Topband: Modeling a shunt fed tower with EZNEC

[John Kaufmann]

Bill Tippett wrote:

> I do use two media for the ground, using Eznec's
> "salt water" parameters for the initial radius of 150' to simulate my
> radial field, and then using Eznec's "average" parameters beyond that.
> Eznec gives me a guideline check ***"WARNING: Wire connected to real
> non-MININEC ground.  See manual."*** but I simply ignore this because my
> model seems to accurately reflect real-world impedance measurements.  


Bill Tippett wrote:

> I do use two media for the ground, using Eznec's
> "salt water" parameters for the initial radius of 150' to simulate my
> radial field, and then using Eznec's "average" parameters beyond that.
> Eznec gives me a guideline check ***"WARNING: Wire connected to real
> non-MININEC ground.  See manual."*** but I simply ignore this because my
> model seems to accurately reflect real-world impedance measurements.  


The two-media model is probably OK for calculating impedance data, as Bill's results appear to confirm.  However, the model has a known problem, confirmed by EZNEC author W7EL, in predicting far-field pattern and gain.  Therefore I would caution people about using the two-media model for anything more than simple impedance calculations.  

The root of the problem is that EZNEC uses a simplistic geometric model for calculating the far field when using two media in the near field of the antenna.  It models near-field antenna radiation as rays reflecting off the media, using different reflection coefficients for the two media, and then performing vector addition of the rays in the far field.  Many of you have seen ray diagrams where fields are represented by arrows.  However, rays are only a simple approximation of what's happening in reality.

The truly accurate way to model near-field radiation with two media is by a full electromagnetic field description of the radiation and not rays. The fields interact with the ground, and the interaction results in re-radiation of the fields into the far field in a modified form.  The calculation involves math which is more complicated than simple rays, although it can be done by numerical methods. 

W7EL himself has suggested modeling radials in the near field as horizontal wires a very short distance above ground (~.001 wavelengths) and using the "high accuracy" ground model in EZNEC.  There is an automated mechanism in EZNEC for creating ground radials of a specified length and number for this purpose.  EZNEC uses the "correct" method of calculating effects of near field wires, unlike the two-media model.

To reiterate, impedance calculations are probably not sensitive to the errors in the simple two-media ground model, but far field gain and pattern are.

73, John W1FV