Cecil, Using a 1/4 wave open stub.

Cecil Moore wrote:
David wrote:

I have constructed a 915 MHz 1/2 wave, end fed antenna that fits onto
the end of a SMA plug based on help from you guys.


How are you matching the very high feedpoint impedance?

I have some pics of the construction if anyone is interested in taking
a look.
Where is the appropriate place to post these jpg files ?


Some of us have web pages from qsl.net for that purpose.
There is also a netnews group for that purpose. I think
it is alt.binaries.

Roy,

I am using your EZNEC V4.019. Would you happen to have the ez
description file for this antenna so I can see how it is modeled ?

I had a "play" and set up 2 vertical have wave lengths and set the
source at the centre I then placed a 1/4 wave wire near the lower
1/4 wave and connect them at the centre point with a horizontal wire but
the program kept complaining that the wires were overlapping etc.

I also tried modeling an open stub J-Pole and seemed to be able to get
some plots that look ok but SWR was 100 across the band. Not sure what
I did wrong here but I assume as I get more familiar with modeling
techniques, I will get more meaningful results.

Thanks

Regards

David

Roy Lewallen wrote:
Jerry Martes wrote:


I'd have expected the "gain" to be closer to 4 1/2 db over the 1/4
wave stub over a ground. is it easy to show where i've missed something?


I think it should be more like 3 dB, but hadn't said anything until I
had a chance to model it. The quarter wave stub doesn't radiate
significantly, so it can be ignored. A half wavelength element should
have about 1.5 dB gain over a quarter wave. Two of them would give
another 3 dB if it weren't for mutual coupling, but the mutual coupling
of collinear elements reduces the gain to about 1.5 dB over a single
element. (See for example Fig. 39 on p. 8-35 of the ARRL Antenna Book,
20th Edition; look up Collinear, Gain and directivity in the index of
other editions; or model it with your favorite program.)

Roy Lewallen, W7EL

David wrote:
Roy,

I am using your EZNEC V4.019. Would you happen to have the ez
description file for this antenna so I can see how it is modeled ?

Sorry, I haven't had time to model it. But I dusted off a very old,
originally ELNEC, model of a 3 element collinear, and uploaded it to my
web site. It should give you the general idea. It's Collinear3.EZ at
http://eznec.com/misc/.

. . .

Roy Lewallen, W7EL

David wrote:
I also tried modeling an open stub J-Pole and seemed to be able to get
some plots that look ok but SWR was 100 across the band. Not sure what
I did wrong here but I assume as I get more familiar with modeling
techniques, I will get more meaningful results.

The open stub J-Pole (Zepp) is modeled as a shorted stub
J-Pole with the source in the middle of the shorted segment.
Mine is for 53.2 MHz. On edge it looks like this:

13.5 ft.
+---------------------------------------------------
|
(S)
|
+-----------------
4.5 ft.

The sections are 0.1 ft. apart. The resonant frequency is
53.2 MHz and the feedpoint impedance is 25 ohms for an SWR
of 2:1. The feedpoint is 30 ft. high over high-accuracy
ground. The gain over ground is 6.2 dBi, omnidirectional.
Free space gain is 3.4 dBi.

Anyone who wants a copy of this 6m J-Pole EZNEC file, please
send me an email to .
--
73, Cecil http://www.qsl.net/w5dxp

----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

On Sun, 18 Sep 2005 07:14:39 -0500, Cecil Moore
wrote:


The open stub J-Pole (Zepp) is modeled as a shorted stub
J-Pole with the source in the middle of the shorted segment.
Mine is for 53.2 MHz. On edge it looks like this:

13.5 ft.
+---------------------------------------------------
|
(S)
|
+-----------------
4.5 ft.

The sections are 0.1 ft. apart. The resonant frequency is
53.2 MHz and the feedpoint impedance is 25 ohms for an SWR
of 2:1. The feedpoint is 30 ft. high over high-accuracy
ground. The gain over ground is 6.2 dBi, omnidirectional.
Free space gain is 3.4 dBi.

You can vary the feedpoint impedance by changing (S).

73,
Danny, K6MHE

email: k6mheatarrldotnet
http://users.adelphia.net/~k6mhe/

Dan Richardson wrote:
You can vary the feedpoint impedance by changing (S).

Moving the elements to one foot apart increases the
feedpoint resistance to about 50 ohms at resonance
but it decreases the gain and distorts the radiation
pattern since one foot is an appreciable distance
on 6 meters.
--
73, Cecil http://www.qsl.net/w5dxp

----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

Thanks for the model, I'll try it.

What is the diameter of the elements. I understand you can also vary the
feedpoint impedance by changing element diameter ? (Zo = 276 log(2S/d)

Regards
David

Dan Richardson wrote:
On Sun, 18 Sep 2005 07:14:39 -0500, Cecil Moore
wrote:



The open stub J-Pole (Zepp) is modeled as a shorted stub
J-Pole with the source in the middle of the shorted segment.
Mine is for 53.2 MHz. On edge it looks like this:

13.5 ft.
+---------------------------------------------------
|
(S)
|
+-----------------
4.5 ft.

The sections are 0.1 ft. apart. The resonant frequency is
53.2 MHz and the feedpoint impedance is 25 ohms for an SWR
of 2:1. The feedpoint is 30 ft. high over high-accuracy
ground. The gain over ground is 6.2 dBi, omnidirectional.
Free space gain is 3.4 dBi.


You can vary the feedpoint impedance by changing (S).

73,
Danny, K6MHE

email: k6mheatarrldotnet
http://users.adelphia.net/~k6mhe/

That model seems to work fine in EZNEC.

I tried a 918 MHz version with long length = 219mm, short length = 72mm,
24mm apart, 1.5mm solid aluminum construction.

The results show excellent SWR and feedpoint impedance close to 50
Ohms.(50.29 -j0.0148), SWR = 1.006
The unit is not so omnidirectional showing a front/back ratio of
5.79dB.(Beamwidth 214.4 Degrees).

The maximum gain was 4.74dB(ref) at 4 degrees (elevation). dB ref =
2.14dBi so I suppose this is saying gain is 4.74dBd.



Dan Richardson wrote:
On Sun, 18 Sep 2005 07:14:39 -0500, Cecil Moore
wrote:



The open stub J-Pole (Zepp) is modeled as a shorted stub
J-Pole with the source in the middle of the shorted segment.
Mine is for 53.2 MHz. On edge it looks like this:

13.5 ft.
+---------------------------------------------------
|
(S)
|
+-----------------
4.5 ft.

The sections are 0.1 ft. apart. The resonant frequency is
53.2 MHz and the feedpoint impedance is 25 ohms for an SWR
of 2:1. The feedpoint is 30 ft. high over high-accuracy
ground. The gain over ground is 6.2 dBi, omnidirectional.
Free space gain is 3.4 dBi.


You can vary the feedpoint impedance by changing (S).

73,
Danny, K6MHE

email: k6mheatarrldotnet
http://users.adelphia.net/~k6mhe/

David wrote:
That model seems to work fine in EZNEC.

I tried a 918 MHz version with long length = 219mm, short length = 72mm,
24mm apart, 1.5mm solid aluminum construction.

The results show excellent SWR and feedpoint impedance close to 50
Ohms.(50.29 -j0.0148), SWR = 1.006
The unit is not so omnidirectional showing a front/back ratio of
5.79dB.(Beamwidth 214.4 Degrees).

That's pretty strange. The only ready explanation is radiation from the
transmission line section. Apparently it's significant with the 0.074
wavelength conductor spacing.

The maximum gain was 4.74dB(ref) at 4 degrees (elevation). dB ref =
2.14dBi so I suppose this is saying gain is 4.74dBd.

If the antenna is in free space, the currents in the elements are
exactly in phase, and the transmission line radiation is negligible, the
gain should be maximum broadside to the antenna, not at 4 degrees
elevation. So one or more of those conditions isn't true. The horizontal
directivity indicates that the transmission line section is radiating,
but I wonder also if you might be modeling it over ground.

The gain is 4.7 dB relative to a dipole in free space. If you're
modeling it over ground, the gain isn't 4.7 dB over a dipole at the same
height above ground, which is why I dislike the term "dBd". (But antenna
manufacturers who can capitalize on the ambiguity love it.) If you're
modeling it over ground and want to know the gain over a dipole at the
same height, model a dipole at that height and compare gains.

Roy Lewallen, W7EL

David wrote:
Thanks for the model, I'll try it.

What is the diameter of the elements. I understand you can also vary the
feedpoint impedance by changing element diameter ? (Zo = 276 log(2S/d)

Open the EZNEC file. In the main window, click the Wires line to open
the Wires Window. Find the Diameter column, where you'll see the
diameter of each conductor.

The equation you give is approximately the characteristic impedance of a
transmission line, not the feedpoint impedance of a J-pole. The
relationship between the two is very tenuous and complex.

Roy Lewallen, W7EL

Roy,

The model I used for the open stub J-pole was not the same as your 3
element col.

I used a vertical long element and vertical short element. They were
joined by a horizontal element at the base. The source was set at 50%
along the horizontal bar as per your diagram.

I changed the model for free space rather than over real ground but this
has not helped the directional aspects of the plot.

The dimensions were scaled form those shown for a 150MHz open ended
J-pole shown on Arrow Electronics site

http://www.arrowantennas.com/j-pole.html



Roy Lewallen wrote:
David wrote:

Thanks for the model, I'll try it.

What is the diameter of the elements. I understand you can also vary
the feedpoint impedance by changing element diameter ? (Zo = 276
log(2S/d)


Open the EZNEC file. In the main window, click the Wires line to open
the Wires Window. Find the Diameter column, where you'll see the
diameter of each conductor.

The equation you give is approximately the characteristic impedance of a
transmission line, not the feedpoint impedance of a J-pole. The
relationship between the two is very tenuous and complex.

Roy Lewallen, W7EL

Two other cautions for anyone modeling a J-pole with any NEC-2 based
program:

1. Be sure to read the EZNEC manual section "Closely Spaced Wires" --
you'll find it in the index.
2. Run an Average Gain test. (See "Average Gain, Detecting source
placement problems" in the manual index.) The small loop in closed-ended
J-poles can be a problem for NEC-2. If the Average Gain test shows a
problem, using the double precision calculating engine (available in
EZNEC+) might help.

Roy Lewallen, W7EL

David wrote:
Roy,

The model I used for the open stub J-pole was not the same as your 3
element col.

I used a vertical long element and vertical short element. They were
joined by a horizontal element at the base. The source was set at 50%
along the horizontal bar as per your diagram.

I changed the model for free space rather than over real ground but this
has not helped the directional aspects of the plot.

The dimensions were scaled form those shown for a 150MHz open ended
J-pole shown on Arrow Electronics site

http://www.arrowantennas.com/j-pole.html

"Jerry Martes" wrote in message
news:b%NWe.1010$9a2.584@trnddc04...

"Richard Harrison" wrote in message
...
Cecil, W5DXP wrote:
"The 180 deg. phase reversing coil is the tricky part."

For UHF, you might prefer to use a 1/4-wave short-circuited stub in
place of a coil to reverse the phase. My 19th edition of the ARRL
Antenna Book shows such an antenna, "the super J-pole on page 16-25. At
other frequencies, this might be called a "Franklin Antenna". It`s a
1/2-wave in-phase with another 1/2-wave, one mounted directly over the
other. Best regards, Richard Harrison, KB5WZZI

Richard

Two half waves in phase, colinear. The 1/4 wave shorted "stub" is perhaps
the easiest way to get a 'good" 180 degree shift to get the segments in
phase. Used from early antenna design times...
73, Steve, K9DCI

Roy,
I'd appreciate your further comment on this. I 'know', from history,
that the spacing of two collinear half waves in phase affects the gain and
that there is an optimum spacing for gain that places the two collinear
elements apart (not close enough for a 1/4 wave shorted stub to be used for
a common feed). I had theorized that this "optimum" spacing results in the
most compressed lobe (without excessive secondary lobe formation) which is
simply due to the far-field phase summation of the two elements radiation.
Is this different than the mutual coupling to which you refer, or is this
another effect?

73, Steve, K;9.D,C'I




"Roy Lewallen" wrote in message
...
Jerry Martes wrote:

I'd have expected the "gain" to be closer to 4 1/2 db over the 1/4
wave
stub over a ground. is it easy to show where i've missed something?


I think it should be more like 3 dB, but hadn't said anything until I
had a chance to model it. ...
Roy Lewallen, W7EL

Steve Nosko wrote:
Roy,
I'd appreciate your further comment on this. I 'know', from history,
that the spacing of two collinear half waves in phase affects the gain and
that there is an optimum spacing for gain that places the two collinear
elements apart (not close enough for a 1/4 wave shorted stub to be used for
a common feed). I had theorized that this "optimum" spacing results in the
most compressed lobe (without excessive secondary lobe formation) which is
simply due to the far-field phase summation of the two elements radiation.
Is this different than the mutual coupling to which you refer, or is this
another effect?

73, Steve, K;9.D,C'I

It's actually the same effect.

If we assume no loss (a reasonable assumption for this kind of antenna),
all the power applied to the antenna is radiated. So any change in gain
is accompanied by a change in pattern -- if there's a single major lobe,
the more gain you have the narrower the lobe is. But if the element
currents are in phase, there will always be a maximum broadside to the
array regardless of the spacing, because the fields from the elements
will always add in phase in that direction.

But why is the gain different for different element spacings? You get
different gains for different spacings, which means that the sum of the
fields changes as you change spacing -- and this means that the field
from each element changes as you change the spacing. If you put 100
watts into the array, each element will radiate 50 watts, again
regardless of the spacing. So why does the 50 watts produce a larger or
smaller field broadside to the array as you change the spacing? What's
happening?

As the spacing changes, the mutual coupling between elements changes.
This alters the feedpoint resistances (and reactances, which aren't
important to this discussion) of the elements. And this in turn modifies
the amount of current flowing on each element for that 50 watts of
applied power. (It can be more reasonably argued that the mutual
coupling changes the current, and that changes the resistance. Or that
the mutual coupling produces a feedpoint voltage which alters the
current and resistance. But you reach the same conclusion via any of
those paths.) The essential fact is that the field gets stronger or
weaker as the element current increases or decreases as a result of
mutual coupling. The combination of the changed pattern shape due to
spacing and the changed maximum pattern size due to mutual coupling
always result in all 100 watts being radiated.

Chapter 8 of the ARRL Antenna Book has a graph of gain vs spacing for
two half wave elements placed end-to-end, compared to a single half wave
element. The gain peaks at about 3.2 dB at a spacing of about a half
wavelength. When the elements are very close, as they are in the super
J-pole, the gain is only about 1.6 dB greater than a dipole. That's why
I felt the J-pole gain wouldn't be as high as claimed.

Note: The current distribution on the elements also changes as a result
of mutual coupling -- see
http://eznec.com/Amateur/Articles/Current_Dist.pdf. But I don't believe
the effect is very significant on a collinear array with thin elements.
Anyone wanting to find out for sure, though, can do so with EZNEC or a
similar program.

Roy Lewallen, W7EL