> it comes and goes. Perhaps power line noise
> can't propagate in the horizontal polarization,
> making horizontal receive antennas relatively
> immune.
If power line noise is severely attenuated so are low-angle signals
that make up virtually ALL 160 meter DX signals. Of course at
times during strong peaks or abnormal conditions like magnetic
storms the difference between low-angle antennas and high angle
antennas diminishes, but even then properly installed verticals are
better (if you can do it).
The largest single most common mistake is designing a system
that is nearly impossible to make work the way the design (in this
case we should say the model) is supposed to behave. Models are
a useful tool, but we commonly assume the perfect world of a
model represents the imperfect and more complex real world.
Let me be specific, and show how we can judge the sensitivity of a
system to errors.
Using Eznec, model a vertical ten foot wire near earth. Center feed
the wire with a source, and insert a load near the source of 100
ohms to represent the low radiation efficiency of any loading or
matching system. Look at the gain and average gain of that wire
over average soil.
You should see about -26dBi response at about 25 degrees.
Average gain is about -31dB.
Place the same wire horizontally ten feet high, and look at average
gain and gain at 25 degrees. You now should see -37dBi gain at 25
degrees in the best direction, and -44dB average gain.
This illustrates the signal response or sensitivity difference of a
given size "perfectly fed" small antenna with only a polarization
change.
The problem illustrated above becomes MUCH more severe if we
have out-of-phase close spaced elements or a small loop antenna
mounted horizontally. That because the out-of-phase antenna
sections reduce sensitivity even more, while the feedline remains
just as sensitive to common mode currents. The K6STI antenna fits
this description.
As we place the antenna higher, the vertical feedline gets longer.
The increased area of the longer vertical feedline actually more than
offsets the increased horizontal antenna sensitivity.
As a matter of fact, it is virtually impossible to NOT have the
feedline a major part of the antenna, no matter what we do!
I saw a similar situation on a lowfer antenna, where it was
assumed a small whip mounted on a mast only used the whip as
an antenna. As the mast was made taller the antenna became
more sensitive, and a "rubber sheet" theory was invented to explain
why. I don't need a special theory, because I understand the
feedline (and mast) is virtually impossible to decouple from elevated
small antennas! This is at least ten dB more true for horizontal
antennas!
The amount of isolating impedance is all but impossible to achieve,
since the feedlines common mode response to signals may be
hundreds of times more than the small antennas sensitivity to
signals! That's why the snake antenna works in some cases, even
though we don't even bother to put anything we call an antenna at
the far end.
In one ten foot long dipole 20 feet high I modelled, it took 50,000
ohms of feedpoint decoupling to reduce feedline contribution of
signal enough to be **equal** to the small dipole response!!! Two
out-of-phase close-spaced elements would require even more
feedline isolation.
This is why I made three important points:
1.) The K6STI antenna is so critical to build, it is very unlikely
anyone can actually make it work as the theory suggests. That
does not mean it won't receive better than something else in some
cases. After all, even low 80 meter dipoles and snake antennas
can receive OK "by accident" on 160 at times.
2.) We often are not actually using the systems or patterns we
think we are when our antennas are in real-world congested
environments and connected to feedlines.
3.) It is almost impossible to receive or transmit a decent long
distance signal with a true horizontal polarization at low antenna
heights.
Note: I am **not** saying a low horizontal antenna won't work in
some cases, only that they would much rather behave like verticals!
73, Tom W8JI
W8JI@contesting.com
|