A New Approach to the Analysis of Adjacent Structure Effects on HDTV Antenna Performance
[ MULTIPLE DTV/NTSC ANTENNA SYSTEMS DESIGN ]     [ NTSC COVERAGE AND DTV SERVICE ANALYSES ]     [ INTERFERENCE ANALYSES ]     [ TRANSMITTER-TO-TRANSMITTER ISOLATION ]     [ NTSC GHOSTS FROM TALL TOWERS AND BUILDINGS ]     [ CHANNEL COMBINING AND FILTERING SYSTEM DESIGN ]     [ RF RADIATION HAZARD ANALYSIS ]     [ TECHNICAL PAPERS ]     [ PATENTS ]     [ ABOUT ]     [Back]

0. Bendov
Dielectric Communications
Cherry Hill, NJ 08003

Introduction
In the US, many television towers support more than one TV antenna. In major markets these towers were designed to support up to 10 TV antennas and additional FM antennas. Some of these antennas are side-mounted below the top platform. Others are mounted on the top platform. Yet other antennas are mounted on separate towers that may be separated by no more than a 100 feet.
During the 60's and 70's, RCA invested substantially in developing the software that was required for multiple antenna tower design. The challenge was not limited to the expected service degradation resulting from the interaction among the antennas themselves and their interaction with the supporting tower. The challenge extended to include earthquake and wind gusting effects.
As part of the RCA program, operating scaled-down models were built to verify the accuracy of the patterns predicted by the software. For example, the entire top platform of the San Francisco tower with 9 TV antennas was built to 1/10 of its actual size.
The RCA software, developed for the design of top-mounted and side-mounted NTSC-TV antennas, has proved itself through repeated use. Unfortunately that software, in its present form, cannot be used for HDTV design.
In this paper, the method used for the analysis of adjacent structural effects on NTSC-TV antennas will be briefly reviewed. The failure of the NTSC approach to meet HDTV requirements will be outlined, and the proposed procedure for estimating the power penalties and the resulting loss of coverage for HDTV antennas in the presence of interfering structures of will be presented.
NTSC
The analysis and design of antennas for NTSC transmission in the presence of adjacent structures has been based on the minimization of three performance degradation factors as a function of azimuthal angle:
  1. Azimuthal pattern distortion at the carrier frequency due to scattering and reflection from adjacent structures.
  2. Gain variation across 4 MHz, sometimes called "video response."
  3. Visible echoes of a > 25 usec delay due to interception and reradiation by other transmitting antennas and towers sharing the same aperture.
Figure 1 is a top view of three antennas sharing the same aperture in a candelabra fashion. The antennas are mounted on a triangular platform and are separated by 50 feet from each other. The channel 4 operating antenna transmits circular polarization and is marked as TDM. The other two antennas are obstructions to the TDM antenna.

The azimuthal patterns in Figure 1 are those of the TDM antenna with the interfering antennas (solid line), and without (dashed line). These patterns are normally plotted only at the carrier frequency of the operating antenna because, during program transmission, most of the picture energy is concentrated around the carrier.
In effect, NTSC transmission is a relatively narrow band transmission even though three separate carriers occupy the 6 MHz channel.
The distorted pattern (solid line) shows that in some directions the Effective Radiated Power (ERP), in the presence of the other antennas, increases over the authorized ERP. In other words, the reflections are assumed to add desirable energy so long as the associated
picture degradation is acceptable.
Just what is an acceptable degradation has not been well defined. Over the years, two rules of thumb have been successfully used to define what is an acceptable picture distortion due to reflection from adjacent structures:
  1. Gain variation limited to < 2 dB over 4 MHz in the critical directions.
  2. Visible echoes prevention by limiting the separation between the operating antenna and the interfering structures to a maximum of 100 feet (< .2 usec delay for a round-trip) and by the insertion of filters in the feed lines of the interfering antennas.
HDTV vs. NTSC
None of the three degradation factors used in the design and analysis of NTSC installations are useful in the design of installations for digital transmission of HDTV where no impairments are visible and where all the distortions and undesired signals, such as reflections, are manifested as reduced service.
The reduction in service can be evaluated from assessing a power penalty against the desired signal for each form of impairment. The sum of the penalties, in dB, is the ratio by which the ERP must be increased in order to restore the intended service. If the ERP cannot be increased, the power penalty translates into a shrinking coverage contour.
Implicit in the application of the power penalty method is that the post-equalizer Bit-Error-Rate (BER) at the HDTV receiver does not exceed 3xl0-6. That may not always be possible under conditions of heavy interference from cochannels, adjacent channels, and impulse noise.
Not only are the three NTSC degradation factors not useful for HDTV, but the use of field strength to define service contours (e.g. Grade B = 64dBu) is not applicable. This point can be understood by comparing the spectrum of NTSC with that of HDTV (without the pilot tone) as shown in Figure 2. As seen on a spectrum analyzer, the NTSC signal shows the precise locations within the 6 MHz channel of the picture, color and aural carriers and their relative magnitude. In contrast, the HDTV spectrum, prior to being subjected to multipath and interference, is flat over the channel. That is, no specific carrier can be defined in the same manner as is done in the case of NTSC signal.
Accordingly, the useful HDTV "carrier" can only be defined as the total signal power, in dBk, within the 6 MHz channel minus the power lost to the in-band uncorrectible distortions such as intermodulation products.
While the total signal power is useful in determining HDTV coverage contours limited by thermal noise, it is not useful in determining service contours when multipath, impulse noise and interferences are present.
The limited use of the total HDTV signal power for defining service contours in the presence of multipath and interference can be explained by referring to Figure 2 again. In Figure 2, the effect of multipath on both the NTSC and the HDTV spectra is shown as a notch. In the case of NTSC, the notch, located between the Visual carrier and Color subcarrier, has relatively little effect on the quality of the received picture, and no effect on the sound quality. In contrast, the notch in the HDTV spectrum will cause a loss of picture and sound even though the total signal power within the 6 MHz channel has suffered only a minor change. Even though the notch may result in the loss of picture and sound, the total signal power, represented by the area under the spectrum curve, remains essentially unchanged.
Whereas the equalizer at the HDTV receiver may be able to correct the multipath notch if it is not too deep and if it is not random in its position within the channel, it cannot do so if the notch (or notches) were to result from impulse noise and interference as they vary randomly and fast, anywhere inside the channel. As the equalizer attempts to correct the distorted channel by raising the gain at the deficient frequencies, it inflicts a penalty on the HDTV signal that must be incorporated in the definition of the service contour.
HDTV
In the HDTV environment, the reflections and scattering of the primary signal due to the obstructions are viewed as undesirable multipath*. In other words, the structural obstructions to the transmitting antenna are best viewed as a filter whose attenuation and frequency response vary with the angle toward the receiver. The effects of the signal lost and frequency distortion due to obstructions in the vicinity of the transmitting antenna, can be quantified as power penalties on the desirable signal.
Penalty Due to the Signal Power Lost to Reflections and Scattering
The difference between the CNR (Carrier to Noise Ratio) of the HDTV channel with and without the obstructions is (Ref. 1):
where:
(CNR)FS = the available CNR when the operating antenna is in free-space in dB.
fB = channel bandwidth.
= channel transfer function in the direction of the receiver and in the presence of the structural obstructions.
The difference between the available CNR in free space and that available in the presence of structural obstructions is the power penalty, in dB.
It is clear from the power penalty equation that the reflections, as well as the free-space antenna patterns, must be processed over 6 MHz rather than at the carrier frequency alone.
Penalty Due to Frequency Distortion in the Passband
The second power penalty on the available CNR is imposed by the equalizer at the receiver. For example, the equalizer attempts to compensate for the notch in Figure 2 by introducing an infinite gain at that frequency. This compensation is at the expense of enhancing the noise level and thereby reducing the available CNR.
The magnitude of the penalty depends on the choice of equalizer at the receiver. Peak Distortion, Mean Square Error and Decision Feedback are examples of known equalizers (Ref 2).
While the exact magnitude of the noise power penalty depends on the choice of equalizer, the worst case is:

where:
N - NFS is the power penalty inflicted by the equalizer on the noise level due to the distortion of the frequency response.
It is clear that any notches in the transfer function will cause the equalizer to raise the noise level to a point where the threshold CNR cannot be maintained, resulting in lost picture and sound.
CNR Coverage Contours
Since knowledge of the power penalties, in addition to the Effective Radiated Power (ERP), are essential in defining coverage contour of obstructed antennas, it would appear that a contour of threshold CNR, rather than field strength or power, will serve as the proper tool for coverage description. The two power penalties and the free-space ERP may be automatically incorporated within one number, the threshold CNR. The threshold CNR (~15 dB) contour can then be drawn for various percentages of time and location availability depending on the margin between the available CNR and the threshold level.
HDTV Example
This example is based on an existing threelegged tower, with a hypothetical, omnidirectional HDTV antenna, mounted 8 feet due east from the line connecting legs B and C. The top view of tower cross-section and the HDTV antenna is shown in Figure 3. The width of the triangular legs is 9 feet and the separation between the legs is 63 feet. The calculations are all at channel 39.
The first step is to calculate the reflections and scattering from the three legs. The result is shown in Figure 4 where both the free-space (magnitude=l for omnidirectional pattern) and the reflection azimuthal patterns are shown. In particular, note the large, undesired, reflection due East from the flat surface of leg A. Also note that the patterns were calculated and integrated over 6 MHz in accordance with the principles set forth in this paper.
The next step is the assessment of the two power penalties. The results are shown in Figure 5. The total penalty is the sum of the penalty due to loss of primary signal resulting from the reflections and the penalty due to the distorted frequency response, which is also a by-product of the reflections. Note that the penalty due to the equalizer is relatively small since no notch such as that shown in Figure 2 was produced.
Finally, the calculated threshold CNR coverage contour is shown in Figure 6. Note the effect of the power penalties on the reduced coverage in some directions. This particular contour is based on the FCC propagation curves and the following receiver model:
Rx Noise Figure = 10 dB.
Rx Antenna Height = 30 feet.
Rx Antenna Gain = 10 dB.
Downlead Cable = 50 feet of RG-59/U.
Rx Antenna Balun Loss = .5 dB.
Threshold CNR = 15 dB.
Inside the contour, the CNR is > 15 dB which is the threshold for yes/no HDTV. Finally, note that the contour is based on FCC(50,99) rather than the FCC(50,90) curves which are normally used for NTSC coverage analysis. The justification for the higher percentage of time availability in the case of HDTV broadcasting was explained earlier (Ref 3).
Conclusion
Most US broadcasters will face the transmission tower capacity bottleneck as part of their transition plan to HDTV. In many locations, HDTV antennas will be installed next to other antennas or below the top of the tower. The HDTV antennas will then be subjected to reflections from nearby antennas and/or the supporting tower itself The procedures outlined in this paper provide the means for quantifying the degradation, in any direction, as a function of the HDTV antenna position in the presence of undesirable structures.

* It is possible that in some directions, the reflections could be "useful" energy if they assist in equalizing the overall frequency response.

References
1. Rummler, W.D, "A Simplified Method for the Laboratory Determination of Multipath Outage of Digital Radio in the Presence of Thermal Noise," IEEE Transactions on Communications, Vol. COM-30, March 1982.
2. Proakis, J. G. Digital Communications, 2nd edition, McGraw-Hill, 1989.
3. Bendov, 0. "The Effect of Channel Assignment on Transmitter and Receiver Requirements for Equivalent HDTVNTSC Coverage," NAB Proceedings, 1994.
Back