To enable high-data-rate applications on wireless networks, microwave engineers are moving beyond traditional analog methods to embrace a variety of digital design approaches.
Right now, many families rely on wirelesslocal- area-networking (WLAN) access points to provide Internet capabilities for multiple computers inside their homes. In the near future, however, such devices will be replaced by cellular femtocells. These smaller base stations are expected to be widely used in homes and small businesses to provide access to fourth-generation (4G) communications services. To enable these higherdata- rate capabilities, however, the network infrastructure as a whole must operate with reduced amplifier distortion, increased power efficiency, and higher output power. RF and microwave designers are preparing to win this battle by leveraging the latest digital technologies ranging from advanced digital modulation schemes, multiplexing, baseband processing, and cuttingedge approaches to signal processing and linearization.
The move to digital-modulation formats and the resulting evolution of those formats has been a rather natural progression for wireless communications technologies. Compared to analog modulation schemes, digital approaches can convey larger amounts of information on limited bandwidths. This capability is a convenient solution to the problem of providing a growing number of users with an increasing number of services over still-limited RF spectrum. In addition, digital modulation is generally credited with advantages like improved security and greater resistance to interference and fading than analog approaches.
Femtocells are among the devices being enabled by newer digital-modulation techniques, thanks to firms like picoChip. According to Rupert Baines, the firm’s VP of Marketing, “The areas we are working on are HSPA femtocells (increasing efficiency by making base stations smaller and cheaper) and Long Term Evolution (LTE; dense modulation with 64 quadrature amplitude modulation , orthogonal frequency division multiple access , and multiple input multiple output ). The interesting point is that we can make base stations smaller and cheaper. That increases the signal-to-noise ratio (SNR) so you can actually achieve the higher rates.”
At Spain’s Mobile World Congress this past February, picoChip debuted two new picoXcell system-on-a-chip (SoC) devices. The PC313 and PC323 single-chip HSPA+ femtocell SoCs serve eight or 24 users, respectively. They can be scaled to accommodate even more users. The SoCs boast Third Generation Partnership Project (3GPP) Release 8 features like 42-Mb/s downlink and 11-Mb/s uplink data rates. They also meet that release’s demand for the inclusion of MIMO technology, which will enable the high data rates and receive diversity needed to achieve the higher real-world user counts required for enterprise and “greater femto” applications. To that end, the PC323 SoC has been designed to accommodate as many as 24 simultaneous users for enterprise, campus, rural, and metro femtocell usage scenarios with a cell radius to 2 km. Two cascaded devices can be used to power systems with 48 users.
An SoC from Percello Ltd. also promises to reduce HSPA+ femtocell cost and size. The single-chip PRC6100 maximizes digital integration by housing a high-speed MIPS24Kc processor and its peripherals, femtocell L1 engine (FLE), and embedded double-data-rate (DDR) memory. As a result, it does not require any external memory or digital parts on the board. The PRC6100 is compliant with 3GPP HNB specifications. It supports eight users and offers HSPA+ data rates of 21.6 Mb/s downstream and 5.76 Mb/s upstream.
The emergence of third-generation (3G) HSPA/HSPA+ and 4G Long Term Evolution (LTE) mobile data services has resulted in a sharp increase in bandwidth demand. This demand, in turn, has spurred the development and implementation of advanced modulation and multiplexing techniques in the microwave backhaul arena. According to Paolo Volpato, Product Strategy Manager at Alcatel-Lucent’s Wireless Transmission Product Unit, “The result of this focus has been the achievement of greater than 1-Gb/s capacity using just a single radio element. Specific techniques used to achieve this breakthrough include the application of high modulation schemes (up to 512 QAM), the exploitation of adaptive code modulation (ACM), and the use of advanced cross polar canceller systems (XPIC) for frequency reuse.”
He continues, “ACM, for example, makes it possible to more efficiently handle in real time dynamic capacity changes in the radio link—changes that can occur as a result of fading or variable weather conditions. In response to these variations, the modulation changes accordingly—ensuring that high-priority services are consistently given precedence over best-effort traffic. Applying ACM, however, poses a challenge: Congestion may occur in the uplink radio direction as a result of limitations in the current generation of nodal hybrid microwave devices that manage the grooming of traffic coming from diverse access sources.”
Volpato explains that these hybrid microwave devices handle time-division-multiplexed (TDM) and data traffic through two fixed portions of the same uplink radio channel. TDM traffic always gets a fixed amount of bandwidth, statically reserved. Because of this condition, adaptive modulation can only be applied to the data portion of the traffic in the current environment. This condition results from the fact that TDM service makes use of fixed slots that are consistently transmitted even when some of those slots are empty. When congestion occurs, all of the data traffic including voice-over-Internet-Protocol (VoIP) is impacted regardless of its relative priority.
The latest generation of microwave packet-radio devices strives to overcome this limitation by treating all types of traffic as packets that can be transmitted using a single Ethernet switch. They also employ statistical multiplexing on radio directions so that channel capacity can be more fully exploited. For example, Alcatel-Lucent’s 9500 microwave packet radio leverages adaptive modulation and supports deterministic behavior in terms of latency and jitter independently from the network load. Services are handled as a function of their priority. Native TDM voice, for example, is converted into packets and combined with VoIP so that a single voice service is handled by microwave-packet-radio devices.
Underlying these system solutions are advances in component design that also rely on digital techniques. To answer the call for reduced amplifier distortion, increased power efficiency, and higher output power, for example, Scintera recently teamed with Richardson Electronics to bring its SC1887 power amplifier (PA) to market (Fig. 1). This PA vows to deliver critically needed linearity improvements without requiring access to in-phase/quadrature (I/Q) baseband signals. The SC1887 Adaptive RF Power Amplifier Linearizer (RFPAL) vows to provide an adjacent-channel leakage ratio (ACLR) to 26 dB. The solution requires no analog-to-digital converters (ADCs) or digital-to-analog converters (DACs). With Scintera’s RF Power Amplifier Linearizer, the complex signal processing is done in the RF domain. This solution is applicable across a broad range of signals including 2G, 3G, 4G, and other modulation types, as its analog signal-processing engine is capable of linearizing highly efficient power-amplifier topologies.
According to Kris Rauch, Scintera’s Vice President of Sales and Marketing, “The IC provides a programmable analog signal processor under digital control by way of an on-chip microcontroller integrated with advanced analog circuitry. The solution can be reconfigured with software to support a wide array of modulation types and signal bandwidths over all operating conditions.”
Chris Marshall, VP of RF/Microwave Products at Richardson Electronics, notes, “The Scintera approach of analog predistortion is grabbing a lot of attention because it is easy to implement. And as an RF-to-RF solution, it can be applied to PA designs that do not have access to the baseband signal. The potential ACLR improvement of up to 26 dB is not as great as digital-predistortion (DPD) designs. But the minimal engineering cost and relatively low unit price make it attractive for smaller production runs and for repeaters and other lowerpower designs, where the DPD overhead becomes significant and the relatively low operating cost doesn’t justify much of a price increase.”
Conintue to next page.
Advanced digital algorithms like DPD and crest-factor reduction are becoming popular techniques to linearize PAs and improve transmitter efficiency from less than 10 to over 40 percent, notes Gina Colangelo, Segment Applications Engineer for Analog Devices’ Communications Infrastructure Segment Team. She emphasizes that DPD requires an observation receiver with a high-bandwidth ADC that downconverts a coupled version of the PA output. The digital version of the transmitted waveform is compared to the received waveform and adaptive algorithms compute/update a set of coefficients to pre-load the next transmitted waveform. As the adaptive algorithm converges, the PA is linearized and the output distortion is significantly reduced. The increase in speed and the move to finer-line processes has made DPD more scalable for multiple antenna transmit systems compared to other analog linearization methods, such as feed-forward linearization.
As mentioned, very-high-speed DACs also are playing a major role in today’s wireless infrastructure, as they are the chief signal generator. Colangelo notes, “They are also now performing many of the functions that were once handled by various additional circuit elements including the baseband processor. In addition, the increasing use of lowvoltage differential signaling (LVDS) for the digital interface allows the data rate to run at 1200 MSamples/s and higher while keeping power and power supply voltages low. High input data rates to the DAC help to increase the transmit path input bandwidth, allowing for higher-order DPD algorithms or wider correction bandwidths. LVDS also radiates less and provides better noise immunity and timing.”
The latest signal-processing DACs, such as ADI’s AD9122 TxDAC+, promise to work in a complex, intermediate- frequency (IF) direct-conversion architecture (Fig. 2). They provide the fully modulated IF I and Q signals that can be fed directly to the antenna through the analog quadrature modulator and PA without additional signal modulation.
By leveraging digital superconductor technology, Hypres has been able to develop ADCs that supposedly surpass current ADC offerings in terms of sampling speed, resolution, and bandwidth (Fig. 3). When applied in an RF or microwave transceiver, this digital superconductor technology allows a very wideband RF or microwave signal to be directly digitized right after the antenna. It does not require downconverters or multiple ADCs, filters, and amplifiers. Richard Hitt, CEO of Hypres, states, “Digitizing such large swaths of spectrum so accurately, quickly, and at the top (near the antenna) of the signal chain creates the path to a true ‘all-digital’ system—a digital baseband that now has a true digital front-end companion. And for RF and microwave engineers, there is no longer a need to devise workarounds to help boost or clean up a signal that has been degraded through its twists and turns in the predominately analog front-end of the signal chain as it heads toward the digital baseband processors.”
As these few examples show, the multitude of cutting-edge digital technologies is offering a seemingly limitless amount of options for next-generation infrastructure designers. As is common with cutting-edge, emerging technologies, some of them are already in competition. In addition, a number of them may not be deemed practical solutions for real-world problems because of expense, dependability, ruggedness, or performance issues. No matter which solutions win for LTE, however, one thing is certain: The future of wireless infrastructure will increasingly rely on digital techniques.