This is a circa-1997 draft of a feature I wrote for Wired magazine, a story that through a combination of factors, not the least of which was my slowness in making changes to a piece that had crested 13,000 words, was never published. I would like to mention however, that this story introduces ultrawideband radio a full 2+ years before even Markoff wrote about it. The point being that I was, at the very least, on the right track.
It’s very long and far from perfect, but it’s been languishing on my hard drive for years, and I thought I’d put it online. Maybe someone will find it useful if not entertaining. (Bob Parks liked it, anyway.)
Delilah and the Bad Boy
Hedy Lamarr and George Antheil were just the first of many renegades to believe in the power of spread spectrum radio. Now, after a 50-year battle with the military, the FCC, and powerful commerical radio interests, spread spectrum is poised to become the solution for high-bandwidth, wireless communications.
+++
In 1940, the designer Adrian and his wife Janet Gaynor invited composer George Antheil to dinner at their Hollywood home, explaining that Hedy Lamarr wanted to see him about her glands. George, who had been writing columns on endocrinology for Esquire to supplement his meager composer’s income, was not fully convinced the invitation was honest, but agreed to attend anyway. He wasn’t disappointed.
“I sat down and turned my eyes upon Hedy Lamarr,” he wrote in his autobiography, Bad Boy of Music. “Here, undoubtedly, was the most beautiful woman on Earth. Most movie queens don’t look so good when you see them in the flesh, but this one looked infinitely better than on the screen.”
George remembered reeling as words spilled from his tongue.
“But your breasts,” he stuttered, “your breasts—”
Hedy whipped out a notebook and pencil. “Yes, yes,” she said breathlessly, “my breasts?”
“They are too small.”
She made a note in her book. “Go on.”
“Well, they don’t really have to be, you know.”
She made another note, taking some time to do it.
“You are a thymocentric, of the anterior-pituitary variety, what I call a ‘prepit-thymus.’”
Hedy kept writing for a moment, and then said, “I know it, I’ve studied your charts in Esquire. Now what I want to know is, what should I do about it? Adrian says you’re wonderful.”
George began to stammer now.
“Go on, go on,” she said, becoming a bit restless. “The thing is, can they be made bigger?”
The conversation that evening sparked a long-lasting friendship between Hedy and George, two intellectual misfits who found refuge under the golden umbrella of Hollywood. The next day, the composer and the actress, known for her roles in Boom Town and Sampson and Delilah, would begin an unlikely collaboration on a torpedo guidance-control system they felt could help the Allied effort in World War II.
Their invention was a flavor of spread spectrum radio, a radically new radio communications technique that European and American engineers and scientists were secretly experimenting with during the 1940s. Instead of transmitting on a single frequency, this super-radio spreads its available power over a wide band of radio frequencies. Since only a fraction of the power is found on any single frequency, it is difficult to detect, and jamming one frequency won’t affect the rest of the transmission. Additionally, spread spectrum signals are modulated in such a way that transmissions resemble noise, making them almost impossible to detect and decode. Spread spectrum’s ability to resist enemy interception and jamming made it a perfect candidate for military communications.
Hedy’s own discovery of spread spectrum probably began after eavesdropping on conversations between her ex-husband, Austrian munitions magnate Fritz Mandl, and his colleagues at their Vienna home, where the young actress lived during the early 1930s before fleeing her oppressive husband and his Nazi pals for England and then Hollywood. Wherever she got the information, Hedy eventually patented, with George, one of the first conceptions of true spread-spectrum radio.
If the two had been versed in radio history and had understood how radio was expected to work, they probably wouldn’t have designed the system they did. But they had never built radios, and weren’t interested in commercial applications. They just wanted to help the US government win the war.
The movie star and the bad-boy composer inadvertently discovered something that has taken 50 odd years and a microprocessor revolution to rediscover. A technology, coupled with computing power, that can transform the regulated scarcity of radio spectrum into a shared sea of bandwith. As George Gilder sees it, spread spectrum, combined with other digital communications technologies, “could open up the entire spectrum as one gigantic broadband pipe … dark fiber in the air.”
In an age where information is currency, digital data networks form the lifelines for billions of people throughout the world, who rely increasingly on wireless communications – whether pagers, cell phones, or new wireless modems. But narrowband radio simply cannot meet the bandwidth needs of the rapidly approaching tomorrow.
Why? Information theory, a set of rules which explain that the world is set up in such a way that to send high bit rates on a narrowband system, you would need a perfect environment — one virtually without noise. In other words, the only way to get narrowband radios to transmit even 128 Kbps, the speed of ISDN, is to turn off any electrical device radiating interference in range of your transmitter and receiver. That means turning off every interfering device — from lights and stereos, to nearby cars and radio stations.
Spread spectrum, however, is designed to exist with noise. The same ability that was once used to hide transmissions from enemy eavesdroppers in wartime can now be used to send high bit-rate communications in noisy, interference-ridden, real-world environments. This technology can share the radio waves without interfering with others — something radio has never seen before.
But there is a problem. Spread spectrum requires a large swath of bandwidth to in which to operate, something only the likes of the US military can get its hands on. The FCC, still operating under the ancient, original principles set down at the time of the commission’s founding in 1934, has divided radio spectrum into very thin bands among wildly diverse users. Today it is nearly impossible to aggregate those bands into the continous acreage that spread spectrum needs. Too many users would have to surrender their frequencies.
So spread spectrum continues to be written off as a bandwidth hog, as well as an expensive and relatively esoteric technology suited only for the fringes of communications.
In the perfect, digital world — spawned on the 24-bit color savannas of George Gilder’s mind — computer-controlled radios with spread spectrum at their cores could revolutionize wireless communications by allowing completely decentralized, yet digitally efficient sharing of the radio waves. It’s wireless bandwidth on demand: when you need spectrum, your radio goes out and finds you some you can use. Translated, that means no more telco monopoly. No more wires. No boundaries.
For spread spectrum zealots, this is the beginning of another war. A war of wires, lines, radios, antennas — carriers. A battle for control of bandwidth. Their weapon? Spread spectrum. Their ammunition? Microprocessors and Moore’s Law. Their pinup? Hedy Lamarr.
ECCO MARCONI….ERGO MARCONI…
The young Italian electrical engineer Guglielmo Marconi was first introduced to radio through an article describing Hertzian waves (now normally called radio waves), which were discovered by H. R. Hertz in THE 1880s. Marconi believed that these waves could be used for signalling and he was soon sending Marconigrams using his wireless telegraph. He even reported the results of the Kingston Regatta in 1898, two years before he received his historic patent (number 7,777).
For radio to be a popular commerical prospect, Marconi realized, he needed to find a way to distinguish between different radios operating simultaneously — otherwise, everyone’s communications would be muddled together. He had to divine a method of combatting this interference. the solution he settled on was to separate users by radio-wave frequency.
This division is similar to the the distinct electromagnetic frequency that defines a particular color of light. Radiate energy at 5×1014 Hz, for example, and you get a particular shade of orange light. Tune your radio to a specific frequency, and you’d hear the communication you were looking for.
Following Marconi, radio, defined as a spectrum of frequencies, was easily quantified, divided, and doled out. It exploded as a popular comunications medium, used by thousands and then millions of different people, each operating within the boundaries of their frequencies. And almost imperceptably, radio changed — from a wild, uncharted electromagnetic sea to simple plots of real estate.
Suddenly, useable radio frequencies became quite valuable, not unlike beachfront property. In 1934, the US government established the Federal Communications Commission to manage the radio spectrum, doling out new radio licenses and arbitrating disputes when users interfered with each other.
Heading down Marconi’s frequency-division path had led to maximizing the use of specific, small bands of frequencies — narrowband communications. And that’s radio as we know it: the electromagnetic energies between 3 KHz and 300 GHz, which serve as home to FM and AM broadcast radio, VHF and UHF television, and their ilk.
But outside narrowband rules and regulations, the military has long experimented with spread spectrum radios that don’t tune in a specific frequency. These radios instead looked for specific radio waves in a wide radio band.
The two systems arrive at different solutions to a problem analogous to conversing in a cocktail lounge, packed with a boisterous crowd just let out from the theater. To continue a narrowband conversation, you and your select audience would have to elbow your way to a table in the corner of the room. The table can be thought of as your frequency. A spread spectrum conversation, however, can continue in the noisy lounge even when the participants are scattered throughout the room. The trick is selecting a language no one else in the lounge is speaking. Despite the din, the only words you hear are the ones in the language you’re listening for.
Spread spectrum conversations don’t require an FCC maítre d’ to perform complicated seating arrangements for each radio user, foregoing that type of order for a more Babelonian seating arrangement where radio users CAN arrange themselves based on whatever unused radio space is currently available and CAN converse using unique, sophisticated mathematical languages.
An 80-ton Decoder Ring
Between the two World Wars, work in radio communications focused primarily on protecting the privacy of conversation, finding some way of camouflaging – or encrypting – transmissions to prevent enemy eavesdropping. One solution was to cloak transmissions so they would sound like random noise to enemy ears. The intriguing and very difficult idea here was to develop two identical noise codes, one for encrypting a signal, the other for decrypting.
At the beginning of World War II, while Americans were tentatively watching the fighting escalate in Europe, Bell Telephone Labs developed a unbreakable communications system, nicknamed Green Hornet, that was used for conversations between President Franklin Roosevelt and Prime Minister Winston Churchill. The Hornet mixed speech samples with a noise-like cryptographic key, recorded on phonograph records pressed by Musak. The entire system, including power source, air conditioner, and a complete set of spare parts, weighed around 80 tons.
The cryptographic keys used to modulate the conversation were prerecorded and shipped to the receiver, where they were then synchronized to the incoming signal to extract, or correlate, the signal. Using each key only once gave Hornet its unbreakable security – independently verified by Bell Labs researcher Claude Shannon and the British mathematician and computer theorist Alan Turing – but the keys were also the Hornet’s greatest weakness. Having to send records between the two radios presented a serious distribution problem and a great potential weakness.
After seeing the Hornet and digesting Shannon’s work, Turing began developing his own secure voice system, which he christened Delilah, after the infamous biblical deceiver. Turing dispensed with the key records and their distribution problems, and instead focused on an electronic key-stream generator. But the generator was a problem he wouldn’t be able to solve until after the war.
The Hornet, like other similar cryptographic communications systems developed at this time, combined a signal with a noise code, masking it from eavesdroppers.
But this is not true spread spectrum. Not yet. True spread spectrum, as the name implies, involves spreading the signal over a wider band of frequencies than would normally be needed. While noise can be used to encrypt the signal, spreading offers an additional benefit: even if an enemy is aware that communication is taking place, jamming certain, discrete frequencies doesn’t destroy the whole signal.
One of the first descriptions of spread spectrum with this anti-jamming capability was a variety known as frequency-hopping, in which the signal jumps rapidly among several different frequencies in a seemingly random pattern. It is the order of the jumps — known only to the transmitter and receiver — that creates the encryption protection.
This frequency-hopping system was initially conceived in 1941 by Hedy Lamarr who, more than five years after Turing’s work with Delilah, would capture the hearts and libidos of American men with her famous portrayal of Samson’s betrayer.
Enter the Endocrinologist
At the end of the evening at the Gaynor’s back in 1940, George Antheil found a dinner invitation scrawled in red lipstick across the windshield of his car. If Hedy’s forwardness threw the married music director, it didn’t throw him far. Who was he to pass up an intimate audience with the starlet in her Benedict Canyon home?
George had toured Europe as a concert pianist, gaining a reputation as one of the foremost avante garde composers of the period. He returned to the States in 1933 and found work in Hollywood as a music director.
In need of extra cash, he met and impressed Arnold Gingrich, editor of Esquire, beginning a fruitful freelance career with the magazine, which included articles on endocrinology, the study of the hormonal glands. It was those columns that, four years later, would lead him to the starlet anxious for advice on growing her bosom.
That evening in Benedict Canyon, the conversation between Hedy and George began at chest-level, but soon turned to the status of the war, which in the summer of 1940 appeared bleak. Europe had been consumed in war and the US was teetering on the edge of non-belligerency while gearing up, in money and men, as the “arsenal of democracy.”
Hedy had been contemplating quitting MGM to join the recently formed National Inventors’ Council in Washington, DC. Her first marriage to Austrian munitions magnate Mandl, had allowed her unique access to discussions on new types of weapons. She felt it her duty to contribute to the war effort in a way that no other in Hollywood could.
“They could just have me around,” she told George, “and ask me questions.”
George convinced Hedy that she could better help public morale by remaining in Hollywood. But he was intrigued with some of Hedy’s other ideas. One of them, he later wrote, “was so good that I suggested she patent it and give it to the US government.”
During the summer of 1940, Lamarr and George began work on the invention, a torpedo guidance-control system, discussing it at length and taking copious notes over the next few weeks. At the time, when a torpedo was fired off a ship, the target vessel often changed course to avoid being hit, necessitating a corresponding correction in the torpedo’s course. But even with radio control, the enemy was often able to detect the signals and jam them, preventing the torpedo from being steered at all. Hedy and George worked up a system that would change frequencies in a predetermined pattern known only to the transmitter and receiver. “Without knowledge of the records,” they explained, “the enemy would be unable to determine at what frequency a controlling impulse would be sent.”
As it happened, George was precisely the man to help Hedy build the noise into her communications system. He had gained his reputation as “bad boy of music” for raucous and inventive scores like his Ballet mécanique, scored for 16 synchronized player pianos. When the piece premiéred in Paris in 1926, he had only one piano, but he made up for that shortcoming with electric bells and airplane propellers.
Their guidance system employed a device similar to the records used in Bell Telephone Lab’s Green Hornet: the piano roll. These rolls are essentially simple data-storage media, not much different in principle from the old computer punch cards or a cassette tape. A player piano is designed to pull data off the roll, turning it into audible notes. Hedy and George used the rolls to program a path of frequencies the transmitter and receiver would communicate over.
They reasoned that since a player piano had to be able to strike 88 keys, their system could use 88 different frequencies, more than enough, they figured, to elude enemy eavesdroppers.
The biggest problem, as with all spread spectrum-type systems of the day, was synchronizing the receiver with the transmitter. For Hedy and George, this meant starting both piano roll records at precisely the same time so that the receiver would be on the right frequency whenever a controlling signal was sent.
Each “note” marked on the piano roll corresponded to a frequency. Synching the transmitter and receiver rolls meant that whenever a new “note” was “played,” the radios would simultaneously switch frequencies. The line of communication was always open, though it seemed to switch quickly and erratically.
On August 11, 1942, they were granted US patent 2,292,387 for a “Secret Communications System.” The two brought their patent to Charles Kettering, head of the Inventor’s Council, who classified it “red hot.” Yet, the patent was routinely issued and never used. Because its authors were out of the scientific loop, their torpedo-control system fell into the shadows where it languished, unnoticed.
Noisemakers
Twenty years would pass before a frequency-hopping spread spectrum system was finally placed into active military service, in the Sylvania-developed BLADES communications system deployed on US Navy ships during the Cuban missile crisis. But here, Hedy’s patent had no impact. Sylvania had been working on the technology for years — and its system was no mechanical contraption of wires and piano rolls, but an electrical system, an early vestige of the emerging age of digital signals and computer processors.
Up until the early 1950s, spread spectrum had to rely on mechanical systems to generate and store the noise codes. The size and complexity this approach engendered severely limited the potential applications. You simply couldn’t tote a system like the Green Hornet around in an airplane or missile.
Just five years later, just before Christmas 1947, two Bell Telephone Laboratory researchers, Walter Brattain and John Bardeen, constructed the first transistor. Their invention obviated the vacuum tube, opening the door for electrical systems that used smaller electrical currents in much smaller spaces. Spread spectrum radio engineers were some of the first to test the new technology.
With a handful of transistors, they could now generate extremely long noise codes with a simple, digitally coded mathematical formula. This electrical kickstart enabled radio engineers to experiment more freely, and they soon began mapping out the capabilities of spread spectrum. In 1953, engineers at Sylvania discovered a key concept: processing gain. In other words, the more you spread, the better spread spectrum worked.
It’s a concept that Claude Shannon had outlined while creating the new field of information theory. In 1948, Shannon, then a 32-year-old researcher at Bell Labs, completely changed the face of the information age with his two-part article “A Mathematical Theory of Communication.”
Shannon had discovered a mathematical definition of information and a way to measure it using entropy — in radio terms, interference or noise. His theories coalesced during his early cryptography research at Bell Labs and his work on the Green Hornet. Shannon’s theories gave electronics researchers rules to follow and theoretical targets to shoot at.
What’s now known as the “Shannon limit” explains why new technologies like ADSL, which don’t use the phone company’s telephone network systems, can transmit data up to 6 Mbps. Shannon found that in a real-world environment, one containing entropy — or noise — there is only so much information you can push through a given bandwidth at a given power level.
What’s most interesting however, is the relationship between bandwidth and power. The Shannon limit shows that increasing bandwidth will directly increase your information capacity, while increasing power has only an indirect effect on capacity — and with decreasing returns at that.
The difference here is simply the difference between a shower and a bath. The shower has higher pressure, or power, but its water flow is much less than that of the bath faucet. A shower manages to get you clean, but a bath, well…. The benefits of bigger pipes should be obvious.
Although it may still seem counterintuitive, Shannon’s limit explains why spread spectrum can easily achieve high data bandwidths with low transmission power, and why traditional narrowband radio, even with tremendous power behind it, is inherently tied to lower bandwidths.
Because noise is a reality every radio system has to deal with, Shannon’s limit shows that to send a 156 Kbps transmission, an analog cell phone using a 30 kHz bandwidth would need a signal power more than 158,000 times as great as the level of noise in the channel. That’s simply not possible, especially with those little cell phone batteries.
A spread spectrum radio, however, using a 1.25 Mhz channel, needs a signal power only a third as great as the noise to communicate. Even though the signal is less powerful than the noise, communication is still possible.
This is the key to the military’s interest in spread spectrum. When you spread far enough, the signal power drops below the average level of noise. The signal effectively disappears. You can’t hear it. You can’t detect it.
In 1951, engineers at ITT’s Federal Telecommunications Laboratory used a spread spectrum system to send a transmission over a Bing Crosby broadcast (the interference) on station WOR in New York. The test was designed to see how well their radio would work in an extremely noisy situation.
While Bing himself might not have appreciated being used as interference, the power of spread spectrum was clear even then. The demonstration showed that you could transmit and receive a signal 1,000 times weaker than the interference. Just like magic.
Today, there are two main flavors of spread spectrum: “Frequency hopping” still works like Hedy and George’s system: though powered by today’s microprocessors, it jumps around its alloted bandwidth quite a bit faster. Hopping works like it sounds: the radio sends a chunk of the signal, then quickly switches frequencies and transmits again.
“Direct sequence” is a bit more complicated: it combines the data signal with a long, seemingly random pseudonoise code, the result of which is then transmitted over the entire available bandwidth. The receiver, armed with a duplicate code, knows exactly what to look for on the radio waves. When it picks up a signal that looks like the code it’s looking for, it pulls it in and reconstitutes your message. Despite the fact that the amount of data being sent is now much greater, the receiver and its microprocessor brain are able to sift through all this information very quickly.
The Obstacles
In some ways, direct sequence spread spectrum works like a giant wireless Ethernet. Everyone shares the same pipes, yet each user only receives the data meant for them. It’s decentralized, invisible to the user, and governed by microprocessors. Anarchic yet elegant. And similar to Ethernet, described by its inventor Bob Metcalfe as a technology that works in practice, but not in theory, spread spectrum doesn’t fit well in an FCC regulatory structure that was singulary defined by Marconi’s century-old ideas.
The Telecommunications Act of 1934 established the FCC and the rules for managing the radio spectrum. The spirit of those rules was largely informed by Marconi’s vision, which was itself informed by the meager technological means of 1900. And so they also reflect a 19th-century worldview vested in protecting A new, invisible resource for what was then considered the public interest, while helping entrench the hegemony of the small clique of powerful radio owners.
The entire FCC-founding document could not keep step with the amazing strides radio, essentially still a young technology, would be taking after World War II. As it turns out, there is more than one way to divide the radio spectrum — yes, you can separate users by frequency, but you can also separate them using spread spectrum.
Of course, no method is magical. No matter how you use the radio spectrum, there is a limit to how much information you can squeeze in. But it is this very limit that should be driving radio — all radio. If spectrum scarcity demands a spoonful of government regulation to maximize its benefit to society, why not then a spoonful of technology?
Techniques like frequency division have little room to grow. Once you dole out spectrum to many different users, it becomes inflexible and resistant to changes or improvements.
Spread spectrum, however, engenders flexibility by transferring that regulatory duty to microprocessors. The radios themselves can negotiate the spectrum better and faster, allowing more users per band, and making spectrum use more efficient by eliminating all the wasted guard bands needed to separate narrowband radio users. And as the radios get better with more powerful chips, the more you can do with the little you have.
Back in the 1970s, it was clear to a number of the engineers who had worked on military spread spectrum that the technology had commercial applications. The 1950s saw the development of two mature military systems — the F9C of MIT’s Lincoln Lab and the ARC-50 of Magnavox — proof that spread spectrum radio worked. But now, spread spectrum’s military incubator became a problem because of its secrecy, and the daunting task of pulling together information on the state of the technology would have to begin if the civilian world would ever be introduced to this radio technique.
Most of this work, including the F9C and ARC-50, had progressed independent of any outside assistance — even radio engineers from other projects. As Robert Dixon, who worked at Magnavox in the late 1950s, explains, “We built the ARC-50 without input from anybody. There was no sharing as far as we were concerned.”
Over at Sylvania, which had been working on spread spectrum since the 1950s, there was a comprehensive manual on spread spectrum — complete with a cover indelibly marked “Secret,” says Mike Marcus, now associate chief of the Office of Engineering and Technology at the FCC. “There was a lot of spread spectrum technology around,” Marcus adds, “but it was all military and classified. There were no good general discussions.”
Nothing in the unclassified literature helped people build even military systems, Dixon adds. “I could see that it had commercial possibilities.” So the engineer finally took it upon himself to write a guide book, Spread Spectrum Communications with Commercial Applications, which was published in 1976.
Then, in 1981, Dixon started Spread Spectrum Sciences, building the PacTel TeleTrac system for vehicle monitoring and location. PacTel had successfully lobbied the FCC for its spectrum. But until the mid-1980s, there was no allocation for general, unlicensed spread spectrum devices, so no one was experimenting.
Getting regulators and corporations to see the benefits of spread spectrum proved to be a dangerous business — not unlike promoting communism in government halls. No matter what argument you threw at them, be it security, efficienct spectrum use, or non-interference, sooner or later you’d end up mentioning the word “share” — it is, after all, one of spread spectrum’s prime benefits. But radio spectrum was, and is, viewed as property. Something to be bought and defended, not shared.
In 1981, Marcus at the FCC put together a Notice of Inquiry soliciting comments on whether the commission should investigate the possibility of allowing unlicensed spread spectrum communications in all bands above 75 MHz at up to 100 watts of power. His proposal was the grail spread spectrum advocates had always hoped for. But Marcus had unwittingly opened a nasty Pandora’s box.
The proposal suggested that, at low enough power levels, spread spectrum signals resemble the electromagnetic noise that PCs emit. So, Marcus figured, why not allow useful transmissions if similar, spurious ones weren’t causing any harm?
“That really had people pissed,” he explains.
Although he refuses to talk about it in detail, others familiar with spread spectrum lore paint a compelling picture of Marcus, who emerges as a regulatory Robin Hood, the guy who saw the potential of unlicensed spread spectrum devices to liberate radio from the ruling forces.
Those who know their history, however, add darker shades to the tale. Spread spectrum evangelist Dave Hughes says Marcus was likely demoted by officials and powerful corporate radio interests who saw his spread spectrum proposal as a dangerous and particularly unfunny stunt.
Unlicensed spread spectrum, sharing the bands that others had paid good money for, was an idea ahead of its time. Mike Kennedy, who worked with Marcus at the FCC and helped draft the Notice Of Inquiry, explains that their idea was quite innocent: “The technology was out there, in government communications, and we thought there was some promise.”
Despite the setback, unlicensed spread spectrum became a reality in 1985. After some serious revisions were added, at the request of intelligence agencies and industry alike, three bands were authorized in the FCC’s Part 15 rules for unlicensed spread spectrum communications at one watt of power.
At the time, the bands weren’t heavily used, and the users in the bands were convinced that spread spectrum devices wouldn’t pose any serious threat. Today, these are what’s known as the garbage bands: 902 to 928 MHz, 2.4 to 2.4835 GHz, and 5.725 to 5.85 GHz. You’ll find everything in these bands: from cordless phones to industrial, scientific, and medical equipment.
Unfortunately, the expected demand for spread spectrum devices never emerged. According to Dewayne Hendricks, who is investigating educational wireless apps for the National Science Foundation, “Everyone focused on LAN stuff. But they were never able to get the radios down to the price point of Ethernet cards.” Some companies built high-end spread spectrum systems that undercut expensive fixed microwave radios by 25 percent, and that was as far as it went. Not exactly a revolution.
That was spread spectrum in 1985: the technology was out there, it was given an opportunity, and it flopped.
The Dig
This is not to say that spread spectrum could or should have taken over the world then and there. The establishment of the unlicensed spread spectrum bands merely defined the battlefield. Now an army would have to be amassed, allies found, and a cause developed to rally around.
Spread spectrum radios depend heavily on their processing power to translate their encrypted signals. In 1985, the fledgling PC industry had only just birthed its first children, and the digital signal processing chips used in spread spectrum radios were still extremely expensive. But by the 1990’s Moore’s Law would change that, and the explosion of computing would beget the fast, inexpensive chips spread spectrum so desperately needed.
By 1984, declassification of spread spectrum documents allowed engineers to get a comprehensive look at the technology, leading to the publishing of the Spread Spectrum Communications Handbook, by Marvin Simon, Jim Omura, Barry Levitt, and Robert Scholtz. This was the bible the technology had heretofore lacked. It helped unify different techniques and their histories. Robert Scholtz was the book’s historian, contributing the research he had done over the past seven years to unify spread spectrum’s origins.
Scholtz’s journey into spread spectrum’s past began in 1977, when Bill Lindsey, editor of a special issue of the IEEE journal asked Scholtz to write a tutorial on the technology.
“The Spread Spectrum Concept,” a communications analysis published that year, brought Scholtz, now a chair of the electrical engineering and systems department at the University of Southern California, in contact with Paul Green, who had been wading in spread spectrum’s state-of-the-art while at MIT’s Lincoln Lab in the 1950s. Green had also been working on a history of the technology, but was hesitant to write up his work because of government secrecy issues. (Even Green’s thesis was secret.)
“He was afraid of letting something out of the bag that not everybody knew about,” Scholtz recalls. “So he asked if I would be willing to do it.”
Scholtz began investigating, and soon it became clear that many different places had likely been working on the technology. With the help of Green and another MIT colleague, Bob Price, and despite the fact that many old papers and documents were still classified by the government, the three began to pull away the veil that had shadowed spread spectrum for more than 40 years.
“It was Bob who uncovered Hedy Lamarr,” Scholtz says. “It had been sort of known, but it hadn’t been popularized, so most radio engineers didn’t know about it.” Price checked out the lead. He did the patent searches and even went to Los Angeles to confirm the story in 1982. He talked with Hedy, taking a publicity shot of her along for good measure. Price wrote up the Hedy Lamarr story and later gave Scholtz permission to use it in The Spread Spectrum Communications Handbook, published in 1984.
“He was quite enthralled with her,” Scholtz says. “Look at the picture of Hedy that’s in the book; you’ll see some handwriting to the side.” Sure enough, it reads: To Bob Price, Best wishes always, Hedy Lamarr.
Foot Pounding
By the 1990s, spread spectrum had discovered its past and was ready to move into the future with its new ally, computing. All it needed was a a few good applications. Those apps emerged because the computing industry quickly became more about communications than the communications industry itself. Personal computing demanded a way to talk to other machines, and spawned networking, LANs, and Ethernet.
But then came mobile computing, necessitating new wireless solutions. that, coupled with the rise of the Internet, has forced a complete reevaluation of our communications infrastructure. The phone companies are feeling it, consumers are feeling it, and spread spectrum may finally see its day.
In the early 1980s, Andrew Viterbe, now XXX at Qualcomm, was researching how to leverage spread spectrum’s ability to share bandwidth within the constraints of a cellular phone system. Qualcomm’s code-division multiple access technology, CDMA, is a variant of direct-sequence spread spectrum (as an international standard CDMA is known as IS-95).
A real CDMA system had to wait until the FCC authorized the new PCS cellular bands, but now that it has, CDMA cell phone networks are sprouting up around the country, offering digital service and a network capacity around 10 times greater than comparable analog cellular systems.
CDMA may be an extensively modified spread spectrum system designed specifically for a cellular network and licensed exclusively by Qualcomm, but another company has sprouted from the woodwork with a wireless data solution that puts the FCC’s unlicensed spread spectrum bands to good use.
Metricom Inc., started in 1985, has rolled out three metro-area, spread-spectrum packet radio systems, along with systems on university campuses, schools, and Sun Microsystems’s corporate campus. The Los Gatos, California-based company’s Ricochet system uses cell phone-sized, frequency-hopping spread spectrum radios to provide wireless Internet access at speeds that approach 28.8 Kbps. It’s currently serving thousands of satisfied customers in the San Francisco Bay Area, Seattle, and Washington, DC.
Although Ricochet has to share the crowded spectrum within the unlicensed 902 to 928 Mhz band, its performance to date has been impressive, especially considering that the network does not rely on multimillion dollar radio stations, but instead on a well-placed army of small radio receivers hung on the tops of utility poles throughout an area. A user’s radio broadcast is picked up by the nearest receiver, which shuttles the data through other receivers until a Wired Access Point is reached. This larger receiver is connected by wire to a Ricochet gateway, and then, to the Internet.
The benefits of the system are its tremendous flexibility and low cost. Adding microcells on a few light poles increases capacity, while providing users with new, improved radios would increase their access speeds.
Despite the apparent ease of a system like Ricochet, it has not taken off nationwide. This is partly, at least in Metricom’s case, because the Ricochet technology is patented and the company has so far only licensed its system to a few parties. But despite its ingenuity, the Ricochet system is not robust enough to provide higher data rates, in the ISDN range, or phone service. That problem is left for the bigger telecom players.
AT&T announced in February, 1997, that it was using the $1.7 billion in PCS licenses it won in the 1995 FCC auction not — as everyone expected — for another cellular service, but a new type of fixed wireless solution. Industry jaws hit the floor.
What AT&T proposed was a unique product for the local access market, long the exclusive territory of the regional Bell operating companies. Instead of ordering a phone line, AT&T wants to offer consumers a pizza-box sized dish that connects to a neighborhood antenna. Phone lines are connected to a box wired to the dish — service works as usual, with the addition of a connectionless 128 Kbps data link to the Net.
The most forward-looking option of the system, however, is a wireless base station technology that would allow an appropriately enabled cell phone to automatically switch from premium mobile rates to cheaper fixed rates when it was in range of the home’s base station. The system’s potential is seductive: one phone, one number — and, of course, one carrier.
The AT&T system, currently being tested in Chicago by a select group of AT&T employees, is not exactly spread spectrum — it’s a proprietary scheme designed for fixed service that the company has declined to discuss in detail — but its underlying principles are the same, illustrating the possibilities of what could be done with spread spectrum technology if radio spectrum were available.
The Learning Society
Radio is strange and difficult beast. Being invisible, it seems a single, distinct entity. But actually its properties change dramatically according to frequency. At the lower end of the frequency scale, the realm of TV and radio, signals propagate well, even through walls. But as you rise into the gigahertz range, what’s known as “line of sight” becomes a serious limiting factor. To communicate, each radio’s antenna must be able to see its receiver’s antenna clearly, without obstructions.
This isn’t a problem for satellite-based radio, just point your antenna toward the heavens — straight up. But even in high-frequency satellite transmissions using microwave spectrum (above 8 GHz), seemingly innocuous raindrops become obstructions as the water droplets begin absorbing the signal.
Given that most of spread spectrum’s assigned radio spectrum is between 2 and 5 Ghz, its limited propagation characteristics aren’t creating a market for spread spectrum pagers. But with a bit of work, spread spectrum radios with directional antennas (which form endpoints on a line of transmission — between two buildings, say) can be used to create cheap, high-speed wireless links that replace costly T-1 data lines rented from the phone company.
As far as Dave Hughes is concerned, spread spectrum, applied to these types of fixed wireless networks, is the answer to the government mandate to wire the nation’s schools. And connect our children. And their teachers. And our communities. And if he has his way, saving America’s children will be the rallying cause that leads spread spectrum out of the badlands.
He, of course, will be Moses. Through a half-million dollar grant from the National Science Foundation, Hughes, as the project’s principle investigator, is using spread spectrum radios operating in the unlicensed Part 15 bands to wire schools with high-bandwidth links at prices far cheaper than the telcos could ever hope to offer. Nothing since the Internet itself has won him over like spread spectrum has. “It’s still the only fundamental technology,” he expounds, “that has the potential for non-interference, security, and high bandwidth that obviates the need for extreme restrictions on the range, power, or location of the radios.”
It’s probably no surprise, at least to seasoned netizens, to find Hughes smack dab in the middle of the spread spectrum revolution. The grizzled, XX-year-old Coloradan first built a reputation online using BBSes to interconnect citizens with each other and the world (Wired 1.2, page 62). His online proselytizing is the stuff legends are made of — most notably his own. He is one of those netizens who transcends his biological manifestation — David Hughes is best, and more widely, experienced not as the man, but as the infinite potentialities of a blinking cursor from which his mind and being pour forth.
And now, Hughes is pumping his vigor into wireless. For him, spread spectrum is yet another weapon in the fight for, as he calls it, “grass roots, up from anywhere, by any culture on the globe, personal and organizational networking.”
And Hughes, with his own inimitable style, is pulling out all stops to make his point. Get him talking and he will hit security and cheap bandwidth, greedy telcos, notable advocates of the technology like Paul Baran (the father of packet switching) and George Gilder (the telecommunications evangelist and FCC groupie). At some point down the line, he’ll make his way back to World War II and a story about a young movie star who developed a spread spectrum system for guiding torpedoes.
In fact, Hedy has worked so well in pricking up the ears of a bored and ill-informed media, that Hughes has even suggested, only half-jokingly, that the nascent spread-spectrum industry attach to their radios a small logo: a sketch of Hedy emblazoned with the words “Hedy Lamarr Inside.” If that doesn’t sell radios, well, nothing can. But regardless of his methods, his vigor is helping push spread spectrum from the sidelines into the mass market spotlight.
As Hughes sees it, “we have enormous pressures to solve the problems of connecting up 55 million kiddies, 3 to 4 million teachers, 84,000 schools, 15,000 public libraries, required by law to have advanced communications capabilities.”
He has chosen a compelling strategy, for while the telcos rail against public, unlicensed wireless which would cut them out of billions of dollars of business, the government has mandated that the schools be wired. The FCC and the telcos agreed to start a Universal Service fund, maintained by the telcos and filled with money from the public’s phone bills, to pay for educational infrastructure upgrades. Now this is fine and dandy for big telco, since the public ends up subsidizing the cost of their own equipment — it’s a closed system in which the telcos really don’t spend or lose a dime.
But suggest, as Hughes has, that the fund could be used to buy wireless equipment, adding that you are, for all practical purposes helping alleviate the problem of Net traffic overloading the phone companies’ aging voice network, and prepare to hear the sound of lax jaws swinging.
In a post on The Well, Hughes recalled making that suggestion before the FCC: “Now that we have, for all practical purposes, an unlimited spectrum, let’s get serious about using the ‘public spectrum’ for ‘public purposes’ like connecting up all schools to the nearest POP; and all the students at home, plus the teachers, to the school servers — free. Just buy the radios, not the services. And save the taxpayers, for the T-1 link from the schools only, about $16 million a month, which otherwise will go into the pockets of the telcos, whose right to such income I really don’t see in the Constitution.
“Well, you know, they (FCC staffers and others to whom I made the pitch) get funny looks on their faces. Here I am trying to do the phone companies a favor, and they don’t appreciate it.”
But above the fray, one simple fact is killing the telcos — wireless is cheaper than wires, especially when it comes to high-bandwidth data connections.
In Belen, New Mexico, Hughes networked the district’s eight schools together using eight pairs of spread spectrum radios made by Solectek. The bid from US West included $8,000 in hardware and used T-1 wirelines to interconnect the school buildings at a cost of $84,000 per year under a five year contract. The cost of the radios was $108,000 up front.
After five years, the US West bid would have cost the district $428,000. The wireless solution still cost $108,000. “That’s real goddamned money!” pops Hughes. The only real drawback to this case study is that the district wasn’t able to use the digital radios to connect to their nearest Internet POP, around 30 miles away. Operating at the FCC-specified 1 watt, the radios aren’t strong enough to reach that far — yet.
Driven by demand for alternatives to expensive leased wirelines from the phone company, the technology is developing at an extremely rapid pace. Freewave Technologies of Boulder, Colorado, (**(303) 444-3862/ swulchin@freewave.com) now offers a $XXX radio rated for 3.2 Mbps at a distance of 7 miles. San Diego’s Solectek (**(619) 450-1220/Amy Starks) builds a $XXX monster with IP routing that will do 10 Mbps to 25 miles. With the increased availability of faster and more powerful digital signal processors, companies are pushing more and more out of that one watt.
Which makes the technology not only critical for schools, but Third World nations where there is no wired network at all.
Third world countries like, for example, Mongolia, where Hughes and his co-investigator, Hendricks, set up a spread spectrum wireless network last year. The phone system there was so bad that, as Hendricks explains, “you couldn’t even sustain 14.4, and people were having a hell of a time getting into their POP.”
The solution was a network of radios to connect four Mongolian institutions, a library, and the US embassy – all via 115 Kbps links to a satellite ground station that connects to the Internet and the world. They even have a home page – www.magic.mn.
The network puts Mongolia, as Hughes says, “more into the future of telecom than Washington, D.C.”
In developing nations, which don’t have the luxury of being able to purchase wired connectivity — at any price — wireless can help build a modern communications infrastructure that wires cannot. And yes, as Hughes loves to point out, it can bring the “Mongolians, on their Manchurian ponies, to the library to surf the Net for free.”
Power Control
Because spread spectrum radios can easily share spectrum with other users, building a successful market for these devices doesn’t depend on the FCC granting the radios a large, exclusive chunk of unused bandwidth. That dream will never materialize anyway.
First off, there isn’t a large enough band of exclusive spectrum available. And unlicensed devices don’t bring in the auction money, so the government would have to just turn over the spectrum for free, something it’s not likely to do when it could sell it off for a few billion dollars. But spectrum that’s already being used could be authorized for spread spectrum devices. And if if it included enough contiguous bands, if it was wide enough, the radios could spread enough to make interference with other users unlikely.
Bandwidth is spread spectrum’s only real limiting factor now. Moore’s Law is helping fuel better chips which make more powerful radios. The better the radios spread, the higher bit rates they’ll have, the less they’ll interfere, and the more appealing they’ll be in the market.
But even investing in technology is a hard road within radio’s regulatory environment. “The computer industry doesn’t have to go to government and say, Can I put out a new model?” says Hendricks. “In radio, you do.” And the R&D costs are still too high to justify the investment in a device that operates in the unlicensed badlands.
Which is why, with the help of a NSF grant, the Tucson Amateur Radio Project built a reference model. In the fall of 1997, they designed a 1-watt radio with an estimated 128 to 500 Kbps throughput at distances up to 20 kilometers.
Theirs is a prime example of what can be done if you’re willing to work with what spectrum is there. But the truth of the matter is that current spread-spectrum radios designed for the Part 15 bands aren’t spreading that much. 902 to 928 MHz? That’s only 26 MHz to spread in. If you could find 100 or 300 MHz, then you’d really begin to see what spread spectrum could do.
To build a really great radio, what you want to do is find pasture — acres of bandwidth you can spread over. You could try going up into the high microwave frequencies — 30 Ghz and the like — but your propagation characteristics (which determine how the radio waves travel) are so bad that rain tends to absorb the signal transmission and your antennas have to be in line — perfectly. Not the kind of setup that makes for a mass market consumer app.
What if you could completely rethink radio? What if you used the whole spectrum? Ignored all the FCC rules. Ignored spreading within a limited, licensed band. Toss everything everyone’s told you about radio into the shredder and start over.
That’s the idea driving a tiny startup in Huntsville, “a city surrounded by Alabama” — the home of Time Domain and a technology called impulse radio. What impulse radio does is quite simple. It uses the whole spectrum — FM and AM radio, television, cell phones — all of it.
The idea for impulse radio came to Larry Fullerton in 1976, while he was studying electronics and engineering at the University of Arkansas. He began working with different types of waveforms for radio communication. Traditional radio waves, the ones Marconi worked with, are sinusoidal — they oscillate up and down, like a sine wave. Fullerton wanted to see if you could communicate with other waveforms, like the wave emitted by lightning, which is simply a big spark, a large surge of energy that doesn’t oscillate.
“He got this bug in his ear,” says Withington, “and spent most of his spare cash on test equipment to operate his home lab. Over a period of about a decade, he developed a fairly thorough concept of how this radio should work.”
Fullerton’s impulse radio uses a waveform called a Gaussian monocycle. The wave is transmitted around a center frequency, say 2 Ghz, and, because of it’s physical characteristics, spreads an equivalent 2 Ghz around that center. Graphed, it loosely resembles an umbrella, as the power levels fall off the farther you get from the 2 Ghz center, until they drop far enough to be unusable.
This is completely different from traditional radio, which oscillates over a group of frequencies at a single power level.
In the late 1980s, the patents Fullerton had applied for began to be issued. The only thing left was to start up a company. And with the financial help of some locals — typical Hunstvillian PhD-types — Fullerton started Time Domain.
The company’s first challenge, however, wasn’t demonstrating a feasible commerical product — it was convincing people that its technology didn’t completely violate the laws of physics.
That was the problem Withington faced when he came on board. A working prototype wasn’t enough, he had to have some big name, scientific backing. So he asked around, and the name that came up was Robert Scholtz. Once Time Domain got Scholtz to examine the technology and publish a paper in 1993, things got a little easier.
“I remember going into a meeting with that paper in my briefcase,” Withington recalls, “and a fellow said, ‘Well I just can’t believe that you guys can do this,’ and I said, ‘Well here’s the paper that describes it.’ I threw it in front of him, and he didn’t even read it. He just looked at who wrote it and said, ‘Oh I know Bob, I’ve known Bob for 30 years. If he says it works, it works.”
One obvious problem for impulse radio is that the band it operates in is so wide, it overlaps users of cell phones, TV, and broadcast radio, to name only a few. The reason Time Domain believes it can get away with stepping over everyone else’s bandwidth is simple: no one will ever know it’s there. Withington explains that the company’s dealings with the FCC so far “relate to applications where people would prefer that the radios not be known to be use. The term is covert. Impulse is pretty good, and for public consumption, I’d rather not say how good.”
Taking Shannon’s limit a little farther than before, impulse radio spreads so much that its transmission power is distributed over that whole band, 2 Ghz. From a narrowband perspective, so little power exists within any thin slice of spectrum that impulse radio is almost undetectable.
Again because of Shannon’s limit, impulse spreads so far that Time Domain can drop the transmit power. How far? How’s less than one milliwatt? At that power level, the company has built a radio that does 156 Kbps to around 10 miles.
For comparison, your cell phone uses about 600 milliwatts. Your watch — around 60 microwatts. That’s 60 millionths of a watt.
The almost insignificant amount of power Time Domain’s radio uses has caused confusion, however. At one demonstration using a 50 microwatt radio, audience members, though wisened in the ways of spread spectrum, thought Withington was cheating. They set up a spectrum analyzer to check, but couldn’t find the radio’s signal.
“Well, we convinced them that they had to move the analyzer closer,” says Withington. “About six inches away, they saw the energy from the transmission.”
Normally, a radio is licensed to operate in a certain frequency band at a specific power level. In addition, there are out-of-band power limits which are set so a radio won’t interfere with its neighbors. But those low interference limits set by the FCC are more than high enough for impulse radio transmissions.
One of the primary obstables preventing spread spectrum from becoming a more popular communications solution is its lack of available bandwidth. Yet here is impulse radio spreading across huge swaths of spectrum — and it’s virtually undetectable. Those FCC rules that have kept spread spectrum on the sidelines simply don’t apply to impulse radio. Sixty-odd years of regulation have been rendered irrelevant. You begin to wonder whether the intelligence agencies saw this coming when Marcus approached them with his plan in 1985.
Impulse technology is still in its infancy, however. Up to now, Time Domain has had to build the necessary circuits itself. Not that it’s difficult — the circuits are actually quite simple. Since impulse radio uses all of the spectrum, it doesn’t need to add extra circuits, like filters which limit a radio’s operation to a certain frequency band.
What it does need is fast silicon. Impulse encodes data by shifting the signal at a resolution of around 20 trillionths of a second. To use higher frequencies, like 10 Ghz, and maintain accuracy, Time Domain needs fast processing chips, faster than anything currently available.
That means courting the computing industry, in this case, Big Blue and its silicon-germanium chips. “It’s such a new process, we had to find a design team and get them trained, says Withington. “IBM didn’t have the people to do it.”
Already clamoring for high-bandwidth wireless solutions, computing interests are shaping the future of radio at the head-end as well. And with IBM’s help, Time Domain hopes to shrink all the radio components onto a few silicon chips. Two to be exact.
The result should be nothing short of a radio miracle, a cheap, silicon-based impulse radio that uses around one-thousandth the power of a cell phone with better performance. All because Marconi’s rules of radio were finally, completely ignored.
Withington is excited himself. “People look at this and they go, ‘My god, you’ve built the simplest possible device and it overcomes all the problems that we’ve been making our circuitry incredibly complex to try and solve, and we don’t come anywhere near the performance you guys achieve.’
“One of the things I’ve learned being here is that the most beautiful circuits are typically the simplest,” he continues. “It is a truly beautiful technology. It’s hard for people who are not in technology to understand that kind of beauty, but it really is fantastic.”
CODA
Radio has always been one of those mysterious technologies that, despite its ubiquity and permanence, never seems to garner much interest outside tight circles of engineers and hackers. As far as the public is concerned, radio works, what else is there?
But spread spectrum will be difficult to brush off much longer. It has traveled from darkened military halls to the wide-open public spaces. From a classified, exclusive technology to one with the potential to include individuals in a way narrowband radio never had.
If you need a historical marker, according to Telegeography Inc., the cost per subscriber of fixed-wireless solutions for telephony has, in the last few years, passed under that of traditional copper wiring. In other words, copper is no longer the communications carrier of choice. Digital wireless networks, like the fixed wireless system AT&T is testing in Chicago, can be a smarter, more economical choice. And as computing power — microchips — obediently follow Moore’s Law to become faster and cheaper, spread spectrum will more often be part of the solution.
This is the radio Hedy started with, this is the radio we end up with. After receiving the Electronic Frontier Foundation’s Pioneer Award this past March for her work in spread spectrum more than 56 years before, the now reclusive, 82-year-old Floridian is still — if inadvertently — helping to popularize a communications technology in the way only Hedy Lamarr has ever been able to do.
In fact, it was Hughes who nominated her because of, as he wrote, the wonderful juxtapositions of her story — the beautiful woman whose career was as far from technology as it gets, her desire to help her country in difficult times — and the example her work might set for others.
It was certainly a mouthful, one designed to drag the spotlight back onto Hedy and spread spectrum, yet it is largely true. And who’s to say that it isn’t exactly fair, for while she does serve as pinup and poster child, her name and face may help wire our schools and ease our bandwidth woes. People, as she is first to admit, have always paid attention to Hedy Lamarr.