NEC-LIST: CFA Measurements Seem Required to Settle the Controversy

All, but especially Jack Belrose and Dick Adler:

I was disappointed not to make ACES this year, but a report still due
kept me here in WDC (where it suddenly is spring!). I would have
enjoyed Jack's talk, and have enjoyed his trials and tribulations in
modeling of the Cross-Field Antenna (CFA), as shared with this forum.

I first heard of the CFA quite a few years ago, from a client who
thought it sounded too good to be true. I discussed the CFA briefly
with Dick Adler back then, and on occasion when this controversy
heated back up, on several occasions (to the extent that I have
considered the CFA "the antenna that wouldnt die"). I didnt do any NEC
modeling of the antenna, but I did consider it in some detail from the
standpoint of the claims vs the fundamental limitations of
electrically small antennas as documented in the IEEE and other
literature by Wheeler, Walters, and others. Here is how it went, in
the simplest version:

1) The antenna qualifies as an electrically small antenna. I dont
think anyone disputes that, and it is always good to have a safe
starting point on this type of discussion.

2) If it has any efficiency at all it will have to be a high-Q
antenna.

3) If it is high Q, it will be difficult to accurately measure its
impedance and to feed it properly.

4) If it is not fed properly (which includes some rather critical
phasing that is hard to deal with in analytical modeling vs guidance
when you have to actually solder something exactly the right length),
it wont accept the RF power being fed to it, presumably through a
transmission line.

5) The power reflected will return to the source (or at least try to)
both through the inside of the coax (if that is the type of
transmission line with which it is fed) and RF currents will (of
necessity) also be on the outside of the outer conductor (and, as Jack
noted, probably elsewhere as well).

6) The antenna wont radiate much, like many other very small
(electrically) high-Q antennas, but the feedline may be a half-decent
radiator, depending on its length and orientation relative to the
ground.

7) The feedline currents may not be in the right orientation to the
ground to launch a very effective vertically polarized ground wave
(i.e., Norton surface wave). But they will launch most of what is
launched that is observed at a distant receiving location. Reradiation
by nearby other structures that may have induced currents with a
vertical component may increase the effective strength of the
far-field vertically-polarized groundwave.

I have now long forgot who asked me first about the CFA, and it no
longer matters. But I have consistently said "dont hold your breath"
until it is an efficient radiator for MF (or HF or VHF) terrestrial
broadcasting. I did suggest that anyone interested, who didnt have
Jack's or Dick's (or Gerry's) skills with the modeling, could purchase
one, and have the manufacturer set it up and specify that it was set
up and fed "properly". One then could feed it with RF power and
measure the input power and the reflected power (all of it) and
compute thereby the net forward "transmit" power. The
vertically-polarized electric field strength produced could be
measured with a calibrated field strength meter as a function of
distance. This could be done more or less in line with the guidance in
IEEE Std. 291-1991 (which we currently are revising to bring it up to
date). I realize that here I am stepping over into both IEEE AP-S and
PROP-LIST territory, but that is required to solve the still-current
CFA conundrum (in my opinion). The propagation (i.e., decrease in
electric field strength with distance) can be modeled easily with the
appropriate Sommerfeld integral equation, after determining the proper
ground constants for the path (presumably over level, flat
terrain). Once this modeling has been completed, and matched in
absolute terms to the effective radiated power (ERP) required to
establish the now-documented field strength along the path, one can
compare the ERP and the net input power and (by subtraction) obtain an
estimate of the "gain" of the CFA for a groundwave path. I put gain in
quotes because the definition of antenna gain is under scrutiny in the
IEEE Antenna Standards Committee (IEEE Std. 145 revision effort, which
I chair). I also put it into quotes because of recalling Dr. Helmuth
Bruckmann, then of CECOM, Ft. Monmouth, NJ, reprimanding me about the
use of the term gain in the context of the Norton surface wave. He
suggested I frame the problem in terms of the antenna current and the
ground constants and the electric field produced (a la Jim Wait)
rather than in terms of gain (a la Ken Norton and basic transmission
loss). But one can still perform the following computation using the
measured net input power, Pin (in dBW) and the ERP = Pin * Gt (in dBW)
to solve for Gt (in dBi):

Gt (dBi) = ERP (dBW) - Pin (dBW).

I believe that one would get, if this were done, something like maybe
Gt = - 15 to - 25 dBi. Of course, if we got more out than we put in,
we should be off to the patent office without delay. But my main point
is that this experiment could be performed with the cooperation of the
manufacturer and someone who would donate a nice flat open site of the
right size. The measurements should be made by a competent
non-stakeholder in the presence of competent witnesses. And if the
antenna does produce the field strengths at the required distances for
successful broadcasting, then that is how it turns out. If not, then
that is how it turns out.

But the modeling my suggestion requires is NOT dependent on anything
that has not been proven over and over again, right Charlie Burrows,
Ken Norton, et al.? We just have to get the sign right in the
Sommerfeld integral, which is no problem due to the dilligence of that
pioneer of our trade, Ken Norton. And that consumate experimentalist,
Charlie Burrows and his partners and the lake propagation experiment
many years ago showed how the electric field died off vs distance.

All the broadcast listener really wants is a good sounding signal
(i.e., good SNR without too much distortion). WYHIWYG (what you hear
is what you get), at least for the audio broadcaster and his clients.

Anyone have the patience to try to follow the above? If so, comments
are welcomed.

George

P.S. I have always found a need for vertical plane aperture (in
wavelengths) to get any vertical beam compression (and any related
gain increase at lower take-off angles). I dont see how that can be
achieved with an antenna whos' selling point is that it is small in
vertical aperture, in contrast to its competitors, the current crop of
MF broadcast antennas. The SRI RELEDOP full-scale pattern measurement
system (see Stehle, et al., IEEE Trans Broadcasting, June 1988) can be
used as low as the AM broadcast band to measure the pattern as well as
the ERP (in dBW) vs theta and phi of a transmitter and antenna
combination, but I believe that the ground-based measurements
suggested above should suffice. RELEDOP would document the actual
vertical-plane pattern, though, even if there is a little occasional
beam lift on the test frequency. Right, Dick?

George H. Hagn
tel:(703) 247-8470
fax:(703) 527-3087
Hagn_at_wdc.sri.com
Received on Thu Apr 01 1999 - 03:53:27 EST