Topband: Elevated Radials

[Guy Olinger K2AV]

On Sat, Oct 30, 2010 at 10:03 AM, Art K6XT <k6xt at arrl.net> wrote:
> My 160 ant is a 1/4 wave 80M GP. A 1/8 wave GP. Currently 18 radials, all
> 1/8 wave on 160... I have space so began adding radials. In the area of 10 radials
> there were no further, significant, changes in measured feed Z so I began to
> think the antenna was reaching a decent balance between gain and the work.

I realize this is turning into a long article on the reflector, and
overall is not really a response to an individual poster.  But we are
into yet another season and it's getting late for construction.

The short response is that four 1/8 wave elevated radials create a
counterpoise that is in the one or two ohm series resistance range and
effectively solves series resistance losses.  What was accomplished
with the ten and more elevated 1/18 wave radials was to create a
smooth field underneath which was to some degree canceling the
downward aimed field from the vertical conductor above. Canceling the
field with something electrically attached to the antenna system will
change the system feed Z so you can see when the point of no further
returns is reached.  This would however be safer measured with field
strength to guard against some local anomaly like buried conductors
distorting the returns and telling you done, when say 14 is the real
number.

Long version:

All elevated radial discussions below relate to elevations of 7 feet
(2 meters) or more.  Most of my modeling of is done with radials at 8
feet or 2.5 meters. Testing is at 7 or 8 feet which is practical and
non-risky reaching by me in the woods tip-toe.  I'm not very tall.
Lower than that is a hazard anyway and antler bait.

All discussions below are irrespective of SWR bandwidth.  The gold
standard dense buried radial and dense (32 elevated 1/4 wave) elevated
radial solutions are both broad AND efficient.  Everything else gets
into a tradeoff between bandwidth and efficiency.  Anyone without
dense solutions should be paranoid about broad SWR as indicative of
unaddressed LOSSES.

None of this applies to over salt water or salt water marshes which
are their own most fortunate separate universe.

On to the subject:

BURIED radials have much higher effective series resistance per radial
than elevated radials, and so it takes more in parallel to drop the
radial system series loss resistance. This is also why sparse buried
radials tend to do poorly.

TOO LOW "ELEVATED" radials quickly assume the attributes of buried
radials as they go low.  There is a clue to this in the method that
non-NEC4 modelers are told to model buried radials:  use "raised"
radials one or two wire diameters above ground with an end segment to
real ground (NOT MINNINEC ground).  So radials "on" or close to the
ground are roughly equivalent to buried radials.

In a sparse elevated radial field, once you have put up enough
elevated radials to lower the series resistance of the paralleled
radials down to one or two ohms, you are at the end of performance
improvements DUE TO REDUCED SERIES LOSS.

DENSE fields of elevated radials engage another phenomenon, a FIELD
CANCELLATION ZONE UNDER THE RADIALS.  This is where the field from the
center conductor above heads for the dirt at one polarity, and the
field from the radials above heads for the dirt with the opposite
polarity at the same amplitude.  In practice this is a not a sharp
cancellation, but it's enough to change feed Z when it's working, and
this is would-be induction loss from the vertical radiator in the dirt
under the radials that does NOT happen.

With dense 1/8 wave radials, you can only accomplish 1/4 of the
cancellation area possible with dense 1/4 wave radials, just because
of the smaller area covered BUT it's the area with the largest field
amplitude to cancel.  The 1/8 wave radials also don't need as many
wires to create a smooth reverse field below  ---  that's a function
of how far apart the radials are at their end.

I would estimate that your ten 1/8 wave radials was well beyond the
point of dividing the series ground induction loss below an ohm, you
probably hit that with FOUR of them over all but the worst dirt.  BUT
ten was at the point where whatever field cancellation that was
possible had been accomplished because the opposite polarity field
from the radials was "smooth enough" to do whatever cancellation it
could underneath.  The math suggests that the same point of
diminishing returns would have occurred with 20 1/4 wave radials at
your place.  Note that in that situation adding a few 1/4 wave radials
can't help in reducing series loss (already at point of no further
returns) and doesn't help field cancellation because the addition is
not dense, therefore not smooth and not equal and opposite the
radiation out there.  Again that matches your experience.

16 to 32 elevated radials is the anecdotal range where 1/4 wave
radials hits the point of no further returns.  For the same separation
between wires to accomplish the same smoothness of counter-field
underneath, 1/8 wave radials need 8 to 16, which contains your 10.  My
unverified SWAG is that 16 elevated 1/8 wave radials will accomplish
1/2 to 2/3 of the field cancellation of 32 elevated 1/4 wave radials
because the densest fields are close under the feedpoint, and that is
the range of most bang for the buck.

32 for 1/4 wave and 16 for 1/8 wave are the safe dense numbers
regardless of dirt.

This is for a vertical radiator that has its current max AT THE
FEEDPOINT.  Current max at the feedpoint will be true for a radiator
that is 1/4 wave and less. If the radiator is longer than a 1/4 wave,
the current max will move up the wire, and the equation for
cancellation becomes very complicated and one needs a model as an
addition tool to estimate what performance improvement occurs by what
changes.

If you CAN go DENSE, then you can live with the current max down at
the feed. This is the basis of the gold standard dense radial fields.
In this regard your 10 1/8 wave elevated radials seems pretty good.
Since a large percentage of the heavy field cancellation possible is
inside the 1/8 wave radius.

If one CANNOT go dense radials, then the current max at the ground is
your ENEMY, because that has the maximum magnetic field ground
induction loss invoked from the VERTICAL RADIATOR. If you move the max
up the radiator, then you diminish the field in the ground directly
underneath.  This is a real improvement in the radiation THAT IS NOT
SEEN IF THE RADIALS ARE DENSE.  The improvement is seen when the
radials are sparse, and DOES compensate for some of the loss due to a
non-dense counterpoise solution.

In 1/8 wave talk, if you can't do 8-16 elevated 1/8 wave radials, then
the radiator needs to be designed to minimize ground loss and moving
the current max up the wire will improve results.  If constrained to
the sparse radial side of the antenna design fence, further minimizing
reflection loss has been conceded to the enemy.  Now you just need
your counterpoise to be EFFICIENT.  You can forget "purity" issues and
go for the best low angle 5/16 or 3/8 wave L, or a large T, trying to
place the current max 1/16 wave below the top of the best vertical run
you can accomplish in your situation.  This results in the least
ground induction by the radiator without giving up low angle
radiation.

We have talked about 4 1/8 wave raised radials being sparse, but
efficient. But what do you do if you don't even have the 90 x 90 foot
box on your property to put up as little as 4 1/8 wave raised radials.

Needing a counterpoise JUST to be efficient opens the door for
non-traditional counterpoises.  It has been very hard to talk about
these in the past because the flip point in ground/radial strategy was
never identified or quantified.  From what I can gather from many,
many communications of one sort or another,

MOST OF THE 160M POPULATION DOES NOT HAVE THE SPACE, CIRCUMSTANCE
AND/OR MONEY FOR DENSE RADIALS. MOST OF THOSE CAN'T DO 1/8 WAVE
EITHER.   RANDOM PARTIAL ATTEMPTS AT DENSE RADIALS ARE LOSSY AND
GENERALLY REPORTED AS UNSATISFACTORY BY THOSE WHO HAVE THEM COMPARED
TO THE "GOOD SYSTEMS" .

So the real discussion for most people is how to do efficient when
dense simply isn't going to or can't happen.  A lot of people have
configurations installed on 160 that are not dense, therefore REQUIRE
efficient, but instead are predictably lossy.  But most of the
literature solutions out there only says more radials.  Nobody is
talking about how to make SPARSE work as efficiently as it can,
especially not talking about a city lot.

You don't hear it said that with a 1/4 wave or less radiator over bad
dirt, sparse 1/8 wave elevated radials will outplay sparse 1/4 wave
because the 1/8 wave will have a less series loss.  You don't hear it
said that over sparse radials, move the current max up the wire.

As to buried radials, insufficiently dense buried radials are a big
loser, and miscellaneous collections of short buried radials combined
with low Z vertical wires can approach the infamous driven iron pipe
in the degree of ground loss. It is very easy to create models in NEC4
that don't radiate even 10% of energy applied at the feedpoint.  I had
one such worm warmer in my suburban backyard in 1965 that had played
magnificently when I lived in a rowhouse apartment and the vertical
was over a whole block of copper roofs in a row, akin to ham life over
saltwater.  I guarantee the suburban version did not even do 10%.  I
was remarkably ignorant about those issues even with my extra class
license earned in 1963.

Four 1/8 wave elevated radials is certainly efficient for series loss
in all but the worst dirt and one of ON4UN's solutions, but eight or
sixteen is better to cancel some ground induction loss from the
radiator, and if you already have the real estate, then why not?  But
you will need something else if you don't have the 90x90 footprint
available on property.

More to come on some efficient space-saving non-traditional
counterpoises, including a single wire 5/16 wave folded counterpoise
that is +/- 32 feet and comes in at two or three ohms series loss over
so-so dirt. I am writing an explanation for a web page which contains
how-to content.

I will be on 160 soon with up 90 and out 110 over one of these,
invoking all the issues above.  It will be something like 90 ohms feed
R in series with 2 ohms series loss R. The current max will be up
about 75' and we shall see how it does.  I will be using the 516FC
because I can't cross the driveway with radials, but can run it
parallel to the driveway.

I have not really talked about bandwidth or SWR.  Elevated radials
less than 1/4 wave will need to have the extra capacitive reactance
accounted for in the antenna system somehow.  IF it is EFFICIENT, it
WILL be NARROWER bandwidth.  LOSS makes a given configuration BROAD.
To see this in action in a model, simply check off "free space" for
the ground definition in most models, rerun and watch the SWR curve
change.

If you can't do the gold standard, part of the issue is dealing with
some degree of narrowness.  Some answers for narrowness:

1) put the center of your narrowness on 1827 and not bother with
contests and other activities that drive you high and low.

2) put in an electrical halfwave of larger hardline from the feed and
use a tuner in the shack. The diameter of the halfwave feed is not a
concern with QRO, just with loss, more important for QRP.

3) put the feed close enough to the shack that you can use a SHORT
piece of RG8-ish coax to a tuner in the shack.

4) use a remote tuner at the feed point. These can be bought, but for
QRO are really expensive.  If you are a 100 watter, a lot of the
remote brick tuners will be quite good for this.  The trick is to put
them RIGHT AT THE FEED.  Up on the post if that's where the feed is.
This solution has a lot going for it.

5) create 3 or 4 series matching transformers, which as a collection
cover the parts of the band you use, and use relays to switch them in
and out. One of the lines should be straight coax. E.g. SMT which
matches at 1810, another at 1827 (straight coax if you hit 50 ohms at
resonance), another at 1845 and another at 1865. This is more
efficient overall than a long coax run with a tuner in the shack.

These SMT's would be a length of RG8, a length of RG11, then RG8 the
rest of the way.  Some matching device at the feed to get you near 50
ohms at 1827 makes this more straightforward.  SMT's have the
advantage of covering distance with the solution, and with mild
matches introduce no extra loss.  AND coax is cheaper than a lot of
things.  There is a free program called smc.exe that calculates these
for you given a target Z equal to a coax characteristic Z and a
miscellaneous R and X to be matched. SMT's and smc.exe have worked
well for me on some number of sites cleaning up misc impedances so
that long coax runs can be traversed matched to Z0 at their lowest
possible loss .

(5) has the disadvantage of not working for the buy it and hook it up
crowd (whether for reasons of time or expertise, not intended as a
slur).  You have to be into the theory and construction to make this
work because your installation will be different than anyone else's,
and you will have to measure and construct your way to the solution.

73, Guy