Greetings, to all those interested in the above subject.
Since I have been carrying the ball, I must continue to run with it.
I have been corresponding with Brian Stewart, Jerry Burke, and this
morning I received a note from J.L. Smith re the CFA Patent (which he
is sending to me); also a note from Clifford Craft (I am not sure
whether he is on the NEC-List).
As Jerry said (in this mornings e-mail addressed to me) there are
certainly some strange ideas going around wrt to the CFA.
This note is based on my view whuch reflects my past experience, on my
CEM analysis of the CFA, and, it begins with direct paraphrasing from
notes sent to me by Jerry and Brian. I hope they will not get upset.
Brian wrote:
"The reason why NEC does not produce radiated power. John, we have
been telling everyone from the beginning that the major difference
between a CFA and a conventional antenna relates to the means of H
field production. The D plate is a capacitor carrying a voltage on it
thus the existence of E field lines between the D plate and
ground. When this voltage oscillates at high frequency then time
varying E field (and thus time varying D (= epsilon times E) exists
between the plates). From Maxwell's 4th equation D' = curl H, the
capacitor therefore generates a magnetic field around it. This
magnetic field can easily be evaluated from application of Stokes'
Theorem resulting in H = I/(2.pi.r) (standard H field continuity
across capacitors) where I is the current flowing into the D plate and
r the distance from the centre of the D plate capacitor to a
horizontal location outside the D plate. This can also be expressed in
terms of the Voltage or D field and Capacity of the D plate and also
frequency if desired. However this H field is not the H field
generated from the current in the wires feeding the D plate (Belrose:
this is very very small for my model). The two are different sources
of H field. The H field from a small wire of length d (separation of D
plate and ground plane!) is approximately H = I d/(4.pi.r.r) assuming
that d < r, a situation which exists in the CFA. My rough back of the
envelope calculations thus indicate that any H field from the D plate
is much= greater than H field from the short wire feeding the D
plate. So the H field from the D plate cannot be ignored (agreed)."
And, he continues: "John, does the NEC code evaluate the H field
vector distribution from the D plate time varying voltage? That is the
crux. Not only must it include H field from wires it must include H
field from the capacitor plate (Belrose: my capacitor plate is a wire
grid). John if the NEC code does not include this scenario could you
help me understand why this is so. If it does could you show me where
to find it and how it is implemented in the H field evaluations? I am
genuinely seeking to establish whether the NEC codes can indeed cater
for the CFA".
"Burke's reply (to me):
"NEC is completely consistent with all of Maxwell's equations. It
computes both E and H fields from the electric currents. They are the
only source there is, since we do not have magnetic currents. The H
produced by time varying E comes out of Maxwell's equations.
Evaluation of near E and H from the same set of currents using the
standard Green's functions (which NEC does) will guarantee that E and
H are consistent with Maxwell's equations (even if the boundary
conditions on the antenna are not satisfied too well)".
"The term H=I d/(4.pi.r.r) in Brian Stewart's message is a part of the
near H field of a current element. If the total near H = I d/(4
Pi)(2/r^2 + j k/r) with r=Sqrt[rho^2+z^2] is integrated along a wire
with constant current over an infinite wire or length >> rho, the
result will be H=I/(2 pi r) for k*r <<1. That is all in Maxwell's
equations and NEC. A small length of displacement current will not
produce any more circling H than the same small length of current on a
wire."
If someone says an antenna works, it is difficult to prove otherwise
without careful measurements duplicating their situation, or a
complete CEM analysis including the environment (elevated ground mat,
unused tuned or untuned nearby towers, feeder transmission lines,
etc.). Burke continues: "But we can say with 100% certainty that
claims about evaluating H due to voltage or varying D is wrong. The
relation of H to varying D and the relation of E to varying B is
contained in Maxwell's equations and is exactly satisfied in the NEC
solution. The only source terms in Maxwell's equations are electric
current (and magnetic current if you can find any) and that is what
NEC uses. Of course on the plates the large D is produced by charge,
but that is related to current through continuity and is taken into
account by NEC".
"The D plates have a conduction current going up and a displacement
current going down. That is not going to be a good producer of H.
Despite what Brian Stewart seems to think, conduction current is just
as good a producer of H as displacement current is. With no net
current there will not be much H. Cite: the Stratton-Chu formulas for
the fields, derived from Green's Theorem (Stratton, Equations (19) and
(20), page 466). They give the total E or H field in a closed region
as an integral over the currents in the region (electric and magnetic
if any) and an integral over the fields on the surface of the region.
In NEC the surface of the region is at infinity, so all currents
contribute to the fields. You do not and cannot include fields (or
voltages) as sources in the region bounded by infinity."
With that background let me continue giving detail based on my CEM
analysis. I now have a model of the CFA antenna which I am satisfied
with, and, using the program facilities I have I cannot improve upon
it.
The cylinder is a 20-sided polygon; each sector is broken up into 8
rectangular grids. The circular disk has radial symmetry, by dividing
the circle up into 32-sectors with 6 irregular rectangular grids in
each sector (the grids become smaller as we approach the center of
this wire grid disk).
Brian has told me that the CFA has very strong near E-fields, and
(according to what we read about the Egyptian antennas) the far fields
are comparable to or greater than produced by the previous used 75 m
tower over an extensive ground system (the Hatley et.al ICAP91 paper).
I told Brian that I would expect very-very strong near E-fields, since
the impedances and circulating currents (and powers) seen by the
sources feeding the D-plate and the E-plate (the cylinder monopole)
are almost unbelievably wild and high. But the far field (radiation
field) according to NEC is very low.
The impedance seen by the source feeding the cylinder, a very-very
short monopole (reactance -j715 ohms canceled, inductor Q factor 300)
is 187 ohms; the forward circulating power for an applied power = 1000
watts is 23,610 watts (!!!); source currents 9.2 amperes.
The impedance seen by the source feeding the disk (reactance -j477
ohms canceled) is -177 ohms; return circulating power 14,970 watts
(!!); source current identical 9.2 amperes.
To compare with measurement, data yet to provided, we have to assume
that we can feed the antenna.
So does this antenna radiate? It must radiate something because of
all that circulating power. In what follows, rather than quoting
calculated E-fields (dB microvolts/m), I compare the E-field for CFA
minus the E-field for the conventional monopole (a 75 m tower), power
difference in dB. The frequency is 1161 kHz.
____________________________________________________
Distance CFA antenna compared with
(meters) the 75 m Monopole
____________________________________________________
3 + 29.5 dB
5 + 22.8 dB
10 + 12.6 dB
20 + 0.5
60 - 16.6
100 - 17.9
200 - 18.8 dB
____________________________________________________
Jerry has suggested I compute H-fields, and so I will do this --- but
from the point of view of power radiated we conventionally want to
determine the unattenuated (by ground loss) FS at 1 kilometer.
The field at 200 m is dominantly in the region where FS decreases by
1/r (the radiation field), and so the power radiated by the CFA
antenna I analyzed is 10.8 watts (radiation efficiency 1.1 percent).
Clearly, the CFA antenna is a most interesting antenna to model. The
impedance of the cylinder monopole by itself is 0.134 - j 727 ohms;
the impedance of the disk by itself is 0.016 - j 480 ohms. Compare
these impedances with the impedances seen by the sources feeding the
CFA.
Now where are we concerning availability of measurements and new
information on the inventor's CFA? Alberto Fassio, an engineer with
RAI Radio Televisione Italiana, has purchased the flared version of
the Egyptian CFA, but he is currently having problems with the local
council which is delaying the installation phase. However we can
expect some measured data eventually. And, this week, a number of my
correspondents (on the NEC-List) are attending the NAB Convention in
Las Vegas, NV, where they will hear an up-date on the CFA antenna told
by two of its inventors, Stewart and Kabbary.
Finally, Brian Austin, Liverpool University, has told me that at the
recent IEE Conference on Antennas and Propagation he met Hately.
Hately did not present a paper but he did tell Austin that one of his
antennas is soon to go into service (???) as the main antenna for a
longwave broadcast station on the Isle of Man. How or why this
decision was made was not divulged. So perhaps we will hear something
about that station. A great signal, or no signal at all in the
service area (??).
If the CFA is a good efficient wideband radiator as claimed by its
inventors, then my analysis to date makes no sense; and we will have
to change completely our current views wrt antenna fundamentals.
In my view the CFA is a narrow band inefficient radiator, having
very-very strong near fields (do not touch it or for transmitter
powers of 100 kW go near to it); and so if the CFA antenna radiates
efficiently it must the radiating environment in which the antenna is
operated (recall my modelled CFA at a low height 0.6 m over a PEC
ground, and it is all by itself).
I will be glad to FAX my detailed analysis, to anyone wanting to
pursue CEM modelling the antenna by a different method (I have used
EZNEC/4 double precision); or to anyone wanting to help us interpret
the results of my analysis.
Regards, Jack
15 April 1999
_____________________________________________
John S. (Jack) Belrose, PhD Cantab, VE2CV
Senior Radioscientist
Radio Sciences Branch
Communications Research Centre
PO Box 11490 Stn. H
OTTAWA ON K2H 8S2
CANADA
TEL 613-998-2779
FAX 613-998-4077
e-mail <john.belrose_at_crc.ca>
_____________________________________________
Received on Thu Apr 15 1999 - 18:46:34 EDT
This archive was generated by hypermail 2.2.0 : Sat Oct 02 2010 - 00:10:39 EDT