Understading the Controlled Current Distribution (CCD) Antenna
The controlled current distribution (CCD) antenna has been around since the late 1970s. Every so often, it arouses a flurry of questions in my e-mail. So I decided to look into the CCD to see what we might reasonably expect of it.
Fig. 1 shows the general outline of the center-fed version of the CCD. It consists of a number of straight wire sections of any practical number that we can designate as N. Except for the feedpoint and the wire tips, we must separate each section with a capacitor. Hence, the total number of capacitors will be N-2. The CCD simulates a continuously capacitively loaded 1 wavelength element by using equally spaced discrete components.
There is also a vertical version of the CCD. One way to create a vertical CCD is to simply use the center-fed antenna in a vertical installation. However, some literature suggests using the antenna against ground as a monopole. Fig. 2 shows the general layout of this configuration.
Beyond the general claim of "improved performance," I have not encountered a very clear account of what advantages the CCD is supposed to offer. The wire length will be 1 physical wavelength. However, the distributed capacitive loading of the element will electrically shorten the wire. If we select the correct value for all of the capacitors (they are all equal), we can arrive at resonance, that is, a feedpoint impedance with negligible reactance. It appears that the idea behind the CCD is to avoid the very high impedance of a 1 wavelength center-fed wire while preserving its gain and directivity.
I sometimes hear the terms "aperture" applied to the antenna. Most texts reserve the term for use with highly directive antennas, such as UHF horns and the like. Nevertheless, the calculation of aperture rests--according to one text--on the wavelength, the directivity, the polarization match, and the impedance match. Since a comparison of a CCD with any other antenna will equate the wavelength, polarization, and impedance match, the remaining factor is directivity. Hence, if the CCD is an improvement over a legitimate comparator, then it will show improvements in directivity. We can examine the directivity of the CCD by looking at its gain and beamwidth.
To simplify the examination, let's look at the CCD on a common frequency. 3.5 MHz is widely used in articles, so that will be our choice. A wavelength at 3.5 MHz is 280.02' long, so we shall make the antenna length 280'. Throughout, we shall use AWG #14 copper wire to reflect typical amateur building practices. Our initial tests will use a free-space environment. In this environment, we must use the center-fed version of the antenna, but we shall not have to be concerned with vertical vs. horizontal orientations.
I chose to model the antenna in NEC-4, which creates a small difficulty. If I use a single wire for the entire element, I can place load capacitors on individual segments of the wire. However, the feedpoint region will not be quite perfect. Fig. 3 shows the alternative models that I used. The upper model uses a single source segment, but that does not ensure that the wire lengths on each side of the feedpoint are equal to the other segments between capacitors. The lower section uses a split source to simulate a source at the center segment junction. Although this move improves the segment spacing, it can result in somewhat erroneous impedance reports for very high impedances, where the impedance might change significantly with only a small change of feedpoint position or total wire length.
For all practical purposes, the difference between the models is not great enough to jeopardize the modeling analysis. The required values of capacitance tend to coincide closely with values in the literature. The next task involves the capacitors themselves.
Literature on the CCD shows that we can build the antenna with almost any number of wire sections and corresponding numbers of capacitors. Since constructing each section involves wiring in a capacitor with appropriate strain relief--a considerable task--I wondered what one might gain by opting for a "large" CCD over a "smaller" CCD, where large and small indicate the number of wire sections and capacitors. Therefore, I created 2 models, one that used 26 wires section and 24 capacitors and another that used 50 wire sections and 48 capacitors.
The final step involves selecting the capacitor value. Since modeling allows easy variation of the capacitor value, we might explore the performance of the antenna in free space using various capacitor values. Let's start with the larger model. Table 1 shows the results of this modeling exercise.
Perhaps the first notable feature of the data is that as we raise the value of the capacitors in the string, the gain increases. So too does the resistive component of the feedpoint impedance. The reactive component shows an initial capacitive value that becomes inductive as we increase the capacitor values. This feature is natural enough, since increasing the capacitance value lowers the capacitive reactance along the wire. Since the wire is long compared to a dipole, lowering the compensating capacitive reactance will leave the feedpoint increasingly inductive.
In fact, more is at stake than just the feedpoint reactance. Note the entries for resonance. At resonance, the feedpoint reactance disappears, marking a balance between the inductive reactance of the wire and the capacitive reactance from the string of inserted components. Let's also examine the data for the "smaller" CCD that uses 26 wire sections and 24 capacitors. Table 2 provides the necessary information, but with fewer steps along the way.
The antenna gain does not depend for its gain--above some minimum number of capacitors--on the array size in terms of the number of capacitors in each leg. The gain of the smaller array with 500-pF capacitors is the same as the gain of the larger array with 1000-pF capacitors. Likewise for the smaller 1000-pF and the larger 2000-pF entries. The feedpoint impedance reports also track each other in the same manner. This result is also very reasonable, since 24 500-pF capacitors have the same capacitive reactance as 48 1000-pF capacitors if we measure at the same frequency.
The relative balance between these two factors alters the current distribution along the 1 wavelength wire. Fig. 4 provides some samples of the distribution curves for the smaller array, but curves for corresponding large-array values are virtually identical. The curve for the array with a low capacitance value that yield a capacitively reactive feedpoint impedance is very steep. At the other extreme, where the capacitor value yields an inductively reactive impedance, the curve shows dual current peaks. Only the capacitor values that yield resonance produce a curve that we tend to associate with a dipole.
The resonant impedance of a center-fed CCD is in the neighborhood of 200 Ohms. Therefore, one may simplify the matching problem by installing a 4:1 balun at the feedpoint and using a standard coaxial cable feedline. Of course, one may also use parallel transmission line to either a tuner or a 4:1 balun at the shack end of the line. Since 200-Ohm feedlines are rare, one likely would need a 300-Ohm feedline cut to the nearest half wavelength (allowing for the line's velocity factor) at the operating frequency to minimize balun losses.
With the right choice of capacitors in the string, the CCD offers bi-directional performance with a resonant feedpoint impedance. In exchange for the more complex construction of the antenna, we obtain a simplification of matching requirements. However, we have not yet assessed how good that performance is. For that task, we need an appropriate comparator.
The CCD vs. the Dipole and 1-Wavelength Center-Fed Wire
In CCD literature, the resonant 1/2 wavelength dipole seems to be the most popular antenna with which to compare the CCD with respect to performance as a horizontal antenna. An alternative comparator is the plain and simple 1 wavelength center-fed wire antenna. The 1 wavelength plain wire is the same length as the CCD. However, the current distribution of the plain wire differs from the current distribution of the resonant CCD. Fig. 5 provides a set of free-space current distribution curves for the resonant dipole and for the 1 wavelength wire for ready comparison to the CCD curves in Fig. 4.
The dipole curve resembles the resonant CCD curve. Of course, neither are perfectly sinusoidal, since the voltage at the center feedpoint never goes to zero. I have not explored the degree to which each departs from that familiar curve. In contrast, the 1 wavelength wire curve most resembles the CCD when the latter uses a very high value for the capacitors and thus loads the wire least.
Table 3 compares the free-space performance of the dipole and 1 wavelength wire along with resonant CCDs of the smaller and larger type. Clearly, the CCD has more gain (by about 1.1 dB) than the dipole. However, the CCD falls about 0.7 dB short of the plain 1 wavelength wire. We should have expected this result from the tabular data on the CCDs as we raised the value of the capacitors and lowered the level of loading. (The capacitors in all models are lossless.) As the capacitive reactance decreased, the gain increased. From a certain perspective, we may view the 1 wavelength wire as a CCD with capacitors having an indefinitely high capacitance and hence negligible reactance.
Fig. 6 overlays the free-space E-plane (azimuth) patterns of the antennas (using only 1 of the resonant CCDs). The CCD pattern most resembles the E-plane pattern of a center-fed plain wire about 0.85 wavelength long, although the CCD beamwidth is close to 10 degrees wider.
Very few prople operate 3.5-MHz antennas in free space. Therefore, we may usefully transplant all of the antennas in Table 3 to a height of 1 wavelength above ground. With this adjustment, we obtain the data in Table 4.
The gain differentials that we found in the free-space models hold up when we move the antennas over ground. Of course, the CCD loading technique does not alter the TO angle of the horizontal antenna relative to either comparator. Table 4 records the horizontal beamwidth values for all of the antennas, and we can see that the CCD beamwidth is closer to the value for the dipole than it is to the value for the 1 wavelength plain wire. Fig. 7 records the elevation and azimuth patterns for the antennas in this test group.
There is no difference among any of the elevation patterns, since those patterns emerge as a function of the height of the horizontal wire above ground. The 2 versions of the CCD show no differences in their azimuth patterns. Hence, the choice among the antennas with respect to performance largely hinges upon the combination of gain and beamwidth.
One of the seeming advantages of the CCD is the fact that in exchange for more complex construction, one obtains a higher gain with a simple resonant feedpoint. However, the tables have shown near-50-Ohm values for the plain 1 wavelength wire. The technique used to obtain such feedpoint values is very simple and very old. Since the terminal impedance of the 1 wavelength wire is very high, we can attach a 1/4 wavelength section of parallel transmission line. Somewhere very close to the end of the line will be a 50-Ohm matching point. The matching sections for the horizontal 1 wavelength wires use 600-Ohm line. The required length for the impedances shown is 68' (compared to a full 1/4 wavelength of 70.25'). Fig. 8 shows the simple set-up. The antenna environment and construction variables will determine the exact line length required, so the examples use a velocity factor of 1.0. Most 600-Ohm lines may have values between 0.95 and 0.98.
The matching section may be much simpler to build than the CCD antenna element. However, the 2:1 SWR bandwidth of the plain wire plus matching section is fairly narrow--perhaps 150 kHz at 3.5 MHz. (The resonant CCD has a 200-Ohm SWR bandwidth of about 300 kHz.) Indeed, the wisest feed system for the plain wire may be parallel line all the way to the shack antenna tuner. However, setting the line length to an odd multiple of a quarter wavelength at the most used frequency may ease the tuner's task by a good margin.
The Vertical CCD and the 1/2-Wavelength Base-Fed Plain Wire
One application suggested for the CCD is as a vertical antenna. We may hang a full center-fed CCD vertically with its feedpoint at any height above ground that we can manage to support. However, the more interesting case is to use a half CCD as a monopole, with the feedpoint at ground level. A monopole CCD simply uses half the number of wire sections and half the number of capacitors as a center-fed horizontal CCD. Hence, the wire will be 1/2 wavelength long. We shall split the 26-section, 24-capacitor CCD and create a 14-section, 12-capacitor CCD.
The appropriate comparator for this antenna is a simple 1/2 wavelength monopole. The base feedpoint will use a matching section to arrive at a near-50-Ohm impedance. Because the natural impedance of a base-fed 1/2 wavelength monopole is lower than the impedance of a horizontal center-fed wire, we must use a parallel transmission line with a lower characteristic impedance. 450-Ohm line works well here, and a 68' length allows us to arrive at the desire impedance level. Again, the models use a velocity factor of 1.0, but an actual line would use the actual velocity factor for the selected line. As well, the variables of installation will likely require some experimentation to find the correct length.
The simplest way to model both antennas is to use a MININEC ground. The presumed advantage of using this ground system is that it allows direct contact between the lower end of the vertical wire and the ground without need for modeling radials. In some contexts, this system can be useful, but not in this case. In fact, the MININEC ground obscures some critical differences between the operation of a CCD and a simple 1/2 wavelength base-fed wire.
To use a NEC-4 Sommerfeld-Norton ground with the antenna wire touching the ground requires that we install some kind of wire below ground. The simplest below ground wire might be a simulation of an 8' ground rod. Like the MININEC ground, this treatment is applicable to both antennas. Whether such a treatment is advisable is something that the data will tell us.
Alternatively, we can install a buried radial system using as few or as many radials as the analysis might dictate. The data in Table 5 show 4-, 16-, and 32-radial systems, using 70' AWG #14 copper radials, for the CCD. The 1/2 wavelength wire only uses a 4-radial system for reasons that become apparent in the data.
Each antenna entry set begins with the MININEC ground. If we were to use this data, we would reach the conclusion that both antennas perform very similarly, with a negligible difference in gain and only a 2-degree difference in the TO angle. However, if we change to the S-N ground with an 8' ground rod, we obtain very different results. With no change in the TO angle, the plain wire outperforms the CCD by about 2.6 dB, a noticeable amount.
To improve performance of the CCD, we must replace the ground rod with radials. The table shows the results for radial fields of various sizes. For the CCD, adding 4 radials increases the gain by over 1.5 dB. Only when we have 32 radials does the CCD challenge the gain of the plain 1/2 wavelength wire. The data for the plain wire shows that replacing the ground rod with 4 70' radials only increases gain by 0.1 dB. The small gain increment tends to suggest that a radial system is optional with a 1/2 wavelength wire. Fig. 9 overlays the plain wire with ground rod pattern and two CCD patterns: one for the ground-rod model and the other for the model with 16 radials. The patterns reveal not only the gain differences, but also the TO angle differences for the two types of antennas.
The difference in the radial requirements between the plain wire and the CCD emerges from the different current distributions on the two types of antennas. The sketch on the left in Fig. 10 shows the distribution along the 1/2 wavelength simple wire monopole. The current maximum occurs at the wire center, which is elevated. Hence, the antenna shows a lower TO angle than the CCD. As well, the current reaches a minimum at the base of the antenna. Essentially, the element is complete at that point and requires only a good RF ground. The 4 radials provide a better distributed RF ground than the simple rod, but the performance difference is small.
In contrast, the CCD design places a current maximum at the base of the monopole installation. Maximum efficiency requires an effective completion of the antenna in the form of a low-loss radial system. The more radials that we have, the more efficient will be the total antenna system, as indicated by the declining resistive component of the feedpoint impedance as we add more radials. Note, however, that as we bring the system to a level of maximum efficiency, the TO angle actually increases. At a high efficiency level, the feedpoint would need a 2:1 impedance transformation device or network for compatibility with the standard 50-Ohm coaxial cable.
An effective monopole CCD system thus requires two forms of construction complexity relative to the plain 1/2 wavelength wire: the installation of capacitors along the monopole and the burying of a large number of radial wires. In contrast, the plain 1/2 wavelength wire needs no modification of the monopole wire and provides good service at a lower TO angle with only a ground rod or the simplest of radial systems. However, the high impedance of the antenna may require a dedicated network or a carefully tuned matching section to arrive at a 50-Ohm impedance.
How "Small" Can I Make a CCD?"
The terms "large" and "small" in connection with the discussion of CCDs in these notes refer to the number of wire sections and the number of capacitors needed to make up a 1 wavelength antenna. In our initial free-space modeling experiments, we saw that there is no significant difference in the performance of a 26-section, 24-capacitor CCD and a 50-section, 48-capacitor version. Since the smaller of the two systems requires less work than the larger, we may naturally pose a question: how small can I make the CCD antenna and still achieve proper performance. Since the goal is a self-resonant 1 wavelength antenna, we have a criterion for success.
Models of the free-space center-fed CCD seemed unable to achieve resonance with 6 or fewer capacitors. However, an 8-capacitor, 10-section model of the antenna did achieve resonance. The feedpoint impedance was considerably lower than the 200-Ohm target that we have used in order to apply a 4:1 balun at the feedpoint. As well, the gain was down slightly relative to larger models, but not enough to be operationally significant. Table 6 shows the results obtain from several different models leading up to the 26-section, 24-capacitor model.
Note that between the worst and the best of the model set, we have only a 0.15-dB difference in gain. As well, the beamwidth changes by only 3 degrees, indicating a stable E-plane pattern. However, the feedpoint impedance climbs from about 170 Ohms to well over 200 Ohms. The differences in the gain levels and the feedpoint impedance values result from the fact that the current distribution curve changes as we move from fewer to more wire sections and capacitors. Fig. 11 provides a sense of the evolution of the distribution curve from essentially a triangle to a smooth curve.
The shapes of the current magnitude distribution curve do not materially affect the shape of the radiation pattern, which remains well-behaved and bi-directional. The absence of any significant change in the beamwidth is a further indication that even the 8-capacitor version of the antenna will perform well.
The answer to the section's lead question then is that about 8-capacitors and 10 wire section form the smallest practical CCD with full performance and resonance. It is likely that the precision of the match between capacitors and between wire sections will play a stronger role for the small CCD than it does for larger versions.
Conclusion
We have taken a short look at the controlled current distribution (CCD) center-fed 1 wavelength antenna with an eye toward understanding its operation and assessing its virtues. The center-fed version of the antenna provides a 200-Ohm resonant feedpoint impedance with the choice of the correct capacitor size, compared to the 70-Ohm impedance of a 1/2 wavelength dipole and the very high impedance of a plain center-fed 1 wavelength wire. The CCD gain and beamwidth values fall between those of the two antennas used as comparators. Models suggest that CCD performance peaks with about 26 wire sections and 24 capacitors. However, CCDs as "small" as 8 capacitors and 10 wire sections may work satisfactorily.
The vertical monopole version of the antenna is perhaps more problematical, since when ground mounted, it requires an extensive radial system for most efficient operation. Unlike the plain 1/2 wavelength wire monopole that has an elevated point of maximum current, the CCD vertical monopole reaches maximum current at ground level and thus requires radials to complete the radiating structure.
Essentially, the CCD simulates with discrete components a continuously loaded element with 501 pF/meter. At this loading level, the gain in free space is 3.10 dBi, with a feedpoint impedance of 245 - j0.2 Ohms. The performance reports are virtually identical to those that emerge from the 26-section, 24-capacitor antenna that used 420 pF capacitors.
The CCD is a viable and potentially useful antenna of its type. Whether the advantages warrant the relatively complex construction compared to simpler wire antennas is a user judgment.
Updated 02-26-2006 © L. B. Cebik, W4RNL Data may be used for personal purposes, but may not be reproduced for publication in print or any other medium without permission of the author.
Return to Amateur Radio Page