[Antennaware] Modeling Software -

[Terry Conboy]

On 2009-02-20 8:13 AM, K9AY wrote:
> Most recently, I relied heavily on modeling to decide how to build my new 
> 160M vertical. After exploring options, I settled on 100 ft of 25G, 
> series-fed, with a six-spoke top hat of #16 wire supported by 13-ft quad 
> spreaders from my boneyard. Modeling suggested that going beyond 100 ft. was 
> not a significant benefit, but anything less would need a larger capacity 
> hat and have a reduced VSWR bandwidth with a simple matching network. Also, 
> I was able to see the greatly increased loading by using a perimeter wire 
> connecting the ends of the capacity hat spokes. The measured zero-reactance 
> frequency is higher than modeled (1725 vs. 1670 kHz). This is easily 
> explained by uncertainty in choosing an equivalent diameter "solid wire" for 
> the 25G tower (I used 5 inches). Feedpoint reactance is quite close, but the 
> resistive component is higher than modeled, 45-50 ohms instead of 38-43 over 
> 1800-1900. I did not see this with modeling of a previous inverted-L over 
> the same radial system, and so far, cannot find a set of realistic 
> parameters in EZNEC that forces the model to produce that impedance. Maybe 
> I'll figure it out later this year when I install and evaluate a large 
> buried radial system!
>
>   
To me, one of the most frustrating limitations of NEC2-based modeling 
software is its inability to accurately deal with models where 
conductors of different diameters are joined, especially at right 
angles.  EZNEC (and probably other programs outside my experience) have 
adjustments for tapered diameter elements, such as those common in 
yagis.  But this does not really address connecting a wire capacity hat 
to a large cross-section tower, such as in Gary's 160m vertical.

I've had some possible success by modeling a tower as a shaped bundle of 
conductors of the same diameter as the top-hat wires.  At least the 
results seem more consistent, but it's hard to say, with only limited 
opportunities to verify them with real world measurements.  There are 
still often problems with the details of the wire junctions to the tower 
and adjacent crossbraces, which require care and checking for 
sensitivity to segmentation and verification of 0 dBi average gain with 
losses zeroed out.

73, Terry N6RY