Depends.
You could just keep adding radials when you can afford more copper until
things stop improving.
(Whatever "stop improving" means to you.)
Copper's expensive.

My SteppIR vertical is on an aluminum roof.
(Just my approach to the problem)

And Walt's right.
What is trivially obvious to us wasn't so in 1937.
Maxwell's equations weren't 100 years old yet.
It had only been a few years since Gibbs wrote them in the modern form we
use.
Radio was barely understood by only a few people.

73
H.

NQ5H
"Wayne" wrote in message
news:L0Nwg.5924$yN3.4270@trnddc04...
These are very good points. I am reading these postings to try to
understand the behavior of actual implementations that lie somewhere
between the extremes you pointed out. In other words, what gets you the
most bang for the buck.... How fast does performance change with
increased radial length and number of radials.


"H. Adam Stevens, NQ5H" wrote in message
...
I'm often confronted with problems as a physicist where one can only get
a handle on upper and lower bounds.
Lower bound:
I'd say the minimum number and length of radials is 3 (must define a
plane) and 1/4 wavelength (satisfies boundary conditions).

Upper (infinite sheet of copper)
As Walt and Reg have debated, the "Cleese extreme" (to steal from Reg's
post) is trying to duplicate the "infinite perfectly conducting plane" of
our elementary physics books.
Cheers and beers
H.

73, NQ5H

On Sun, 23 Jul 2006 11:35:42 -0500, "H. Adam Stevens, NQ5H"
wrote:

Depends.
You could just keep adding radials when you can afford more copper until
things stop improving.
(Whatever "stop improving" means to you.)
Copper's expensive.

My SteppIR vertical is on an aluminum roof.
(Just my approach to the problem)

And Walt's right.
What is trivially obvious to us wasn't so in 1937.
Maxwell's equations weren't 100 years old yet.
It had only been a few years since Gibbs wrote them in the modern form we
use.
Radio was barely understood by only a few people.


Hi OM,

In fact, how "many" people knew is immaterial to what was known a good
twenty five years before the BLE paper.

From my "Standard Handbook for Electrical Engineers," 1912,
Sec. 21, Radiotelegraphy, Method of Exciting the Antenna, part 283
Antenna Ground Connections:
"The outward and inward movement of the lines of electric force
during the oscillations in the antenna give rise to earth
currents. These earth currents are most intense in the immediate
neighborhood of the antenna, and if the earth is a poor conductor
a large waste of energy ensues. To guard against this loss, a
radiating network of wire is place beneath and around the antenna.
In the case of a flat-top antenna, the radius of this wire net
should not be less than the length of the horizontal portion of
the antenna."

I shouldn't have to point out that a handbook is not the place where
new science appears, but where tested science is aggregated. Earth
currents, screens, and lost power were not unfamiliar a century ago.

What is "Bleeding obvious" about the BLE paper, is that it puts
numbers to the quoted paragraph above in the face of its mocking:
At risk of upsetting a great number of patriotic USA citizens, all BLE
hero-worshippers
It is quite evident that the merit of the BLE paper serves the true
spirit of Lord Kelvin, and that in the context of this group, it is
USA citizens who honor his precepts in the face of this last piece of
British trolling of Reggie's who is more interested in juvenile
posturing than celebrating his heritage's expression in a fine work.

73's
Richard Clark, KB7QHC

Reg Edwards wrote:
"H. Adam Stevens, NQ5H" wrote

I'm often confronted with problems as a physicist where one can only

get a

handle on upper and lower bounds.
Lower bound:
I'd say the minimum number and length of radials is 3 (must define a

plane)

and 1/4 wavelength (satisfies boundary conditions).

Upper (infinite sheet of copper)
As Walt and Reg have debated, the "Cleese extreme" (to steal from

Reg's

post) is trying to duplicate the "infinite perfectly conducting

plane" of

our elementary physics books.
Cheers and beers


==========================================

Yes Adam, a logical way of looking at it.

Associated with any number there is always another number which is
sometimes, but not often enough, used to describe its uncertainty.

But nearly always it takes much longer to determine the uncertainty
than it does to arrive at the first number, especially if the first
number is the result of a measurement.
----
Reg.



What is the uncertainty of the uncertainty? If the uncertainty is a
number, then, "Associated with any number there is always another
number which is sometimes, but not often enough, used to describe its
uncertainty." You're going to end up with an infinite string of
uncertainties if you keep this up, Reg.
73,
Tom Donaly, KA6RUH (Who never could understand Sartre.)

"Wayne" wrote
understand the behavior of actual implementations that lie somewhere
between
the extremes you pointed out. In other words, what gets you the
most bang
for the buck.... How fast does performance change with increased
radial
length and number of radials.
===================================

That's exactly what program RADIAL_3 is intended to help you with.

All the user has to do is convert radiating efficiency into bucks. The
higher the radiating efficiency the more bucks it will cost, the
greater the length of wire, and the more painful the back ache.

It's a matter of diminishing returns.
----
.................................................. ..........
Regards from Reg, G4FGQ
For Free Radio Design Software go to
http://www.btinternet.com/~g4fgq.regp
.................................................. ..........

"Reg Edwards" wrote in message
...

"Wayne" wrote
understand the behavior of actual implementations that lie somewhere
between
the extremes you pointed out. In other words, what gets you the
most bang
for the buck.... How fast does performance change with increased
radial
length and number of radials.
===================================

That's exactly what program RADIAL_3 is intended to help you with.

All the user has to do is convert radiating efficiency into bucks. The
higher the radiating efficiency the more bucks it will cost, the
greater the length of wire, and the more painful the back ache.

It's a matter of diminishing returns.
----
.................................................. .........
Regards from Reg, G4FGQ
For Free Radio Design Software go to
http://www.btinternet.com/~g4fgq.regp
.................................................. .........



So if you calculate it in Pounds, is the back ache worse?
;^)

73
H.

Thanks Richard...however the info you present doesn't deal with the real
issue that has been raised by Reg's program: (you are adressing another area
of possible disagreement)

Do 26 radials 5 metres long perform substantially as well as 26 radials 16
metres long at 3.62 mhz with the two soil constants = 25 (in my case), with
radial wire size #14 and antenna wire size #10 (I think I used 2mm and 4mm
in my calculations.)

The issue has never been (for me) how closely does Reg's program match a 1/4
wave vertical with standard length radials. The "real" question is does
Reg's program accurately reflect the performance of vastly shorter radials
than the typical "wives' tale" (Reg's characterization) 1/4 wave length
radials.

I have always accepted that some shortening of earth based (on or under)
radials (below the assumed 1/4 w or longer as in BL&E) was acceptable. The
problem is, Reg's program allows incredible shortening, predicting high
efficiency at the same time.

I have a 1000' of wire left to put down. If Reg is right, I can put down 66
radials 5 metres long and get outstanding surface coverage. If more
"orthodox" texts are correct, then I should stick with 16 metre (approx 50')
length and then I can only put down 20 more radials than the 26 I have now.

Let's not get distracted (although your point for the 20 degree antenna
shows yet another departure from BL&E).

Here's the fundamental contradiction between Reg's program and the orthodox
approaches:

================================================
Reg says (given the values I have stated above), that 5 metre long radials
will peform (substantially) as well as 16 metre long radials, all other
things being equal. Thinking this makes my head hurt.
================================================

BL&E seems to contradict this (although I haven't found the precise
comparison I'm looking for yet).

Tom, W8JI's, measurements seem to contradict this.

NEC-4 should be able to tell us how much current is in a radial and how that
current is distributed along the length of the wire. If it disappears into
inconsequential levels within the first 5 metres, then it confirms Reg's
assertion. If it does not, i.e., it remains at substantive levels well
beyond 5 metres, then it contradicts Reg's program, and agrees with BL&E, as
well as W8JI. (I thought NEC-4 could do this problem, maybe my assumption is
completely wrong.)

================================================== ============
If NEC-4 can't do this "current along a radial" analysis (buried or on the
surface, take your pick), then we need experimental data that shows us the
same thing: how fast does the current along a radial decrease to
inconsequential levels. If it is within the first 5 metres, Reg is right. If
not, he's wrong. It's as simple as that.
================================================== =============
Why is it so hard to get this answer?

hasan, N0AN


"Richard Fry" wrote in message
...
"Reg Edwards" wrote
"Richard Fry" wrote
N. B. for/to REG EDWARDS (G4FQP): I hope that you will be
motivated to follow through on one or the other of these offers,
and that you will post a comparison of the results of your
ready-to-run, "radial_3" DOS program as compared to the
BL&E datum, for equivalent conditions.
= = =
What equivalent conditions? Where can they be found? What was the
ground resistivity and permittivity on BL&E's site?

I am not motivated to do anything except reply to your remarks.
...
Reg.
_______________

OK, I'll do it then. Attached is a plot of BL&E's numbers versus yours,
for
the conditions stated there. Ground resistivity and permittivity were
estimated using the FCC's M-3 chart to select values of R and K at the
BL&E
test site from those shown in your program.

Other parameters for radial_3 calculations were taken from the physical
and
electrical descriptions in the BL&E paper. The OD and depth of the
radials
were estimated.

You and BL&E agree fairly well for a 90 degree vertical, but not well at
all
for a 20 degree vertical.

I'll be glad to explain how I generated my plots, and even send you the
spreadsheet, if you want.

RF

"hasan schiers" wrote ...
Do 26 radials 5 metres long perform substantially as well as
26 radials 16 metres long at 3.62 mhz with the two soil
constants = 25 (in my case), with radial wire size #14 and
antenna wire size #10 (I think I used 2mm and 4mm in my calculations.)

Some insight into this is provided by the following.

Figure 37 in the BL&E paper shows about 77.5% of the theoretical maximum
inverse field for a 77 degree vertical used with 30 radials of
0.137-wavelength each.

Figure 36 shows 90.6% of the theoretical maximum inverse field for the same
vertical with 30 radials of 0.411-wavelength.

For 30 each 0.137-wavelength radials and a 77 degree vertical, Reg's program
radial_3 calculates an efficiency corresponding to 87.2% of the theoretical
maximum inverse field. (Other parameters for radial_3 were as shown in the
plots I posted earlier.) This is significant when considering that the
inverse field varies by the square root of this power difference.

NEC-4 should be able to tell us how much current is in a radial
and how that current is distributed along the length of the wire. If it
disappears into inconsequential levels within the first 5 metres,
then it confirms Reg's assertion. If it does not, i.e., it remains at
substantive levels well beyond 5 metres, then it contradicts Reg's
program, and agrees with BL&E, as well as W8JI.

BL&E data show that if few radials are used they may as well be "short,"
because system performance isn't improved greatly by making them much
longer. Quoting the BL&E paper (p.760), "These diagrams show that the
ground system consisting of only 15 radial wires need not be more than 0.1
wave length long, while the system consisting of 113 radials is still
effective out to 0.5 wave length.

But there was no experimental evidence from BL&E showing that radiation
efficiency ever _improved_ with shorter radials, as apparently calculated by
radials_3..

RF

Wayne,

The best study I've seen is in both the ARRL Antenna Handbook and in ON4UN's
Low Band DX'ing Handbook. I think it was a 3 station that came up with a
method (consistent with BL&E) that gave a simple formula for putting down
the optimum number and length of radials, for a given length of radial wire
available. I used that study to originally arrive at 50' long radials at
80m. This gave tip to tip separation of about 3 or 4 feet on 80m, which met
his criteria. The material above specifically answers the question: how do
you get the best bang for the buck for a given amount of available radial
wire.

Read that material..or at least get the formula and apply it to your
available wire...that will get you were most of us are with respect to
optimizing radials.

Now, Reg has come up with his program that flies in the face of these other
studies, indicating one can obtain comparable performance with MUCH shorter
radials (5 metres instead of 16 metres) and that is what started this whole
thread. We await some sort of comfirmation from several sources that Reg's
numbers are correct. If they are, Reg will become famous.

Currently here is how things line up:

1. BL&E doesn't seem to agree with Reg's numbers (on the issue of short
radials)
2. Tom, W8JI's, recollection of his measurement don't either.
3. NEC-4 is in the process of analyzing the short radial comparability claim
as we speak.

The entire issue is: does the current in the radials described above taper
off as quickly as Reg predicts, or not? If it does, the short radials will
be comparable and Reg is right. If it doesn't, Reg needs to fix his program
in that particular section.

We await more data, or someone to extract from BL&E a precise answer to the
actual question: how fast does current fall in a radial as you move away
from the base of a 1/4 wave ground mounted vertical with shallowly buried
radials.

In the mean time, you can get started with the formula I referred to above.
If Reg is right, you used more wire than needed. If not, you have your wire
in place and are ready to go.

....hasan, N0AN
"Wayne" wrote in message
news:L0Nwg.5924$yN3.4270@trnddc04...
These are very good points. I am reading these postings to try to
understand the behavior of actual implementations that lie somewhere
between the extremes you pointed out. In other words, what gets you the
most bang for the buck.... How fast does performance change with
increased radial length and number of radials.

On Mon, 24 Jul 2006 08:32:08 -0500, "hasan schiers"
wrote:

The best study I've seen is in both the ARRL Antenna Handbook and in ON4UN's
Low Band DX'ing Handbook. I think it was a 3 station that came up with a
method (consistent with BL&E) that gave a simple formula for putting down
the optimum number and length of radials, for a given length of radial wire
available. I used that study to originally arrive at 50' long radials at
80m. This gave tip to tip separation of about 3 or 4 feet on 80m, which met
his criteria. The material above specifically answers the question: how do
you get the best bang for the buck for a given amount of available radial
wire.

You can find the article and the formula in the August 2004 QST.

Danny, K6MHE

On Mon, 24 Jul 2006 08:47:36 -0700, Dan Richardson
wrote:

You can find the article and the formula in the August 2004 QST.


OOOPS! Make that August 2003!

Sorry,

Danny, K6MHE