On Sat, 08 Oct 2005 16:45:24 GMT, "Frank"
wrote:

A nominal 2.5 kHz shift, at 3.575 MHz (0.07%), is about all I can detect.

Hi Frank,

Is this shift constant over all axial lengths, or variable?

73's
Richard Clark, KB7QHC

I just took the ARRL antenna modeling course and bought EZNEC+ V4.0 to
provide the tools to do the examples. One chapter was devoted to
limitations and common shortcomings of the software. I think the
course is great and I would select the same software if I was doing it
again. I wonder if you are referring to the free demo package?

On Fri, 7 Oct 2005 20:38:43 -0700, "RST Engineering"
wrote:

In my humble opinion, EZNEC is a POS for anything other than a simple "wire
antenna". Try to do a true ground plane with bent radials? Try to do a
simple patch? Try to do a J? Try to do ANY configuration other than a wire
dipole and the sucker chokes.

Jim



"David" wrote in message
...
Does anyone know how I can model a coaxial sleeve antenna
on EZNEC ?

A nominal 2.5 kHz shift, at 3.575 MHz (0.07%), is about all I can detect.

Hi Frank,

Is this shift constant over all axial lengths, or variable?

73's
Richard Clark, KB7QHC

Hi Richard,

Axial extension 1.2" from antenna end -- Results as follows:

Extension Length (ft) Resonant Freq. (MHz)

0 3.5765
5 3.576
10 3.575
20 3.574
50 3.5725
100 3.5720

For a total shift of 4.5 kHz (0.125%) from zero to 100 ft

73,

Frank

On Sat, 08 Oct 2005 06:58:43 -0700, Wes Stewart
wrote:

On Sat, 08 Oct 2005 13:36:37 GMT, "Frank"
wrote:

"David" wrote in message
...
Does anyone know how I can model a coaxial sleeve antenna
on EZNEC ?

I know how to enter a standard Vertical 1/2 wave but not sure how
to represent the coax going up through the tubing(sleeve) in the lower
portion.

Thanks in advance.

Just model a vertical dipole. The presence of a coaxial cable, within the
lower conductor, will not effect its performance.

But the presence of the coax below the antenna sure will.

Let me expand on this. If---big if---the sleeve is the correct length
and lossless, so that it acts as a quarter-wave choke, then in theory
it does decouple the coax from the radiator.

A few points however:

1. The Vp of the choke will depend on the effective dielectric
constant of the air/outer jacket combination. The choke so formed
will likely be fairly lossy as the outer jacket is probably not the
lowest loss dielectric around.

2. Since the choke Vp will be lower than an air-dielectric one, it
must be physically shortened to optimize the choking function. This
combined with the usual diameter differences between the sleeve and
the upper radiator (rod) makes the vertical dipole asymmetric, i.e.
unbalanced, when the rod length is adjusted for resonance. In effect
the feedpoint is moving up and down as the rod length is adjusted. In
practice, this is no big deal but offered for completeness.

3. You can model this antenna to some extent. Below are some
coordinates for a model I used. The wire size was #12, all lengths in
inches and the frequency was 100 MHz. I used Multinec invoking EZNEC 4
for the calculations. In Multinec I made the Z a variable and could
vary the height above ground programmatically. The sleeve is
represented by four stubs; more would perhaps more accurately
represent a cylinder, but I don't believe it's necessary. Since the
"coax" connects to ground, MiniNEC ground (avg) was used. To observe
NEC limitations, all segments are (very nearly) the same length and
adjacent segments are aligned. The sleeve length was adjusted to
minimize the current on the "coax" below the antenna (wire #4) and the
rod length adjusted for zero reactance.


End 1 End 2
X Y Z X Y Z Dia Segs.

Source wire
0.00 0.00 354.09 0.00 0.00 355.09 #12 1

Upper radiator (rod)
0.00 0.00 355.09 0.00 0.00 382.47 #12 29

"Coax" inside sleeve
0.00 0.00 326.84 0.00 0.00 354.09 #12 29

"Coax" below sleeve
0.00 0.00 0.00 0.00 0.00 326.84 #12 337

Top of sleeve
0.00 0.00 354.09 0.00 2.00 354.09 #12 2
0.00 0.00 354.09 0.00 -2.00 354.09 #12 2
0.00 0.00 354.09 2.00 0.00 354.09 #12 2
0.00 0.00 354.09 -2.00 0.00 354.09 #12 2

Sides of sleeve
0.00 2.00 326.84 0.00 2.00 354.09 #12 29
0.00 -2.00 326.84 0.00 -2.00 354.09 #12 29
2.00 0.00 326.84 2.00 0.00 354.09 #12 29
-2.00 0.00 326.84 -2.00 0.00 354.09 #12 29



4. If SWR is your criteria for "goodness", you will be in for a big
surprise. The above example places the midpoint of the antenna at 3
wavelengths above ground. The feedpoint Z is a nice 56.6 +j0 but the
elevation for maximum radiation is 71 degrees above the horizon. If
satellite work is your goal, this is your antenna.

5. Despite the "choke", the transmission line is part of the antenna.
Skeptics of the model above can remove the sleeve and put a trap at
end 2 of the coax and see similar results.

On Sun, 09 Oct 2005 03:39:32 GMT, "Frank"
wrote:

Axial extension 1.2" from antenna end -- Results as follows:

Extension Length (ft) Resonant Freq. (MHz)

0 3.5765
5 3.576
10 3.575
20 3.574
50 3.5725
100 3.5720

For a total shift of 4.5 kHz (0.125%) from zero to 100 ft

Hi Frank,

Thanx.

Was this for a coaxial line running from the drivepoint, then parallel
to the lower leg, down? Was the parallel separation 1.2" as you
describe above?

Was this line modeled as a third wire connected at the drivepoint and
dropping as I describe?

73's
Richard Clark, KB7QHC

"Richard Clark" wrote in message
...
On Sun, 09 Oct 2005 03:39:32 GMT, "Frank"
wrote:

Axial extension 1.2" from antenna end -- Results as follows:

Extension Length (ft) Resonant Freq. (MHz)

0 3.5765
5 3.576
10 3.575
20 3.574
50 3.5725
100 3.5720

For a total shift of 4.5 kHz (0.125%) from zero to 100 ft

Hi Frank,

Thanx.

Was this for a coaxial line running from the drivepoint, then parallel
to the lower leg, down? Was the parallel separation 1.2" as you
describe above?

Was this line modeled as a third wire connected at the drivepoint and
dropping as I describe?

73's
Richard Clark, KB7QHC

Hi Richard,

My model is only valid providing the coax has no current on the exterior of
the shield. All I did was extend one leg of the dipole by the length
specified, but not actually connected to it; with a 1.2" gap. As per the
following sketch.

----------o---------- ----------
dipole extension

Do you think it makes sense? It is the only way I could think of modeling
it with NEC.

73,

Frank

On Sun, 09 Oct 2005 09:17:31 -0700, I wrote a fairly detailed
analysis:

[snip]

Apparently, I wasted my time.

On Mon, 10 Oct 2005 09:54:51 -0700, Wes Stewart
wrote:

On Sun, 09 Oct 2005 09:17:31 -0700, I wrote a fairly detailed
analysis:

[snip]

Apparently, I wasted my time.

Hi Wes,

I wouldn't say so. I took the time to enter your data into EZNEC and
confirm your results. I haven't closely examined the implications of
construction variations for the skirt wires, much less the "choking"
action that is notably missing.

73's
Richard Clark, KB7QHC

On Mon, 10 Oct 2005 16:05:16 GMT, "Frank"
wrote:

My model is only valid providing the coax has no current on the exterior of
the shield. All I did was extend one leg of the dipole by the length
specified, but not actually connected to it; with a 1.2" gap. As per the
following sketch.

----------o---------- ----------
dipole extension

Do you think it makes sense? It is the only way I could think of modeling
it with NEC.

Hi Frank,

No. The extension lies in the null of the dipole. As such, it would
be no mystery that it has so little influence.

To model it correctly requires some form of cage or skirt of wires as
Wes has provided in this thread.

73's
Richard Clark, KB7QHC

I have tried entering the model shown in EZNEC 4

The "View" of the antenna looks fine but when I try to run the
simulation it reports that it cannot because the gain is negative.

I then changed from azimuth to elevation simulation but this time it
reported the maxim 500 segments were exceeded.

I then reduce the segments slightly but now the program reports Runtime
error M6201: Math Sqrt Domain error.

ie. Cannot get the model to work using EZNEC. I'll see if it will run in
MNANA.

Richard Clark wrote:
On Mon, 10 Oct 2005 16:05:16 GMT, "Frank"
wrote:


My model is only valid providing the coax has no current on the exterior of
the shield. All I did was extend one leg of the dipole by the length
specified, but not actually connected to it; with a 1.2" gap. As per the
following sketch.

----------o---------- ----------
dipole extension

Do you think it makes sense? It is the only way I could think of modeling
it with NEC.


Hi Frank,

No. The extension lies in the null of the dipole. As such, it would
be no mystery that it has so little influence.

To model it correctly requires some form of cage or skirt of wires as
Wes has provided in this thread.

73's
Richard Clark, KB7QHC