By Jeffrey Krauss,
President of Telecommunications
and Technology Policy
The FCC has begun a proceeding on a new, wildly impractical concept for spectrum management it calls "Interference Temperature." It plans to set up an experiment to test the concept. The guinea pig will be the CARS band, 12.7 to 13.2 GHz, still widely used by cable operators. But the framework for the experiment has nothing to do with the new concept.

The 12.7 to 13.2 GHz band was allocated for point-to-point microwave use by cable operators to distribute programming from a headend to multiple hubs in a metropolitan area. Some cable MSOs have replaced their CARS microwave with fiber, but others continue to use it. Other users of this band include a few satellite earth station uplinks, and low-power, short distance portable microwave links used by broadcasters for electronic news gathering.

The Interference Temperature concept works like this: If you have a complete database of radio spectrum usage by licensed transmitters at every point in space and time, and you know the receiver sensitivity of every nearby receiver, and you know the level of man-made and atmospheric noise, then you can (in theory) determine the amount of "spare" spectrum capacity and allocate it to temporary low-power transmitters to share that spectrum.

It's complicated. The number and location of the allowable temporary users depends on a calculation to determine how much additional interference the existing receivers can tolerate. So long as the new interference, coupled with the existing noise level, stays below a certain "temperature," the existing receivers can continue to operate without a problem. When that "temperature" is exceeded, the existing radio links start to fail.

Making this concept work depends on an extensive database of receiver locations and receiver characteristics, and a network of sensors to measure noise and channel occupancy. You can conceive of how this might work. And you quickly realize that it's not practical.

But the FCC isn't merely asking for comments on the feasibility of the concept. It is also proposing an experiment to try it out in practice. The problem is that the experiment has nothing to do with a database of receiver locations and characteristics; nothing to do with existing noise levels. It's simply a raid on the CARS band.

The FCC proposes to allow low-power digital transmitters, like those used for WiFi local area networks, to share the CARS band and another point-to-point microwave band around 6 GHz. But how will they avoid causing interference to existing microwave links? Make them use Dynamic Frequency Selection (DFS) to listen before they transmit. DFS isn't a brand new concept. For the past few years, WiFi equipment vendors negotiated with the U.S. Defense Department over the issue of extending the IEEE 802.11a band (5725 to 5875 MHz) down to 5470 MHz. But portions of the extended band are used by DoD radars, including sensitive spaceborne National Defense radars. The compromise is that the new 802.11a devices, when they eventually hit the market, will employ DFS. They will constantly listen for signals from satellite radars, and will quickly shift transmissions to other frequencies when they detect a radar signal.

If DFS works for 5 GHz, will it work for 13 GHz? Well, maybe. Remember, it's the CARS receiver that needs interference protection. But the new low-power device would listen for transmissions from the CARS transmitter. It's easily possible to have the CARS transmitter-to-low-power-device-path blocked by a building, but no blockage on the low-power-device-to-CARS-receiver-path. The FCC assumes that CARS receiver antennas will be mounted on towers and rooftops, while low-power device transmissions will be at ground level, and the highly directional antenna on the CARS receiver will help to block out interference.

Well, maybe. But the FCC doesn't have a clue about the characteristics of CARS receiver installations. The FCC licenses cover transmitters, not receivers. There is no FCC database with adequate information about CARS receivers to determine whether the FCC's interference assumptions are valid. Still, my guess is that properly engineered CARS systems with adequate rain fade margin will probably be immune to this new interference, at least most of the time.

But what's the point? It's a stretch to claim that this "experiment," relying on Dynamic Frequency Selection, is a valid test of the Interference Temperature concept. Anyway, it really isn't an experiment. If it goes forward (and by the way, that depends on comments that you submit in response to the FCC's Notice of Proposed Rulemaking), it will be a permanent reallocation of the spectrum. There is no plan or test protocol to establish hypotheses, test their validity, and then terminate the experiment if the hypotheses are not confirmed.

Usually, when the FCC reallocates spectrum, it does so in response to a request that describes the new use for the spectrum, and explains why that need cannot be satisfied in existing allocations. In this case, nobody has asked for this reallocation. There is no demonstration of need. Who knows how it will be used? Maybe the consumer electronics industry will discover a whole new unmet demand that this technology can meet. Well, maybe.

Have a comment? Contact Jeff via e-mail at: