As set forth by the International Telecommunications Union's (ITU's) IMT2000 broad set of guidelines for third-generation (3G) Personal Communication Services (PCS) and cell phones, current standards and systems must either evolve, or change completely, to satisfy the new 3G objectives. After nearly 15 years of experience, cell-phone manufacturers and telephony infrastructure companies all have ideas about what the next generation should be like. As we all know, though, more than one path will take us there. That makes the challenges for design engineers particularly exciting or excruciatingly difficult depending upon their point of view.
Just what exactly is a 3G phone? Lest we forget, cell phones have undergone a natural evolutionary process which so far has produced two generations of increasingly capable and complex systems. Today's second-generation (2G) digital phones are already being upgraded to a mid-generational phase known as 2.5G phones (Fig. 1). These new phones implement many features which are objectives for 3G phones. So while we aren't yet at the third generation, we're getting closer through an evolutionary process
Market research studies predict as many as 500 million cellular subscribers by the end of 2001. At that point, the total number of cellular subscribers will actually exceed the total number of hard-line-based subscribers. Many forecast a growth to over one billion subscribers by 2003 or 2005 at the latest. This means steady growth of the cellular business and lots of opportunities for continued product design.
The table summarizes the features and technical specifications of all cell phones to date, from the original Advanced Mobile Telephone Service (AMPS) analog phones to the latest iteration 2.5G phones. Although all of these phones support a common standard, 3G phones could potentially have a common standard. But, that's a real long shot. While efforts are under way to resolve the various conflicts among the different parties, it's expected that 3G phones will support multiple new standards as well as continue supporting those from previous generations, including analog standards.
The ITU didn't define specific interfaces or technical details for 3G phones. Instead, the union put forward a broad set of guidelines stating desirable features and capabilities that virtually everyone agrees with. Where the disagreement lies is in how to achieve them. As usual, the devil is in the details when it comes to addressing the design objectives. But isn't that what engineers are for?
One design issue is global roaming. This is a good objective as it would permit any new 3G phone to work essentially anywhere in the world. As you deplane from your New York to London flight, you can flip open your phone and make calls immediately.
Neat. Yet while this seems to be a great objective, it probably isn't the most important in the whole scheme of things. Recent surveys of cell-phone users have shown that over 80% couldn't care less about global roaming. It's probably more important for Europeans than for U.S. citizens who really only need U.S. roaming.
The IMT2000 guidelines set the objectives of 144 kbits/s as the minimal data rate achievable from a fast-moving vehicle. A data rate of 384 kbits/s should be possible with pedestrian speed. Furthermore, a 2-Mbit/s rate should be possible from a fixed location. Such high data rates make it possible to achieve all of the desired data and multimedia features suggested by the guidelines.
According to Scott A. Schmok, strategic marketing manager of Motorola's Wireless Infrastructure Systems Division, one really exciting outcome of such high data rates is that 3G phones will have capabilities far greater than fixed hard-line phones, which use the Public Switched Telephone Network (PSTN) and current access methods. These include the Integrated Services Digital Network (ISDN), cable modems, and xDSL lines. It's anyone's guess just what the full implications of this will be.
Already, 2.5G cellular systems are addressing the e-mail and Internet access issues. Results are mediocre at best because of the data-rate limitations of current 2G systems that provide data transfer rates of 9.6 or 14.4 kbits/s. Third-generation systems will provide much higher data rates, making the performance in e-mail and Internet access far more acceptable.
In the meantime, 2.5G enhancements, such as Enhanced Data Rates for GSM Evolution (EDGE), the General Packet Radio Service (GPRS) system, and the High-Speed Circuit Switched Data systems, are expected to provide 3G high-speed data access for the GSM and IS-136 TDMA systems during the late 2.5G period. I might check my e-mail by cell phone, but utilizing it to access the web on a tiny monochromatic screen seems ridiculous. Still, I can picture myself using a laptop computer that's plugged into an Internet-enabled 3G phone. Multimedia applications will be easier and more acceptable when the operating systems and better (larger/color) screens become available at a reasonable cost.
Once high data rates are achieved, many additional multimedia applications become practical. Navigation and location services are an example. The U.S. Federal Communications Commission's (FCC's) mandate for installing an emergency 911 number on telephones is expected to be implemented before the third generation is reached. Known as Enhanced 911 service (E911), the FCC expects 60% of the new cell phones to have E911 capability by Oct. 1, 2001. With the expected higher data rates, even more sophisticated navigation and location services can be employed.
Video is another potential 3G cell-phone application. Along with the higher data rates and new compression algorithms, such as MPEG-4, two-way video becomes a possibility.
No Common Air Interface
One thing that the ITU didn't specify was a single air or radio interface standard, as if that could really happen, especially in the U.S. Again, it would have been a good objective, but probably it's impractical if not altogether impossible in the U.S. where multiple standards happily coexist and backwards compatibility will become an issue for the extended future. As Findley J. Shearer, technical marketer of Motorola, put it, "It's an unstable time when the standards are in flux." Multiple air interfaces will continue to exist. But, progress will be made. Eventually, the total number of standards will be pared down to only a few.
A single air interface is probably a better goal for the fourth generation (Fig. 1, again). Generally, it's agreed that the most desirable air interface standard for 3G phones is some form of wideband code-division, multiple-access (W-CDMA) transmission. The suggested channel bandwidth of 5 MHz with a direct-sequence spread-spectrum (DSSS) chip rate of 3.84 Mbits/s permits the higher data rates demanded by the more advanced services that are proposed for 3G systems. Wider bandwidths of 10, 15, or 20 MHz are expected to also be available. These will further improve the data-rate potential.
One proposal for such a W-CDMA system was originally developed by the Japanese standards organization, the Association of Radio Industries and Business (ARIB). The European Technical Standards Institute (ETSI) proposed a similar W-CDMA method. Both standards are considered as one despite minor differences.
Another competing standard is CDMAOne, which includes backward compatibility with the 2G and 2.5 IS-95 standards. Proponents of the CDMAOne system developed a revised and upgraded version of W-CDMA known as CDMA 2000.
Unfortunately, no formal agreement has been reached on a single W-CDMA standard. Because of the conflict between the competing W-CDMA proposals over duplex methods and other issues, an Operator's Harmonization Group (OMG) was formed in mid-1999 to work out the differences. The 3GPP (ETSI W-CDMA) and 3GPP2 (cdma2000 W-CDMA) groups are working toward an agreeable consensus.
Then there's the IS-136/UWC-136 path, the highly refined time-division multiple-access (TDMA) standard so widely used in the U.S. The original IS-54 standard has been greatly improved to form IS-136, which many believe is still the way to go for 3G phones. Like GSM, IS-136 phones will be enhanced with a series of progressively more capable data-handling methods. The GPRS standard steals extra time slots in the TDMA frame to increase the data throughput rate. The target rate is 115 kbits/s, but the potential is as high as 160 kbits/s. GPRS systems are expected to be in operation by late 2000 or early 2001.
The EDGE standard is a newer technology that will offer faster data handling in GSM and IS-136 phones. It too steals time slots in the TDMA frame to boost the data rate. Yet it also uses 8PSK modulation that kicks the speed up to 384 kbits/s. It's anticipated that some form of W-CDMA will predominate in 3G phones, but UWC-136 will also be widely supported because of the enormous infrastructure supporting it.
Cellular operators and equipment manufacturers will continue to face a series of critical issues on the path to 3G systems. These include the frequency spectrum, technical standards, interoperability, and infrastructure costs. As usual, resolution to these critical issues will be dealt with by the standards organizations, advocacy associations, and the various working groups of operators and manufacturers.
The electromagnetic frequency spectrum, like other natural resources, is finite in scope. That's why it's regulated by governments worldwide. While cellular phone services have assigned spectrum blocks throughout the world, there's constant pressure to increase the amount of spectrum available for cellular services. Greater bandwidth has been the general solution to adding subscribers and achieving the higher data rates demanded by advanced services. The FCC's allocation of the 1900-MHz bands for PCS was a step in the right direction. It has allowed for the accommodation of many more subscribers as well as permitted higher data rates.
One of the big issues at May's World Radio Conference (WRC2000) in Istanbul, Turkey, was increased spectrum space for cellular services. The allocations proposed at WRC1992 in the 2.1- to 2.3-GHz range were reblessed. Three 160-MHz bands were set aside especially for IMT-2000 service.
As we mentioned earlier, market researchers project a total of over one billion mobile subscribers by 2003 or 2005, depending upon whose figures you believe. This will place continued pressure on the FCC and other regulatory bodies to find new usable bands.
An example of possible future space for cellular services is the 700-MHz range (747 to 762 and 777 to 792 MHz), previously occupied by UHF TV channels 60 through 69. Some say this spectrum isn't suitable for 3G applications. But, the FCC will soon auction it off to the highest bidders. The planned Sept. 6, 2000 auction was postponed. Check the FCC's web site for more information. Most likely, the 2.1- to 2.5-GHz bands will be used for most future cellular applications.
Technical standards will always be a critical issue as operators and manufacturers continue to discover and invent new technical interfaces and debate their merits. Because the performance differences between the various air interfaces appear to be miniscule, any one of a number of standards will produce comparable operation to the others. So while one standard doesn't appear to be technically superior to the others, multiple standards will continue to coexist.
Currently, the greatest number of operators and equipment manufacturers believe that some form of W-CDMA offers the best potential for achieving 3G goals. On the other hand, excellent standards, like IS-136, GSM, and their various 2.5G enhancements, will continue to provide exceptional service to their subscribers.
Cellular operators have an enormous investment in equipment, systems, and services. These won't go away overnight. Plus, new systems will come online only when it makes sense profit-wise.
Interoperability, or the ability of a cell phone to access services at any time in all locations, will continue to be a critical issue facing all parties. Because of the many existing standards and new emerging standards, interoperability becomes "the" elusive objective.
Currently, such problems are being solved by multimode/multifrequency cell phones and basestations. Support for multiple standards will continue to be the norm as we work toward the 3G objective of global roaming.
Infrastructure Must Come First
Before advanced 3G cell phones can be used, the supporting infrastructure must be in place. New air interfaces and network structures require an enormous investment from the operators. Most operators must continue to support existing first- and second-generation systems because of the huge installed base of subscribers. Third-generation systems will require an increased investment. As long as the number of subscribers continues to grow, such investments will make sense on a business basis.
Yet, operators will search for new ways to expand or modify existing systems so that the 2.5G and 3G objectives will be met. Newer basestations for this infrastructure will be made more modular, expandable, and software-programmable in order to permit rap-id changes, additions, and upgrades.
Engineers designing for 3G handsets and basestations face a variety of development challenges from the design complexity to controlling costs, meeting time-to-market goals, supporting multiple standards and bands, and hardware/software tradeoffs. Other challenges include functionality, packaging, power limitations, testing, infrastructure issues, component shortages, and software development.
Because of the increased functionality and new standards, 3G cell-phone equipment will be far more complex. The 2.5G and 3G phones must support multiple operational modes and frequency bands. While the all-digital design permits more of the product to be put into software, the complexity level will go far beyond that of older first- and second-generation equipment.
Engineers designing 3G cell phones and basestations will be constantly fighting to keep the costs at or below current levels. The more complex designs will require larger, more complex chips and other components, thereby raising costs. While the expected high volume will help, increased software development costs will still force costs upwards.
The PC industry has spoiled consumers with the reality of increasing speed and functionality at the same or lower cost. Many consumers will expect to be given their new 3G cell phones for free with a fixed contract for services, or else to purchase them at very low prices. Some will gladly pay more for the new multifunction high-speed phones, but others will resist. Then again, as Mark Williams, RF applications engineering manager at Motorola, points out, handset cost has a small effect in the grand scheme of things. The large costs come from the infrastructure. Handset costs are buried in the consumer charges for services in most cases.
Aside from controlling costs, a designer is going to have to complete his or her design quicker in order to beat the competition. This is a problem for most electronic engineers these days. The cell-phone industry, however, is larger and more competitive than most. Those who reach the marketplace first typically reap the benefits of new buyers and increased market share.
More CAD Would Help
One way to meet this challenge is to implement more CAD methods. EDA software suppliers are beginning to offer more products to simulate and design RF circuits and systems.
Cell-phone manufacturers will continue to build phones and basestations that support multiple standards. Older first-generation analog phones and newer 2G and 2.5G phones will continue to coexist with 3G phones.
So that wide roaming capability can be supported and even expanded, multimode, multiband phones will be necessary. This means that most new phones will operate in two, three, or even four modes and frequency bands.
Already, such phones are widely used. Typical combinations include analog/CDMA, analog/TDMA, GSM/CDMA, and analog/TDMA/GSM. In the U.S., multiple frequency bands are available. In addition to the widely used 800-MHz U.S. band, many new cell-phone systems employ the newer 1900-Hz PCS band. These phones typically use the GSM or U.S. IS-136 TDMA standards, although some CDMA operation is available.
Supporting multiple-frequency bands requires dual, triple, or wideband RF front ends. Supporting multiple air interface standards requires more complex vocoders and DSP chips for baseband voice and data modulation and recovery.
The challenge is to use more software and less hardware in cellular/PCS designs. While the transmitter and receiver RF sections will continue to be hardware-only, more and more software will be involved in implementing the voice, data, and networking functions. This trend will continue as designs move increasingly closer to the ideal software radio.
One major development is the trend toward greater use of direct-conversion transreceivers. Direct-conversion or zero-IF (ZIF) receivers set the local oscillator frequency to the frequency of the incoming signal, creating a difference IF of zero (Fig. 2). In other words, using a single downconversion, the original baseband information is recovered. Such a technique eliminates not only the IF amplifiers themselves, but also the expensive and space-consuming IF filters. Filtering at baseband is simpler as low-pass rather than bandpass filters can be used.
A critical software issue is the browser for Internet access. The Wireless Applications Protocol (WAP) microbrowser will be used in late 2.5G and most 3G designs until someone comes up with a better design for this application. While it could be a modified version of Microsoft's Windows CE, it probably will be something new that's designed especially for this market.
Adding the new features sought by the 3G/IMT 2000 objectives will make cell phones more than just phones. The devices will become the ultimate mobile information appliance. Users are already beginning to enjoy the paging/messaging services currently provided by many operators, and many are starting to use their cell phones for e-mail and Internet access. Data rates are limited, but now steps are being taken to improve the 2.5G phones with higher data rates by employing GPRS, HSCSD, and EDGE.
The greatest challenge of all, though, will be adding the 3G features. It goes without saying that faster e-mail and Internet access will make 3G phones far more acceptable. But the high-speed data rates also will permit location services, like E911 and even video. Such systems are even more challenging due to their complexity.
The Packaging Challenge
Since the World War II walkie-talkie configurations, cell-phone handsets have shrunk down to the size of a deck of cards. Moving from analog to an all-digital design has facilitated amazing size decreases. Still, the complexity of the 3G phones could cause the more full-featured phones to grow in size. Phones supporting video and location services are an example of phones that would, by necessity, have to increase in size because of the keyboards, CCD video camera, color LCD screens, and attendant battery requirements.
As a result, packaging will continue to be one of the greatest challenges facing cell-phone designers. The movement toward an all-software radio design will help, but supporting multiple standards and frequencies as well as adding new functions may necessitate larger physical configurations. In the meantime, IC manufacturers are squeezing more and more devices on their chips using systems-on-a-chip (SoC) approaches and smaller packages (see "New Products For 3G Phones," p. 110).
A related issue is ergonomics and safety. Many local and state governments are examining the possibility of banning cell-phone use while driving. There are just too many cell-phone related accidents. One possible solution is to package the cell phone as a belt clip-on with a headset and use voice-recognition dialing. The headset design separates the RF power output from the user's head and should also eliminate the complaints and unsubstantiated claims that cell phones cause brain cancer.
Two areas of packaging that will continue to provide challenges are pc boards and displays. Multilayer pc boards will be necessary, and their layout will continue to be a challenge because of RF requirements.
As for displays, the current monochromatic LCD is more than adequate for simple messaging, paging, e-mails, and limited Internet access. When increased data rates are achieved, however, greater use of Internet access will result. Users will undoubtedly demand larger higher-resolution screens. Even a color LCD screen isn't out of the question, although current active-matrix designs are inadequate for use in full sunlight. Unless you are ready for a laptop-size battery pack, they simply consume too much power for a portable device.
The new 3G designs will put an ever increasing load on the cell-phone power supply, namely the battery and the power-management circuits. High speed almost always involves higher power dissipation. Furthermore, high-end functions, such as larger and even color screens for Internet access and location services, will make demands higher than ever (see "Needed: Sophisticated Power Management," p. 116).
So while manufacturers have made excellent progress over the years in increasing standby time, talk time, and battery life, that trend might become more difficult to maintain with 3G systems. Certainly, increased power demands will largely be met by the on-going reduction in chip supply voltages as well as improved batteries. Also, newer, improved power-management chips will help ensure reasonable talk and standby times, plus battery life.
Another important issue is the availability of the proper test and measurement equipment, without which no 3G designs would be possible (see "Meeting The 3G Test Challenge," p. 118). Some new test-equipment companies have been founded on the premise of serving only the 3G market and beyond.
It appears that a great many of the challenges for 3G equipment design lie in pushing an increasing amount of functionality into at least one embedded processor. As analog-to-digital and digital-to-analog converters (ADCs and DACs), DSP chips, and communications processors become faster, more functions can be programmed in software. ASICs and FPGAs will still be required by some CDMA designs because of the need for speed.
By making the cell phone more programmable, many different features and functions can more easily be supported. On the other hand, that pushes more of the design into the software area, which continues to represent a challenge to many engineers. Improved development software for DSP and other chips will help ensure that new designs come to market on time. A highly desirable design objective is flexibility—that is, making the handset or the basestation upgradable through software changes. It's easy to imagine a cell-phone user dialing a number in order to receive a download of the latest embedded software or operating system versions.
It will be a while yet before the true software radio is realized. Faster ADCs and DSP chips will be required. Each year, however, we grow closer to the ideal as DSP moves closer to the antenna with each new generation.
So what do DSP chips actually do in a cell phone? They primarily handle baseband functions, which include control of the voice codecs, filtering, equalization, echo cancellation, error detector and correction, and encryption, as well as speech coding (compression) and decoding (decompression).
Voice recognition also will be implemented in DSPs. In many of the current 2G and 2.5G CDMA phones, DSP chips haven't proven to be fast enough to perform some of the synchronization and processing required by the spread-spectrum technique. Special ASIC chips have been created to deal with this problem. Some functions have been implemented with FPGAs too.
It isn't anticipated that the first full-3G phones and systems will be available until some time in 2001. The earliest full implementation will probably take place in Japan. In the meantime, operators and cellular-equipment manufacturers will continue to augment, enhance, and otherwise improve the current 2.5G phones. In any case, current phones and systems will gradually evolve into full-fledged 3G designs over the coming years.
It could be as late as 2005 before full-3G objectives are met. This gradual progress will be caused by the slower infrastructure growth due to the massive investment that's required. Most operator companies will delay the build-out.
Study groups are already beginning to think in terms of the next or fourth generation. What is a 4G cell phone and what does it feature? Well, we can certainly expect such phones to operate at even higher microwave frequencies than current models. Plus, even higher data rates will be possible. What we may be looking at is a color video phone with voice recognition operating in the 2-GHz+ range. Or, perhaps it will be just a headset with hearing-aid-size RF and processing circuitry.
|Cellular/PCS Companies And Organizations|
Spectrum and regulatory
fax (202) 293-9287 or
fax (212) 398-0023
+33 4 92 94 42 00
fax +33 4 93 65 47 16
+41 22 730 5111
fax +41 22 733 7276
fax (703) 907-7727 or
Project and Third Generation
Partnership Project 2
fax (714) 545-4601
fax (703) 799-5819
+353 1 269 5922
fax +353 1 269 5958 or
+44 207 518 0530
+44 207 518 0531
Agilent Technologies Inc.
fax (408) 654-8550
Analog Devices Inc.
fax (781) 937-1058
fax (408) 487-2600
fax (408) 988-0297
fax (949) 483-4078
Infineon Technologies Corp.
fax (408) 501-2424
fax (800) 943-0087
fax (858) 658-2100
fax (408) 452-1549
fax (732) 544-8347
fax (503) 627-5593
Useful web sites
White papers, news,
glossary, products, etc.
Digital wireless basics
Tutorials on spread spectrum
(CDMA), software radio, etc.
Tutorials and presentations
on TDMA and EDGE