amount of work. If you find it interesting you could
do something similar at 28 MHz. The 20dB limit may be reached with an
even shorter length of radial. Radial resonances ( which BL&E never
dreamed of ) will be much more pronounced especially with high ground
resistivity.

Has anybody ever generated an input table for 120 radials?

It's all in the cause of Science. Thanking you in advance.
----
Reg.

Reg Edwards wrote:
Frank,

After correcting the resonant lengths for the change in frequency from
7 to 8.07 MHz there is almost exact agreement between NEC4 and
Radial_3.

Keep a record of length, Rin, jXin for each incremental increase in
length. They could be useful.

When using Radial_3 set the number of radials to 1. The input
impedance of the radial system will then be same as the input to the
single radial and will be displayed with a greater number of
significant figures.
(snip)

Does Radial_3 assume that each radial is independent of its neighbors,
regardless of how close or far?

"John Popelish" wrote
Does Radial_3 assume that each radial is independent of its
neighbors,
regardless of how close or far?
=======================================

No John. The input impedance of a set of radials is not the sum of
the individuals all in parallel. Input impedance is a non-linear
function of N, the number of radials.
----
Reg.

"Reg Edwards" wrote in message
...
Frank,

After correcting the resonant lengths for the change in frequency from
7 to 8.07 MHz there is almost exact agreement between NEC4 and
Radial_3.

Keep a record of length, Rin, jXin for each incremental increase in
length. They could be useful.

When using Radial_3 set the number of radials to 1. The input
impedance of the radial system will then be same as the input to the
single radial and will be displayed with a greater number of
significant figures.

When you compare results between the two programs set the frequency on
Radial_3 also to 8.07 MHz. The resonant frequency of the Radial_3
antenna is slightly higher - it's something to do with the end-effect
and the fact that a vertical antenna needs pruning by a few percent to
make it resonate at the theoretical value of 75/Height MHz. Since at
present we are concerned only with the radials it is better to use the
same frequency for both programs.

I think that will complete all I have in mind. It may involve you
with a tedious amount of work. If you find it interesting you could
do something similar at 28 MHz. The 20dB limit may be reached with an
even shorter length of radial. Radial resonances ( which BL&E never
dreamed of ) will be much more pronounced especially with high ground
resistivity.

Has anybody ever generated an input table for 120 radials?

It's all in the cause of Science. Thanking you in advance.
----
Reg.

I find this very interesting Reg, and learn more about NEC
all the time. I think I could probably do a 120 radial model;
particularly with segment length tapering. With linear
segmentation, and 0.025 m segment length, the 36 radial
model has almost 15,000 segments. It seems that
radial segments can be sloped to their final depth, so
can probably reduce the segmentation requirement.

Frank

Reg Edwards wrote:

[snip]

. . . Radial resonances ( which BL&E never
dreamed of ) will be much more pronounced especially with high ground
resistivity.

Reg,

I have no idea what BL&E might have dreamed of, but I did find one
curious item on the fifth page of their paper (page 757 in the original).



Where there are radial ground wires present, the earth current consists
of two components, part of which flows in the earth itself and the
remainder of which flows in the buried wires. As the current flows in
toward the antenna, it is continually added to by more displacement
currents flowing into the earth. It is not necessarily true that the
earth currents will increase because of this additional displacement
current, since all the various components differ in phase.



Whether this is "resonance" I cannot say. However, it is pretty clear
they understood that the radial currents did not monotonically increase
as the distance from the antenna decreased. There was some sort of
variation.

Their figure 42 seems to show significant "resonance", but there does
not appear to be any discussion of that behavior.

73,
Gene
W4SZ