On 9/10/2019 4:56 PM, Bob Shohet, KQ2M wrote:
Sure I did my homework ahead of time but my inclination was to build it and put it up anyway just to see
how it performed. I felt better about it after I modeled it even though I was using software which was not
designed to provide a perfect representation of my qth with stacking heights over my wildly varying
topography in each direction. And while the antenna on 15 and stacking combos kicked butt at 109’,
a similar stacking proposition for 10 meters with an antenna at 65’ stacked with others at
100’ and lower heights did NOT work as the model predicted. In fact the 10 meter antenna at
65’ performed poorly in a stack with all other antenna heights and by itself in every direction BUT
towards Europe.
So “put it up” and try it had great results on 15 and mixed results on 10, whereas the model which
predicted a great stacking pattern in all directions for both antennas on 10 and 15 also had mixed results.
Ultimately I removed the sidemount @ 65’ for 10 – it was useless to rotate the antenna at that
height for stacking purposes at my qth – it was never better than another antenna combo. So, no model is
foolproof.
Not quite. This post, and the logic that goes with it, is a great
example of several important concepts/principles.
1) Nearly all real world problems are complex, whether technical,
personal, financial, legal, or political. Wise people know this, and
will reject input that tries to make things too simple.
2) The application of well known fundamental principles to solve a
problem will only be as good as the knowledge of, and the modeling of,
ALL of the variables.
3) Anyone who thinks that "there's a difference between theory and
practice" doesn't know enough about one or the other or both.
4) Nearly all commonly used design equations/formulas are simplified
versions of the full equation. The simplifications are based on certain
assumptions, and if the assumptions are satisfied, the formula will give
a good result.
Example: Commonly used equations for Zo and VF of transmission lines are
simplified, based on the assumption that F is a high number, leaving F
out of the formula, yielding the same value for all frequencies. But Zo
and VF are NOT constant with frequency, and become increasingly variable
at low frequencies. Indeed, VF varies enough that stubs for 40M and
below cut using the simplified formula will 1-2% off frequency.
Example: NEC assumes flat earth and uniform ground characteristics, and
uses ground characteristics in the model. If the ground is not flat, or
if it is not uniform, or even if its characteristics are not known, or
are not entered correctly, NEC will yield some error.
Example: Most hams have conductive elements within their direct field
that can act as parasitic elements of an antenna. Coax from horizontal
dipoles, towers, trees, even wiring in nearby homes can interact with
verticals. If these conductive elements are not known or not considered,
NEC will yield some error. I've posted here several times the
interaction between my tall tower and 160 verticals, and that I add
chokes along the coax from high dipoles to avoid interaction.
Example: HFTA does NOT model interaction of stacked antennas, it only
sums their lobes, assuming that each is a dipole! When a given antenna
is selected (dipole, x-element Yagi) HFTA simply increases the gain. And
the result obtained from the model will only be as good as the data
entered. My terrain is quite irregular -- I don't know how good the data
is nor how closely spaced, and there are limits to the number of data
points.
Example: HFTA can yield errors with some irregular terrain. N6BV has
always recommended running calcs at multiple closely different heights
and azimuths to expose these errors. When I used it to site my antennas
I ran in 10 ft increments to find "sweet spots," then +/- 5 ft. Likewise
for azimuths -- I modeled in 5 degree increments to major directions.
Example: HFTA propagation data is statistical, over long periods of
time, so includes wide variations over the solar cycle(s). AND data for
some bands is interpolated from data for adjacent bands. I think I
recall Dean telling me that 160M data is interpolated.
Example: The combined pattern of antennas in a stack will depend
entirely on the phase response of the combination, including ground
effects. The phase response of an antenna like a Yagi will vary not only
by brand name or generic design, but by every element of the design, and
it varies with angle in both the H and V plane. If we want to model the
behavior of a stack on something as simple as flat earth, we must enter
an accurate detailed model into our design software, and the model must
include accurate details of all matching elements, phasing lines, and so on!
73, Jim K9YC
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|