|
0. Bendov
|
|
Dielectric Communications
|
| Introduction |
| The mandated transition from analog to digital television has created an installation challenge in all major markets. The number of transmitted channels will have to be doubled for a yet unknown period while new tower availability remains stagnant. The stagnation in new tower building is partly due to local opposition, as is the case in Denver, and partly due to lack of physical space, as is the case in New York City. Such circumstances present a challenge to consolidate antennas onto one tower and to consolidate several channels into one antenna. |
| The consolidation of antennas and channels poses severe technical problems. Some of the challenges, those related to HDTV transmission, are new and thus not well understood. Because of lack of practical field experience in designing collocated HDTV antennas a conservative design philosophy must be applied. |
| Design considerations such as group delay & amplitude response over 6 MHz vs. antenna-to-antenna spacing, Tx-to-Tx isolation and wind sway will be illustrated. |
| Design implementations in San Francisco, New York, Baltimore and Chicago will be shown. |
| Design Considerations |
| A. Wind Sway |
| High winds can produce significant variation in received SNR (Signal-to-Noise Ratio). The variation in SNR is particularly significant for antennas with built-in beam tilt of more than 1° as shown in Figure 1. |
| As shown in Figure 1, a typical antenna with built-in beam tilt of 2.5°, when subjected to wind sway of ±1°, would result in SNR variation of >15 dB below the horizontal. |
| Antennas on the same platform will most likely sway at different rates, thereby adding dynamic multipath to the main signal. |
| In analog TV, the AGC of the receiver may help to limit the visible effect to color variation. In digital TV, depending on the modulation used and severity of the sway, an undetermined number of viewers will intermittently lose picture and sound. |
| B. Null Fill |
| Nearby structures scatter a portion of the incident signal in three dimensions. A 10% scatter is not unusual. In Figure 2, 6% of the main signal is scattered toward the first null of the elevation pattern, which has been filled to some degree in the factory. |
| The amount of null-fill by the factory is important. Insufficient null-fill could result in zero signal at that angle when the magnitude of the scattered signal equals that of the null-fill. |
| An additional complication arises for digital TV. The closer the multipath is to the main signal, the worse the spectral distortion over 6 MHz. This will be discussed in more detail in section F. |
  |
| C Pedestal Height |
| The bottom radiator of antennas mounted on a wide platform, typically in a candelabra configuration, should be raised to clear the 1st Fresnel zone reflection as shown in Figure 3. The proper height varies with channel and with platform size. For a 50' platform, the pedestal height may vary from 6' for UHF channels to 12' for VHF channels. |
| D. Undesirable RF Leak Level |
| Transmitting antennas in close proximity to each other also act as receiving antennas. The received level will depend on the bandwidth of the receiving antenna and the number and proximity of the transmitting antennas. |
| As shown in Figure 4, the received signal will be intercepted at the transmitter, usually by a balanced filter. A portion of the intercept is absorbed in the filter's reject load and the remainder is reflected back and radiated as multipath. |
  |
| That is the extent of the problem for analog TV. In digital TV, there is an additional concern. |
| The directional coupler samples both the reflected and the desired signal as they travel toward the antenna. The coupler feeds the HDTV equalization circuit of the transmitter. The level of the reflected undesired signal should be held to below 40 dB. |
| E. Directional Pattern Distortion |
| The most common UHF-TV antennas are slotted-pipe antennas which, when designed for directional pattern applications, would typically require a "backbone" for structural support. Figure 5 shows a typical configuration of such an antenna mounted on a support pipe. |
| The support pipe, the transmission line to the upper antenna, and the other antennas on the platform each contribute to the reshaping of the "designed" pattern into the "as-built" pattern as shown in Figure 6. |
  |
| The left side of Figure 6 shows how the "designed" pattern is distorted by the "backbone" pipe and by the transmission line to the upper antenna. The right side shows the "as-built" pattern when the scatter from the other two antennas on the platform is included. |
| Note that the deviation of the "as-built" pattern from the "design" pattern is not uniform in all directions. It is generally far worse in the sector where the pattern significantly deviates from being omnidirectional. Therefore, omnidirectional antennas generally provide better performance in a candelabra configuration. |
| As was the case in Figure 2, the calculation is shown at a single frequency inside the channel. Single frequency calculation is adequate for analog TV which is a essentially a narrow band transmission. For digital TV, the calculation needs to be done across the channel and it must include both amplitude and phase information. |
| F. HDTV Power Penalty |
| In analog TV, the effects of shadowing, scattering and leakage among collocated antennas is manifested in two ways: variation of the ERP (Effective Radiated Power) from the designed level and video distortion. Video distortion could be a combination of lost color, lost resolution and multiple ghosts. |
| In digital TV there is no video distortion to contend with. Instead, the only visible degradation is a total loss of image and sound. The sum total of all the degradation must therefore be expressed in term of reduction of SNR. The reduction in SNR is identical to a power penalty on the available ERP. In other words, the raw ERP out of the antenna must be reduced by the penalty to obtain realistic coverage calculation. |
| To obtain the power penalty, the amplitude and phase response of the antenna are first evaluated across the channel. |
| Figure 7 is an example of the response of a perfectly omnidirectional antenna facing one other antenna 10' away. In addition to the modulation in ERP, there is also a variation in the amplitude/phase response in each direction. |
| Note that the maximum rate of variation in the phase response exactly coincides with the maximum rate of variation in the amplitude response. Further, the direction in which the maximum rate of variation occurs, 180° in Figure 7, is also the direction in which the scatter suffers the maximum delay. |
| A typical response of an omnidirectional antenna facing two additional antennas on a 50' triangular platform is shown in Figure 8. |
  |
| The antenna response can be translated into power penalty1. The accuracy of the penalty calculation will depend on the particular equalizer used in the receiver because that is where the spectral distortion penalty is exercised. |
 |
| Figure 9 shows the translation of the spectral response and the modulation of the ERP, shown in Figure 7, into two power penalties. It is the sum of the two penalties that constitutes the total power penalty on the radiated ERP. |
1 0. Bendov, "A New Approach to the Analysis of Adjacent Structure Effects on HDTV Antenna Performance." 1995 NAB Proceedings. |
