For 3.0, A Big Stick May Not Be Enough
Since the beginning of broadcast time, TV broadcasters have relied solely on The Big Stick — a high-power transmitter feeding an antenna on the tallest available tower, building or mountain.
But as they begin exploring the potential of the new ATSC 3.0 standard and the myriad new services it enables, they are also experimenting with single frequency networks to extend the reach of the Big Stick and insure reliable delivery of the new signals to smartphones, tablets and in-car receivers, particularly when those devices are in motion.
The SFNs comprise multiple lower-power transmitters broadcasting on the same frequency with antennas on shorter strategically placed towers typically used by wireless carriers.
“A kilowatt on the corner is worth a lot more than a megawatt on the mountain, especially when you’re trying to hit mobile and gateways in the home,” says Fred Baumgartner, director of next-gen TV implementation for ONE Media LLC, a unit of Sinclair Broadcast Group.
“If you’re in Nebraska, a tall tower is great. But in downtown Baltimore, you want a lot of tiny little transmitters everywhere.”
Baumgartner served as director of broadcast engineering for Qualcomm from 2005 to 2010 and oversaw the creation of a 500-transmitter SFN network for MediaFLO USA, the revolutionary mobile TV service for smartphones that Qualcomm launched in early 2008 but shut down by early 2011 due to consumer indifference.
While MediaFLO was a commercial failure, the company was successful in providing mobile TV service over UHF spectrum using an SFN, which is what many broadcasters are hoping to do now with 3.0.
SFNs are already used widely for digital TV broadcasts in Europe and Asia but have been deployed in only a few circumstances for digital TV in the U.S. That’s partly due to a difference in modulation schemes.
The orthogonal frequency division multiplexing (OFDM) used by the European DVB-T standard is designed to work with multiple carriers and lends itself to SFNs, while the vestigial side-band (VSB) modulation used in ATSC 1.0 is designed as a single-carrier system.
As such, using it in an SFN is generally only feasible in markets where terrain shielding or other geographic considerations prevent unwanted on-channel interference; only 19 SFNs in the U.S. have been created for ATSC 1.0 broadcasts.
“You can do SFNs but it’s very, very difficult with VSB,” says Triveni Digital Chief Science Officer Rich Chernock. “It’s much easier with OFDM.”
Because 3.0 also uses OFDM modulation, it too can take full advantage of SFNs.
Sinclair, a big believer in using 3.0 to deliver mobile services, has already been testing SFNs in the Baltimore and Washington markets since 2015.
Earlier this year it teamed with Nexstar, Univision, Tribune and Northwest Broadcasting — members of a consortium called Spectrum Co. — to launch a 3.0 SFN deployment in Dallas.
The Dallas project involves pooling spectrum from several stations in the market to support both ATSC 1.0 and 3.0 operations, a model Spectrum Co. wants to replicate in markets around the country.
“The beauty of single frequency networks is that there is nothing that you can’t solve with more power, more transmitters, and of course, more money,” says Baumgartner.
“Want to up the data rate, want to get your coverage improved? Add a transmitter. The cool thing about OFDM is that it does not cause interference and cancellation. It loves multipath [interference]. The more transmitters you can put out there, the more directions the signal comes from, the better it gets. Anything else we’ve ever done, AM-FM, traditional TV, the opposite is true.”
SFNs could also be used to simply improve reception for traditional TV viewing in the living room. While many stations applied with the FCC last year to maximize their transmission power as part of the spectrum repack process, a few are now questioning the wisdom of making the investment to fully maximize given the reception improvements possible with SFNs.
While it might cost $1 million for a station to buy a higher-powered transmitter for maximization, says one station engineer, the same benefit could potentially be had for as little as $300,000 by creating a small SFN. That might make financial sense even if it was used only to deliver traditional TV programming to indoor antennas, not mobile TV or datacasting services.
Cleveland, Ohio-based Osborn Engineering, which has been consulting with large station groups like Sinclair, Tegna and Nexstar on their spectrum repack plans, has noticed this shift in thinking as stations look ahead to 3.0. operations.
“Stations are thinking about SFNs, and do we really have to maximize?” says Jeff Andrew, director of broadcast communications for Osborn. “Do we have to invest in higher-power transmitters?”
That said, Andrew expects that mobile and portable applications, not improved traditional TV reception, will wind up being the real driver for stations to invest in SFNs.
“If that wasn’t the case, people probably wouldn’t invest in just having a little bit better signal,” says Andrew, although he thinks SFNs might take the place of translators to improve coverage in problem areas within a market.
Dick Fiore, CEO of transmitter manufacturer Hitachi-Comark, agrees.
Hitachi-Comark supplied the UHF transmitters for the Spectrum Co. SNF deployment in Dallas, and the company is seeing a lot of interest in SFNs from other broadcasters. But it’s focused on the new business models possible with 3.0 such as mobile TV and datacasting.
“If you asked me to apply an SFN architecture to put out 4K pictures, I don’t see the benefit,” says Fiore. “There are no additional revenue streams.”
He also hasn’t seen any broadcasters looking to employ SFNs as a cheaper way to improve traditional TV coverage over maximization.
“I have not at this point seen any broadcaster choose to reduce their ERP [effective radiated power] and go with an SFN system,” Fiore says. “I think we will see people maximize, and then employ an SFN if that’s their business model.”
Every broadcast market is unique from an RF perspective, says Fiore, and he has heard projections of SFNs ranging in size from three sites all the way up to 20 different sites for a major market.
He thinks the average size will probably be five or six transmitter sites, counting a “big stick,” and he expects that the SFNs will be created in a partnership among broadcasters like the one in Dallas. That’s partly because of the financial burden, and partly because of the need to pool spectrum in order to maintain ATSC 1.0 services while creating a new platform for 3.0 broadcasts.
“You’re talking about aggregating a lot of frequencies that are going to go with you into these systems, in order to keep the lighthouse [station] on the air,” Fiore says.
One important consideration in designing an SNF is the amount of geography one needs to cover. OFDM uses a small delay known as a “guard interval” between carriers to avoid interference, and the longer the guard interval is, the farther apart the transmitters can be in an SFN. The tradeoff is that increasing the guard interval decreases the overall data capacity of the signal.
Sinclair’s testing has found that in Dallas and Baltimore you have to give up “surprisingly little capacity to get a reasonable guard interval,” says Baumgartner.
But he says the impact of the guard interval on data capacity would be a much bigger consideration in a geographically huge market like Lincoln, Neb., where SFN sites would likely be a large distance away from each other.
Gaining Real-World Experience
Sinclair and the other broadcasters in the Spectrum Co. consortium chose the Dallas market to test 3.0 broadcasts over an SFN for several reasons, Baumgartner says.
The market has flat terrain, which eliminates complications from terrain shielding, and there was good availability of UHF slots. Another consideration was that Spectrum Co. partner American Tower Corp. has facilities there. There are also a lot of high-tech companies in and around Dallas.
“As we start to stitch things together, this is going to be an interesting place to play,” Baumgartner says.
In Dallas, Spectrum Co. has placed 100-kilowatt SFN transmitters in three different sites to supplement two 1-megawatt 3.0 transmitters broadcasting off two tall towers. (One of the SFN transmitters will be operating at a slightly lower ERP to avoid local interference.)
The stations that will be turning off ATSC 1.0 to launch 3.0 include Univision’s KSTR, a UniMas affiliate, and Cunningham Broadcasting’s KTXD, an independent.
The 1.0 programming from those stations will then be channel-shared across several stations including Univision’s KUVN and Tribune’s KDAF (CW).
Dish Network is also participating in the trial and in April began broadcasting 3.0 on ch. 56 spectrum it acquired in the 700 MHz spectrum auction. The Spectrum Co. stations are still working through contractual agreements and regulatory requirements but should begin 3.0 broadcasts in September, with KSTR going first.
Technical considerations actually aren’t the hard part of setting up an SFN, says Baumgartner. Instead, it’s the business relationships between the stations that are pooling resources and spectrum.
“Once you peel the engineering part of the onion away, it’s about the contracts,” he says. “It’s about who moved where, and who will let what program operate at what bit rate, and everything else.”
The SFN sites in Dallas, which are owned by American Tower, all represent different tower sizes and different power supplies. The tower in Fort Worth is a 1,000 foot, 11-foot face guyed tower that was a former broadcast TV tower, while the Garland site is a small guyed tower with a 40-inch face that hosts a low-power FM radio station and a bunch of cellular gear.
When Sinclair put six-inch transmission line along the side, it “looked like the transmission line was holding the tower up,” Baumgartner quips. The tower in Denton is a 200-foot self-supported tower that AT&T originally put up for transmitting long-line data.
Spectrum Co. could have spent additional money to standardize power across the three sites, says Baumgartner, but chose not to as a proof-of-concept for how SFNs will eventually be rolled out. The general rule-of-thumb is to find workable homes for SFN antennas in existing sites which provide good coverage into attractive areas for new ATSC 3.0 services.
Each of the SFN sites in Dallas is built to handle up to six transmitters, meaning the system could eventually be scalable to 12 channels in two sets of six UHF channels.
Sinclair is putting only the first two transmitters in each site for now. It’s using Comark Parallax UHF GTP transmitters with Dielectric wideband UHF antennas, transmission line, combiners and filters. Sinclair owns Dielectric.
For studio-to-transmitter links (STLs) it is using one-gigabit fiber. The repack process means the SFN transmitters will have to operate on new channels, but Baumgartner says the transmitters and associated filters are all re-tunable.
Six transmitters operating at 5 kW TPO (transmitter power output) seems to be the “magic number” or “sweet spot” in running a 3.0 SFN, says Baumgartner, given power, equipment size and safety considerations.
The Comark Parallax transmitters were custom-built in order to fit two transmitters per cabinet (normal is only one per cabinet).
While Comark isn’t disclosing the exact price of the units, the company says that a typical 5 kW liquid-cooled transmitter is $225,000 each, including the RF system, cooling system, installation and proof-of-performance testing. That price doesn’t include the transmission line or antenna.
Picking Your Spots
Osborn Engineering is talking to broadcasters in San Francisco and Raleigh-Durham, N.C., about setting up potential SFNs.
Like Dallas, they are being viewed as collaborative efforts where certain stations would launch 3.0, and other stations would “host” their existing 1.0 programming, Andrew says.
Osborn had previously worked with wireless carriers AT&T, Verizon and Sprint in putting up more cellular towers to boost their coverage. Andrew says the push to launch 5G services may benefit broadcasters’ 3.0 efforts as they could probably locate SFN antennas on the same towers that will be used for 5G.
He says that Osborn is also exploring providing the capital to create SFN sites in some markets and then lease them to 3.0 broadcasters.
“They have to formulate their own business plans, but we’ll help with the infrastructure,” Andrew says.
Pinning down the costs of creating SFNs is difficult at this point. An FTI study commissioned several years ago by the Pearl station consortium pegged the number at about $450,000 to $600,000 per station.
Andrew says that an average cost to create an SFN would be about $1.5 million to $2 million per DMA, with costs obviously varying depending on existing tower leasing options. The average cost per SFN site would be in the $250,000 to $300,000 range.
Andrew notes that broadcasters could start small and then expand their SFN as more of the population adopts 3.0 receivers.
“It’s like the rollout of cellular technology, that came in stages,” Osborn says. “You don’t have to build out every SFN site immediately.”
Baumgartner says that SFNs are “all about shadows and angle of attack,” as penetrating into a building requires signals from all sides.
Fundamentally, he says there are three kinds of SFN sites: boomers, with the highest-power transmitters; boosters, with lower-power transmitters; and then distributed antenna systems, which would provide coverage in difficult areas like tunnels and stadiums. These could consist simply of laying transmission cable designed to “leak” signal into a subway tunnel.
When MediaFLO created its SFN in Salt Lake City, for example, it put its 25-kilowatt boomer on Farnsworth Peak, the main TV transmitter site in the market located southwest of downtown.
It then located 8 kW boosters, one in Ogden to the north, one in Provo to the south and one downtown. That provided good mobile TV coverage up and down the Interstate 15 corridor, Baumgartner says.
Everywhere you went you were getting the signal from at least two sides. For ATSC 3.0 broadcasters, their boomers would obviously be their one-megawatt “big stick” transmitters, which is a nice advantage.
“You’re starting out at a million watts at 1,000 to 2,000 feet,” notes Baumgartner.
The overall number of SFN sites one would need depends on not only geographic but demographic considerations. In designing its SFNs, MediaFLO didn’t start with traditional notions of broadcast market coverage.
“We started with where’s the population, and where’s the money,” Baumgartner says. “That’s the place you want to start concentrating your coverage into.”