Topband: modeling BOGs (or whatever we call them)

[Guy Olinger K2AV]

Running way, way, way behind in responses to postings. Getting hopefully
well-thought-out responses the same day or even the same week or month is
apparently not always going to happen.

Short Version: Thank you Rudy. And there is more work to be done.

Long Version:

Please do not consider this to be dismissive of Rudy's work in any sense.
That's not at all what I intend. His (and our) proposition that an inert
BOG just laying there can be grown into the ground over time, and thereby
harshly deteriorate performance, is absolutely confirmed in our collection
of anecdota.

For long term performance, it is necessary to fix (make permanent) a BOG's
*electrical* relationship to ground by some mechanical design or process.
Lacking that, regular effective maintenance/adjustment must be kept up in
all but very arid environments. The deterioration in BOG performance is
without sudden drops like someone cut a wire. So it's very sneaky, and in
many cases sneaks toward extremes to the point of losing 10-15 dB and even
reversing pattern.

The QST article is rather severely edited for space. The full version will
apparently be in July QEX. For those who do not subscribe to QEX, or don't
want to wait, the full version can be found directly at
http://www.arrl.org/files/file/QST%20Binaries/June2016/QST-in-Depth-0616-Serverns.pdf

NOTE: My comments refer to the FULL Severns article and set of graphs found
at the arrl.org URL above, and presume the reader is looking at the full
article, not QST.

I will remain differing with Rudy on whether NEC is accurate. I personally
think that has to do with how we separately define "accurate". I would say
that NEC with regard to ground and MF antennas invoking ground is
*approximate*, and only to a point carries a correct graphical "shape" of
expected results in presentations.

That does not mean I am at odds with Rudy's measurements. I will trust his
measurements long after I give up on NEC. Just looking at his graphs, I see
those as evidence that NEC is *not* accurate in the sense that *I* use the
word "accurate".

Brown, Lewis and Epstein, and Rudy, make current measurements along the
line that show some degree of wavelength related wiggle. NEC4.x shows no
wiggle. If NEC were calculating with all the variables at play, NEC would
show the wiggle even if located differently or with different amplitude or
distance between nodes. That makes the NEC curve only approximate, however
useful that approximation may or may not be. Without knowing WHY the wiggle
is absent in NEC curves, it's hard to make a case to the authors to fix
that.

Some of the divergences between NEC 4.2 and his measurements portrayed in
Rudy's graphs would render an antenna diminished if NEC was followed
blindly. I have learned the hard way not to follow NEC-based programs
blindly and have the scars to prove it. I believe Rudy's measurements. To
my eye NEC cannot reproduce Rudy's measurements.

Further, Rudy is measuring at a single location, which is NOT a criticism.
I give all due respect for the time and effort testing there. A project
like this can be full of mechanical-blowing-up-the-electrical-results
issues needing to be avoided. Realizing an issue at some point well into
the procedure can force one to redo everything all the way back from the
start. Having the experiment in one's own back yard, with zero travel time
to the experiment, under one's own self-permission is a huge advantage to
reforming/restarting/finally completing the experiment.

Or sadly, as seen in other cases, running into "issues" having exhausted
resources and time, one has to give up on the project.

To further carry this exercise to the level needed to publish is even more
work. But it IS a single location, and we have to remember that. We must
excuse location centric for cause because picking sites and repeating the
testing in a dozen places all around the USA would create a huge,
time-consuming and expensive undertaking. Even then the case can be made
that not everything happens in the USA. Then how expensive does it get?

Our testing (I call them the Rowdy Raleigh Radio Researchers out of
earshot) only in the 12 county area around Raleigh/Durham North Carolina
showed huge variations in the primary electrical length of a 151' (46m)
dipole laid on the ground. The often referenced FCC ground conductance maps
calls us all 2 milliSiemens. However...

The velocity factor of that Dipole On Ground (DOG) varied from 0.45 to 0.8
across all sites and placements. That's +/- 22% (twenty-two percent not two
point two). Consider what would happen to a Yagi if the manufactured
element dimensions could only be guaranteed to lengths +/- 22 %. An
intended 15 meter yagi +/- 22% could actually be on 17 or 12 meters. Or if
all yagi elements did not have their errors vary in unison, could render
the antenna completely dysfunctional. Enter into the world of "wonderful"
to "d*mned waste of time" customer performance reviews. Sound familiar?

This measured variation in eastern North Carolina VF was not a gradually
changing figure with area changes in geography. Reorienting the compass
bearing of the DOG around its center in the same back yard, or placing the
DOG at another part of the same back yard, or just linearly sliding the DOG
up it's line for 50 feet could generate large variation in VF. This even
without buried pipes, wires, or septic fields in the yard. What effect that
may have had on Rudy's graphs if measured by his procedures in a dozen
locales scattered around the US is anyone's guess. It certainly would have
been varied. Varied quite enough to take a SINGLE instance "good" layout
for a 160 BOG in a specified location (like Rudy's back yard) thrown down
anywhere else and produce results varying from "works wonderful" to
"doesn't work worth a d*mn".

What do you do to take the BOG construction and have it respond to a normal
"wild variation" in VF of specific chosen spots of ground to lay out a
BOG?  Is it to measure the ground characteristics? Rudy hints that NEC 4 is
accurate if the ground characteristics are accurate. There is a long and
difficult discussion that could be had to show that even the FCC does not
believe this in their administrata for commercial LF/MF AM broadcast
stations, and they have a we-will-not-get-on-your-case cobble to get around
it. But Rudy does get approximate correlation in his back yard. And I trust
his measurements.

But even if we let that stand without challenge to whether it works
everywhere, there is another problem. Rudy's methods are full-on lab and
academic quality. And he has the equipment, software and expertise to do
it. Certainly not a criticism in *any* sense, Rudy attacks the problem with
1) a sophisticated knowledge base from an enviable employment experience,
and 2) a practiced experimenter's hand using 3) expensive equipment and 4)
expensive software, and 5) with a gift or two for excellence in technical
writing and publication skills.

Then there is Joe Average Ham, hereafter called Joe A H.

While Rudy's methods are full on lab and academic with adequate equipment
for those methods, Joe A H lacks the means to use *those* methods to do a
BOG on his own property. We need to arrive at something workable for Joe A
H with stuff *commonly available* to Joe A H.

Rudy's stuff and outstanding background makes him rare among the army of
Joe A H.

Rudy is using export controlled NEC 4.2. NEC 4.x is the only software from
the NEC family that can deal with buried conductors. One must pay a fee for
a license from a government agency to use Unix NEC on a Unix platform. If
you are not a naturally Unix person, then the high end professional EZNEC
Pro4 has a NEC 4.2 build that runs inside the EZNEC shell for Windows,
That's yet more $$ for the EZNEC Pro4 license, which is enabled by a key
that goes in a USB jack on your PC. You will not find any midnight copies
of Roy Lewallen's (W7EL) high end pro stuff on a Russian web site. (Read
around sometime about how hamdom screwed K6STI and shut him and his
excellent programs down.)

In addition to NEC 4.2, Rudy is using a VNA, and equipment sufficient for
accurate ground conductivity measurements. All his stuff and programs
together cost more than a high end state of the art HF transceiver.

Many Joe A H cannot find that kind of money in their budget for anything
other than necessities, if even that. And if they did have transceiver
level money, they would spend it on the transceiver, not the test equipment
and software. So who will be providing the instructions that allows Joe A
H, with typical Joe A H equipment, to hit the nail on the head with a BOG,
and maintain it?

I *personally* have found these expensive investments to be very
worthwhile, even if just for hobby and entertainment value since I find
this stuff extremely fun and interesting. But that's just me, and some
folks look at me just a bit askance...

I have had it unkindly hammered home to me that merely modest means
precludes the availability of Rudy-worthy equippage for practical Joe A H
construction of a BOG. Or for that matter, that even having time to do it
up to Rudy-grade standards just isn't going to happen. Again this is *not*
a criticism or dismissal of Rudy's article.

This is as well a hard-as-nails lesson regarding FCP kinds of things for
Joe A H. We need to write for Joe A H, design for Joe A H, and learn how to
do it with tools that can realistically belong to Joe A H.

To maintain a BOG that is working, Rudy's conclusions from his experiments
in the article and a large pile of anecdotal trial and error known to me,
some posted here, show that one cannot allow the wire to change its
effective height with respect to ground by allowing natural processes like
accumulating rotting leaves, etc, to gradually bury the wire, or bury it
deeper. Getting the BOG working well in the first place is a separate
story.

For MF ground-low-velocity-factor antennas, NEC requires a single
monolithic uniform ground medium. Real underfoot ground is most often
anything but uniform. NEC using "high accuracy ground" frequently
underestimates ground loss and can miss VF by a mile. The reasons for this
are not yet clear. And we got no Daddy Warbucks interested in the problem
to pay for the likes of the RCA funded Brown Lewis & Epstein study by
people PAID to keep at it and do it right with equipment and support
provided by their employer.

Thank you, Rudy.

And there is more work to be done.

73, Guy K2AV.