There has certainly been more than enough debate about the efficacy of the 8VSB modulation system for digital television. Early reports of difficult reception and problems with multipath have led some parties to believe that the United States has made a serious mistake in adopting VSB as opposed to the OFDM standard being adopted by Europe and Asia.
The truth is, 8VSB signals are not as difficult to receive as many would maintain. I know; I have done extensive testing over the past year-and-a-half with a variety of antennas in several different indoor and outdoor locations. As a result, I can report that it is indeed possible to receive 8VSB signals — at least those transmitted on UHF frequencies — in a variety of indoor and outdoor locations by paying careful attention to the type of antenna used and overall signal levels.
While 8VSB doesn't work everywhere, it's still early in the game; and further developments in 8VSB receivers — and enhanced 8VSB data streams for “training” those receivers — are showing promise for solving even the most difficult reception problems. Aside from these developments, there are plenty of things that can be done to improve the antennas used to watch DTV.
Let the games begin
My initial tests with antennas began as I was trying to watch Monday Night Football in HD in October of 1999. Oddly enough, the primary reason I began experimenting with a few Radio Shack antennas was to try to get a more reliable signal to test out a 50" plasma panel and a 32" 16:9 direct-view TV.
By January 2000, I had installed a semi-permanent antenna mount on the rear of my deck to facilitate swapping out various antennas I had accumulated from Radio Shack, Terk and Channel Master.
Along the way, I had discovered a couple of interesting things. First, the best antennas I tested weren't always the biggest. Second, I found that optimum reception of two stations with antennas mounted on the same tower — KYW-26 and WPVI-64 — often came from different directions.
The best antenna for receiving WPVI-64 turned out to be a plain little screen design with four collinear UHF bow ties mounted on it. This antenna helped my Panasonic TU-DST50W first-generation set-top box lock up the three strongest Philadelphia UHF DTV stations and even let me see snippets of a fourth, WCAU-67 (running low power). I decided to start conducting more scientific tests of antenna performance. News that additional DTV stations from Philadelphia (WHYY-55) and Allentown 25 miles northwest (WFMZ-46 and WLVT-62) would soon be coming on-air provided further incentive.
As the weather warmed up, I mounted five feet of mast and a rotor next to the chimney on my house and installed the previously mentioned Channel Master model 3021 collinear with an accessory UHF preamp. Eighty feet of RG-6 routed from my basement to connect to the Panasonic tuner completed the installation.
With the help of Dave Smith, an engineer at KYW-TV, I was able to borrow a spectrum analyzer and began to shoot some waveforms from the local stations. By April 2000, I was able to watch all the local DTV stations that were available for viewing, as well as pick up an 8VSB carrier or two from New York, 65 miles to the northeast.
I secured review models of numerous TV antennas. I positioned each antenna on my rear-deck test mast and then atop the roof mast, measuring signal strength on each received DTV signal with a Sadelco Mini-Max 800 meter.
I also purchased a precision 75Ω step attenuator from JFW Industries and began conducting what I called “headroom” tests for every DTV set-top box I could get my hands on. This test involved peaking the antenna for best signal, then adding attenuation in either 1dB or 10dB steps until the signal started to break up. This reverse-engineering process, combined with spectrum analyzer measurements, gave me a rough idea of the carrier-to-noise performance of each STB, as well as which antennas worked best for 8VSB reception.
Like most first-time DTV viewers, I went for the biggest antenna on the tallest mast for my first DTV antenna setup.
Too much fun to be work
As spring turned into summer, I had identified a few antennas that seemed to work well for UHF DTV reception. I also decided to try to build a compact UHF antenna that would be optimized for gain around channels 40-45. (We have a lot of local DTV channels above 40 in the Philadelphia area.)
Eric Hodges at Tektronix provided a model 3661 spectrum analyzer to enable me to perform real-world comparisons between my designs and the ever-growing piles of aluminum in my yard.
It is well known in antenna design circles that a full-wave dipole is a moderately broadband, antenna-driven element that is relatively easy to match. In fact, the folded bow tie dipole configurations used by many of today's commercial antennas are simply modifications of the folded dipole design.
In short order, I had a working five-element yagi using a folded dipole that was doing an excellent job of picking up all the Philadelphia DTV stations while mounted on 10 feet of mast on my rear deck. The 8VSB waveforms I was seeing on the 3661 showed low to moderate multipath on the received signals, although the C/N levels weren't really high.
Further modification to the antennas resulted in flatter-looking waveforms and improved C/N levels. I got to work cutting and gluing, and soon had a three-element version of the same antenna, which tested nearly as well as the five-element version.
Coupled to either antenna, my “deaf” Panasonic TU-DST50W set-top box proved quite sensitive. The received signals had only about a 20dB C/N ratio in some cases, but reception was rock steady.
Adding a high-quality UHF preamp (Channel Master model 7775) and deliberately inducing multipath taught me something else. The Panasonic tuner could also handle a fair amount of signal multipath, provided the overall C/N ratio remained high. If multipath produced notches in the received 8VSB waveform that dipped below the STB's minimum discernable signal level then there wasn't any way I was going to get reception.
By now, I had started to cannibalize the pile of antennas from my previous tests to make up improved versions for UHF DTV reception. The result was a modification to Channel Master's model 3022 suburban yagi that improved gain and multipath performance on the upper UHF channels. The 3021 screen antenna didn't need further tweaking — it was still the best performer on hand.
During September and October of 2000, I tested my antennas indoors in my family room and basement. Results were encouraging — the strongest stations could be pulled in with the 12" long three-element yagi, provided the C/N was strong enough. (The CM7775 UHF preamp helped in many locations.)
I also took a pair of antennas to the famous “Schubin Site,” an apartment located in the upper west side of Manhattan that is the residence of industry pundit Mark Schubin. His reception problems are legendary, and a borrowed Hewlett-Packard spectrum analyzer showed lots of multipath and low signal levels.
After some testing, I placed one of my antennas on the floor for a moment to pick up my camera. Instantly, the local Fox channel, WNYW-44, locked up on his RCA DTC-100 set-top box. The antenna was oriented northwest, with Fox's transmitter atop the Empire State Building about 40 blocks south-southeast.
Further testing was bringing in WCBS-56 intermittently, but in another direction. By the time I had to depart, I had hard evidence that an 8VSB signal with considerable multipath could be received indoors — again, provided that the C/N ratio was high enough.
Subsequent tests in my basement and office revealed I could get reliable reception from KYW-26, WTXF-42, WHYY-55, WPVI-64 and WLVT-62 with not much more than the three-element yagi and a UHF preamp. These results were duplicated with the Antiference Silver Sensor, a tabletop log-periodic UHF antenna that is quite compact.
A trip to ABC Television's Advanced TV Labs in New York resulted in successful reception of WNYW-44 and WCBS-56 simply by placing the 12" three-element yagi on a filing cabinet near a rear window and aiming it at a building across the street — no preamp was needed. By anyone's reckoning there were numerous signal paths to this location. Once again, though, the high C/N was sufficient to lock up a Sencore AT984 receiver, as well as a Panasonic TU-DST50W STB.
By now, I had a few of my antennas fabricated out of something more substantial than PVC pipe. I took two prototypes down to the DTV 2000 conference in Atlanta. Most of what transpired there was detailed in the January 2000 issue of Broadcast Engineering. Brad Dick, editor of BE, made a bet with me that he'd buy me dinner if I could receive an Atlanta DTV station indoors at the hotel.
The site was a tough one — the Hyatt-Regency, which sits on a steep incline. The mini-trade show took place in its basement exhibit hall, and the ceiling of that room is below ground level. That's not a particularly great place to try indoor reception, but with the assistance of the kind folks at Sencore, I set up my five-element yagi and several feet of coax to another AT984 receiver.
I was as surprised as anyone a few minutes later when local Fox station WAGA-27 appeared. Not only that, the signal was fairly steady — it only dropped out when we walked in front of the antenna. No preamp was used, and the AT984's readout showed we had a fluctuating signal level from 18dB C/N to as high as 26dB C/N.
Neither of the other two Atlanta DTV stations was strong enough to receive indoors, as both measured below 18dB C/N. Of course, the Fox signal had lots of multipath, and reception wasn't easy. In fact, we measured one notch as deep as 12dB on the accompanying Sencore 8VSB waveform analyzer. Still, the AT984 held firm and we watched several Fox programs before pulling the plug. (Brad, I'm still waiting for dinner!)
Antennas and fitness
My reception tests continued into this year, as next-generation DTV set-top boxes started coming to market. My first experience with these was with the terrestrial-only Samsung SIR-T150, with a complete Broadcom front end and 8VSB demodulator. In a side-by-side test with my old Panasonic tuners and an RCA DTC-100, the Samsung box exhibited at least 6dB more sensitivity across the UHF spectrum.
In fact, with this tuner and either of my two Channel Master UHF antennas, I can watch 14 different UHF DTV stations at home over partially obstructed paths as long as 70 miles. Subsequent tests of set-top boxes from Panasonic (TU-HDS20), Princeton (HDT2000) and Sony (SAT-HD100) have confirmed that the new receivers are indeed more sensitive than their ancestors.
A trip to the Los Angeles area in late March gave me another chance to test out my antenna designs. Several high-profile clients of a home theater firm were having little or no success with reception of UHF DTV signals in Beverly Hills, Malibu and Bel Air. The causes were varied, but in Bel Air the problem turned out to be an inefficient antenna and too much feedline.
The Beverly Hills location was a real puzzler — the house was located almost 300 feet below and 1/4 mile away from a nearby canyon rim and about 20 miles from the Mt. Wilson transmitter site. Previous attempts by the installer had resulted in intermittent or no reception of Los Angeles DTV signals.
Armed with a spectrum analyzer, preamp, some of my own antennas, a modified Channel Master 3022 and the Samsung receiver, I quickly determined that signal levels were abysmally low with lots of multipath.
Still, the modified Channel Master 3022 coupled to a CM7775 UHF preamp turned the trick. I situated it on a rise about 60 feet higher than the house, using 10 feet of mast and a tripod support. All eight of the available DTV stations were subsequently received reliably with no dropouts by aiming the antenna directly at Mt. Wilson.
Even though the feedline consisted of 250 feet of RG-6, the Samsung SIR-T150 worked quite well with the signals it was being fed. (Note that the antenna was still over 200 feet below the rim of the canyon.) Similar success was had in Malibu and Bel Air with driveway reception and 10 feet of mast. In some cases, the antenna had to be rotated slightly to improve some of the channels, but the beam angle never varied by more than 30 degrees for all stations.
Subsequent to these tests, I had the pleasure of hearing several reports by John Tollefson of PBS and James Kutzner of MSTV on outdoor and indoor reception of DTV signals at NAB2001. All of the reports confirmed (in a much more scientific way) what I had learned: 8VSB works much better if there is sufficient C/N ratio present at the input of the receiver.
That's not to say you can receive 8VSB “anywhere, anytime.” The system still needs work. Enhanced 8VSB tests are now under way to enable receivers to better learn and remember a particular signal's multipath characteristics and how to compensate for them. Research into the infamous short delay 0dB echoes continues at several of the 8VSB chip designers.
In fact, an old bit of wisdom still holds true: The easiest, most cost-effective way to add extra gain to your station is to make improvements to your antennas.
I'd like to acknowledge several people who have provided invaluable assistance and insight during my 8VSB reception tests: Mike Strein and the engineering staff at ABC's Advanced TV Labs; Tom Duff and the NBC Engineering staff; Ed Williams at PBS Engineering; Bob Seidel and Bob Ross at CBS Engineering; Bill Weber and John Doran at WHYY-TV; David Smith at KYW-TV; Mark Schubin; Duffy Paul and Wayne Massengill at Channel Master; and Gil Akroyd, formerly of WLVT-TV.
If you'd like to read more about my antenna work, go to my wffkkebsite at www.projectorexpert.com where you'll find several articles detailing antenna tests and news about DTV set-top receivers.
Peter H. Putnam is president of ROAM Consulting.