the low-angle maximum is scarcely affected. Over typical soil, the phase change occurs at launch angles around 15°. For launch angles below this, performance is seriously degraded. (See Fig 1A.) This poor performance occurs at angles comparable to the most probable arrival angles for sky-waves from distant stations. [See Chapters 3 and 23 of the 18th edition of The ARRL Antenna Book for detailed elevation-angle statistics.—Ed.]
For horizontal polarization over imperfect ground, the reflection coefficient at small elevation angles is slightly smaller than at higher angles. The phase change for the ground-reflected wave is substantially 180°, so the launch angle for maximum radiation is scarcely changed. The reduced reflection coefficient leads to a small loss of gain, by a dB or so for low launch angles (see Fig 1C). For high launch angles, the ground-reflected wave is weaker and the
Fig 1—At A, elevation pattern for half-diamond GP-type loop. At B, elevation pattern for a dipole at 15 meters (0.19 X) and at 40 meters (a/2). At C, elevation pattern for a dipole at 80 meters (l X). Frequency in all cases is 3.75 MHz, for four ground conductivities: sea water (solid line), good ground (dashed line: o = 10 mS/m, e= 30), poor ground (dotted line: a = 1 mS/m, e = 15), and very poor ground (dashed-dotted line: a = 0.1 mS/m, e = 3).
antenna's gain is less, by up to 4 dB compared to perfect ground. The gain does remain some 2 to 3 dB higher than it would be if ground reinforcement were absent. (See Fig IB.) For practical purposes, and particularly for antennas designed for low launch angles, the effects of real ground may often be ignored if the polarization is horizontal.
The launch angle for maximum gain from a horizontal dipole, however, depends on the height in wavelengths of the antenna above the ground. For dipole heights less than 0.3 X, ground reinforcement occurs at high launch angles; that is, \f/ close to 90°. This is an antenna that Doug DeMaw, W1FB, refers to as a "cloud warmer." (See Fig 2A.) When the dipole height is 0.5 a, the radiation pattern has a single lobe in the plane orthogonal to the plane of the dipole, and there is a null overhead. For ground reinforcement to occur at small launch angles, say \|/ < 15°, the electrical height of the dipole must be greater than one wavelength. For 80 meters, this is impractical, since the height would have to be greater than 80 meters! The best antenna (vertical or horizontal polarization) for 80-meter DX is obviously a compromise.
Full-wavelength loops have been popular with radio amateurs for many years. They are used as elements for quad and delta-loop beam antennas for the higher frequency bands. The usual configurations are shown in Figs 3A, B and C, popularly called quad, diamond and delta-loop antennas, respectively. When fed as shown, these antennas are horizontally polarized in the plane orthogonal to that of the loop.
The loops are symmetrical about the center line, shown
Fig 2—At A, overlay of elevation patterns for 3.75-MHz halfdiamond loop, dipoles at 15.24 meters (0.19 X) and at 40 meters (A/2) over average ground (o = 3 mS/rn, e = 13). At B, elevation-pattern overlays for only the half-diamond loop and the dipole at 15.24 meters. See Note 6.
dashed. If we rotate the loops clockwise through 90°, and if the referenced center line is now the ground plane, we have a half-loop, which works together with its image in the ground plane (Figs D, E and F). One end of the half-loop is grounded and the other is fed against ground. The azi-muthal pattern at the fundamental resonant frequency is maximum in the plane broadside to the plane of the loop, and the radiated field is vertically polarized. Sometimes wire loops are supported by trees; sometimes by metal masts. Often the mast also supports a Yagi antenna at the top. It has long been known that a metal tower can affect the impedance and radiation pattern of any wire antenna that it also supports. The effect of support structures on the radiation patterns of antennas has been reported, however, for only a few configurations and antenna types.'
Whatever loop configuration is employed, a ground-plane (GP) type loop must be well grounded. My half-delta loop is grounded by a 3-meter ground rod and four buried quarter-wave radials. The fed end is close to the house, so only three radials could have been used. The transceiver is grounded to the power mains ground, which is connected to the copper water-pipe system in the house. Since the house water supply is a well, my water-pipe system is not connected to an extensive and well-grounded city water system. Further, the lead-in pipe from the well is plastic.
One day I decided to tie the backyard chain-link fence into the ground system. This vinyl-coated, steel-fabric fence is fastened to a tubular steel framework, made up of a continuous pipe running the 200-meter perimeter of my backyard. The fence pipe is supported by tubular metal posts set in concrete every 6 meters. A pipe set in concrete in moist earth can make a good ground. Such a ground is known as a Ufer ground, after the engineer that studied this method of grounding.8 9 To ensure a continuous loop of pipe, a buried jumper wire was connected across the gate; and the fence was jumpered to the copper water-pipe system at both sides of the house.
Curiously, the received background noise level (judged
Fig 3—At A, B and C are sketches of horizontally polarized quad, diamond and delta loops. At D, E and F are sketches of vertically polarized ground-plane-type half-quad,'half-diamond and half-delta loops.
to be local mail-made noise) on the 80-meter band dropped by an S-unit or more when the fence was tied in. I checked this on several occasions. My whole backyard (about half acre) is my ground system!
About 15 years ago, after conducting a series of extensive experimental modeling studies, I erected an 80-meter halfdelta loop in the back yard.101 already had a 15-meter freestanding tower in place (with a 3-meter pipe mast extension to support TV antennas) located just outside the window of the half-basement ham shack. The most logical arrangement was to mount a tree-supported wire loop so that the vertical part of the half-delta loop was as far as possible from the tower, with the sloping wire running toward the tower. But since the wire antenna was in the backyard, the lower end of this sloping wire did not come directly to ground, so that people could walk safely beneath it. This wire ran through an insulator attached to the tower at a height of about 2.5 meters, and from there to ground level. The feed was between the lower end of this wire and ground. See Fig 4A. The spacing between the wire and the tower was about 30 cm.
I noticed that when using the loop there was significantly more of a problem with TVI (recall that the tower also supported TV antennas) than when I used a half-wave drooping dipole suspended from the same tower. Since the halfdelta loop is vertically polarized and the tower is about the right height (18.3 meters) to be approximately resonant, I wondered whether this so-called "isolated tower" was in fact isolated, or whether it was a part of the radiating system.
I modeled the antenna and tower, using EZNEC. Because the program does not permit any wire to be connected to imperfect ground, I simulated the ground connection for the half-delta loop using two sets of two quarter-wave resonant radials elevated a meter above ground. These radials were directed in the plane orthogonal to the loop. See Fig 4A. For the tower, I simulated the ground connection using four quarter-wave resonant radials elevated 5 mm above ground. [EZNEC data files for all these models by VE2CV are located on the Hiram BBS, in an archive file called VE2CV.ZIP—Ed.]
According to EZNEC, the loop and tower system was resonant at 3.7 MHz (the measured resonant frequency was 3.65 MHz). The tower does carry a significant current, 0.75Z690 A for a 1Z0°A source current at 3.75 MHz. The current at the grounded far end of the vertical element of the half delta loop is 0.97Z69° A. Since the currents on the vertical parts of the radiating system (vertical wire and tower) are approximately in quadrature, you might expect a cardioid-like pattern—and this is what EZNEC predicts. See Figs 4B and C.
The antenna was oriented so that the plane of the loop was in the N-S plane, with the tower and feed in the north. By chance, my radiation pattern was very suitable for a control station of the Trans Canada Pow Wow Club—with best coverage a bit to the north of the E-W direction. But this was by chance rather than by design, since I had not intended the tower to be a part of the antenna system!
This discovery of this interaction sparked a detailed study, since I wondered whether a half-diamond loop, where the loop is centered on the tower, might be better. First, let's look at some characteristics of half loops in "ideal" situations.
Fig 4—VE2CV's half-delta loop for the 80-meter band. At A, wire model showing currents on the wire loop and the induced current on the tower. At B, elevation pattern and at C, the principal-plane azimuthal pattern. The patterns were calculated for average ground in front of the antenna.
Was this article helpful?