Modeling question
 

On a learning curve here. I'll admit I am after a fast answer or two.

Using 4nec2 to model what will likely be a 2m antenna used at the home
QTH. Its an extended double zepp. ie two 5/8's in phase with a hairpin
like centre section, the whole thing made of 1/2" copper and fed at 200
ohms with a 4:1 balun or 100 ohm Q section. I have also included the
mast but havent really adjusted for best gain. There will be a quad on
the other side that isnt modeled yet.

Looking good so far but;

- Is the diameter of the source wire an issue?

- How do I allow for the use of copper pipe 90 degree elbows used in
construction. If I model them as (say) a extra single wire at 45 degrees
to mimic the radious I run into segment len/diameter limits.

Any thoughts or feedback welcome. Nec file attached inline. (Apologies
for the line breaks) I have yet to go through convergence testing and
playing with numbers of segments in the shorter sections.

Cheers Bob W5/VK2YQA

CE
GW 1 13 0 0 15.19 0 0 16.38 0.012
GW 2 3 0 0 16.38 0.28 0 16.38 0.012
GW 3 3 0.28 0 16.38 0.41 0 16.38 0.012
GW 4 1 0.41 0 16.38 0.41 0 16.45 0.012
GW 5 3 0.41 0 16.45 0.28 0 16.45 0.012
GW 6 3 0.28 0 16.45 0 0 16.45 0.012
GW 7 13 0 0 16.45 0 0 17.64 0.012
GW 8 1 0.28 0 16.45 0.28 0 16.38 0.012
GW 9 81 0.501 0 0 0.501 0 17.7 0.025
GE 1
EK
EX 0 8 1 0 1 0
GN 2 0 0 0 14 8.e-3
FR 0 1 0 0 146 0
EN

Modeling question
 

"Bob Bob" wrote in message
...
On a learning curve here. I'll admit I am after a fast answer or two.

Using 4nec2 to model what will likely be a 2m antenna used at the home
QTH. Its an extended double zepp. ie two 5/8's in phase with a hairpin
like centre section, the whole thing made of 1/2" copper and fed at 200
ohms with a 4:1 balun or 100 ohm Q section. I have also included the mast
but havent really adjusted for best gain. There will be a quad on the
other side that isnt modeled yet.

Looking good so far but;

- Is the diameter of the source wire an issue?

- How do I allow for the use of copper pipe 90 degree elbows used in
construction. If I model them as (say) a extra single wire at 45 degrees
to mimic the radious I run into segment len/diameter limits.

Any thoughts or feedback welcome. Nec file attached inline. (Apologies for
the line breaks) I have yet to go through convergence testing and playing
with numbers of segments in the shorter sections.

Cheers Bob W5/VK2YQA

CE
GW 1 13 0 0 15.19 0 0 16.38 0.012
GW 2 3 0 0 16.38 0.28 0 16.38 0.012
GW 3 3 0.28 0 16.38 0.41 0 16.38 0.012
GW 4 1 0.41 0 16.38 0.41 0 16.45 0.012
GW 5 3 0.41 0 16.45 0.28 0 16.45 0.012
GW 6 3 0.28 0 16.45 0 0 16.45 0.012
GW 7 13 0 0 16.45 0 0 17.64 0.012
GW 8 1 0.28 0 16.45 0.28 0 16.38 0.012
GW 9 81 0.501 0 0 0.501 0 17.7 0.025
GE 1
EK
EX 0 8 1 0 1 0
GN 2 0 0 0 14 8.e-3
FR 0 1 0 0 146 0
EN

Bob, while your model does not produce any errors with NEC 2, the
"Average Gain Test" (AGT) returns:

Freq.(MHz) Wavelength(m) Avg. Power Avg. Power (dB) Comments
-------------------------------------------------------------------------
146.00000 2.05340 0.78364 -1.05885
Questionable

All GW cards return a warning: " Seg.Len/Rad 8. Also the pole
segmentation is considered too large.

It appears that accurate modelling of such a structure in NEC 2 is not
possible unless you reduce the size of your conductors.

73,

Frank

Modeling question
 

Hi Frank

The pgm I was using complained about Seg.Len/Rad 6 for which I
adjusted the seg lengths for accordingly. I guess the warnings are
different for each software front end.

More reading and expermenting to do! Tnxs for your input.

Cheers Bob


Frank's wrote:


Bob, while your model does not produce any errors with NEC 2, the
"Average Gain Test" (AGT) returns:

Freq.(MHz) Wavelength(m) Avg. Power Avg. Power (dB) Comments
-------------------------------------------------------------------------
146.00000 2.05340 0.78364 -1.05885
Questionable

All GW cards return a warning: " Seg.Len/Rad 8. Also the pole
segmentation is considered too large.

It appears that accurate modelling of such a structure in NEC 2 is not
possible unless you reduce the size of your conductors.

73,

Frank

Modeling question
 

Hi Frank

The pgm I was using complained about Seg.Len/Rad 6 for which I adjusted
the seg lengths for accordingly. I guess the warnings are different for
each software front end.

More reading and expermenting to do! Tnxs for your input.

Cheers Bob

Hi Bob, you are probably right about the different software standards. The
NEC-2 Users Manual, p8, first paragraph, states: "Delta/a must be 'about'
8 for errors 1%". I guess the developers of 4nec2 considered the error
for Delta/a 6 to be acceptable.

Your model, in free space with no pole, does deviate a little from a thin
wire dipole of the same length. Possibly the stub effects the pattern some.

Both of Cebik's books make excellent reading for NEC modeling, although
oriented to Nittany's products. Also www.cebik.com is a good source of
information.

73,

Frank

Modeling question
 

Hi Bob, you are probably right about the different software standards. The
NEC-2 Users Manual, p8, first paragraph, states: "Delta/a must be 'about'
8 for errors 1%". I guess the developers of 4nec2 considered the error
for Delta/a 6 to be acceptable.

Yes, and now I am wondering where I got this value of 6 instead of 8...

However when delta/a (len/radius) 6 a warning is generated, when
delta/a 2 an error is generated. When using more thick wires, you can
add the EK card. With this more high delte/a values are possible.
(warning if delta/a 2, error if delta/a 0.5).

Greetings from the 'developer',

Arie.

p.s. I can also recommend www.cebik.com

Modeling question
 

"Arie" wrote in message
ups.com...
Hi Bob, you are probably right about the different software standards.
The
NEC-2 Users Manual, p8, first paragraph, states: "Delta/a must be
'about'
8 for errors 1%". I guess the developers of 4nec2 considered the
error
for Delta/a 6 to be acceptable.

Yes, and now I am wondering where I got this value of 6 instead of 8...

However when delta/a (len/radius) 6 a warning is generated, when
delta/a 2 an error is generated. When using more thick wires, you can
add the EK card. With this more high delte/a values are possible.
(warning if delta/a 2, error if delta/a 0.5).

Greetings from the 'developer',

Arie.

p.s. I can also recommend www.cebik.com

Not sure if "EK" will help much here since the "wire" diameter is
lambda/100. Seems to be a borderline case for a wire grid model. It is
like building a 20m dipole from 10" diameter pipe.

Also noticed the pole touches the ground, which is not allowed in NEC2.
Don't understand why it does not produce a warning. Usually when I connect
a wire to a non-perfect ground the ground grid turns purple (NEC-Win Pro),
and I get crazy results, but not in this case. Not that this is relevant in
this model.

Frank

Modeling question
 

Hi Frank/Arie

Not really knowing what I was doing grin I tried EK and got the same
bad average gain figure.. I also removed most of the lower part of the
mast since its effect will be fairly minimal. No help there either.

Tnxs for your comments gents. There is still the two questions though;

- Does one bother to model in a "curved" wire for copper pipe joiners? I
note Mr Cebik didnt mention it when he was talking about J poles on 2m.

- Is the diameter of the source wire important?

Thinking now about reducing the copper pipe diameter simply so I can
model it properly. Then again its a pretty broad design and tuning would
be easy...

Cheers Bob VK2YQA

Frank's wrote:
Not sure if "EK" will help much here since the "wire" diameter is
lambda/100. Seems to be a borderline case for a wire grid model. It is
like building a 20m dipole from 10" diameter pipe.

Also noticed the pole touches the ground, which is not allowed in NEC2.
Don't understand why it does not produce a warning. Usually when I connect
a wire to a non-perfect ground the ground grid turns purple (NEC-Win Pro),
and I get crazy results, but not in this case. Not that this is relevant in
this model.

Modeling question
 

"Bob Bob" wrote in message
...
Hi Frank/Arie

Not really knowing what I was doing grin I tried EK and got the same bad
average gain figure.. I also removed most of the lower part of the mast
since its effect will be fairly minimal. No help there either.

Tnxs for your comments gents. There is still the two questions though;

- Does one bother to model in a "curved" wire for copper pipe joiners? I
note Mr Cebik didnt mention it when he was talking about J poles on 2m.

- Is the diameter of the source wire important?

Thinking now about reducing the copper pipe diameter simply so I can model
it properly. Then again its a pretty broad design and tuning would be
easy...

Cheers Bob VK2YQA

Hi Bob, attempting to model a curved wire will possibly lead to
over-segmentation. It is also important to avoid junctions of dissimilar
wire diameters as it can lead to errors. The source wire should therefore
be the same diameter as the rest of the structure.

73,

Frank