[TowerTalk] Practical Ground Screens for NVIS antennas

[Jim Lux]

At 10:42 AM 7/26/2005, Bob Kellow, W5LT wrote:
>Seems like this problem should be an easy one to model with AO or similar
>modeling program.
>Tell the program the antenna is in free space, then define the antenna and
>then the screen in the Z-lane below it. The diagonal version would be
>easiest to model. Anyone want to give it a try?
>73s
>Bob, W5LT

The trick is in adequately modeling the earth losses.  Wires close to the 
ground are always challenging to model, and small changes in the spacing 
(from earth) or the dielectric properties of the soil can change things a 
lot.  And, then, once you're done, you'll have a pile of model data that is 
hard to relate to actual use.   As you say, it would be pretty straight 
forward (although time consuming) to iterate over some ranges of spacings, 
lengths of wires (or zig zags) and soil properties.

On the other hand, if you set up in a field, and do a couple dozen tests 
with various arrangements, you'll not only get the RF data (which could be 
used to validate the computer modeling results) but you'll also get that 
invaluable data on how long it takes to lay out the wires in various 
patterns, etc.

In the context of a potential QST article, too, it would be a pleasant 
change to see both the model data AND actual field test results with 
halfway decent experiment design.  It would be fairly easy to get actual 
quantitative results (as in dB differences) instead of a "I worked umpty 
ump countries with a model X transceiver at some unspecified time and date" 
or a "it sounds loud" or "it's full quieting into the repeater".  The NVIS 
path is moderately stable (compared to, for instance, a DX beacon with 
multiple hops and traversing many different ionospheric conditions along 
the way).