I S S U E 4 4

T E C H N I C A L I N S I G H T F O R W I R E L E S S P R O F E S S I O N A L Sw i r e l e s s . i o p . o r g

SimulationminimizesRF failure

A P R I L / M A Y 2 0 0 6

Fractals ease integration Acquisition boosts RFIC design Quality RF drives ULCHs

If you’re looking for RF and microwavedesign software that’s fast, accurateand easy to use, CONCERTO is anadvanced modelling package withunbeatable capabilities:

UNBEATABLE ANALYSIS – with threecomplementary techniques – finitedifference time domain (FDTD), methodof moments frequency domain and finite element frequency domain –CONCERTO gives you the right analysis tool for every job.

UNBEATABLE SPEED AND STABILITY –in independent tests CONCERTO wasfaster than other software, remainingstable even with the largest of models.

UNBEATABLE FLEXIBILITY –not only can you build new models while you’re running CONCERTO, you can also use its suspend&restartfeature to suspend one FDTD analysis in order to carry out another.

Don’t take our word for it – order your free trial and check outCONCERTO for yourself.

Method of moments for antenna designand installed performance assessment

www.vectorfields.com

Concerto

Vector Fields Ltd 24 Bankside, Kidlington, Oxford, OX5 1JE, U.K. Tel: +44 (0)1865 370151 Fax: +44 (0)1865 370277 Email: info@vectorfields.co.uk

Vector Fields Inc 1700 North Farnsworth Avenue, Aurora, IL, 60505. USATel: (630) 851 1734 Fax: (630) 851 2106 Email: info@vectorfields.com

A Cobham Group company

FDTD for transient and wideband applicationsPBG structure courtesy of QMC London

Tetrahedral finite elements for accurateeigen mode analysis

Unbeatable performance 3D EM modelling with suspend&restart

L E A D E R 5TV standards battle begins

U P F R O N T 6Most operators are unhappy with their OSS

N E W S 7ip.access wins £85 m

A N A LY S I S 9Antennas: Fractals are the future of antenna design

O P I N I O N 12Mark PaxmanThe business case for WiMAX is looking good, but operators must keep a close eye on cost.

H A N D S E T S 13RF transceiver plays crucial role in ULCHdevelopmentAdvances in transceiver technologies can reduce the bill-of-materials in ultra-low-cost handsets.

P R O T O T Y P I N G 17Virtual systems help weed out bugsVirtual system prototype simulations can improve the debugging process during the development ofmulti-core wireless chips.

P R O D U C T F O C U S 20Simulation tools minimize RF failure in handsets ● APLAC engine powers Analog Office

P R O D U C T S 24Design and simulation software

T H E F U T U R E 26TD-CDMA delivers mobile TVJon Hambidge of IPWireless explains the advantagesof delivering mobile television services via unpaired3G spectra.

3C O N T E N T S

APRIL/MAY 2006ISSUE 44

Institute of Physics Publishing Ltd,Dirac House, Temple Back, BristolBS1 6BE, UK.Tel: +44 (0)117 929 7481Editorial fax: +44 (0)117 925 1942Advertising fax: +44 (0)117 930 1178Web: wireless.iop.orgE-mail: wireless@iop.org

EDITORIALEditor: Hamish Johnston;hamish.johnston@iop.orgProduction: Paul JohnsonArt director: Andrew GiaquintoTechnical illustrator: Alison Tovey

ADVERTISINGSales manager: Simon Allardice;simon.allardice@iop.orgProduction co-ordinator: Jayne Boulton;jayne.boulton@iop.org

MARKETING AND CIRCULATIONProduct manager: Angela Peck;angela.peck@iop.org

PUBLISHERSarah Chilcott; sarah.chilcott@iop.org

PUBLISHING DIRECTORRichard Roe

ISSN 1471-3888

SUBSCRIPTIONSFree to qualifying readers (see subscription cardenclosed or register online at the above Webaddress). Subscriptions £96/7139 ($147 in theUS and Canada) for six issues to readers who donot meet qualifying criteria. Orders to WirelessEurope, IOPP Magazines, WDIS Ltd, Units 12 &13, Cranleigh Gardens Industrial Estate, Southall,Middlesex UB1 2DB, UK. Tel: +44 (0)20 86067518; fax: +44 (0)20 8606 7303; e-mail:wirelessreaderservice@iop.org

©2006 IOP Publishing Ltd

The contents of Wireless Europe do not representthe views or policies of the Institute of Physics, itscouncil or officers unless so identified.

Printed by Warners (Midlands) plc, Bourne,Lincolnshire, UK.

BPA Worldwide Business Publication Auditmembership applied for January 2005

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

Applying fractals to antenna design. p9.

Handset sales grow by 21% worldwide. p6.

Jon Hambidge: delivering mobile TV. p26.

On the cover:Modelling the effectsof hand positions onRF performance. p20.Image: SPEAG.

CST – COMPUTER SIMULATION TECHNOLOGY | www.cst.com | info@cst.com

CHANGING THE STANDARDS.

A major breakthrough A major breakthrough in 3D EM simulation.in 3D EM simulation.Tried, tested and trusted by today’s top design engineers,the release of CST MICROWAVE STUDIO® 2006 is creatingwaves in the world of simulation technology. Embedded in CST DESIGN ENVIRONMENT ™, the market leading timedomain 3D EM tool can be coupled with all CST solver technology including circuit and thermal simulation.CST MWS is the first commercial HF 3D EM code to offer the advantages of time and frequency domain, hexahedral andtetrahedral meshing, united in one interface.

Experience the benefits of complete technology, experienceCST MICROWAVE STUDIO® 2006.

5L E A D E R

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

N E W S I N B R I E F

Huawei grows business in EuropeHuawei’s bid to become a major player in the European wireless market hasbeen boosted by three recent deals. In the first, Huawei will supply UMTSnetwork equipment to KPN Mobile of the Netherlands. China-based Huaweiwill work with KPN to build a unified 2G/3G core network with both circuit-switched and packet-switched capabilities. The radio access network will besupplied by Ericsson and the network will cover all of the Netherlands.Meanwhile in Poland, Huawei will supply the operator P4 with end-to-endUMTS/HSDPA network infrastructure including the core and radio-accessnetworks, handsets and streaming services. Commercial services areexpected to be launched at the end of 2006. Huawei has also signed a five-year deal to supply 3G handsets to Vodafone. The Vodafone-branded phoneswill be marketed in 21 countries from September 2006.

ULCH production could reach 48 million by 2010Annual sales of ultra-low-cost handsets (ULCHs) could reach 48 million in2010 or 5.3% of total phone sales, claims a report from Informa Telecoms &Media. A ULCH is currently defined as a phone costing less than $40 andInforma’s Gavin Byrne predicts that this will drop to $28 in 2010. The UltraLow-Cost Handset Strategic Report also predicts that ULCHs will be a keydriver of market growth in Africa and India, where they will account for morethan 10% of sales by 2010.

Icera gains $40million for modem developmentIcera has secured $40 million in third-round funding to further develop itswireless modems for mobile handsets and data cards. UK-based Icera wasfounded in 2002 and introduced its Livanto soft modem chip in 2005.Livanto is based on Icera’s Deep Execution Processor (DXP) architecture andcomprises a complete wireless modem that is realized entirely in softwarewithout the need for any hardware acceleration. Livanto supports a range ofadvanced mobile standards including HSDPA.

Motorola boosts investment in ChinaMotorola Ventures has pumped money into three Chinese companies in anongoing effort to benefit from the rapid growth of China’s mobilecommunications industry. In one funding deal the company has formed analliance with Legend Silicon for the design, development and manufacture ofdemodulator chips for mobile TV handsets. The second investment involvesthe formation of a strategic partnership with the Shenzhen ShenxunInformation Technology Development Company (SXIT). Motorola willcooperate with SXIT on the development of mobile platforms and applicationsincluding multimedia messaging services (MMS), assisted-GPS and location-based services. Motorola has also increased its investment in ShanghaiNewMargin Venture Capital Enterprise, which is a venture-capital firm.

HSPA will serve 1billion by 2012Nearly 1 billion subscribers will use high-speed packet access (HSPA)technology in 2012, claims a report from Analysys Research. The report alsopredicts that HSPA will increase data revenues for UMTS operators by756 billion in 2012, which is almost half of projected 3G revenues. HSPArefers to the HSUPA and HSDPA upgrades to UMTS, which boost data rates inthe uplink and downlink respectively. Entitled The Global Market for HighSpeed Packet Access, the report was commissioned by the UMTS Forum,which is an industry body that promotes UMTS. “By 2012, 3G HSPA users willbe generating average traffic per user seven to nine times that of non-HSPA3G subscribers as a result of higher-quality user experience and lower tariffs,”said Jean-Pierre Bienaimé, chairman of the UMTS Forum.

There was no getting away from mobile television at thisyear’s 3GSM World Congress – and after a compellingdemonstration of Qualcomm’s MediaFLO technology, I wasconvinced that TV will be a killer application.

The most pressing issue facing the mobile-phone industryis the precise role that it will play in the delivery of mobileTV services. As one might expect, the mobile operator wasthe linchpin of the mobile TV business models presented at3GSM. After all, operators have the customer base, theydecide which devices to offer their customers and they havethe billing systems to capture payment.

The problem is that most of the leading technologicalcontenders for mobile TV can be delivered without theinvolvement of a mobile operator. The DVB-H, DAB andMediaFLO mobile TV standards are broadcast rather thancellular technologies and operate best in the UHF bands.While mobile operators would have the expertise and real estateto roll-out broadcast networks, they don’t own the airwaves.

IPWireless’s Jon Hambidge believes that the prospect ofacquiring more radio spectrum is enough to turn mobileoperators away from a non-cellular technology (see page 26).Of course operators could partner with a broadcast network,but this would further dilute their share of the revenues.Nokia estimates that two-thirds of mobile TV revenue will goto content providers, with the rest going to the distributors.Current wisdom is that consumers will spend about 610 permonth on services, leaving mobile operators with a euro ortwo if they choose to partner with a broadcast network.

Hambidge believes that the prospect of such meagrereturns has made operators keen on delivering TV servicesover 3G. Fortunately, many 3G operators already ownsignificant chunks of vacant 3G spectra – the so-calledunpaired bands – which could be exploited.

TDtv to the rescueAt the behest of CTOs at leading 3G operators, IPWirelesshas created its TDtv system. The TD-CDMA air interface isused to broadcast TV in the unpaired bands. As well as beingin full control of the broadcast system, TDtv has the benefitof being based on 3GPP-approved technology.

While TDtv may seem a more natural technology formobile operators, questions still remain. The unpaired bandsare up at 2 GHz, where signals fade rapidly and have troublepenetrating buildings. It is also much cheaper to manufactureUHF equipment than devices operating at 2 GHz – althoughHambidge says that in the long term TDtv will become justanother layer in a UMTS chip, reducing the PCB area andcost required to equip a phone with TV.

The key unknown, however, is who will move first oncommercializing mobile TV. If it’s mobile operators, TDtvmay be the technology of choice. Other broadcasters wouldlikely favour DVB-H.

Hamish Johnston, Editor

TV standards battle beginsE D I T O R I A LN E W S

6 U P F R O N T

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

Most operators areunhappy with their OSS

Three of Nokia’s latest handsets: theFinnish company sold about 265 millionmobile phones in 2005, making it thelargest handset maker in the world.

Annual global sales of mobilephones grew by 21% in 2005compared to 2004, according toa survey by the analyst Gartner.Sales reached an all-time high of816.6 million units in 2005, with

Nokia leading the way with32.5% of units sold – up from30.7% in 2004. Motorola andSamsung hung onto second andthird position respectively, withMotorola increasing its marketshare by more than 2% to17.7%. LG and Sony Ericssonwere fourth and fifth respectively.

2005 was a year of transitionfor Siemens, which ended 2004in the number four position with7.2% of annual sales. It then soldits handset division to BenQMobile in October 2005. BenQshipped 11 million handsets inthe fourth quarter of 2005,putting it in sixth position.

Asia/Pacific was the largestregional market in 2005, with204 million units sold. WesternEurope bought 164 million, andNorth America 148 million.

Only two out of 25 Europeanmobile operators are happy withtheir current operational supportsystem (OSS), claims a recent sur-vey by the market-research firmVanson Bourne. The reportrevealed that 92% of operatorssurveyed said that their OSS wasinadequate in its support of net-work evolution, 3G technologiesand the provision of new services.

While a majority of respon-dents said that an OSS shouldplay an important role in radioand transmission network plan-ning, only one operator said thattheir OSS could adequately sup-port these tasks.

Network inventory was seen asthe most important OSS func-

tion by 72% of respondents, butmore than two-thirds of opera-tors found current inventory sys-tems difficult to update. Almost90% of operators said that theyemployed a home-grown inven-tory system that is often based ona simple PC application such asMicrosoft Excel.

A majority of operatorsexpressed a need to combine allexisting inventory systems –including those covering networkinfrastructure equipment – intoone converged system.

The survey polled networksupport managers at 25European operators and was per-formed on behalf of Comptel,which supplies OSS platforms.

Handset sales continue to soar...

There are now 315 W-CDMA3G terminals on the marketaccording to a survey by theGSA, which promotes the inter-ests of GSM and W-CDMAequipment suppliers. The GSAreported that the number of dif-ferent commercial handsets anddata cards supporting W-CDMAhas increased by 129 in the pastsix months. UMTS terminalsaccount for 80% of W-CDMAdevices, while the remainder areFOMA handsets, used in Japan.

HSDPA is supported by 25W-CDMA devices, more thandouble six months ago. Multi-mode GSM/EDGE/W-CDMAaccess is available on 48 handsetmodels, up 40% in six months.

The GSA says there are 102commercial W-CDMA networksworldwide, with more than50 million subscribers. This isstill less than one-third of the169 million subscribers that aresaid to be using cdma2000,which is a rival 3G technology.

...as more W-CDMA models are released

Airspan Networks has revealeddetails of a wireless USB devicecompatible with the 802.16emobile WiMAX standard. An“industry first”, the 16eUSB sup-ports multiple-input multiple-output (MIMO) antennadiversity, beam forming, han-dover and idle and sleep modes.

The device operates in fourbands between 2.3 and 5.4 GHz,where mobile WiMAX servicesare expected to operate. Airspan’spresident Eric Stonestrom said:

“This will allow a user to haveaccess to WiMAX networks vir-tually anywhere in the world.”Development was sponsored bythe Japanese telecoms operatorYozan, which operates a WiMAXnetwork in Tokyo based on the802.16d fixed-wireless standard.Yozan intends to launch mobileWiMAX services in 2007.

Airspan has also announcedthat mobile WiMAX is now sup-ported on its AS.MAX line ofWiMAX base stations.

Airspan unveils mobile WiMAX devices

7B U S I N E S S

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

ip.access wins £8.5m

Gold Sponsor Silver Sponsor

22-25 May 2006, Vienna

Esbjörn Nordesjö, Head of Network Engineering, 3 SwedenGeoffrey Sole, Principal Engineer, 3 UK

Nicholas Payant, Technology Advisor – Radio Systems, Bell CanadaSvein Haglund, Technical Project Manager, MCP

Pedro Oliveira, Project ImplementationManagement Group Manager, Sonaecom

Bernd Margotte, Network Roll Out Manager, Swisscom MobileMagne Pettersen, Research Manager, Telenor

Paul-Flemming Hermann, In-Building & SpecialSolutions Expert, Access Network, Telia Mobile

Hans-Hubert Röhrig, Senior Expert 2G/3G RF Repeater& In-building Systems, T-Mobile International

Ilhan Akinci, Cell Planning and Optimisation Engineer, TurkcellSotiris Soukias, Radio Systems Engineer, Vodafone GreeceDaniel Almodovar Herraiz, Access Technologies Evolution –

Converged Networks, Vodafone Group Research Development

IIR is delighted to announce details of the 10th Annual In-Building Solutions European Summit. This event attracts a formidable groupof technical experts who share their ideas and discuss the specific

challenges of indoor RAN planning. Don't miss this opportunity to develop and consolidate your knowledge and contacts in this field.

www.iir-conferences.com/in-building2006

The conference programme includes major contributions from:

MEMS capacitor maycut multimode costsA duplexer based on RF microelectromechanical system (MEMS)technology could reduce the sizeand cost of multimode GSM/W-CDMA handsets, claims areport from the analyst firmStrategy Analytics. Created by US-based WiSpry, the duplexer isbased on MEMS variable-capaci-tor technology and will be releasedcommercially later this year.

The device addresses animportant challenge in the real-ization of multiband, multimodehandsets – the development oflow-cost, compact and low-lossfilter and switching systems. “If itperforms as claimed, it willreduce the number of passivecomponents needed in hand-sets,” said Chris Taylor, directorof Strategy Analytics’s RF andWireless Components Service.

ip.access has secured £8.5 millionto expand its 2G in-building base-station products and to develop3G base stations for homes andsmall businesses. Chief executiveStephen Mallinson told WirelessEurope that ip.access intends todeliver indoor 3G base stations byearly 2007. This equipment willbe single-mode UMTS ratherthan dual-mode GSM/UMTS.“There is a greater need for indoorUMTS coverage,” he explained.

Funds were provided byScottish Equity Partners, IntelCapital and Rothschild & CieGestion. UK-based ip.access’snanoGSM equipment is distrib-uted by Siemens and deployed byseveral operators worldwide.

The company has also joinedforces with the UK-based cablecompany NTL to explore thepotential of deploying GSM base

stations in the homes of cablesubscribers. The base stationswould be backhauled via cableand would operate at very lowpower using spectra that wasreserved for the GSM/DECTguard bands. A bidding process iscurrently underway for licencesin these bands.

4G downlink hits2.5 Gbit/s in JapanA packet-data transmission rateof 2.5 Gbit/s has been achievedon a fourth-generation radio–airinterface developed by Japaneseoperator NTT DoCoMo. Trans-mission was achieved in thedownlink between a stationarytransmitter and a receiver travel-ling at 20 km/h.

The interface employs the vari-able spreading factor spreadorthogonal frequency divisionmultiplexing (VSF-spreadOFDM) radio interface, which isbeing developed by NTTDoCoMo as a 4G air interface.The downlink was enhancedusing six-channel MIMOantenna diversity and a 64-QAMmodulation scheme.

The system achieved a fre-quency spectrum efficiency of25 bit/s per Hertz of bandwidth.

Stephen Mallinson of ip.access intends todeliver in-building 3G base stations in 2007.

8 B U S I N E S S

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

Contract manufacturing keeps growing

The definitivenetworking event forthe handset industryWhy you MUST attend:

• NETWORK with key decision makers: 67% board-level & businessdevelopment attendance

• DISCOVER where the industry is heading: Mobile TV, VoIP, HSxPA and Ultra LowCost Handsets

• LEARN about the hottest trends in the industry: Mobile Music & 3G, User Interface design

• MEET Network Operators, HandsetManufacturers, Component and SoftwareSuppliers

9-11 MAY 2006HILTON HOTEL, PRAGUE, CZECH REPUBLIC

Part of the World Handset Forum SeriesOver

140 Senior

Executives at

Last Event

Book now at www.informatm.com/whf

Learn from High-level PresentersCovering the EntireHandset Value Chain

Sony Ericsson

T-Mobile International

HTC

O2 Group Technology

NEC Europe

Hutchison Whampoa

NMT

Nokia

Symbian

visit the website at: www.imsvision.com

Launching IMS Based Services and Achieving Multiparty Interoperability

Conference & Exhibition 25th – 27th April 2006, Hotel Majestic, Barcelona, Spain

2006

INCLUDING 16 OPERATOR PRESENTATIONS

SELL OUTin 2005!

Post-Conference Workshop AMaking The Business Case For

IMSLed by ??

IMS As A Guiding Principle

To Readjust The Operators Role

And Strategies

Led by Telecom Italia,Swisscom & Vodafone

Full Day - Friday 28th April

Post-Conference Workshop

Operators Only

REGISTRATION HOTLINE +44 (0)20 7017 5506

• Defining and prioritising which IMS services to deploy

• Examining the requirements and issues needed for FMC with IMS

• The evolution of personal devices to suit IMS based services

• Achieving a cost effective and seamless migration of subscribers

• IMS Interoperability issues between infrastructure vendors and network operators

• Enabling third party developers to create their own IMS based applications

• Featuring new case studies from USA, Asia and Europe

Hear over 35 high-level presentations covering

newly researched topics such as:

Brand

New Fo

cus

for 2

006

Post-Conference Workshop AMaking The Business Case For

IMSLed by ??

Making The Business

Case For IMS

Led by MotorolaHalf Day - Monday

24th April (Morning)

Delivering securityin the IMS

environmentLed by Gemplus

Half Day - Monday24th April (Afternoon)

Pre Conference

Workshop A

Pre Conference

Workshop B

UK-based Actixbuys RadioplanActix has acquired Radioplan,which provides automated net-work optimization software forcellular operators. UK-based Actixoffers a range of network opti-mization systems.

Actix chief executive DaveWilkinson said: “Actix andRadioplan products are a naturalfit; like us Radioplan automateskey wireless engineering processesto realize potentially huge reduc-tions in the cost of engineeringwireless networks.”

Radioplan is based in Germanyand has customers in Europe, theUS and Asia. The acquisitionbrings Actix’s product-develop-ment team up to 130 employeesbased in eight countries.Radioplan was a privately heldcompany and terms of the acqui-sition have not been made public.

The contract manufacturing ofmobile handsets will continue togrow and will account for 47%of all devices made in 2009, saysa report from analyst firmVisiongain. Called Outsourcingin the Mobile Handset Market,the report predicts that out-sourced handsets will account for36% of sales in 2006. In 2005,the leading contract manufactur-ing company Flextronics pro-duced 47 million handsets, morethan 5% of handsets produced.

According to Visiongain, themove to contract manufacturingis being driven by leading net-work operators. “Own-brandhandsets are one way for opera-tors to increase data usage, andthe trend towards greater cus-tomization will impact upon themanufacturing side of the indus-try,” says the report.

Vodafone’s 804SS 3G handset for the Japanese market: the demand for own-brandhandsets is growing contract manufacturing, claims Visiongain.

9A N A LY S I S

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

Ruben Bonet explains why fractalgeometries will help antenna designersmeet the size and performancedemands of future phones.

Designers must grapple with several seem-ingly incompatible demands when creatingantennas for mobile phones. Modern hand-sets require internal antennas that are smallbut support multiple bands; that are low costbut increasingly complex; and that can beproduced in large volumes while being cus-tomizable to the design requirements of ter-minal manufacturers. And if that isn’tenough of a challenge, antennas must deliverthe exceptional reliability and consistencythat all operators, manufacturers and end-users demand and expect.

It is ironic, therefore, that the one elementthat is so fundamental to the wireless world– the antenna – has for so long been a com-ponent that was shoe-horned into a handsetdesign at the end of the development process.Fortunately this is changing and rather thanbeing seen as an afterthought, today’s anten-nas are assuming a central role in the handset-design process. As form factors continue toevolve and an increasing number of connec-tivity standards are incorporated within hand-sets, the pressure on antenna producers toanswer these challenges within a single com-pelling device will be immense. Only by tak-ing a new approach that returns to the veryfundamentals of antenna geometry, can theseburgeoning requirements be met.

Demanding timesThese demands are being driven by the rapidpace of change in handset design and func-tionality. Clamshell, slider and swivel phonesall require antennas that deliver peak perfor-mance for each new device – and perfor-mance must be unhindered by the shape ofthe phone. Antenna performance can be

affected by all the materials and electricalprocesses surrounding it. When a device isdesigned to operate in more than one shapeconfiguration – a slider phone can be bothopen and shut for example – the antennamust work just as effectively in all configu-rations. Indeed, a slider phone is two com-pletely different devices, as far as the antennais concerned. Handset materials andprocesses will also differ from manufacturerto manufacturer and from device to device,so one antenna will never be appropriate forall handset types.

Form-factor complexities are further com-

pounded by the need for multimode hand-sets to support a proliferation of cellular andwireless standards. A modern handset mayneed to operate on four GSM bands, whilealso supporting a CDMA band, UMTS,WiFi, Bluetooth and GPS. While all of thesemodes may not use a single antenna, there isa strong desire to minimize the number ofantennas required. Another key design pres-sure is the move from external antennas tointernal handset components. This putssevere pressure on the antenna industry tosignificantly reduce antenna size and increaseperformance. The appearance of new stan-

A N T E N N A S

Fractals are the futureof antenna design

The mathematics of fractals can be used to design antennas that make efficient use of the limited space insidehandsets, while boosting performance and lowering costs.

1 0 A N A LY S I S

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

dards on the horizon – including WiMAX,DAB, DVB-H and UWB – suggest that thischallenge is only going to increase as antennatechnologies are becoming increasinglyimportant to handset designers.

Research and development activity onantennas has taken two approaches – mat-erials and geometry. Materials-based R&Dhas focused on developing new materials thatcan support the growing multiband require-ments of modern antenna design. The geom-etry-based approach seeks to use existingmaterials but with a new focus on the geom-etry of the antenna to enable the optimiza-tion of size and performance.

The materials approach requires an invest-ment in new processes and raw materials,which can push up the cost of antenna devel-opment. This is not very appealing to hand-set manufacturers, which work very hard tominimize the bill of materials in a device andstreamline the manufacturing process. Inaddition, antennas that are developed using amaterials approach tend to have a fixedshape, which limits handset design.

Geometry-based antenna innovationsallow designers to take a more innovativeapproach to the overall design of the hand-set while optimizing antenna performance –and ultimately the connectivity of the device.As a result, geometry can take handset designin a new direction where the device and theantenna are together part of a symbioticdevelopment process. Existing manufactur-ing processes can be used in the productionof the antenna, which reduces both the up-front investment and the time to market.

Natural technologyOne promising geometry-based approachinvolves the use of fractals. Fractals are oftenrepresented as a geometric pattern that isrepeated at ever smaller scales so that eachdetail of the pattern is a smaller copy of thewhole. Fractal shapes are often self-similarbecause individual elements look the sameas each other and like the whole object, allindependent of the length scale.

Often referred to as a technology of nature,fractals can be used to describe a wide rangeof natural phenomena. For example, light-ning and snowflakes take fractal formsbecause those shapes are structurally efficient.Plants such as ferns and capillary patterns inskin have evolved fractal forms because theyprovide an efficient way to exploit and fill allthe available space.

Fractal properties are defined by the under-lying mathematics and can involve irregulargeometries – including shorelines and space-filling shaped antennas. Indeed, fractal tech-nology is playing an important role in a new

approach to antenna design that optimizesantenna size and performance to create pow-erful multiband antennas.

Easy integrationAntenna-design benefits can be achieved bytaking advantage of the fundamental prop-erties of fractal geometries. Fractal antennasare based on geometrical/design innovation,not on new materials or processes. They canbe produced without any change in currentelectronic component production processesor materials – and this makes fractal antennaseasy to integrate into a multitude of devices.

The ability of fractals to fill available 2-Dor 3-D space in an efficient manner meansthat small antennas can be designed todeliver the same performance as larger anten-nas. The multiscale nature of fractals can beexploited to create multiband antennas.

In 2002 the Spanish company Fractus usedits FracPlane proprietary fractal technologyto produce an internal quadband antennaover the handset ground plane. This single-component antenna can support four 2Gmodes – GSM850, GSM900, GSM1800and PCS1900 – in a compact 2.7 cm3 vol-ume.

Notable benefits can be achieved whenhandset designers work with the antennasupplier well before the device layout is final-ized. This enables the best possible integra-tion of the antenna and the device. This isbecause the antenna size and performance aredefined earlier in the development process,which results in optimal integration and easeof production. It is at this level that technol-ogy innovators can help handset developersget the most from new technologies.

Internal antennas, as we know them today,

will be around for a long time, but beforelong we can expect to see the first handsetswith antennas that are integrated on the elec-tronics package. The first are likely to beantennas providing short-range wireless con-nectivity, which do not involve as many tech-nology hurdles in the development process.

Embedding connectivity into the veryheart of the device heralds a massive changein the antenna market. Antennas could beproduced by electronic component manufac-turers as simply additional elements of com-ponent packages. Terminal manufacturerswould no longer require in-house RF exper-tise – or could choose to devote this valuableresource to other higher value projects.

Wider impactAntenna-in-package (AiP) technology couldalso make a significant impact on the widerconsumer electronics market. This is becauseit will allow manufacturers with little RFexperience to add wireless connectivityquickly and easily into a wide range of prod-ucts without having to increase the numberof components.

The development of AiP technologies is alogical solution to the challenges of reducingthe size and cost of mobile phones. Researchsuggests that fractal technology will play acritical role in the development of AiP byproviding high-performance antennas thatare small enough to embed within electronicspackages. The first AiP devices based on frac-tal technology are now at the trial stage withmajor component manufacturers and areexpected to reach the market within the next12 months. ■

Ruben Bonet is chief executive of Fractus.

Fractus’s dual-band Reach Xtend antennas are for 802.11a/b/g/j/n operation in the 2.4 and 5 GHz bands.

If you‘re looking for a test solution for Bluetooth components that‘s fast, flexible and cost-effective, there‘s only one way to get it: the R&S®CBT series.

◆ Fast: The R&S®CBT performs a complete basic rate transmitter test on three channels in less than two seconds.◆ Flexible: With the R&S®CBT you can perform tests strictly in compliance with Bluetooth SIG test recommendations, includ- ing for EDR. For example, the R&S®CBT is the first test solution to offer EDR loopback mode for receiver tests under quasi-real- istic conditions. Alternatively you can vary all test parameters for customized test strategies within broad limits.

Need a Bluetooth® booster? It‘s ready for lift-off.R&S®CBT – the Bluetooth® tester as fast as EDR

◆ Cost-effective: Pared down to a production environment configuration, the R&S®CBT 32 gives you the performance of our high-end testers at a price that makes it economical to test even low-cost products.

You can‘t get this kind of performance anywhere else. So if it‘s time for a change in your testing, get in touch with us.

www.cbt-we.rohde-schwarz.com

1977_Anz.CBT_213x282_e.indd 1 20.03.2006 20:32:44 Uhr

1 2 O P I N I O N

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

M A R K P A X M A NThe business case for WiMAX is looking good in certainmarkets, but operators must still keep a close eye oncost, says Mark Paxman.

WiMAX is really achieving credibility. Fixed-wireless equipment is available today andis performing well – and mobile WiMAX is not that far off. Perhaps as a result of thisearly success, an emboldened WiMAX marketing machine has turned up the volume.So it is a good time to ask how closely the marketing hype matches the technology.

The WiMAX Forum has proposed several key markets for WiMAX, namely fixedbroadband, cellular backhaul and mobile data. In many regions, however, suchWiMAX services will be up against fierce competition from existing broadband,backhaul and cellular technologies.

But all is not lost and WiMAX is emerging as a fantastic solution where thealternatives are poor. This includes many markets in the developing world where fixedinfrastructure simply doesn’t exist, or is over-regulated and uncompetitive. Some areasof the Middle and Far East, Eastern Europe and Latin America fall into this category.In these regions the costs associated with building a WiMAX network and thetechnology’s very short time-to-service conspire to make WiMAX look good from a

financial point of view.Closer to home, in developed markets WiMAX is

a credible contender for most applications, but“buyer beware” is the key. The hype and tech-nological performance must be carefully scrutinized,particularly if the business case doesn’t quite add up.

Indeed, developing WiMAX business cases is aparticularly complex task. The WiMAX standardoffers a wide array of performance-enhancingoptions and features and, to further complicatematters, some equipment vendors have included

their own non-standard features. This results in very large spreadsheet models, andmakes performance trade-off analysis a particularly complex and time-consuming task.

However, it’s worth persevering with complex business models, because choosing theright combination of features can really make a difference to the viability of a businesscase. For example, at PA Consulting we have been studying advanced antenna systemsthat use multiple antennas at each WiMAX base station to improve range or capacity.As a result, the deployment of multiple antennas can lead to substantial financialbenefits in some circumstances.

Ultimately, the best technology choice depends on what the operator wants toachieve with its wireless service – be it good indoor coverage; high capacity; reliablehandover; or self-installation of customer premises equipment. And it will be the cost,not the performance, that operators must ultimately bear in mind.

Like all wireless technologies that came before it, WiMAX claims to be all things toall people. In practice, however, it will thrive in situations where it makes economic –and not just technical – sense. Prospective WiMAX operators must take a hard look atthe figures before committing to WiMAX, or indeed any other technology. ■

Mark Paxman is a managing consultant at PA Consulting’s Wireless Technology Group.He can be contacted at mark.paxman@paconsulting.com.

Operators must takea hard look at thefigures beforecommitting toWiMAX, or indeedany other technology

Components to helpyou design tomorrow.

Xinger® 0404,nano-profile baluns> Nano-profile, sub-miniature SM format,

only .039 x .039 x .026” (1 x 1 x .65mm)> Low loss: 0.6dB typ> Typical return loss: >16dB> Complies with 802.11 a, b, g; MIMO;

Bluetooth®; Zigbee...covering 2400-5900MHz

> Many more advantages!

Xinger® 0603, ultra-low profilebaluns, jumper, and crossovers> Low-profile, SM package measures a mere

.059 x .029 x .027” (1.5 x .75 x .7mm)> Low loss: 0.7db typ> Typical amplitude balance: <0.6dB> Complies with 802.11a, b and g

Uniband I and II, WiMAX, UWB, Bluetooth®, Zigbee...covering 2300 MHz to 6.0 GHz

> Many more advantages!

Xinger® 0805, ultra-low profile baluns> Measures a mere .08 x .05 x .03”

(2 x 1.25 x .7mm)> Low loss: starting at 0.6dB typ> Typical amplitude balance: <0.6dB> Compliant with all current consumer band

allocations, from 400-6000MHz...802.11 a, b and g Uniband II and III, WiMAX, UWB, Bluetooth®, Zigbee...covering 400-6000 MHz

> Many more advantages!

Call for our free samples or our completeCCG engineering kit (qualified inquiries).

In Europe, call 44-2392-232392 ISO 9001 certified Visa/MasterCard accepted (except in Europe)

Yesterday.

Today.

>800-411-6596 >www.anaren.com

Gary Levy explains howadvances in transceivertechnologies canreduce the bill-of-materials in ultra-low-cost handsets.

There are about 3 billion peo-ple in emerging markets wherewireless networks are the mostcost-efficient communicationaccess technology. Commercialsuccess, however, will only beachieved in these markets if thetotal cost of handset ownershipis addressed. This vast businessopportunity has not gone un-noticed and there has been asurge of enthusiasm about thedevelopment of ultra-low-costhandsets (ULCHs).

Affordability is the most important driver of subscribergrowth in developing markets. The pricing of mobile services isdetermined by handset cost (including distribution), the ser-vice offering and the local regulatory environment and tax struc-ture. Today, the general consensus is that a ULCH shoulddeliver dual-band GSM voice and SMS services with no packet-data capability and minimal features for a retail price under $30.At the same time, certain performance parameters – notablysensitivity – must be achieved with significant margin.

Operating costs will be kept low in ULCH service areas byusing refurbished network equipment, minimal infrastructuredeployment and scant after-sales service. As a result, operatorsoffering ULCH services must pay close attention to user satis-faction and handset quality. To avoid costly handset returns,ULCH handsets must be high-quality and high-performance.This is also in the best interests of leading handset makers, whocould damage their brand if they produced substandardULCHs. Therefore handset manufacturers must also pay criticalattention to quality and handset yield.

An obvious way of reducing handset price is to use lower-cost components at the expense of performance, feature-set orquality. However, designing a ULCH terminal is not a trivialtask and the product definition can have counterintuitive effects

upon the radio chipset. While the RF portion of the handsetmay represent less than 10% of the overall bill-of-materials(BOM), a series of recent advancements in RF technology coulddirectly benefit the GSM/GPRS ULCH market. As a result,handset designers must have a good understanding of RF whendesigning ULCH terminals. Designers must also be aware of theRF front-end and reference-clock interface as well as perfor-mance, manufacturing and yield-related characteristics of trans-ceiver integrated circuits (ICs).

Multiple strategiesULCH design and manufacture involves neither a specific tech-nology nor prescribed methodology. Rather, it is a process thatuses a variety of strategies to lower the total cost of handset own-ership at an accelerated rate. In general, ULCH designers shouldselect a radio with the lowest BOM, highest integration and low-est external component count. An affordability analysis must alsobe done at the system level, where sensitivity is a critical ULCHparameter that is related directly to transceiver performance.

System sensitivity is important because it has a direct impacton the efficiency of the network infrastructure as well as the userexperience. Sensitivity also affects handset design issues includ-ing radio front-end component selection and cost. Networkoperators and handset manufacturers alike focus on sensitivity

1 3H A N D S E T S

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

RF transceiver plays crucialrole in ULCH development

PGA

PGA

SPI

I

Q

DAC

SYSCLKAFC

RFOH

RFOL

PA

digitalLOW-IF

IADC

QADC

Si4210 GSM/GPRS XCVR

reference oscillator

XTAL1 XTAL2

DCXO

SAWs

RFIA

RFIE

RFID

RFIP

inputmatch

850

900

1800

1900

850900

18001900

PGA

PGA

sensitivity = –174dBm/Hz + 52.3dB + LASM + LSAW + NFXCVR + 10log (S/N)min

φDET

VCO + frequencysynthesizer

0/90

ante

nna

switc

h

digi

tal f

ilter

anal

ogue

inte

rface

base

band

Fig. 1. Sensitivity derivation for a GSM radio, which shows the contributions from various components.

www.anritsu.comEurope + 44 (0) 1582 433433 United States +1 972 644 1777 Japan + 81 46 296 1264

Trouble Communicating?Innovative Test and Measurement equipment is critical for the development,implementation and consistent performance of communication systems. With theintegration of NetTest’s unique OSS service assurance solutions, Anritsu now provides thebroadest portfolio of solutions globally for existing and next-generation wired andwireless telecom networks.

The drive to provide customers unparalleled levels of support sees our product rangespan the entire network life cycle, from development and conformance testing tooptimization and service assurance.

With the widespread planning and introduction of technologies such as HSDPA andHSUPA, Anritsu is a committed, proven developer with a leading position as applicationsevolve and new test challenges arise. Partnerships with global telecom leaders andparticipation in standards organizations allows us to provide the most complete range oftest and measurement and service assurance solutions available today, helping ourcustomers meet their business objectives.

Discover how Anritsu keeps the world communicating, visit: www.eu.anritsu.com/totalnetwork

1 5H A N D S E T S

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

because it affects call quality.A phone that can consistentlyreceive a smaller signal than itscompetitors will have bettervoice clarity and fewer call dis-ruptions.

Improved sensitivity impr-oves the coverage enjoyed by ahandset within a given networkcell – including at cell bound-aries – and provides betterimmunity to signal fading. Thiscan be exploited to lower infra-structure equipment expendi-tures, and therefore operatorsmay choose to offset the poten-tially modest ULCH averagerevenue per user (ARPU) by minimizing their capital investmentsin base stations and related infrastructure equipment.

Handset manufacturers focus on sensitivity performancebecause it relates to manufacturing design and production mar-gins as well as handset call quality. For example, the antennadesign will have a significant effect on isotropic receive sensitivityperformance. While internal antennas are increasingly popular,they increase design complexity and the location, size, form-factorand performance of the antenna will often be compromised toconform to the mechanical requirements of the phone.

Transceiver impactThe transceiver’s impact on ULCH product design is illustratedin figure 1, which shows the signal flow within a GSM radio.During the receive operation, the antenna switch (ASM) directsthe incoming signal from the antenna to the appropriate surfaceacoustic wave (SAW) filter for frequency band selection. Thesecomponents introduce an insertion loss (IL) to the received sig-nal that must be accounted for in the design. An input matchresides between the SAWs and the transceiver to maximizeeither the power or voltage transfer from the SAW output tothe transceiver’s low-noise amplifier (LNA) inputs.

The transceiver then down-converts the RF signal to an inter-mediate frequency, further mixes the signal down to basebandfrequencies – or alternatively converts the signal directly downto DC – and then selects the proper receive channel. Thereceived signal is degraded by the transceiver during this opera-tion and the degree of degradation is defined as the noise figure(NF). Voice and data information are then passed across thebaseband interface.

The transceiver selected for a ULCH terminal – indeed forany handset – should provide the lowest possible NF. A signifi-cant margin between the required NF and the actual value canbe exploited to lower the overall BOM. This is because thehandset designer can select front-end components along thereceive signal chain, including the ASM and SAWs, that offerthe best trade-off between cost and IL.

Such a trade-off could include the use of dual SAWs, whichcombine two SAWs into a single module, in place of two dis-crete SAWs. This can achieve a 20% cost reduction when com-pared to two discrete SAWs and dual SAWs occupy about 45%

less space on the PCB. On the downside the use of dual SAWscan result in a slight performance degradation, includingincreased IL and passband ripple, which can affect system sen-sitivity and reduce blocker attenuation that can potentiallyaffect GSM full-type approval (FTA). However, for transceiverswith ample NF margin and strong blocking performance, thesepotential effects are likely to be imperceptible at the system level.

Superior NF performance also allows designers to use lowercost, multi-layer ceramic (MLC) inductors for the LNA inputmatch. Wire-wound or coil inductors are normally used for theinput match, as they deliver a high quality factor (Q) for the bestmatch. While a 75% cost reduction can be achieved by usingMLC inductors, the result is a reduced Q, which often translatesinto a higher transceiver NF. This problem can be avoided byusing a transceiver with a low NF that is also designed specifi-cally for a low-Q input match. These features combine to min-imize the NF degradation to a tolerable level, allowing the use oflow-cost MLC inductors.

At the reference oscillator interface, transceiver timing is syn-chronized to an external, crystal-based source. The transceiverthen provides a system clock to the baseband. Expensive voltage-controlled, temperature-compensated crystal oscillator(VC-TCXO) modules were traditionally used to maintain thetarget oscillation frequency.

Cheaper oscillatorsHowever, a move is underway to smaller and lower-cost altern-atives to the VC-TCXO. Numerous handsets now employtransceivers with either voltage-controlled crystal oscillators(VCXO) or fully-integrated digitally-controlled crystal oscilla-tors (DCXO). While the VCXO replaces the VC-TCXO witha standard AT-cut crystal, varactor and capacitors, integratedDCXO circuitry provides the lowest BOM by integrating allexternal components except a 3.2 × 2.5 mm crystal that costs$0.35. Integration eliminates the need for varactor calibrationon the manufacturing line and the overall cost is very favourablewhen compared to a VC-TCXO costing $0.95.

This type of integration is illustrated in figure 2. While it isnot a new solution, the integrated DCXO approach has enjoyedsuccess in high-volume production by only a small number oftransceiver vendors.

ante

nna

switc

hPA

SAWsinputmatch

850

900

1800

1900

850900

18001900

DAC

GSM/GPRS/EDGEtransceiver

tosynthesizer

XCVR VCXO XCVR DCXO

LDO

varactor

VC-TCXO module

2, 1,

AFC

Q

I

dataenable

clock SPI

SYS clock(13 or 26 MHz)

crystal(13 or 26 MHz)

base

band

Fig. 2. Conceptual diagram illustrating integration trends at the reference oscillator interface.

1 6 H A N D S E T S

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

For transceivers with a proven DCXO and large on-chip capac-itive tuning range, the use of larger 3.2 × 5.0 mm crystals candeliver an additional near-term 10% cost saving. Since ULCHterminals have minimal features, the small size increase shouldnot be a problem. Low-cost larger crystals are likely a transientphenomenon caused by the vagaries of the return on technologyinvestment and shifting production line volumes. The newerChinese and Taiwanese crystal vendors are a size-generationbehind established vendors and will need to aggressively discounttheir 3.2 × 5.0 mm products. However, in the longer term, the3.2 × 2.5 mm crystals will become cheaper than their larger coun-terparts thanks to higher material costs associated with largerpackages. Regardless of the economics of various crystal sizes, atransceiver with broad DCXO production success will deliverspace savings on the PCB and the lowest BOM.

CMOS and dieAt the transceiver manufacturing level, the process technologyshould be chosen wisely to ensure a feasible business modelwhile meeting ULCH cost requirements. Complementary metaloxide semiconductor (CMOS) has emerged as the best processtechnology for producing integrated circuits for cellular hand-sets. CMOS transceivers benefit from lower manufacturing costand a smaller die area thanks to a smaller lithography size –when compared to silicon germanium (SiGe) BiCMOS andsilicon BiCMOS.

The design of CMOS ICs does sometimes present challenges.When implementing certain types of complex systems – partic-ularly those that require both stringent RF performance met-rics and large digital integration – CMOS does not always offera desirable Q or the best insulation against substrate couplingand high-frequency parasitic degradation. In the last severalyears, however, some CMOS GSM transceivers have overcomethese design challenges. These devices have achieved significantcommercial success and delivered cost and integration-relatedbenefits. Since CMOS is also used for the baseband functions inmobile phones, this innovation is critical in the pathway to truesingle-chip integration of the RF, baseband and power-management functions.

Transceivers designed with a high-volume manufacturingmethodology and that capitalize upon the high integration capa-bilities of CMOS will be best able to boost yield. For example, atransceiver that improves handset yield by 1% can effectivelyshave $0.30 off the price of a $30 ULCH, which is a substantialsaving. Greater integration lowers the external component countand insulates the handset from component, circuit trace andPCB variations. If all the RF systems can be fully integrated,complete system testing can be used to ensure that all partsshipped meet stringent performance and quality requirementsbefore they become a problem for the handset manufacturer.

Overall handset affordability and network efficiency is thuscritical in driving GSM adoption by the next 1 billion users.Within this emerging ultra-low-cost handset paradigm, the RFtransceiver performance, radio solution BOM and reliability canaccelerate opportunities for growth and provide economicadvantages for both suppliers and end customers. ■

Gary B. Levy is a Product Manager for Silicon Laboratories.

Introducing the industry’s fastest, most flexible,and most cost-effective RF solutions.

RFwww.keithley.com

■ Faster test execution speed.

■ Quickly change test requirements with SDR architecture.

■ Unmatched ease-of-use with touch-screen GUI.

Visitwww.keithley.info/ad/RF

to learn more.

Also, please request your free copyof the booklet “Speaking RF: Wireless Terminology“ there.

DON’TWASTE YOUR

TIME, MONEY AND EFFORT

ON YESTERDAY’S RF SOLUTIONS

Belgium: info@keithley.nl • Phone: 02-3630040Finland: finland@keithley.com • Phone: 09-88171661

France: info@keithley.fr • Phone: 01-64532020Germany: info@keithley.de • Phone: 089-849307-40

Italy: info@keithley.it • Phone: 02-553842Netherlands: info@keithley.nl • Phone: 0183-635333

Sweden: sweden@keithley.com • 08-50904600 Switzerland: info@keithley.ch • Phone: 044-8219444

United Kingdom: info@keithley.co.uk • Phone: 0118-9297500

For all countries not listed contact the KIEX department in Germany:info@keithley.de • Phone: +49-89-84 93 07-0

The Only Available Lab Quality,Handheld RF Power Meter

RF SIGNAL GENERATOR ■ RF SIGNAL ANALYZER ■ RF POWER METER

A G R E A T E R M E A S U R E O F C O N F I D E N C E

When multiple processor coresare included on a single chip,it becomes much more diffi-cult to verify the design – andto debug the operation – ofthe entire system. With prod-uct lifetimes growing shorterand shorter, it is no longereconomically viable to wait fora system-on-a-chip (SoC) tobe fabricated before trying toget a wireless design to work.The complexity of the designensures that many bugs endup in silicon, and the limitedvisibility into the internalbehaviour of a fabricated chipmakes debugging at this stagea daunting task.

This article presents a methodfor system-level debugging of amulti-core system using a vir-tual system prototype, whichallows a cycle-accurate simulation of a complete system to exe-cute in real time on a PC. Figure 1 shows the overall architectureof an actual multi-core chip for wireless applications. As is typicalin the wireless industry, this chip uses both general-purpose centralprocessing units (CPUs) and a dedicated digital signal processor(DSP) to achieve both parallelism and specialization. It alsoincludes a multi-layered memory system and several peripheraldevices for off-chip communication.

Dual coreThis particular chip includes two ARM926E CPU cores, eachwith separate caches for instructions and data. One ARMprocessor runs the Linux operating system and is for generalapplications such as virtual machines, graphics environmentsand messaging. The second ARM is designed to work tightlywith the DSP core and to handle communications and controlof the overall system. It can be assigned special tasks and processapplications as needed, too. Having two CPUs allows a broaderrange of processing traffic to be handled in real time, a criticalrequirement for wireless applications.

The two ARM cores are complemented by a StarCoreSC1200 processor, which handles multimedia acceleration and

wireless modem processing functions as well as other DSP tasks.This core includes two independent execution units, each ofwhich can perform multiply-accumulate (MAC) operations aswell as other arithmetic functions commonly used in signal-processing algorithms. The DSP is designed to provide muchof the processing for voice data as well as to service multimediadecoding such as MP3, MPEG-4, and H.264.

As in most wireless chips, a hierarchical bus network providesdedicated data paths, reducing traffic and providing access tocommon memory blocks for communication. At the top level,six system buses connect the processors to their respective dedi-cated memory subsystem and peripherals. All of these busesuse the advanced high-speed bus (AHB) protocol, which isdefined by ARM for use with its processor cores.

Dedicated AHB buses allow simultaneous data and instruc-tion access to memory for all three processors, eliminating thebottlenecks common to multi-core devices. Bridges provide thelink from the AHB system-level buses to lower levels of the bushierarchy. Three dedicated memory buses provide access to thememory blocks, while two low-speed peripheral buses connectto timers, interrupt controllers and serial interfaces for off-chipcommunications.

1 7P R O T O T Y P I N G

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

Virtual systemshelp weed out bugs

Gabe Ahn and PrashantRao explain how virtualsystem prototypesimulations canimprove the debuggingprocess during thedevelopment of multi-core wireless chips.

ARM926EJS P1virtual processor model

I Cache D CacheStdBus I/F StdBus I/F

ARM926EJS P2virtual processor model

I Cache D CacheStdBus I/F StdBus I/F

P2 m

emor

y

shar

ed m

emor

y

P1 devices P2 devices

UART

TIMER

INTC

UART

TIMER

INTCmemoryblock

memoryblock

memoryblock

memoryblock

StdBusbridge

StdBusbridge

Arb.CtrlDRAM

StdBusbridge

StdBusbridge

console 2console 1

D CacheI CacheStdBus I/F StdBus I/F

StarCore SC1200virtual processor model

D ROM P ROM

AHBBuses

P1 m

emor

y

Fig. 1. A typical multi-processor core used in wireless systems. This design includes two ARM processors and aStarCore processor.

1 8 P R O T O T Y P I N G

Of course, there are times when the processors must commu-nicate among themselves to exchange data or control informa-tion. All such communication happens through shared memorythat is accessible from anywhere in the system. Established tech-niques such as semaphores and mailboxes provide safe passage ofinformation among the processors and peripherals.

Traditionally, chip architects and designers have attempted tofully verify and debug the chip before fabrication using soft-ware-based models. Perhaps the most common approach is theuse of instruction-set simulator (ISS) models, which executethe same binary code as the actual processor in the chip.However, ISS models are typically too slow to run a large bodyof software, and higher performance models sacrifice accuracyfor speed. The majority are only instruction-accurate, meaningthat the cycle-by-cycle behaviour of the processor interfaces maynot match that of an actual processor core.

Until recently the only alternative to ISS models was the useof cycle-accurate simulation models for processor cores. Thesemodels sacrifice speed for accuracy, running so slowly that onlysmall portions of critical code can be executed in simulation.

Ideally, multi-core wireless designers require a simulationmethod that offers very high simulation speed, full cycle-accuracy, and an integrated environment that supports precisecontrol and debug of interaction among different processors.The only known solution that meets all of these requirementsis the technology of virtual system prototypes.

A virtual system prototype is a software-simulation-based,timing-accurate, electronic systems level (ESL) model of the sys-tem. It is first used at the architectural level and then as an exe-cutable golden reference model throughout the design cycle.The prototype can contain cycle-accurate virtual processor mod-els that run the same compiled and linked target code as doesthe real hardware, which provide an accurate prediction of thesystem’s real-world behaviour. Cycle-accurate models for buses,peripherals and other portions of the hardware design can beadded so that the virtual system prototype models the completebehaviour of a multi-core wireless SoC.

Architectural explorationVirtual system prototypes enable architectural exploration byallowing hardware and software components to be mixed andmatched. Accurate measurements that model real-world behav-iour allow system architects to make accurate hardware/softwaretradeoffs early in the development process. When the initial sys-tem architectural model is built, it becomes the executable sys-tem specification that drives the concurrent development ofthe detailed hardware and software implementations.

By offering both high speed and cycle-accuracy, virtual systemprototypes play a critical role in SoC development. Virtual sys-tem prototypes can be executed on standard PC platforms, andtherefore can be easily distributed to system architects, softwareengineers and hardware designers, even within geographicallydispersed teams.

All multi-core SoC design projects must overcome severalgeneral debugging challenges. With the complex interactionsof multiple processors and peripherals, there are many commu-nication links that require deep observation and debugging inorder to ensure that high-quality software is delivered.

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

1 9P R O T O T Y P I N G

The system illustrated in fig-ure 1 is an example of a designthat relies heavily on synchro-nization through memory, andas a result suffers from com-mon pitfalls includingimproper arbitration of mem-ory accesses and unwanteddata corruption. Some othersystems use direct communica-tion through dedicated hostports, a convenient and effec-tive way of managing commu-nications. Systems using thisalternative method are proneto synchronization problems.If they are not carefullydesigned and verified the resultcan be stall conditions andeven deadlocks, which areespecially difficult to debug.

It is critical that designers areable to accurately assess allaspects of system performanceand to debug any deficiencies

so that they can be corrected by changes to the architecture orimplementation. Typical issues relate to bus bandwidth and thelatency associated with heavy traffic, both problems for real-timeapplications such as speech analysis and synthesis, in which per-formance guarantees need to be met.

The typical problems that must be debugged and fixed duringverification of a wireless SoC are listed in table 1. Virtual systemprototypes make it easier to debug these problems by providing acomprehensive verification environment with consistent proces-sor models fast enough to execute actual code. It is much easier toset up complex test scenarios than with other software-basedmethods and, because of links to third-party debuggers, mucheasier to have full visibility into the design while single-stepping.

Designing, verifying and debugging a complex, multi-core,wireless SoC is not a simple task. Poor debugger visibility intothe hardware combined with cost and schedule pressures meanthat designers must take a software-based approach.Unfortunately, traditional software techniques suffer from speedand accuracy problems, limiting the ability to test and debugcommon problems related to inter-processor synchronization,contention for shared resources and performance.

Virtual system prototypes enable early software developmentand debugging, and also provide better visibility into the SoC.As seen in an example design with two ARM CPU cores andone StarCore processor core, the ability to control, measureand debug complicated multi-core interactions is a key to pro-ject success. The payback for implementing such a solution isenormous and its benefits include effective architecture explo-ration, concurrent hardware and software development, and amuch greater chance of a production-worthy chip on the firstimplementation in silicon. ■

Gabe Ahn is at VaST Systems Technology; Prashant Rao is at StarCore.

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

Shared memory applicationsData corruption

Race conditions

Direct communication linksDeadlocks

Stalls

Starvation

Processor performanceCache

Pipeline stalls

Starvation

System bus performanceCongestion

Peripheral performanceLatency

1: Problems withverifying SoCs

2 0 P R O D U C T F O C U S

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

ANTENNA SIMULAT ION

Simulation toolsminimize RF failurein handsetsBy Peter Futter of SPEAGThe design, development and testing of com-plex RF devices such as mobile phones can bea costly and time-consuming process.Traditionally, design engineers and devicemanufacturers performed antenna designand RF integration separately. The introduc-tion of RF and specific absorption rate (SAR)standards in the mid 90s by the CTIA, ISOand other groups, has imposed significantnew restrictions on mobile-phone perfor-mance. This has led to an ongoing drive toincrease the synergy between the design andRF integration phases.

Many modern handsets operate in multi-ple bands and have multiple antennas, whichmust be accommodated within ever-shrink-ing space allocations. The resulting RF chal-lenges have reinforced the need for closersynergy between the development and prod-uction phases. In particular, it is essential toassess and predict potential device failures atan early stage prior to the production phase.

Most antenna optimization software cannotaddress transmitter design, integration andcompliance issues on its own. This is becausetraditional software suffers from two impor-tant limitations: insufficient computationalperformance and the poor handling of com-plex computer-aided design (CAD) models.

The electromagnetic (EM) simulation of adetailed mobile-phone structure covering allthe relevant operational frequency bands caninvolve many computational unknowns.Simulation times on the order of hours anddays are also required, depending on the levelof spatial detail to be resolved.

GUI restrictionsThe graphical user interface (GUI) and mod-elling engines of some simulation tools expe-rience a significant slowdown as thecomplexity and number of parts in the CADmodel increases. This can make it impossi-ble to handle models that contain more thanabout 100 CAD parts.

These limitations have forced designers touse oversimplified handset models during theantenna optimization process. Models areusually restricted to the antenna and printedcircuit board (PCB) only. More complexproblems are usually not considered for sim-ulation because of the additional computa-

tional resources needed to obtain modelswith the required accuracy.

Schmid and Partner Engineering (SPEAG)has addressed these challenges in itsSEMCAD X, EM- and thermal simulationsoftware platform. Based on the finite-difference time-domain (FDTD) computa-tional technique, the software offershardware-accelerated models for FDTD.SEMCAD can achieve simulation speedsaround 300 000 FDTD cells computed persecond (300 Mcells/s). This can be achievedusing a standard PC with an aXware accel-erator card. The aXware card allows the sim-ulation to run faster than on the latestdesktop PC. Thanks to the accelerator,SEMCAD X can run up to 20 times fasterthan other tools. In figure 1, data are shownfor two competing software products andSEMCAD X EIGER, which is the mostrecent version of SEMCAD runing on astandard PC. Also shown are results forSEMCAD X running on a PC equipped withan aXware card and on a Cluster in a Box(CIB), which is a 64-bit PC with two accel-erator cards connected in parallel.

SEMCAD X offers highly efficient routinesfor importing CAD models and SPEAG’s

QTech rendering system performs the quasireal-time processing and visualization ofhighly complex CAD derived models com-prising 10 000 or more CAD parts. As a result,the system can simulate detailed mobile-phone structures within several minutes.

As handsets become ever more complexand diverse, SEMCAD X addresses thesegrowing design complexities by bridging thecapabilities of virtual prototyping, failuresynthesis and whole multiphysics problems.

Failure synthesis involves the simulationof a range of different scenarios to assesspotential shortcomings in the handset. Forexample, a user could hold a phone in sev-eral different ways, each having a differenteffect on performance. Figure 2 shows howSEMCAD X can be used to model the effectsof different hand positions on RF perfor-mance. This exercise will soon become partof routine virtual RF failure synthesis

The ultimate aim of failure synthesis is toavoid potential problems – or even devicefailure – and ultimately to improve the over-all design of the phone. These shortcomingscan cause suboptimum RF performance ofthe device, failures in standards compliance,or even complete device failures.

Fig. 2. Modelling the effects of hand positions on RF performance will soon be part of routine virtual RF failure synthesis.

100

250200150100500minutes

48

8.2

5.2

main competitor 1

main competitor 2

SEMCAD X Eiger

SEMCAD X aXware

SEMCAD X CIB

208

Fig. 1. Runtime comparison for a transmitter-absorber benchmark simulation demonstrating that hardware-acceleratedFDTD SEMCAD X can achieve simulation speeds about 10 times faster than those possible on the latest desktop PCs.

2 1P R O D U C T F O C U S

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

These defects may occur for various rea-sons. Handsets are produced in large volumesand manufacturers will often use more thanone supplier for a specific component. Thiscan result in variations in the size, structureand composition of components and thesedifferences can lead to defects. Other defectssuch as poor contacts, short circuits orgrounding problems can be introduced dur-ing the manufacturing process. Some faultsare related to the use of new materials suchas metal coatings, which are used for shield-ing components, or composite dielectric–metal materials that are used to give hand-sets a metallic look and feel.

Three-step processThe handset-simulation process involvesthree steps. The first step is to import CADdata into the EM tool’s modelling environ-ment. Based on OpenGL, SPEAG’s QTechrendering engine allows the rapid processingof more than 10 000 CAD parts. This elimi-nates the need for the time-consuming sim-plification of large CAD models.

The generation and verification of the EMmodel is the second step. Creating the correctEM model – in terms of structure, materialparameters, grid resolutions etc – is especiallychallenging when dealing with many CADparts. SEMCAD X employs an automatedgrid generator and voxel creator for obtainingreliable and effective grids for the most com-plex of models.

To be meaningful, virtual failure analysismust be based on a numerical model that isan accurate representation of the physicalmodel. It is therefore imperative that thenumerical model is validated using the appro-priate performance parameters. These includeimpedance bandwidth, radiation perfor-mance, near-field scans and far-field radiationpatterns, and data from SAR measurements.

The third and final step involves failuresynthesis and begins after the numericalmodel has been validated. One approach is todefine a failure synthesis matrix that containsa range of different scenarios to be simulatedand certain significant parameters to berecorded. A failure synthesis matrix is usedfor assessing the effects of potential manufac-turing defects and variations. These includethe effects of different materials, ground con-nections, shields and mounting tolerances.A scripting engine based on the Python pro-gramming language was used to create asemi-automatic system for simulating theeffect of these parameters on handset perfor-mance behaviour. This type of virtual testingoffers a much more cost-attractive way ofperforming failure analysis than physical test-ing and can be conducted much earlier in the

handset development process. The next-generation CAD environment is

designed to interact with the manufacturingprocess by allowing virtual prototyping ofnew technology efficiently. SEMCAD Xoffers the most advanced, reliable and fastestinteractive design environment for assessingvirtual failure analysis by integrating optimalperformance features such as hardware accel-eration for high computation speeds, specialdedicated 3D rendering techniques and the

Python-based automation/scripting engine.Upcoming new features such as an effectivegenetic algorithm based on automated opti-mization algorithms will offer even more reli-able, cost and time-efficient solutions invirtual prototyping. The next version ofSEMCAD X, V11.0 Edition Moench, sched-uled to be released later this year, will providea special feature that will further simplify thevirtual failure analysis process.www.semcad.com

ANTENNA SIMULAT ION

www.andrew.com

Andrew offers industry leading solutions.Andrew Corporation is a one-stop, end-to-end RF systems and solutionssupplier with global resources. Our product portfolio encompassestraditional wireless networks, third-generation technologies, Internetservices, and specialized applications in microwave, satellite, radar,and HF communications. Our passion for innovation, commitment toquality, and dedicated customer service shows in everything we do.

RF Site Wireless Network Satellite CoverageSolutions Infrastructure Solutions Communications and Capacity

CLEARADVANTAGES

FROM ANDREW

CLEARADVANTAGES

FROM ANDREW

EMEA’s Premier Telecoms Billing And Customer Care Event Is Back …

And With A New Look!

London, Olympia 2 • June 5-9 2006Attend this event to access all the information you require to achieve

systems and strategy excellence for:

� Outstanding IT, Billing, Charging and Rating � Competitive Business and Operational Support� World Class Information and Customer Management � Best Practice Revenue and Margin Management

For full details of the conferenceprogramme and exhibition highlights visitwww.billingsystems2006.com today!

From maximising the real ROI of fixed-mobile solutions to developing effective propositions to increase customer retention, Fixed-Mobile Convergence 2006 brings you exciting case studies from high-level operators.

For more information visit www.fixedmobile-events.com, e-mail events@vibevents.com or call us today on +44 (0) 20 7753 4268 quoting reference WEA to book your place at this cutting-edge event.

FIXED-MOBILE CONVERGENCE Wednesday 7th and Thursday 8th June 2006, Central London, UK

www.fixedmobile-events.com

Endorsing Organisation:

2 0 0 6

Media Partner:

The Second

20-22 June 2006Crowne PlazaPragueCzech Republic

Achieving Cost Effective Implementation Of HSDPA and HSUPA ForDelivering Real Time Applications With End To End QoS

Case StudyPresentations andPanels From:

Part Of The

Over 130 delegates at September 2005 event

www.informatm.com/hsxpaREGISTRATION HOTLINE +44 (0)20 7017 5506

Leading International Speakers will cover:

• Evaluating HSUPA System Performance

• Deploying VoIP On HSxPA and EvaluatingIts Demand

• HSUPA Business Case And TechnicalImplementation

• Maximising HSxPA Potential WithStreaming And Office Based Applications

• HSxPA QoS Measurement, Testing AndOptimisation

• Marketing HSxPA Applications To ReachThe Corporate And Residential Market

• Examining The Impact Of HSDPA/HSUPAOn Network Dimensioning

ORANGEMOBILKOMFRANCE TELECOM R&DINTERDIGITALMOBTELNORTELALCATELO2ICERALGOPTIONQUALCOMMAIRCOMVODAFONETELFORT

24th - 27th April 2006 • Sana Lisboa Hotel, Lisbon

Learn how to accurately measure and optimiseend-to-end network performance, service qualityand the customer experience amid the increasein network traffic and service complexity

◆ Examine the latest tools and techniques that are being used to ensure accurate performancemeasurements across evolving networks and services

◆ See how to carry out cost-effective performancemonitoring in your next generation networks andjustify the investment required

◆ Hear how to measure and optimise end-to-end network performance, service quality and customer satisfaction

www.iir-conferences.com/performance

2 3P R O D U C T F O C U S

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

APLAC engine powersAnalog OfficeBy Hamish JohnstonA revamped layout editor and a new tool formodelling chip interconnects are key featuresof Analog Office 2006 from Applied WaveResearch (AWR). Aimed at designers ofradio-frequency integrated circuits (RFIC),this is the first version of the software to bereleased after AWR’s acquisition of APLACof Finland in September 2005. WirelessEurope caught up with AWR’s Chris Paris tofind out how APLAC’s considerable RFdesign experience is being used by AWR.

“We have been staggered at how goodAPLAC’s harmonic balance engine is, espe-cially for large problems with more than1000 active devices,” said Paris, adding thatthe addition of the APLAC software makesAnalog Office a very powerful platform fordesigning RFICs, which can integrate thou-sands of devices on one chip. “The APLACengine is probably the only one that will con-verge when you get up to tens of thousandsof transistors,” he explained. “APLAC wasfunded very well by Nokia and the Finnishgovernment for many years. About 90 man-years of engineering have gone into theAPLAC simulator and some of the algo-rithms are absolutely unique.”

As part of the APLAC deal, AWR alsoacquired foundry-approved process designkits (PDKs) for several leading Europeanchipmakers including STMicroelectronics,Infineon and Philips. “This is something thatwe did not have at AWR,” said Paris. “TheSDKs are very important because it doesn’tmatter how good the design platform is if themodels are of no use.”

APLAC was spun out of Nokia in 1998and retains strong ties with the world’slargest mobile-phone maker. Indeed, AWRsays that APLAC products have been used todesign more than 30% of mobile-phoneRFICs worldwide. Buying APLAC is strate-gically important for AWR. The company isa relative newcomer to the RFIC designmarket, having launched Analog Office in2003. “AWR came from a microwave back-ground,” explained Paris. An important dif-ference between monolithic microwaveintegrated circuits (MMICs) and RFICs isthat the latter have many more transistors.As a result RFIC design requires a muchmore sophisticated layout editor. “This issomething that we spent a lot of time work-ing on,” said Paris.

AWR has always been perceived as a com-

pany that supplies “front-end” simulationtools, according to Paris. AWR users wouldusually transfer their simulation results to“back-end” design tools produced byCadence, Mentor or Synopsys. “We haveadded a number of features that you wouldnormally see in these back-end tools. Wehave industry-leading simulation technologyat the front end, and now we have a back-endlayout tool that allows all RF design workwithin one tool.”

Three-month successAccording to Paris, a major Japanese trans-ceiver manufacturer used this new feature todesign an RFIC in three months. The designwas successful on the first pass and the chipmet all design specifications once manufac-tured. “We believe that this represents a sub-stantial improvement over what anyone elsehas achieved – first-pass success in threemonths is almost unheard of.”

As a relatively new software system AnalogOffice has taken advantage of modernobject-oriented database technologies.“Other products use separate databases forlayout, schematic, simulation and systemdesign, whereas we have just one object-oriented database,” said Paris. While a com-ponent such as a capacitor only appears oncein the database, it displays different attributesto the various tools that comprise AnalogOffice. For example, a capacitor wouldappear as a collection of polygons whenviewed by the layout tool, whereas the simu-lator would see the S-parameters associatedwith the device, and the schematic tool

would see a capacitor symbol. “This is anabsolutely unique feature and means that theoverall design process takes less time.”

Analog Office 2006 also includes iNet2,which is an updated version of AWRs iNetmethodology for the modelling and analysisof RF interconnects. “There are many dif-ferent ways that you can model the connec-tion between the two elements on an RFIC,”explained Paris. These range from a simpleshort circuit with no RF properties to ahighly complex 3D electromagnetic model.According to Paris, iNet2 allows designers tochange interconnect models “on the fly”. Ateach stage of the design process, designerscan choose the level of interconnect com-plexity based on the desired accuracy.

For example, an electromagnetic modelmay be chosen for a connection between twoelements where a significant degree of crosstalk could occur. A line connecting two RFtransistors could be modelled as a transmis-sion line if it operates at high frequencies,whereas a simple lumped element would beappropriate if operation is at low frequencies.

Paris describes Analog Office as an “opendesign platform that can link to software frommost leading providers”. He believes that thisis an important feature given the currentdebate regarding which EM simulator is thebest for the job at hand. “Some designers usethe Sonnet 2.5D simulator exclusively, whileothers favour Ansoft’s 3D HFSS over both2.5D and 2D simulators. We provide ourusers with the option to link into six or sevenEM tools as well as our own.”www.appwave.com

RFIC DES IGN

AWR’s Analog Office 2006 features a new layout editor and an updated system for modelling interconnects.

2 4 P R O D U C T S

FLO/PCB links to AllegroFlomerics has released version 3.0 of itsFLO/PCB thermal design software. FLO/PCB for Allegro includes a bidirectionallink to Cadence’s Allegro PCB Editorsoftware. According to Flomerics, designerscan call up a menu item on Allegro andquickly generate a thermal model of theirdesign. This can be analysed and problemsidentified early in the design process.

The FLO/PCB/Allegro interface transfersrelevant information regarding PCBgeometry and components. This includesthe number of metallic layers, the functionof each layer, the coverage of copper on theboard and the location and powerdissipation of each component. Theinterface also allows the user to select thelayer that defines the size of the package.Placement updates made in FLO/PCB canbe passed back to the Allegro PCB Editor. www.f lomerics.com

CST Microwave Studioconnects with layout toolsThe latest release of CST Microwave Studioallows users to import multilayer structures inGerber format from Cadence, Mentor, Zukenand other leading EDA tools. According toComputer Simulation Technology (CST),complete 3D models can be studied withtransient or frequency-domain full-waveanalysis. This can provide important insightsinto cross-talk, radiated emission and signaland power integrity. Users also benefit fromfast and efficient full 3D simulation ofmultilayer structures and the automatedcreation of full 3D models. There is no loss oflayout details during import.www.cst .com

HSPA library includes uplinkA high-speed packet access (HSPA) wirelesslibrary is now available for use with theAdvanced Design System (ADS) electronicdesign software from Agilent Technologies.This first phase of the HSPA librarycontains a set of preconfigured test benchesand sources for the development of high-speed uplink packet data access (HSUPA)equipment as defined in Release 6 of the3GPP specifications. The second phase ofthe library release is expected in the thirdquarter of 2006 and will cover high-speeddownlink packet access (HSDPA).

The HSPA wireless library providespreconfigured simulation setups, signalsources and fully coded bit error rate

analysis for simulation of the circuitry usedin 3GPP fast uplink designs. It helps speedthe development cycle by allowing systemdesigners to analyse a system’s performancebefore all of its components are designed.The HSPA wireless library works within theADS environment and with Agilent’sPtolemy system-level simulator. The librarysupports technologies specific to HSPAincluding the short transmission timeinterval (TTI), hybrid automatic repeatrequest (HARQ) and uplink data rates of upto 5.8 Mbit/s.eesof . tm.agi lent .com

Multiphysics modelling canhandle third-order elements Significant performance enhancements andthe ability to address new classes ofproblems are highlights of version 3.2a ofCOMSOL Multiphysics. This latest versioncan run on PCs equipped with 64-bitWindows XP Professional, which canprovide as much as 128 Gbyte of RAM and

16 Tbyte of virtual memory. This increasesboth the size of models that the softwarecan handle and the number of problemsthat can be addressed using a direct solver,rather than an iterative solver.

COMSOL has also improved thepackage’s Electromagnetics Module, whichis an add-on to the basic software.Previously this module worked only withfirst-order vector elements when creatingthe polynomial that forms the basisfunction that mathematically describes theproblem. This latest version adds second-and third-order vector elements on any typeof mesh including unstructured, mapped,extruded and revolved. These complexfunctions allow more accurateinterpolations and extrapolations among thefinite elements, resulting in higher accuracyand faster convergence. According toCOMSOL, the solvers can achieve an orderof magnitude improvement in performanceover previous versions.www.comsol .com

CLR library is pad-scalableThe CLR Library 4.0 from Modelithicsnow includes pad scalability on selectedcomponents. The Modelithics GlobalModel has input parameters that allow usersto specify custom pad dimensions to meettheir design requirements.

Pad-scalable models added so far includeModelithics models CAP-ATC-0402-001,CAP-JOH-0201-001, IND-TKO-0402-001and RES-KOA-0805-001.The 4.0 versionalso includes 13 new models from the

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

DESIGN & SIMULAT ION

2 5P R O D U C T S

following vendors: Dielectric Labs, ATC,AVX, TDK and Johanson Dielectrics.

Modelithics has also announced that itwill distribute and support its active andpassive models for the Eagleware/AgilentGENESYS and HARBEC simulators.Previously, the GENESYS and HARBECmodels were distributed by Eagleware,which was acquired by Agilent in September2005. Modelithics will continue todistribute and support models for theAgilent ADS software. www.model i th ics.com

Designer hits version 3Ansoft has launched a new version of itsDesigner software. Ansoft Designer v3offers an expanded version of the company’sSolver on Demand technology and a newdesign-verification system. According toAnsoft these new features allow the softwareto accelerate the design process of RFsystems and RF/analogue integrated circuitsas well as increasing the accuracy of signal-integrity studies.

The Solver on Demand technology hasbeen expanded to include Ansoft’s HFSS3D full-wave electromagnetic-fieldsimulation software. The simulation ofdesigns with large transistor counts and/orhigh spectral content can now be done by aco-simulation involving HFSS and Nexxim,which is Ansoft’s high-capacity, transientand harmonic balance, circuit-simulationsoftware.

Version 3 has new features for designverification including a user-defined designrule checking (DRC) system, whichaddresses the design for manufacturing(DFM) requirements. Designer v3 operateson PCs running Windows 2000, XP andServer 2003 Standard Edition.

Ansoft has also released a new devicelibrary that describes high-performancecompound-semiconductor transistortechnologies from NEC. When used withAnsoft’s Nexxim and Ansoft Designersimulation software, the library will supportthe design and development of low-noise,high-gain radio frequency components andintegrated circuits.

The library includes the followingnonlinear device models for Nexxim:Gummel-Poon model for low-noise bipolartransistors; Gummel-Poon model forheterojunction bipolar transistors (HBTs);and the EEHEMT1 for heterojunction fieldeffect transistors (HJ-FETs).

The library also includes severalS-parameter models and device footprintsincluding wide-band amplifier, dual-gate

MES/MOS FET, low-noise bipolartransistor and SiGeHBT twin transistor.www.ansoft .com

Cluster computing slashesEM analysis timesSonnet Software has updated Release 10 ofits Sonnet Suites Professional high-frequency electromagnetic (EM) software.Release 10.53 includes the new emClusterComputing option, which Sonnet sayssignificantly reduces the EM analysis timesfrom days to hours.

Version 10.53 also improves uponprevious thick-metal models, whichincreases its accuracy over a wider frequencyrange. According to Sonnet, this featuredelivers greater accuracy for loss at lowfrequencies. The efficiency of the sub-sectioning algorithm has also beenimproved.www.sonnetsoftware.com

Wizard offers 3D simulationMician will soon release Version 6.0 of itsµWave Wizard electromagnetic softwaretool. The new version will feature a new 3Dfinite-element method (FEM) simulationtechnology in addition to the software’sexisting mode-matching and 2D FEMcapabilities.

According to Mician, Version 6.0 canachieve full 3D visualization ofelectromagnetic fields and calculate lossesarising from finite wall conductivity. It alsofunctions as an integrated eigenmode solver.The package has new and flexible libraryelements that aid the modelling ofarbitrarily shaped objects. A fully integratedformula editor is included and the elementeditor has been extended to manageelements with complex properties in anintuitive manner.

Version 6.0 of µWave Wizard will bereleased by the end of May 2006. www.mician.com

w i r e l e s s e u r o p e wire less. iop.org A p r i l / M a y 2 0 0 6

DESIGN & SIMULAT ION

2 6 T H E F U T U R E

Why should TV be delivered via TD-CDMA?In mid-2005 the chief technology officers(CTOs) of at least five leading operators toldus that they see mobile TV as the next big ser-vice. However, for strategic reasons CTOs arenot happy with the technologies that are cur-rently available – namely DVB-H and possiblyMediaFLO. Neither of these technologies usesexisting mobile radio spectra and this poses seri-ous difficulties for operators. For example, amajor operator wishing to launch TV serviceson 10 networks across Europe could be forcedto partner with a different broadcaster in eachcountry. Operators could acquire their ownDVB-H spectrum, but this could be problem-atic and the last thing most operators want todo is to buy more spectrum.

The overall feeling we get from operators is that they want tokeep mobile TV on their 3G spectrum. They also want perfor-mance that is as good as DVB-H because ultimately they mayhave to compete with DVB-H services.

How has IPWireless responded to this opportunity?Operators have asked us if we could add something to ourTD-CDMA platform that would allow them to offer mobileTV services using their unpaired 3G spectrum. Over the lastnine months we have been working with operators in the itera-tive process of defining TDtv technology.

While this work was underway, Release 6 of the UMTS stan-dard was finalized and included the definition of multimediabroadcast and multicast services (MBMS). This is a very detailedspecification and includes definitions for user authenticationand data encryption. However, the most important thing aboutMBMS is that it introduces a point-to-multipoint channel,which allows an infinite number of users to access a service. Thisis unlike current 3G unicast services, which must devote sepa-rate bandwidth to each individual user.

How will MBMS and TD-CDMA be used to deliver mobile TV?These technologies can be used to turn 3G into a broadcast sys-tem. TDtv devotes the entire unpaired band to the downlink,which allows for the full utilization of available spectrum. Allof the interactive elements of the service are done over theW-CDMA paired bands. A key benefit of this approach is thatall of the interactivity is defined within the 3G standard.

We have worked with a company called Vidiator, whichdevelops server technologies for streaming services. The aim isto integrate Vidiator’s new Broadcast Multicast Service Center(BMSC) into our TDtv solution. The BMSC is a box that sits

in the 3G network and supports all of thisinteractivity.

How much will TDtv cost to roll out?Because TDtv is a downlink-only technology, itcan be rolled out as a low-cost upgrade to a 3Gnetwork. Operators have told us that costsshould not exceed 610 000 per cell-site. Ibelieve that TDtv can be rolled out effectivelyby upgrading every second or third 3G site.Operators will be able to take advantage of verylow cost repeaters and other technologies.

Will TDtv deliver mobility and indoor coverage?TDtv supports mobility and can be used onhigh-speed trains travelling at speeds up to220 km/h. Later this year we expect to take it

up to 300 km/h. DVB-H is not as sophisticated in handlinghigh-speed movement and I understand that it runs into prob-lems above about 120 km/h.

The TDtv receiver performs point-to-multipoint combiningof signals from multiple cell sites and the signals from up to sixcell sites can be processed at any given time. Indoor coveragecomes down to the quality of the network – both the technologyand how it is deployed. We have recently announced a trial ofTDtv with Orange in the UK and we plan to work throughmany of the coverage issues. However, we are confident thatDVB-H and TDtv will deliver similar levels of in-building cov-erage for about the same total cost of ownership.

What sort of handsets will receive TDtv? Early devices will either be smartphones or PDAs with SDIOcards. These cards cost about $100 and therefore this is not acommercial solution but rather a phase-one technology for mar-keting trials. The second phase will involve a two-chip solu-tion, which will be very similar to what is used for DVB-H.Today’s DVB-H phones have two radios, which are combinedthrough an applications processor. We expect that three ven-dors will have two-chip TDtv solutions on the market in the sec-ond quarter of 2007.

TD-CDMA and W-CDMA standards are identical at all levelsexcept at the air interface. As a result TD-CDMA can make use ofmuch of the existing software on a W-CDMA chip. TD-CDMAwill ultimately become a layer technology on the W-CDMA chipand we have already partnered with a chipmaker to this end. Ourgoal is to add TD-SCDMA capability while increasing the costof any existing W-CDMA chip by less than $10. ■

Interview by Hamish Johnston, editor of Wireless Europe.

A p r i l / M a y 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

TD-CDMA delivers mobile TVJon Hambidge of IPWireless explains the advantages of

delivering mobile television services via unpaired 3G spectra.

Jon Hambidge: “Operators want tokeep mobile TV on their 3G spectrum.”

From SMS to MMS, ftp to http and video, the Agilent 8960

offers realistic network simulation that puts your products

to the truest test. Helping you develop better devices and

slashing your time to market.

With the 8960 on your bench, you can complete protocol

stack development and integrate the stack with cellular

device hardware faster and more efficiently. You can also

test Internet connectivity with real data traffic flows.

And perform development testing without ever having to

travel to a real network!

What's more it opens up a real-time environment with

standards based measurement integrity for all the latest

3GPP and 3GPP2 technologies.

The 8960 test set. Real testing for real results: real

results, real fast.

Get realGet realNetwork emulation has never been more realisticNetwork emulation has never been more realistic

© Agilent Technologies, Inc. 2006

For more information, please go to:www.agilent.com/find/8960devicedesign

Agilent 8960 + Lab Applications• Tests HSDPA, 1xEV-DO, all 3GPP and 3GPP2 technologies• Offers full-featured messaging capability• Allows real-time protocol logging and analysis

Agilent J7830A Signaling Analyser Real-time• Real time monitoring and troubleshooting of mobile networks• 2G, 2.5G and 3G• UMTS, HSDPA, SS7, CDMA, UTRAN, GERAN, Core and much more.

http://www.semcad.com