Next for the Cellular Phone

By Anthony Ramirez
The New York Times

March 15, 1992

Nature created an abundant electromagnetic spectrum, which ingenious scientists have exploited for radar, television and lasers. But that is small comfort to cellular telephone operators. The Federal Communications Commission has been stingy with the radio frequencies that make wireless telephone calls possible.

But through clever design that largely prevents any two users from electronically bumping into each other, cellular telephones and the networks that serve them have coped with spectrum scarcity -- up to a point. That point is rapidly approaching in America's largest cities, where more and more miffed cellular users find their calls garbled, or blocked entirely, when a cellular channel fails to open up.

To increase capacity as well as to provide for improved voice quality, privacy and longer battery life, cellular operators plan to convert to digital transmission in Los Angeles, New York, Chicago and other large cities this autumn. The conversion is not without controversy within the cellular industry, as competing methods of digital transmission vie for the lucrative digital equipment market, certain to be worth hundreds of millions of dollars over the next few years.

One digital method, known as time division multiple access, is available now and can increase network capacity threefold. A method of staggering phone calls in sequence, it has been adopted by the industry, at least for now, and will be used for digital conversion later this year.

Another method, known as code division multiple access, is a year or more away, but might increase capacity fifteen- to twentyfold. A method of uniquely identifying information, it is the subject of furious technical debate over capacity and compatibility.

Still a third method, known as extended time division multiple access, is also a year or more away. This method, a variation of time division, might raise capacity eighteenfold, but without the suspected flaws of the code division method.

These methods imply that a large cellular system like that of the Los Angeles metropolitan area, with a theoretical capacity today of about 700,000 users, might eventually have enough capacity for as many as 14 million cellular users. That is almost two phones for every man, woman and child in greater Los Angeles -- without an increase in the amount of radio spectrum.

Such astounding increases in capacity come from digital coding. In analog transmission, electrical impulses carried by the telephone vibrate in the same way as the human voice disturbs the air. Like a mime imitating a man climbing uphill, the action is similar to, or analogous to, the original action. But much like a bad mime, an analog signal can perform poorly. The result is static, brought about by electromagnetic disturbances from such sources as a vacuum cleaner or a fluorescent light.

In digital transmission, a conversation is converted into the ones and zeros of computer code. Like painting by the numbers, the process is a bit soulless. But as long as the format is followed faithfully, reproduction is truer and static free, as in compact disks.

Moreover, analog signals can easily be heard with radio scanners by hackers bent on mischief, say, or criminals bent on fraud. With digital transmission, eavesdroppers hear only a loud and unintelligible hiss, a "bit stream" of computer code.

In every cellular market, there are two cellular franchises. Each operator, whether an independent company or the local-telephone provider, gets 25 million hertz of spectrum, or enough for about 832 channels of 30,000 hertz apiece. (Hertz is a measure of radio frequency.)

In the honeycomb cell structure of the network, which gives the cellular telephone its name, two users sharing the same channel out of the 832 channels per cell is statistically rare.

Frequencies are shuffled and reused so often and the cells are so physically distant that the user overlap is substantially less than 1 percent. But the larger the number of users, the more likely such sharing will occur. It is called "co-channel interference," or cross talk, in which callers are unwillingly thrust onto the same wavelength and thus overlap their conversations.

Dropped calls, when cell transmitters fail to "hand off" to nearby cell transmitters, also increase in number and so do blocked calls, when channels do not become available.

Time division multiple access reduces those problems by tripling capacity. It was first developed in the 1960's for use in satellite-to-earth transmission. As was the case with cellular telephones, analog signals were tried first, but because satellites and rockets were expensive to build and launch, time division was invented to increase capacity.

In time division, the cellular network puts three callers on a channel instead of one. Like a barbershop trio singing the opening notes of "Mr. Sandman" in sequence , three callers carry on three conversations in separate "time slots," with the network precisely timing the signals. Since they are all on the same channel, but not talking at the same time, there is no annoying electronic cross-talk.

All leading makers of telecommunications equipment, including Ericsson-G.E. and A.T.& T, are preparing time division multiple access telephones. Digital dual-mode phones are expected to be equal to or slightly higher in price than analog cellular phones.

In extended time division multiple access, now under development by Hughes Network Systems Inc. in Germantown, Md., and by others, the network accommodates either six or 18 callers, using different channels at precise time intervals.

Extended time division takes advantage of the fact that human conversation is actually silence half or more of the time, as even the fastest speaker pauses between syllables and leaves breathing space at the end of sentences. In between pauses, extended time division inserts other conversations.

By reducing the rate at which information is coded into computer ones and zeros, extended time division exploits the pauses as well as creating more time slots, thus allowing more conversations. Moreover, like one megaphone being passed around in what is now a barbershop choir of 18, the radio channel is used only when there is information on it, further enhancing efficiency.

Code division multiple access, under development by Qualcomm Inc. in San Diego, works on entirely different principles. The method takes the same 25 million hertz of spectrum, but divides it into only 20 channels of 1.25 million hertz apiece.

When code division digitizes human conversation, it attaches unique identifiers to each information burst and uses the identifiers to reassemble the conversation.

Code division is like mingling at a United Nations cocktail party and finding the one English speaker. The unique identifiers of English allow one conversation among many to be singled out.

The reason that code division is hotly debated is that its signal structure is so different from that of time division. Code division's channel is more than 40 times larger than that of time division.

In order not to anger millions of cellular subscribers by making their analog cellular phones obsolete, cellular operators want to sell phones that can be used in analog or digital mode. Time division will work on phones that receive analog signals, but it is unclear whether code division will allow such dual-mode phones to work efficiently. More field tests are necessary, engineers say.

Such arcane technical controversies aside, everyone in the industry agrees that digital will help solve one of its most annoying problems -- brief talk time. The smallest cellular phones are about the size of a package of Twinkies but consume so much power that a subscriber can talk for less than 45 minutes before recharging. With digital, even the tiniest phones consume from one-third to one-twentieth less power for transmission purposes.

That means that even on a Twinkie phone, a cellular caller might someday gab continuously for half a day, pausing only for breath.

Copyright 1992