Re: Segments

In reference to RC Looke's question regarding
bowtie antennas:

NEC does not like sharp angles therefore
you should try to avoid them when modeling
an antenna. (excuse the art...probably most
editors won't display it right anyways)

 \ /
    \ /
        \ /
            \ /
               \ /
   sharp ( \ /
--------------S--------------
                 / \
               / \
             / \
           / \
        / \
     / \

       "Bad model"

A way to more accurately model this
antenna is shown below.

       \ /
           \ /
              \ /
                 \ /
   l arger( | |
      ---------|-------S-------|---------
                  | |
                 / \
              / \
          / \
       / \

         "Good model"

Basically, instead of having the wires meet right
at the sharp angle, extend the area around the
source and then start the wire at the coordinates
which will give you a larger angle.

Also, you should break each wire into
smaller wires so that you can place
more segments in the area close to
the angle.

Hope this helps,

Joe Finnerty
jdf112_at_psu.edu
Paragon Technology

>Would anybody care to comment on choosing the optimal segments per wire in
>NEC-2 ?
>
>While simulating a di-pyramidal over a 2 octave band, I noticed significant
>differences in impedances when trying to optimize segments.
>
>For example, here's a sample of impedance vs. segments per wire at a
>mid-frequency ( aperture is about 1/2 wavelength ).
>
> Segs Zant
>----------- ----------------
> 25 102.3 -j274.6
> 50 62.7 -j41.1
> 75 51.1 -j33.2
> 100 45.8 -j29.7
> 150 41.3 -j26.6
>
>Note: At above 150, I reach max segs in my software.
>
>
>I wonder if I can assume that the highest possible segmentation always yields
>the most accurate results. I'll be using the impedance data to build matching
>networks, so I'm wondering to what degree can I trust the generated data.
>
>If you want to help me out with this one, but need more info, I'll be happy to
>send you my model.
>
>Thanks,
>RC
>
>
>
>
Received on Thu Nov 30 1995 - 09:43:00 EST