Re: NEC-LIST: Image Plane Log Periodic Array

In a message dated 3/29/03 2:45:04 AM Eastern Standard Time,
ccc_at_space.mit.edu writes:

> AFAIK, there's nothing magic about a triangular mesh or an overall
> hexagonal shape. You could use a square mesh and a rectangular or
> square overall shape.
>
> -Chuck.
>

Chuck is right on this. Triangles (for the mesh) just use fewer wires, but
this can be key if you are time or wire number-limited. Otherwise, no
difference. Distributing current over surface is the name of the game.

However, meshing GP's is one of those areas in which people build to test
the model, and then tweak the model(!). You can see this limits the
predictive utility of modeling with any great precision.

You do reach a sensitivity in under-sampling the mesh more quickly with
uniform triangles, so if you want to can as many wires as possible it pays to
watch the gains of a sample monopole model as you pull back the mesh sampling
.

Most folks like to sample much higher near known points of high current,
often near edges (on a box for example) and feedpoint(s).

The image LP will be quite sensitive to the size of your GP, at the lowest
frequencies, so don't skimp on the mesh just because of this. If you can do
it, try meshing for the highest frequency first, and keep that mesh for the
others.

A good rule of thumb is to look at the currents and question whether some
departure from a slowly varying function makes sense. If not, hit it with
higher mesh density there.

73,
Chip N1IR

-- 
The NEC-List mailing list <nec-list_at_gweep.ca>
http://www.gweep.ca/mailman/listinfo.cgi/nec-list