plate dipole is similar to a rod dipole, with the rods replaced by plane sheets of metal,
which can be of any shape, but in this example they are assumed to be rectangular, with a
small spacing between the parallel edges. A P2 is like a rectangular plate dipole with the
inside missing, so that the P2 is delineated by wires of appropriate thickness running
along the perimeter of the plate dipole. The spacing and thickness of the wires in the
parallel edges of a P2 are carefully chosen to form a transmission line of the necessary
characteristic impedance. Please see the earlier papers on the P-type antennas in the
Archive IV and V of antenneX.
The plate dipoles considered in this short article are modelled in NEC2 by wire grids. It is still not clear how accurate this approximation is to a continuous sheet of metal. In order to get some feel for potential inaccuracies, the modelling has been made at various wire grid hole sizes.
In the course of many discussions between the authors related to the theoretical understanding and practical applications of Dan Handelsman's Prismatic Polygon, David Jefferies undertook to assign the construction and analysis of a P2 to one of his students. This antenna, pictured in Figure 1, is a full wavelength (fw) loop in which the two radiators are fed by transmission lines (tx lines) from a common feepoint at the antenna center.
Although this antenna is planar, and not three-dimensional (3-D) as are the true Prismatics, it still has a very wide bandwidth as illustrated in Figure 2 and is simple to construct. The bandwidth is about 2.8:1 and the R/X sweep, Figure 3, is characteristic of the entire family of Prismatics - although the R and X excursions are wider and the SWR bandwidth is narrower than with the more complex 3-D antennas.
The gains and radiation patterns of vertically polarized P2s, shown in Figures 4 (azimuth) and 5 (elevation), are typical of dipoles, albeit with extremely wide bandwidths. Jefferies theorized that a "Plate dipole" (PD), in which the upper and lower halves of the P2 were replaced by solid metal plates, could perform similarly. In addition, this configuration would make it amenable to "sandwiching" the conductors between dielectrics in order to reduce its size.
This article, the first of two on this topic, will reveal the results of modeling studies on the construction and performance of these PDs and the second will go into a theoretical analysis of why they function as they do.
A simple dipole, of what may be considered "thin wire", has the SWR BW seen in Figure 6. The resonant length at 300 MHz is 0.473 m. "Thin" is a relative term - the antenna wire is 2.5 mm (corresponding to #10 AWG) at a wl of 1 meter or 0.0025 wl in diameter. However, the diameter is nowhere as thin as dipoles we would use on say 80 meters. This antenna's R and X plot is seen in Figure 7. The relative bandwidth of this antenna is about 12%.
A "thick" dipole, having a wire diameter ten times as large or 2.5 cm (0.025 wl) and with a resonant length of 0.448 m, has a SWR BW seen in Figure 8 and a RX plot as in Figure 9. This diameter was chosen since it was the greatest that fell within the limits of accurate NEC modeling. The relative bandwidth about the first resonant frequency is 27%.
While Figure 6 shows a relatively narrow SWR sweep of the thin dipole, Figure 8 gives you an idea of the extended bandwidth response of such a thick dipole. As expected, the antenna is also resonant at its third harmonic and shows a high SWR in between.
The peak SWR between the resonances is inversely proportional to the diameter of the wire. Stated differently, the peak is directly proportional to the Q factor of the antenna. Theoretically, we should be able to construct a dipole of sufficient wire diameter to minimize the Q such that the intermediate SWR peak falls below 2 and thus provides a bandwidth of 3:1 encompassing the primary and secondary resonance.
Given these two dipoles as standards of reference, let us now examine how well two other types of antenna perform. First, we shall look at the ultra-wideband reference standard or the P2.
P2 - Two-Radiator Prismatic
The antenna shown in Figure 1 and characterized in Figure 2 and Figure 3 is composed of 1.5 cm thick wires and is exactly 1 meter in perimeter. A very wide bandwidth is achieved with a simple open matrix structure involving only four wires and two transmission lines. This antenna was designed for a nominal reference Zin of 75 ohms to be comparable to that of a simple dipole. A wider BW, approaching a value of slightly less than 3:1, can be attained if one designs for a higher Zin such as 100 - 200 ohms.
This antenna, with a height of 0.4 m, a width of 0.1 m and with tx lines of Zo = 300 ohms connecting the radiators to the feedpoint, is designed for an omnidirectional radiation pattern when vertically polarized. The Az pattern is circular at the lower edge of its passband and is just slightly out of circularity at the upper end, as shown in Figure 9. When the two radiators are separated by a relatively large distance, in terms of wl, the P2 wants to radiate broadside to the plane containing the radiators or the plane of the rectangle. A lot of design compromises go into producing antennas with circular radiation patterns at the high frequency end and these invariably involve some narrowing of the bandwidth.
Nevertheless, this P2, with wires significantly thinner than the thickest dipole's - 0.015 vs. 0.025 wl - has a BW of 260 - 715 MHz or about 2.8:1.and approaches the limit of what can be achieved by this simplest of all the Prismatic antennas.
The Plate Dipole or PD
Jefferies suggested substituting solid plates for the upper and lower halves of the P2 in Figure 1. A plate dipole is similar to a rod dipole, with the rods replaced by plane sheets of metal which can be of any shape. In our examples they are assumed to be rectangular, with a small spacing between the parallel edges.
A P2 is like a rectangular plate dipole with the inside missing. It can therefore be delineated by wires of appropriate thickness running along the perimeter of the plate dipole.
The spacing and thickness of the wires in the parallel edges of a P2 are carefully chosen to form a transmission line of the necessary characteristic impedance. Please see the earlier articles on the Prismatic type antennas.
A solid plate is impossible to model with NEC but a reasonable substitute may be found if one uses a grid of wires. Such an antenna is shown in Figure 10. It is still not clear how accurate this approximation is to a continuous sheet of metal. In order to get a feeling for the potential inaccuracies, the modeling has been made at various wire grid hole sizes.
Various grid densities were tried including the one illustrated which has a 10 x 20 grid matrix per radiator. This antenna has almost 1000 wires and fairly long run times at each frequency modeled. In the end, it was found that a grid matrix of 10 x 6 squares per radiator gave reliable results, though with a narrower bandwidth as we shall see below.
In order to compare the performance of this antenna with that of the "thin" dipole discussed above, the wire matrix was also modeled with 2.5 mm diameter copper wires. The antenna's dimensions are a length of 0.1765 m and a width of 0.147 m per radiator half with a feed gap of 0.02 m between the plates. Therefore the overall height is 0.353 + 0.02 or 0.373 m.
Figures 11 and 12 show how this PD performs. The maximum BW of 300-860 MHz or 2.9:1 was attained at a Zref of 90 ohms.
For the sake of comparison, a PD was modeled with a grid density of 5 x 4 squares per radiator and with exactly the same dimensions. Its performance is shown in Figures 13 and 14. Even with this coarse grid structure, the bandwidth, at a Zref of 85 ohms, is substantial at 290-785 MHz or 2.7:1.
The Az radiation patterns of the fine mesh, wider bandwidth, PD are given in Figures 15 and 16. The pattern is circular at 300 MHz and becomes substantially broadside at 900 MHz. At the high frequency end it is out of circularity by 4.1/0.84 dB. The P2s, as we have seen in Figure 9, have substantially more circular patterns at the high ends of their passbands and can be designed for almost total circularity - within 1 dB.
Lastly, we come to the major bugaboo or limitation on the performance of simple dipole-equivalent, wide bandwidth, antennas such as the biconicals. Figure 17 is the elevation pattern of the vertically polarized thick dipole we have discussed above at its second resonance at 900 MHz. Note that the greater part of its radiation takes place at its secondary elevation lobe and squints off the boresight in free space. At frequencies higher than about 2.5 x the lowemost frequencies of their passband, biconicals and similar antennas become useless due to this effect. The PD and the P2 maintain their patterns at the boresight at the upper edges of their passbands and more complex antennas such as the P4 - where there are 4 radiators arranged in a 3-D square configuration - maintain useable patterns at bandwidths exceeding 4:1.
Plate dipoles or PDs can be designed to match the performance of the open matrix P2. It is expected that, when such an antenna is built with solid metal plates for each radiator, the bandwidth will approach 3:1. Such an antenna would be easy to construct at UHF and higher out of PC boards and should be amenable to sandwiching between dielectric plates in order to reduce its size.
The next article will give us a better theoretical understanding of why the PD performs as it does. -30-
|BRIEF BIOGRAPHY OF AUTHOR
Dan Handelsman, N2DT
Dr. David J. Jefferies
School of Electronic Engineering, Information Technology and Mathematics
University of Surrey
Guildford GU2 7XH
Click Here for the Authors' Biography
~ antenneX ~ April 2002 Online Issue #60 ~
Send mail to email@example.com
with questions or comments.
Copyright © 1988-2015 All rights reserved - antenneX©
Last modified: December 31, 2009