These next-generation technologies may seem functionally similar, but they diverge when it comes to the markets they serve.
The development of Long-Term Evolution (LTE) and WiMAX has become, well, long term. Both technologies use advanced methods like orthogonal frequency-division multiple access (OFDMA) and multiple-input multiple-output (MIMO) (see “\\[\\[All-Hail-OFDM21880|All Hail OFDM\\]\\]”). They’re also fully IP-based (Internet Protocol), offering high-speed data capability to deliver fast Internet access and advanced applications like video.
But do these standards really represent the fourth generation (4G) of wireless technology? The International Telecommunications Union (ITU) defines LTE and WiMAX as 3G. According to some experts, both are almost 4G, more like 3.9G. Perhaps 4G still awaits us in the future.
While these two systems may seem to compete, with one eventually winning at the expense of the other, that’s not always the case. Competition exists for data services in some geographic areas, but it’s not all-encompassing.
Both developed from different backgrounds and have now essentially found their separate places. LTE is the clear-cut cellphone successor to the UMTS/WCDMA/HSPA and cdma2000 3G technologies, while WiMAX is finding use in broadband wireless connectivity and back haul.
LTE is the Third Generation Partnership Project (3GPP) name for the worldwide 4G cell-phone standard. It’s the planned and agreed-upon successor to the current 3G technologies.
The technology evolved from the original GSM voice technology to GPRS and EDGE for data to the current UMTS WCDMA and HSPA advanced 3G methods. Most of the standard is complete at this point, but is not yet finalized. It’s currently working through the tedious steps of the ITU standardization process, and final completion is expected later this year.
Most cellular operators have agreed to adopt LTE as the 4G standard, getting almost everyone around the world on board. That includes carriers like Verizon and Sprint in the U.S., who use cdma2000, which differs from the UMTS WCDMA standard. China, though, wants an LTE variation based on time-division duplexing (TDD) rather than frequency-division duplexing (FDD). More than 30 operators have agreed to adopt LTE and plan to integrate it into their systems in the coming years.
Right now, the only operational LTE systems are trial test systems in various parts of the world. We will see some formal LTE in 2010, but most carriers suggest deployment in scale beginning in 2011 or 2012 and beyond.
The still new 3G systems are working well and are still rolling out. You can’t blame the operators, who want to get a little mileage and profit out of their most recent 3G investments before implementing a newer and even more complex and expensive technology.
LTE is the inevitable choice. Its greater subscriber capacity and higher data rates support current and forthcoming services such as video, as well as data-intensive services like Internet access on the handset.
It’s designed for mobile operations with downlink (DL) data rates as high as 100 Mbits/s and uplink (UL) rates of 50 Mbits/s. Compare that to the current maximum 14-Mbit/s DL and 5.7-Mbit/s UL for most HSPA 3G services. Evolved HSPA standards define maximum DL and UL rates of 84 Mbits/s and 22 Mbits/s, but such systems aren’t available. With LTE, these high rates enable fantastic access to video, Internet services, games, and other high-intensity applications.
Current WCDMA and HSPA channels use 5 MHz. LTE is designed to fit into different bandwidths, including 1.4, 3, 5, 10, 15, and 20 MHz. Each bandwidth uses 128, 256, 512, 1024, 1536, and 2048 fast Fourier transform (FFT) subchannels, respectively.
LTE can function on virtually any current cellular frequency assignment that has enough bandwidth. Not all frequencies can accommodate the wider bandwidth modes. In the U.S., look for LTE in the 2.1-GHz bands and in 700-MHz assignments.
LTE is strictly FDD, where the send and receive bands are separate from one another. The downlink and uplink frequency separation varies widely from about 12 to 18 MHz on the lowest bands and as much as 340 to 560 MHz on the higher frequencies. The highest data rates need wide bandwidths, and they use lots of spectrum space.
Quadrature phase-shift keying (QPSK), 16-phase quadrature amplitude modulation (16QAM), and 64-state QAM (64QAM) are the modulation methods, depending on bandwidth availability and data-rate needs. Different levels of MIMO are supported, including 4x4, 2x2, 2x1, and 1x1 for the downlink. Typical cell capacity is in the 200-subscriber range with a bandwidth of 5 MHz. Wider bandwidths can increase this to greater than 400 users per cell site. Spectral efficiency is in the 5-bit/Hz range, but that depends on the physical-layer (PHY) details.
Continue to page 2
Incidentally, the OFDMA is used only in the downlink. The uplink radio technology is single-carrier frequency-division multiple access (SC-FDMA). This method was selected mainly to reduce the power consumption in the handsets. The peak to average power ratio (PAPR) is lower than OFDM, but that also means simpler RF power amplifiers can be used. MIMO support is provided for 1x2 and 1x1.
While LTE is basically a cell-phone technology, it also will be employed in high-speed broadband wireless connections. Expect to see it in USB dongles and data cards, as well as embedded in laptops and netbooks, where it will compete directly with WiMAX if available. LTE femtocells will ultimately bring that highspeed service directly into homes.
While LTE has yet to be deployed, it’s on the threshold. It appears initial LTE systems will arrive in 2010. But even with no LTE infrastructure, the ITU and 3GPP are working on the true 4G technology. Called IMT (International Mobile Telecommunications) Advanced, it defines a more aggressive version of LTE.
Specifically, IMT Advanced or LTE Advanced seeks mobile data rates in excess of 100 Mbits/s and a fixed or nomadic rate of 1 Gbit/s. Apparently, those specifications will be met with more bandwidth (>20 MHz), larger MIMO formats, and other methods. Nonetheless, it will be years before anyone sees that standard or its deployment. So, let’s get on with the LTE rollout (see “\\[\\[Chip-Makers-Target-LTE21878|Chip Makers Target LTE\\]\\]”).
LTE DEPLOYMENT ISSUES
Besides the overall cost of adopting a whole new technology for their cellular system, operators face other key issues. They involve the difference between switched and packet-based systems, the need for more basestations, the role of voice services, and backhaul deficiencies. All current cellular systems are circuit-switched systems, while LTE is packet-based.
This means carriers will need to keep their circuit-switched systems and equipment in place to maintain service for older and current customers. But at the same time, the new packet-based system must be implemented and made compatible with all existing systems. The older systems will be phased out eventually, leaving the simpler packet-only system.
Also, LTE initially will use the higher cellular frequencies around 2.1 and 2.6 GHz in the U.S. At such frequencies, cell site range will be much less than the range of a typical 800- or 900-MHz basestation. Therefore, more basestations will be needed, significantly increasing costs. One hope lies in the use of the new 700-MHz band assignments owned by some carriers, which will provide coverage that’s comparable to, or even slightly better than, current basestations.
LTE is a packet data system that isn’t necessarily designed for voice service. Not surprisingly, then, one issue facing carriers is determining how to implement voice over LTE. Three systems are being considered.
One is a circuit-switched (CS) fallback that lets the operator use existing circuit-switched 3G technology to handle the LTE voice. Another is the IP Multimedia Subsystem (IMS). In this case, the network can handle any service. 3GPP favors IMS, but few carriers use it. The third system is Voice over LTE Generic Access (VoLGA), which tunnels circuit-switched voice over the LTE network.
It would be nice to have one standard. However, each carrier will choose the method that makes the most out of existing systems until a formal standard is agreed upon. Until then, the lack of a single standard will affect nationwide and worldwide roaming.
Finally, the backhaul for most cellular networks is inadequate, even for heavy 3G services usage, and it could really crater with full-blown LTE deployment. This means phasing out those millions of T1 and T3 lines and adding faster microwave backhaul or even fiber where it can be accommodated.
WiMAX IS HERE NOW
WiMAX, short for Worldwide Interoperability for Microwave Access, had a major headstart over LTE. The first IEEE 802.16 WiMAX standard was approved in 2003. The 802.16d standard, what’s now called fixed WiMAX, came in 2004, and the 802.16e mobile WiMAX version was approved in 2006. Over the years, it has undergone slow but steady development and adoption.
Developed as a broadband wireless access (BWA) technology to replace or compete with DSL and cable TV for Internet access, WiMAX has seen little success in the U.S., but seems to be finally coming into its own. Its greatest achievements have occurred in developing countries where installing new wired telecom systems is impossible or impractical.
Many BWA WiMAX installations in India, South America, and elsewhere are used for telephone and Internet service and backhaul. Korea also uses a close cousin to 802.16e WiMAX called WiBro for BWA. Japan’s equivalent is known as XGP.
Sprint and Clearwire are looking to jumpstart adoption in the U.S. by rolling out WiMAX with BWA services, especially in smaller cities. Clearwire’s Clear service now serves more than four dozen cities in 16 states, with more on the way. The basic service costs $30 for data rates of 1.5 Mbits/s downlink. Voice over Internet Protocol (VoIP) service also comes with the data modem for $50 per month. Data cards for laptops are available, too.
Continue to page 3
Expect further WiMAX growth as the government pushes stimulus money into broadband access for the more remote, underserved areas in the U.S. Cable TV companies such as Comcast and Time Warner are also adopting WiMAX for BWA services in some parts of the country.
The U.S. uses the 2.3- and 2.5-GHz bands for WiMAX, and Korea uses 2.7 GHz. The rest of the world utilizes the 3.5-GHz bands, though the U.S. has an assignment in the 3.65-GHz band, and assignments in the newly opened 700-MHz spectrum are available. While most frequency assignments fall in the 2- to 11-GHz range, WiMAX has assignments in the 11- to 66-GHz range as well. These see limited use, though, and it will likely remain that way.
WiMAX uses OFDM and OFDMA with FFT sizes of 256 and 1024 subchannels on the 802.16d fixed version. For scaling in the 802.16 mobile version, it uses those sizes plus 128-, 512-, and 2048-subchannel versions. Typical bandwidth sizes are 3.5, 5, 7, 10, and 20 MHz for 802.16d, and 5, 8.75, 10, and 20 MHz for 802.16e.
Modulation methods include binary phase-shift keying (BPSK), QPSK, 16QAM, and 64QAM. Modulation is adaptive to the range and other environmental conditions. WiMAX also incorporates the MIMO capability of the 2x2 Tx/Rx format. Total cell capacity is in the 100- to 200-subscriber range. Typical spectral efficiency is 3.75 bits/Hz.
WiMAX is primarily a data service. Implementation of VoIP over WiMAX is occurring, but it’s more of a fixed service than an option for mobile handsets. The best way to think of WiMAX is as a super-long-range version of Wi-Fi. It’s not as fast, but it has a range of many miles. You will see WiMAX USB dongles, data cards, and even embedded WiMAX in laptops and netbooks. Development of WiMAX femtocells sits on the horizon, too.
Data rates depend on the service you buy, the bandwidth, the modulation, and other factors, but they typically run from about 1 to 2 Mbits/s in common consumer installations. Maximum rate is about 75 Mbits/s under maximum bandwidth and other conditions. Range extends from one to five miles depending on basestation placement and numbers.
For backhaul, maximum range is about 30 miles under ideal conditions. As for duplexing, either FDD or TDD can be used, but most WiMAX is TDD. Finally, the IEEE is currently working on a newer version called 802.16m that promises data rates from 100 Mbits/s to 1 Gbit/s. This is a good matchup with the ITU’s IMT Advanced requirements. WiMAX hopes to be part of that definition.
PRODUCTS OF NOTE
Analog Devices offers a line of RF products that can be used to implement LTE or WiMAX. Its AD935x transceivers are designed for both WiMAX and LTE customer premise equipment (CPE) and terminal implementations, as well as small basestations like femtocells and picocells.
The newest versions from the company, such as the AD9356 and AD9357, operate from 2.3 to 2.7 GHz and 3.3 to 3.8 GHz, respectively. Both feature dual transmit and receive chains (for MIMO) and include on-chip 12-bit analog-to-digital converters (ADCs), digital-to-analog converters (DACs), and frequency synthesizers. The ADRF670x and ADRF660x also address this space (Fig. 1).
The ADRF670x is an analog I/Q modulator with an RF output switch and phase-locked loop (PLL) with an integrated voltagecontrolled oscillator (VCO). Featuring a bandwidth to 500 MHz, it’s designed for use in IF upconversion transmit signal paths. It has high linear output power as well. The ADRF660x is a line of active RF mixers for receiver path downconversion. They include an RF input balun for single-ended 50-O input and a PLL synthesizer with integrated VCO. The differential IF output supports frequencies to 500 MHz.
The PC9608/9 LTE development system from picoChip targets small-form-factor LTE basestations optimized for metro, enterprise, and residential applications. It promises to expedite time-tomarket for basestation designs.
The system, which integrates baseband, software stack, and RF, allows designers and carriers to test the deployment of new network architectures. Many analysts and operators believe that LTE will require a dense network of small cells optimized for high-capacity data services to deliver on its full potential.
Most LTE implementations will be with standard macro basestations. However, operators realize that the higher-frequency assignments of most LTE will require more smaller basestations. That includes home femtocells.
The PC9068/9 incorporates a carrier-class software-defined LTE modem that complies with 3GPP standards and can be rapidly optimized and customized via an application programming interface. The system uses picoChip’s PC203 multicore picoArray silicon and supports the TDD and FDD versions of LTE (Fig. 2).
Wavesat’s Odyssey 9000 targets handsets, dongles, data cards, and other mobile wireless devices. The integrated systemon- a-chip (SoC) offers a fully programmable architecture that can easily adapt to LTE’s evolving standards. It also can easily implement WiMAX and the Japanese XGP equivalent. The hybrid chip combines highly efficient DSPs and hardware acceleration blocks that easily support the LTE 100-Mbit/s downlink and later 150-Mbit/s versions.
The first chip version, the OD9010, comes with a complete LTE protocol stack, including MAC, PDCP, RRC, and NAS layers. A reference design is also available. The OD9050 will offer 3G and LTE within the same chipset.
“We are working now with a few lead operators and OEMs to have fully interoperable LTE devices in place to meet the demand as LTE networks go operational worldwide beginning next year,” says Rah Singh, president and CEO of Wavesat.
Agilent Technologies recently published a great new textbook on LTE called LTE and the Evolution to 4G Wireless, Design and Measurement Challenges. It was written by a group of 30 Agilent engineers and scientists. This up-to-date book addresses almost everything you need to know about LTE. Go to www.agilent.com to get a copy.