The IEEE Broadcast Symposium has been the place to learn about new developments in broadcast technology for almost 60 years. Many of the papers at the 59th IEEE Broadcast Symposium covered digital TV and radio broadcasting. I'll present brief highlights from some of the papers this month. For a copy of the CD containing the text of many of the papers, contact the IEEE Broadcast Technology Society.
MOBILE MEASUREMENT TECHNIQUES
John Kean from NPR Labs discussed techniques for field measurement of FM and IBOC (HD Radio) signal coverage during his tutorial. At least two of the techniques should be useful for mobile DTV measurements as well.
Kean constructed a circular ground plane with a whip antenna in the middle that was mounted on the roof of an automobile. As his charts show, this dramatically reduced the pattern distortion caused by the vehicle and thus improved measurement accuracy.
One of the problems with mobile measurements is that local multipath creates a lot of "noise" in signal level measurements. Kean reviewed techniques for removing multipath fading without removing log normal fading. Mathematical solutions are too complex and time consuming. His hardware solution—applying a 2 hertz fourth-order active low-pass filter to the detected output from the receiver—worked fine and was much simpler.
SIGNAL LEVELS FOR RELIABLE DTV RECEPTION
The signal level required for reliable DTV reception has been a popular topic for IEEE Broadcast Symposium papers over the last 20 years. Pascal Marcoux's paper "Revisiting the Field Strength Requirements for DTV in the Canadian Context" uses data from previous papers by Oded Bendov, Yiyan Wu, Guy Bouchard and others (see the paper for all references), and measurements on current consumer antennas and receivers to calculate recommended "Grade A" and "Grade B" signal levels for UHF and high-VHF.
The finding that consumer antennas have gain 2.9 dB below specification (on average) should come as no surprise. "Low noise amplifiers (LNAs)" may not help as measurements showed only 2 out of 11 worked as specified.
The "Grade A" DTV reception level is used for indoor reception and compensates for the use of indoor antennas, offset somewhat by shorter cable, about 6 dB of loss due to height reduction and about 8 dB of building penetration loss. Measurements showed additional margin was needed.
Additional signal strength is needed to cover more man-made noise than expected, noise added when the equalizer needs to correct for multipath, and mismatch loss and degraded noise figure due to poor antenna and receiver VSWR.
Table 1
Table 1 shows the results presented in Marcoux's paper. Note that for "Grade A" indoor reception, more field strength is required at VHF than at UHF.
LONGLEY-RICE INACCURACIES
I've discussed Sid Shumate's improvements to the Irregular Terrain Model (ITM) Longley-Rice propagation software used by the FCC for all coverage and interference calculations in previous columns. Most of those columns focused on simple "housekeeping" changes to the current ITM implementation designed to correct errors in the program's subroutines and calculations. In his IEEE tutorial Shumate gave a brief description of these and other corrections to the ITM C++ code.
The major improvement in accuracy comes from changing the source code used for calculating line-of-sight (LoS) losses beyond free space loss and two-ray multipath. Existing ITM code, which uses diffraction equations designed for use beyond the radio horizon, doesn't provide accurate results.
Shumate quotes Dr. Harry Anderson's 2003 book, which notes "At distances from the transmitter to the horizon, the path loss actually includes a weighted portion of the diffraction loss beyond the horizon. Having the path loss to a receiver determined by terrain obstacles beyond that receiver location is clearly non-causal and violates physical reasoning for a single path two-dimensional model."
Shumate's Irregular Terrain with Obstructions Model (ITWOM) includes the use of clutter, rather than diffraction, to calculate losses in the LoS range and on the far side of obstacles. ITWOM does away with the averaging present in ITM, except for Troposcatter paths. LoS and diffraction losses are determined on a point-to-pixel basis, eliminating the "smearing" present in ITM.
The tutorial highlighted the difference by comparing ITM and ITWOM coverage maps for WETA. ITM significantly over-predicts signal strength and misses many weak signal plots. I also noticed that the impact of the antenna elevation pattern is much clearer in ITWOM than ITM. Shumate handed out CDs at the symposium containing the paper, along with additional data, copies of his other papers and an Excel spreadsheet to test the code on a point-by-point basis. Contact him to see if copies are still available.
MECHANICAL BEAM TILT
During the session "Broadcast Antenna Technology," Merrill Weiss presented a paper showing why the method currently used for calculating the impact of mechanical beam tilt on antenna azimuth and elevation patterns is inaccurate.
When mechanical tilt is used, the simple approach is to determine the amount of mechanical tilt at a given azimuth by extrapolation between the azimuth of maximum tilt and the azimuth of no tilt (90 degrees each side of the maximum). This tilt is then added or subtracted from the calculated depression angle to determine the angle from the antenna if it were not tilted.
For example, consider an antenna mechanically tilted 2.0 degrees down at zero degrees true. At an azimuth of 45 degrees, using simple extrapolation the mechanical tilt will be 1.0 degree. If the depression angle to a receiver on the 45 degree radial is 3.5 degrees, the relative field for the elevation pattern at 2.5 degrees (down) will be combined with the azimuth relative field at 45 degrees to determine the effective radiated power.
This simple approach may be accurate enough for most studies, but not in the design of distributed transmission systems where antenna elevation pattern nulls are needed to limit contours or prevent interference.
Weiss used a basketball as a prop to show it is necessary to use spherical geometry to calculate an accurate pattern. In the simple analysis, the assumption is the antenna mechanical tilt, regardless of the amount, is always pointed down directly towards the receiver. In the example above, that's not the case—an observer looking at the antenna from the receive point at 45 degrees would see it tilted slightly to the right.
By mapping the 3D antenna pattern without tilt on a sphere, then rotating the sphere to add the tilt and extrapolating and remapping the points to create a new 3D pattern, the relative field will be correct at all points. Although Weiss' paper is not part of the IEEE Broadcast Symposium Proceedings, now that you understand the concept you should be able to work through the somewhat laborious math involved.
DTV PERFORMANCE
A manufacturers panel discussion on TV antennas was also held during the "Broadcast Antenna Technology" session. Prior to the discussion, the IEEE BTS honored Jules Cohen for his contributions to broadcasting, including work in ACATS and the creation of the ATSC standard. In accepting the award, he described the work on the U.S. DTV system and reception problems, particularly at VHF.
After his talk, members of the audience offered comments on DTV reception. With regards to VHF in particular, it was stated that the FCC's interference rules are hurting stations.
To protect distant cells, which in reality may have no reception (see the previous discussion on required signal strength and Longley-Rice errors), stations have to limit their signal and reception in the primary signal area 20 to 30 miles from the transmitter.
One audience member recommended waiving the half-percent interference limit, picking another number, and considering all the different reception environments—indoor, outdoor, big antenna, small antenna, etc. Power increases could be done under experimental licenses to see if they caused any real interference problems.
Kerry Cozad from Dielectric Communications made the point that, instead of looking at "coverage," we need to focus on "reception." Compare predicted versus measured field strength. If the signal level is low, modifying the antenna elevation pattern (broader pattern or smoother pattern) can provide gains up to 10 dB a few miles from the antenna.
It won't, however, make much difference over 5 miles (VHF) or 10 miles (UHF). Side-mount or directional antennas could be an issue for DTV. Now that TV broadcasting has transitioned to digital, signal saturation may become much more important than coverage contours.
This is only a small sample of the information presented during the three-day 2009 IEEE Broadcast Symposium. There were several other interesting papers and tutorials that I do not have space to mention here, but will try to fit into future RF Technology columns over the next few months.
Comments are welcome! E-mail me atdlung@transmitter.com.
