Hi Mauri,
> what you say it's a bit pessimistic in numbers but fundamentally true
> because any electrical antenna, expecially when shorted, suffers a great
> amount of extra losses if close to other objects (proximity) or it's not
> normal to the ground or to other things. Knowing it, the best to do is to
> minimize inherent antenna losses meanwhile trying to avoid extra losses.
If you consider the fields around an antenna, you'll have a big
problem following the logic that loaded radials, or a few elevated
radials, are a "good thing".
Under the radials, unless the radials are a LONG distance away
from earth both in terms of wavelength and physical length of the
radials, there is considerable field magnitude under the radials.
I can give an example. I was walking with a 30 foot long wire tied to
my belt more than 30 feet below an elevated radial on a 5 kW BC
station.
That wire, dragging on the ground, arced through my pants and
underwear and burned a quarter size hole in my flank. If the field
below the radials is zero or even greatly reduced, that never would
have happened.
The more you concentrate those fields, the higher loss becomes.
There is no way around that problem except more wire or more
distance.
The fields are NOT dispersed and they are NOT cancelled with
small radial systems close to earth, and that is especially true with
loaded radials.
> It's amazing to measure, but with large antennas (where it's not possible
> to use an anechoic chamber) I'm always a bit skeptical about results
> validity. it's very possible to be misleaded mesuring something else than
> wanted (couplings, reflections), mixing the fields or doing mesures at
> useless angles of radiation. Nowadays I definitely find the join of logic,
> experience, long term reports and computer modeling at least valid as very
> extensive measures done by RF professionals (a kind on the way to become
> extinct).
While you question a direct comparison, and rightfully so since no
measurement is absolute, virtually all of the supporting data isn't
even that good. It is based on a long chain of approximations,
which is virtually worthless.
The total potential error in my instruments was about one dB using
recently calibrated equipment.
Until several different people do A-B tests with proper peer review,
the results are just opinions. I'm willing to donate time and space
to measurements, if an "opposing view" person wants to help.
Otherwise I'm done making tests. I know what works for me.
Also, I've seen (first hand) two different cases where considerable
FS was gained by only one change, going to a conventional
system.
Computer modeling is OK, but consider this. The early elevated
radial models were based on NEC-2, which was NEVER
accurately validated for wires close to ground at HF.
If you compare the Hagn-Barker field strength measurements of low
dipoles (two separate occasions in 1970) to NEC-3, at .01
wavelength the model disagrees with the REAL results by about 7
dB.
That would correspond to a 160 meter radial at 6 feet.
We know that error exists, it is documented. If the model is wrong
with a simple dipole at low height, it is most certainly NOT
trustworthy for low radial performance.
Quite coincidentally, modeling predicted the signal at WVNJ with
an error of about 5 dB or so. In that case the radials were about 30
feet, or .035 wavelength, high.
In all cases NEC predicts more field strength than is actually
obtained.
It also predicts the current in my Beverage antennas incorrectly.
Since current reduction in a low long antenna like a Beverage is
dominated by losses in earth below the antenna, that also raises
questions. It predicted about 3 dB less current loss than the
antenna actually had.
I'm not saying that models are useless, just that we have to be very
careful to be sure they work when predicting loss in something as
non-homogenous as earth. If they miss on a dipole near earth, you
can bet they will miss on a radial.
> On my house roof top (65ft) I've a 38 ft tower with an HF yagi on the top.
> The structure has two separate shunt feeds, one for 80 (gamma) and one for
> 160m (omega) with distinct loadaed counterpoises. When resonating the
> counterpoises the ammeter current reading doubles on 80m and the far field
> increases proportionally. On 160m, resonating the counterpoise increases
> 1.4 times the current and it's hard to notice a field increase, but surely
> it doesn't decrease. On both cases the counterpoises set the point were
> they are connected as the lowest voltage with all the inherent benefits.
Because something improves something in one case, it doesn't
mean it provides optimum performance or even improves things in
other cases. You went from almost no ground system to
"something" that was resonant. W8XO went from a fairly good
ground system to something different that added a resonant wire.
I would not expect parallel results, since the situations are totally
different. Plus you have no idea, as does Dave, how close to
optimum either system is. Neither of you did an A-B test to a full
size system.
It's also a warning flag that you saw a 1.4 times increase in
current, which is equal to a 3dB power increase, yet measured no
definite FS increase on 160 meters.
Perhaps a resonant counterpoise is better than NO counterpoise,
but everything I've seen indicates a conventional system is always
and reliably the best.
I don't play the lottery with antennas. I want them to work.
> When using a vertical radiator over a properly elevated counterpoise, the
> situation is quite a bit different than in a real ground plane,
> approaching rather that of an half wave vertical dipole over ground.
What is "properly elevated"?
If the wires are near earth, in theory four radials are not nearly
enough. Elevating a 160 meter radial ten feet does very little to
decrease fields in the earth. Shortening and "loading" the radial
makes things even worse, not better.
Because there is no direct connection to earth, it does not mean
there are no "displacement currents". It certainly does not mean
losses disappear under or near the antenna. Only a large ground
screen will do that.
> angle radiation and the efficiency at various elevations. On the contrary
> of horizontally polarized antennas where up to extra 6 dB gain are
> achieved at some angles (at price of nulls), elevating a vertical antenna
> reduces low angle radiation although proximity losses decrease.
I have dipoles at 50 feet and 310 feet on 160 meters, as well as tall
verticals with conventional ground systems. In over 600 signal tests
with ZL2REX and VK3ZL over the last year on 160 meters, over
70% of the time the omni-vertical was better than a high dipole. If
you toss out sunrise peaks, the vertical was better almost 100% of
the time. At no time ever was the low dipole better. It is a bird
perch and 100 mile distant radiator, nothing more.
Modeling predicts the high dipole has more absolute FS at all
useful angles when broadside to the dipole, yet the dipole in its
optimum directions can't surpass the vertical 300 miles away on
most nights!
Results certainly vary with location and a particular antenna, but
low angle losses for the vertical must be a lot less than indicated
by the model I used.
This debate will go on and on until a few people roll up their sleeves
and push back from the CRT long enough to do some
measurements like Brown, Lewis, and Epstein did. Until then I'll go
with what I know works, not with what might work is some cases.
73, Tom W8JI
w8ji@contesting.com
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|