Richard,
With great respect and a smile and as much friendly courtesy as I can
possibly muster via this infernally clueless medium of email. :>)
I am quite jealous of the fingertip access you seem to have to a large
collection of very useful older material, this one no less than any of
the others you have sent my way.
I am publishing this at large, because I have some number of
individuals who want to follow this conversation on the reflector.
This is an answer to prior posts, both yours and theirs.
Your responses are representative of the seeming brick wall I seem to
hit in moving forward with new information and understandings that are
emerging, and how these suggest other 160m antenna configurations.
Although you would seem to think otherwise, I am familiar with all the
points you have referenced. And while I cannot specifically claim
older familiarity with the BLE study you put forward, I did spend a
lot of time some years ago in other parallel work trying to discover
why my multiband vertical was a killer antenna on the Washington DC
copper rowhouse roofs and such complete crap over my Alexandria
backyard sod. My problem is NOT that I have never heard of or don't
understand what you are putting forward.
My problem IS that I know those specifications quite well enough,
drank the kool-aid a long time ago, but am finding irrefutable
contradictions that will not go away in the realm of less than optimum
radials, of sparse and miscellaneous "radial" implementations. And I
am therefore forced to discover the point at which these
contradictions can be harmonized at the edges of the realm of proven
and long demonstrated material.
I will certainly agree with you that *ABSOLUTE* RBN readings on a
single station are just relative readings, subject to endless
influences, and quite useless versus primary field strength readings.
***BUT*** RBN network readings CAN be used to great advantage when
*COMPARISONS* of two or more closely located stations are done
frequently over a period of time and most variables are held in
common. Contest weekends in which these stations participate are ideal
for this.
By comparing geographically close stations with just a few controlled
divergent aspects one can see what the divergences do. A particular
RBN network application graphs the two stations coming into a single
receiver that only records signal to noise and call signs when heard
calling CQ across the weekend in 24 hour chunks, producing as many as
a hundred comparison points. This comparison chart graphing all the
readings allows one to easily visualize the essential permanent gain
separation between the stations. Power level can be factored out, or
the stations can deliberately run the same power. The raw data is
available and regular statistical methods can be run.
The BLE study, taking the two-radial readings at face value without
any elaboration would imply that all else held constant that the RBNs
should indicate a 3.6 dB vicinity of difference between a station with
60 1/4 wave radials and another with two. I have yet to see any RBN
comparison of a two radial on ground vertical come anywhere remotely
close to within 3.6 dB of a 60 radial vertical system. It has so far,
amazingly without exceptions known to me, been of the order of ten dB
or worse, and often QUITE worse, sometimes in the 20 dB order of
magnitude. This explains the anecdotal bad results for two radial
systems. If one has had his 100 watt 160m signal somehow reduced to
ONE watt, we are talking about serious QRP. 100 watts for some
contacts can be a very difficult power level for a QSO with GOOD TX
antenna systems, such as where the far end does not have RX antennas,
is bedeviled by noise, etc. In such a situation a reduction to one
watt means no QSO.
Which is exactly the picture the discouraged anecdotes I hear present.
Good science will harmonize theory, studies AND some degree of
anecdotes, which are studies in miniature, however scattered and
uncontrolled. Most anecdotes are difficult to use because of the need
to re-interview and ascertain additional aspects. That makes them
like trying to plow and plant in stony dirt, a LOT of work to explain
anecdotes. But if one has a ton of anecdotes contradicting some
extrapolation of a study, one ultimately has a problem with the
extrapolation.
I HAVE printed out the BLE study (35 pages, pp 753-787, Proceedings of
IRE, June 1937) and spent time in it. I really am surprised that you
or anyone reading this work would have assigned the work at 2 radials
any significance or considered it in any way definitive. The AUTHORS
themselves clearly did NOT.
Of nineteen graphs with traces for different radial counts, fifteen
have four traces corresponding to 15/30/60/113 radials, with a very
clear purpose to show the advantages of 120 times 0.4 wavelength
radials versus 60, and especially versus- 30 and especially especially
versus 15. Only four graphs include a fifth trace for two radials.
None include separate traces for 4 or 8 radials. Indeed most of those
graphs, 15-30-60-113 radials, give a sense of awful, poor, good,
excellent to those radial counts. Where does that leave two radials?
In figure 19 and its descriptive text, they describe burying "miles"
of number 8 copper at a depth of six inches. The interesting plow in
the picture that they used would have had to be pulled behind a horse
or a tractor. The difficulty of this exercise means that they put
down two radials, measured JUST A FEW of the parameters, dropped it
and moved on directly to 15 radials, ENTIRELY SKIPPING measurements at
4 and 8 radials. The proof of this comes in a Sherlock Holmsian
understanding of text on page 774:
"When only two radial wires, separated 180 degrees, were used, the
resistance [as measured at the feedpoint] was independent of wire
length since the current vanished from the wires within a few feet of
the antenna."
Current confined mostly to the buried radials is the classic
understanding of the efficiency of radials by density and length.
They had a few measurements at two radials at that point which they
kept and put on the four out of the nineteen graphs, BUT DID NOT
BOTHER TO COMPLETE THE SUITE of measurements for two radials. Then
they skipped the measurements at 4 and 8 radials that likely was
originally intended to be a complete suite of measurements at each of
2-4-8-15-30-60-120 radials.
There was not enough of what they
wanted out of the radial study at
2, 4 and 8 to even bother measuring!!!
And here we are many years later trying to use their study, which
SERIOUSLY duns 15 and 30 radials, to legitimize a ham putting down 2,
4 or 8 buried radials.
Why the 113 instead of 120? This is how I know that they started with
two and worked outward. My 95% gut lurch understanding: The 120 was
never staked to start. By the time they got to 60 they had
irregularity in the end spacing, and rather than cram radials, they
skipped a few around the circle to keep the last count as uniform
about the compass as possible.
I know the concepts proffered by the study quite well, with as many
parallel works as there are, and lived by them as gospel for many
years. What I am stating in these posts is that hamdom's common
collection of ideas about small counts of buried or on ground radials
is CONTRADICTED by modern methods, and really not supported by the
original works. You are obviously quite free to disagree, but kindly
don't presume I differ because I don't understand the material.
I will repeat my conviction that the BLE study was really looking to
document the high end radial systems for ground wave, proving in what
became 120 x 0.4 wavelength FCC standard, which was controversial
then, not the given that it is today. They spent almost no time on
low end radials, and certainly no time at sky wave. They deliberately
skipped four and eight radials, instead using 15 radials as the
carefully documented "bad boy" low end choice. Who would ever use two
or four or eight radials for a broadcast station? They proved that 15
and 30 buried radials are clearly inferior. Do we really want to
defend the idea that two buried radials is only down 3.6 dB from 60
radials to that station in Minnesota that could barely work anyone?
When did minus 3.6 dB ever do that kind of TX signal strength damage
to anyone?
Further comment on your comments below.
73, Guy
Nov 8, 2011 at 6:27 AM, Richard Fry <rfry@adams.net> wrote:
> Guy: Below are responses to a few of the statements in your recent post on
> this topic.
>
>>I'll have to counter that the two radial figures in your quoted RCA study
>>are not about defining what goes wrong with minimalist and
> miscellaneous downward extrapolation of dense and uniform radial systems.
>
> The purpose of the BL&E experiments was to determine how buried radial
> systems of various configurations affect the performance of monopole
> radiators of various electrical heights. The measured data in their IRE
> paper speaks for itself. It shows everyone including ham operators the
> effects of using sparse/short radial systems as well as the benefit of using
> dense/long radial systems.
The full study only ran on 15 radials and above.
>>Tests for a given frequency need to be done on that frequency.
>
> The BL&E tests were designed to determine the "efficiency" of monopole
> antenna systems, as shown by the groundwave field they produce close enough
> to the monopole so that the effects of frequency and earth conductivity are
> negligible. The data gathered is applicable to other frequencies for
> antenna systems having the same conductor dimensions in wavelengths.
>From the BLE study, p. 767, emphasis added.
"In the past, experimental curves similar to the theoretical field
intensity curve show in Fig. 1 have been made by maintaining a fixed
antenna height and varying the frequency over a wide range. These
curves have generally been flat in the vicinity of quarter-wave
antenna heights. However, such results have been questionable,
because of the fact that the ground systems become a different
fraction of a wavelength long each time the frequency is changed,
***>>> while at the low frequencies, the attenuation in the first mile
is less than for the high frequencies <<<***. Thus, in planning the
experiments about to be described, it was decided that the frequency
must remain fixed, while the antenna height itself was adjusted. The
frequency of operation was 3000 kilocycles."
Which means that in their experience, for a given transmitted power,
lower frequencies will measure more than higher frequencies at a
distance of one mile, all else held constant. This contradicts your
assertion above and common understandings which deem that loss a
constant across frequencies for a given tract of land. My question
then immediately is that if the H plane field readings at a mile are
NOT constant, then how can one use those readings to tie down sky
wave, other than MEASURING sky wave? Enter RBN stage left and after
all these years we finally we have our shot at it. Or we do if we can
give up our conviction that was all settled in the 30's and 40's.
>>Sky wave behavior is ASSUMED to be the same as at the ground. This is risky
>>because of the loss mode presumptions used with the ground based field
>>strength. ... And that is not a rub on this RCA study. All these studies
>>have this blind spot, as if measurement at the ground is in it's own world
>>and seemingly unrelated to skywave.
>
> Unloaded, uniform cross-section monopoles of all physical heights have close
> to sinusoidal current distribution along their heights, with essentially
> zero current at the top. For monopoles of 1/4-wave height and less this
> produces an elevation pattern having a relative field equal to the cosine of
> the elevation angle: zero field toward the zenith, and maximum field in the
> horizontal plane. That pattern shape does not change with the number/length
> of buried radials used with the monopole, or the conductivity of the earth
> around the antenna site. Therefore the h-plane field measured close to such
> a monopole (as done by BL&E) also leads to knowledge of the radiation at
> angles above the horizontal plane.
This is going to be the point where the extension of commercial ground
wave work breaks down for sparse and miscellaneous "radials", because
there is no old data to validate sky wave. All the validation work has
been done at the ground, and the accumulation of RBN data
overwhelmingly points to something additional at work at sky wave for
antennas not approaching commercial standards, something not
encountered or measured in the old studies.
The BLE study is all about validating what we know today as the FCC
standard ground radial field for commercial MF BC stations. Any use of
the study for less than 15 radials is unwarranted at ground wave, and
what remains is unvalidated for skywave.
I don't see anyone back then connecting all the dots between right at
the antenna H plane readings and field strength readings for sparse
radials and sky wave. I don't think they had the means, and it was
not possible for them to have this discussion. In our time we need to
have either unlimited helicopter resources, OR have RBN. We finally DO
have RBN.
>>There are just too many grotesque differences as measured on RBN, even
>>after all the various (and rightly pointed out) caveats are dealt with.
>
> RBN is more a measure of skywave propagation conditions than the radiation
> patterns of the antennas used. About the best that can be done using RBN
> for antenna evaluation is by relative comparison for the same transmitted
> power, modulation, carrier frequency, propagation path, date, and time of
> day.
That's how we have done our comparisons. And "about the best" is more
than enough to discredit old assumptions for sparse and miscellaneous
radials. By knowing the stations, we can pair off stations for the
same transmitted power, modulation, carrier frequency, propagation,
path date, and time of day, PLUS we can then graph these hundreds of
observations for an entire contest weekend. It's skywave only (don't
care a rotten rat about ground wave), and we throw in a correction if
the models of the antennas suggest that there is a significant
radiation pattern difference in the direction of the RBN RX station.
IF the RBN differences were subtle, I would not be writing this and
would have moved on to basket weaving or something. The RBN
differences are STARK. There just isn't enough "stark" on those old
studies. They weren't serious about low end radials. We ARE.
>> More radials is better is no urban myth. Dense and uniform really does
>> work the best. But two askew radials often seems to do orders of magnitude
>> worse than such material as the RCA study would suggest. One 100w station
>> in Minnesota went from barely working and being heard in North Carolina,
>> to working Europe, just by picking up his two 1/4 wave radials off the
>> ground, and folding them back elevated to the feed. The old stuff is
>> simply not accounting for true anecdotes like that.
>
> Two radials of any length, lying on or buried in the ground have high r-f
> loss when used as the r-f ground for a monopole (see the BL&E paper).
So why do the BLE graphs of measurements for 2 vs. 60 radials at one
mile only show 3.6 dB difference? I agree that two quarter-wave
radials, even ABOVE ground are non-intuitively seriously lossy in all
but the best circumstances.
> The data in that BL&E paper was limited to buried radials, but that doesn't
> mean that the effectiveness of using several symmetrically arranged,
> elevated horizontal wires as a counterpoise (instead of using buried wires)
> was/is unknown. For example the 1952 textbook RADIO ANTENNA ENGINEERING by
> Edmund Laporte has several pages describing them. Such systems have been
> installed by a few AM broadcast stations in the US, and their performance is
> equal to that of a monopole of that height using 120 x 1/4-wave buried
> radials. NEC software shows this, as well.
NEC software, including the NEC4 variety, uses an approximation method
that is adjusted for ground wave and ground field strength
measurements. In other words, the models were adjusted to GENERATE
the ground measurements, not validate them.
_______________________________________________
UR RST IS ... ... ..9 QSB QSB - hw? BK
|