Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > RF/Microwave

Maximize throughput in HSDPA networks via receive diversity

Posted: 01 Mar 2007 ?? ?Print Version ?Bookmark and Share

Keywords:HSDPA W-CDMA Wi-Fi? HSUPA Sirific 802.11?

By Sathwant Dosanjh and Abdellatif Bellaouar
Sirific Wireless Ltd

Cellular network operators are investing heavily in next-generation infrastructures with the goal of enticing subscribers into a mobile multimedia environment. To capitalize on their investment, operators are counting on a significant boost in average revenue per user (ARPU) from increased access of audio-visual content through high-speed cellular networks. However, carriers know that they can't rely solely on their own high-speed infrastructures to ensure a satisfying wireless multimedia user experience. That's why they're driving OEMs to develop high-speed feature-rich handsets, with a push towards small form-factor designs.

Addressing the demand for higher throughput, the 3rd Generation Partnership Project (3GPP) introduced high-speed downlink packet access (HSDPA) in Release 5. Otherwise known as 3.5G cellular technology, HSDPA is an enhancement of wideband code division multiple access (W-CDMA). This new feature is backwards compatible with W-CDMA, requires no new spectrum to roll out in the network, and promises as much as a sevenfold increase in peak data rates in the downlink direction.

As most experienced wireless users know, data throughput achieved in the field usually falls far short of the advertised peak rates. This is even the case with HSDPA-enabled terminals. To realize HSDPA peak rates under actual radio conditions, these cellular devices must utilize more advanced radio solutions. In particular, the application of receive diversity can more than double the effective HSDPA throughput across the entire cell and significantly improve network efficiency and capacity. This performance improvement results in a more robust end-user experience while web browsing, downloading files, and streaming media.

Enabling wireless broadband
HSDPA applies a shared-channel transmission concept, which makes more efficient use of available channelization codes and power resources than a dedicated channelas used in standard W-CDMA networks. This shared channel is defined in W-CDMA 3GPP Release 5 as the high-speed downlink shared channel (HS-DSCH). The increased efficiency of code and power use in the HS-DSCH boosts cell capacity by more than twice that of a dedicated channel, thus paving the way for higher data rates.

HSDPA networks further increase system capacity by implementing short transmission time intervals (TTI) with fast scheduling techniques. The basestation uses these efficient scheduling methods to monitor the link quality to each mobile user on the network and prioritizes data traffic accordingly. As a result, network latency is reduced by a factor of up to 20.

HSDPA networks also dynamically and quickly adapt channel-coding and modulation depending on radio conditions. In clean radio environments, HSPDA systems use 16-bit quadrature amplitude modulation (16-QAM) to increase throughput. Under adverse radio conditions, HSDPA networks dynamically switch to quadrature phase shift keying (QPSK) modulation. Standard W-CDMA transmission, on the other hand, is limited to QPSK modulation only. W-CDMA networks must therefore employ power-control techniques to compensate for changes in radio conditions. Because HSDPA networks don't rely solely on power control to optimize gain, they're more power efficient than W-CDMA networks. This improvement in spectral efficiency can yield an increase in system capacity of up to 300 percent.

These performance enhancements enable HSDPA networks to achieve a significant throughput advantage over W-CDMA networks. In Category 10 operation, HSDPA systems can theoretically reach throughput levels up to 14.4Mbps. HSDPA data rates are characteristic of those achieved in IEEE 802.11 networks, which have proven sufficient to support such services as Web browsing, file server access, and even streaming audio and video (Figure 1).

Figure 1: Each new generation of technology brings higher throughputs and a new set of applications.

There's a price to be paid for higher data rates. The typical constellation for 16-QAM has a smaller symbol decision area compared to QPSK. As a result, the received 16-QAM signal is more susceptible to spectral impairments, which translates into a tighter error vector magnitude (EVM) requirement for HSDPA receivers.

The EVM performance of the RF transceiver is adversely affected by four primary factors: phase and magnitude response of the receive filter; phase error due to the frequency synthesizer; I/Q mismatch; and dc offset (in direct conversion receivers).

The application of programmable digital filtering in the transceiver can considerably improve the receiver's phase and magnitude response. Implementing digital equalization techniques in the transceiver can equalize the phase error. Digital calibration methods can also be applied to correct I/Q mismatch and dc offsets. Ensuring that the dynamic range of the analog-to-digital converters (ADCs) is optimized will also help mitigate EVM degradation. These digital-centric design enhancements are best implemented using nanometer CMOS technology, which provides the greatest integration, performance and cost benefits.

Making such improvements to RF transceiver design is a must for better EVM performance. However, a more robust RF subsystem design is required to mitigate the effects of multipath interference and fading conditions.

Receive diversity
In high-density mobile environments, such as in cities and other urban areas, the mobile terminal is often subject to multipath interference. In such cases, the received signal contains multiple noisy time-delayed copies of the desired signal (Figure 2). This type of interference can cause deep fading and even nulls at the receiver. Under such adverse radio conditions, data throughput and network efficiency can be greatly compromised.

Figure 2: Time-delayed copies of a signal are often received in noisy environments.

To reduce the degradation of signal integrity resulting from multipath interference, receive diversity must be incorporated into the mobile device's RF subsystem. Diversity operation mitigates deep fades by enabling receivers to concurrently receive and process independent RF signals from two distinct antennas to maximize signal quality and reception. Receive diversity results in fewer fades in the combined signal, thus allowing the decoder in the baseband processor to perform better. The result is improved quality of service (QoS) throughout the entire cell and a boost in data rates of more than twice that of single-antenna designs.

Receive diversity reduces basestation power requirements because less power needs to be transmitted to maintain a high-quality link between the basestation and the handset. With receive diversity, the mobile device can "see" and process two signals instead of one, reducing the likelihood of the basestation having to transmit more power to contend with poor signal quality. This means that the cell coverage for existing subscribers can be extended and the saved system resources can be allocated to new subscribers. Simulation results demonstrate the advantage of receive diversity throughout the cell for both QPSK and 16-QAM in a Category 6 HSDPA network (Figure 3). At the middle of the cell, receive diversity improves throughput by more than twice that of non-diversity receivers.

Figure 3: The advantage of receive diversity becomes obvious through simulation.

Multipath effects can also cause intersymbol interference (ISI), whereby temporal spreading and overlap of individual pulses reduce the receiver's ability to decode data correctly. ISI can be reduced using a rake, or other more advanced, equalizer in the digital baseband processor. These equalizers counter the effects of ISI by independently tuning and decoding each time-delayed component. Linear minimum mean square error (LMMSE) equalizers have been shown to yield superior throughput compared to conventional rake receivers for HSDPA systems of different cell geometries (Figure 4).

Figure 4: The throughput advantage of advanced equalizers combined with receive diversity is shown.

The concept of receive diversity is not new. Wi-Fi systems have employed simple antenna diversity methods for years. These simple diversity solutions use the spatial separation between two antennas to receive two uncorrelated signals, and then select the signal with the highest signal-to-noise ratio (SNR).

More advanced forms of receive diversity are being employed in cellular devices to enhance performance. These advanced diversity receivers simultaneously receive two uncorrelated signals on two distinct antennas. Both signals are then processed using two separate receive paths. The baseband processor applies special algorithms, such as maximal ratio combining (MRC), to combine the two received signals and improve the composite signal quality. When MRC is applied, the receive paths are weighted, in proportion to the SNR of each path, to maximize the total SNR. In other words, the received signals on the two paths are combined in proportion to the strength of each path to yield one optimized signal.

Receive diversity
Up to now, receive diversity technology has been slow to make its way into handsets because it requires extra components, including an additional antenna and receive path, which increases cost and consumes additional board space. For 2/2.5G networks (such as GSM, GPRS and EDGE) this extra cost couldn't be justified because the maximum achievable data rates of these networks constrained carriers from offering enhanced data services, such as streaming media.

With the introduction of higher data rates in 3.5G networks, carriers can now offer a plethora of premium multimedia services. Furthermore, the cost of implementing receive diversity is decreasing as RF components become more integrated and antenna technologies mature. However, there are still a number of challenges facing RF system designers in their efforts to improve performance and reduce the cost and PCB size of multiband 3.5G receive diversity solutions.

Antenna placement in a dual-antenna radio plays a significant role in achieving optimal performance. To take full advantage of diversity technology, the signals received at each antenna must be independently faded, i.e. uncorrelated. A correlation factor is used to measure the dual-antenna configuration's effectiveness. The lower the correlation factor, the more effective the diversity system is at improving the SNR and gain of the combined signal.

The correlation (p) between two antenna output signals is a function of the radiation patterns from the antenna and the distribution of the incoming field, and is defined as:

where Vo1 and Vo2 denote the open-circuit voltages of the two antennas, respectively.

To maintain a low correlation factor, it's important to minimize coupling between the two antennas. The correlation factor directly corresponds to average power that the basestation must transmit to maintain a constant level of QoS at the mobile terminal. As the antenna correlation approaches unity, the basestation must provide more power to maintain the same level of QoS. As a result, the system's diversity gain drops. Given this behavior, an RF subsystem that supports diversity typically targets an antenna correlation factor of less than 0.6.

Low correlation can be achieved by applying some combination of the following antenna diversity techniques:

  • Spatial separation between antennas

  • Utilizing antennas with different polarization characteristics

  • Utilizing antennas with different radiation patterns

With spatial separation, the two antennas are placed far enough apart to minimize coupling. Due to the area restrictions in handsets, and the fact that the wavelength of an 800-MHz cellular signal is on the order of 30 cm, just using antenna separation to achieve low correlation isn't suitable. Diversity solutions that are dependent on spatial separation alone are more suitable for laptop applications, which have more relaxed area constraints. Given the pcb area constraints of handset designs, the use of antennas with different polarization characteristics or radiation patterns to achieve optimal diversity gain is preferred. With this technique, antennas of different types and/or shapes can be used to achieve low correlation.

The antennas must also contend with the EMI, which can be particularly problematic in laptop applications. EMI can be prevented by locating the antennas at the top edges of the laptop screen. The trade-off in this technique is that insertion loss is increased due to the antenna traces' additional length. This insertion loss can be recovered by applying off-chip LNAs at the transceiver's inputs.

In addition to the antenna placement and coupling issues, both antennae are often required to provide multiband support, between 800MHz and 2,500MHz, and still provide the necessary antenna gain and efficiency specified by the system. The main antenna must also contend with specific absorption rate (SAR) requirements because it's also used for W-CDMA transmission on the uplink. Especially in handsets, this places restrictions on the location and type of antenna used. Given the considerable antenna constraints that exist in a dual-antenna radio subsystem, it's important that the entire system design takes these issues into account.

Designing a multiband 3.5G RF subsystem with Rx diversity
The greatest challenge in creating a low-cost, space efficient RF subsystem that supports receive diversity is in transceiver design. Until recently, high-speed diversity-enabled handsets have been both cost and space prohibitive due largely to the lack of a single-chip 3.5G transceiver that supports multiband receive diversity.

Addressing this need, one vendor has successfully designed a radio subsystem that supports tri-band HSDPA/W-CDMA receive diversity operation and quad-band EDGE using a single-chip transceiver (the Sirific Wireless EM3210 radio evaluation module). This multimode RF subsystem fits into a PCB area of 6.1cm? (excluding antennas), and requires less than 172 components.

Beyond HSDPA
HSDPA mobile devices with receive diversity provide optimal downlink throughput levels across the entire network. With data rates up to 14.4Mbps, HSDPA networks can support higher network capacity and more data intensive services. This increased capacity and throughput means more users and data traffic on the cellular infrastructure. For carriers, all of this translates into higher ARPU.

The next logical request from cellular operators is to enhance W-CDMA network performance in the uplink direction. With the addition of high-speed transmit capabilities, carriers will be able to offer services such as streaming media broadcasting, fast file uploading and real-time Internet gaming to end users. In Release 6, the 3GPP is addressing this demand with high-speed uplink packet access (HSUPA), which will pave the way for mobile devices to transmit data at rates greater than 5Mbps.

About the authors
Sathwant Dosanjh
is the director of systems and product development at Sirific Wireless. He holds BASc and MASc degrees in electrical engineering from the University of Waterloo in Waterloo, Ontario. Dosanjh can be reached at
Dr. Abdellatif Bellaouar is the vice president and technical lead for RFIC design at Sirific Wireless He is a graduate of the University of Paul-Sabatier in Toulouse, France and holds a masters and doctorate in microelectronics.

Article Comments - Maximize throughput in HSDPA network...
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top