Reg:

Here is the results of my analysis of a 99 radial monople:
Height 9 m, radial length from 0.5 to 10m, all conductors
#14 AWG copper, ground Er = 16, resistivity 150 ohm - m.
Radials 25 mm below ground. Antenna efficiency includes
the surface wave.

Radial Radial Radiation Ant
Length Z Resistance Efficiency
(m) (ohms) (ohms) (%)

0.5 63.0 - j 33.6 13.2 17.6
1.0 47.7 - j 18.2 13.2 21.7
1.5 41.8 - j 13.7 13.2 24.1
2.0 38.3 - j 12.5 13.2 25.7
2.5 35.7 - j 11.2 13.2 27.1
3.0 33.4 - j 10.8 13.2 28.3
3.5 31.6 - j 10.5 13.1 29.4
4.0 29.8 - j 10.1 13.1 30.6
4.5 28.2 - j 9.5 13.1 31.7
5.0 26.8 - j 8.9 13.1 32.8
5.5 25.7 - j 8.2 13.1 33.8
6.0 24.7 - j 8.0 13.2 34.7
6.5 23.9 - j 6.9 13.2 35.6
7.0 23.2 - j 6.2 13.3 36.4
7.5 22.6 - j 5.5 13.4 37.7
8.0 22.1 - j 4.8 13.5 38.0
8.5 21.6 - j 4.2 13.7 38.8
9.0 21.2 - j 3.5 13.8 39.5
9.5 20.9 - j 2.9 14.0 40.2
10.0 20.7 - j 2.2 14.2 40.9

Note that the radiation resistance is computed
from the total radiated power (including surface
wave) divided by the RMS base current squared.

The radial input impedance is derived from
the difference between the antenna input
impedance and the radiation resistance. A
fraction of an ohm can be attributed to the
copper losses in the monopole. Also
some of the imaginary part of the radial impedance
must be due, in part, to the input impedance
of the vertical section.

With 0.5 m radials the surface wave accounts for
2% of the total radiated power. With 10 m radials
the surface wave accounts for 5% of the TRP.


Frank
CM Reg's 99 radial Vertical

CM (WG)

CE

GW 2 1 0 0 0 0 0.0968 -0.025 0.00082

GW 35 4 0 0.0968 -0.025 0.026 0.5 -0.025 0.00082

GW 70 4 0 0.0968 -0.025 0 0.5 -0.025 0.00082

GW 105 4 0 0.0968 -0.025 -0.026 0.5 -0.025 0.00082

GR 1 33

GE -1 2

GN 2 0 0 0 16 0.0067

FR 0 1 0 0 8.07 0.01

LD 5 0 0 0 5.8001E7

WG

EN



CM Reg's 99 radial

CM (GF)

CE

GF

GW 1 90 0 0 9 0 0 0 0.00082

GE -1

EX 0 1 90 00 83.83328192 0

LD 5 0 0 0 5.8001E7

RP 1 101 1 0000 200 0 -2 1 200

RP 0 91 1 1000 0 0 1 1

RP 0 19 73 1002 -90 0 5.00000 5.00000

NE 1 1 46 1 200 45 90 1.0 1.0 1

NH 1 1 1 1 200 89 90 1.0 1.0 1

EN

I'm curious: How did you calculate the total radiated power including
surface wave, and how did you determine what fraction of the radiated
power is in the surface wave?

Roy Lewallen, W7EL

Frank's wrote:
. . .
Note that the radiation resistance is computed
from the total radiated power (including surface
wave) divided by the RMS base current squared.
. . .

Frank,

You don't mention frequency. I assume it is still 8.07 MHz.

There's something seriously wrong!

The only change you have made (or I think you have changed) is to
increase the number of radials from 36 to 99.

Yet, for a length of 10 metres, the resistance of the radials ground
connection has INCREASED from a few ohms (for 36 radials) to 20.7 ohms
(for 99 radials).

This is impossible! It should either decrease to an even lower value
or at least remain the same.

Although I am not particularly interested in radiation resistance,
there is also something seriously wrong with Rrad. Rrad for a
1/4-wave vertical ought to be in the region of 34 ohms - not as low
as 13 ohms.

I think you use Rrad to calculate radials input resistance in which I
AM very interested.

I think you subtract the antenna input impedance, from the total
impedance of antenna + radials, to obtain the radials input
resistance.

Rrad + conductor resistance is the feedpoint resistance of the
antenna. You make it about 34 - 13 = 21 ohms too low.

If you subtract 21 ohms from YOUR radials input resistance values,
then the EXPECTED very low input resistance values for 99 radials are
obtained. But, of course, the radials input resistance should never
become negative.

If you are unable to find where the error arises then use my value of
Antenna Feedpoint Resistance = 33.8 ohms (which I have just
calculated.)

Could you please investigate your results and apply corrections? If
you are unable to determine the resonant input resistance of the
9-metre vertical antenna ( jX = 0) then use my value of 33.8 ohms
which, as likely as not, will not be exactly correct.

.................................................. .....................
.................................

Then change antenna height to exactly 3 metres and change frequency to
about 25 MHz. The exact frequency being that at which the antenna is
1/4-wave resonant with the feedpoint reactance being zero. Repeat
measurements for 99 radials.

Such measurements will be far more accurate than if they were made in
the field.

I have some nice graphs of input impedance, R + jX, versus radial
length for work you have already done. They tell me quite a lot.
----
Reg.

=======================================
Reg:

Here is the results of my analysis of a 99 radial monople:
Height 9 m, radial length from 0.5 to 10m, all conductors
#14 AWG copper, ground Er = 16, resistivity 150 ohm - m.
Radials 25 mm below ground. Antenna efficiency includes
the surface wave.

Radial Radial Radiation Ant
Length Z Resistance Efficiency
(m) (ohms) (ohms) (%)

0.5 63.0 - j 33.6 13.2 17.6
1.0 47.7 - j 18.2 13.2 21.7
1.5 41.8 - j 13.7 13.2 24.1
2.0 38.3 - j 12.5 13.2 25.7
2.5 35.7 - j 11.2 13.2 27.1
3.0 33.4 - j 10.8 13.2 28.3
3.5 31.6 - j 10.5 13.1 29.4
4.0 29.8 - j 10.1 13.1 30.6
4.5 28.2 - j 9.5 13.1 31.7
5.0 26.8 - j 8.9 13.1 32.8
5.5 25.7 - j 8.2 13.1 33.8
6.0 24.7 - j 8.0 13.2 34.7
6.5 23.9 - j 6.9 13.2 35.6
7.0 23.2 - j 6.2 13.3 36.4
7.5 22.6 - j 5.5 13.4 37.7
8.0 22.1 - j 4.8 13.5 38.0
8.5 21.6 - j 4.2 13.7 38.8
9.0 21.2 - j 3.5 13.8 39.5
9.5 20.9 - j 2.9 14.0 40.2
10.0 20.7 - j 2.2 14.2 40.9

Note that the radiation resistance is computed
from the total radiated power (including surface
wave) divided by the RMS base current squared.

The radial input impedance is derived from
the difference between the antenna input
impedance and the radiation resistance. A
fraction of an ohm can be attributed to the
copper losses in the monopole. Also
some of the imaginary part of the radial impedance
must be due, in part, to the input impedance
of the vertical section.

With 0.5 m radials the surface wave accounts for
2% of the total radiated power. With 10 m radials
the surface wave accounts for 5% of the TRP.


Frank
CM Reg's 99 radial Vertical

CM (WG)

CE

GW 2 1 0 0 0 0 0.0968 -0.025 0.00082

GW 35 4 0 0.0968 -0.025 0.026 0.5 -0.025 0.00082

GW 70 4 0 0.0968 -0.025 0 0.5 -0.025 0.00082

GW 105 4 0 0.0968 -0.025 -0.026 0.5 -0.025 0.00082

GR 1 33

GE -1 2

GN 2 0 0 0 16 0.0067

FR 0 1 0 0 8.07 0.01

LD 5 0 0 0 5.8001E7

WG

EN



CM Reg's 99 radial

CM (GF)

CE

GF

GW 1 90 0 0 9 0 0 0 0.00082

GE -1

EX 0 1 90 00 83.83328192 0

LD 5 0 0 0 5.8001E7

RP 1 101 1 0000 200 0 -2 1 200

RP 0 91 1 1000 0 0 1 1

RP 0 19 73 1002 -90 0 5.00000 5.00000

NE 1 1 46 1 200 45 90 1.0 1.0 1

NH 1 1 1 1 200 89 90 1.0 1.0 1

EN

Roy, I use a couple of methods. First, the RP card as follows:

RP 0 19 73 1002 -90 0 5.0 5.0

Where, the last digit of "XNDA" is "2", and the average gain
is calculated providing the following output:

AVERAGE POWER GAIN= 3.00224E-01 SOLID ANGLE USED IN AVERAGING=( 2.0000)*PI
STERADIANS.

POWER RADIATED ASSUMING RADIATION INTO 4*PI STERADIANS = 1.15944E+01 WATTS

The input power for this particular test was 38.619 W.

As a verification I numerically integrate (Excel) the total radiated E-field
over a hemispherical region.
The elemental area I use is r^2*sine(theta)*d(theta)*d(phi). Of course
the E field is normalized to 1 m, and "r" is obviously "1". d(phi), and
d(theta) are both one degree.
The results are in very close agreement with the radiation, as above,
in 4*PI Steradians/2. I therefore have a fairly accurate figure for
the total radiated power without the surface wave.

For the next step I use the following RP card:
RP 1 101 1 0000 200 0 -2 1 200. The problem is, of course,
that the results are only available in cylindrical coordinates.
Where the total ground wave, and sky wave, is computed in the far field
(Where I have taken the far field to be 200 meters at 8 MHz)
from Z = 200 to Z = 0, in steps of 2 meters (From theta(zenith) 45 to
90). I then normalize these data to a radial distance of 1 m; taking only
those data points close to (theta) integral degrees. Picking out
these data points, taken from the NEC output file, in Excel is a
fairly tedious process. I then "cut and paste" these normalized
data into my "integrating" Excel spread sheet; replacing the
previously computed "sky wave" data from 45 to 90 degrees.
These results then give me the total radiated power, including
the surface wave. I can then easily compute the contribution,
to the total radiated power, by the surface wave.

When I replace the data, in the integrating spread sheet, the data
at 44 degrees is very close to the 45 degree field strength, obtained
from the cylindrical coordinates.

Hope you managed to follow my rambling description. To say the
least it requires a lot of tedious data manipulation with Excel.
If you are interested I can e-mail my spread sheets, NEC input
and output files, etc.

Regards,

Frank (VE6CB)


"Roy Lewallen" wrote in message
...
I'm curious: How did you calculate the total radiated power including
surface wave, and how did you determine what fraction of the radiated
power is in the surface wave?

Roy Lewallen, W7EL

Frank's wrote:
. . .
Note that the radiation resistance is computed
from the total radiated power (including surface
wave) divided by the RMS base current squared.
. . .

On Sat, 19 Aug 2006 00:52:40 GMT, "Frank's"
wrote:

....
Hope you managed to follow my rambling description. To say the
least it requires a lot of tedious data manipulation with Excel.
If you are interested I can e-mail my spread sheets, NEC input
and output files, etc.

Frank,

You might want to consider a scripting environment to build, run, and
summarise NEC runs.

Windows lacks a good scripting language for this type of application.
Fortunately there are others available, and at no charge. Two that I
have used are PERL and Python.

I prefer PERL mainly for historical reasons and the huge base of
packages to extend the PERL base, though Python is probably a better
designed language, and less likely to change radically (as PERL looks
likely to do with V6). Importantly, both natively support complex
number type. For instance, you can write in PERL:

#calculate Zo and gamma
my $a=$R+i*2*PI*$this-{f}*$L;
my $b=$G+i*2*PI*$this-{f}*$C;
$this-{zo}=($a/$b)**0.5;
$this-{gamma}=($a*$b)**0.5;

PERL is excellent for reading megabytes of NEC output and extracting
key figures for summarisation.

An example, I recently wanted to explore the relationship between
predicted beamwidth (in E and H planes separately) and gain for DL6WU
long boom Yagis. I built and ran literally hundreds of models with
pattern reporting at 0.1deg intervals, producing over 50MB of output.
Models were automatically generated from a PERL port of the DL6WU
design algorithms, and NEC runs of the generated models were
automated. Half power point was found by linear interpolation between
0.1deg points around the half power points. If I did that by hand, I
would be working for months, whereas a half hour of scripting produced
a solution that was more accurate and reusable.

Is there a place for Excel? Certainly, I usually create tab delimited
summary files with PERL scripts and drop them on Excel to the final
ad-hoc analysis and presentation. DPLOT is also a pretty neat tool for
graphical analysis and presentation. For example, the regression
analysis of the data from the study above
http://www.vk1od.net/dl6wu/new_pa2.gif was calculated and displayed
using DPLOT.

Worth the investment in learning PERL (or the like)!

http://www.perl.org/ (You need the Activestate PERL for Windows.)

http://www.python.org/

http://www.dlot.com/


Owen
--

On Sat, 19 Aug 2006 22:44:04 GMT, Owen Duffy wrote:
On Sat, 19 Aug 2006 00:52:40 GMT, "Frank's"
wrote:
...
Hope you managed to follow my rambling description. To say the
least it requires a lot of tedious data manipulation with Excel.
If you are interested I can e-mail my spread sheets, NEC input
and output files, etc.

You might want to consider a scripting environment to build, run, and
summarise NEC runs.

Windows lacks a good scripting language for this type of application.
Fortunately there are others available, and at no charge. Two that I
have used are PERL and Python.

And, there's REXX (several -- recommend Regina for one) which has
_great_ string handling capabilities.

Jonesy
--
Marvin L Jones | jonz | W3DHJ | linux
38.24N 104.55W | @ config.com | Jonesy | OS/2
*** Killfiling google posts: http//jonz.net/ng.htm

On 20 Aug 2006 00:52:41 GMT, Allodoxaphobia
wrote:


And, there's REXX (several -- recommend Regina for one) which has
_great_ string handling capabilities.

I haven't followed the fortunes of REXX since I taught some "Safe
REXX" courses in the ealy '90s.

I do recall that in the 80s, IBM wanted to hand it over to the ANSI to
be made a standard. That should have nobbled its progress sufficiently
to be overtaken by developments originating from the Unix world.

But yes, I see that enthusiasts have kept it alive. Does it have
(PC)REs, might be an option for people who don't want to embrace REs.

Owen
--

Reg,

The frequency is still 8.07 MHz. I was also puzzled by the radiation
resistance, but I arrived at the value differently than in previous models.

Anyway, I think I have discovered the error. In computing the surface
wave at 200 m I assumed the loss would be insignificant. This does not
appear to be true. For example; at a 26 degree elevation angle
normalization of the E-field to one meter produces essentially the same
result from either: 200 meters, or 400 meters. At zero degree elevation
angle, where the surface wave dominates; normalization from 200 m
or 400 m produces significantly different results -- as shown below:

Elevation Distance E- field
Angle (m) normalized
(deg.) to 1 meter
(V)

26 200 83.0
26 400 83.3
0 200 67.5
0 400 45.5

I had this at the back of my mind when making the calculations.
To be honest it is pretty much a no brainer, since
it is well known that the surface wave diminishes rapidly at
the higher frequencies.

All your comments are noted. For the moment I would like to perform an
integration of the Poynting Vector in the near field. Hopefully it will
provide a more realistic radiation resistance. Now to figure out
how to do this in Excel.

Frank





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

You don't mention frequency. I assume it is still 8.07 MHz.

There's something seriously wrong!

The only change you have made (or I think you have changed) is to
increase the number of radials from 36 to 99.

Yet, for a length of 10 metres, the resistance of the radials ground
connection has INCREASED from a few ohms (for 36 radials) to 20.7 ohms
(for 99 radials).

This is impossible! It should either decrease to an even lower value
or at least remain the same.

Although I am not particularly interested in radiation resistance,
there is also something seriously wrong with Rrad. Rrad for a
1/4-wave vertical ought to be in the region of 34 ohms - not as low
as 13 ohms.

I think you use Rrad to calculate radials input resistance in which I
AM very interested.

I think you subtract the antenna input impedance, from the total
impedance of antenna + radials, to obtain the radials input
resistance.

Rrad + conductor resistance is the feedpoint resistance of the
antenna. You make it about 34 - 13 = 21 ohms too low.

If you subtract 21 ohms from YOUR radials input resistance values,
then the EXPECTED very low input resistance values for 99 radials are
obtained. But, of course, the radials input resistance should never
become negative.

If you are unable to find where the error arises then use my value of
Antenna Feedpoint Resistance = 33.8 ohms (which I have just
calculated.)

Could you please investigate your results and apply corrections? If
you are unable to determine the resonant input resistance of the
9-metre vertical antenna ( jX = 0) then use my value of 33.8 ohms
which, as likely as not, will not be exactly correct.

.................................................. ....................
................................

Then change antenna height to exactly 3 metres and change frequency to
about 25 MHz. The exact frequency being that at which the antenna is
1/4-wave resonant with the feedpoint reactance being zero. Repeat
measurements for 99 radials.

Such measurements will be far more accurate than if they were made in
the field.

I have some nice graphs of input impedance, R + jX, versus radial
length for work you have already done. They tell me quite a lot.
----
Reg.

=======================================
Reg:

Here is the results of my analysis of a 99 radial monople:
Height 9 m, radial length from 0.5 to 10m, all conductors
#14 AWG copper, ground Er = 16, resistivity 150 ohm - m.
Radials 25 mm below ground. Antenna efficiency includes
the surface wave.

Radial Radial Radiation Ant
Length Z Resistance Efficiency
(m) (ohms) (ohms) (%)

0.5 63.0 - j 33.6 13.2 17.6
1.0 47.7 - j 18.2 13.2 21.7
1.5 41.8 - j 13.7 13.2 24.1
2.0 38.3 - j 12.5 13.2 25.7
2.5 35.7 - j 11.2 13.2 27.1
3.0 33.4 - j 10.8 13.2 28.3
3.5 31.6 - j 10.5 13.1 29.4
4.0 29.8 - j 10.1 13.1 30.6
4.5 28.2 - j 9.5 13.1 31.7
5.0 26.8 - j 8.9 13.1 32.8
5.5 25.7 - j 8.2 13.1 33.8
6.0 24.7 - j 8.0 13.2 34.7
6.5 23.9 - j 6.9 13.2 35.6
7.0 23.2 - j 6.2 13.3 36.4
7.5 22.6 - j 5.5 13.4 37.7
8.0 22.1 - j 4.8 13.5 38.0
8.5 21.6 - j 4.2 13.7 38.8
9.0 21.2 - j 3.5 13.8 39.5
9.5 20.9 - j 2.9 14.0 40.2
10.0 20.7 - j 2.2 14.2 40.9

Note that the radiation resistance is computed
from the total radiated power (including surface
wave) divided by the RMS base current squared.

The radial input impedance is derived from
the difference between the antenna input
impedance and the radiation resistance. A
fraction of an ohm can be attributed to the
copper losses in the monopole. Also
some of the imaginary part of the radial impedance
must be due, in part, to the input impedance
of the vertical section.

With 0.5 m radials the surface wave accounts for
2% of the total radiated power. With 10 m radials
the surface wave accounts for 5% of the TRP.


Frank
CM Reg's 99 radial Vertical

CM (WG)

CE

GW 2 1 0 0 0 0 0.0968 -0.025 0.00082

GW 35 4 0 0.0968 -0.025 0.026 0.5 -0.025 0.00082

GW 70 4 0 0.0968 -0.025 0 0.5 -0.025 0.00082

GW 105 4 0 0.0968 -0.025 -0.026 0.5 -0.025 0.00082

GR 1 33

GE -1 2

GN 2 0 0 0 16 0.0067

FR 0 1 0 0 8.07 0.01

LD 5 0 0 0 5.8001E7

WG

EN



CM Reg's 99 radial

CM (GF)

CE

GF

GW 1 90 0 0 9 0 0 0 0.00082

GE -1

EX 0 1 90 00 83.83328192 0

LD 5 0 0 0 5.8001E7

RP 1 101 1 0000 200 0 -2 1 200

RP 0 91 1 1000 0 0 1 1

RP 0 19 73 1002 -90 0 5.00000 5.00000

NE 1 1 46 1 200 45 90 1.0 1.0 1

NH 1 1 1 1 200 89 90 1.0 1.0 1

EN

You might want to consider a scripting environment to build, run, and
summarise NEC runs.

Windows lacks a good scripting language for this type of application.
Fortunately there are others available, and at no charge. Two that I
have used are PERL and Python.

I prefer PERL mainly for historical reasons and the huge base of
packages to extend the PERL base, though Python is probably a better
designed language, and less likely to change radically (as PERL looks
likely to do with V6). Importantly, both natively support complex
number type. For instance, you can write in PERL:

#calculate Zo and gamma
my $a=$R+i*2*PI*$this-{f}*$L;
my $b=$G+i*2*PI*$this-{f}*$C;
$this-{zo}=($a/$b)**0.5;
$this-{gamma}=($a*$b)**0.5;

PERL is excellent for reading megabytes of NEC output and extracting
key figures for summarisation.

An example, I recently wanted to explore the relationship between
predicted beamwidth (in E and H planes separately) and gain for DL6WU
long boom Yagis. I built and ran literally hundreds of models with
pattern reporting at 0.1deg intervals, producing over 50MB of output.
Models were automatically generated from a PERL port of the DL6WU
design algorithms, and NEC runs of the generated models were
automated. Half power point was found by linear interpolation between
0.1deg points around the half power points. If I did that by hand, I
would be working for months, whereas a half hour of scripting produced
a solution that was more accurate and reusable.

Is there a place for Excel? Certainly, I usually create tab delimited
summary files with PERL scripts and drop them on Excel to the final
ad-hoc analysis and presentation. DPLOT is also a pretty neat tool for
graphical analysis and presentation. For example, the regression
analysis of the data from the study above
http://www.vk1od.net/dl6wu/new_pa2.gif was calculated and displayed
using DPLOT.

Worth the investment in learning PERL (or the like)!

http://www.perl.org/ (You need the Activestate PERL for Windows.)

http://www.python.org/

http://www.dlot.com/


Owen

Thanks for the info Owen, also the comments from Jonsey. The
scripting languages look very interesting. From your sample
code above it does look very logical and easy to use; not
to mention that it can handle complex numbers.

Regards,

Frank

Worth the investment in learning PERL (or the like)!

http://www.perl.org/ (You need the Activestate PERL for Windows.)

http://www.python.org/

http://www.dlot.com/


Owen

Thanks for the info Owen, also the comments from Jonsey. The
scripting languages look very interesting. From your sample
code above it does look very logical and easy to use; not
to mention that it can handle complex numbers.

Still thinking about the scripting languages, but wanted to finish the
analysis
with Excel. While trying to set up the "Cross product matrix" I
ran into a major formatting problem with Excel. When I convert
the polar form of a complex number (two cells) to rectangular form
(single cell); the cell size expands to 42 digits in scientific format.
The problem lies in the fact that Excel treats complex numbers as
text, and they cannot be formatted in the normal way. I have
searched several bookstores, and looked at every major book
on Excel, and VBA macros. Complex numbers are treated in only
a superficial way, or not at all.

Microsoft does discuss the problem concerning XL2000 at:
http://support.microsoft.com/?kbid=213294
The VB code does not make a lot of sense to me, also in particular
the statement: "=FormatComplex(A1,"0.00","0.0000")" does not
appear to work, nor does their "Sample VBA Procedure". This may
be due to the fact that I am using Excel 97.

Anyway, this is not really the forum for this discussion, but others
attempting to deal with the NEC output file analysis, may have
some ideas. It looks like it may be some time before I can
arrive at any results -- probably with some scripting code.

Frank