[TowerTalk] Practical Ground Screens for NVIS antennas

Jim Lux jimlux at earthlink.net
Mon Jul 25 14:10:16 EDT 2005


At 10:24 AM 7/25/2005, K4SAV wrote:
>Interesting question.  I have never reasearched this one, even though I
>have thought about doing it, so I tried a few quick simulations.  At
>first the suggestion of making the reflector 5% longer seem to make
>sense, but ... read on...
>
>Using an 80 meter dipole, 133 ft long, at various heights, with a
>reflector at 1 ft above ground with various lengths.
>
>In the table below, first column is the reflector length, then seperate
>columns for different driven element heights, gain figures are for
>straight up (zenith).
>
>R length   20ft   30ft    40ft   50ft
>    76       4.79   6.17  6.69  6.77
>    73       4.94   6.24  6.73  6.78
>    70       5.15   6.34  6.78  6.8
>    66       5.6     6.54  6.88  6.84
>    63       5.77   6.67  6.96  6.87
>    60       3.51   5.93  6.66  6.84
>     0        3.67   5.61  6.4    6.59
>
>At first it may seem surprising that the peak gain seems to be for a
>reflector that is a little shorter than the driven element.  This is
>because the reflector is being tuned to a lower resonant frequency by
>the ground capacitance (which is higher than that for the driven
>element). Actually, my simulations show the resonant frequency of this
>132ft wire, 1 ft off ground is 2.6 MHz.


Excellent use of modeling..
Note that even in the worst case (20 ft up) the difference is only 1.5 -2 
dB, and the fact that it varies fairly quickly indicates that it's quite 
sensitive to small changes.

FWIW, it's not only just bulk capacitance to ground, but that the field 
from the wire is partially immersed in a dielectric.

Running in NEC4, and assuming a copper wire 1mm in diameter and 132 ft 
long, I got a resonant frequency of about 3.28MHz, with the wire 30cm (1ft) 
off the ground.  With the wire 100 meters off the ground, I got 3.62 for 
the resonant frequency, and about 3.63 in free space.

Not sure why there's the difference in resonant frequency from your model. 
(Running NEC2, I get 3.26 for the resonant frequency of the 1 ft high 
case)  If I drop the wire to 3 cm, the resonance drops to 2.7 MHz (in NEC4, 
with 45 segments or with 101 segments).  If I change the epsilon to 12, the 
resonance is pretty much the same (within 10 kHz). Change epsilon to 30 
(i.e. soak that soil down!) and the resonance moves up to 2.73.  Back to 
13, but doubling conductivity to 10 mS/m has about the same effect (fres = 
2.73)

So, we're seeing 1% changes in resonant frequency with fairly large changes 
in soil properties. Jerry's changes in length were on the order of 5% 
(which would have a similar effect in resonant frequency change).

I suppose I should run a model for 126 ft?  (same 3cm off the ground gives 
fres=2.84 MHz)


>These numbers haven't been checked against any other references.  It
>represents a quick simulation done with EZNEC, using Real/High Accuracy
>Ground, Cond = .005 s/m, diel const = 13.   Since we are dealing with a
>wire only 1 ft above ground, I would expect slightly different numbers
>to be produced by NEC4, but the trend should be the same.
>
>In practice, these simulations are not going to be accurate because of
>the variations in ground parameters.  You probably couldn't even measure
>the resonant frequency of your ground reflector and keep it constant
>because of varying conditions, such as rain.  I simulated only one
>reflector, but if you build this, you will want more than one.
>
>If you are not interested in DX and only in generating maximum straight
>up signal, the max gain appears around 55ft height for a 3.6MHz antenna.
>
>Jerry, K4SAV






More information about the TowerTalk mailing list