What the Heck Is The “Duplex Gap” And Why Has It Blown Up The July FCC Meeting?

Difficult as it is to believe, there are times in policy when issues do not break down simply by partisan interest or into neat categories like incumbents v. competitors or broadcasters v. wireless carriers. Sometimes — and I know people are not gonna believe me on this – issues break down on pure substance and require lots of really hard choices. Of course, because these issues are highly technical and complicated, most people like to ignore them. But these kinds of issues are also usually the hardest and most intractable for people who actually care about what the world looks like and how this policy decisions will actually work in reality.


So it is with the question of whether to put broadcasters in the duplex gap as part of the repacking plan in the incentive auction. Did your eyes glaze over yet? Heck, for most people, it’s gonna take a paragraph or two of explanation just to understand what that sentence means. But even if you don’t know what it means, you can understand enough for this basic summary:


  1. Just about every stakeholder in the auction — wireless carriers, broadcasters, wireless microphone users, tech company supporters of using unlicensed spectrum in the broadcast bands, public interest groups — all told the FCC not to put broadcasters in the duplex gap.


  1. Nevertheless, the Auction Team proposed putting broadcasters in the duplex gap, based on a set of simulation that suggested that the FCC would only get back 50-60 MHz of spectrum to auction if they protected the duplex gap. The Chairman circulated a draft order adopting the Auction Team’s proposal.


  1. Everybody freaked out. The Chairman found he did not have 3 votes, or possibly not even 2 votes, to adopt his proposal on duplex gap. The freak out is so intense and so bad that the FCC actually waived the Sunshine Period for this itemso that interested parties can continue to talk to FCC staff and commissioners until the night before the meeting. The FCC also released additional data showing the impact would be limited to a relatively small number of cities.


  1. That helped some, but not enough. Despite progress on negotiations, the FCC clearly did not have time to get to the right solution in the 5 days between the release of the new data and the actual vote. Also, a bunch of people were pissed that the Auction Team hadn’t released the data sooner, and hadn’t provided more explanation of the underlying model and the assumptions behind it. On Tuesday, the Republican Chairs of the House Energy & Commerce Committee & the Telecom Subcommittee wrote Wheeler a letter chastising him for having a bad process and calling on Wheeler to pull the item from the agenda entirely. On Wed., the day before the vote, Wheeler wrote back defending the process but agreeing to pull the item (and the associated item on whether or not to change the spectrum reserve) until the August Meeting three weeks from now.


In Policyland, this passes for high drama. It is, to say the least, highly unusual. Enough so that even folks who find technical issues like this complicated and boring to the point of insanity are asking: “what the heck just happened there? Who lost and who won?” The equally complicated answer: “no one lost or won, we’ve got a serious debate about a technical problem which has consequences no matter how you resolve it” is not nearly as satisfying as “the carriers” or “the tech companies” or whatever.


I explain and unpack all of this below, as well as consider possible impacts and ways to resolve this. But again, I want to stress this is a super hard problem. This is about competing goals and the difficulty of predicting the future with any certainty. It’s also about trust and stuff, which is hard to come by in Washington even at the best of times. This is not subject to simplistic plotlines like “Oh, the Auction Team are out of control” or “The broadcasters and unlicensed supporters are just being stubborn.” (Wait, the NAB and the unlicensed guys and the wireless microphone guys are on the same side? And they agree with Verizon? WTF?) This stuff is hard.


More below . . .


Now Will You Tell Us What A “Duplex Gap” Is?


For those who have not been buried in this for the last however many years, the proposed 600 MHz licenses has 3 parts. The top part is the 5 MHz “uplink” that allows the mobile device to upload information to the cell tower (and from there back to the Internet). The bottom part is the 5 MHz “downlink” that allows the cell tower to download information from the Internet to the mobile device. In between, you have an 11 MHz “duplex gap” that is empty. This acts as a guard band between the uplink and the downlink to keep them from interfering with each other.

The duplex gap is at particular set of frequencies, with the uplink above it and the downlink below it. So for the entire country, there is one duplex gap at one set of frequencies, with its location depending on how much spectrum we pull back. You can see a nifty chart of the band plan in this article here.

Visually, imagine that the band plan looks like a multi-lane highway (Say Route 1) with a grassy separator between the northbound lanes (uplink) and the southbound lanes (downlink). The grassy duplex gap keeps traffic from the northbound left lane from running into the southbound left lane (and vice versa). Because the grassy duplex gap is pretty wide, the rules allow bicycles (unlicensed) and pedestrians (wireless mics) to each have a path on the grassy duplex gap. Put a broadcaster in the duplex gap and you not only loose the bike path and the walking path, you also loose the left lane on the southbound (downlink) side.


So How Would A Full Power Broadcaster End Up In The Grassy Duplex Gap?


After the auction is done, the FCC needs to find spaces for all the surviving broadcasters, a process called “repacking.” In the 2012 Spectrum Act that created the incentive auction (part of the Middle Class Tax Relief and Job Creation Act of 2012), Congress put a number of limitations on how the FCC could repack the surviving full power broadcasters. Full power broadcasters need to retain their existing viewing area, and can’t be moved down to the VHF channels 2-7 (unless they get paid to move down there) because it creates signal problems.

Additionally, this becomes a jigsaw puzzle because broadcasters have to have sufficient geographic space for their transmitters so that they don’t interfere with each other, and sufficient empty “white space” between each other so that they don’t interfere with each other. Depending on how many broadcasters actually participate in the auction and either cash out or get moved to VHF, the FCC may not have enough room to repack the broadcasters in the existing broadcast band. Or it may have enough room, but doing so would cause so much interference to the wireless block up for auction that they would no longer be for sale.


Is That What They Call Impairment?


Yes. Another complicating factor here is that broadcasters in the surviving market, or in the neighboring market, may create interference issues for wireless providers, or may require wireless providers to refrain from using their 600 MHz license in a particular geographic area. (We had a similar problem with the Lower 700 MHz A block licenses in the 700 MHz Auction back in ’08. That covered returned channel 52, and winners of A block licenses had to leave a big hole in their service territory to protect broadcasters still operating on Channel 51. As you can see from the auction page, the 700 MHz A block licenses went for a lot less than the unimpaired 700 MHz B block licenses.)

Impairment can come from needing to protect broadcasters in the market or in neighboring markets. Remember, broadcasting uses really huge transmitters to cover a large geographic area, and they need geographic distance from each other – there are only so many places to put them. Additionally, carriers need to protect broadcasters in Canada and Mexico from any cross-border interference. Canada and Mexico are not having an incentive auction, and while the FCC is in negotiations with the relevant agencies in both countries, we don’t have any kind of agreement with them yet. So it is all on us to make sure that (a) we don’t repack our TV stations in a way that would interfere with Canadian or Mexican broadcasters, which we are obligated by pre-existing treaty to do; and (b) we can’t allow wireless carriers to interfere with TV stations across either border.

So figuring out impairment for licenses is very tricky, and will shift constantly throughout the auction.


How Can Wireless Carriers Possibly Bid If They Can’t Know How Impaired the License Will Be?


An excellent question, and one carriers and the FCC have struggled with for some months now. The FCC has come up with a bunch of fixes to try to meet carrier concerns on getting stuck with damaged good after paying top dollar. The two biggies are:

  1. A) The FCC divides licenses into 2 categories, “Category 1” licenses that have 0-15% impairment, and “Category 2” licenses, with 15-50% impairment. Licenses with more than 50% impairment are not auctioned.
  2. B) The FCC has set a “national average” threshold for each clearing target. If the carriers bid enough money to clear enough broadcasters to release a block of spectrum (the clearing target), the block released must also meet a national average of no more than X percentage of the population subject to impairment (this measure of weighting spectrum by population covered, rather than by geographic area covered, is called “MHz/pop” and is the standard unit for evaluating the value of a license). If the total amount of spectrum cleared nationally has too much impaired spectrum to meet the goal, than the clearing target fails in the same way it would have failed had the carriers not bid enough money to clear the larger block.

The FCC initially proposed 20% impairment nationally, but after much pushback from carriers the FCC raised that to a requirement that no more than 14% of the spectrum (weighted by national population, not geography), can be impaired. For that to happen, the bulk of the spectrum released must not only be the better Category1 (0-15%), it needs to be relatively unimpaired Category 1.

Unfortunately, by making the number depend on a national average based on population, large cities have a disproportionate impact. A really bad problem in Los Angeles and San Diego can bring down the national average to a point where even if the rest of the country is in pretty good shape, we still don’t clear the larger block.


And This Relates To The Duplex Gap How?


The Auction Team has been fiddling with how to predict what rules have what impact on the auction outcomes for a long time now. As they have worked to resolve what rules proposals to use, their projections become more useful. Mind you, even with all the rules known, these projections still require a fair amount of guess work and assumptions. And these are only probabilities. They are not — and cannot — be certainties. But without making these projections, we can’t even test whether decisions potentially have a huge impact or a small impact on the overall auction results.

Needless to say, however, the decision to adopt the higher threshold to reassure carriers that the vast majority of licenses would be Category 1 licenses rather than Category 2 impaired licenses by requiring 14% national average impairment rather than 20% had a huge impact on the question of how to repack broadcasters and whether to put a broadcaster in the duplex gap or not.

So at the end of May, just before Memorial Day, the Auction Team released a bunch of projections on the outcome of the auction under two sets of assumptions: “protect the duplex gap” v. “not protect” the duplex gap. The Auction Team simulations show that if you prohibit putting a broadcaster in the duplex gap, it makes it very unlikely that the auction will clear 84 MHz or more — which means that we only get back 50-60 MHz to auction, which means that we only have 10-20 MHz in the reserve for competitive carriers instead of 30 MHz, and generally have less spectrum to auction. By contrast, if the FCC has flexibility to put a broadcaster in the duplex gap, then the auction will not have any problems clearing the national impairment threshold no matter what the return.


So What’s The Big Deal? Just Put A Broadcaster In The Duplex Gap If It Makes Such a Big Difference?


The problem is that everyone hates doing this, but for different reasons.


Back in May 2014, the FCC adopted a Framework Order that spelled out a set of goals: get back as much quality spectrum as possible to auction to enhance competition and meet future mobile broadband needs, preserve a robust free over-the-air broadcast service – including the ability to use wireless microphones for electronic newsgathering, and provide enough space for unlicensed operation in the guard bands and elsewhere to ensure that enough capacity for TV white spaces operation in every market to encourage investment and deployment in the service.


To provide enough capacity for unlicensed to encourage investment and deployment of next generation Wi-Fi, the FCC agreed with the advocates for TVWS that the FCC should provide a minimum of three, usable 6 MHz channels for white spaces operation in every market. To achieve that goal, the 2014 Framework Order proposed to take 3 steps:


  • Allow use of TVWS in the duplex gap, subject to protecting LTE and broadcast from interference;


  • Allow TVWS to share Channel 37 with wireless medical telemetry service devices (WMTS), subject to protecting WMTS from interference;


  • Ensure at least one vacant broadcast channel survives in the repacked broadcast band for unlicensed use.


In addition, the FCC eliminated the two channels (36 and 38) reserved for wireless microphone use. To ensure enough capacity for wireless microphones used by mobile news teams, the FCC proposed to allow the wireless microphones used for electronic news gathering to share the duplex gap with unlicensed. The duplex gap, as noted, is 11 MHz, which should provide 6 MHz for unlicensed and enough room for wireless mics and sufficient internal guardbands to protect LTE and broadcast operations.


Placing a broadcaster in the duplex gap eliminates the use of that duplex gap in that market for unlicensed and wireless microphones. As both TVWS users and wireless microphone users were counting on the availability of the duplex gap, this causes huge problems. In particular, since one of the markets likely to be impacted is Los Angeles, it means the second largest market in the U.S. this potentially impacts millions of people.



OK, That’s Why NAB and Unlicensed Advocates Hate This. Why Do Carriers Hate This?



Putting a broadcaster in the duplex gap in a way that minimizes overall impairment requires impairing the downlink. Because most mobile users (at this point) do much more downloading than uploading, impairing the downlink does a lot more damage to the value of the license than impairing the uplink does. Which is why the record shows that carriers initially said that as between impairing the uplink and impairing the duplex gap/downlink, carriers vastly preferred locating a full power broadcaster in the uplink.


Nevertheless, despite this initial broad resistance, it is important to note that some carriers (AT&T, Competitive Carrier Association) have reluctantly accepted the necessity of putting a broadcaster in the duplex gap in order to reach a national clearing target of 84 MHZ or more. Some carriers (Verizon), have continued to oppose any placement in the duplex gap. Even those wireless carriers now reluctantly supporting a broadcaster in the duplex gap (such as AT&T) have recommended limitations on the total number of markets where the Commission should do this.


Part of the reason for this is that there is a theoretical fix to the carriers’ concerns about impairment to the downlink, but it requires a major advance in filtering technology. (“Filters” in wireless means screening out unwanted interference and preventing your device from leaking interference to others.) The more optimistic you are about your ability to engineer your way out of this at reasonable cost, and how long you’re willing to wait before deployment, the more likely you are to accept a broadcaster in the duplex gap. Additionally, CCA and its members need the auction to clear 84 MHz to get the 30 MHz they want in the spectrum reserve. They will take a hit in a few markets to get to 84 MHz and a fully stocked reserve.



How Much Damage Does Locating A Broadcaster In The Duplex Gap Actually Do?



That’s an excellent question, and very hard to determine. In part, it depends on how many markets get a broadcaster in the duplex gap. When the Auction Task Force released its first batch of simulation results on May 20, it wasn’t clear how many major markets might get impacted. Take out enough major markets and major chip manufacturers won’t bother to develop mobile chips or put them in every device. Consider the difference between Wi-Fi and the 3.65 GHz “licensed lite” band, which has large exclusion zones that cut out the coasts and the Great Lake region.


Wi-Fi works everywhere, so just about everything these days has a Wi-Fi chip in it. That makes the price of Wi-Fi chips dirt cheap, because of economies of scale. By contrast, only wireless ISPs and others interested in broad area networks in rural areas use the 3.65 GHz band. It does good stuff, but no one makes equipment for it except a handful of manufacturers making small runs of high-power stationary transmitters. No one bothered to make low-power chips for portable devices because the exclusion zones cut out about 50% of the market for consumer devices. Equipment remains expensive because there are no economies of scale.


If broadcasters in the duplex gap hit a large number of markets, it would raise the serious danger that TVWS would go the way of 3.65 GHz, useful for some purposes but nowhere near what you would need to create next generation Wi-Fi services. Similarly, impacting a large number of markets would make significant live news coverage effectively impossible. Consider that when the Supreme Court made its big decisions on Obamacare and same sex marriage last month, you had dozens of live news crews from all the major cable and broadcast networks doing live coverage in front of the Supreme Court. Without sufficient capacity for wireless microphones, it is hard to see how you could do the same kind of coverage by lots of news crews.


Fortunately, the subsequent release by the Auction Team showed that for most of the major clearing targets, we are only talking about 6 or 7 markets impacted. Of these, the most significant is Los Angeles, which for various technical reasons I won’t get into is a total mess. Most of the impacted markets are small markets like Wilkes-Barre Scranton along the Canadian border. While losing a major market like LA or Cleveland would certainly hurt, it isn’t nearly as catastrophic as losing Southern California and the Northeast Corridor – as a lot of people initially thought would happen.



Wait, You Mean All This Angsty Tsuris Comes From a Total of SIX Television stations, Most of Them In Small Markets? Can’t You All Just Find Some Way To Deal With The Problem?



Another excellent question! Since it became clear the FCC was going to propose putting broadcasters in the duplex gap, we’ve seen a huge interest in trying to find solutions. But because this is a complicated, hard problem, every proposal has some pluses and drawback. And, because we are trying to predict the future, we can’t really know how well any one of these will actually work in reality. To give a decent sampling (briefly):


Buy out the broadcaster that would go in the duplex gap. The Book of Ecclesiastes tells us “money answers all” (Ecl. 10:19). The proposed auction rules contain a mechanism called the “extended final round.” If the auction closes (i.e., the carriers have stopped bidding) and the carriers are close to a clearing point, the FCC will tell the carriers “hey, if you collectively bid this much more, you will hit the next clearing target” and will go one more round to see if the broadcasters are willing to bid that much more to release more spectrum.


The FCC could apply a similar mechanism here. If the carriers reach a monetary clearing target, but they cannot clear that amount because of the national impairment limit, the FCC could say “hey carriers, if you bid this much more, you can clear out the stations that are blocking the impairment and release more spectrum.”


Advantages: broadcasters are always happy to get more money. Carriers do not want a broadcaster in the uplink either. They can decide for themselves if they would rather have more spectrum with an impaired downlink or less spectrum that is relatively unimpaired.


Disadvantages: Carriers might decide they can live with the impaired downlink, which would still create huge problems for unlicensed spectrum and wireless microphones. Also, since everyone knows which markets are likely to be impacted, it invites gaming by broadcasters who want to leverage their “hold out value.” A station in Flint or Scranton is likely to want a lot more to get bought out if it knows that the success of the entire auction depends on buying out enough stations in these small markets to save the whole auction.


Find other space for unlicensed and wireless microphones. Keeping the duplex gap clear is a means to an end for unlicensed users and wireless microphones, not an end in itself. The whole point was to make sure we had enough channel capacity for unlicensed use and wireless microphones. If we can find that space another way, then most of the issue of putting a broadcaster in the duplex gap goes away.


There are two possible sources for “alternative equivalent spectrum” for unlicensed and wireless microphones. The first is expanding the “vacant channel.” In the 2014 Order, the FCC stated it would preserve one open UHF channel in every market for unlicensed us and wireless microphone use. Under the current proposal, any low-power TV (LPTV), TV translator, or broadcast auxiliary service (BAS) applying to relocate out of the repurposed 600 MHz band would need to show that by doing so it does not take up the last vacant channel in a market usable by unlicensed or wireless microphones. In theory, the FCC could expand that to hold open 2 vacant channels rather than just 1 vacant channel.


The other source of new equivalent spectrum is blocks of spectrum in the market so impaired (more than 50% impairment) that the Commission will not sell them at auction. The Commission’s simulations show that even with putting a broadcaster in the duplex gap, the FCC will still have a bunch of licenses too impaired to auction in the impacted markets, like LA. But factors that create impairment for full power LTE broadcasters do not create problems for the much lower power unlicensed devices and wireless microphones. Those blocks may operate just fine in impaired blocks.


Advantages: Allows wireless carriers to get cleaner spectrum while avoiding severe impact on next generation Wi-Fi and electronic news gathering. May even create more capacity for next gen Wi-Fi and wireless mics.


Disadvantages: If it were easy to find space outside the duplex gap, we would have done it in 2014. Finding a vacant broadcast channel puts the secondary services (LPTV, TV translators, BAS) at risk. Preserving 2 vacant channels makes it that much harder to find homes for relocating secondary broadcast services.


And while blocks that are too impaired for LTE may provide enough space, they may not. TVWS requires a 6 channel block. An LTE channel is 5 MHz. You would need two adjacent LTE channels to get a channel usable for unlicensed. Even if you can find two impaired blocks adjacent to each other, there is no guarantee that unlicensed will be able to operate in the impaired LTE channels. It depends on what kind of impairment it is, and whether the impairment significantly impacts unlicensed (or wireless mics) or not.


Finally, that still leaves carriers with an impaired downlink from putting a broadcaster in the duplex gap.


Change the rules on impairment or the modeling assumptions to make the problem go away. The national impairment threshold is obviously not a law of physics. In fact, the Commission originally proposed a 20% national limit. The Commission could go back to that proposal, or could avoid having a national threshold altogether, or could say the national threshold doesn’t apply if it only impacts X number of markets so that we can hit the national clearing target but just have fewer Category 1 licenses in the 6-7 markets where it matters.


Additionally, the current model considers blocks that can’t be sold because of impairment as “100% impaired.” But there is a reasonable argument that these licenses shouldn’t count toward impairment at all. After all, the purpose of the national limit is to ensure a high ratio of Category 1 licenses to Category 2 licenses. Blocks not auctioned don’t change the ratio at all. So why not exclude them from the national impairment limit?


Advantages: Problem solved! No more simulations projecting we will fail to meet the clearing target, so no more need to put a broadcaster in the duplex gap.


Disadvantages: Reality probably doesn’t work this way. Yes, we still don’t have a lot of information on the assumptions governing the simulation software and how they were operationalized, so we can argue that the Auction Team is being too conservative or should have taken into account this or that or modeled possible other rule changes. But remember that the Auction Team did not lower the proposed national threshold on a whim. It did it in response to a lot of carrier concern that this would produce too low a ratio of Category 1 licenses to Category 2 licenses. Going back to the old national threshold means going back to solving that problem. Now maybe carriers would prefer a lot more Category 2 licenses to impairment of their downlink. But since none of the carriers suggested this as a viable approach to solve the problem, I kinda doubt it.


Also, whatever assumptions we use, we still have to put that leftover broadcaster somewhere. Putting it in the uplink channel may create more impairment, or may impair more blocks. The physics of lowband spectrum is funny that way. The very thing that makes it valuable – much better propagation characteristics and penetration capability – also makes it harder to control interference with other users in the band.


Finally, even if we argue the Auction Team assumptions are too conservative, they at a minimum reflect a worst case scenario we need to know how to address if it does come up when we hold the actual auction. A rule that absolutely prohibits putting a broadcaster in the duplex gap anywhere under any circumstance may create the problem the Auction Team simulations exist, but with no way to address it.


I’m not saying we shouldn’t get more information on the simulations, or even a chance to play with the software. We might find there is some combination of rule tweaks that can address the problem, or minimize it further. We might also find some big problem with the assumptions or the software. After all, no one is perfect, and even really brilliant people can make mistakes when working a complex multivariable problem that they have to translate into a mathematical formula that will allow a computer to make projections. But unless we postulate a rogue Auction Team conspiring to screw over every single stakeholder out of revenge for three years of being pounded by competing teams of industry lobbyists, the odds that the simulations are so wrong we can just ignore the danger of not reaching the national impairment clearing target are pretty small.


So we can’t just wish the problem away. The map is not the territory, and the simulations are not the reality. The simulations help us prepare for reality. And when they flag a potential problem that could have major impact, even if the chance is more remote than the initial projections show, we need to take that seriously or risk the consequences.



So What Happens Now?


Happily, the Chairman decided to withdraw the item from the July meeting and shift it to the August meeting. Given progress made on finding an acceptable approach in the 5 days since the Auction Team released the additional simulation data, this should give stakeholders, staff and the FCC Commissioners who actually vote this stuff time to decide which of these approaches works best and can generate the most support in the stakeholder community.


Pulling the item until August was, of course, also good politics. Between the July meeting and the August meeting, Wheeler and Pai will be back in front of the House Energy and Commerce Committee for another hearing on FCC process. By complying with Upton’s and Walden’s demands to pull the item from the July meeting, Wheeler has done a lot to defang the issue. I’m not saying it won’t come up, but Wheeler (and the Democrats on the Committee) can point out that they did what Upton and Walden wanted and that the process is working to find a solution.


Additionally, because the Chairman also pulled the spectrum holdings item, we get another 3 weeks to push for expanding the reserve from 30 MHz to 40 MHz, and to adjust the trigger for the reserve so that it goes live at the right time.


On the downside for me, it means another 3 weeks of working on this stuff, less the time I will be at Pennsic 44 playing with Baby and Junior.


Stay tuned . . .

Comments are closed.