John S. (Jack) Belrose, VE2CV 17 Tadoussac Drive Aylmer, Quebec J9J 1GI Canada ELEVATED RADIAL WIRE SYSTEMS FOR VERTICALLY POLARIZED GROUND-PLANE TYPE ANTENNAS Part 7 -Monopoles esonant quarter wavelength radials been recognized because of the way in which (electrical length at a quarter wave- current densities in the ground are more or less R length) can be used with practical ele- uniformly distributed over the area of the insu- vated ground-plane type antennas and to simu- lated counterpoise (which is, in effect, a large late "connection" to ground for numerical mod- capacitance ground). Figure 1 shows wires eling programs such as NEC-2, which does not running radially outwards on insulated supports allow a wire to touch lossy ground. without connection being made to earth plates or ground stakes at the outer ends of the sys- tem. An alternative view of the way an elevated Introduction radial system works is that it allows the collec- tion of electro~nagneticenergy in the form of From the earliest days of radio, the merits of displacement currents, rather than conduction elevated counterpoise and radial systems have currents flowing through lossy earth. Certainly, Figure I. T-antenna over a counterpoise earth [after The Admiralty Handbook of Wireless Telegraphy, 1~38~1. Cornrnunicafrons Quarterly 29 COUNTERPOISE EARTH GROUND Figure 2.2BML's "flat top" antenna system of 1921. when only a few elevated radials are used, we in the early days of radio. Unfortunately, after cannot consider the counterpoise forming a about 1937, the use of elevated radial systems large capacitance ground. became an almost forgotten art, because of H.H. (Bev)Beverage, 2BML (now a silent research by Brown et. a1.3 in favor of the buried key), in a 192 1 publication by RCA described radial system. These authors carried out a very an aerial and counterpoise system suggested by extensive measurement program to determine Alexanderson and using a coupled ground wire. the impedance and radiation efficiencyof a ver- Details of 2BML's antenna system are shown tical monopole as a function of ground system in Figure 2. Having permission to operate parameters. A result of their work was their above 200 meters, Beverage chose to tune his recommendation that 120 radial wires, each one antenna system to 280 meters (1070 kHz). With about half a wavelength long, should be used to a fair ground, his measurements showed a sys- maximize radiation efficiency;0.4 h is a length tem resistance of around 70 ohms and 0.5 quoted in many text books as optimum. This is ampere antenna current. But with the elevated a curious result (a half a wavelength?) in light counterpoise attached, the system resistance of what we know today.4 dropped to I0 ohms. The ground lead tap on Elevated counterpoise types of ground sys- the inductor was adjusted to cancel the capaci- tems were extensively studied through experi- tive reactance of the elevated counterpoise. ments by Doty, Frey, and Mills.5 The radial With both the earth and counterpoise connect- systems they used were not unlike that shown ed, a system resistance of 4 ohms and an anten- in c.f. Figure 1 (except a perimeter wire joined na current of 8 amperes was obtained. With 8 the ends of the radial wires), i.e. the counter- amperes of RF current going into the antenna, poise radial wires filled a rectangular area over the counterpoise current (Ic)was about 6 the ground, rather than being of equal or a reso- amperes, and the earth ground current (Ig)was nant length. Christman, KB81,6 was perhaps the about 2 amperes. first to show by numerical modeling that a ver- It was stated that most amateurs, and Inany tical monopole with as few as four horizol~tal broadcast stations (KG0 was one of them) radial wires could be at a height quite low over were already using the counterpoise at that real ground before the radiation efficiencyof time. In December 192 1, the amateur radio sta- the antenna was significantly degraded. tion of The Radio Club of America, 1 BCG, Unfortunately the topic of ground mounted and also employed a counterpoise system contain- elevated radials, and antennas with elevated ing 30 wires each 73.5 meters long (0.3 1 h for feed, has remained clouded by controversy, an operating wavelength of 230 meters).? e.g., posting on the Internet News Group on The development of such ground systems Radio Amateur Antenna; and, c.f. Belr~se~.~ was, by necessity, empirical in nature, as and ~awker.~More recently weberl0 has sophisticated instrumentation and standardized described measurements he made on systems antenna testing procedures were not available employing elevated radials, which revealed 30 Winter l99ti Figure 3(A). Wire model and current on the wires for a 20.2-meter vertical wire antenna with one resonant radial (length 19.184 meters for a height of 2.5 meters over average ground; (B) vertical radiation pattern; and (C) principle plain azimuthal pattern (resonant frequency 3.75 MHz). ....@dB--., EZNEC/4 120 ..... '.. 60 . ... '. .. ...... ... .., I..:.: . 1 . .0 deg. ...I................ ..... ..:... ...: ... ., ... ......... ..... .... ... ..... ... .... ... .................. ... ... ... ... ,.. .... ... .... .... ' . '.. ............:... ... ..... ... Outer Ring = -0.39 dBi ..... ...... 8 deg. .... Communicoiions Quarterly 3 1 I - ii. i / j ! 1 i .@@Ill .a001 .@#I .O1 1 Height (wavelengths) Figure 4. Predicted gain for a 20-meter monopole versus number of 20-meter radials and height in wavelengths of radials over average ground; and for a vertical half-wave dipole for reference, where height is the lower end height of the dipole. Frequency is 3.75 MHz. unequal radial currents, and he employed short antennas, I noticed that a vertical antenna numerical modeling to aid his understanding of with one radial exhibited a marked directional the subject (see also comment by ~evernsl and effect. The direction of maximum gain (maxi- Appendix 1 of the present article). mum groundwave field strength) was the direc- To continue, if four elevated radials can be tion toward which the radial ran. A frontback used to realize performance for a monopole ratio of 10 dB was observed for a short center- comparable with many buried radials, then ele- loaded whip with a quarter wavelength insulat- vated radials can be used for real antennas and ed radial lying on the ground. to simulate ground "connection" for numerical This discovery lead to the development of a modeling programs like NEC-2, which does not simple transportable antenna, dubbed, by permit a wire to connect to lossy ground. Christman,6 the "VE2CV Field Day Special" This article is an overview of the subject of antenna. This antenna was simply a quarter elevated radials. wavelength vertical wire with a tree support (or a pipe mast for 40 meters and down), with one radial elevated above head height (2.5 meters) Initial Experience directed toward the azimuth of principle inter- est. The directional pattern can be changed by NEC-2 is currently enjoying wide application installing more than one radial and selecting the by radio amateurs using PC based programs appropriate radial by connect/disconnect relays. such as EZNEC, available from Roy Lewallen, Figures 3A through C show the patterns for W7EL,12 and EZNECl4 pro for those regis- such a simple antenna over average ground tered for NEC-4. The ability of the NEC codes (ITU-R definition of average ground, viz. = 3 to accurately treat the air-ground interface mSIm; E = 13). The patterns are for a frequency (based on the Sommerfeld-integral formula- of 3.75 MHz, quarter wavelength vertical (reso- tion) is assumed to be validated by the develop- nant length 20.2 meters), with a single resonant ers of the program3. radial (length 19.184 meters) for a radial height About 20 years ago, while conducting experi- of 2.5 meters over average ground. The input ments to determine the efficiency of electrically impedance according to NEC-4 is a good match 32 Winter 19911 .00001 .0001 .001 .O1 .1 Height (wavelengths) I Figure 5. Predicted impedance (resistance and reactance) for a quarter wavelength, 20-meter, 3.75-MHz monopole versus radial height over average ground, for four fixed-length (20-meter) radials and for four resonant radials. Communications Quarterlv 33 - 1 I i 0- -1 - s - -2- .Ic c? -3 -- I -4 -. -5 1 .00001 .0001 .001 .01 .1 Height (wavelengths) 1 Figure 6. Predicted gain for a quarter wavelength monopole with four quarter wavelength radials and with resonant radials versus height (wavelengths) of radials over average grounds; and for a half-wave dipole where height is the height of the lower end of the dipole. Frequency is 3.75 MHz. for 50-ohm coax (57 ohms), but a current balun changed in a way closely predicted by NEC-4. should be used. For four, eight, and 16 radials on the ground, Note that Figure 3C shows the vertical, hori- his measured gain differences, referenced to the zontal, and total fields-dashed, dotted, and relative field strength measured for 60 radials continuous lines-since there has been some on the ground, were -5.5 dB, -2.7 dB, and -1.3 dB. discussion about the horizontal field not cancel- Since NEC-4 does not allow a wire to touch ing when there are fewer than three radials (three lossy ground in more than one place (NEC-2 radials is the minimum number that can be does not allow a wire to touch the ground at arranged symmetrically about the monopole). all), I modeled a radial on the ground by assuming an insulated radial system with radi- als 5 millimeters high. With four, eight, 16, 32, A detailed study and 64 quarter-wavelength radials, EZNECl4 predicts gains for an antenna over average For the case studies to be reported here, 1 ground of -4.66 dBi, -2.16 dBi, -0.6 dBi, +0.03 chose to model a 3.75-MHz monopole over a dBi, and +0.23 dBi.