For most of the era of analog television, engineers calculated coverage by plotting at least eight radials, including one passing over the community of license. The terrain height between 3.2 and 16 km (2-10 miles) along each radial was averaged to calculate the transmitter site’s height above the average terrain. This height above average terrain was then used to determine the distance to the defined field strength using a graph in the FCC rules, which, through curve fitting, was reduced to formulas computers could use. For example, 64 dBuV/m is the field strength at the analog UHF Grade B contour, which defines a UHF station’s service contour. The distances along each radial were calculated and then connected to make the contour map.
As engineers who’ve studied propagation have pointed out, the results were usually quite accurate. Indeed, these FCC field strength graphs are still in the FCC regulations—47 CFR 73.699 Figures 9-10. They are still used to define the service area of TV stations. While there is no charter for the 50 percent location and 90 percent time variability used to calculate DTV contours, the FCC has provided a way to calculate DTV contours using the existing FCC contours. The latest FCC rules are available at ecfr.gpoaccess.gov. Select Title 47 and then navigate to Part 73, table of contents, Section 73.699, the propagation charts and the small links to PDF versions of the charts.
One reason the FCC F (50, 50) and F (50, 10) charts have survived so long is that they are based on field measurements throughout the country. As you will see in my next column the latest ITU propagation model for VHF and UHF also relies on tables based on measured values.
These charts served broadcasters well for many years, but they didn’t account for terrain along radials not studied and terrain less than 2 miles or more than 10 miles from the transmitter site wasn’t considered at all. The FCC curves provide a good way to determine general coverage area, but don’t allow a way to study coverage and interference at individual locations. When the FCC started looking at ways to provide every station a second channel for DTV broadcasting, the charts and contours weren’t sufficient to evaluate the tradeoffs between coverage and interference.
The FCC wisely rejected proprietary propagation models when considering how to determine analog and DTV coverage and the interference between stations. While it may not be easy to get it working on a personal computer, if you want you can download the source code for the programs the FCC uses for their analysis. The FCC program is based on the Irregular Terrain Model Version 1.2.2 from ITS in Boulder, Colo. The ITM source code, originally written in FORTRAN, is now available in the C programming language from the ITS Web site, flattop.its.bldrdoc.gov/itm.html. Most people refer to the ITM model by its creators’ names—Anita Longley and Phil Rice—or Longley-Rice.
In past columns, I’ve criticized the FCC’s implementation of Longley-Rice, while realizing some of the short cuts the FCC took in implementing ITM were necessary to keep the computer time required to do interference studies, which often involve calculating coverage for many stations, to a reasonable length. I summarized the limitations in my presentation “RF Delusions” at NAB two years ago. Slides are available at www.xmtr.com/rfdelusions.
It is important to recognize that while there may be problems with the FCC’s Longley-Rice implementation, providing more accurate inputs to the model will fix the many problems, and upon proper showing, the FCC is likely to accept the more accurate results.
IT’S ALL IN THE DATA
Before comparing propagation models, it’s important to understand how propagation software works. First, let’s consider the data available. We know the height of the transmitting and receive antennas and the frequency being studied. We know the terrain, from USGS or shuttle radar terrain mapping mission data. If calculating field strength in addition to path loss, we need to know the effective radiated power towards the receiver or obstruction in between. This is where the FCC software’s use of a standard elevation pattern, ignoring actual electrical and mechanical beam tilt, can cause problems.
If you listen to a distant radio station or watch a distant TV station over the air, you may have noticed reception varies with the weather. Longley-Rice and most other propagation models allow input of surface refractivity and the effective curvature of the Earth. A propagation model based on conditions that exist in a coastal environment is unlikely to perform well when calculating coverage in an inland valley. To keep things simple, the FCC Longley-Rice implementation is based on parameters for a “Continental Temperate” area. Other models, such as TIREM and ITU P1546-3, use different calculations for propagation over water and over land, including mixed paths that include both.
(click thumbnail)Fig. 1: Data entered into propogation software can predict the path loss and signal level at a specific point, like the path between the Mount Wilson antenna farm in Los Angeles and Valencia, Calif., shown here.Given this input data, propagation software can predict the path loss and signal level at a specific point. For line of site paths between outdoor antennas, the calculations are relatively simple. For obstructed paths, like the path between the Mount Wilson antenna farm in Los Angeles and Valencia, Calif. shown in Fig. 1, the calculations become much more difficult.
When a line of sight path is not available, the signal will be attenuated but enough may make it to the receive antenna to allow reception. Knife-edge diffraction over the first obstruction will allow some signal to make it over the obstruction. Ideally, the propagation model would know whether the obstruction is bare rock, a smooth hill, or a hill with vegetation on it. These conditions affect how much signal will make it over the obstruction. If no signal makes it over the obstruction, reception may still be possible using troposcatter, where the signal is reflected from the troposphere. That propagation mode was used before satellites for long distance VHF and UHF links beyond the line of sight.
Propagation models vary in how they evaluate obstructed paths. Some average the results, resulting in an easier to read coverage plot of many points. Others switch modes abruptly, resulting in a display of coverage that may be more accurate at a specific point, but more difficult to understand as signal level changes drastically over a small distance. When averaged over a large area, the accuracy of the two models may be the same, but differences at individual locations may be large.
Evaluating point-to-point coverage in rural and suburban areas is easy compared to mobile or indoor reception in urban environments. To calculate mobile reception in an urban environment, we also need to know the height of buildings around the receive antenna for accurate results. We also need to consider the potential for a signal reflected from a building that is not on the line between the transmitter and the receiver. Longley-Rice and other propagation models are unable to consider obstructions outside the straight-line path. Analyzing indoor reception requires all of this data and more—the location of windows, the direction they are facing, the height of the receive antenna above ground and the attenuation of the walls of the building.
Next month I’ll discuss some of the improvements to the Longley-Rice model Sid Shumate has been working on, as well as TIREM, a proprietary propagation model many consider superior to Longley-Rice, and the ITU model P1546-3, which you can download for a modest fee from the ITU store. If space permits, I’ll also look at some of the ray-tracing software cellular operators are using for street level propagation analysis.
Doug Lung is regional vice president of technology for a major station group.