cfameas.jpg (31419 bytes)Introduction
cfaf.gif (1194 bytes)rom several accounts we have heard that CFA antennas have been deployed and/or are planned, in several locations around the world, namely:

The Egyptian deployment, so far has yielded the most data. We have some Australian data, and we await data from other commercial sources [9,10,11].

The co-inventors: Mr. Hately, Dr. Kabbary, and Dr. Stewart [2,3,4,5,7] make promising claims for this new form of antenna. Despite the seemingly successful use at several sites [9,10,11], the antenna is shrouded in controversy [5,11,12], not the least of which is caused by the incredible performance claims versus the relatively small antenna size.

The "theory of operation" [2,4,7], stated by the co-inventors, is also controversial. Mainstream academic researchers do not agree with the form that the theory takes. However, it is perhaps the intent of the theory that is important. The inventors state that the CFA generates its radiation by direct Poynting Vector synthesis in a zone of interaction, near the antenna.

The CFA is composed of two components that interact to produce the zone of interaction. The first is the E-plate which principally generates E-fields in the near zone (ratio of E / H > 1) and the second is the D-plate, which is stated to produce H-fields (ratio of H / E > 1) in the near zone. According to the Hately-Steward-Kabbary theory, the D-plate must be fed in phase-quadrature with the E-plate so the correct time and spatial relationship exists to generate the Poynting vector (S = E x H). In general terms this is not implausible and the analysis involved in proving or disproving this statement would be quite complex and beyond my means.

Measurements which have been performed for the D-plate have shown that the H field is generated and extends outside the radius of the D-plate, see reference [7 II] The generation of radiation by direct Poynting vector synthesis, if it does occur, is likely to be complex as there are several source of E-fields and several sources of H-field and the direction of these fields is also complex. Neither the E-plate would exclusively generate an E-field, nor would the D-plate generate an exclusive H-field.

Subsequent to the first deployment of the CFA in Egypt, the CFA has been modified to include a conical top-hat which is connected to the E-plate. Dr Kabbary states that the conical section improves the zone of interaction resulting in improved performance at lower elevation angles.

This article will relate the measured performance of the CFA against the theoretical performance of a point source for the same input power. We are not seeking to explain, but just to elucidate.

Antenna Performance
Antenna analysis shows us that the performance from an antenna is measured in terms of its efficiency and its gain. In order for one antenna to perform better than another antenna it must have a higher efficiency or comparable efficiency and gain over the other antenna.

Antenna gain is traditionally derived by the following means:

  1. Size: As a linear antenna is increased in size from very short up to approximately 5/8 of a wavelength, the forward linear gain increases from 1.5 to 3.12. With longer lengths the radiation pattern breaks-up and no-longer looks like a dough-nut and contains multiple lobes with energy no longer being directed in the main direction perpendicular to the antenna element. For example, long wires exhibit gain and the longer they get the more the beam converges towards the axis of the wire.
  2. Stacking: Antennas are stacked and fed from the same source to form arrays of multiple sources. If you use two sources you can obtain a linear gain of up to 2 over the single source.
  3. Parasitic Elements: Parasitic elements can be introduced to provide gain, these elements absorb and re-radiate energy and act like multiple sources. Parasitic elements form the basis of the Yagii-Uda antennas. Elements may be close-spaced to obtain super-gain antennas, or wide-spaced to obtain traditional array gains.
  4. Reflection: If an antenna is located over a perfect reflector then gain can be achieved by means of the image formed by reflection. In the case of a vertically-polarized antenna (E-field vertical to earth), the antenna can achieve up to 6 dB gain from the co-location with the ground. Some of this gain occurs because the antenna is no longer permitted to radiate its power into a full sphere, but instead, it is radiating into a hemisphere (the same power in half the space), the remaining gain is derived from the virtual source, an apparent improvement of 3 dB. This means that a vertical antenna near ground can have a field strength which is up to 2 times greater than the field-strength in free-space—this is a 6 dB improvement, or 4 times the apparent power.
  5. Area: Antennas with a large capture area exhibit large gain. This is evidenced by parabolic antennas deployed in microwave bands. The larger the aperture, the more the gain and the narrower the radiation beam.

As the isotropic radiator is the ideal source, antenna gains are quoted in reference to the isotropic (point) source, hence the term dBi. The gain of an antenna over an isotropic source is brought about by the modification of the antenna's radiation pattern. Higher gain antennas (including arrays, etc) have narrower radiation patterns than the isotropic, i.e. their power is being directed in some directions but not in others. For example, a dipole has nulls in the radiation pattern above each end, while a monopole has a null towards the zenith and no radiation below the ground level, and microwave dishes have very narrow beams.

Transmitting antenna performance may be characterized as the Field Strength at a specified distance with a specified input power. In Imperial systems we arrive at so many milli-Volts per meter (mV/m) at one mile as the standard, whilst in the SI units we arrive at (mV/m) at one kilometer. Traditionally the technique involved an input power of 1kW.

The next section shows how to relate the received field-strength to the input power, distance and transmitting antenna gain. This was shown in reference [8]

The maximum power flux from an antenna is described by (Eqn 1)

rh-equ01.gif (1420 bytes)

Where Gt is the gain of the transmit antenna in the direction of the maximum, Pt is the transmitted power [W] and R is the distance of observation [m]. If you want to consider the perfect (isotropic) point source, the gain is set to unity (or 0 dBi).

The flux equation can be used to obtain the Field Strength at the observation point by (Eqn 2):

rh-equ02.gif (1352 bytes)

Where rh-equ02a.gif (845 bytes) is the intrinsic impedance of free-space ~ 120 p (~ 377 ohms).

We can substitute (Eqn 1) into (Eqn 2) to arrive at (Eqn 3):

rh-equ03.gif (2410 bytes)

The Field Strength from an antenna is related to the input Power, the gain and the distance of observation by (Eqn 3,3A).

Ground Effects
The ground, by far, has the most interesting influence on low vertical antennas which are deployed at LF, MF and HF. (See reference [1] and reference [6] - the later is very comprehensive.)

Traditional linear antennas are surrounded by large reactive near-fields. Some of these near-fields penetrate the lossy ground and reduce the antenna's efficiency as a result. This is why the FCC and other broadcast authorities state that 120 radials shall be employed under commercial monopole installations [4]. This situation can be alleviated by elevating the radials [13].

As stated in the previous section (and reference [8]), the ground can provide a gain factor of up to 4, or an improvement in field strength of 2, in reference to the same antenna operating in free space (or the absence of ground effects). This gain is a cheap gain as it requires no hard work, other than mounting the antenna.

K. A. Norton [1] refers to the Field Strength of an antenna above ground as 2 Eo / d, because the Field-strength is twice that of free space and it decreases linearly with distance for distances between 1 wavelength up to 10 wavelengths and more.

We can account for the effect of the ground by modifying the field strength by multiplying by 2, or by multiplying Gt by 4—it doesn't matter which, as the result is numerically equivalent.

In the LF, MF and lower part of the HF band, the ground appears largely as a resistor, very little contribution to ground-loss is afforded by the relative dielectric constant [1,2,3] and below 600 kHz, the ground appears almost like a perfect conductor.

The Field (Eqn 2) is modified by a ground-loss Attenuation term (A) in (Eqn 4):

rh-equ04.gif (1443 bytes)

In simple terms the ground can be treated as a loss-less plane conductor over short distances, at greater distances the field requires corrections which include ground losses, and at greater distances requires some correction due to the earth's geometry.

The plane-ground case has been included in Norton's empirical attenuation term (A) and which can be approximated by:

rh-equ05.gif (2818 bytes)

The numerical distance of one (rh-equ05a.gif (851 bytes) = 1) is the distance up to where ground losses do not substantially contribute to a decrease in field strength predicted by the inverse-square law for propagation of power.

The numerical distance rh-equ05a.gif (851 bytes), for vertical polarization, is defined by (Eqn 6):

rh-equ06.gif (2376 bytes)

rh-equ07b.gif (837 bytes) is in mhos / m (or Siemens).

The angle b, for vertical polarization, is described by the (Eqn 7):

rh-equ07.gif (2266 bytes)

rh-equ07a.gif (848 bytes) is the relative dielectric constant, with air and vacuum being 1.

The actual distance where the numerical distance equals one (rh-equ05a.gif (851 bytes) = 1) can be quite large. For 500 KHz over seawater, this distance is 21,300 miles. For 1120 KHz over land, with rh-equ07b.gif (837 bytes) =5 mS/m and rh-equ07a.gif (848 bytes) = 21, rh-equ05a.gif (851 bytes) = 1 at 231.6 miles.

If you plot E / d curves on log-log paper they should form a straight-line. The change in slope from the inverse-distance line (power inverse square-law) depends upon the ground-loss term A. Thus different loss factors are easily identified by fitting the slope of the E/d line. In the LF, MF and lower HF bands, ground loss depends almost entirely on rh-equ07b.gif (837 bytes).

Traditionally field-strength measurements are either quoted in milli-Volts per meter (mV / m) or referenced to decibel mirco-Volts / m (dBµV/m). The conversion is as follows:

rh-equ08.gif (1601 bytes)

I currently know, at the time of this writing (June 2000), only two disclosed sources of measurements for the Egyptian CFAs, one commercial null-result in the case of the Sydney CFA [12], plus various experimental and some professional engineers' positive and null results. There are reported to be measurements of the CFA at San Remo, Italy, but they have not yet been disclosed to the public. There are only two absolute Field-Strength measurements for the Sydney CFA provided by Dr Kabbary and four relative field-strength readings provided by Steve Olney.

Because of this scarcity of data, I have requested that all who are working, or have worked, on the CFA, to forward their results to me at, so I may collate and publish findings in a convenient place when such data becomes available. The questions I need answered are:

·  If you have constructed a CFA did you get it going?

·  If you got it going what, if any, field-strength measurements versus distance have you made?

·  Can you provide near-field measurements as well?

·  Have you measured the current distribution?

The early sets of CFA measurements, which support the CFA favourably, were performed on the Egyptian CFAs and were supplied by Dr Kabbary [9,7IV] The second set of Egyptian CFA measurements were supplied by Sylvio Damiani [10]

The first set of measurements, I believe, were performed circa 1992, while the second set of measurements where performed in June 1999 and have been certified by Sylvio's seven-engineer team. A facsimile of the Brazilian Engineering Report may be found at .

I have corresponded with Sylvio and he stands by these measurements and states that the FIM-41 Field Strength meters were calibrated and in agreement with each other. Given his background, his correspondences and his attitude to stand by these measurements, I have no doubt of Sylvio's veracity. So perhaps we should keep an open mind and consider that these Egyptian CFAs are indeed working as the inventors claim.

On the other side of the argument, by reports, the recent Sydney CFA is an apparent null-result, with radiation terms stated as up to 9 dB down on the replacement antenna. We must note that the distant measurements were relative field strengths performed at large numerical distances.

My own initial investigations into the CFA also yielded a null-result, with signals approximately 24 dB down, but I am prepared to reopen my CFA investigations, given my recent analysis of the CFA data and correspondences from many experimentalists around the world.

Table 1 has been organised in increasing distance from the Tanta antenna site and shows the comparison of the Electric Field Strength from the CFA and the tower which the CFA has subsequently replaced.

Freq kHz


Tower mV/m

CFA mV/m

































Table 1: CFA measurements for 1 kW @ 1 mile for various CFA sites attributed to Dr Kabbary

In Figure 1, I have plotted the results of table 1, and included the E / d and 2 E / d curves for a 1/4 wave monopole (Gain=1.5).

figure1x.gif (53319 bytes)

Figure 1: various field-strength measurements at 1 mile for 1 kW input

In Figure 1, for reference and to set the gain in context, I have plotted the apparent gain of the CFA and towers above the isotropic radiator.

figure2x.gif (42952 bytes)

Figure 2: Apparent gain from measurements in Table 1











Kafe el Zaiat





Kafe el Sheikh










El Mansoura




















Shourbra - Cairo





EUTR Cairo















Table 2: Tanta far-field measurements, attributed to Dr Kabbary NAB/99

Figure 3 includes the E /d curves for the Tanta CFA and the reference tower, which the CFA replaces. Note that the CFA follows the E / d curve for ground conductivity 25 mS / m. Some of the measurements are at large numerical distances and may require correction due to the earth's curvature.

figure3x.jpg (38850 bytes)

Figure 3: Tanta 1161kHz measurements

Figure 4 contains the apparent gain of the CFA and the towers over an isotropic radiator, again for reference and context.
>>> Click for Figure 4: Tanta apparent gain

The Egyptian data in Table 3 were supplied by the Brazilian Measurement team after the CFA had a top hat included in the design, and after the CFA was co-located with another CFA on a building approximately 7.5m above the ground.


E V/m









Table 3. Brazilian, Top-hat CFA measurements, June 1999.

Figure 5 contains the E /d plots for the Brazilian measurements on the elevated top-hat Tanta, 1161Hz, CFA.
>>> Click for Figure 5: Brazilian measurements at Tanta

While Figure 6 shows the apparent gain of this CFA over an isotropic antenna!
>>> Click for Figure 6: Apparent gain from the Brazilian measurements at Tanta

Table 4 contains the Sydney Data attributed to Dr Kabbary.


E V/m





Table 4, Sydney CFA measurements

Figure 7 shows the gain of the Sydney CFA in reference to E / d and 2 E / d, and Figure 8 shows the apparent gain of the Sydney CFA in dBi.
>>> Click for Figure 7: The Sydney measurements
>>> Click for Figure 8: The apparent gain of the Sydney measurements

Not wanting to draw any conclusion about why the measurements appear so favourable for the Egyptian CFA I will only summarize the results.

The initial CFA measurements, attributed to Dr Kabbary, et al show that the CFA has gain which varies between 0 and 7.1 dBi, and was measured to be approximately 3 to 9 dB above the apparent gain of the towers. These data seem to fit the soil conductivity rh-equ07b.gif (837 bytes) = 25 mS/m. There is a suggestion that the output power may have been 25 kW, rather than the stated 30 kW, but my theoretical values assume 30 kW input.

The data for signal strength at 1 mile, with 1kW input, show that the CFA apparent gain varies between 3.4 to 7.1 dBi, approximately 2 to 4.4 dB gain over the towers.

The data from the Brazilian measurements also seem to fit s = 25 mS/m. The Brazilian data show the CFA as varying between 3.6 to 11.1 dBi, with a modal value near 10.7 dBi, which is higher than the first set of Egyptian measurements.

Any difference between the earlier measurement and these might be explained by the top hat, the mounting height and perhaps the co-location on top of the building with the other CFA.

The Sydney data is too small of a sample. Its shows the CFA gain as varying between 4.9 and 7 dBi (not unlike the initial Egyptian measurements) but the measured E / d curve does not fit any particular ground model. The first measurement is too high and a little too-close and may have been influenced by the near-field of the antenna, but a subsequent report informs us that a peak reading was adopted from a fluctuating field strength, so this reading may just be reported high. If I had more measurements I may be able to fit the data to an appropriate value of s.

As indicated by the relative field strength measurements in reference [13] and, the Sydney CFA seems to be a failure. However, note that these measurements were performed at large numerical distances and I would prefer to see ones taken closer in. The terrain may also be affecting the signal strength, because unlike Cairo, Sydney cannot be considered as flat. I am uncertain of the direction and terrain profile surrounding the distant Sydney measurements.

The general variation of all apparent gain data may be attributed to variation in terrain or obscuration by buildings. Note how the CFA gain profiles tend to follow the monopole gain profiles. This is what you would expect if the variation is due to the same effect, such as terrain profile variation.

Lastly, you may copy the spreadsheet from here  (Click Here for Spreadsheet) as referenced [14] and explore effects of ground and use your own measurements and input power. The spreadsheet contains macro functions which have been used to calculate Norton's ground attenuation parameter, because it is a lot easier to call functions than to layout intermediate results for the intermediate variables.


  1. The Calculation of Ground-Wave Field Intensity Over a Finitely Conducting Spherical Earth", K. A. Norton, Proceeding of the I.R.E, December, 1941.
  2. "The Cross-Field Antenna in Practice", Bryan Wells, G3MND, Electronic and Wireless World, November 1989.
  3. "CFA Experiments", Bryan Wells, G3MND, Electronics and Wireless World, March 1990.
  4. "CFA: Working Assumption?", Electronics and Wireless World, December 1990.
  5. "CFA - RIP?", Electronics and Wireless World, May 1993
  6. "Medium Frequency Propagation Prediction Techniques and Antenna Modelling for Intelligent Transport Systems (ITS) Broadcast Applications", Nicholas DeMinco, Institute for Telecommunication Sciences,
  7. "The Cross-Field-Antenna Part I", by Maurice C. Hately GM3HAT and Ted Hart, W5QJR,, Sep 1998.
    "The Cross-Field-Antenna Part II", by Maurice C. Hately GM3HAT and Ted Hart, W5QJR,, Oct 1998
    "The Cross-Field-Antenna Part III", Mr Hately, Dr Kabbary, Dr Stewart, and Ted Hart,, Nov 1998
    "The Cross-Field-Antenna Part IV", Mr Hately, Dr Kabbary, Dr Stewart, and Ted Hart,, Dec 1998
  8. "MF HF Groundwave Model", Ralph Holland, VK1BRH,, June 2000
  9. "Egyptian CFA Field Strengths", Dr Kabbary, NAB99 paper.
  10. Correspondence: "Brazilian Field Strengths for Tanta 1161 Hz CFA", Sylvio M. Diamiani, June 1999. See the report at
  11. Archive: "Brazilian Field Strength Report of Tanta CFA Station", Jack Belrose,
  12. "The Sydney CFA - For Sale", Jack L. Stone, June 2000,
  13. "Short Verticals and Ground Systems", Ralph Holland, Amateur Radio, October 1995,
  14. Data as provided by "CFA.xls" Click Here for Spreadsheet
  15. "The Crossed Field Antenna" M. C. Hately", Journal of the Royal Signals Institution, Vol. XX, Summer 1991 ISSN 0374-3519

Brief Biography of Author Click here

Print/Save Article

Send mail to with questions or comments.
Copyright © 1988-2015 All rights reserved - antenneX©
Last modified: December 31, 2010