Frequency-by-frequency licensing of the last hundred years will go the way of the vacuum tube.

In an obscure and largely overlooked Notice of Inquiry, the FCC has begun to overhaul the very foundations of radio communications.

The first practical radio transmitters, early in the 20th century, used a simple “spark gap” technology that spread signals over a wide swath of frequencies. This was not a problem when few transmitters existed, but as their numbers increased, the then-useful part of the spectrum soon became crowded.

Within a few years, engineers were using recently-invented vacuum tubes in conjunction with a circuit that limits a radio signal to a specific frequency. That solved the immediate congestion problem, as each transmitter could be assigned a frequency different from others in the vicinity.

Now, a century later, we still use that same system. Every licensed transmitter, whether flea-powered walkie-talkie or megawatt TV station, is assigned a specific frequency. Over the decades, as more transmitters came into use, the licenses gradually filled up each part of the spectrum. The engineers, though, kept finding ways to use ever-higher frequencies, and thus steadily pushed the supply of spectrum ahead of the demand. Back in 1984, when I started doing FCC work, there were plenty of unallocated frequencies below 1 GHz, open spaces up to 40 GHz, and almost nothing above. Today everything up to 40 GHz and beyond is filled in solid, with active use extending up to 95 GHz.

Worse, the tactic of opening ever-higher frequencies has now run out. Those pesky laws of physics limit the frequencies above 95 GHz to short distances, straight lines, and dry climates. The spectrum is effectively full and there is nowhere else to go, say the supposed experts.

But wait.

Take a scanner radio and tune through the bands. The local AM, FM, and TV stations will show activity, as will the cell phone bands and the amateur radio bands. The rest, though, is mostly empty. Even bands that are heavily licensed, such as those used for two-way radios, are actually used only infrequently, and will show sparse activity at most. To be sure, the scanner radio will miss satellite signals, fixed microwave links, “passive” applications (such as radio astronomy), and a few others. And some users, like a first responder who picks up a microphone to call for backup, see a vacant channel as a good thing. But the fact remains: although small pieces of spectrum are overcrowded, much of it goes unused most of the time.

The FCC has taken early steps toward smoothing out the load – for example, allowing licensees to lease out their excess spectrum to others. Industry engineers have come up with “cognitive radios” that respond to the ongoing state of the radio-frequency environment by changing frequency band, modulation, etc. to make better use of temporarily vacant spectrum. Every Wi-Fi laptop automatically sniffs out quiet channels within the Wi-Fi bands. Radios in another, less-used laptop band test the air to avoid radar devices that share the band. Still another band requires transmitters to share frequencies cooperatively. TV “white spaces” devices, if and when they appear, will either sense the spectrum directly or will check in with a database to find locally vacant frequencies.

The FCC now wants to expand automatic spectrum-sharing to help alleviate the shortage. Rather than offer specific proposals, however, it is on a hunt for ideas. The broadly-worded Notice of Inquiry asks for comment on several topics:

  • The current state of spectrum-sharing radios. Techniques for sharing include detecting and identifying other users’ transmitters, exchanging information among users to determine whether a frequency is vacant, and detecting changes in the noise floor to see if there is room for additional traffic. But straight sensing turned out to be an unexpectedly hard problem in the TV white space proceeding. Another hard problem is monitoring the noise floor. The FCC proposed something similar in its former interference temperature proceeding, since abandoned.
  • Use of geolocation and real-time database for checking frequencies. The white space proceeding will test this idea, if the FCC ever solves the problem of database design. After a two-year struggle, still on-going, an FCC official has informally indicated that a solution may be announced “real soon.”   Also, geolocation usually relies on GPS, which does not work well indoors.
  • Building interference suppression into radios. Most experts would agree that good receivers are a key element of efficient spectrum use. Yet the FCC earlier considered the imposition of standards on receivers, only to drop the idea
  • Improving interference prediction. Spectrum-flexible radios will have to avoid causing interference to other users. Effective technical rules must rest on good predictions of how radio signals at various frequencies behave, in different environments. The mathematical models needed for this work have lagged behind other aspects of radio technology.
  • Policy radios.” The next step beyond cognitive radios, these are transmitters programmed with broad policy constraints on spectrum usage. The concept is still in its early stages.

After inviting discussion on these topics, the Notice of Inquiry turns to nuts-and-bolts practical questions. How should the FCC test the new radios for technical compliance? How should it license them? How can it facilitate spectrum sharing between licensees and other users? What frequency bands can make best use of the new methods?

The end result of this undertaking, years from now, will be an utterly different radio spectrum – as different from today’s as today’s is from the spark-gap era. Transmitters will automatically hop among frequencies, stepping into vacant channels temporarily and then moving on. The result will be vastly more traffic in the same amount of spectrum.

We commend the FCC for reaching out to the public early in the process, even though the agency’s own ideas are still embryonic at best. The many people who will be affected by the coming changes – manufacturers, service providers, end users, regulators – will welcome the opportunity to help shape the outcome.

Comments and reply comments are due 60 and 90 days, respectively, after the Notice of Inquiry appears in the Federal Register.  We will let you know when that happens.