Designing an All-Channel TV Antenna, February 1966 Electronics World

Yes, the old days of over-the-air broadcasts and analog television could be a pain. Perfect adjustment of the antenna and TV controls one day could be totally useless the next day or even later the same day. Atmospheric and physical variations can change suddenly and significantly, affecting both radio and television. Proper separation and processing of horizontal and vertical synchronization of the video, color and intensity, and audio by the TV's electronic circuits depended on the right combination of antenna, and lead-in cable. The advent of semiconductors in place of vacuum tubes helped stabilize the television's role in viewing quality and lessened the overall irritation level, and the introduction of cable-based and then satellite-based programming distribution to reduce irritations to the point where most people never had a problem. That said, I would happily return to the days of yore and suffer the aforementioned inconveniences, to enjoy a time when the content was significantly less rude and crude, and patriotic and traditional family-based shows were by far the rule rather than the exception. More than two decades have passed since Melanie and I watched broadcast TV of any sort with the exception of a couple programs - "24," "Chuck," and "NUMB3RS." These days we only watch shows from the 1950's through 1980's on DVD sets. Our sole TV is a 26", circa 2007 model with a built-in DVD player. It's not the technology we reject, it's the content.

Evolution of the design of an antenna that will cover both v.h.f. and u.h.f. TV channels and that employs but a single downlead.

The emergence of u.h.f. television broadcasting brought about by Federal legislation has presented antenna designers with a new challenge. The advantages of an all-channel (v.h.f. plus u.h.f.) antenna in areas with intermixed telecasting are apparent. Only one lead-in need be run from antenna to set. Installation time is reduced since it is not necessary to mount separate v.h.f. and u.h.f. antennas. The possibility of scattering from one antenna affecting the performance of the other is eliminated. Although separate v.h.f. and u.h.f. input terminals are still present on most sets now available, single-input sets will be the inevitable result of improved tuner design. For multiple-input sets, efficient signal splitters can be used at the receiver to provide separate v.h.f. and u.h.f. signals (also to drive an FM tuner).

Fig. 1 - (A) Resistive and (B) reactive components of the input impedance of a center-fed linear dipole antenna element.

Fig. 2 - Radiation patterns of linear dipoles (center-driven to produce in-phase currents in both halves of the element) having various lengths and with l/d ratio greater than 100. Antennas are indicated by the dark horizontal lines shown.

The horizontal scale shows the ratio of dipole length (l) to wavelength (λ). The vertical scale on the upper graph is the resistive part of the impedance; on the lower graph, the reactive part. Note that only near the resonances of the dipoles which occur approximately at l/λ = (2n+1)/2(n = 0,1,2, etc.) is the impedance almost totally resistive and at the same time small enough to be easily matched to the resistive characteristic impedance of a common low-loss transmission line. The frequency band over which the reactance can be maintained at low values is dependent upon the cross-section of the dipole. However, even for very "fat" dipoles as indicated by small values of the length-to-diameter ratio, l/d, the variation in input impedance may not be tolerable. The relative bandwidth is often defined as the difference between the frequencies at which the reactance is equal to the resistance at resonance divided by the geometric mean of these frequencies. The curves of Fig. 1 correspond to a length-to-diameter ratio of 100. The relative bandwidth at the first resonance is approximately 0.18. For l/d = 300, the bandwidth is 0.15.

Not only does the impedance of a single dipole change considerably with frequency but the radiation pattern as well. Fig. 2 shows some graphs of relative response of linear center-fed dipoles to signals arriving from various directions around the antenna. As long as the dipole length is less than one-half wavelength, the pattern is a characteristic figure-of-eight and does not change much with frequency. However, as the l/λ ratio continues to increase, the pattern shape becomes multi-lobed and changes rather rapidly as frequency changes. These pattern variations are generally undesirable since the usual requirement for maximum signal reception is to have large response in a single direction and minimum response from all other directions.

The problems of maintaining a resistive impedance and unidirectional pattern over extremely wide frequency bands have been subjects of extensive research. Considerable progress in this area has been made at the Antenna Laboratory of the University of Illinois since 1954. One of the most useful designs to come out of this work is the "log-periodic" dipole array which was developed by D. E. Isbell in 1959. A schematic diagram of this antenna is shown in Fig. 3. The antenna is composed of a number of linear dipoles which are all connected to a common transmission-line feeder. The distinguishing feature of this array is the fact that the dipoles are all cut to different lengths with the length gradually diminishing from one end of the feeder to the other.

Note that the spacings between dipoles likewise decrease from a largest value at the longest dipoles to a smaller value at the shortest dipoles. When properly controlled, this tapering of dipole lengths and spacings is capable of minimizing variations in input impedance and radiation pattern over extremely wide bandwidths. The theoretically ideal situation is for the ratio of lengths of any two adjacent dipoles to be constant throughout the array and for the ratio of spacings between adjacent pairs of dipoles to be given by the same constant. As in most practical situations, some departures from the ideal are permissible before the antenna performance is seriously affected but, generally speaking, the closer the practical design to the theoretical ideal, the smaller the performance variations with frequency.

The operation of the log-periodic dipole array is similar to that of a stagger-tuned amplifier in that the dipoles are resonant at various frequencies (depending upon the length) and yet the overlapping responses of the multiplicity of elements produces a total effect which changes little with frequency. The resonant frequencies of the several dipoles in a typical log-periodic dipole array are shown in Fig. 3. Conventional linear dipoles for the v.h.f. band are constructed from 3/8-inch diameter aluminum tubing with maximum lengths of approximately 111 inches. Thus the largest l/d ratio encountered in these antennas is approximately 300. Hence, a single dipole which is resonant at 69 mc. in the middle of the low v.h.f. TV band will have a bandwidth of 0.15 x 69≅10 mc. This is adequate to cover only two channels and even that at a reduced performance on the band edges (64 and 74 mc.) where only80 percent of the received power would be delivered to a load which is ideal at resonance.

Fig. 3 - Log-periodic dipole array for low-band v.h.f. TV.

Fig. 4 - Feeder voltage and dipole current phasors illustrating the effect of transposition of the feeder that is used.

The manner in which the interconnection of several dipoles of different length can overcome the problem of impedance variation can best be understood by considering the antenna to be operating as a transmitter. A basic result of antenna theory states that the input impedance and radiation patterns of an antenna are the same for transmitting and receiving. In the transmitting case, a generator is connected to the feedpoint as shown in Fig. 3. Let us suppose that the frequency of the generator is such that l₁/λ = 0.5 where l₁ is the length of the longest dipole. The dipoles near the feedpoint will have considerably smaller values of l/λ and Fig. 1 shows the input impedance of these dipoles to be primarily a high reactance. Little energy will be transferred to these dipoles from the feedline. However, as the wave traveling along the feeder encounters longer dipoles, more and more energy will be transferred. One secret of a successful log-periodic design is to insure that most of the energy is coupled into the dipoles before reaching the end of the feeder. The termination at the last dipole, Z_l, can be important in this regard, particularly for an antenna containing only a few dipoles. It is preferable to use a reactive load across the longest dipole since any resistive component will absorb power and reduce the antenna gain. This load may take the form of an open- or shorted-stub, sometimes in conjunction with a small capacitor.

The several dipoles which absorb the greater power from the feeder are often called the "active region." Consideration will be given in the next section to the manner in which the currents in these dipoles can be made to radiate a highly directive beam. But now consider what happens when the frequency is changed. In particular when the next shorter dipole corresponds to l/λ = 0.5, the active region has simply moved up the feeder by one dipole spacing. The input impedance at the front of the active region is the same as before and the feeder plus shorter dipoles between the active region and the feedpoint all have the same or approximately the same wavelength dimensions as before. Hence the input impedance at this higher frequency will be the same as at the lowest frequency. Since the dimensions of the active region are the same in wavelengths, the radiation pattern produced is also unchanged.

Although the above considerations of scaling indicate that the performance characteristics of the antenna will repeat at any two frequencies related by the scale factor for adjacent dipoles, there is no guarantee that the performance will remain constant between these frequencies. This must be accomplished by proper choice of scale factor, depending upon the bandwidth of the individual dipoles in the array. For dipoles with l/d in the range from 300 down, scale factors as low as 0.9 have proved satisfactory. When scale factor and dipole bandwidth are properly related, then the active region moves smoothly from dipole to dipole as frequency changes and variations in antenna performance over an extremely wide frequency band are very small.

Although a smoothly moving active region is very important in achieving frequency-independent performance, there are other important considerations as well. The single-lobe radiation pattern which has a high ratio of front -to-back response requires that several dipoles be interconnected in the proper manner. Consider the transmitting case once again. The generator connected to the input terminals of the antenna launches a traveling wave on the feeder. The voltage on the feeder undergoes a phase delay as an observer moves away from the feedpoint. The phase shift between dipoles is increased above the value obtained on an unloaded feeder because of the additional capacitive reactance which the short dipoles present across the line. Thus, although the spacing between dipoles in the active region usually corresponds to less than 60 electrical degrees, the feeder voltage phase shift is more nearly 90 degrees.

For determining the pattern of the array, it is the dipole currents which are of prime importance. The transposition of feeder conductors between adjacent dipoles introduces an additional 180 degrees of phase shift between dipoles as compared to using a straight feeder. The combination of these effects is to produce a phase delay in the dipole currents which is toward the feedpoint of the array rather than away from it. As a result, the direction of maximum radiation ( as a transmitting antenna) and maximum response (as a receiving antenna) is along the axis of the array in the direction of the feedpoint. This phenomenon is often called "backfire" radiation to distinguish it from the more conventional "end-fire" antennas, such as multiple-element yagis. Yagis have maximum response off the end which is farthest removed from the feedpoint.

Fig. 5 - Adaptation of array for moderate gain on two bands. Note the use of the two groups of antenna elements in array.

Fig. 6 - An all-channel antenna using half-wave dipoles for low v.h.f. band and 3/2-wave dipoles for high v.h.f. and u.h.f. bands.

Fig. 7 - All-channel antenna using capacitor-loaded dipoles.

Fig. 8 - An all-channel, 15-element antenna using capacitor-loaded dipoles. Capacitance and load values were determined empirically for best performance during original design work.

The manner in which the "backfire" phasing of the dipoles is achieved is illustrated in Fig. 4 by the arrows labeled "feeder voltage phasors" and "dipole current phasors." The data shown in Fig. 3 for feeder voltage phases is for a frequency near the center of the low v.h.f. TV band so that the active region is located at the intermediate-length dipoles. The feeder voltage phases at the dipoles are characteristic of a wave traveling away from the generator, i.e., phase increasing toward the left. This is shown in Fig. 4A by the rotation of the phasors in a counterclockwise direction. Due to the transposition of the feeder between adjacent dipoles, the dipole currents in alternate dipoles have a phase which is advanced (or retarded) by approximately 180 degrees. Reversing the direction of alternate phasors, as shown in Fig. 4B, the direction of phase progression in the dipole currents is opposite to that of the feeder voltage. An important feature of backfire phasing is that the radiation is directed away from the longer dipoles and the result is a high front-to-back ratio. To preserve the high front-to-back ratio it is necessary to observe a minimum feeder length of approximately one-third to one-half the maximum wavelength. Minimum front-to-back size for a backfire antenna with a good front-to-back ratio at channel 2 is, therefore, about 60 inches.

Although the combination of tapered dipole lengths and spacings plus backfire phasing is ideal for continuous coverage of an extremely wide band, the frequency allocations for television do not call for this type of operation. In fact, for an all-channel antenna, it is necessary to cover three separate frequency bands, 54-88, 174-216, and 470-890 mc. There are several approaches to adapting the above principles for this type of intermittent coverage, however, and we shall consider briefly the advantages and disadvantages of some of them. Of course, it would be possible to build a log-periodic dipole array which would cover the entire band from 54 to 890mc. even though this represents a frequency span of over 16 to 1. The disadvantage to this approach is that many dipoles would be required and the total length of the feeder would be rather large even for achieving moderate antenna gain. It was demonstrated early in the work on log-periodic dipole antennas that the dipoles which were resonant at frequencies somewhat removed from the bands which were to be covered could be removed from the antenna without adversely affecting the performance in these bands. While it is true that this will eliminate several dipoles and thereby shorten the required feeder, as shown in Fig. 5, the gain remains relatively constant for all frequencies in the desired bands. This may produce satisfactory reception in some locations, but generally the properties of wave propagation with increasing frequency make it desirable to have larger gain on the upper frequency bands.

The dipole impedance characteristics shown in Fig. 1 suggest a way in which each dipole on the antenna can be made to operate effectively at more than one frequency. Since the formation of the active region is primarily dependent upon the impedance properties of the dipoles, it is reasonable to expect that an active region could be formed in the vicinity of dipoles which are three-half-wavelengths long. As a matter of fact, even higher-order resonances which occur at other odd integer multiples of a half-wavelength could be used for this purpose. In order to achieve an active region which moved smoothly with frequency, only an appropriate adjustment in the spacing between dipoles was required; the higher the order of the resonances, the closer the spacing.

Inspection of the patterns of Fig. 2 reveals another problem with operating in the higher-order resonances. Note that the pattern of a linear dipole which is three-half-wavelengths long has six distinct lobes rather than two. Considering the transmitting case once again, it is easy to see that the power radiated in the side lobes is wasted. The side lobes can be greatly reduced by shaping the dipole. The simplest configuration is a V-dipole made by tilting the dipole halves toward one another. The combination of V-dipoles in a driven array with backfire phasing can be made to have no observable side lobes and a front-to-back ratio greater than 20 decibels. Since the dipole is larger in wavelength when used at the higher resonances, the beams produced are much narrower, resulting in the increased gain which is desired in the higher frequency bands.

Fig. 6 shows one way in which the higher modes can be utilized in an all-channel antenna. In this example 3/2-wave dipoles are used throughout the high v.h.f. and u.h.f. bands. Since the maximum theoretical bandwidth of a 3/2-wave antenna is 3 to 1, further modification of the dipoles is required. The desired result is achieved through the reactance-loading discussed in the following section.

Consideration of the resonances of the dipoles shown in Fig. 6 suggests a way in which performance of the antenna might be improved, especially in the high v.h.f. band. Note that the 3/2-wave resonances of several of the dipoles fall at frequencies outside the 174-216-mc. band. Hence, although these dipoles are required on the antenna to adequately cover the 54-88 mc. band, they are not contributing effectively toward achieving the higher gain which is desired in the upper v.h.f. band.

In order to force more of the dipoles to resonate within the 174-216-mc. band, a way is needed to cause the ratio of first to second resonance frequencies to depart from the customary value of 3. Recent research has shown that proper placement of a lumped reactance along the dipole arms can be made to produce the desired shift. When a capacitor is inserted in the dipole, the frequency variation in the reactance produces a different effect at the second resonance than at the first. Ratios as low as 2 have been achieved in this way. The antenna shown in Fig. 7 illustrates how more of the capacitance-loaded dipoles can be made to resonate in the higher v.h.f. band. The more active dipoles, the greater the gain. An additional benefit of the capacitor is to cause the resonances to occur at longer dipole lengths. Greater length produces greater gain.

Reactance loading can be used to achieve yet another result. The directivity of 5/2-wave dipoles is greater than 3/2 -wave and that of 7/2-wave greater than 5/2-wave, etc. Hence it would be desirable to use these longer dipoles to cover u.h.f. frequencies. Since the wavelength is so much shorter at u.h.f., operating in even these higher resonances does not require an inordinately long dipole. However, the tilt of the dipoles for the higher modes must be increased to maintain low side lobes. A section of dipoles which operates on the higher modes for u.h.f. is easily added to the front of a section designed to operate in no higher than 3/2-wave resonance at v.h.f. and can, in fact, be integrated to enhance the v.h.f. performance.

To achieve this, however, an adaptation of the linear dipoles is required. Ordinarily the longer u.h.f. dipoles would resonate at frequencies in or near the v.h.f. band. This would prevent the less sharply vee'd dipoles from receiving any energy from the generator (considering the transmitting case). It is possible to effectively disconnect the larger dipoles from the feeder at v.h.f. frequencies by once again inserting a capacitor in the dipole. The capacitor (formed merely by a break in the element) is designed to have sufficiently high reactance to decouple the sections of the dipole at v.h.f. frequencies and yet low enough reactance to couple at u.h.f. frequencies. The outboard portion resonates above the upper end of one of the v.h.f. bands and therefore will enhance the v.h.f. gain in the manner of a conventional parasitic element. Fig. 8 shows an antenna which makes use of these latter modifications, along with radiation patterns (Fig. 9) which have been measured in each one of the three bands. The reduction in beamwidth illustrates the increased directivity and gain which is achieved by operating in the higher modes.

Fig. 9 - Horizontal directivity patterns of fifteen-element antenna.

Although the use of a well-designed, all-channel antenna will alleviate the difficulties often encountered in a multiple antenna installation, there are still several precautions which should be observed to insure realization of the full performance potential of the antenna. Experience with v.h.f. installation is no guarantee of equal performance on u.h.f. since the higher frequencies behave quite differently. The first rule of a good u.h.f. installation is to use best quality, low-loss lead-in. This is essential for u.h.f. but will also increase the v.h.f. signal level at the set. Encapsulated lead-in is recommended for general use. Low-loss stand-offs which hold the lead-in away from any metal are best for u.h.f. In those cases where electrical interference is a problem, a low-loss shielded twin-lead should be used.

If amplification is required on an all-channel antenna, be sure to use an all-channel amplifier. The u.h.f. signals are generally attenuated severely by a v.h.f. amplifier and vice versa. If only u.h.f. or only v.h.f. amplification is desired, the amplifier can be used on the output of the signal splitter.

The narrow patterns of the all-channel antennas require careful orientation of the antenna toward the station. If stations are located in more than one direction, a good quality continuous-turning rotator is a wise investment.

With the rapid growth of u.h.f. and color television with their more stringent performance requirements, most antenna installations which are more than three years old are probably not adequate. The new all-channel antennas provide a good answer to the question of how to improve the reception of today's television programming.

Designing an All-Channel TV Antenna February 1966 Electronics World

Designing an All-Channel TV Antenna
February 1966 Electronics World