i 
A STUDY OF THE VIABILITY OF RF MEMS FOR THE IMPLEMENTATION OF A 
SINGLE-CHIP PROGRAMMABLE RF FRONT END FOR SOFTWARE DEFINED 
RADIO FEMTO CELLS 
 
Except where reference is made to the work of others, the work described in this 
dissertation is my own or was done in collaboration with my advisory committee. This 
dissertation does not include proprietary or classified information. 
 
__________________________ 
Jorge Sigfrido Hurtarte 
 
 
 
 
Certificate of Approval: 
 
 
 
 
________________________________         ________________________________ 
Prathima Agrawal             Richard C. Jaeger, Chair 
Ginn Distinguished Professor         Distinguished University Professor 
Electrical and Computer Engineering          Professor Emeritus  
        Electrical and Computer Engineering 
 
 
 
 
 
________________________________         ________________________________ 
Paul M Swamidass             George T. Flowers 
Professor, Aviation and Supply Chain         Dean 
Management; College of Business          Graduate School 
   
                           
 
 ii 
 
A STUDY OF THE VIABILITY OF RF MEMS FOR THE IMPLEMENTATION OF A 
SINGLE-CHIP PROGRAMMABLE RF FRONT END FOR SOFTWARE DEFINED 
RADIO FEMTO CELLS 
 
Jorge Sigfrido Hurtarte 
 
 
A Dissertation 
Submitted to 
the Graduate Faculty of 
Auburn University 
in Partial Fulfillment of the 
Requirements for the 
Degree of 
Doctor of Philosophy 
 
 
 
Auburn, Alabama 
May 9, 2009 
 
                           
 
 iii 
A STUDY OF THE VIABILITY OF RF MEMS FOR THE IMPLEMENTATION OF A 
SINGLE-CHIP PROGRAMMABLE RF FRONT END FOR SOFTWARE DEFINED 
RADIO FEMTO CELLS 
 
 
 
 
 
Permission is granted to Auburn University to make copies of this dissertation at its 
discretion, upon request of individuals or institutions and at their expense.  The author 
reserves all publication rights. 
 
 
 
 
 
 
______________________________ 
Signature of Author 
 
 
 
______________________________ 
Date of Graduation 
                           
 
 iv 
DISSERTATION ABSTRACT 
 
A STUDY OF THE VIABILITY OF RF MEMS FOR THE IMPLEMENTATION OF A 
SINGLE-CHIP PROGRAMMABLE RF FRONT END FOR SOFTWARE DEFINED 
RADIO FEMTO CELLS 
 
Jorge Sigfrido Hurtarte 
 
Doctor of Philosophy, May 9, 2009  
(M.B.A., Auburn University, 2002) 
(B.S., University of California, Irvine, 1983) 
 
 186 Typed Pages  
Directed by Richard C. Jaeger 
 
The RF front end circuitry makes up over 60% of the cost of a base transceiver 
station (BTS, also referred as Node B in 3G networks).  Included in this cost is that of the 
Power Amplifier unit, RF filters, and the GPS unit for clock recovery and 
synchronization.  Because of the evolving mobile wireless communications standards 
(2G/2.5G/3G/LTE/4G and their various implementation standards including CDMA2000, 
WCDMA, TD-SCDMA, WiMAX, etc), the BTS original equipment manufacturer 
(OEM) is faced with the monumental task of designing a BTS RF Front End for each 
mobile network standard that they wish to support.  Moreover, the OEMs are forced to 
manufacture and maintain in the field all of these BTS variants. 
                           
 
 v 
One way to address the frequency re-use and power range limitations of today?s 
base stations is to have more base stations covering smaller geographical areas by 
implementing what is called a femtocell, which also implies a significant cost reduction 
of the RF front end.  Thus the RF front end represents the single biggest opportunity for 
cost reduction and logistics simplification for the base station of the future: the software-
defined radio based cellular femtocell.   The problem is that today?s BTS RF front end 
implementations are not programmable but fixed to work with a specific RF frequency 
(e.g., 1,900 MHz) for a specific mobile network standard (e.g., CDM2000 or W-CDMA).  
To band-aid this problem, the OEMs typically implement multiple RF front ends to give 
the illusion of a ?multi-band? programmable system at twice, triple or quadruple the cost 
of a single-band system. 
RF MEMS (micro-electro-mechanical systems) is a technology that has the 
potential to not only help solve the BTS cost and logistic issues, but also to complement 
SDR architectures by providing a true low cost, high performance, programmable RF 
front end.   Because of its potential in the implementation of low cost, high performance, 
programmable RF front end filters and antennas, the author is investigating the possibility 
of implementing a single-chip programmable RF MEMS front end for SDR-based 
cellular pico and femto base stations for commercialization by 2015 or earlier.   This 
dissertation summarizes the author?s findings, conclusions, and future work 
recommendations. 
                           
 
 vi 
ACKNOWLEDGMENTS 
 
 
 
The Author would like to thank his advisor Dr. Richard C. Jaeger for his support 
and advice throughout the author?s Ph.D program.    The author would also like to thank 
the members of his advisory committee, Dr. Prathima Agrawal and Dr. Paul Swamidass 
for their support and feedback on the dissertation.    In addition, the author would like to 
thank Dr. Santanu Das, President and CEO of TranSwitch Corporation for his relentless 
support to academia and in particular for his flexibility in allowing the author to pursue 
his Ph.D studies while also working at TranSwitch. 
The author would like to also thank the Vodafone-US Foundation for their 
Fellowship and financial support. 
                           
 
 vii 
Style manual or journal used:  Guide to Preparation and Submission of Theses 
and Dissertations. 
Computer software used: Microsoft? Office Word 2003. 
                           
 
 viii 
 TABLE OF CONTENTS 
 
LIST OF FIGURES .......................................................................................................... xii 
Chapter 1 The State of the Mobile Wireless Communications Industry..................... 1 
1.1 General Overview ............................................................................................... 1 
1.2 China................................................................................................................... 7 
1.3 Japan ................................................................................................................... 9 
1.4 South Korea ...................................................................................................... 13 
1.5 India .................................................................................................................. 15 
1.6 Europe............................................................................................................... 18 
1.7 North America .................................................................................................. 20 
1.7.1 Overview................................................................................................... 20 
1.7.2 Femtocells................................................................................................. 21 
1.8 Chapter 1 Summary .......................................................................................... 23 
Chapter 2 The Mobile Wireless RF Spectrum Allocation......................................... 24 
2.1 Introduction....................................................................................................... 24 
2.2 Data Rate Requirements ................................................................................... 24 
2.3 Spectrum Issues ................................................................................................ 27 
2.4 New Spectrum................................................................................................... 29 
2.5 Chapter 2 Summary .......................................................................................... 34 
Chapter 3 The Mobile Wireless Channel Multiple Access Techniques.................... 35 
3.1 Introduction....................................................................................................... 35 
3.2 Overview of Multiple Access Protocols ........................................................... 35 
3.3 OFDM-based Hybrid Protocols ........................................................................ 38 
3.4 Chapter 3 Summary .......................................................................................... 41 
Chapter 4 Quality of Service Requirements.............................................................. 42 
4.1 Quality of Service (QoS) Attributes ................................................................. 42 
4.2 QoS Classes ...................................................................................................... 43 
4.3 QoS Architecture .............................................................................................. 44 
4.4 Chapter 4 Summary .......................................................................................... 47 
Chapter 5 Trends in Base Station Architectures........................................................ 48 
5.1 Introduction....................................................................................................... 48 
5.2 Distributed Base Stations.................................................................................. 50 
5.2.1 CPRI and OBSAI Architectures ............................................................... 51 
5.2.1.1 CPRI...................................................................................................... 52 
5.2.1.2 OBSAI................................................................................................... 53 
5.3 Chapter 5 Summary .......................................................................................... 55 
                           
 
 ix
Chapter 6 4G Requirements ...................................................................................... 56 
6.1 4G Definition .................................................................................................... 56 
6.2 4G Requirements .............................................................................................. 56 
6.3 Backhaul Interworking...................................................................................... 58 
6.4 Flexible Platforms Requirement ....................................................................... 60 
6.5 4G Handsets...................................................................................................... 61 
6.5.1 Hardware................................................................................................... 62 
6.5.2 Software .................................................................................................... 62 
6.6 Chapter 6 Summary .......................................................................................... 63 
Chapter 7 Technologies and Research Aimed at Addressing the 4G Requirements. 64 
7.1 Wireless Transmission Techniques................................................................... 64 
7.2 Network Technologies...................................................................................... 65 
7.3 Mobile Terminal Techniques............................................................................ 65 
7.4 Mobile Systems Technologies. ......................................................................... 66 
7.5 Hardware Technology Roadmap ...................................................................... 67 
7.5.1 Modems..................................................................................................... 67 
7.5.2 Radio Technology..................................................................................... 69 
7.6 Chapter 7 Summary .......................................................................................... 70 
Chapter 8 RF Front End Technology Requirements and Challenges........................ 71 
8.1 The Vision for the RF Front End of the Future ................................................ 71 
8.2 Cost Considerations .......................................................................................... 73 
8.3 Vanu, Inc........................................................................................................... 76 
8.4 Chapter 8 Summary .......................................................................................... 80 
Chapter 9 Current Technology Limitations of Software Defined Radios................. 81 
9.1 SDR Defined..................................................................................................... 81 
9.2 Current SDR Limitations .................................................................................. 82 
9.2.1 Antenna..................................................................................................... 84 
9.2.2 Circulator or Duplexer. ............................................................................. 85 
9.2.3 A/D Converter........................................................................................... 85 
9.2.4 D/A Converter........................................................................................... 86 
9.2.5 Receiver Anti-alias Filtering..................................................................... 86 
9.2.6 DSPs and Equivalent Technologies .......................................................... 87 
9.2.7 RF Power Amplifiers ................................................................................ 87 
9.3 Chapter 9 Summary .......................................................................................... 87 
Chapter 10 Multi-Band Flexible (MBF) Transceiver Architectures ........................... 89 
10.1 Introduction....................................................................................................... 89 
10.2 Receiver Single-Carrier Designs....................................................................... 90 
10.2.1 Analog Quadrature Receiver Design ........................................................ 90 
10.2.2 Digital IF Receiver Architecture............................................................... 92 
10.3 Multi-Carrier (Multi-band) Receiver Design.................................................... 93 
10.4 Flexible Transmitter Architectures ................................................................... 95 
10.5 The Problem of the Diplexer............................................................................. 98 
10.5.1 The Isolation Issue .................................................................................... 99 
10.5.2 Potential Solutions to the Isolation Issue................................................ 103 
10.6 Chapter 10 Summary ...................................................................................... 106 
                           
 
 x
Chapter 11 The Current State of RF MEMS Technology......................................... 109 
11.1 Introduction..................................................................................................... 109 
11.2 ?More than Moore?......................................................................................... 110 
11.3 The State of RF MEMS Technology .............................................................. 110 
11.4 Switches .......................................................................................................... 111 
11.5 Applications .................................................................................................... 115 
11.6 Reliability........................................................................................................ 118 
11.7 Packaging........................................................................................................ 119 
11.8 Varactors......................................................................................................... 122 
11.9 Inductors ......................................................................................................... 125 
11.10 RF MEMS and CMOS Integration ............................................................. 128 
11.11 Chapter 11 Summary .................................................................................. 128 
Chapter 12 A MBF Transceiver Architecture with RF MEMS................................. 129 
12.1 The Ultimate Goal........................................................................................... 129 
12.2 A Single Chip Front-Haul Processor Architecture ......................................... 130 
12.3 A  Five-Year Horizon Possibility ................................................................... 134 
12.4 Chapter 12 Summary ...................................................................................... 137 
Chapter 13 Femtocell Market Overview ................................................................... 139 
13.1 Femtocell Market Trends................................................................................ 139 
13.2 Market Size and Forecast................................................................................ 143 
13.2.1 Overview................................................................................................. 143 
13.2.2 GSM / GPRS Femtocells. ....................................................................... 144 
13.2.3 CDMA Femtocells.................................................................................. 144 
13.2.4 W-CDMA Femtocells............................................................................. 145 
13.2.5 CDMA2000 Femtocells.......................................................................... 145 
13.3 Market Growth Potential................................................................................. 145 
13.4 Market Size by Geographies of the World ..................................................... 147 
13.5 Femtocell Equipment Manufacturers.............................................................. 149 
13.6 Femtocell Adoption Roadmap ........................................................................ 150 
13.7 Chapter 13 Summary ...................................................................................... 150 
Chapter 14 Summary and Future Work..................................................................... 152 
14.1 Summary......................................................................................................... 152 
14.2 Future Work.................................................................................................... 154 
14.2.1 A New Venture ....................................................................................... 154 
14.2.1.1 Market............................................................................................. 155 
14.2.1.2 Financing......................................................................................... 155 
14.2.1.3 Management Team.......................................................................... 155 
14.2.1.4 Product ............................................................................................ 156 
14.2.1.5 Unique features ............................................................................... 157 
14.2.1.6 Trademarks, Patents, Copyrights, Licenses, and Royalties ............ 158 
14.2.1.7 Governmental approvals ................................................................. 158 
14.2.1.8 Production....................................................................................... 158 
14.2.1.9 Suppliers ......................................................................................... 159 
14.2.2 Future Research and Development ......................................................... 159 
14.3 Business Case.................................................................................................. 160 
                           
 
 xi
References....................................................................................................................... 161 
Appendix A Wireless Communications Technologies................................................ 163 
Appendix B TD-SCDMA Standard ............................................................................ 165 
Appendix C Glossary of Terms................................................................................... 166 
Appendix D Preliminary Financial Plan and Sales Projections .................................. 168 
D.1 General Assumptions:..................................................................................... 168 
D.2 Specific Assumptions...................................................................................... 168 
D.3 Financial Tables.............................................................................................. 170 
                           
 
 xii
LIST OF FIGURES 
Figure 1.1 Various Upgrade Paths for 2G to 3G Technologies ....................................... 2 
Figure 1.2 WiMax Network ............................................................................................. 3 
Figure 1.3 Cellular versus Mobile WiMAX..................................................................... 4 
Figure 1.4 Worldwide Mobile Subscribers by Technology ............................................. 5 
Figure 1.5 Top Mobile Operators around the World........................................................ 5 
Figure 1.6 Mobile Subscribers by Geographical Region ................................................. 6 
Figure 1.7 Mobile Subscribers by Asia Pacific Sub Region ............................................ 7 
Figure 1.8 Wireless Subscribers in China ........................................................................ 9 
Figure 1.9 3G Data Access Speeds ................................................................................ 10 
Figure 1.10 NTT DoCoMo?s i-mode Services Evolution ................................................ 12 
Figure 1.11 India?s Wireless Subscribers......................................................................... 16 
Figure 2.1 Capabilities of IMT-2000 and systems beyond IMT-2000........................... 25 
Figure 2.2 NTT DoCoMo Achieves 1 Gbps Packet Transmission ................................ 26 
Figure 2.3 IMT-2000 Spectrum Allocation WARC ?95 ................................................ 28 
Figure 2.4 Spectrum requirements for IMT-2000 and for IMT-Advanced.................... 30 
Figure 2.5 New Spectrum Bands Identified for IMT at WRC-07.................................. 31 
Figure 2.6 Global Allocation in the 3- to 6- GHz Band ................................................. 32 
Figure 2.7 Global Allocations for Mobile Services in the UHF Band........................... 33 
Figure 3.1 Classification of the Multiple Access Protocols ........................................... 37 
Figure 3.2 History of Mobile Communications Adopted Access Techniques............... 38 
Figure 3.3 Hybrid System Overall Aspects.................................................................... 40 
Figure 4.1 UMTS QoS Classes ...................................................................................... 43 
Figure 4.2 QoS Architecture .......................................................................................... 45 
Figure 5.1 Various Forms of Base Stations.................................................................... 49 
Figure 5.2 Macro and Pico Cell Base Stations............................................................... 50 
Figure 5.3 Partitioning and Interfaces defined by CPRI ................................................ 53 
Figure 5.4 Open interfaces defined by OBSAI .............................................................. 54 
Figure 6.1 4G Requirements .......................................................................................... 57 
Figure 6.2 Core Network Backhaul Access Technologies............................................. 59 
Figure 7.1 Hardware Technology Roadmap .................................................................. 68 
Figure 8.1 Miniaturization Challenge in Mobile Phones ............................................... 73 
Figure 8.2 Vanu?s Anywave Multi-Standard Base Station............................................. 75 
Figure 8.3 Vanu?s System Level Base Station Architecture .......................................... 76 
Figure 8.4 Vanu?s Anywave Multi-Standard Base Station ............................................ 79 
Figure 8.5 Vanu?s Selectable Carrier Frequency Architecture ...................................... 80 
Figure 9.1 Ideal Software Defined Radio Architecture.................................................. 83 
Figure 9.2 Basic Specifications for a Handportable Software Defined Radio............... 85 
Figure 10.1 Simplified Linear Receiver Architecture ...................................................... 91
                           
 
 xiii 
Figure 10.2 Digital IF-based Linear Receiver Architecture............................................. 92 
Figure 10.3 Multi-carrier Receiver Architecture Based on Digital IF ............................. 94 
Figure 10.4 Transmitter Architecture Employing a Digital IF Output............................. 95 
Figure 10.5 Multi-carrier transmitter architecture employing a digital IF output............ 97 
Figure 10.6 Switched Antenna Diplexer ........................................................................ 105 
Figure 10.7 Use of a Circulator to Provide Transmit/Receive Isolation........................ 106 
Figure 11.1 Different Configurations of MEMS Devices.............................................. 112 
Figure 11.2 Performance Comparison of FETs, PIN Diode, & RF MEMS Switches ... 114 
Figure 11.3 Application Areas of MEMS Switches, Varactors, and Hi-Q Inductors .... 116 
Figure 11.4 RF MEMS Applications in Multi-band RF Front Ends.............................. 117 
Figure 11.5 MEMS Packaging ....................................................................................... 121 
Figure 11.6 Varactor and Switched Capacitor Development Around the World .......... 126 
Figure 12.1 RF Front-End Design of a 3-band MEMS Wireless Telephone System .... 130 
Figure 12.2 A Reconfigurable RF Front End Block Diagram ....................................... 132 
Figure 12.3 Multi-band Flexible Base Station Architecture .......................................... 132 
Figure 12.4 ITRS Roadmap for RF MEMS ................................................................... 136 
Figure 13.1 Femtocell vs Macrocells Deployment and Topologies............................... 140 
Figure 13.2 Femtocell vs Macrocell Deployment and Topologies ................................ 141 
Figure 13.3 Worldwide Femtocell Revenue per Unit .................................................... 143 
Figure 13.4 Worldwide Femtocell Units by Technology............................................... 146 
Figure 13.5 Worldwide Femtocell Revenue by Technology ......................................... 146 
Figure 13.6 Worldwide Femtocell Revenue by Geographic Region ............................. 148 
Figure 13.7 Worldwide Femtocell Units Percentage by Geographic Region ................ 148 
Figure 13.8 Femtocell Equipment Manufacturers.......................................................... 149 
Figure 13.9 Femtocell Adoption Roadmap .................................................................... 150 
 
                           
 
 1 
Chapter 1 The State of the Mobile Wireless Communications Industry 
 
1.1 General Overview 
The wireless mobile communication industry is the fastest growing segment of 
the telecommunications market.    At the same time the wireless mobile industry is in a 
state of transition from the legacy analog network, 2G, to the new 3G network.   Japan 
has been leading the way in the implementation of new 3G services, including EVDO 
and HSDPA for data access.  A series of new standards have emerged over the past few 
years for mobile wireless, including WCDMA, CDMA-2000, TD-SCDMA, EVDO, and 
HSDPA.  Fig. 1.1 shows the migration paths across the various generations of wireless 
mobile technologies.  
The 3rd Generation Partnership Project (3GPP) is a collaboration initiative 
between groups of telecommunications associations, to make a globally applicable third 
generation (3G) mobile phone system specification within the scope of the International 
Mobile Telecommunications-2000 project of the International Telecommunication Union 
(ITU).  3GPP specifications, also known as UMTS, are based on evolved Global System 
for Mobile Communications (GSM) specifications. 3GPP standardization encompasses 
Radio, Core Network and Service architecture.    
3GPP2 (not to be confused with 3GPP) is the standardization group for 
CDMA2000, the set of 3G standards based on earlier 2G CDMA technology  [1].  Not 
                           
 
 2 
shown in Fig. 1.1 is the migration from 3G to 4G yet to come in the next three to five 
years (see also  Chapter 6). 
In today's market, the subscriber / end user is at the center of everything, and the 
 
 
 
 
Source:    [2] 
Figure 1.1 Various Upgrade Paths for 2G to 3G Technologies  
 
services are driven by user demands.  Initially more focused on short message service 
(SMS), multimedia message service (MMS), and content downloading, the 3G market 
has moved rapidly toward video sharing, mobile video, and IPTV, all of which require 
very high data throughput and highly efficient radio. 
                           
 
 3 
The latest UMTS evolution brings improved spectral efficiency at lower latency 
and higher data speeds with almost 100 times improvement from 3G, which is the 
promise of 4G services.  In addition, because of the increasing popularity of broadband 
access services like xDSL and cable for graphics and video streaming, a new broadband 
wireless access (BWA) competing and complementary technology has recently emerged:  
WiMax (Worldwide Interoperability for Microwave Access).    This technology has the 
potential to compete not only with 3G mobile wireless networks primarily in data 
applications, but also in voice applications thanks to voice-over-IP standards (VoIP).  
Fig. 1.2 shows a typical WiMax 802.16 network which also requires base stations 
connected to 802.11 ?hot spots? to reach the end users. 
 
 
 
Figure 1.2 WiMax Network 
                           
 
 4 
Fig. 1.3 shows a more complete view of the various competing wireless standards 
and highlights that the trend for the co-existence of multiple wireless standards will 
continue for many years ahead.   
 
 
Source:   Dr. Hossam Hmimy, Ericsson, 2006 
Figure 1.3 Cellular versus Mobile WiMAX 
 
Fig. 1.4 not only shows that the predominant technology for subscribers continues to be 
GSM but also that the 3G technologies are starting to increase in volume and will become 
more predominant in the next decade. 
Fig. 1.5 shows the top mobile operators around the world.  Of particular interest is 
the fact that the Chinese operators have the most subscribers followed by Latin America.  
In terms of operator?s market share, China Mobile ranks at the top with Vodafone in 
second place.    However, China Mobile is a domestic operator only while Vodafone, a 
UK company, is worldwide. 
Of particular interest are the wireless telecommunication trends that are  
                           
 
 5 
 
Source:   Infonetics Research, September 25, 2008 
Figure 1.4 Worldwide Mobile Subscribers by Technology 
 
 
Figure 1.5 Top Mobile Operators around the World 
                           
 
 6 
happening in China, Japan, South Korea, India, Europe and the United States.    The 
author spent several months at each of these regions of the world, in particular China and 
Japan, researching such market trends and published his findings in various papers at 
Auburn University  [3]. 
Fig. 1.6 shows the number of mobile subscribers by geographical region between 
2005 and 2011.  It can be seen that Asia Pacific leads the world in terms of subscribers 
with EMEA ranking in second place.   Asia Pacific is also the fastest growing market and 
is expected to add over 200 million subscribers per year between 2009 and 2011.   India 
will have the highest subscriber growth adding about 100 million subscribers per year as 
shown in Fig. 1.7Figure 1., followed by other countries in Asia Pacific (primarily South 
East Asia). 
 
 
Source:   Infonetics Research, September 25, 2008 
Figure 1.6 Mobile Subscribers by Geographical Region 
 
 
                           
 
 7 
 
Source:   Infonetics Research, September 25, 2008 
Figure 1.7 Mobile Subscribers by Asia Pacific Sub Region 
 
1.2 China 
As of 2006, China had been conducting standardized on-location tests on all three 
internationally recognized 3G technologies.  The three technologies are: TD-SCDMA 
standard jointly developed by China?s Datang Mobile Communications Equipment 
Corporation and Siemens from Germany  [See Appendix B]; W-CDMA standard based 
on the global mobile communication network system, a system commonly used in 
Europe; and CDMA2000 standard, popular in USA using CDMA standards.  Among the 
three, China owns the intellectual property of TD-SCDMA. 
On May 23, 2008, China, the most populous nation in the world, announced a 
historical re-structuring of their telecommunications carriers.   The announcement called 
for the six major Chinese carriers (China Mobile, China Telecom, China Netcom, China 
Unicom, China Railcom, and China Satcom) to re-structure into three New carriers:  New 
China Mobile, New China Telecom, and New China Unicom.    Each of these New 
                           
 
 8 
carriers would be focused on different 3G standards: New China Mobile would promote 
TD-SCDMA (China?s own 3G standard), New China Telecom would focus on W-
CDMA, and New China Unicom on CDMA2000.   Interestingly, this re-structuring came 
as no surprise to the author.  On May 18, 2006, the author interviewed a high-level 
official at Huawei Technologies who speculated that, while China was working on its 
own 3G standard (TD-SCDMA), the most like scenario would be that the Chinese 
government would end up supporting all three 3G standards concurrently. 
Fig. 1.8 shows the trend on the number of wireless subscribers in China from 
1999 to 2008.   The total number of wireless subscribers in 2007 was estimated between 
510 to 540 million, which would amount to a penetration of just over 40% considering 
that China had a population of 1.33 billion as of mid-2008.   Moreover, another industry 
estimate predicts that by 2010 the number of subscribers will increase to 738 million, or 
roughly 55% of the population  [4].    As of 2008 over 75% of the subscribers in China 
were for 2G GSM services and the rest for 2G CDMA services.   China is in the very 
early stages of 3G technology deployment but it is expected that after 2009 subscribers 
will rapidly move to 3G services.   
The recent China telecommunications re-structuring and wireless mobile market 
trends support the belief that multiple wireless standards will co-exist in China for many 
years to come.   The adoption of the three 3G standards in early 2008 also indicates that a 
vast volume of 3G base stations and handsets will be deployed in China during the next 
five years at cost targets which will be much lower than most Western countries. 
 
                           
 
 9 
  
Figure 1.8 Wireless Subscribers in China 
 
 
 
1.3 Japan 
As of August 31, 2008, Japan had just over 104 million subscribers of mobile 
wireless phones  [5], representing a penetration rate of 81% for a population of 128 
million as of mid 2008.    About 89% of the 104 million are for 3G services using either 
W-CDMA or CDMA2000 technologies.   What is most interesting about Japan?s 3G 
subscribers is that 100% have access to the Internet via their mobile phones.    The data 
access rates vary according to which technology is used as shown in Fig. 1.9.  Japan?s 
leadership in the adoption of the latest wireless mobile technologies was further validated 
                           
 
 10 
by DoCoMo?s announcement in June of 2005 that it was already testing 4G wireless 
networks, which would provide down link data speeds of about 1 Gbps  [6]. 
 
 
Source: 
 www.mobilein.com/Perspectives/eJapan_Article_062206.pdf by Jeorge S. Hurtarte 
 
Figure 1.9 3G Data Access Speeds 
 
On January 22, 2001, The Prime Minister and His Cabinet released the following 
statement to the Japanese people:  
?Japan must take revolutionary yet realistic actions promptly in order to create a 
"knowledge-emergent society," where everyone can actively utilize IT and fully 
enjoy its benefits. We will strive to establish an environment where the private 
sector, based on market forces, can exert its full potential and make Japan the 
world's most advanced IT nation within five years.?   [7] 
 
The e-Japan initiative had as one of its goals to ?establish one of the world's most 
advanced Internet networks within five years, and enable all the people who need it to 
have ultra high-speed access networks? using fixed-line and wireless networks.    In the 
                           
 
 11 
wireless mobile industry, Japan was the first country in the world to install the latest third 
generation (3G) wireless network infrastructure. 
During the five-year e-Japan project of 2001-2006, three were the wireless 
carriers that took the lead in its implementation:   
? NTT DoCoMo with a W-CDMA 3G wireless network infrastructure launched 
in October 2001.  NTT DoCoMo was the first wireless carrier to introduce the 
?always-on? data services capability with its i-mode feature, and the first one 
to introduce financial services to the mobile phone with its FeliCa technology.  
NTT DoCoMo?s market share in 3G services by was 49% by August 31, 
2008.   
? KDDI AU  with a CDMA 2000 3G wireless network infrastructure launched 
in April 2002.   KDDI AU was the first one to introduce the high speed data 
service in Japan in November 2003 using EVDO technology at 2.4 Mbps peak 
rates (the effective data rate is about 500 kbps).  KDDI AU?s market share in 
3G services was 32% by August 31, 2008. 
? Vodafone (J-Phone before being acquired by Vodafone) with a W-CDMA 3G 
wireless network infrastructure which was launched in December 2002.  
Vodafone was later acquired by Softbank in early 2006.   Softbank?s 3G 
market share was 17% by August 31, 2008. 
 
Fig. 1.10 shows the evolution of i-mode related services as introduced by NTT DoCoMo.  
The number of i-mode services stood at 46 million subscribers as of August 31, 2008.    
Thanks to their 3G wireless infrastructure all newer phones are ?always-on? in the 
sense that as long as they are powered up, they are always connected to the high speed 
Internet.   Thus, email messages can be received instantaneously and there is no need to 
log in to any e-mail service provider account.    Just a quick look at NTT DoCoMo?s 
phone line-up can provide a good insight of the various feature trends in Japanese 
phones, which include larger LCD displays, IR information exchange, bar code readers, 
built-in FM radios, Moji icons, etc.  NTT DoCoMo?s phone line up even includes some 
                           
 
 12 
models for children aimed at promoting the use of mobile communications at an early age 
 [8].  
 
 
Source:   www.nttdocomo.co.jp 
Figure 1.10 NTT DoCoMo?s i-mode Services Evolution 
 
Japanese people have access to a wide variety of 3G services which include: 1)  e-
mail, 2) internet access, 3) location-based services which enable the user to check the 
weather forecast, traffic and store information, and other convenient information for local 
areas as well as the map information to the user's current location, 4) dynamic video 
                           
 
 13 
content, 5) financial services that enable to phone to act like a ?credit card? and many 
other 3G-enabled services  [9]. 
Japan?s leadership in the adoption of 3G technology has led the way to a wide 
variety of data-intensive services and features which in turn has given way to a lot of 
innovation in the phones.   Japan?s top-down initiative by its government will continue to 
drive the Japanese people towards even more advanced wireless mobile technologies like 
4G ahead of any other nation in the world.    It can also be inferred from such market 
trends that the amount of features that will be added to Japan?s 4G-enabled phones will 
tax the amount of the printed circuit-board space available and will also drive down the 
cost budgets for the electronics related to the RF front end of such phones.  See also  [10] 
for an article written by the author on Japan?s world leadership in the implementation of 
3G infrastructure and its related phones. 
 
1.4 South Korea 
South Korea?s mobile phone penetration was roughly 88% in early 2008 for a 
population of 49 million people, and the mobile phone penetration is expected to grow to 
95% by 2010.   After Japan, South Korea launched the second 3G network to go 
commercially live by SK Telecom in South Korea on the CDMA2000 1xEVDO 
technology in January 2002.   By May 2002 the second South Korean 3G network was 
launched by KTF on EVDO, and thus the Koreans were the first to see competition 
among 3G operators. 
South Korea has long led the way in the rapid adoption of wire-line high speed 
broadband access and took a world leadership position with the adoption of xDSL 
                           
 
 14 
broadband access technologies.   Non-surprisingly, such leadership on broadband access 
prompted South Korea to also lead the way on yet another broadband access technology:  
WiBro, which stands for Wireless Broadband.    WiBro is the South Korean service name 
for the IEEE 802.16e (mobile WiMAX) international standard. 
WiBro was devised to overcome the data rate limitation of mobile phones (for 
example CDMA 1x) and to add mobility to broadband internet access (for example 
ADSL or Wireless LAN).  In February 2002, the Korean government allocated 100 MHz 
of electromagnetic spectrum in the 2.3 - 2.4 GHz band, and in late 2004 WiBro Phase 1 
was standardized by the TTA of Korea.  In late 2005 the ITU incorporated the elements 
of WiBro in the IEEE 802.16e standard.   Two South Korean Telcos (KT, SKT) launched 
commercial service in June 2006. 
In November 2004, Intel and LG Electronics executives agreed to ensure 
compatibility between WiBro and WiMAX technologies.   This agreement signaled a 
major step towards the adoption of both WiBro and WiMAX as viable competing 
technologies versus 3G mobile technologies.   This fact was further accentuated when in 
September 2005, Samsung Electronics signed a deal with Sprint Nextel Corporation of 
the United States of America to provide equipment for a WiBro trial. 
WiBro base stations will offer an aggregate data throughput of between 30 to 50 
Mbps and will cover a radius of 1-5 km allowing for the use of portable internet usage.  
WiBro will provide mobility for moving devices up to 120 km/h (74.5 miles/h) compared 
to Wireless LAN having mobility up to walking speed and Mobile Phone having mobility 
up to 250 km/h.   The technology will also offer improved Quality of Service (QoS; see 
also  Chapter 4).  The inclusion of QoS allows for WiBro to stream video content and 
                           
 
 15 
other loss-sensitive data in a reliable manner. These all appear to be (and may be) strong 
advantages over the fixed WiMAX standard (802.16a). 
Wireless broadband access is set up like a cellular system; using base stations that 
service a radius of several miles/kilometers (see Fig. 1.2).  Base stations do not 
necessarily have to reside on a tower.   More often than not, the base station antenna will 
be located on a rooftop of a tall building or other elevated structure such as a grain silo or 
water tower. A customer premise unit, similar to a satellite TV setup, is all it takes to 
connect the base station to a customer.  The signal is then routed via standard Ethernet 
cable either directly to a single computer, or to an 802.11 hot spot or a wired Ethernet 
LAN. 
South Korea?s leadership in the adoption of broadband access technologies has 
established WiBro/WiMAX as yet another wireless mobile technology that has potential 
as a viable competing technology against 3G mobile wireless (see also Fig. 1.3).   This 
fact further validates the fact that the base stations and handsets of the future will 
continue to support multiple wireless standards. 
 
1.5 India 
The mobile market in India continued its strong growth through 2007 and looked 
to be carrying a 50 percent annual growth rate into 2008.  While the market was initially 
totally a GSM domain, CDMA technology was introduced as a Wireless Local Loop 
(WLL) service, which after a long battle with the regulator was eventually accepted as a 
legal mobile service.  The subscriber base has developed into a fairly stable mix of GSM 
subscribers, (74 percent market share at end-2007) and CDMA subscribers (24 percent). 
                           
 
 16 
India had almost 350 million mobile subscribers (including GSM & CDMA) in 
early 2008, representing a penetration rate of only 30%.    However, the number of 
subscribers surpassed by far all estimates of 2004 as shown in Fig. 1.11.   
 
 
 
Figure 1.11 India?s Wireless Subscribers 
 
The huge potential of the mobile wireless market in India was attracting foreign 
investment.  Vodafone arrived on the scene with a considerable impact, having 
successfully acquired Hutchison Telecom's 67% stake in Hutchison Essar for US$11.1 
billion, one of the largest ever single foreign investments in India.   UK incumbent BT 
also boosted its presence in India in 2007, expanding the footprint of its managed 
                           
 
 17 
network services and IT operations in the country through its acquisition of i2i 
Enterprise. 
As of mid 2008, there was a capacity crunch in India primarily paced by capacity 
limitations at the State-owned Bharat Sanchar Nigam Ltd (BSNL).  In August 1, 2008, 
India?s government announced its plans to hold an international auction for 3G and 
WiMAX telecoms services in the world?s second-largest mobile phone market, with the 
minimum price for a pan-India license set at Rs20bn ($472M)  [11].  Wireless licenses 
would be auctioned, and the process was expected to be completed within several 
months.  Up to five operators would be selected to provide wireless services in each 
coverage area.     As of March 2009, the 3G action was still pending. 
Government-owned service providers, Bharat Sanchar Nigam Ltd (BSNL) and 
Mahanagar Telephone Nigam Ltd. (MTNL), would each receive a license, gaining a head 
start of several months over competitors in rolling out 3G services.   At least two CDMA 
operators would be allowed to offer 3G services in each license area.  
Indian spectrum in the 2.1-GHz band was designated for 3G services, with as 
many as 10 blocks available for auction in each region.  Spectrum for EVDO data 
services would be auctioned in the 450- and 800-MHz bands, and eventually in the 1,900-
MHz band.  Some pockets in the country are expected to have WiMax earlier than 3G as 
private operators such as Reliance, Tata and Bharti have been aggressive on this 
technology front.  
Clearly, India is still at the early stage of mobile phone penetration and the growth 
of 2G services will continue to be phenomenal.  3G services will be next with W-CDMA, 
CDMA-2000 and WiMAX co-existing for the many years ahead. 
                           
 
 18 
 
1.6 Europe 
In general the EU countries have completed the transition to UMTS W-CDMA 
3G networks and are quickly becoming at par with Japan, which has led the 
implementation of 3G in the world.    With mobile penetration at or nearing saturation 
across Western Europe, the deployment of new technologies and services such as 
HSDPA and broadcast mobile TV is providing operators with the opportunity to enhance 
their product portfolios and increase the average-revenue-per-user (ARPU) levels.   The 
country with the highest penetration rate in the world is Luxembourg with 157.3 mobile 
subscribers for every 100 people. 
The total mobile subscription in Europe reached 753.5 Million by the end of 
December 2007 at a CAGR of nearly 17.6% during the period 2002-2007.   In terms of 
subscriber base, the largest market in Europe is Russia with a share of 22%.  In 
comparison, Ireland forms the smallest market with approx 0.7% share in 2006.  Total 
mobile subscribers in the European region are expected to increase at a CAGR of around 
2.35% during 2008-2012.   Eighty percent of Europeans will use 3G by the end of 2013.   
Mobile Internet usage will triple from 2007 to 2013 to 125 Million Users.   [34] 
The following are some mobile market trends to watch in the EU: 
? Mobile Broadband.  Internet access via the mobile phone at broadband data 
rates in excess of 500kbps. 
? Data networking.  The use of shared applications such as Microsoft 
NetMeeting, as well as access to corporate intranets and to corporate database 
applications such as customer relationship management (CRM) and enterprise 
resource planning (ERP) applications or mobile workforce management 
applications. 
? Browsing.  Content that is made available over cellular networks without 
charge, other than those charges that apply for standard network access, 
                           
 
 19 
including mobile-specific content (free-to-air information) and general Web 
access (Web browsing). 
? Paid information. Push or pull electronic content, delivered over cellular 
networks, for which a premium is paid over and above standard network 
access charges. 
? Entertainment services.  Downloading or accessing games, cartoons, music, 
video clips and other forms of entertainment over a cellular network.  
Messaging based on third-party content and interaction with other media (such 
as TV programs) is included in this service category. 
? M-commerce.  Transaction-oriented services, including e-pay facilities, 
mobile shopping portals, mobile banking and share trading, and bookings and 
ticketing, but excluding games and other forms of content. 
? Video-telephony.  Real-time, audio-visual, person-to-person communications.  
? Fixed-Mobile Substitution (FMS).  While voice telephony generates by far the 
greatest proportion of mobile operator revenue in Western Europe, mobile 
voice telephony will strongly influence mobile operators? revenue prospects 
for at least the next three years. Fixed?Mobile Substitution (FMS) 
encompasses traffic substitution, as voice calls migrate from fixed to mobile 
services, and line substitution, as households and businesses dispense with 
their fixed-line voice services.  FMS continues to represent a critical threat to 
fixed operators.  Fixed-line substitution is particularly damaging because it 
removes fixed operators' opportunity to generate revenue from either voice or 
broadband services. 
? HSUPA. High-Speed Uplink Packet Access (HSUPA) is a 3G mobile 
telephony protocol in the HSPA family with up-link speeds up to 5.76 Mbit/s.  
The specifications for HSUPA are included in Universal Mobile 
Telecommunications System Release 6 standard published by 3GPP.   
Vodafone, T-Mobile, 3, Telecom Italia and all other major European carriers 
introduced HSUPA service in late 2007.    Vodafone was the first network 
operator in Germany to launch in 2007 a notebook data card that delivers 
online content downloads at speeds of up to 7.2 megabits per second plus 
extremely fast uploads. It supports upload rates of up to 1.45 megabits per 
second using the new HSUPA technology, which is four times faster than ever 
before.  It can be used to send very large image and video files or PowerPoint 
presentations with a notebook - at faster rates than standard wired broadband 
connection - while on the move.  
 
2G digital GSM technology started in Europe and has since evolved into various 
technologies.    Europe is now leading the way in the adoption of the most recent 3G 
High-Speed Uplink Packet Access (HSUPA) is mobile telephony protocol, which 
requires that the base stations and mobile phones be upgraded for forward compatibility 
                           
 
 20 
with new protocols such as HSUPA.    As later described in this dissertation, the constant 
need for network and phone handset upgrades only highlights the need for a 
programmable RF front end architecture which can complement a software 
configurable/defined radio bases band architecture. 
 
1.7 North America 
1.7.1 Overview 
The total number of mobile subscribers in North America rose to 278 million at 
the end of Q1 2008, from 257 million a year earlier.  The penetration rate saw a 5.3 
percentage-points gain to 82.5%. Obviously this figure is close to the 84.9% penetration 
rate in the USA, the region's largest market by some distance with 92.7% of North 
America's mobile customers, although it is dragged down by the low Canadian rate of 
just 60.5%.  
By far the fastest growing technology in North America is W-CDMA, which saw 
a 336.7% gain in customers in the 12 months ending on March 31, 2008. The total 
number of W-CDMA subscribers surpassed the 10 million mark during Q1 2008 to finish 
the quarter at 11 million.  This is still below the number of iDEN customers, which stood 
at 16.6M at the end of Q1, but with iDEN losing 5.1M customers year on year and W-
CDMA gaining 8.6M, it seems likely that iDEN will lose its status as North America?s 
third most important technology before the end of 2008. 
CDMA remains the dominant standard with 52.7% of the total North American 
customer base, or 146M subscribers, up from 130M (50.7% of the total) a year earlier. 
                           
 
 21 
The number of GSM subscribers broke through the 100M barrier to reach 104M 
at the end of Q1, but growth more than halved from 15.0% in the 12 months ending Q1 
2007, to 7.0% in the subsequent 12 months, as a consequence of the success of W-
CDMA.  Annual CDMA customer growth was also down compared to the prior twelve 
months with a 4.4 percentage-points fall in the rate, although it remained in double digits 
with a 12.2% yearly gain  [34]. 
 
1.7.2 Femtocells 
Home base station solutions, also called femtocells have recently attracted a lot of 
attention.  The idea is to give the mobile user a small box similar to a wireless router that 
provides mobile cellular coverage at home.   Femtocells are a relatively recent trend in 
the United States.   The Femto cellular device (being smaller than pico, the term used by 
mobile operators that refers to smaller cell sites) has a cellular antenna to boost the 
available signal as well as an Internet connection.  The device uses the user?s Internet 
connection to connect to the mobile provider?s network and route the phone calls. 
The 3G femtocell is a tiny, consumer-friendly 3G base station optimized for the 
home.  It is a low-cost, high-performance W-CDMA (or any other 3G standard) femtocell 
delivering voice and HSxPA services to the home while improving capacity in the entire 
cell.  The physical box is a standalone unit with an Ethernet interface to the user?s 
broadband connection and can also be embedded into other residential gateway products 
(e.g. ADSL or Cable/WiFi routers).   Some of the most recent femtocells being 
announced are based on a software-intensive parallel DSP core and as such are software 
                           
 
 22 
upgradeable in the field to be compatible with the latest W-CDMA and IP network 
features. 
Fixed line carriers and DSL service providers are seriously threatening the mobile 
business by offering new products centered around Fixed Mobile Convergence, which in 
essence utilizes wireless access technology (like WiMAX) to capture mobile users when 
at home, and divert their mobile originated voice and data calls to the fixed line network. 
Mobile operators thus have found themselves in a predicament: the traditional 
macro cellular networks are not able to compete with what DSL service providers can 
offer, neither from a cost structure point of view, nor in terms of access speeds.  A home 
base station or femtocell can help mobile operators in many ways.  First they can 
compete directly with other home access solutions which rely on other types of access 
technology such as WiFi or WiMax.  This is because low cost mobile handsets are widely 
available, whereas dual mode handsets with WiFi capabilities are pricey and still limited 
in variety.    
The femtocell will also enable the operator to cater to subscribers where the 
service is most needed.  Compare this with traditional macro deployment where the 
infrastructure is first built in places where subscribers are "likely" to appear and even 
before subscribers start using the network. A femtocell also circumvents the issue of 
outdoor-to-indoor propagation which typically kills the capacity of 3G and 3.5G systems. 
 The deployment costs associated with a femtocell are marginal in comparison 
with Macro cell deployment.  This is because the customer is providing the "real estate" 
to keep the box, the power, as well as the backhaul (e.g., DSL subscription).  
 
                           
 
 23 
1.8 Chapter 1 Summary 
One key conclusion that can be derived from this wireless market analysis is that 
industry standards are changing dynamically, which in turns forces the 
telecommunications carriers to frequently change or upgrade their networks with the 
latest technology equipment.    In turn, the original equipment manufacturers need to also 
design new equipment with the new standards to stay in business.   
Another key conclusion is that the market size for mobile wireless service still has 
a long way to go especially in China, India and Latin America, and that these regions will 
require lower equipment price points as compared to Western Countries. 
As data-intensive services become available due to the higher data rates possible 
with 4G and WiMAX/WiBro technologies, the base stations and hand sets will require a 
smaller footprint and lower cost RF front end electronics. 
The competition for customers between fixed line and wireless mobile operators 
has also brought about the advent of the femtocell or ?personal base stations? which will 
also have to adopt the various 3G/4G standards.   Femtocells will also face a new price 
point in the market as compared to pico base station thus putting more pressure on the 
cost reduction of the RF front end of the femtocells.   
The combination of above wireless mobile industry trends dictates the need for a 
flexible, low cost, small size, programmable RF front end that can quickly adapt to the 
changing wireless standards, protocols, and feature sets of the future. 
                           
 
 24 
Chapter 2 The Mobile Wireless RF Spectrum Allocation 
 
2.1 Introduction 
The ITU-R Working Party 8F is responsible for the overall system aspects of 
IMT-2000 and beyond.   The ITU-R believes that systems ?beyond IMT-2000? (e.g., 4G) 
should support a steady and continuous evolution of new applications, products and 
services through improvements in data rates and enhancements to the existing IMT-2000 
radio interfaces.    In addition, it further states that a framework for systems beyond IMT-
2000 should be realized by the functional fusion of existing (legacy), enhanced, and 
newly developed elements of new mobile access, nomadic local wireless, and so forth, 
with high commonality and seamless inter-working.   
 
2.2 Data Rate Requirements 
The new capabilities of systems beyond IMT-2000, for example, 4G systems, will 
include much higher data rates than the current 3G systems.   For example, the 
capabilities of some of the IMT-2000 terrestrial radio interfaces are already being 
extended up to 10Mbps and it is expected that these will be expanded even further.    For 
systems beyond IMT-2000, there may be a requirement for a new wireless access 
technology for the terrestrial component and the ITU predicts that the data rates for such 
new systems will be in the range of 100 Mbps for high mobility such as mobile access, 
                           
 
 25 
and up to approximately 1 Gbps for low mobility such as nomadic/local wireless access 
(see Fig. 2.1).   In fact, as early as 2005, NTT DoCoMo announced that it was already 
testing systems capable of 1Gbps (see Fig. 2.2). 
 
 
Source:   ITU-R,  M.1645, ?Framework and overall objectives of the future development 
of IMT-2000 and systems beyond IMT-2000.? 
 
 
Figure 2.1 Capabilities of IMT-2000 and systems beyond IMT-2000 
 
                           
 
 26 
These data rate figures and their associated mobility should be seen as the targets 
for research and investigation of the basic technologies necessary to implement the 
framework.   Note that 100 Mbps and 1 Gbps refer to the overall cell throughput in wide 
and local access, respectively.    
 
 
Source:   www.nttdocomo.com.jp 
Figure 2.2 NTT DoCoMo Achieves 1 Gbps Packet Transmission 
                           
 
 27 
Due to the predicted data rate requirements, additional RF spectrum will be 
needed in order to deliver the new capabilities of systems beyond IMT-2000.    It also 
possible that upstream (uplink) and downstream (downlink) may have different 
maximum transmission speeds.  In addition to IMT-2000 and ?beyond IMT-2000? 
systems, other different radio access and communications systems will continue to co-
exist, for example, WLANs, WPAN, WMANs, digital broadcast, and fixed wireless 
access (FWA) like WiMax. 
 
2.3 Spectrum Issues 
Spectrum is a crucial issue in the design and manufacture of wireless systems and 
devices.   The characteristics of propagation and technical parameters of the frequency 
bands may influence or determine requirements of devices for systems operating in the 
frequency bands and service characteristics such as mobility requirements and coverage.  
In addition, spectrum is a scarce resource and, therefore, it should be utilized efficiently.   
Thus, suitable and affordable frequency bands should be allocated to each wireless 
system for the purpose of efficient use of spectrum. 
The ITU-R plays an important role in the allocation/identification of wireless 
services and the WRC, supported by the ITU-R, is the only entity that can make decision 
on the global use of spectrum.  To meet the technical and service objectives of systems 
beyond ITM-2000, more spectrum is required than that already identified for IMT-2000 
(see Fig. 2.3) and pre-IMT-2000.   Therefore, the ITU-R allocates new spectrum in terms 
of bandwidth (spectrum amounts), affordable frequency bands, and associated sharing or 
compatibility problem. 
                           
 
 28 
 
 
 
 
 
 
 
 
 
 
 
Source: 1995 World Radio Communication Conference 
Figure 2.3 IMT-2000 Spectrum Allocation WARC ?95 
 
The working methods of the ITU-R WP8F (Working Party 8F) pursuing new 4G 
systems spectrum allocation are as follows: 
? Identify potential future services for 4G and calculate associated traffic 
demands. 
? Develop methodology for calculating spectrum requirements for 4G and 
estimate required spectrum amounts (bandwidth) for 4G that reflect future 
traffic demands. 
? Identify candidate frequency bands for 4G mobile and nomadic systems, 
respectively, and conduct sharing and compatibility research in the identified 
bands. 
? Develop technical and regulatory framework/requirements for 4G with regard 
to sharing/compatibility studies. 
? Derive frequency bands for 4G and the required bandwidth in each frequency 
band. 
 
 
                           
 
 29 
2.4 New Spectrum 
New frequency bands for the new mobile systems beyond IMT-2000 concentrate 
on frequency bands below 6 GHz.   Several technical and regulatory aspects support 
these frequencies bands.   On the technical front, mobile systems usually suffer from 
larger Doppler shift as the operating frequencies are increased.   Moreover, path-loss in 
the transmission also influences system performance and coverage, and hence, system 
deployment costs.   Path-loss has an explicit and direct relationship with the frequency 
bands.   To obtain bandwidth sufficient enough to support services with very high data 
rates, frequency bands for future mobile networks need to be above frequency bands 
allocated for IMT-2000, that is, above 2.7GHz.   While at this point there is no conclusive 
technical evidence on what the upper frequency limit for the band should be, 6 GHz 
appears to enjoy support among the active members of ITU working on spectrum 
identification and planning.    Thus, frequency in the band from 2.7 to 6 GHz could be 
regarded as the most promising frequency bands for future mobile communications 
systems. 
Report ITU-R M.2078 calculated that for a single network deployment per 
country the overall spectrum requirements for the future development of IMT-2000 and 
for IMTAdvanced to be around 1280 MHz in a low market setting and 1720 MHz in 
higher market settings for the year 2020.  This results in an additional spectrum demand 
of around 500 MHz and 1 GHz respectively in all ITU Regions as shown in Fig. 2.4. 
 The total traffic volume (voice and data) in mobile networks is expected to grow 
significantly.  For example, Report ITU-R M.2078 expresses growth factors of about 2 to 
3 by 2010 for Europe compared to today.  The traffic increase originates from both voice 
                           
 
 30 
and non-voice traffic.  Fixed voice substitution, broadband Internet access and mobile TV 
will be major contributors along this way.  Therefore, sufficient network capacity is a 
must to satisfy the end-user expectations in terms of service quality, namely throughput 
and latency. 
 
Region 1: Europe;  Region 2:  Americas;  Region 3: Asia. 
Source:   ITU-R M.2078 
Figure 2.4 Spectrum requirements for IMT-2000 and for IMT-Advanced 
 
In November 2007, the WRC-07 identified globally harmonized spectrum for use 
by International Mobile Telecommunications (IMT-2000 and IMT-Advanced).  
Additional spectrum was allocated for IMT systems in various new bands, resulting in 
392 MHz of new spectrum in total in Europe and 428 MHz in the Americas  [12] (see also 
Fig. 2.5): 
? 20 MHz in the band 450?470 MHz (globally) 
? 72 MHz in the band 790?862 MHz for Region 1 (Europe) and parts of Region 
3 (Asia) 
? 108 MHz in the band 698?806 MHz for Region 2 (Americas) and some 
countries of Region 3 (Asia) 
? 100 MHz in the band 2.3?2.4 GHz (globally) 
? 200 MHz in the band 3.4?3.6 GHz (no global allocation, but identified in 82 
countries) 
                           
 
 31 
Note: These bands will not be available immediately for next generation mobile 
networks (NGMN) usage, but opened to the market following transition periods of up to 
several years. Additionally, the allocations regarding the bands 790-862 MHz and 3.4 - 
3.6 GHz in Region 1 will only come into full effect in 2015 and 2010 respectively.   
 
 
 
 
Source: IMT at WRC-07 
Figure 2.5 New Spectrum Bands Identified for IMT at WRC-07 
 
 
 
 
Current global allocation of the 3- to 6- GHz band is shown in Fig. 2.6, while the global 
allocation of the UFH band is shown in Fig. 2.7. 
 
                           
 
 32 
 
Frequency Bands ITU-R Allocations Current Operating Mobile Systems
3,100-3,300 MHz Radio location Airborne Warning and Control Systems (AWACS)
Earth exploration Synthetic aperture radars (SARs)
satellite (Active)
Space research (active)
3,300-3,400 MHz Radio location AWACS
SARs
3,400-3,600 MHz Fixed, mobile, Broadband wireless access (BWA)
fixed satellite, radio location such as WiMAX , 802.16
European countries: electronic news gathering
(ENG) / Outside broadcasting (OB)
3,600-4,200 MHz Fixed, mobile, Fixed Satellite Services (FSS), FWA, including
fixed satellite microwave link
4,200-4,400 MHz Aeronautical radionavigation
4,400-4,500 MHz Fixed, mobile Point-to-point microwave link
4,500-4,800 MHz Fixed, fixed satellite, mobile Global geostationary Earth orgit (GEO) satellite
bands, point-to-point microwave link
4,800-4,990 MHz Fixed, mobile, Point-to-point microwave link
radio astronomy
5,000-5,150 MHz Aeronautical radionavigation
5,150-5,350 MHz Aeornautical radionavigation, 5-GHz wireless LAN
fixed satellite
(Earth-to-space), mobile
5,350-5,470 MHz Earth exploration, 5-GHz wireless LAN
radiolocation,
space research, mobile
5,470-5,650 MHz Maritime radionavigation
5,650-5,725 MHz Radio location, 5-GHz wireless LAN
space research, mobile
5,725-5,825 MHz Radio location, amature 5-GHz wireless LAN, dedicated short-range
communications (DSRC)
5,825-6,000 MHz Fixed, fixed satellite Broadcasting relay
(Earth-to-space), mobile
  
Source:    [15]  
Figure 2.6 Global Allocation in the 3- to 6- GHz Band  
                           
 
 33 
 
Frequency Bands ITU-R Allocations Operating Mobile Systems (Current and Future)
406-430 MHz Mobile, fixed. TRS including TETRA, iDEN
Europe: PMR
U.S.: PMRS
440-470 MHz Mobile, fixed Cellular (NMT 450),
Radio location (440-450 MHz) TRS inlcuding TETRA, iDEN
Radio sonade (460-470 MHz)
740-806 MHz Mobile, broadcasting No current mobile services
806-960 MHz Mobile Cellular including CDMA (IS-95, 1x), GSM
(IMT-2000 additional bands) TRS including TETRA, iDEN
1,710-2,025 MHz Mobile PCS1900, GSM 1800, IMT-2000 including FDD, TDD
2,110-2,200 MHz (IMT-2000 additional bands) and satellite component
IMT-2000 core bands DECT (Europe, 1,880-1,900 MHz)
(1,885-2,025 MHz
and 2,110-2,200 MHz)
2,300-2,400 MHz Mobile, fixed WiBro (Korea), TDS-CDMA (China)
2,400-2,500 MHz Mobile (ISM bands) IEEE 802.11b/g, Bluetooth
2,500-2,690 MHz Mobile, fixed, satellite (BSS: IMT-2000
2,605-2,655 MHz,
MSS: 2,500-2,520 and 
2,670-2,690 MHz)
 
Source:    [15] 
Figure 2.7 Global Allocations for Mobile Services in the UHF Band 
                           
 
 34 
2.5 Chapter 2 Summary 
Wireless systems frequency spectrum utilization for pre-IMT-2000, IMT-2000, 
and ?beyond IMT-2000? will continue to co-exist for many years to come.    The ITU?s 
envisioning of a backward compatibility of services as new systems are deployed will 
also dictate that the new systems and devices become network agnostic and thus should 
be able to operate with as many as possible of the existing networks.    The RF front end 
of the radio interfaces of the future should be able to be ?plug and play? and be agnostic 
to any spectrum frequency domain while capable of operating at frequency bands in the 1 
to 6 GHz domain for backward and forward compatibility with pre- and post-IMT-2000 
systems.   Moreover, the radio RF front-end should be capable of allowing data rates of 
up 1 Gbps for low mobility and 100Mbps for high mobility systems. 
                           
 
 35 
Chapter 3 The Mobile Wireless Channel Multiple Access Techniques 
 
3.1 Introduction 
The understanding of the various multiple access techniques for mobile wireless 
communications is relevant to this dissertation because it also validates the thesis that a 
variety of multiple access techniques will continue to co-exist in systems beyond IMT-
2000.    Thus, the platform for the implementation of future base stations and handsets 
will have to be flexible enough to accommodate a variety of multiple access techniques.       
Software Define Radio (SDR) technology addresses the multiple access 
techniques in that it provides for configurability of the protocol via DSP techniques.    
However, the author believes that a ?smart programmable RF front end? (SP-RFFE) will 
also be needed to complement the functions being implemented by SDR techniques.   
Thus, before elaborating on the functions that the SP-RFFE will implement, it is critical 
that an overview of the various multiple access techniques is provided to understand how 
the SP-RFFE might be able to assist in addressing any implementation issues regarding 
the various multiple access techniques. 
 
3.2 Overview of Multiple Access Protocols 
Agreement among users on the means of communication is known as protocol.   
When users use a common medium for communications, it is called multiple access.  
                           
 
 36 
 Thus, the multiple access protocol is defined as the agreement and set of rules 
among users for successful transmission of information using a common medium.   
Whenever some resource is used and thus accessed by more than one independent user, 
the need for a multiple access protocol arises.   In the absence of such a protocol, 
conflicts occur if more than one user tries to access the resource at the same time.   
Therefore, a multiple access protocol should avoid or at least resolve these conflicts.   
Thus, a multiple access technique is defined as a function sharing a (limited) common 
transmission resource among (distributed) terminals in a network  [15]. 
Starting in 1970 with the ALOHA protocol, a number of multiple access protocols 
have been developed.    One classification breaks these various protocols in three 
categories: the contention-less protocols, the contention protocols, and the hanging (or 
swimming) protocols.   Fig. 3.1 shows the various protocols under each of these three 
classifications and Fig. 3.2 shows the history of mobile communications generations in 
terms of adopted access technique. 
                           
 
 37 
 
Source:    [15] 
Figure 3.1 Classification of the Multiple Access Protocols 
 
 
 
                           
 
 38 
 
Source:    [15]  
Figure 3.2 History of Mobile Communications Adopted Access Techniques. 
 
3.3 OFDM-based Hybrid Protocols 
Of all of the multiple access protocols, OFDM-based techniques look the most 
promising for systems beyond IMT-2000 (e.g., 4G).    OFDM stands for Orthogonal 
Frequency Division Multiplexing.   The reasons for the high potential of OFDM are as 
follows: 
? Multi-carrier techniques can combat hostile frequency selective fading 
encountered in mobile communications.   The robustness against frequency 
selective fading is very attractive especially for high-date transmission. 
? The OFDM scheme is a mature technology due to much research and 
development for high rate wireless LANs and terrestrial digital video 
broadcasting.  A great deal of know-how has been developed on OFDM. 
? By combining OFDM with CDMA, we can have synergistic effects, such as 
enhancement of robustness against frequency selective fading and high 
scalability in possible data transmission rate. 
                           
 
 39 
OFDM, which is a form of multi-carrier modulation, is used as a multiple access 
scheme in wireless communications.   In OFDM, densely spaced sub carriers with 
overlapping spectra are generated using fast the Fourier transform (FFT).  Signal 
waveforms are selected in such a way that the sub carriers maintain their orthogonality 
despite their spectral overlap.   Usually, CDMA or TDMA is used with OFDM to achieve 
multiple access capability. 
Basically, orthogonal frequency resources (i.e., orthogonal sub carriers) can be 
shared among users, which is the simplest multiple access technique based on OFDM 
modulation.   Besides, there are a number of hybrid multiple access techniques that can 
be found in the literature.   Hybrid means an amalgamation of OFDM and multiple access 
techniques (with the main accent on spread spectrum) to provide an efficient multi-user 
scenario with very high data rate.    Some of the most popular techniques are: 
? MC-CDMA (multi-carrier code division multiple access) 
? OFDMA (orthogonal frequency division multiple access) 
? OFDM-CDMA-SFH (OFDM-CDMA slow frequency hopping) 
? VSF-OFCDMA (variable spreading factor-orthogonal frequency and code 
division multiplexing) - proposed by NTT DoCoMo (see also Fig. 2.2) 
 
The hybrid approach of above techniques solve a variety of issues that cannot be 
addressed by CDMA or OFDM techniques alone.   Fig. 3.3 summarizes some of the 
problems and solutions using the various hybrid techniques. 
 
                           
 
 40 
Problem Solution
1. OFDM system only supports one user Use OFDM/CDMA, which supports multiple users
2. CDMA does not permit very high data rates, OFDM counters this by S/P conversion, allowing
owing to frequency selective fading at high data flat fading at subcarrier level.
rates
3. CDMA systems suffer from near-far effect in This is solved in the hybrid system using SFH
the uplink
4. CDMA systems use DS-SFH to control the The OFDM-FH system hops on frame basis allowing
near-far effect.   But DS-SFH mostly supports coherent modulation
noncoherent modulation owing to hopping at bit
level.   However, coherent modulation is possible,
but maintaining phase coherence between hops is
difficult.
5. CDMA systems cannot indefinetely support The hybrid system allows any number of users by 
multiple users due to MAI and SI problems caused increasing the number of hops.  Hence, Number of
by too many users. Users - Number per CDMA System Number of Hops.  
Bandwidth should, however, be available.
6. CDMA poses synchronization problems at very OFDM systems have an easier synchronization
high chip rates problem due to using cyclic prefixes.
 
Source:    [15] 
Figure 3.3 Hybrid System Overall Aspects 
 
Each of the presented protocols has its own merits but each of them should strive 
to solve the following issues related to the wireless mobile communications environment: 
? The hidden terminal problem: two terminals are out of range of each other, 
hidden by a hill, a building, or some physical obstacle opaque to UHF signals, 
but both are within the range of the central or base station; 
? The near-far effect: transmissions from distant users are more attenuated than 
transmissions from users close by; 
? The effects of multi-path fading and shadowing experienced in radio channels; 
? The effects of co-channel interference in cellular wireless systems caused by 
the use of the same frequency band in different cells. 
 
                           
 
 41 
Many of the protocol properties mentioned above are conflicting, and a trade-off 
has to be made during the protocol design.  The trade-off depends on the environment 
and the specific use for the protocol one has in mind. 
 
3.4 Chapter 3 Summary 
The various multiple access techniques for mobile wireless communications will 
continue to co-exist as the migration to systems beyond IMT-2000 occurs.    Because of 
the various ODFM hybrid approaches being proposed, it is clear that a flexible and 
programmable platform would have to be used for the implementation of the various 
multiple access schemes.     SDR-based platforms have the potential to address such 
flexibility and programmability requirements.    SDR systems use DSP techniques and 
there is plenty of research and industry work being done in this area.   The author of this 
dissertation believes that some of the SDR-DSP functions may be able to be assisted via a 
smart programmable RF front end (SP-RFFE), which in essence, complements the SDR 
platform.   Future chapters examine possible functions that can be moved from the SDR-
DSP section of the radio to the SP-RFFE.    The idea behind the SP-RFFE is that it has 
the potential of becoming some sort of a ?wireless channel sniffer? whose specific 
features and functions will be later defined by the future work being proposed in this 
dissertation. 
                           
 
 42 
Chapter 4 Quality of Service Requirements 
 
4.1 Quality of Service (QoS) Attributes 
In addition to understanding the various spectrum and data access techniques 
which impact the architecture of the radio, it is critical to gain a good understanding of 
the QoS (quality of service) requirements for the next generation mobile systems beyond 
IMT-2000. 
Usually, the QoS attributes for a particular application or service are: required 
throughput, maximum acceptable delay, maximum acceptable delay jitter, and maximum 
acceptable bit error rate.     Any change in radio architecture must take into account these 
QoS parameters so as to optimize their performance.  
The following issues must be taken into account as viewed from the end user?s 
point of view: 
? The end user only cares about the degree of QoS, and not about how it is 
provided; 
? It is the user that decides whether he is satisfied with the provided QoS or not; 
? The number of ?user defined/user controlled? parameters has to be kept at a 
minimum; 
? A derivation/definition of QoS attribute from the application requirements has 
to be simple enough; 
? End-to-end QoS has to be provided. 
 
                           
 
 43 
4.2 QoS Classes 
In UMTS standardization the following four service classes are defined:  
background class, interactive class, streaming class, and conversational class.   The 
primary difference between these four classes is the delay sensitivity of the traffic.   For 
example, the conversational class is most delay sensitive since the users would not 
tolerate any appreciable time of silence by the far end party.   The background class is the 
most delay tolerant.    Fig. 4.1 summarizes the UMTS QoS classes and their main 
characteristics.   The table also shows some typical applications for each class of service. 
 
?  Traffic class  Conversational 
class 
 
Real Time 
Streaming class 
 
 
Real Time 
Interactive class 
 
 
Best Effort 
Background 
class 
 
Best Effort 
Fundamental 
characteristics 
- Preserve time 
relation (variation) 
between 
information 
entities of the 
stream 
- Conversational 
pattern (stringent 
and low delay ) 
- Preserve time 
relation 
(variation) 
between 
information 
entities of the 
stream 
- Request 
response pattern  
 
-Preserve 
payload content 
-Destination is 
not expecting the 
data within a 
certain time 
-Preserve payload 
content 
Example of the 
application 
Voice, video 
games, voice 
telephony 
 Streaming 
video 
Web browsing, 
network games  
Telemetry, 
background 
download of 
emails 
 
Source: 3GPP 23.107 
Figure 4.1 UMTS QoS Classes 
 
                           
 
 44 
4.3 QoS Architecture 
Network Services are considered end-to-end  [13], this means from one Terminal 
Equipment (TE) to another TE. An End-to-End Service may have a certain Quality of 
Service (QoS) which is provided for the user of a network service.  As indicated above, it 
is the user that decides whether he is satisfied with the provided QoS or not. 
To realize a certain network QoS a Bearer Service with clearly defined 
characteristics and functionality needs to be set up from the source to the destination of a 
service.  A bearer service includes all aspects to enable the provision of a contracted QoS.  
These aspects are among others the control signaling, user plane transport, and QoS 
management functionality.  A UMTS bearer service layered architecture is depicted 
below in Fig. 4.2.  Each bearer service on a specific layer offers its individual services 
using services provided by the layers below. 
The UMTS Bearer attributes are: 
Traffic class (conversational, streaming, interactive, background)  
Maximum bit rate (kbps)  
Guaranteed bit rate (kbps)  
Delivery order (y/n)  
Maximum SDU size (octets)  
SDU format information (bits)  
SDU error ratio  
Residual bit error ratio  
Delivery of erroneous SDUs (y/n/-)  
Transfer delay (ms)  
Traffic handling priority  
Allocation/Retention Priority  
Source statistics descriptor (speech/unknown) 
 
                           
 
 45 
  
Source: 3GPP 23.107 
Figure 4.2 QoS Architecture 
 
 
The Radio Access Bearer (RAB) shown in Fig. 4.2 is the entity responsible for 
transporting radio frames of an application over the access network in UMTS.  The 
parameters of a RAB, namely the maximum bandwidth and the allowed frame sizes, can 
be configured according to the requirements of the application using it.  The 
understanding of the RAB is of most interest to this dissertation as it has direct 
implications in the radio design. 
                           
 
 46 
3GPP has defined the concept of Radio Access Bearer (RAB) as a user plane 
connection provided by the UMTS Terrestrial Radio Access Network (UTRAN) between 
a User Equipment (UE) and the Core Network. The general characteristics of a RAB 
(data rates, QoS, etc) are normally set by the Core Network (CN) based on subscription 
and/or requirements of the media or set of medias using the RAB.  The actual 
configuration for a RAB is decided by UTRAN based on the RAB information received 
from the CN. 
The RAB configuration has a direct impact on network resource usage.  The more 
suited the RAB configuration is to the actual pattern of the data being transferred, the 
more efficient the RAB is in terms of usage of network resources.  Choosing proper RAB 
configurations is key to UTRAN, given the high cost of last-mile transport (Iub interface) 
and the rather limited radio resources  [14].   The RAB bandwidth determines the QoS 
received by the application, and the set of allowed frame sizes for the RAB determines 
the amount of bandwidth wasted for padding.  Given a certain application, it is crucial to 
define its RAB well adjusted to its requirements;  too small bandwidth will result in a bad 
quality, while too large bandwidth or improper frame sizes will result in a waste of 
resources.  From the above it follows that a RAB design optimized for the various 
applications (VoIP, streaming video, web browsing, etc) is an important issue in UMTS 
networks. 
The application may specify its QoS requirements to the network by requesting 
the radio access bearer (RAB) with any of the specified traffic type (conversational, 
streaming, interactive, background), maximum transfer delay, delay variation, bit error 
                           
 
 47 
rates, and data rates.  In practice, it should be possible to define the main RAB 
characteristics from the service quality requirements: 
? The transmission rate of RAB should be determined by the bandwidth 
requirement of the information source. 
? The choice of dedicated or shared RAB should be based on the requirement 
for the maximum delay and delay jitter. 
? The SNIR requirement, channel coding, and interleaving for the RAB shold 
be based on the BER requirements. 
 
4.4 Chapter 4 Summary 
The QoS problem involves integrating delay-sensitive applications such as voice, 
audio, and video onto a single network with delay-insensitive applications, such as e-
mail, fax, and static file transfer.   3GPP has defined the concept of Radio Access Bearer 
as a user plane connection provided by the UMTS Terrestrial Radio Access Network 
between a User Equipment and the Core Network.    The RAB parameters must be 
optimized for each type of traffic and QoS requirements.      
The radio architecture for next generation mobile networks needs to consider the 
optimum way to implement the RAB function.   Should it be handled by SDR or should it 
be handled by a ?co-processor? co-located next to the radio interface, namely, the RF 
Front End?   As explained in the next chapter, SDR technology may be prohibitive for 
certain applications where ASICs may be most suited for cost reasons.   Yet, the 
possibility of having some ?SDR? functions embedded closer to the RF front end may 
prove advantageous for a wide variety of applications without adding too much cost to 
the architecture. 
                           
 
 48 
Chapter 5 Trends in Base Station Architectures 
 
5.1 Introduction 
This chapter provides an overview of recent trends in the architecture and design 
characteristics of base stations and what technology challenges most be overcome in the 
design of next generation base stations.   This overview is presented here to stress the fact 
that the radio front end of the future (c. 2012) will have to be that of a programmable, 
smart RF front end which is flexible enough to work with all the various spectrum 
allocations, data access methods, and QoS requirements for various applications while 
being small in size and very low cost.  
The first generation of W-CDMA base stations were launched in 2000.  Since 
then, the increasing size of W-CDMA networks has led to greater challenges in terms of 
network deployment, and 3G base stations are now required to balance high performance 
with low cost,  possess flexible deployment capabilities and be future-oriented.  To meet 
these demands, new generation base stations featuring digital power amplification, multi-
carrier technology, HS D/U PA support and open architecture have come into being. 
Additionally, increased diversity in base station form is being witnessed.  
Traditional base stations provide macro and micro base station services.  Large-sized, 
high capacity and powerful macro base stations are suitable in most cases, while smaller,
                           
 
 49 
less powerful micro (and pico) base stations are applicable for densely-populated areas 
and indoor coverage (see Fig. 5.1 and Fig. 5.2).   
 
 
 
 
Source:   Advanced Topics in Wireless Communications, SMU 
Figure 5.1 Various Forms of Base Stations 
 
 
                           
 
 50 
 
                    Source:   http://en.wikipedia.org/wiki/Base_station 
Figure 5.2 Macro and Pico Cell Base Stations 
 
 
5.2 Distributed Base Stations  
Distributed base stations are becoming increasingly prevalent given their great 
power capability and high capacity.  For small areas such as a cluster of households, 
concept base station Pico cells are proving highly suitable.  While an increasing number 
of base station types are emerging that boast an array of features, it is distributed base 
stations which no doubt represent ?next-generation base station? development.   This is of 
                           
 
 51 
interest for this dissertation as it validates the view that the RF front end will have to 
evolve on its merits towards a much more flexible, smaller size, and lower cost form 
factor platform as explained below. 
 
5.2.1 CPRI and OBSAI Architectures 
RRH (Remote Radio Head) and baseband pool technologies are popularized in 
recent products, and this in turn greatly promotes the development of distributed base 
stations.  Previously, base station manufacturers were accustomed to making their 
products in isolation.  This resulted in an increased variety of modules and specifications 
for connecting them, and module suppliers had to make different processors or radio 
frequency devices to adapt to different base station products.  Costs, therefore, rose.  Not 
surprisingly, it was logical to move toward standardized base stations interfaces that 
allow inter working between a large number of third-party module manufacturers and 
digital interfaces, with possibilities for multiplexing modules.  This gave rise to two 
organizations spontaneously formed by industry vendors. 
Nokia, NEC, LG and Samsung established OBSAI (Open Base Station 
Architecture Initiative) in October 2002.  The organization?s objective was to standardize 
wireless base stations? architecture, internal interfaces, control modules, transport, 
baseband and radio frequencies.   However, OBSAI was disadvantaged by possessing 
only one strong base station vendor - Nokia.    
In June 2003, Ericsson, Huawei, NEC, Nortel Network and Siemens jointly set up 
an organization called CPRI (Common Public Radio Interface), which also aimed to 
identify a universal standard of key internal interfaces; specifically interfaces between 
                           
 
 52 
baseband and RF modules.  The organization?s strength was increased by the fact that the 
five CPRI members were all large-scale base station vendors. NEC ? a founding member 
of OBSAI ? turned to support CPRI less than a year after its inception. Numerous 
manufacturers have continued to join the organization, which now has over 100 
members. 
 
5.2.1.1 CPRI 
According to CPRI, a base station can be divided into two units: BBU (Baseband Unit), 
which is also called as REC (Radio Equipment Control) and RRU (Radio Frequency 
Unit), which is also called RE (Radio Equipment).  CPRI?s interface mainly features 
baseband and radio frequency separation.  Small in size, easy to install and offering all-
round function provision with low power consumption,  the BBU can coexist with current 
sites and support expansion in the cascade mode.   RRU is small, light, simple and can be 
directly mounted on a pole or a wall near the antenna thus maximizing its convenience. 
CPRI defines interfacing between BBU and RRU, thus making it possible for different 
manufacturers? BBU and RRU modules to be interconnected.  Fig. 5.3 shows the 
partitioning of the CPRI architecture.  
 
                           
 
 53 
 
Source:   http://www.cpri.info/spec.html  
Figure 5.3 Partitioning and Interfaces defined by CPRI 
 
5.2.1.2 OBSAI 
OBSAI divides a base station into four functional modules: transport, processing, 
radio and control.  The transport module provides base stations with standard external 
network interfaces including IP and ATM. The processing module is used to process 
baseband signals.  The radio module receives, transmits and amplifies RF signals, and 
converts them from digital baseband signals to analog RF signals.   The control module 
provides a system clock for timing, manages and configures base station devices, 
monitors the operational status of the other base station modules and submits a report to 
the network equipment management system.   Fig. 5.4 illustrates the OBSAI-defined 
open interfaces. 
                           
 
 54 
 
Source:   http://www.obsai.org/obsai/obsai/ 
Figure 5.4 Open interfaces defined by OBSAI 
 
The superiority of a distributed base station is easily apparent. Equipment 
networking is more flexible due to the distributed base station?s smaller size, and the 
equipment involved enhances the value of site resources given their current short supply.  
The construction of equipment rooms and their daily maintenance requires numerous 
personnel and a high level of financial investment from operators.  Apart from cost 
savings, RF module and baseband transport separation grants greater capacity allocation 
flexibility and base station expansion potential.  Transport via optical fibers can be 
realized between different base stations or management modules, which will greatly 
improve an entire network?s performance.  These reasons underpin the trend for 
equipment manufacturers to produce this type of base station. 
                           
 
 55 
However, given that the CPRI and OBSAI standards were created by 
manufacturers, they remain non-mandatory.  In terms of the standards themselves, CPRI 
only defines layer-1 and layer-2 of the CPRI interfaces between BBU and RRU, with 
nothing defined for layer-3.   In layer-2, manufacturer-defined fields exist, but there is no 
definition for O&M.   From a practical perspective, there is still a long way to go before 
inter working between RF and base band modules and the openness of current repeater 
equipment can be achieved. 
 
5.3 Chapter 5 Summary 
Two trends in the design of base stations are clearly happening and both are 
challenging the architecture of the radio.   One trend is related to size reduction of the 
radio and the other is related to the separation of the radio from the base band module.   
Two industry initiatives are addressing these trends: the CPRI and OBSAI initiatives.     
The fact that the radio is being separated from the base band also validates the view that 
the radio will in fact become ?smarter? and more flexible so that it can communicate with 
different base station implementations as well as with the various air interface standards.   
                           
 
 56 
Chapter 6 4G Requirements 
 
6.1 4G Definition 
Reference  [15] attempts to define 4G as ?a fully IP-based integrated system of 
systems and network of networks achieved after the convergence of wired and wireless 
networks as well as computer, consumer electronics, communication technology, and 
several other convergences that will be capable of providing 100 Mbps and 1 Gbps, 
respectively, in outdoor and indoor environments with end-to-end QoS and high security, 
offering any kind of services anytime, anywhere, at affordable cost and one billing.?  
This definition matches well the characteristics of 4G as summarized by Dr. Hossam 
Hmimy, from Ericsson, in a lecture at TranSwitch Corporation, Shelton, CT, as shown in 
Fig. 6.1. 
 
6.2 4G Requirements 
4G (also known as ?beyond IMT-2000?) handsets and base station architectures 
will have to satisfy the following requirements: 
? Adaptive and scalable air interfaces; 
? Reconfigurable ambient networks; 
? Highly available backbone technologies (e.g., fiber ring, MPLS); 
? User friendly multimedia interfaces and context-aware technologies; 
? Flexible platforms 
                           
 
 57 
 
 
Source:   Dr. Hossam Hmimy, Ericsson, 2006 
Figure 6.1 4G Requirements 
 
As explained earlier, 4G requires data rates of 100 Mbps for full mobility and 1 
Gbps for low mobility indoor coverage.   To satisfy this requirement, a highly reliable 
wireless access technology is needed (e.g., OFDM, MC-CDMA) as well as efficient 
coding schemes, reconfigurable radio, multiple antennas, and adaptive power control.    
In addition, to enable adaptability to various network conditions and scalability (in terms 
of data rate and QoS), new air interfaces (like HSDPA+, ultra-wideband, optical wireless, 
etc) will have to be investigated in terms of coexistence with other radio systems, 
multiple access capability, resistance to interferers, low cost, low power, and other 
implementation issues. 
                           
 
 58 
Reconfigurable networks imply the design of new communications mechanisms 
at different layers.   The reconfiguration actions have an impact on various levels of 
mobile systems architecture and introduce high complexity that has to be handled by 
some reconfigurability management intelligence, which could be distributed.    This 
reconfigurability requirement mandates compatibility to legacy networks and that 
compatibility into future IP-based networks and future network topologies (such as ad 
hoc networks) also be embedded in them.    One possible scenario of these reconfigurable 
ambient networks is that more and more algorithms and concepts will have to be 
implemented at higher layers (that is, above the physical layer) in an efficient manner as 
variations in topology increase as well as user requirements and resource and power 
constraints persist.    Such migration of functions into higher the higher layers (2 and 3) 
will continue to put a strain to the processing power of DSPs which are at the heart of 
SDR architectures.     Thus, the issue of distributed reconfigurable management 
intelligence becomes more relevant. 
 
6.3 Backhaul Interworking 
Wireless systems need to be interconnected with each other via the backhaul wire 
line network (See Fig. 6.2).   While most of today?s existing backhaul technologies 
include legacy T1/E1 interfaces, the future all-IP-network will require backhaul 
technologies that are IP-compatible (e.g., Ethernet-over-PDH, Ethernet-over-SDH, etc).    
This trend is already happening as the backbone network migrates to IP.      
The network concepts are also moving from network-centered towards person-
centered solutions (wireless personal area networks, or WPAN).  WPANs will enable a 
                           
 
 59 
person to establish connectivity from his or her personal wireless devices to the outer 
world.    Since the WPAN personal networks may utilize various wireless technologies, 
mechanisms for the provision of interoperability between existing and forthcoming 
wireless infrastructures become mandatory.  Thus, the coexistence of the various wireless 
devices will require further research of the impact of this requirement at the physical 
layer, medium access control layer, radio frequency channel models, and data traffic 
models. 
 
 
Source:  Courtesy of TranSwitch Corporation, 2008 
Figure 6.2 Core Network Backhaul Access Technologies 
                           
 
 60 
6.4 Flexible Platforms Requirement 
The above 4G requirements, that is, adaptive and scalable air interfaces, 
reconfigurable ambient networks, and interoperability and access of the WPANs to the 
backbone network, dictate a fundamental need for flexible (programmable, 
reconfigurable, adaptable) platforms as a key requirement for 4G networks. 
Flexible platforms will be necessary to cope with the broad range of end user 
requirements in a context-aware environment.   A flexible platform will require a flexible 
architecture.    A flexible architecture implies a high degree of re-configurability achieved 
both in hardware and software.     Beyond the current FPGA and NPU (network 
processing units) technology, newer and more advance solutions will have to be 
developed and optimized for packet-based traffic.    Mixed architectures utilizing DSPs 
and ASICs will also be utilized for some solutions but the increased functionality that 
will be required of these components will most likely necessitate more innovative (and 
possibly distributed) solutions.   For example, 3G terminals already require some 100 
MIPS (million of instructions per second) as compared to 10 MIPS for a 2G terminal, and 
the requirements for a 4G terminals, as explained below, will be much greater.  Thus, a 
key area for further research is how much of the layer 1 and 2 processing can be pushed 
back as close to the RF front end module as compared to the base band module so as to 
offload the various DSP, FPGA, or ASICs in the base band module to process other layer 
3 functions.    
In addition, the introduction of ?multi system? flexible architecture platforms will 
require highly reconfigurable antennas and RF subsystems given the variety of spectrum 
frequencies the network of networks vision in the 4G requirement.   Thus, new RF 
                           
 
 61 
architectures will be needed (for example, zero-IF, tunable MEMS filters for optimum 
cost and performance, etc).  
In summary, a flexible platform for 4G systems will have to meet the following 
requirements: 
? A flexible service platform 
? Co-existence with other wireless systems 
? Co-existence (access to) various core network backhaul access technologies 
? Automatic service discovery 
? End-to-end QoS and security 
? Compatibility with wireless legacy networks 
? Upgradability to newer or enhanced wireless networks 
? Scalable connectivity 
? Capacity enhancements-ready 
? Spectrum efficiency enhancements 
 
6.5 4G Handsets 
The analysis of 4G requirements would not be complete without an overview of 
the requirements for 4G terminals (that is, the handsets).   Per  [15], the critical success 
factors for future 4G terminals are service convergence, user-centric interface, and 
portable intelligence.   Of these, the most relevant to the subject of this dissertation is that 
of service convergence.   Service convergence is important not only in the sense of a 
terminal?s ability to support multiple services, some of them concurrently, but also from 
the standpoint that many services from different providers will have to be compatible 
with a large variety of terminals with different capabilities.    This implies that both the 
base station and the terminals will have to be capable of automatic service and capability 
discovery. 
Two clear market trends are seen in the realm of terminals:  single-mode versus 
multi-mode (multiple functions) terminals.    Based on current market trends, it appears 
                           
 
 62 
that both of these terminal options will continue to co-exist in the future.   However, the 
promise of 4G is more data-centric applications like video, audio, GPS, email (and other 
applications) to the handset.   Thus, below are the expected basic requirements of 
hardware and software technologies for the 4G terminals (the trends relevant to this 
dissertation paper appear in bolded fonts): 
6.5.1 Hardware 
? Multimedia; 
? Modem, processor: enhanced processing power, embedded applications 
processor; 
? Display, camera, storage, battery: all requiring high-quality and high-
performance because of the rich-media services; 
? Flexibility to various environments; 
? RF, antenna: support of multiband and multimode, support of advanced 
technologies (e.g., SDR, MIMI, smart antennas); 
? Modem, system-on-chip (SOC) (i.e., modem and RF integrated onto one 
chip) 
? Multiple connectivity: supporting several communication standards (e.g., 
W-CDMA, CDMA-2000, Wireless LAN, Bluetooth, WiBro, etc); 
? Active development of alternative technologies; 
? Display, keypad: maximize portability and performance (e.g., flexible display, 
virtual keypad, virtual display); 
? Storage, battery: overcome the limitations of current technology (e.g., small 
DHH, MRAM, FRAM, fuel cell); 
? Convergence; 
? Camera: auto-focusing, several mega pixel resolution, and optical zoom; 
? Various technologies and services converged into the mobile terminal such as 
media, biometric, health, leisure, and fashion. 
 
6.5.2 Software 
? General purpose, flexible; 
? Connectivity protocol: seamless connectivity among W-CDMA, CDMA, 
WLAN, Bluetooth, and WiBro; 
? Middleware: wider acceptance of middleware platform, such as Java and 
Brew; 
? Operating system (OS), codec; open OS (Symbian, WinCE, Palm) and general 
purpose Codec (H.264); 
? Standardization is very competitive; 
                           
 
 63 
? Middleware: Java versus Brew; 
? OS, Codec: Symbian verus MS, H.264 versus MPEG4, MP3 versus 
AAC/WMA 
 
At present, terminals are only capable of accessing one network at a time and a 
terminal is capable of switching a radio network among some standards, like GSM, 
UMTS, CDMA, WLAN, and WiMAX.   Parallel radio or multiple simultaneous access 
technologies will allow access to different radio networks at the same time.   For 
example, the user could be connected to the cellular network engaged in a voice service 
while simultaneously retrieving information from a WLAN.    Antenna evolution 
promises increased throughput and more optimal utilization of bandwidth using MIMO-
capable base stations and mobile terminals.   Terminals will be able to communicate with 
a large number of other devices using different techniques (e.g., radio and optical 
interfaces)  [16].   Also, many types of sensors will be embedded in future devices to 
allow control of applications and to sense the current user context. 
 
6.6 Chapter 6 Summary 
The various 4G requirements discussed in this chapter show the fundamental need 
for flexible (programmable, reconfigurable, adaptable) platforms as a key requirement 
for 4G networks.   This need will also propagate into the RF front end section of both the 
base stations and terminals because it will need to support multiple bands and air 
interfaces for seamless connectivity among W-CDMA, CDMA, WLAN, Bluetooth, 
WiBro, and other standards. 
 
                           
 
 64 
Chapter 7 Technologies and Research Aimed at Addressing the 4G Requirements 
 
The Mobile IT Forum  [17] has identified the following techniques and 
technologies aimed at addressing the future 4G requirements.    A few of the relevant 
technologies for this dissertation are later discussed in more detail. 
 
7.1 Wireless Transmission Techniques. 
The following are several technologies and research areas aimed at improving the 
high speed and large-capacity wireless transmission techniques. 
? Frequency re-farming:  Re-arrangement of allocated frequency bands to allow 
more efficient spectrum utilization and study of new bands 
? Advanced adaptive techniques to increase spectral efficiency:  Adaptive 
modulation and coding, adaptive control for automatic repeat request (ARQ), 
adaptive antennas 
? MIMO techniques for exploiting spatial multiplexing:  Fundamental for 
attaining very high data throughput 
? Multi-carrier techniques: Frequency orthogonal techniques (e.g., OFDM) as 
the baseline modulation technique for 4G systems used in general in 
combination with CDMA and TDMA techniques give more design flexibility 
and interference protection 
? Interference and fading mitigating techniques:  These two major impairments 
in typical wireless scenarios may cause serious degradation to systems 
performance.  Interference cancellation, equalization, and multi-antenna 
techniques, among others, are key approaches to be investigated 
? Error control techniques:  Advance FEC and ARQ schemes need to be 
incorporated to 4G systems to guarantee a desired signal quality regardless of 
the channel conditions 
? High-speed packet radio:  Important techniques are needed to improve the 
utilization of radio resources, allowing higher systems capacity and higher 
data rates 
                           
 
 65 
? Handover techniques:  As seamless communications is one of the most 
distinctive features of future systems, fast and reliable handover techniques 
for a large number of scenarios should be investigated 
 
7.2 Network Technologies 
The following are several technologies and research areas aimed at improving the 
network for 4G. 
? Radio access network technologies:  These techniques will enable the 
dynamic establishment of multiple continuous or isolated access cells. 
? Robust networks:  Variable-capacity networks able to adapt to the variable 
traffic demand conditions while keeping performance unchanged is a 
challenging area for research.   Multiple antennas, congestion control, and 
multihop communications have been identified as promising techniques for 
this purpose. 
? Ad hoc networks:  Rather than the centralized approach of conventional 
wireless communications where communications use a central node (base 
station) to access the network, an ad hoc network can be seen as a distributed 
approach where the signal can hop through several nodes (terminals ) to reach 
the destination.  Routing protocols, user and service discovery procedures, and 
security issues are the main research items. 
? Seamless networking techniques:  Allows seamless interconnection of a 
packet-based backbone network with radio access or other networks. 
? Approach link techniques:  A number of techniques used to connect base 
stations or remote stations to control stations by using optical means (fiber, 
free-space optics), microwaves, and so forth. 
? High-speed transport technology:  Techniques to transport and switch 
efficiently very high-speed packet data. 
 
7.3 Mobile Terminal Techniques 
The following are several technologies and research areas aimed at improving the 
mobile terminals for 4G. 
? Circuit and component technology:  Faster processing circuits, denser 
integration capabilities, smaller batteries, and display technologies with better 
image quality are identified an important goals to be fulfilled to fully comply 
with 4G system requirements and user expectations. 
? Battery technology:  Enhancement of battery capacity, efficiency, operational 
time, and size is crucial to succeed in producing future handheld terminals.  
                           
 
 66 
Low-cost batteries and environmentally friendly materials are also very 
important facts to consider. 
? Human interface:  This is recognized to be a very important research area, 
where natural ways of interaction between user and machine are of the utmost 
importance in order to make future 4G systems attractive.   Terminal form 
factor, wearable terminals, visual and voice interfaces, and the use of 
biometric information are among the most important research issues. 
? Terminal security techniques:   Techniques aiming to provide levels of 
security at least similar to those found in networks using public key-based 
infrastructure should be developed.  User authentication and data protection 
are also highly relevant research topics. 
? Terminal software:  Common operating systems, software platforms, and 
applications compatible with SDR should be developed. 
? Multi-system wireless terminal technologies:   It is envisaged that future 
terminals will be able to handle seamlessly short-range, wide-area, and 
broadcast communications using different technologies implemented in the 
terminal.  There is a series of technical challenges that need to be solved, 
including multi-system radio for multiple modulations, protocols and 
frequencies, multi-band antennas, practical realization of SDR and seamless 
operation between the systems. 
? Software defined radio:  Fundamental technology applicable in both base 
stations and mobile terminals permitting the use of multiple access 
technologies in a very efficient fashion.   Issues to be studied include 
transceiver architecture, basic implementation technologies, and basic 
transmitter and receiver technologies realizable in software, e.g., multimode 
modulation and demodulation techniques, multi-band transmit and received 
techniques, and so forth. 
 
7.4 Mobile Systems Technologies. 
The following are several technologies and research areas aimed at improving the 
mobile systems for 4G. 
? Quality of service:  These techniques encompass different methods to control 
and ensure fundamental communications parameters, like desired transmission 
speed, bandwidth, delay, and packet loss rate. 
? Mobility control:  Techniques aiming to provide and maintain QoS in 
environments characterized by mobility.  These include handovers and mobile 
Internet protocols. 
? Mobile multicast techniques:  Broadband transmission techniques for real-
time and non real-time broadcasting of information. 
? Location determination and navigation:  Reliable techniques for providing 
user coordinates, speed, and other navigation information in outdoor and 
                           
 
 67 
indoor scenarios.   A large number of applications could potentially exploit 
this kind of information. 
? Security, encryption, and authentication:  Techniques aiming to secure mobile 
communications by preventing any type of illegal use of the wireless 
resources. 
 
Of the above list, two technologies are most relevant to the topic of this 
dissertation paper:  multi-system wireless terminal technologies and software defined 
radios.     As indicated above, there is still a series of technical challenges that need to be 
solved, including multi-system radio for multiple modulations, protocols and frequencies, 
multi-band antennas, practical realization of SDR and seamless operation between the 
systems. 
 
7.5 Hardware Technology Roadmap 
Fig. 7.1 shows a Hardware Technology roadmap that shows several key hardware 
related components and technologies that are relevant to this dissertation as follows. 
 
7.5.1 Modems 
As the need for processing power is increased, the competition between integrated 
single-chip architecture and functionally divided multi-chip architecture will become 
more pronounced.   Single-chip architecture has the advantages of requiring less size and, 
in some situations, lower power consumption.   On the other hand, a multi-chip 
architecture is powerful in processing multimedia applications.     The need of fixed (e.g., 
ASIC) or flexible (e.g., reconfigurable) structures has to be assessed carefully when 
                           
 
 68 
   
Source:  [15] 
 
Figure 7.1 Hardware Technology Roadmap 
                           
 
 69 
designing the platforms for implementing the modem.   Cost and space limitations also 
need to be taken into account when considering single and multi-chip architectures. 
 
7.5.2 Radio Technology 
It is anticipated that new advanced technology will be developed to support multi-
band/multimode applications like W-CDMA, CDMA2000, EV-DV, HSDPA, DMB, 
WLAN, WiBro, Bluetooth, etc.   SDR and multi antenna technologies will become the 
core technologies, offering flexibility to the changing radio environment and higher 
performance.  In particular, SDR enables easier expansion of base station facilities, 
flexible reconfiguration of air interfaces, and terminals to support multiple standards.   
The flexibility offered by this technology is paramount to make possible truly seamless 
networks, and to support multiple communications schemes.   It also provides an 
effective means for expanding functions for multimode operations, as terminals and base 
stations can be reconfigured in the field.    In addition, multiple antennas at both ends of 
the wireless channel can be used to increase the data rate via spatial multiplexing and to 
improve the link quality through diversity. 
Thus, future terminals and base stations could become extremely complex, 
particularly if the trend of developing multipurpose terminals capable of providing 
various types of connectivity (e.g., multi-standard) and supporting diverse services 
persists.     A number of challenges for the design of such terminals have been identified 
and among these is RF integration and the issue of centralized versus distributed 
architectures as highlighted above. 
 
 
                           
 
 70 
7.6 Chapter 7 Summary 
As discussed in this chapter and as also shown in the hardware roadmap of Fig. 
7.1, it is anticipated that the modem processor and the DSPs required for SDR should 
eventually converged into a single piece of silicon.     It also shows that the radio 
technology will evolve into a multi-band RF chip.    Another possible scenario shown in 
the roadmap is the integration of both RF and DSP into a single chip.   However, the 
exact definition and implementation of the future radio architecture is not yet clear as 
more research and industry products in this area are developed.  
                           
 
 71 
Chapter 8 RF Front End Technology Requirements and Challenges 
 
8.1 The Vision for the RF Front End of the Future 
In order to provide wide coverage, mobility, and roaming, the base stations and 
terminals of the future (4G and beyond) will have to include cellular 2G (e.g., GSM) and 
3G (e.g., UMTS) transceivers working  in the 900 to 1,800 MHz bands.   High-speed 
local access could be provided by an 802.11 (Wi-Fi) transceiver in the 2.4/5 GHz bands.   
Basic short-range communications capabilities could be additionally carried out with a 
Bluetooth transceiver operating at 2.4GHz.  RFID is already being integrated in some 
handsets and allows very short-range instant connections for several purposes (e.g., 
payments, identification, and information retrieval).   RFID operates typically in the 13-
MHz band, though low and high-frequency versions are available (100 KHz, 1 GHz, 2 
GHz, 5 GHz).    
In addition to the above functions, and following the trend of designing terminals 
with as many features as possible, one can also see the integration of a conventional FM 
radio receiver (already a trend in many mobile phones), a terrestrial or satellite broadcast 
TV receiver (100-MHz to 1-GHz band), and a GPS/Galileo receiver (1.1- to 1.6 GHz 
band).   Though all of these devices will not be used simultaneously, several could be 
switched on in certain applications.   Terminal design could become the bottle-neck for 
the development of multi-standard, multi-featured terminals, as radio frequency 
interference between different transceivers could have detrimental effects on the 
                           
 
 72 
performance, in particular taking into account the physical limitation of having receivers 
and transmitters enclosed in a very small space.    Methods of interference management 
and control should be carefully applied at the design stage.    The situation becomes 
critical if one considers the fact that the transceivers (e.g., for communications) and /or 
receivers (e.g., radio, TV, GPS) may need to use different antennas. 
Moreover, in order to attain the expected level of performance, particularly high 
data throughputs, cellular and local area systems have to resort to the use of multiple 
antennas.   At high frequency bands (e.g., 5-GHz or higher) it is feasible to place several 
antennas on a small (handheld) terminal ? for example, an array of four or even eight 
antenna elements.   Typically, these antennas have to be distributed to cover virtually all 
the available main plane (largest surface) of the terminal, in order to maximize antenna 
element separation.  Obviously, future designers will face a very difficult task trying to 
efficiently integrate multiple receiving and transmitting elements into small form factors.   
This could even be impossible for certain combinations of multi-standard and desired 
form factors.   Fig. 8.1 shows the challenges already being faced by designers in 
integrating the various discrete components associated with the RF front end part of a 
mobile phone. 
                           
 
 73 
 
Source:   Dr. C. Nguyen, FCS 2005 Tutorial:  ?MEMS for Frequency and Tuning 
References?  
 
Figure 8.1 Miniaturization Challenge in Mobile Phones 
 
8.2 Cost Considerations 
There is another important factor against implementing several transceivers on 
one terminal: cost.   Even in the case where multiple-integration is technologically 
feasible, the relatively high cost associated with having several standards implemented 
remains one of the central issues to be solved.     However, there exists a different 
approach that has been extensively studied in recent years, and in the view of their 
proponents it could be, cost- and technology-wise, the most efficient way to go towards 
                           
 
 74 
making complex terminals a reality.     This solution is known as Software Defined 
Radio.    
Ideally, in the broadest sense, SDR allows on-the-fly modification of the RF front 
ends, base band processing, and even the MAC layer of the terminal, aiming to realize a 
given air interface by reconfiguring the system.   The degree of flexibility brought by 
real-time re-configurability opens up a new world of possibilities for users, operators, 
services providers, and terminal manufacturers.   Users can establish a connection in any 
network, allowing simple local and global roaming.   Users could also benefit from the 
low-cost terminals that this technology would eventually bring.    As new advanced 
services are introduced, software modifications can catch up with the new requirements.   
Hardware and software updates could be easily and wirelessly carried out by the user or 
by the operators.   Manufactures can also take advantage of SDR since large volumes of 
terminals (or base stations) with identical hardware (and fewer number of components) 
would have to be produced.   Even upgrades or changes in the terminals could be carried 
out simply.    Also operators and service providers could exploit this flexibility to better 
match their operation and services to use demand.   Vanu Inc.,  [18] has already made 
significant advances towards the implementation of SDR-based architectures for base 
stations, and the benefits of such architecture are clearly shown in Fig. 8.2.  
From the above synopsis of the challenges that terminal and base stations 
designers face, it can be seen that the RF Front End represents one of the biggest 
opportunities for cost reduction and logistics simplification for the base station and 
terminals of the future. While there is a lot of academic and industry research focused on 
software-defined radio architectures, SDR still does not solve the RF Front End cost 
                           
 
 75 
issue.  In fact, for SDR architectures to achieve their promised ?programmability? 
mantra, they need a true ?programmable? RF Front End.   The problem is that today?s RF 
front end implementations are not programmable but fixed to work with a specific RF 
frequency (e.g., 1,900 MHZ) for a specific mobile network standard (e.g., CDM2000 or 
W-CDMA).  To band-aid this issue, OEMs typically implement multiple RF Front Ends 
to give the illusion of a ?multi-mode? programmable system at twice, triple or quadruple 
the cost of a single-mode system. 
 
 
Source:  www.vanu.com 
Figure 8.2 Vanu?s Anywave Multi-Standard Base Station 
 
 
 
                           
 
 76 
8.3 Vanu, Inc. 
Even industry-ready software defined radio architectures have not yet solved the 
requirement to have multiple RF front end ?heads? for each of the bands supported.  
Vanu Inc.,  [18] has developed a very innovative software configurable architecture for 
base stations still requires multiple RF front end ?heads? for its software-based 
architecture as shown in Fig. 8.3. 
 
 
Source:  www.vanu.com 
Figure 8.3 Vanu?s System Level Base Station Architecture 
 
                           
 
 77 
Vanu?s story is best summarized by extracting it from their website  [18]: 
?Traditionally, wireless infrastructure manufacturers have developed 
single-standard, high-power, high-capacity networks to address cellular coverage 
needs. These dedicated networks required extraordinary financial investments that 
took a long time to generate positive returns. And as new technologies emerged, 
so did the need for another large dollar investment to build a new network to 
support an additional standard. Each technology, or wireless standard, required its 
own dedicated hardware platform and network build-out. 
Vanu, Inc. has developed the first commercially available software radio 
access network (RAN) that lets wireless carriers operate multiple standards 
simultaneously on a single platform. That means GSM, CDMA, iDEN and 
more? all operating simultaneously from a single base station. Vanu was 
founded in 1998 to enable wireless operators to operate and expand their networks 
with the ease, flexibility and cost-effectiveness of modern computing practices, 
using its innovative software radio technology, the Vanu Anywave? solution. 
The number of wireless standards and licensed frequency bands in use are 
multiplying, and the pace of innovation is accelerating. Carriers can?t afford to 
implement a dedicated network for each standard they wish to deploy, and the 
faster rate of change makes it harder to earn a return on investment within a 
reasonable period of time. Yet competitive pressures and consumer trends are 
requiring operators to support multiple standards, creating bigger financial 
challenges in the process. 
Additionally, the traditional high-power, high-capacity networks do not 
provide cost effective solutions for in-building enterprise and home coverage, the 
fastest growing wireless environments today. Carriers need a way to cost-
effectively adopt ? and profitably operate ? multiple standards using their existing 
spectrum, tower, antenna, shelter, etc. And this needs to be accomplished while 
successfully delivering RF coverage solutions for outdoor and indoor settings. 
Vanu has addressed these needs with its Anywave software radio technology. 
Vanu?s Anywave is the wireless industry?s first software radio access 
network (RAN) to obtain FCC certification and commercial deployment. Rather 
than processing wireless standards in single-purpose components ? the traditional 
method that requires costly, specialized signal processing hardware ? Vanu 
Anywave defines and processes each technology entirely in software. 
As a result, expanding an Anywave network with a new wireless standard 
is a fast and cost-effective, remote software download from a central location. 
Leveraging a common software platform that sits on IT grade or NEBS-compliant 
commercial off-the-shelf (COTS) computing servers, the Vanu Anywave solution 
enables carriers to operate multiple standards simultaneously on a single platform. 
Furthermore, because the Anywave solution is hardware agnostic, scalable up and 
down and highly portable, the solution can be packaged to fit the broadest range 
of coverage needs, including: 
? Outdoor, large-cell coverage that accommodates both urban and rural 
environments  
                           
 
 78 
? Outdoor pico cell filler coverage to selectively increase capacity in high-
density traffic locations  
? Pico cells with or without a distributed antenna system (DAS) for 
enterprise applications  
? Femto cells with broadband IP backhaul for home use  
? Mobile cell sites for temporary venues or disaster recovery applications  
New standards, new frequency bands, new competitors and demand for 
new services will continue to escalate. The Vanu Anywave solution enables 
operators to ?future-proof? their networks. If additional wireless standards are 
acquired or desired, they can easily be accommodated on the same network 
infrastructure through a remote, software download without the time-consuming 
expense associated with costly site visits. The same is true when the operator 
wants to add system capacity. Additionally, because the software-based Vanu 
Anywave RAN runs on COTS hardware, it?s able to ride the coattails of Moore?s 
Law where computing capacity doubles every 18 months?that?s price 
performance that only gets better over time. 
The Vanu Anywave system can help operators expand their offerings and 
reduce costs, which will help them to compete more effectively in a constantly 
changing marketplace. It?s what carriers have hoped for in the past, and what will 
enable them to succeed ? today and in the future.? 
 
Fig. 8.4 shows Vanu?s Anywave system architecture in more detail.  A typical 
Vanu Anywave RF head supports up to 16 carriers, depending on the cellular standards 
being used.   This design can, in many cases, provide significant heat dissipation and 
power loss savings compared to the power-inefficient combining networks used in legacy 
line card BTS architectures.   
The analog electronics in the RF head offers sufficient linearity and spurious free 
dynamic range (SFDR) to support any of a variety of cellular standards.  Selectable 
carrier frequency and bandwidth allows multiple carriers in different wireless standards 
to operate simultaneously (Fig. 8.5), for example, the transmission and reception of GSM 
and CDMA channels at the same time through the same RF head. The carriers in use can 
even be changed dynamically, for instance to match the capacity of each standard to the 
number of customers currently in the coverage area of the BTS. The transceiver in the RF 
                           
 
 79 
head performs digital channelization and down-conversion on receive and digital up-
conversion and summing on transmit to reduce the bandwidth requirements on the RF 
sample interconnect. 
 
Source:  www.vanu.com 
Figure 8.4 Vanu?s Anywave Multi-Standard Base Station 
 
                           
 
 80 
 
Source:  www.vanu.com 
Figure 8.5 Vanu?s Selectable Carrier Frequency Architecture 
 
8.4 Chapter 8 Summary 
Vanu?s architecture clearly illustrates the advances being made as of 2008 in base 
station architecture using SDR techniques.    It also highlights the importance of RF 
?heads? with ?selectable carrier frequency and bandwidth? which allows multiple carriers 
in different wireless standards to operate simultaneously.   The challenge that both 
industry and academia research now face is how to both reduce the cost and size of these 
RF ?heads? and scale these down so that the same architecture can be implemented in the 
much smaller-sized femtocells and mobile handheld terminals of the future. 
                           
 
 81 
Chapter 9 Current Technology Limitations of Software Defined Radios 
 
9.1 SDR Defined 
Software defined radio (SDR) is the term adopted by the SDR Forum  [19] which 
is an international body looking at the standards aspects of software radios.    The term 
software radio has become associated with a large number of different technologies.   The 
term is usually used to refer to a radio transceiver in which the key parameters are 
defined in software and in which fundamental aspects of the radio?s operation can be 
reconfigured by the modification of that software.       
A multi-standard terminal (MST) refers to a terminal which is capable of 
operation on a number of differing air interface standards, and it may also be 
implemented using SDR techniques.  The benefits of a reconfigurable multi-standard 
terminal include economies of scale, global roaming, service upgradeability and adaptive 
modulation and coding to cope with the prevailing channel or traffic conditions.    The 
operational characteristics of an ideal multi-standard terminal include the following 
operations: 
? Software-Definable Operation.  This is a key requirement to achieve above 
stated benefits as the terminal must have the ability to be reconfigured either: 
during manufacture, prior to purchase, following purchase (e.g., after-market 
software), in operation (e.g., adaptation of coding or modulation), or 
preferably all four.   This impacts primarily the digital and base band section 
of the terminal and will required the use of reprogrammable hardware as well 
as programmable digital signal processors in a power and cost-effective 
manner.
                           
 
 82 
? Multi-band operation.  The ability to process signals corresponding to a wide 
range of frequency bands and channel bandwidths is a critical feature of a 
MST.    This will impact heavily on the radio frequency segments of the 
terminal, and it is this area which is arguably the main technology limitation 
on software defined radio implementation at present (although processor 
power consumption and cost are still both major issues for SDR 
? Multi-Mode Operation.   This is the ability to change mode and, consequently, 
modulation, coding, burst structure, compression algorithms, and signaling 
protocols. 
 
The concept of a multi-band or general coverage terminal is really an 
extension of the basic software defined radio concept into that of a broadband flexible 
architecture radio, since the basic re-programmability and adaptability aspects of 
operation do not depend upon multi-band coverage.    It would be possible for 
example, to construct a useful SDR-based radio which operated in the 800-/900-MHz 
area of spectrum and which could adapt between AMPS, GSM, DAMPS, PDC, and 
CDMA. 
 
9.2 Current SDR Limitations 
There are many issues that must be addressed in determining if a software defined 
radio is realistic and also to what extent it is flexible.   For example, it is possible to 
create a single-band software defined radio with a narrowband channel restriction 
relatively easily  [20].   Coping with a wider channel bandwidth and operating in multiple 
bands in different parts of the spectrum is much more difficult, but nevertheless essential, 
for a combined GSM/PCS/WCDMA/TD-SCDMA/WiBRO/WiFI/Bluetooth handset, for 
example. 
Fig. 9.1 will used next to discuss the issues and current limitation of software 
defined radio architectures.   Note that the A/D converter in Fig. 9.1 is assumed to have a 
                           
 
 83 
built-in anti-alias filter and that the D/A is assumed to have a built-in reconstruction 
filter.     
 
Source:   [21] 
Figure 9.1 Ideal Software Defined Radio Architecture 
 
The ideal software defined radio has the following features: 
? The modulation scheme, channelization, protocols, and equalization for 
transmit and receive are all determined in software within the digital 
processing subsystem.     A DSP appears in Fig. 9.1, but there exists a variety 
of applicable signal processing hardware solutions for this element. 
? The ideal circulator is used to separate the transmit and receive path signals, 
without the usual frequency restrictions placed upon this function when using 
filter-based solutions (e.g., a conventional diplexer).   This component relies 
on ideal (perfect) matching between itself and the antenna and power 
amplifier impedances and so is unrealistic in practice, based upon typical 
transmit/receive isolation requirements.   Since the primary alternative (a 
diplexer) is very much a fixed-frequency component within a radio, its 
elimination (or ?transparency?) is a key element in a multi-band or even multi-
                           
 
 84 
standard radio.   Note also that the circulator would have to be very 
broadband, which most current designs are not. 
? The linear (or linearized) power amplifier ensures an ideal transfer of the RF 
modulation from the DAC to a high-power signal suitable for transmission 
with low (ideally no) adjacent channel emissions.   Note that this function 
could also be provided by an RF synthesis technique, in which case the DAC 
and power amplifier functions would effectively be combined into a single 
high-power RF synthesis block. 
? Anti-aliasing and reconstruction filtering is clearly required in this 
architecture (but is not shown in Fig. 9.1).   It should, however, be relatively 
straightforward to implement, assuming that the ADC and DAC have 
sampling rates of many gigahertz.   Current transmit, receive, and duplex 
filtering can achieve excellent roll-off rate in both hand portable and 
(especially) base-station designs.   The main change would be in transforming 
them from a band-pass (where relevant) to low-pass designs. 
 
 
To better analyze the difficulties in achieving the ideal software defined radio, the 
set of specifications in Fig. 9.2 can be used as an example.    Those specifications can 
also be used as representative for a software defined radio femtocell.   The current 
difficulties and limitations preventing the realization of the specifications in Fig. 9.2 will 
be summarized next. 
 
9.2.1 Antenna.   
A frequency range of almost 5 octaves is required, together with a realistic 
gain/loss figure of around 0 dBi.   Combine this with the usual (handset) requirements of 
small size, near-omnidirectional coverage pattern (typically, excluding the users head), 
and low cost, and the physical realization of this component becomes extremely 
challenging. 
                           
 
 85 
Parameter Value Notes
Frequency Coverage 100 MHz - 2.2 GHz This would cover most PMR, cellular, PCN/PCS, 
mobile satellite, and UMTS bands worldwide
Receiver dynamic range 0 dBm to -120 dBm This must not only cope with fading and in-band
(based on a 25-kHz interferers, but any signals in the above frequency
equivalent channel bandwidth) range
Transmit power output 1W This is reducing as time progresses and health
fears increase, but most systems stillrequire this
power level (many PMR systems require more).
Transmit adjacent channel power -75 dBc This figure is slighly in excess of most known
specifications in this area (e.g., TETRA)
Transmit power control range 70 dB Most CDMA systems, for example, require a
large power control range
Transmit power ramping range 75 dB DECT requires 68 dB and is probably the toughest
requirement in this area
Channel bandwidth 5 MHz Based on the 3GPP WCDMA standard for UMTS
Receiver image rejection 60 dB Based on an interpretation of the TETRA specifi-
cation  
Source:   [21] 
Figure 9.2 Basic Specifications for a Handportable Software Defined Radio 
 
 
9.2.2 Circulator or Duplexer.   
This is discussed in more detail in the next chapter of this dissertation.   However, it 
needs high isolation and broadband coverage range.   In the case of a conventional, filter-
based duplexer, this latter requirement (that is, broadband coverage range) is essentially 
impossible to achieve with current technologies. 
 
9.2.3 A/D Converter.    
The sampling rate of this converter, if it is to Nyquist sample directly at RF, 
would need to be at least 4.4 GHz and, in reality, much more (to allow for a realistic anti-
alias filter roll-off and real-world converter performance).    For a 4G spectrum with 
frequencies in the 3- to 6-GHz, the sampling rate is much higher.   If, however, the 
converter is permitted to under sample, the required sampling rate drops dramatically.   
The required sampling rate could fall to 20 MSPS (based on two-times Nyquist band-pass 
sampling), assuming that the RF filtering and ADC analog input were up to the task.    
                           
 
 86 
This would lead to an input bandwidth requirement extending to 2.2 GHz and a 
resolution of around 20 bits (from the receiver dynamic range requirement.   Note that if 
a synthesizer and conventional down conversion are employed (in place of band-pass 
sampling), this resolution is available at very low cost in the form of digital audio 
converters.   Up to 200 kHz of channel bandwidth can be accommodated in this way, 
relatively easily and cheaply (based on I/Q down conversion prior to the A/D converters). 
 
9.2.4 D/A Converter.   
This component is currently realizable, although with a relatively high-power 
consumption, again assuming that conventional up conversion is employed and that 
power control is employed either prior to or within the linear power amplifier.   A 
resolution of 12 to 14 bits at 20 MSPS would be required.   IF output devices are also 
now increasingly common, and the available IFs are increasing as technology improves.   
Current devices are capable of operation at an IF in the hundreds of megahertz region; 
however, this will improve over the coming few years to the point where RF output 
frequencies (e.g., 800/900 MHz and 1.9/2.1 GHz) will become a reality, at a realistic 
cost. 
 
9.2.5 Receiver Anti-alias Filtering.   
Based on the two times Nyquist sampling converters discussed above, an 
attenuation of 60 dB is required around 18 MHz from the channel edge.   This would be 
extremely difficult, if not impossible, to achieve in a bandpass filter capable of tuning 
from 100 MHz to 2.2 GHz.  With the architecture of Fig. 9.1, this component presents a 
serious challenge and strongly indicates that a synthesizer-based down-conversion 
                           
 
 87 
mechanism would almost certainly need to be employed in a software defined radio for 
the foreseeable future.   Improvement in sampling rates (for a give converter resolution) 
will, however, allow this requirement to be relaxed and may enable some limited forms 
of  SDR to be realized without such high-performance filtering needing to be included. 
 
9.2.6 DSPs and Equivalent Technologies 
Technology in this area is progressing very rapidly and the primary issue at 
present is that of power consumption (for handset operation).   Combination of 
reconfigurable hardware (e.g., FPGAs) and fully software programmable processors are 
likely to yield the best performance in terms of power consumption, although other, 
newer architectures are also strong challengers in this area (e.g., massive parallel arrays). 
 
9.2.7 RF Power Amplifiers 
Considerable research has been directed at the linearization of power amplifiers in 
recent years, and a number of candidate techniques exist.   Many narrow band systems 
have employed the Cartesian loop technique, achieving up to -70 dBc inter-modulation 
product levels  [20].   For broader bandwidth systems, RF pre-distortion, digital pre-
distortion and feed-forward techniques have also been used.   At present, digital pre-
distortion is a realizable solution and fits well with the architecture of a software defined 
radio.    
 
9.3 Chapter 9 Summary 
The specification outline in Fig. 9.2 and the components required to realize it are 
clearly not all available with current technology and may not be achievable, in many 
cases, for a considerable period (if ever).   It is therefore necessary to examine other 
                           
 
 88 
architectures and/or restrictions in the specifications contained in Fig. 9.2, in order for a 
software-defined radio to become a reality in the short or medium term.      In the next 
chapter of this dissertation, the author proceeds to examines in more detail the various 
architectures for the implementation of SDR-based multi-band flexible transceivers, with 
particular attention to the RF front end issue highlighted above. 
                           
 
 89 
Chapter 10 Multi-Band Flexible (MBF) Transceiver Architectures 
 
10.1 Introduction 
The preceding chapters have explained the compelling reasons for flexible radio 
architecture.    It has been shown that future 4G systems will require the co-existence of 
multiple modes and multiple radio frequencies and that this requirement will be needed at 
both ends of the network: both at the base station and the handset.   It has also been 
shown that while software defined radio (SDR) technology is very promising for the 
implementation of flexible radios, there are other radio design issues that are not being 
addressed by DSP-based SDR techniques and thus require further research. 
This chapter now focuses on a very specific research area that must be solved 
within the next three to five years in order to address the specific issue of multi-frequency 
filter technology required for a true multi-band flexible (MBF) transceiver.    This chapter 
starts by presenting at a high level several possible MBF architectures at both the receiver 
and transmitter and concludes by showing that the requirement for a single-package 
multi-band tunable filter must be addressed as an alternative for MBF transceivers. 
The concept of flexibility in a receiver breaks down into two main areas:  that of 
the flexibility in the modulation format, coding, and framing, and that of flexibility in 
terms of RF frequency (that is, the ability to cover multiple bands, or provide general 
coverage, which is defined as covering all bands between a declared minimum and 
maximum frequency).   This latter area, frequency flexibility, is certainly the more 
                           
 
 90 
challenging of the two and is a concept which is the subject of much research.    The 
former area has been much more widely addressed and most commercial 
communications receiver designs employ many of its basic principles, even if they do not 
yet aim to provide a wide choice of modulation formats. 
 
10.2 Receiver Single-Carrier Designs 
 Receiver single-carrier designs have been available for some time and are 
widely explained in the literature (see for example  [22]).   They are briefly reviewed here 
for completeness of the study as such single-carrier designs do not address the 
requirements of future generation RF front ends as explained in the preceding chapters. 
 
10.2.1 Analog Quadrature Receiver Design 
A simplified, single-band flexible receiver architecture is shown in Fig. 10.1.   Its 
flexibility stems from the use of a DSP as the base band demodulation function; it can 
thereby demodulate any modulation format within its processing and data conversion 
bandwidth.   Note that if a variety of different modulation formats are to be received by 
the same radio architecture, then the desired channel bandwidths must be carefully 
considered.   For example, if the receiver is to handle both GSM (9200-kHz bandwidth) 
and PDC (40-kHz bandwidth), then the IF Filter, anti-alias filters, and A/D input 
bandwidth (and hence sampling rate) must be chosen based on the wider of the two 
bandwidths (more than 200 kHz for the IF filter and more than 100 kHz each for the anti-
alias filters and A/D converter input bandwidths). 
 If this is done, then it is possible for the IF chain and A/D converters to 
experience a wide dynamic range of signals when in PDC mode, since at least six PDC 
                           
 
 91 
channels could appear in the IF bandwidth.   This would lead to a requirement for a 
greater instantaneous dynamic range in the IF and A/D converters than might otherwise 
be necessary.   This problem may be overcome by the use of flexible base band filtering 
employing, for example, switched-capacitor techniques using MEMS technology, 
although such techniques have other inherent problems (to be discussed in the next 
chapter).   
 
 
Source:   [21] 
Figure 10.1 Simplified Linear Receiver Architecture 
 
 
Note that the flexibility of the processing of the DSP allows many of the 
conventional receiver functions to be implemented in that part of the system.   Some 
examples included: 
? Detection/demodulation; 
? Fast AGC (e.g., by feed-forward techniques); 
? AFC, either by the use of internal (to the DSP) oscillators for frequency 
translation, or by pulling of the external frequency standard (not shown in 
Fig. 10.1); 
? Companding for analog voice; 
? De-interleaving and decoding/error correction of data. 
 
                           
 
 92 
10.2.2 Digital IF Receiver Architecture 
An alternative, single-band flexible receiver architecture is shown in Fig. 10.2.   
 
 
Source:  [21] 
Figure 10.2 Digital IF-based Linear Receiver Architecture 
 
Here the quadrature down conversion function is contained within the DSP, and 
this has the advantage that perfect quadrature accuracy can be obtained, without the 
presence of DC offsets.    This is usually performed by ensuring that the final IF (labeled 
baseband IF in Fig. 10.2) is at a frequency sufficiently high that some channel selection 
can be performed, but sufficiently low that a sensible A/D and DSP processing bandwidth 
results.   This compromise is around 10-50MHz  [21] but continues to increase as A/D 
technology advances.   The minimum frequency is determined by the requirement that at 
least a single channel must be capable of being Nyquist sampled at the A/D converter (10 
MHz being the minimum requirement, approximately, for 3GPP W-CDMA).   Some 
allowance should be made for frequency drift of the receiver local oscillator in selecting 
this frequency, if frequency correction is to be performed within the DSP.   For wideband 
systems (e.g., CDMA), this is generally negligible.    Allowance must also be made for 
                           
 
 93 
the roll-off of the IF filter and hence the potential for adjacent-channel energy to enter the 
front-end.   This will also force the base-band IF higher in frequency. 
It should be noted, however, that while the digital IF receiver architecture shown 
in Fig. 10.2 is simpler than that shown in Fig. 10.1 due to the down conversion being 
implemented in the DSP, both architectures still require the use of band-select, IF, and 
anti-alias filters tuned to the bandwidth of interest.  
 
10.3 Multi-Carrier (Multi-band) Receiver Design 
 
The multi-carrier receiver concept can be an extension of either architecture 
shown in Fig. 10.1 or Fig. 10.2.   The architecture shown in Fig. 10.1 would be less 
preferred versus the digital IF radio architecture of Fig. 10.2 as it would require too many 
components to implement, and its cost may be prohibitive for hand held applications.    
Thus, the architecture of Fig. 10.2 can be extended to multiple quadrature down 
conversions as shown in Fig. 10.3.   In this case, the multiple quadrature down 
conversions are performed in the digital domain using separate numerically controlled 
oscillators (NCOs).   Channel selectivity is provided using digital low-pass filtering on 
the resulting I and Q base band signals;   as a consequence, the selectivity achieved can 
be very good.   This approach to a multi-carrier receiver problem, such as a cellular BTS, 
has the significant advantage of a considerable savings in RF hardware over an approach 
involving a number of separate receivers (as shown in Fig. 10.1).   In the case of 
surveillance receivers, it allows a larger number of channels to be monitored 
simultaneously at a relatively modest cost and with a small device.   A multiple-receiver 
design would quickly become unwieldy in this case. 
                           
 
 94 
 
Source:   [21] 
Figure 10.3 Multi-carrier Receiver Architecture Based on Digital IF 
                           
 
 95 
10.4 Flexible Transmitter Architectures 
It is now possible, with modern DACs, to obtain an output at a useable IF 
frequency (i.e., many tens of megahertz).  This brings with it the option of performing the 
required quadrature up conversion processing in the digital domain as done for the 
receivers.   This has the advantage of achieving near perfect image rejection and local 
oscillator suppression.   The transmitter architecture required to accomplish this is shown 
in Fig. 10.4. 
 
 
Source:   [21] 
Figure 10.4 Transmitter Architecture Employing a Digital IF Output 
 
 
The implementation shown in Fig. 10.4 employs interpolation filtering and 
provides similar benefits as provided by an analog up conversion technique.   The outputs 
of the interpolation processes feed a digital quadrature up converter, which utilizes a 
numerically controlled oscillator (NCO) as the local oscillator signal. 
The use of an NCO permits frequency-hopping to take place digitally, if desired, 
and this can typically provide a much shorter hop time than with an analog PLL.   If this 
approach is chosen, however, it is important to note that the analog IF filter must now be 
                           
 
 96 
widened to cover the whole bandwidth over which hopping may occur (typically the 
whole frequency allocation).   It is therefore no longer able to remove close-in DAC 
spurs, and these must be sufficiently low to meet the required system specification. 
The output of the digital up-converter feeds an IF output DAC, which, if over-
sampled, may be operating at a rate of many hundreds of megahertz.   The output of this 
DAC contains the wanted band, plus a range of harmonic and alias products.   These are 
typically removed using a band-pass filter (e.g., a SAW filter); however, some simple 
analog low-pass filtering to eliminate some of the higher harmonics may also be 
beneficial (the higher-frequency stop band attenuation of some filters can be poor).   
Once the wanted IF band has been selected by this first IF filter, one or more stages of 
conventional frequency up-conversion may then be employed to translate the signal to its 
final RF allocation. 
This architecture has the advantage that only a single DAC is needed; although 
the requirements for one or more IF filters (with typically a tight specification) will 
frequently more than offset the cost saving of a second DAC.   In addition, the 
performance of most DACs with an IF output will be poorer than with a base-band 
output, thus making a given specification more difficult to achieve.   Comparing the 
performance of a typical DAC, when utilizing a 5-MHz base-band signal and a 20-MHz 
IF signal, indicates that a 5 to 10dB reduction in spurious-free dynamic range can result.   
This assumes that the DAC is designed to operate at a suitable sample rate, such that the 
20-MHz IF can be accommodated with a reasonable, but not an excessive, margin. 
Fig. 10.5 shows a logical extension of the digital IF transmitter architecture 
detailed above.   In this case, multiple carriers (three are shown) are separately up-
                           
 
 97 
converted, each by its own NCO.   There are then summed digitally, prior to digital-to-
analog conversion.   Since this is now a multi-carrier signal, the peak-to-mean ratio of the 
signal is likely to have increased, unless steps are taken to counteract this effect.   The 
DAC must posses a sufficient dynamic range to cope with this and hence this architecture 
is generally the most demanding in terms of DAC performance.    
 
 
Three-carrier version shown 
Source:   [21] 
Figure 10.5 Multi-carrier transmitter architecture employing a digital IF output 
                           
 
 98 
 The circuitry following the DAC is similar to that described earlier for a digital 
IF based transmitter.  The main differences in this case are: 
? The filtering must now be sufficiently wide to cope with a number of not 
necessarily adjacent carriers, but while still having a similar roll-off 
(typically) to that of the single carrier system.   This therefore places greater 
demands upon the filter design. 
? The dynamic range of the active circuitry (mixers, amplifier, and so forth) 
must be greater in order to cope with the greater peak-to-mean ration of the 
multi-carrier signal. 
 
The NCOs may be used for frequency hopping, as outlined earlier for the digital 
IF architecture.   In this case, however, there is no disadvantage in terms of the analog IF 
filter, as this must already cover the whole band of interest and hence the DAC spurs 
must be low enough to meet the system requirement (or this architecture cannot be used). 
 
10.5 The Problem of the Diplexer 
The diplexer filter in a mobile radio transceiver has, for many years, been the sole 
method of achieving the necessary removal of the transmitter output signal from the 
receiver input, in a full-duplex radio.    This component is normally essential in order to 
realize the benefits of a standard telephone conversation in a FDD system (i.e., to be able 
to speak and listen simultaneously).   In addition, it has been a feature of many TDD and 
TDMA systems, due to the requirement for the transmit and receive frames to overlap, 
when a long turnaround time for the transmit/receive signals is present (e.g., when the 
user is close to the edge of a cell). 
The use of a diplexer (or duplexer) filter has a number of significant 
disadvantages that must either be tolerated or circumscribed in order to enjoy its benefits.   
These can be summarized as follows: 
                           
 
 99 
? Size: Their physical construction is such that they are often bulky, and even in 
hand portables they can consume a relatively significant amount of space. 
? Construction: Their function, and hence their required form of construction means 
that it is unlikely that they will be successfully integrated along with the silicon 
components within a transceiver (at least in the foreseeable short or medium 
term).  They are therefore a barrier to achieving a single-chip, full-duplex radio. 
? Spectrum inefficiency:  The use of a diplexer requires a frequency split between 
transmit and receive bands.   A proposed technique to eliminate the diplexer 
should mean that his split could be eliminated, thus allowing both transmission 
paths to operate on the same frequency (know as on-frequency duplex).  This in 
turn could lead to a doubling of the number of channels available in a given 
bandwidth.   Note that the required performance from the technique increases 
markedly when attempting to achieve on-frequency duplex, as the performance 
requirement changes from one of eliminating overload in the receiver path to one 
of suppressing the transmit signal to a level below the minimum required receive 
sensitivity, by more than the co-channel protection ratio of the modulation 
formant in question.   This is an extremely tough requirement in most systems. 
 
10.5.1 The Isolation Issue 
One example of a problem in the transmit-receive frequency split occurs in the 
220-MHz SMR band in the United States.   The issue is that of a small split between 
transmit and receive bands within the 220-MHz allocation; a given pair of transmit and 
receive channels is only separated by 1 MHz and this is a very small percentage of the 
frequency of operation.   By contrast, the 1,800-MHz DCS band has a split of 90 MHz 
which is around 5% of the operating frequency, compared to only 0.45% for the SMR 
band.   It is this small percentage, which dictates the specification required of the 
diplexer filters in order to allow full duplex operation.  Creating a filter with a suitable 
high rejection over such a narrow frequency band, at say VHF, would result in a very 
large and expensive item if indeed it is realizable at all.    Such filters would be nonsense 
in hand portable and prohibitively expensive and unacceptably large in a mobile handset. 
Thus, a radically new approach to this problem is therefore required, via advances in 
                           
 
 100 
miniaturization technologies (like RF MEMS) or the aforementioned on-frequency 
duplex concept.    
If it is assumed that a 2.5-W output power (+34 dBm) is required from the mobile 
and that the receiver is well designed and hence has a dynamic range of 80 dB, then the 
maximum level of transmit signal permitted in the receiver front end is -40 dBm (for an 
overall receiver sensitivity of -120 dBm).   The rejection required therefore is +34-(-40) 
dB = 74 dB.  This level of rejection must mainly be provided by some form of 
cancellation without sapping significantly additional power from the supply or adding 
unreasonable levels of complexity. 
There are a number of issues which arise when considering a multi-band flexible 
(MBF) architecture radio.   In particular, the requirement for flexibility, in a multi-mode 
radio capable of operating with a number of radio systems (even in the same area of 
spectrum), introduces new duplexer issues which must be addressed.   The problems 
occur since the different systems may use different multiple access schemes [e.g., 
frequency-division duplex (FDD), time-division duplex (TDD), time-division multiple 
access (TDMA), or code-division multiple access (CDMA)] and may have different 
transmit/receive splits (or none at all in the of TDD systems). 
The required transmit/receive isolation for continuous-time, full duplex 
transmission (i.e., not TDD and not TDMA with non-simultaneous transmit/receive 
timeslots) is based on the transmit power level and required receive sensitivity, along 
with the receive A/D dynamic range and the selectivity of the receiver (digital) filtering. 
Consider the example of a handset full-duplex transceiver, with a 1-W (+30 dBm) 
maximum output power capability and a minimum receive sensitivity of -110 dBm (for a 
                           
 
 101 
given modulation bandwidth).   If it is assumed that a 10-dB C/I (carrier to interference 
ratio) is the minimum for adequate demodulation of the chosen modulation scheme, then 
the minimum isolation which must be provided by the duplexer, for on-frequency duplex, 
is  [21]: 
 ZI,OFD = PTx  -  ( Pmin  -  DC/I )   
   = +30 ? (  -110 ? 10 ) 
   =  150 dB 
This is an extremely stringent requirement and would prove almost impossible to meet by 
any currently known and economic technique. 
If however, a duplex frequency split is now introduced, the situation becomes 
more realistic.   Consider the example above, but now with a duplex frequency-split 
introduced, such that the receiver IF digital filtering can reduce the unwanted residual 
transmitter signal appearing in the front-end received signal, to a negligible level.   This 
makes no assumption about any analog IF filtering, which may well ease the burden on, 
for example, ADC dynamic range, as the general case of fully flexible receiver 
architecture is assumed here.   There are now two isolation considerations which must 
both be met, however each is potentially much less stringent that that considered above. 
The first consideration is overload of the receiver?s front-end and receiver strong 
signal handling capability.    This breaks down into the analog part (LNA, mixers, and so 
forth) and its IMD performance and the A/D converter and its dynamic range.  In this 
case (split-frequency duplex), the required isolation may be derived as follows.   In this 
case, the IMD power generated by the receiver non-linearity (or clipping) must not 
exceed the specified minimum sensitivity plus the required C/I for the modulation format 
                           
 
 102 
in question.  In practice a margin of at least 3 dB would be desirable; however, the 
simplified analysis below assumes no margin, hence  [21]: 
 PIMD  =  Pmin  -  DC/I  
The IMD power resulting from the front-end non-linearity, based upon the simple 
assumption of a two-tone and purely third-order non-linearity, is given by: 
 PIMD  =  2 ( PTone  -  PIP 3 ) 
The tone power in this case is provided by the unwanted leakage of the transmit signal 
into the receive signal path, hence: 
 PTone  =  PTx  -  Z I, SFD1     
Combining above equations gives: 
 Z I, SFD1     =  PTx    -    [(Pmin  -  DC/I / 2]    -  PIP 3   
Here PTx  is the transmitter output power, Pmin  is the minimum specified signal power, 
DC/I  is the minimum carrier-to-interference ratio for the modulation format in question, 
and  PIP 3  is the third-order intercept point of the receiver front-end analog components.   
If this example is used and a receiver input intercept point of +30 dBm is assumed (a 
reasonable upper limit for a linearized receiver front-end in a handset), the required 
isolation reduces to 60 dB.   This is still a very high value, but may not be completely 
beyond the bounds of possibility for a future isolation technology. 
Note that above analysis also assumes an adequate A/D converter dynamic range 
is available, where this dynamic range is given by the difference between the unwanted 
(residual) transmitter output signal impinging upon the receiver and the maximum 
permitted interference level.   Strictly speaking, this is the required spurious-free dynamic 
range (SFDR) rather than the signal-to-noise ratio (although this may also be important, 
                           
 
 103 
depending upon the degree of averaging and/or filtering which can be employed in the 
digital domain, to extract the wanted signal).  It is given by: 
 DA/D =   ( PTx   -   Z I, SFD1 )    -   (  Pmin  -  DC/I ) 
   =   90 dB 
This is again high, but not out of the question, particularly in a narrowband system.   As 
ADC technology improves in the future, it will become increasingly realistic, even in 
broadband (and hence high sample-rate) systems. 
The second requirement is that the transmitter noise floor must not mask the 
received signal.   This results in the following isolation requirement: 
 Z I, SFD2     =  NTx    -    (Pmin  -  DC/I ) 
Here NTx  is the transmitter output noise power (in the receiver bandwidth).   A typical 
figure for this noise power is around ? 75 dBm, based upon the minimum received power 
levels used above (and hence channel bandwidths).   With this figure, the required 
isolation is 45 dB, making the first consideration (on receiver linearity) dominant in this 
case. 
 
10.5.2 Potential Solutions to the Isolation Issue 
There are a number of partial or complete solutions to the isolation problem: 
? Tx / Rx Switch.  It is possible to implement a purely switch-based duplexer 
facility, and this has many advantages.   First, it can be made very 
broadband (multi-octave, if necessary) since filtering is not necessary.  
Second, it places no restriction on the system duplex frequency split, since 
no frequency-selective components need to be involved.   Finally, it will 
allow on-frequency duplex (i.e., TDD) for the same reason.   It  may be 
implemented using simple PIN-diode switch technology and is therefore low 
cost, although transmit-receive isolation is an issue and it may be necessary 
to disable the transmitter while in receive mode to ensure that the transmitter 
noise floor does not de-sense the receiver. 
                           
 
 104 
? Switched Diplexer.   Recent advances in integrated diplexer techniques have 
led to the possibility of implementing a switched diplexer, in which the 
transmit and receive paths can be switched between two (or more) paths.   
This type of system is shown in Fig. 10.6.     A two-way or multi-way 
change-over switch is employed, to select the required filter pair for a five 
transmit and received band allocation.    The example shown in Fig. 10.6 
utilizes GaAs PHEMP switches, as these were reported to have a number of 
advantages over PIN diodes at the required frequencies of operation.     
These advantages included: low current/low voltage and single supply 
operation together with having no requirement to resonate out parasitics.    
The switches provided more than 20 dB of isolation (excluding filter 
isolation), together with an insertion loss of less than 1 dB for the overall 
diplexer circuit.   This system, however, has a number of disadvantages, 
including band limiting (i.e., it is not commensurate with an ideal general-
coverage SDR), relatively high loss (typically) due to losses in the switches, 
and limited power handling (again due to switch-related issues, such as 
saturation and IMD). 
? Circulator.   A second solution is to use a circulator, as shown in Fig. 10.7.  
The main drawbacks of this approach lie in the frequency range limitation of 
most high-isolation parts and the achievable isolation from low-cost, small-
sized components, suitable for handset applications.   Typical isolation 
values for these parts, even band-specific items, are in the range of 10 to 25 
dB.   This is adequate for their current, primary application in protection of 
the transmitter from the wide range of antenna VSWR conditions.     
However, even the higher value is not adequate for the duplex function 
under consideration here.  
? Duplexer elimination schemes.  This includes some new methods of 
achieving transmit/receive isolation.   It is possible for example, to use 
cancellation-based techniques in order to remove the transmit signal from 
the receive signal path, although these techniques themselves have a number 
of disadvantages.   They are complex and have difficulties, coping with 
external reflections,   They also generally require complex antenna 
arrangements, which are not currently compatible with small handset 
designs.  Research is, however, being performed in this area and a solution 
may be developed in the future. 
 
                           
 
 105 
 
 
(a) Block diagram of the diplexer module; (b) Diplexer circuit schematic 
 
Source:   [21] 
Figure 10.6 Switched Antenna Diplexer 
                           
 
 106 
 
 
Source:   [21] 
Figure 10.7 Use of a Circulator to Provide Transmit/Receive Isolation 
 
10.6 Chapter 10 Summary 
Current radio receiver designs are, in general, inherently narrowband and can only 
achieve general (or broadband) coverage by the switching or alteration of narrowband 
elements.   Certain designs, such as those used in many scanning receivers, do not 
attempt to overcome some of the fundamental receiver problems, such as blocking and 
image rejection, but rely on the user being able to eliminate interference by positioning of 
the set,  or some other mechanism such as the use of a directional antenna.    Where this 
is not possible, the user most tolerate the problem and the restriction in frequency usage 
which results, as the price of achieving wideband coverage. 
While SDR technologies using DSP have made significant inroads into the utopia 
wideband flexible radio, there are still two basic problems to be solved: 
                           
 
 107 
? The diplexer filter required in a full-duplex transceiver must be specifically 
and carefully designed for its intended frequency band of operation.   This 
filter is usually either a helical component or formed from a dielectric (such 
as ceramic) and hence is almost impossible to tune in any sensible fashion 
over a reasonable range of frequencies.    A multiple-band transceiver would 
therefore require a number of diplexer filters and this would be very quickly 
become prohibitive, both in terms of cost and size. 
? The front-end pre-select filter (also known as band-select or cover filter), 
utilized to reflect the image signal and other particularly strong out-of-band 
signals, must also be either tunable or eliminated in order to allow multi-
band coverage.  Electronic tuning of this filter is a more realistic proposition 
than that of the diplexer filter mentioned above;  however, the change in 
technologies (from, perhaps, lumped-element to dielectric-based) across, 
say, 100 MHz to 2 GHz, would make this very difficult.   The alternative to 
the use of such a filter would require the front-end amplifier [or low-noise 
amplifier (LNA)] to be able to handle the full dynamic range of signals 
within the broad coverage range.   This may include TV transmissions of 
many kilowatts and microcellular transmissions of a few milliwatts, and 
hence a very high dynamic range amplifier is required.   Such an amplifier 
could be created by backing off a high-power linear amplifier (of say, 10 
W), but this is unrealistic in a hand portable radio.   It is therefore necessary 
to utilize a more conventional low-noise amplifier and eliminate its 
distortion when dealing with high input signal strengths. 
 
As discussed in this and preceding chapters, the addition of wide channel 
bandwidth and, in particular, multiple operating bands significantly increases the 
difficulty of producing a flexible receiver design.    The widening of the channel 
bandwidth has the following consequences: 
? The number of narrowband carriers which can enter the IF and base band 
chains is significantly increased, thus increasing the potential dynamic range 
required in these parts of the system.  In going from an IF of, say, 200 kHz 
(for GSM) to 4 or 5 MHz (for UMTS in Europe), the number of 25 kHz 
channels (e.g., for TETRA) that could enter the IF increases from 8 to 200. 
 
In going from a single-band to multiple bands, the receiver faces a number of 
further problems: 
? RF pre-selection filtering becomes difficult or impossible, since the filter 
must now be tuned to each band of interest.   Alternatively, a bank of 
switched filters may be employed, but this can be quickly become unwieldy 
                           
 
 108 
for a truly flexible system.   This latter technique has been used in a number 
of scanning systems in the past.    RF MEMS is a technology that has 
recently gained more interest as a way to realize smaller economical tunable 
filters, and it is discussed in more detail in the next chapter. 
? The channel synthesizer must tune over a far wider range than for a single-
band system. 
? The diplexer in a full-duplex transceiver must have a variable frequency of 
operation and a variable transmit/receive frequency split.   Since the 
diplexer is currently realized in ceramic, SAW, or helical resonators in most 
portable systems, this is clearly impossible with current techniques and thus 
represents another R&D opportunity for RF MEMS.    Again, the main 
obvious alternative is the use of multiple units, with switching to determine 
which is in use at a given point in time.   As before, this can quickly 
becomes unwieldy. 
 
One last point to make, which emphasizes the ongoing need for tunable front-end 
pre-select and diplexer filters, is related to the consequences of eliminating them, namely:   
? All image rejection from the filter is lost, thus leaving the receiver prone 
to signals appearing at its image frequency,  
? All radio signals within the bandwidth capability of the antenna will 
impinge upon the front-end low-noise amplifier in the receiver.    This 
amplifier will therefore require a very high dynamic range to prevent 
overload from strong, unwanted signals (e.g., broadcast TV transmissions 
in a hand portable communications receiver). 
? Without a diplexer, the full power of the transmitter output signals may 
impinge upon the receiver input (depending upon what is used to replace 
the diplexer).  The receiver must therefore be able to cope with these 
signals, or else utilize an alternative method of eliminating them. 
 
Thus, it can be safely concluded that tunable pre-select and diplexer filters (in 
addition to IF anti-alias filters) will continue to be required for multi-band flexible, SDR-
based radio architectures for a long time to come as newer 4G and beyond technologies 
get developed and deployed.     Therefore, both research and development opportunities 
exist now and in the future for achieving lower cost, smaller size, tunable filter banks for 
MBF transceivers. 
                           
 
 109 
Chapter 11 The Current State of RF MEMS Technology 
 
11.1 Introduction 
This chapter provides an update on the current state of RF MEMS technology and 
attempts to derive some conclusions as to the viability of such technology for the 
commercial realization of RF tunable filters for multi-band flexible SDR-based 
architectures for applications in both base stations and handsets. 
RF MEMS stands for Radio Frequency Micro-Electro-Mechanical Systems, and 
refers to components of which freestanding or moving sub-millimeter-sized parts provide 
RF functionality.  RF functionality can be implemented using a variety of passive and 
active RF technologies.  Besides RF MEMS (passive) technology, ferrite (passive), 
ferroelectric (passive), GaAs (active), GaN (active), InP (active), RF CMOS (active), and 
SiGe (active) technology are available to the RF designer.  Each of the RF technologies 
offers a distinct trade-off between cost, frequency, gain, scale of integration, linearity, 
noise figure, packaging, power handling, reliability, and supply voltage.  
RF MEMS components, such as inductors, resonators, switches, switched 
capacitors, and varactors, offer low loss, high linearity power handling, and high-Q 
factor, but require a high supply voltage and wafer-level, liquid crystal polymer (LCP) or 
low temperature co-fired ceramic (LTCC) packaging.  RF MEMS components can be 
fabricated in class 100 clean rooms using 6 to 8 lithography steps, whereas state-of-the-
art MMIC and RFIC fabrication processes require 13 to 25 lithography steps
                           
 
 110 
11.2 ?More than Moore? 
?More than Moore?  [24], or the combination of sensors/actuators, radio, power 
and interfaces with signal processing and storage, is recognized as one of the most 
significant trends in the microelectronics industry.  Strong drivers for the More-than-
Moore domain are automotive applications, ambient intelligence (including bio-medical), 
and wireless applications (including environment monitoring and security).  In this 
framework, RF-MEMS appear as a key enabling technology for dramatically improving 
miniaturization, low-power, low-cost and system performance of wireless systems. 
 
11.3 The State of RF MEMS Technology 
The status and trends in RF MEMS components, systems and services can be 
summarized as follows. 
RF MEMS switches and variable capacitors have reached maturity and are ready 
for industrial applications.  Applications range from reconfigurable front-ends, cognitive 
radios, antenna switches, matching networks to delay lines and others. Research will 
however continue to explore new switching technologies, applicable to power 
management and low power RF switching.  Among the interesting candidates, the NEM-
FET appears to be very attractive with a high on/off ratio. 
Resonators and oscillators are the next candidates for industrial application where 
they could replace bulky ceramic filters, quartz oscillators and others.  Full maturity is 
not reached yet, but interesting breakthroughs are appearing in terms of precision, 
temperature stability and frequency range. The next breakthrough is expected from nano 
electro-mechanical FET (NEM-FET) resonators and oscillators.  Tune-ability, reliability 
and long-term stability remain challenging. 
                           
 
 111 
Smart antennas are attracting a large interest, due to their wide field of 
applications in automotive and avionics.  Besides the use of ?classical? MEMS for 
generating the variable phase shift, very innovative solutions appear that are based on 
liquid crystal or meta-material phase-shifters. These new approaches are promising and 
should be carefully monitored. 
Some challenges, however, remain in multi-physics modeling, MEMS-IC co-
design and system modeling.  Several initiatives exist for providing access to technology 
platforms, support for design, testing and packaging, and others. 
The focus of this chapter is on the viability of RF MEMS for the realization of RF 
tunable filters for multi-band flexible SDR-based radio architectures.    Thus, the primary 
areas of interest related to RF MEMS are switches, varactors, and inductors and the 
subsystems that can be commercially realized with these RF MEMS components. 
 
11.4 Switches 
There are two distinct parts to an RF MEMS switch: the actuation (mechanical) 
section and the electrical section (see Fig. 11.1).   The forces required for the mechanical 
movement can be obtained using electrostatic, magnetostatic, piezoelectric, or thermal 
designs.   The switches can also move vertically or laterally, depending on their layout.    
Electrostatic-type thermal switches and magnetostatic switches have been demonstrated 
at 0.1 ? 100 GHz with high reliability (100 million to 60 billion cycles) and wafer-scale 
manufacturing techniques.    As for the electrical part, a MEMS switch can be placed in 
either series or shunt configurations and can be a metal-to-metal contact or a capacitive-
contact switch.    This means that one can build at least 32 (2 x 2 x 2 x 4) different type of 
                           
 
 112 
MEMS switches using different actuation mechanisms, contact, and circuit 
implementations.  
 
 
Source:   [25] 
Figure 11.1 Different Configurations of MEMS Devices 
 
                           
 
 113 
Electrostatic actuation is the most prevalent technique in use today due to its 
virtually zero power consumption, small electrode size, thin layers used, relatively short 
switching time, 50-200 uN of achievable contact forces, and the possibility of biasing the 
switching using high-resistance bias lines.    However, in many cases, it requires an 
actuation voltage of 30-80 V, and this necessitates the use of CMOS up converters to 
raise the input 3-5V control voltage to the actuation voltage.   In many designs, a thermal 
actuation is coupled with an electrostatic (voltage) hold, or a magnetostatic actuation 
(current in a coil) is coupled with a permanent magnetic field.   Both result in virtually 
zero power consumption once the switch is actuated.   Still, they require a substantial 
amount of current for the switching cycle and, therefore, must be biased using low-
resistance (gold or Al) lines.    The low-resistance lines couple to the microwave 
transmission lines; therefore, careful design must be done for complicated switching 
networks requiring a large number of switches and bias lines. 
Fig. 11.2 shows a comparison between electrostatic MEMS switches and GaAs 
PIN diode and transistor switches.   It is hard to make an accurate comparison over a 
wide range of RF power levels since the size of diode and transistor switches can be 
easily increased for high power applications.    This, in turn, has a substantial effect on 
the switch isolation, insertion loss, switching speed, and power consumption.    Still, it is 
evident that MEMS switches, with their extremely low up-state capacitance (series 
switches) and their very high capacitance ratio (capacitance contact switches), offer a far 
superior performance compared to solid-state switches for low to medium power 
applications. 
                           
 
 114 
Fig. 11.2 also summarizes some of the key advantages of RF MEMS switches 
over the conventional GaAs PIN and FETs, namely: no DC power consumption, much 
higher cutoff frequency, very high isolation, lower insertion loss, and higher third-order 
intercept point.   These superior parameters can replace the GaAs switches in cellular 
telephones resulting in much lower DC-power consumption and longer battery life.   RF 
MEMS switches can be used in phase shifters and low-loss tunable circuits (like 
matching networks and filters), and in high-performance instrumentation systems.  
 
 
 
Source:   [25] 
Figure 11.2 Performance Comparison of FETs, PIN Diode, & RF MEMS Switches 
 
 
                           
 
 115 
11.5 Applications 
Fig. 11.3 summarizes the application areas of RF MEMS devices and the lifetime 
and number of cycles required.     The same figure also shows some of the subsystems 
and circuits that can benefit from this technology.    
Fig. 11.4 shows a multi-band RF front end and the specific circuitry where RF 
MEMS technology has very promising applications in implementing tunable matching 
networks, tunable and switchable RF filters (Filter Bank), diplexers and multi-band 
antennas. 
RF MEMS switches are extremely linear devices since they do not contain a 
semiconductor junction and therefore do not have a nonlinear current versus voltage 
relationship.   MEMS switches and varactors can be designed to be mechanically stiff 
(large spring-constant design) and can tolerate large RF voltage swings without 
generating any significant levels of inter-modulation products.   They are thus near-ideal 
elements for tunable filters and matching networks. 
The third-order intercept point of MEMS switches was measured independently 
by Rockwell Scientific and HRL (metal contact) and Raytheon (capacitive contact) and 
was found to be at least +66 dBm (capacitive contact) to +80 dBm (metal contact).   
These numbers are 25-35 dB higher than what can be obtained with PIN diode and FET 
switches.   There is no doubt that the high linearity of MEMS devices will allow the 
design of complex communications and radar systems that are currently not possible with 
standard GaAs devices. 
 
 
                           
 
 116 
 
 
Source:   [25] 
Figure 11.3 Application Areas of MEMS Switches, Varactors, and Hi-Q Inductors 
 
                           
 
 117 
 
 
 
 Figure 11.4 RF MEMS Applications in Multi-band RF Front Ends 
                           
 
 118 
11.6 Reliability 
As of 2008, many metal-to-metal (DC) contact and capacitive switches have 
achieved > 60 billion cycles under low power conditions (0.5 to 5 mW).    All tests are 
done in nitrogen atmospheres and under clean-room conditions.    The failure 
mechanisms are complex and are different for capacitive or DC-contact switches.   Also, 
the failure mechanisms depend on the RF power used and can be due to thermal issues, 
dielectric breakdown, self-actuation, or critical current density problems.    It is believed 
that once a switch crosses over 10 billion cycles, then it can be easily taken to 100 billion 
cycles since the failure mechanisms (dielectric charging, pitting, hardening, dielectric 
formation, etc) typically present themselves in the first billion cycles. 
Capacitive switches with their large contact area can handle more RF power than 
metal-to-metal switches and are therefore the preferred switch for applications requiring 
30-300mW.   However, most capacitive switches operate at 8 GHz and above due to their 
relatively small down-state capacitance (2-5pF), and DC-contact switches are the only 
switch of choice at 0.1-8GHz.   The lifetime of metal-to-metal contact switches is highly 
dependent on the RF power used, and most electrostatic switches result in only 10-1000 
million cycles at an RF power of 10-100 mW.   This is due to the relatively small contact 
forces (50-200 uN) and small contact areas that can be achieved using electrostatic 
actuation.   On the other hand, thermal MEMS switches with their large contact forces (2-
3 mN) have been proven to reliably switch to > 50 million cycles under an RF current of 
350 mA (which is equivalent to an RF power of 6 W in a 50-? system).   However, 
thermal actuation consumes a large amount of DC power (50-200 mW) and is not 
suitable for most applications. 
                           
 
 119 
It is believed that both capacitive and DC-contact switches can be taken to > 100 
billion cycles under low power conditions, and to billions of cycles under medium to 
high-power conditions.    This will be done with advances in contact metallurgy, thermal 
analysis, high-quality dielectric materials, stress control, and mechanical design. 
The operation of RF MEMS devices can be severely affected by the presence of 
water vapor, oxygen, contaminants, and other hydrocarbons.    The reason is that the 
contact or pull-up forces are quite small (950-500 uN in most cases), and there is simply 
not enough force to either (a) puncture through any contaminants or dielectrics that 
deposit or form on the metal contract surfaces or (b) overcome the adhesive forces of 
water molecules on the capacitive  contact.   It is for these reasons that metal-to-metal 
contact MEMS switches typically fail as open circuits, and capacitive switches fail in the 
down-state position (stiction). 
 
11.7 Packaging 
RF MEMS devices are tested in nitrogen atmospheres and under clean-room 
conditions.    While this is acceptable in the laboratory, careful attention must be placed 
on the packaging of RF MEMS device for commercial applications.   An important 
question is the long-term viability of non-hermetic packages for RF MEMS.   This is 
critical because a non-hermetic package is inexpensive (about $0.2-0.5 per unit), while 
hermetic packages require high temperatures and very smooth surfaces and are, in 
general, quite expensive (about 42-50 per unit).  The answer to this question will 
eventually determine the cost and the application areas of RF MEMS devices.   
The only other way to reduce the cost of RF MEMS switches is to use wafer-scale 
packaging solutions.   Fig. 11.5 shows some of the different wafer-scale packaging and 
                           
 
 120 
inter-connect schemes for MEMS.   In this case, the MEMS switches are first released, 
packaged on-wafer, and then scribed into individual units.   Several different technologies 
are available for packaging MEMS devices as follows: 
? Epoxy Seals 
? Metal-to-metal solder bonding 
? Glass-to-glass anodic bonding 
? Glass frit bonding 
? Gold-to-gold thermo-compression bonding 
 
It is important to note that many conventional MEMS devices are packaged at low 
pressures (1-50 mtorr), while RF MEMS device should be packaged at standard pressures 
and in nitrogen or non-reactive gases (argon, etc).   This results in a low mechanical-Q 
structure due to gas damping, and it increases the reliability of the switch.   Also, MEMS 
switches and varactors are built using at least one metal layer (Au or Al) and therefore 
cannot be subsequently placed in high-temperature LPCVD or thermal oxidation furnaces 
for packaging purposes. 
High-temperature glass-to-glass anodic and frit bonding have been the most 
commonly used techniques in hermetically packaged MEMS sensors  [26].  Anodic 
bonding also requires a large voltage between the glass wafers (600-1200 V).  The reason 
why these methods are successful with MEMS accelerometers, gyros, pressure and 
temperature sensors, and so on, is that these devices are generally fabricated using thick 
polysilicon layers and can withstand higher post process temperatures, do not have 
surfaces that touch each other, are not sensitive to release of organic gases during the 
packaging process, and are calibrated every time one turns on the device.   There options 
do not exist for RF MEMS switches. 
                           
 
 121 
 
(a) Different wafer-scale packaging and interconnect schemes for MEMS switches  
(b) Top view of a possible RF transition.  Feed-through or via technologies are needed to 
pass the RF signal through the sealed package 
 
Source:    [25] 
Figure 11.5 MEMS Packaging 
                           
 
 122 
There are several problems with these techniques when applied to RF MEMS.  
Most bonding techniques outgas organic materials inside (and outside) the MEMS cavity 
during the bonding process due to wetting compounds in the glass, gold, or epoxy layers.   
This has a serious detrimental effect on the reliability of the MEMS switches, both DC-
contact and capacitive switches.  Hermetic bonding processes also require a temperature 
of 250-500 oC to achieve a good seal.   This is not compatible with a release structure that 
may be thin (0.5-2um) and quite long (250-350 um).   The high temperature may bow the 
membrane by ? 1-5 um, making the switch unusable.   Also, many bonding techniques 
such as glass-to-glass anodic bonding or gold-to-gold thermo compression bonding are 
sensitive to the planarity and cleanliness of the wafer (glass case) or to the surface 
roughness and exact height of the plated gold rings on the wafer (gold case). 
While there are many packaging techniques, the packaging of MEMS switches 
provides a larger challenge than standard MEMS devices due to hermeticity, temperature, 
and out gassing constraints.     Also, MEMS devices must be packaged in nitrogen or dry-
air atmosphere for high reliability.   Thus, the question remains as to whether it is 
possible to produce a reliable switch using a low-cost near-hermetic package and more 
research is needed to determine what leakage rate is permissible and what gases have 
detrimental effect of the reliability of RF MEMS devices. 
 
11.8 Varactors 
The development of MEMS varactors has not progressed at the pace of MEMS 
switches.   This can be explained by the abundance of discrete high-Q silicon and GaAs 
varactors usable up to 30 GHz.   Planar silicon varactors, even at 0.5-5 GHz, result in a Q 
of 30-60 using the latest SiGe and high-fT silicon technologies.  In fact, the performance 
                           
 
 123 
of planar oscillators is not limited by the silicon varactor, but by the loss of the planar 
inductor.   Another reason for the limited development of MEMS varactors is their 
capacitance ratio of 1.2-2.5, while standard solid-state varactors have a capacitance ratio 
of 4-6.   Also, it is hard to build MEMS varactors with capacitance values of 5-50 pF, 
which are needed for 30 to 600 MHz applications.   MEMS varactors suffer from 
Brownian, acceleration, acoustic, and bias noise effects, especially for low-spring-
constant designs that are necessary for 3 to 5 V applications. 
There is still, however, a pressing need for MEMS varactors.   First, they have the 
potential of very high-Q (100-400) operation, especially at mm-wave frequencies.    Also, 
they can be designed to withstand large RF voltage swings and therefore result in very 
high IIP3 tunable networks.   MEMS varactors do not contain PN junctions and therefore 
do not pass current under high-power operation.   They can be inexpensively produced on 
glass, ceramic or high-resistivity silicon substrates with applications in low-cost 3 to 60 
GHz tunable networks and filters.   Still, MEMS varactors may take a long time before 
they replace (in high volume handset applications, for example) silicon and GaAs 
varactors, especially below 5 GHz, since solid-state devices result in excellent 
performance and do not have any special packaging requirements. 
Currently, there are three different technologies for building MEMS variable 
capacitors.   The first one is based on the parallel-plate approach (vertical design), and a 
variable capacitance is achieved by changing the gap between the capacitor plates.   The 
second approach is based on an inter-digitated (horizontal) design; again, the gap is 
changed to achieve a variable capacitance.   The third approach consists of building a 
fixed capacitance bank and using MEMS switches to select the required total capacitance.   
                           
 
 124 
These different techniques have yielded MEMS variable capacitors for applications in the 
range of 0.1 to 100 GHz 
The reliability of analog MEMS varactors (non-switched) is much better than 
standard capacitive and series switches.  In the analog mode, the contact areas do not 
touch and one can easily obtain billions of cycles.  The mechanical reliability of thin gold 
or Al membranes and cantilevers has been shown to exceed 1010 cycles for small 
deflections (1-2 um).  The main problem is charging of the dielectric layer (if used), 
which can create a progressive increase in the applied voltage.   This can be solved by 
eliminating the dielectric layer between the capacitor plates and by limiting the applied 
voltage so as to ensure that a pull-down condition is never achieved.   Analog MEMS 
varactors do not need to be hermetically sealed since they do not suffer much from 
charging or surface-contact problems resulting from humidity.   However, if long-term 
operation and billions of cycles are needed, then it is best to hermetically seal the 
varactor.   Still MEMS varactors must be handled and packaged in clean-room 
environments since contaminant particles may cause failure if they come in contact with 
the inter-digitated fingers, or if they settle on or around the capacitor plates.   The 
reliability of switched capacitors (or any design where a metal layer contacts a dielectric 
layer) is identical to that of MEMS capacitive switches.  The switched capacitors must be 
hermetically sealed in clean-room environments for long-term reliable operation. 
Analog parallel-plate and cantilever MEMS varactor designs result in very high-Q 
operation at mm-wave frequencies, but with a limited capacitance ratio of 1.3-1.9:1.  The 
inter-digitated and switched capacitors have shown impressive performance and 
versatility with a Q of 50-70 at microwave frequencies (0.3-4 GHz).   Both switched 
                           
 
 125 
capacitor and interdigital designs can be scaled to higher frequencies using a compact 
low-inductance approach.  Low-loss MEMS varactors are an essential component in 
tunable networks and filters, and very high-Q designs with capacitance ratios of 3-4 are 
being developed already.  Fig. 11.6 shows a summary of various analog varactor and 
switched capacitors development around the world. 
 
11.9 Inductors 
As indicated earlier, RF MEMS high-Q inductors are also key components for 
tunable LC filters in multi-band communications systems.   RF MEMS can be two 
dimensional (planar) or three-dimensional.   The work differs from the RF MEMS switch 
and varactors effort because micromachining techniques are used to create high-
performance static fixed-value inductors.   To date, there has not been a practical 
implementation of a tunable high-Q inductor (as in the case of RF MEMS varactors), and 
such inductors are currently synthesized using RF MEMS switches and a set of fixed-
value inductors. 
High-Q inductors are essential for many different passive and active circuits and 
can substantially reduce the phase noise or power consumption of oscillators and 
amplifiers.   Also, they result in low-loss matching networks and filters as the loss is 
inversely proportional to the Q.    Micro-machined inductors are based on three different 
technologies depending on the frequency range: 
? Three-dimensional solenoid-type inductors result in large-value designs 
suitable for 0.2- 6 GHz 
? Substrate etching underneath the inductor reduces the parasitic capacitance to 
the substrate (suitable for 1-100 GHz). 
? Self-assembly of the inductor away from the substrate to reduce the parasitic 
capacitance to the substrate (suitable for 1-100 GHz). 
                           
 
 126 
 
Source:    [25] 
Figure 11.6 Varactor and Switched Capacitor Development Around the World 
                           
 
 127 
Recently, standard inductor designs on CMOS and SiGe substrates have resulted 
in a Q of 12-18 at 2 GHz and 16-22 at 6 GHz.   The reasons are many: modern 
electromagnetic simulation software allows the user to optimize the inductor geometry 
for the highest inductance and lowest associated series resistance.   Also, the substrates 
used today have a relatively high resistivity (10-2000 ?-cm), thereby reducing eddy 
current losses underneath the inductor.   The inductors are integrated on the top metal 
layer using a 3-?m-thick gold layer and are separated from the silicon substrate using 3- 
6-?m thick oxide layers.   The oxide layers reduce the parasitic capacitance to the 
substrate, and they allow the integration of large value inductors without having problems 
with the inductor resonant frequency.   Micro-machined inductors will therefore be only 
used for high-Q applications (Q >30) in low-noise oscillators, high-gain amplifiers, on-
chip matching networks, and integrated LC filters.   Still, thick-metal electroplating or 
substrate etching is an additional step in the fabrication process and results in an 
increased cost.    
The substrate etching technique is useful for high-frequency applications or for 
large-value inductors, and the thick copper electroplating technique is ideal for low-
frequency applications (0.2-5 GHz).  The elevated copper inductor technique results in 
small size inductors and outstanding Q from 1 GHz to 10 GHz  [27].   The main 
limitations of high-Q inductors are their increased fabrication cost and special handling in 
the case of substrate etching.   However, it is expected that RF MEMS inductors will 
eventually be integrated with CMOS and SiGe transistors for low phase-noise oscillators, 
high-efficiency power amplifiers, and tunable LC filters for MBF SDR-based transceiver 
architectures. 
                           
 
 128 
11.10 RF MEMS and CMOS Integration 
Because RF MEMS are typically fabricated using low-temperature processes, 
they are therefore compatible with post-CMOS, SiGe, or GaAs integration.   In fact, since 
most of the MEMS switches, varactors, and inductors are surface micro-machined, they 
can even be integrated on glass, quartz, or polished ceramic substrates.  The integration 
of RF MEMS with GaAs was demonstrated by Rockwell Scientific as far back as 2001 
 [28]. 
In conclusion, RF MEMS can be integrated with CMOS or SiGe electronics for 
low-power wireless front-end systems.  Also, CMOS is ideal for the development of 
voltage up-converter circuits next to a MEMS switch (thus allowing for a low supply 
voltage operation), or to integrate an intelligent controller next to a MEMS varactor for 
improved tuning range and temperature stability.     
 
11.11 Chapter 11 Summary 
In the near term (1-3 years), the requirement for hermitically sealed packages 
(with its related higher cost) may restrict the use of RF MEMS to the lesser cost-sensitive 
RF front ends (macro, micro, and pico),  with applications for low cost femtocells and 
handsets only following after significant advances in reliable on-wafer hermetic 
packaging techniques are commercially available.    Also, a hybrid approach, that is, 
combining already-reliable RF MEMs devices with SiGe or GaAs devices may accelerate 
the deployment of RF MEMS into such lower cost applications. 
                           
 
 129 
Chapter 12 A MBF Transceiver Architecture with RF MEMS 
 
12.1 The Ultimate Goal 
In this chapter the author presents a system level architecture for a single chip RF 
front end for a 4G multi-band flexible (MBF) transceiver radio using RF MEMS 
technology.    The required feature set for the single chip is outlined and the necessary 
technology, processes, packaging, and other R&D issues to achieve it are discussed.     
The proposed architecture is unique in that it offers a single chip ?front-haul 
processor? (with built-in RF MEMS-based wireless interfaces) capable of processing 
either 1) the base band of one or more high-usage frequencies and 2) passing through the 
IF of any air interface standard into the base-band processing unit via a wireless or wire 
link. 
Fig. 12.1 shows a block diagram of an RF front end for a three-band wireless 
telephone system.   As indicated in the previous chapters, RF MEMS may be used to 
allow the development low-power systems based on tunable antennas, low-noise tunable 
oscillators, tunable filters, and tunable matching networks for wideband and multi-band 
transceiver radios.   The ultimate goal is to integrate the entire front end into a single 
chip.    
                           
 
 130 
12.2 A Single Chip Front-Haul Processor Architecture 
The primary requirement for a 4G MBF transceiver RF front end is that it is fully 
tunable to all the required RF frequencies of interest (see  Chapter 2 and  Chapter 6).  As  
 
 
Source:    [21] 
Figure 12.1 RF Front-End Design of a 3-band MEMS Wireless Telephone System 
 
                           
 
 131 
shown in the simplified circuit diagram of Fig. 12.2 , changing the frequency of operation 
means that the antenna resonant frequency must be reconfigured, together with the band 
pass filter center frequency, the local oscillator frequency, and the matching networks for 
the LNAs (low noise amplifiers) and the power amplifiers.   In certain cases the 
frequency of operation remains fixed, but the power amplifier output impedance changes 
with time and one must retune the output matching network to result in the highest 
system efficiency.  In many portable applications, the antenna input impedance is 
strongly dependent on the position of the portable device, and a low-loss reconfigurable 
matching network at the input of the antenna would result in a substantial performance 
improvement.     Also, some systems require a very wideband but non instantaneous 
frequency coverage, such as 2-18 GHz or 0.1-6 GHz, and this can be done efficiently 
with the use of reconfigurable antennas, matching networks, and filters.   Reconfigurable 
MEMS circuits can also be used to generate a large range of impedance loci which are 
necessary for transistor and diode characterization (gain, noise, conversion loss, etc). 
As indicated above, the author is hereby proposing an architecture which utilizes 
the various concepts and blocks outlined in the various chapters of this dissertation.    As 
can be seen in Fig. 12.3, the proposed architecture is unique in that it offers a single chip 
?front-haul processor? (with built-in RF MEMS-based wireless interfaces) capable of 
processing either 1) the base band of one or more high-usage frequencies and 2) passing 
through the IF of any air interface standard into the base-band processing unit via a 
wireless or wire link. 
                           
 
 132 
  
Source:    [21] 
Figure 12.2 A Reconfigurable RF Front End Block Diagram 
 
 
Figure 12.3 Multi-band Flexible Base Station Architecture 
                           
 
 133 
At the heart of this architecture is the ?front-haul processor? (FHP or SM-RFFE, 
smart programmable RF front end) with built-in RF MEMS wireless interfaces as shown 
in Fig. 12.3.  The FHP is a relatively smaller processing unit (as compared to the base-
band processors) which contains the necessary SDR software to either 1) handle the base-
band processing of one or more high volume air interface standards like GSM and W-
CDMA, or 2) pass through the intermediate frequency (IF) of any air interface standard 
(e.g., UMTS, WiMax, 4G, etc) to the base-band processing unit via a wireless (Wi-Fi or 
TBD) or wireline interface (for example, CPRI or OBSAI interface standards).    
The advantages of this architecture are: 
1. Off loading of the main base-band processing unit when handling the high 
volume air interface base band processing.   In this configuration, the FHP 
in essence becomes a base station on its own merit.   The back haul 
connection of the FHP becomes either the Wi-Fi or wire line interface to a 
nearby Base Station Controller.   In this configuration, the FHP is well 
suited for a Femtocell application supporting a single air interface 
standard. 
2. Providing flexible FHP-to-BB wireless or wire line interfaces which are 
easily configurable in the FHP?s SDR software and built-in programmable 
RF MEMS circuitry. 
3. Achieving economies of scale due to the single chip programmable 
flexibility as described in the various chapters of this dissertation. 
 
In order to achieve the commercial implementation of the proposed FHP for MBF 
SDR-based transceiver architectures the following technological challenges need to be 
addressed and it is believed that all are within commercial realization within the next 
three to five years: 
1. Integration of RF MEMS on top of standard CMOS or SiGe 
semiconductor processes.    This is critical to achieve the lowest possible 
cost. 
2. A full set of RF MEMS libraries for the implementation of switches, 
capacitors, inductors and oscillators for the implementation of filters and 
antennas using standard CMOS or SiGe processes as indicated in 1 above. 
                           
 
 134 
3. Reliable packaging solutions for RF MEMS suitable for industrial 
temperature range at medium power applications such as those required by 
pico or femto cells, and smart terminals. 
4. Base Station Controller application software modifications required to 
treat the ?dual personality? front-haul processing unit, which can behave 
as a) standalone base station and b) as IF-pass-through processing unit for 
multiple frequency bands. 
 
12.3 A  Five-Year Horizon Possibility 
In regard to the integration of RF MEMS on top of standard CMOS or SiGe 
semiconductor processes, the author believes that significant advances have already been 
made in this area and that, while the high volume fabs like TSMC or UMC are not yet 
offering the required solutions, there are other ?specialty fabs? already offering viable 
alternatives  [29].    Also, the International Technology Roadmap for Semiconductors 
(ITRS) has provided a technology roadmap update  [30] shows that the integration of RF 
MEMS components, including resonators, is expected to be a reality starting in 2011   
(see Fig. 12.4).  
Mr. Volker Herbig, Product marketing manager at X-FAB Semiconductor 
Foundries AG said  [31]: 
?The MEMS foundry market is certainly growing, but more slowly than expected. 
From a technology standpoint, integration of CMOS and MEMS is hot.?  
 
A partial list of fabs already in the MEMS processing business follows.   Also 
including are various quotes from their CEOs published in  [31] on December 22, 2008 
regarding their business trends. 
 
Micralyne, Edmonton, Alberta, Canada 
http://www.micralyne.com/?gclid=CIXKpLGZ1ZcCFQgRFQod1UzLDg  
                           
 
 135 
 ?Micralyne has experienced very significant growth recently - 40 percent to 50 
percent revenue growth - and the pipeline of new opportunities looks very strong.  
VCs are investing more and the failures associated with the optical bust are now a 
distant memory.  Furthermore, the VCs are not investing in MEMS fabrication 
capacity. Therefore, most new start-ups that incorporate MEMS are going fabless.  
Mainly captive fabs are not able to deviate much from their internal processes or 
might not have the flexibility to address issues that arise during the early, less 
defined stages. There?s also a concern that external customers will be treated as 
second class citizens compared to internal product lines.  While we can?t ignore 
these companies, we believe we can compete with them very effectively. ? 
 
 
Innovative Micro Technolgoy, Santa Barbara, CA http://www.imtmems.com/  
?We are also in high volume production for a switching application for telecom 
and shipping on the order of two million working MEMS switches each week for 
that customer. Other areas showing strong potential are IR imaging, RF and MM-
wave, and biomedical applications.  Our customers range from those with a spec 
to those with a design, to those that have prototyped or even produced working 
devices. In all cases, we collaborate closely on both process transfer/development 
and design for manufacturability. Our skills and experience in these areas enable 
smooth startup or transfer to our facility of processes even where the tool sets do 
not match up. Tight collaboration is key and several of our customers have 
personnel permanently or periodically assigned at our facility. We believe in 
strong and open partnering as the solution.? 
 
Semefab, Fife, Scotland  http://www.semefab.co.uk/  
?Semefab is a provider of MEMS foundry services and therefore is open to all 
companies seeking silicon processing for MEMS applications. The key 
application fields where Semefab is seeing growth are in microfluidics, RF 
MEMS (albeit still in R&D), biomedicine, accelerometers and thin film 
applications such as pressure sensors. The defense industry is slowly waking up to 
the possible applications of MEMS in security applications and is seen as a major 
contributor to business by 2010.? 
                           
 
 136 
 
Source:    [30] 
 
Figure 12.4 ITRS Roadmap for RF MEMS 
                           
 
 137 
X-FAB Semiconductor Foundries AG, Erfurt, Germany http://www.xfab.com/  
?The MEMS foundry market is certainly growing, but more slowly than expected. 
From a technology standpoint, integration of CMOS and MEMS is hot.  
Companies need to offer a wide variety of process options. X-FAB is continually 
expanding its portfolio, and is investigating or implementing several new 
modules, including DRI etching for bulk micromachining, electroplating, and 
wafer-level packaging/encapsulation.   X-FAB aims to develop batch processing 
for MEMS, and looks for synergies between CMOS operation and MEMS 
operation, such as doing MEMS wafer batch processing instead of single wafer 
processing. We use a KOH batch etch instead of a DRI single wafer etch 
whenever possible. X-FAB also reduces packaging costs by providing wafer-level 
packaging. ? 
 
In addition, captive fabs are expanding their MEMS capabilities and capacity as 
highlighted in  [32]: 
?In 2006, only the two companies in the top 30 MEMS manufacturers, TI and HP, 
were manufacturing MEMS at 200 mm diameters. By 2011, there should be at 
least 12 players: ST?s and Freescale?s 8-inch lines are already operational: Bosch, 
Omron and Canon should follow in the next 18 months. And WTC (Wicht 
Technologie Consulting) knows of a further six companies among the top 30 who 
expect to, but have not yet publicized their plans to move to this diameter. 
The foundry offer also currently expands in 200 mm. Dalsa, tMt, DNP and 
Jazz Semiconductor have started to offer services, while APM, Omron, TSMC 
Silex come on line by 2009.     
RFMD is currently building an 8-inch facility dedicated to RF MEMS and 
NXP will also probably start at this diameter. WiSpry will partner with Jazz 
Semiconductor.? 
 
12.4 Chapter 12 Summary 
The author believes that the proposed single chip front-haul processor (FHP) 
architecture of Fig. 12.3 using RF MEMS or a hybrid approach combining RF MEMS 
and traditional SiGe or GaAs devices is achievable within five years.   
The proposed FHP is a relatively smaller processing unit (as compared to the base 
band processors) which contains the necessary SDR software to either 1) handle the base 
band processing of one or more high volume air interface standards like GSM and W-
                           
 
 138 
CDMA, or 2) pass through the intermediate frequency (IF) of any air interface standard 
(e.g., UMTS, WiMax, 4G, etc) to the base-band processing unit via a wireless (Wi-Fi or 
TBD) or wireline interface (for example, CPRI or OBSAI interface standards).    
The business potential of such device is explored in the next chapter as it relates 
to the femtocell market which is in its infancy stage. 
 
                           
 
 139 
Chapter 13 Femtocell Market Overview 
 
 
13.1 Femtocell Market Trends 
As shown in Fig. 13.1 and Fig. 13.2, femtocells will enable mobile operators to 
push IP out to the edge of the mobile network and generate opportunities for value-add 
services, preserve 3G macro cellular capacity for voice services, and reduce operational 
costs by running backhaul over DSL lines. However, the business model is not yet proven 
and the benefits to consumers are not universally compelling; mobile operators will have 
to develop sophisticated bundles covering phones, data cards/devices, femtocells, and 
voice, data (and likely video) services to attract subscribers. 
The lack of a femtocell standard will inhibit adoption in the next 12 months; any 
such standard would have to incorporate existing standards including 3GPP, 3GPP2, and 
possibly WiMAX for radio access, plus SIP and IMS for IP services, and could also 
include WiFi for local area wireless Ethernet, TR069 for CPE device management and 
IPSec for security.  Although the Femto Forum now has 40 members, including major 
integrated vendors, it is not a standards body and creating a coherent set of accepted 
practices from all the technologies which need to be integrated - not to mention the 
agendas of vendors with competing intellectual property - means that the market will still 
be in a very early phase of adoption through CY 2009. 
                           
 
 140 
 
Source:   Infonetics Research, September 25, 2008 
Figure 13.1 Femtocell vs Macrocells Deployment and Topologies 
 
 
The more widespread availability of HSDPA (and, within the next 12-18 months, 
HSUPA) is driving adoption of mobile broadband; this in turn, is a driver for femtocell 
adoption: once users are hooked on high speed mobile data services, subscribing to a 
femtocell service could follow to ensure coverage and capacity consistency for those 
services in the home or office. 
Not all mobile operators will be able to leverage the potential business benefits of 
3G femtocells; for operators in low-ARPU markets their customer demographics will not 
support the investment required by the end-user to purchase a femtocell and use value-
added mobile data services, in addition to having a DSL line. 
                           
 
 141 
 
Source:   Infonetics Research, September 25, 2008. 
Figure 13.2 Femtocell vs Macrocell Deployment and Topologies 
 
 
In these countries, typically in developing countries, there may be opportunity for 
2G femtocells, driven by cheaper mobile voice calls within the home, but this is not 
likely to be a booming market as 2G coverage does not generally suffer from the in-
building penetration issues that can beset 3G networks. 
3G femtocells could have far-reaching consequences for the telecommunications 
industry; mobile operators have invested heavily in macrocell network construction, but, 
in this scenario, will increasingly divert investment towards indoor femtocell base 
stations, which will have profound implications for cellular network infrastructure 
vendors; potentially, femtocells could allow mobile network operators to make dramatic 
                           
 
 142 
reductions in investment in conventional 3G macrocellular networks and dedicated 
broadcasting networks (such as DVB-H). 
Femtocells still have a number of technology challenges to overcome: it is not 
proven that they will not interfere with signals on the macro cellular network when mass 
deployed, network management, and troubleshooting is not clear-cut as femtocells 
connect directly to fixed broadband lines, which may be operated by another service 
provider, and the efficient scaling up and remote operation of femtocells remains a 
potential challenge. 
The unit price pressures that femtocells will face in the coming years are shown in 
Fig. 13.3.    There it can be seen that the price range will be between $50 to $100 
depending on technology and feature set.    This, however, does not necessarily mean that 
the fabrication cost of the femtocell will be this low as the price to the end user can be 
subsidized by the service subscription as currently done with the handsets.    However, it 
clearly shows that significant pressures will be put in the cost of the components required 
to implement the femtocell function. 
As most adoption is expected to come from consumers, the fall in average 
revenue per unit for femtocells is critical to stimulate mass market adoption; only when it 
falls below $100 in late CY 2009 or early CY 2010 do shipments ramp up into the 
millions. 
                           
 
 143 
  
Source:   Infonetics Research, September 25, 2008 
Figure 13.3 Worldwide Femtocell Revenue per Unit 
 
13.2 Market Size and Forecast 
13.2.1 Overview 
The worldwide femtocell market is expected to reach $1.5B in revenue and 24.7M 
units by CY 2011.  In 2Q08 55% of total revenue came from femtocell gateways and 
45% from femtocells; by CY 2011 it is forecasted this split will shift to 6% and 94% 
respectively; the ratio of femtocell gateways to femtocells in CY 2008 was approximately 
1:1,000 and by CY 2011 this will grow to approximately 1:10,000. 
Operators will likely only launch femtocell services in focused areas of high 
population density in the early years of service; if they were to launch over a geographic 
area which was too widely dispersed, subscriber adoption would be patchier, and the 
proportion of femtocell gateways required to support these users would not be cost-
effective; operators will instead use a build-and-grow model, migrating this template to 
                           
 
 144 
new areas with a focused Capex expectation based on number of gateways required to 
support adoption within that area. 
 
13.2.2 GSM / GPRS Femtocells.  
In CY 2008, the worldwide GSM / GPRS femtocell revenue is expected to hit 
$605K and units are expected to reach 3.5K, based on predicted service launches in 
Western Europe and possibly in Asia Pacific; by CY 2011 revenue will hit $195.5M, a 
2008-2011 CAGR of 586%; units are expected to reach 3.7M, a 2008-2011 CAGR of 
920%; femtocells based on 2G/2.5G technology have a limited appeal, as they support 
predominantly voice centric services, and voice coverage is typically satisfactory over 
2G/2.5G networks; the main adoption drivers for users are home-based cellular call 
tariffs, a valid driver in developing countries with low household income. In such 
countries, the necessity of a broadband line for backhauling femtocell traffic could also 
be cost-prohibitive. 
 
13.2.3 CDMA Femtocells.  
In CY 2008, the worldwide CDMA femtocell revenue is expected to hit $1.9M 
and units are expected to reach 10.2K; by CY 2011 revenue will hit $4.6M, a 2008-2011 
CAGR of 33%; units are expected to reach 92.7K, a 2008-2011 CAGR of 108%; with 
most of the world's CDMA user base migrating to EVDO over this period, the long-term 
opportunity for CDMAone femtocells is limited and will quickly diminish despite being 
the earliest market to ship units. 
 
 
 
 
                           
 
 145 
13.2.4 W-CDMA Femtocells.   
In CY 2008, the worldwide W-CDMA femtocell revenue is expected to hit 
$291.9K and units are expected to reach 5.8K; by CY 2011 revenue will hit $836.9M, a 
2008-2011 CAGR of 1321%; units are expected to reach 13.4M, a 2008-2011 CAGR of 
1222%; this will be driven by Western European mobile operators: for example, 
Vodafone, Orange, T-Mobile, TeliaSonera, and O2 have all announced femtocells trials, 
and some of these operators are expected to launch limited services before the end of this 
year. 
 
13.2.5 CDMA2000 Femtocells.   
In CY 2008, the worldwide CDMA2000 revenue is expected to hit $207.8K and 
units are expected to reach 1.2K; by CY 2011 revenue will hit $454.1M, a 2008-2011 
CAGR of 1198%; units are expected to reach 7.6M, a 2008-2011 CAGR of 1788%; 
Sprint has already launched a service in North America, and Verizon is likely to follow 
suit, driving early market growth.  Over the mid to long term of the forecast period, more 
Asian mobile operators will also enter this market, with at least one service expected to 
go live in late 2008 to early 2009. 
 
13.3 Market Growth Potential 
The femtocell market is just starting to take off.    The growth will be exponential 
starting in CY 2010 as shown in Fig. 13.4 and Fig. 13.5.  Unit volumes are expected to 
grow from 4.7 million units in CY 2010 to over 24 million in CY 2011 with 3G 
technologies (W-CDMA and CDMA2000) being predominant.   Revenue should thus 
reach about $1.5 billion in CY 2010, or almost quadruple that of CY 2009.  While there 
                           
 
 146 
are no forecast projections in  [34] beyond CY 2011, it can be inferred from this analysis 
and other literature that the growth will continue to be exponential at least through CY 
2014. 
 
Source:   Infonetics Research, September 25, 2008 
Figure 13.4 Worldwide Femtocell Units by Technology 
 
 
 
Source:   Infonetics Research, September 25, 2008 
Figure 13.5 Worldwide Femtocell Revenue by Technology 
                           
 
 147 
13.4 Market Size by Geographies of the World  
The strongest early market for femtocells are in developed countries, as 3G 
rollout is more pervasive and it is believed that 3G femtocells have stronger adoption 
benefits for consumers, because 3G suffers from poorer in-building penetration than 2G. 
Also, 3G femtocells will provide more consistent coverage and capacity for mobile data, 
which will be more rapidly adopted by consumers in developed countries with higher 
average household income. 
The large potential of developed countries in both Asia Pacific and Western 
Europe will drive adoption and equipment revenue growth through the forecast term 
provided in reference  [34] and beyond; femtocell adoption will be much later in 
developing countries in MEA, Asia Pacific, and CALA. 
In 2Q08, all revenue came from North America. In CY 2008 it was expected to 
breakdown as: 83% from NA, based on the early femtocell initiative from Sprint in North 
America, with 12% from EMEA and 5% from APAC as services just start to go live, and 
0% from CALA; by CY 2011 (see Fig. 13.6), this shifts with 20% of femtocell revenue 
being from NA, 37% from EMEA, 35% from APAC, and 8% from CALA driven by the 
multitude of mobile operators in Western Europe and developed markets in Asia Pacific 
(e.g., Japan, South Korea, Taiwan, Singapore, Australia, New Zealand) launching 
services.   Fig. 13.7 shows the unit volume projections by geographic region through CY 
2011. 
 
 
                           
 
 148 
 Source:   Infonetics Research, September 25, 2008 
Figure 13.6 Worldwide Femtocell Revenue by Geographic Region 
 
 
 
 
 
Source:   Infonetics Research, September 25, 2008 
Figure 13.7 Worldwide Femtocell Units Percentage by Geographic Region 
 
 
                           
 
 149 
13.5 Femtocell Equipment Manufacturers 
 
Shown in Fig. 13.8 is a list of equipment manufacturers that have either 
introduced or announced the introduction of femtocell equipment as of late 2008.   It can 
be seen that design activity is on going for all femtocell technologies, including 2G and 
3G.  It is also important to note that both Tier 1 manufacturers (like Alcatel Lucent, 
Huawei, Ericsson, and others) as well as Tier 2 and Tier 3 manufacturers are pursuing 
this new market. 
 
 
Source:   Infonetics Research, September 25, 2008 
Figure 13.8 Femtocell Equipment Manufacturers 
 
 
                           
 
 150 
13.6 Femtocell Adoption Roadmap 
 
Femtocell adoption has already started as shown in Fig. 13.9.   However, mass 
deployment will only start in 2013 and beyond when the operators see the Capex and 
Opex benefits of femtocells.  By then, femtocell will become a significant element in the 
4G deployment strategy aimed at dense urban environment and less intensive macro 
cellular rollout. 
 
 
Source:   Infonetics Research, September 25, 2008 
Figure 13.9 Femtocell Adoption Roadmap 
 
 
13.7 Chapter 13 Summary 
The femtocell market concept is a relatively new event which started in 2006.  
While initial adoption has commenced as of early 2009, mass market adoption in the tens 
of millions of units is not expected to occur until 2013, when it will be come a significant 
element in the 4G deployment strategy aimed at the dense urban environment and less 
intensive macro cellular rollout.     
Because femtocells are aimed to be collocated at the user?s home, price points 
will be much lower than picocells and have already reached the sub $100 level.    As a 
                           
 
 151 
result, there will be significant pressure on the cost of the components needed to 
implement femtocells.    
Additionally, it should be noted that femtocell RF front ends will require low 
transmission power and will be powered by a wall-mounted supply source.   Both of 
these elements are necessary for RF MEMS as discussed in  Chapter 11. 
It is precisely because of the above market and product characteristics that the 
author has selected this new market as a case study for a financial and revenue analysis 
for the proposed ?front-haul processor? (FHP) discussed in  Chapter 12.    This financial 
and revenue analysis can be further utilized as the starting point for the new venture 
proposed in  Chapter 14. 
 
 
                           
 
 152 
Chapter 14 Summary and Future Work 
 
14.1 Summary 
The RF front end circuitry makes up over 60% of the cost of a base transceiver 
station (BTS, also referred as Node B in 3G networks).  Included in this cost is that of the 
Power Amplifier unit, RF filters, and the GPS unit for clock recovery and 
synchronization.  Depending on the architecture complexity, number of supported 
sectors, power range specification, and wireless network standards supported, a base 
transceiver station can cost between USD $10,000 to $40,000.   
In addition, because of the evolving mobile wireless communications standards 
(2G/2.5G/3G/LTE/4G and their various implementation standards including CDMA2000, 
WCDMA, TD-SCDMA, WiMAX, etc), a BTS original equipment manufacturer (OEM) 
is faced with the monumental task of designing a BTS RF Front End for each mobile 
network standard that they wish to support.  Moreover, the OEMs are forced to 
manufacture and maintain in the field all of these BTS variants. 
The cost issue and logistics complexity of supporting a multitude of BTS variants 
is further accentuated when looking ahead into future in the next 5-10 years.   One way to 
address the frequency re-use and power range limitations of today?s base stations is to 
have more base stations covering smaller geographical areas (that is, add more cell 
clusters) by implementing what is called a Pico BTS.  The very concept of a Pico BTS 
                           
 
 153 
again implies a significant cost reduction of the BTS to below $5,000 and a Femto base 
station will have to meet a cost target in the range of USD $200. 
Thus the RF Front End represents the single biggest opportunity for cost reduction 
and logistics simplification for the base station of the future: the SDR-based Cellular 
Femto BTS.   While there is a lot of academic and industry research focused on software-
defined radio architectures, SDR still does not solve the RF Front End cost issue.  In fact, 
for SDR architectures to achieve their promised ?programmability? mantra, they need a 
true ?programmable? RF Front End.   The problem is that today?s BTS RF Front End 
implementations are not programmable but fixed to work with a specific RF frequency 
(e.g., 1,900 MHz) for a specific mobile network standard (e.g., CDM2000 or W-CDMA).  
To band-aid this issue, OEMs typically implement multiple RF Front Ends to give the 
illusion of a ?multi-band? programmable system at twice, triple or quadruple the cost of a 
single-band system. 
RF MEMS (micro electro-mechanical systems) is a technology that has the 
potential to not only help solving the BTS cost and logistic issues, but also to 
complement SDR architectures by providing a true low cost, high performance, 
programmable RF Front End.   Some of the advantages of RF MEMS include: very good 
isolation and insertion loss, virtually no control circuit power dissipation in ON or OFF 
state, with proper design it is capable of broadband and high power switching, switching 
speed is sufficient for RF control circuit applications, and relatively low cost as it is 
manufactured by standard semiconductor manufacturing techniques.   
Because of its potential in the implementation of low cost, high performance, 
programmable RF Front End filters and antennas, the author has shown in his research 
                           
 
 154 
via this dissertation work that the possibility of implementing a Single-Chip 
Programmable RF MEMS FRONT END for SDR Cellular Pico BTS for 
commercialization by 2011 is a true possibility. 
 
14.2 Future Work 
The author thus recommends to take the findings of his industry and technical 
research to the next R&D step, which is the formation of a new business venture as 
follows. 
 
14.2.1 A New Venture 
This is a new business venture expected to be founded in the second half of 2009. 
The venture addresses the wireless telecommunications industry, which is undergoing 
migration changes from 2G/2.5G to 3G (third generation) and next to 4G and beyond.   In 
addition, new services are being added like high speed web browsing from mobile 
handsets, multi-media messages, IPTV, video conferencing with mobile handsets, etc. 
The FHP (?front haul processor?) product is currently in the concept stage and 
requires additional detailed feature specification.   The product specification details will 
be done as part of the first phase of the new venture during calendar Q3 of 2009.  A final 
specification will be available in Q4 2009.    
It is expected that it will take approximately 24 calendar months, and 120 man-
months to develop and prototype this product.   The first product customer samples are 
expected to be available in Q4 2011 and initial production is expected to commence in 
mid 2012. 
 
                           
 
 155 
14.2.1.1 Market 
The wireless telecommunications infrastructure is a global market, with very high 
growth rates in underdeveloped regions like China, India and Latin America.  The 
targeted market is Femtocels for the wireless mobile communications market.    See 
 Chapter 13 for the femtocell market overview. 
 
14.2.1.2 Financing 
The venture will first seek ?angel? investors and then will be followed with VC 
investors.  Also, it is possible that financing can be obtained from the OEM system 
integrator.    This last financing has the additional benefit in that it provides a first 
customer to the venture since the founding of the company.  It is expected that ?angel? 
financing will be enough to produce the first full functional requirements specification by 
Q4 2009.   VC funding will be sought after a solid partnership with an OEM integrator 
and RF MEMS and NPU suppliers.    It may also be possible to obtain investment from 
the suppliers as well. 
 
14.2.1.3 Management Team 
Jorge S. Hurtarte (Founder) 
Jorge "Jeorge" S. Hurtarte, P.E., has more than 20 years of semiconductor 
industry experience in design engineering and manufacturing operations. His 
responsibilities at TranSwitch have encompassed worldwide VLSI chip design, software 
and technology development; EDA and manufacturing supplier selection; supply chain 
management and manufacturing logistics; and worldwide Sales. Prior to joining 
TranSwitch in January 1999, Mr. Hurtarte spent 15 years at Rockwell Semiconductor 
                           
 
 156 
Systems, now Conexant Systems. During his tenure there, his responsibilities ranged 
from semiconductor product engineer to director of an engineering group, ultimately 
becoming a vice president of the company. Mr. Hurtarte holds a BS degree in electrical 
engineering from the University of California, Irvine, and a MBA from Auburn 
University. He is a graduate of Harvard Business School's Advanced Management 
Program, and is a Registered Professional Engineer in the State of California. He served 
on the Board of Directors of the Fabless Semiconductor Association during 2002-2005. 
TBD 1 (Co-Founder):   NPU Architect / Specialist 
This individual should have a proven track record in NPU architecture and NPU-
based designs in telecommunications applications. 
TBD 2 (Co-Founder):    RF MEMS / Analog Designer / Specialist 
This individual should have a proven track record in RF MEMS and analog 
circuit design for RF front ends for wireless mobile systems. 
TBD 3 (Co-Founder):  VLSI Designer / VLSI Technology Specialist 
This individual should have extensive VLSI SoC design experience preferably in 
the telecommunications and wireless base stations industry. 
 
14.2.1.4 Product 
Purpose of the product 
The FHP is a base transceiver station (BTS) line card semiconductor device with 
embedded RF MEMS wireless interface integrated with a real time processing unit and 
the necessary software to automatically configure itself to operate with any 2G/3G /4G 
wireless communications systems.   
                           
 
 157 
The FHP is a relatively smaller processing unit (as compared to the base band 
processors) which contains the necessary SDR software to either 1) handle the base band 
processing of one or more high volume air interface standards like GSM and W-CDMA, 
or 2) pass through the intermediate frequency (IF) of any air interface standard (e.g., 
UMTS, WiMax, 4G, etc) to the base-band processing unit via a wireless (Wi-Fi or TBD) 
or wireline interface (for example, CPRI or OBSAI interface standards).    
The advantages of this innovative architecture are: 
1. Off loading of the main base-band processing unit when handling the high 
volume air interface base-band processing.   In this configuration, the FHP in 
essence becomes a base station on its own merit.   The back haul connection 
of the FHP becomes either the Wi-Fi or wire line interface to a nearby Base 
Station Controller.   In this configuration, the FHP is well suited for a Femto 
cell application supporting a single air interface standard. 
2. Providing flexible FHP-to-BB wireless or wire line interfaces which are easily 
configurable in the FHP?s SDR software and built-in programmable RF 
MEMS circuitry 
3. No need to carry multiple part numbers in production inventory as one device 
does it all with software programmability 
4. Ability to download new software updates as wireless systems standards 
change or evolve 
 
To facilitate the design in of the FHP device, a reference design will be available 
as well as a full set of software application programming interfaces (APIs) and technical 
documentation. 
 
14.2.1.5 Unique features 
The unique features of the FHP product are: 
Real time automatic adaptation to any 2G /3G/4G wireless network standard. 
? On chip base-band processing for femtocell application 
? Off loading of the main base-band processing unit when handling  the high 
volume air interface base-band processing.   In this configuration, the FHP in 
essence becomes a base station on its own merit.   The back haul connection 
                           
 
 158 
of the FHP becomes either the Wi-Fi or wire line interface to a nearby Base 
Station Controller.   In this configuration, the FHP is well suited for a Femto 
cell application supporting a single air interface standard. 
? Providing flexible FHP-to-BB wireless or wire line interfaces which are easily 
configurable in the FHP?s SDR software and built-in programmable RF 
MEMS circuitry 
? No need to carry multiple part numbers in production inventory as one device 
does it all with software programmability 
? Ability to download new software updates as wireless systems standards 
change or evolve. 
 
14.2.1.6 Trademarks, Patents, Copyrights, Licenses, and Royalties 
No patents, trademarks, service marks, or copyrights have been obtained yet.  
However, the unique concept of a ?dual personality? of the front haul processor deserves 
a trademark filing to begin with. 
No license or royalty agreements are in place yet.  It expected that license 
agreements will be need for the NPU of choice.   All other circuits will be designed by 
the venture. 
It is expected that several patents, trademarks, and license agreements will be 
pursued as the definition and development of the product proceeds. 
 
14.2.1.7 Governmental approvals 
No governmental approvals are required to ship the product.   However, the 
customers using the product may need some form of type approvals will be needed and 
provided by the telephone carriers where their products will be sold. 
 
14.2.1.8 Production 
The company will be a fabless company.   All manufacturing will be 
subcontracted out. 
                           
 
 159 
14.2.1.9 Suppliers 
The primary suppliers are EDA (engineering design automation) tool providers 
like Cadence, Synopsys, Magma, Verisity, etc, and IP (intellectual property) providers 
like ARM, Tensilica, etc.    Also, since the company is fabless, fab and packaging 
suppliers will be contracted to perform all manufacturing and test operations. 
 
14.2.2 Future Research and Development 
The following are the primary R&D milestones and completion timeframe.  The 
dates shown are subject to change if funding is delayed beyond Q4 2009. 
 
Product Outline Specification      Q3 2009 
Functional Requirements Specification    Q4 2009 
RF MEMS Technology Fab partner selection   Q3 2009 
RF MEMS front end design      Q2 2010 
NPU IP core selection      Q3 2009 
SDR OEM partner selection for system software integration Q4 2009 
VLSI design and verification      Q2 2010 
RF MEMS and NPU integration     Q3 2010 
FHP system verification      Q4 2010 
Tape Out        Q1 2011 
Silicon-level verification      Q2 2011 
Initial Samples       Q3 2011 
Reference Design Specification     Q4 2009 
Reference Design Development     Q1 2010 
                           
 
 160 
14.3 Business Case 
A full business case should be developed to include a Financial Plan with Sales 
Projections, Income Projections, Cash Requirements, Sources of Financing, competitive 
analysis, financing and exit strategies, etc.   A preliminary Financial Plan with Sales 
Projections is shown in Appendix D. 
                           
 
 161 
References 
 
[1] Holma, Harri and Toskala, Antti, WCDMA for UMTS, Wiley, Third Edition, 
2004 
[2] Dharma Prakash Agrawal and Qing-An Zeng, Introduction to Wireless and 
Mobile Systems, Thomson, Second Edition 
[3] http://www.eng.auburn.edu/users/hurtajs/index.html 
[4] http://www.reuters.com/article/pressRelease/idUS17950+29-Jan-
2008+BW20080129 
[5] http://www.tca.or.jp/eng/database/daisu/yymm/0808matu.html 
[6] http://www.nttdocomo.com/pr/2005/000672.html 
[7] http://www.kantei.go.jp/foreign/it/network/0122summary.html 
[8] http://www.nttdocomo.co.jp/english/product/ 
[9] http://www.nttdocomo.co.jp/english/service/ 
[10] www.mobilein.com/Perspectives/eJapan_Article_062206.pdf 
[11] http://www.ft.com/cms/s/0/4821e822-5ff2-11dd-805e-000077b07658.html 
[12] http://www.ngmn.org/index.php?id=47 
[13] 3GPP 23.107 
[14] http://www.it.uc3m.es/banchs/papers/ew04.pdf 
[15] Young Kyun Kim and Ramjee Prasad, 4G Roadmap and Emerging 
Communications Technologies, Artech House Personal Communications, 2006 
[16] D. O?Brien and M. Katz, ?Optical Wireless Communications Within Fourth-
Generation Wireless Systems,?   Journal of Optical Networking, Vol. 4, May 2005, pp. 
312-322. 
[17] Mobile IT Forum Website, http://www.mitf.org/public_e/archives/index.html 
[18] Vanu, Inc. Website, http://www.vanu.com/  
[19] http://www.sdrforum.org/  
[20] P.B. Kenington, ?Emerging Technologies for Software Radios,? IEEE Electronics 
and Communications Engineering Journal, Vol. 11, No. 2, April 1999, pp. 69-83 
[21] P.B. Kenington, 2005, RF and Baseband Techniques for Software Defined Radio, 
Artech House Mobile Communications, p. 58, 2005 
[22] The ARRL Handbook for Radio Communications, The National Association of 
Amateur Radio, Eighty-Thrid Edition, 2006 
[23] R. Lucero,et al., Design of an LTCC Switch Diplexer Front End Module for 
GSM/DCS/PCS Applications. IEEE, 2005 
[24] RF-MST Cluster Workshop, European Commission, Information Society and 
Media Directorate-General, Fodele / Heraklion, Crete, Greece, June 30, 2008 
[25] G. M. Rebeiz, RF MEMS Theory, Design, and Technology, Wiley, 2003, p. 3 
[26] D. Ando, K. Oishi, T. Nakamura, and S. Umeda, Glass direct bonding technology 
for hermetic seal package, in 10th IEEE International Conference on Microelectro-
mechanical Systems, 1997, pp. 186-190 
                           
 
 162 
[27] J. B. Yoon, C. H. Han, E. Yoon, and C. K. Kim, High-performance three-
dimensional on-chip inductors fabricated by novel micromachining technology for RF 
MMIC, in IEEE MTT-S International Microwave Symposium Digest, June 1999, pp. 
1523-1526. 
[28] M. Kim, J. B. Hacker, R.E. Mihailovich, and J. F. DeNatale, A Monolithic 
MEMS switched dual-path power amplifier,  IEEE Microwave Wireless Comp.  Lett., 
Vol. 11, No. 7, pp. 285-286, July 2001 
[29] J. Hurtarte, E. Wolsheimer and L. Tafoya, Understanding Fabless IC Technology, 
Elsevier Publishing, 2007, page 216. 
[30] http://www.itrs.net/Links/2008ITRS/Update/2008Tables_FOCUS_A.xls#'2007_R
FAMS8'!A1 
[31] http://www.smalltimes.com/display_article/260006/109/ARTCL/none/none/1/ME
MS-fab-services/ 
[32] http://www.wtc-consult.com/deutsch/issue-1-v3.html 
[33] http://www.datangmobile.cn/en/TD-SCDMA/standard.htm 
[34] Infonetics Research, September 25, 2008 
[35] J. A. Timmon and S. Spinelli, New Venture Creation, 6th Edition, 2004 
[36] Dr. C. Nguyen, Tutorial:  MEMS for Frequency and Tuning References, FCS 
2005 
 
 
                           
 
 163 
Appendix A Wireless Communications Technologies 
 
2G 
 
TDMA: One of the original digital standards, TDMA divides call spectrum into different 
time slots to transmit multiple calls simultaneously.  Most TDMA networks in the U.S. 
have been overlaid with GSM and upgraded to support GPRS and EDGE; some in Latin 
America, however, remain on the legacy standard. 
 
iDEN: A TDMA-based technology that also supports data, short messaging and dispatch 
radio (two-way radio). Nextel in the U.S. uses the iDEN standard, though with its 
acquisition by Sprint, Nextel?s subscribers are likely to migrate to CDMA-based 
technologies. 
 
GSM: A TDMA-based technology that uses SIM cards to identify individual subscribers. 
GSM is the predominant 2G standard in use throughout Europe, Asia and the Middle East 
and Africa, though most GSM carriers have added 2.5G GPRS across the network and 
most handsets are GPRS capable as well. These networks will eventually migrate to W-
CDMA/UMTS, and many carriers are well underway with their network upgrades. 
 
PDC: A TDMA-based technology that is widely used in Japan. Most carriers in Japan 
have already upgraded their PDC networks to W-CDMA and are actively migrating their 
subscriber bases, skipping the 2.5G node entirely. 
 
CDMA: Pioneered by Qualcomm, CDMA uses a spread spectrum technique to overlap 
every transmission on the same carrier frequency, assigning a unique code to each signal. 
This requires fewer cell sites for the same number of calls. 2G CDMA networks 
(sometimes referred to as ?CDMAone?) have already moved to 2.5G CDMA2000 1xRTT 
and are already starting to roll out 3G CDMA2000 1xEV-DO. 
 
2.5G 
 
GPRS: GPRS is an update to GSM that supports data packets. Most GSM carriers are 
now fully GPRS-capable, with GPRS handset shipments now far outpacing 2G GSM 
shipments. 
 
EDGE: EDGE is a 2.5G enhancement that increases data throughput to 384 Kbps. 
 
CDMA2000 1xRTT: 2.5G standard that has replaced 2G CDMA, adding enhanced data 
capabilities and throughput. 
 
                           
 
 164 
3G 
 
W-CDMA/UMTS: 3G standard that is beginning to replace GSM, TDMA and PDC. W-
CDMA, as the name implies, brings CDMA technology to the GSM world, and thus, it 
requires new network hardware in addition to different handsets (whereas GPRS was a 
mostly a software upgrade in the network). Note that Japanese carriers have already 
rolled out W-CDMA and are actively migrating their subscriber bases from PDC; carriers 
in Europe and the U.S. are implementing W-CDMA/UMTS now. 
 
CDMA2000 1xEVDO: 3G CDMA standard that boosts throughput for the data side of 
the handset. The EVDO build out has accelerated in recent quarters, and EVDO services 
are now available from carriers in the U.S. and Asia. 
 
CDMA2000 1xEVDV: 3G CDMA standard that boosts throughput for both the voice and 
data sides of the handset. Qualcomm (and therefore its handset customers) have shelved 
plans for EVDV chipsets and is instead focusing on EV-DO. 
 
TD-SCDMA: 3G standard being developed from the ground up for China. The Chinese 
Ministry of Information Industry is spearheading the effort to drive this standard along 
with Chinese carriers, networking OEMs, and handset vendors. 
 
                           
 
 165 
Appendix B TD-SCDMA Standard 
 
TD-SCDMA (Time Division - Synchronous Code Division Multiple Access), proposed 
by Datang Telecom Technology and Industry Group on behalf of the Chinese 
government, was approved by the International Telecommunication Union (ITU) in May 
2000, and by 3rd Generation Partnership Project (3GPP) in March 2001 as one of the 3G 
mobile communications standards. 
 
TD-SCDMA adopts high-edged technologies such as Smart Antenna, Software Radio, 
Joint Detection, Banton Handover, High-speed Transmission technology for Downlink 
Packet Data. Compared with other mobile systems, TD-SCDMA boasts outstanding 
technological benefits: 
 
High Spectrum Efficiency: TD-SCDMA, with the highest spectrum efficiency, better 
supports dense services in populated areas. It is capable of making full use of fragmented 
spectrum, and effectively alleviates spectrum resource shortage and limitations at 
carriers? side. 
 
High Capacity: Adoption of high-edged technologies dramatically lowers interference 
and increases system capacities. 
 
Highly Suitable for Operators Asymmetrical Data Services: As TD-SCDMA is 
competent for dynamically adjusting data transmission rates with uplink and downlink, it 
specially suits to handing asymmetrical IP data services. 
 
Low Costs: Adoption of Smart Antenna technology decreases transmission Power in TD-
SCDMA system and dramatically lowers the cost of the system products. 
 
Source:  [33] 
 
                           
 
 166 
Appendix C Glossary of Terms 
 
2G:  Second generation  (Analog) wireless infrastructure 
 
3G:  The third generation of wireless networks. These networks should be able to support 
peak data rates of 144 Kbps at mobile user speeds, 384 Kbps at pedestrian user speeds 
and 2 Mbps in fixed locations (peak speeds), although some initial deployments were 
configured to support just 64Kbps. ITU coordinates 3G standards through its IMT-2000 
project and incorporates the key standards bodies, 3GPP and 3GPP2. 
 
ADSL:  Asymmetrical digital subscriber lines, a method of transmitting data over 
traditional copper telephone lines. Data can be downloaded at speeds of up to 1.547 Mbps 
and uploaded at speeds of 128 Kbps. 
 
ASON:  Automatic switching optical network, the functions of the control interface of 
which allow it to develop an optical communication channel customized to users? needs 
according to their requests on an unmanned basis and support multi-channel 
development, capacity expansion, configurability and network intelligence development. 
 
CAPEX:  Capital expenditures 
 
CDMA: Code division multiple access, one of the standards for 2G mobile 
communications. It is a spread spectrum technology standard that assigns a pseudo-noise 
(PN) code to all speech and data bits, sends a scrambled transmission of the encoded 
speech over the air and reassembles the speech in its original format. By assigning a 
unique correlating code to each transmitter, several simultaneous conversations can share 
the same frequency allocations. 
 
CDMA2000:  A technical specification for the provision of enhanced capacity for voice 
under the CDMAOne standard with a maximum data rate of 2Mbps, comprising 
particular specifications such as 1xRTT (radio transmission technology), 1xEVDO (data 
only version) and 1xEVDV (voice and data version). 
 
DSP:  Digital Signal Processor (or Processing) 
 
DWDM:  Dense wavelength division multiplexing, a technology that enables a single 
optical fiber to carry multiple data channels (or wavelengths). Commercial DWDM 
systems now exceed 100 channels. 
                           
 
 167 
DSL:  Digital subscriber lines, the collective name given to a number of techniques used 
for transmitting digital data over the local loop or subscriber line. These are also known 
as xDSL. Examples are ADSL, HDSL, VDSL, MDSL and RDSL. 
 
EDGE:  Enhanced data for GSM evolution 
 
EVDO:  Evolution Data Only 
 
FCC:  Federal Communications Commission 
 
FRE:  Radio Frequency Front End circuit 
 
GSM:  A global system for mobile communications, a digital cellular phone system 
standard that originated in Europe. It is deployed in more than 170 countries and uses a 
TDMA radio propagation scheme. 
 
IP:  Internet protocol, as more specifically defined in RFC 791, the primary purpose of 
which is to define packet architecture and address format. 
 
HSDPA:  High-Speed Downlink Packet Access.  Third-Generation Partnership Project?s 
high-speed downlink packet-access 
 
NGN:  Next generation network, a data packet-based network capable of providing a 
variety of services by using multiple broadband transmission technology to support 
unlimited access to different service providers and of supporting ordinary mobile 
communication to provide uninterrupted services to users. 
 
PHS:  Personal handy phone system, a digital mobile telephone system using technology 
developed according to Japanese standards and operating on the 1900Mhz frequency. 
 
SNR:  Signal-to-noise ratio 
 
Softswitch:  Abbreviation for software switch, softswitch is an application protocol 
interface which is used to link a traditional PSTN to IP networks and manage traffic 
containing a mixture of voice, fax, data and video. 
 
TD-SCDMA:  Time division synchronous code division multiple access, a 3G technology 
developed in China to support voice and data transmission. 
 
WCDMA:  Wideband CDMA, a UMTS standard for 3G digital mobile networks 
adopting CDMA technologies to provide enhanced capacity for voice with a theoretical 
maximum data rate of 3Mbps. 
 
WiMax:  Worldwide interoperability for microwave access 
 
WTO:  World Trade Organization
                           
 
 168 
Appendix D Preliminary Financial Plan and Sales Projections 
 
 
This appendix summarizes the rough Sales and Financial Plan estimates for the 
start-up company explained in section  14.2.1 (A New Venture).   The numbers will be to 
be verified and fine tuned as part of the Future Work recommended in the same 
referenced section. 
 
D.1 General Assumptions: 
? It will take 18 months of start up work before Year 1 Sales start. 
? The new venture will be funded with venture capital equity 
? The company?s addressed market and R&D development is a explained in 
section  14.2.1. 
D.2 Specific Assumptions 
The following tables show the specific assumptions related to Years 1, 2, and 3 
related to: 
? Units (from Fig. 13.4),  
? Company?s market share projections  
? ASP and COGS projections 
? Sales projections 
? Headcount and Salaries assumptions 
                           
 
 169 
Annual Unit Growth % Projections
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 200% 400%
FHP-CDMA2000 200% 200%
FHP-GSM/GPRS 200% 300%
FHP-4G 500%
Annual Units Projections
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 15,000,000     30,000,000      120,000,000    
FHP-CDMA2000 8,000,000       16,000,000      32,000,000      
FHP-GSM/GPRS 5,000,000       10,000,000      30,000,000      
FHP-4G -                  1,000               5,000               
Total 28,000,000     56,001,000      182,005,000    
ASP Projections ($)
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 20.0$              15.0$               10.0$               
FHP-CDMA2000 18.0$              13.0$               8.0$                 
FHP-GSM/GPRS 15.0$              10.0$               7.0$                 
FHP-4G -$                30.0$               25.0$                
 
Gross Margin Projections (%)
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 50% 40% 30%
FHP-CDMA2000 40% 30% 25%
FHP-GSM/GPRS 30% 25% 20%
FHP-4G 80% 70%
Unit Cost Projections ($)
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 10.0$              9.0$                 7.0$                 
FHP-CDMA2000 10.8$              9.1$                 6.0$                 
FHP-GSM/GPRS 10.5$              7.5$                 5.6$                 
FHP-4G -$                6.0$                 7.5$                 
Market Share Projections (%)
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 10% 20% 30%
FHP-CDMA2000 15% 25% 30%
FHP-GSM/GPRS 5% 10% 15%
FHP-4G 0% 50% 60%  
 
 
                           
 
 170 
Market Share Unit Projections
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 1,500,000       6,000,000        36,000,000      
FHP-CDMA2000 1,200,000       4,000,000        9,600,000        
FHP-GSM/GPRS 250,000          1,000,000        4,500,000        
FHP-4G -                  500                  3,000               
Total 2,950,000       11,000,500      50,103,000      
Sales Projections ($K)
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 30,000,000$   90,000,000$    360,000,000$  
FHP-CDMA2000 21,600,000$   52,000,000$    76,800,000$    
FHP-GSM/GPRS 3,750,000$     10,000,000$    31,500,000$    
FHP-4G -$                15,000$           75,000$           
Total 55,350,000$   152,015,000$  468,375,000$  
COGS Projections ($K)
Product Category Year 1 Year 2 Year 3
FHP-WCDMA 15,000,000$   54,000,000$    252,000,000$  
FHP-CDMA2000 12,960,000$   36,400,000$    57,600,000$    
FHP-GSM/GPRS 2,625,000$     7,500,000$      25,200,000$    
FHP-4G -$                3,000$             22,500$           
Total 30,585,000$   97,903,000$    334,822,500$  
Average Gross Margin % 45% 36% 29%
Average COGS % 55% 64% 71%  
 
 
 
D.3 Financial Tables 
 
Based on above general and specific assumptions, the author used the 
methodology taught and the Excel software provided by New Venture Creation  [35] and 
produced the following financial tables.   Again, these tables are very rough and for 
illustration purposes.   The tables will need to be updated as the New Venture?s concept 
matures and the team founders agree to the numbers together with possible investors. 
                           
 
 171 
START-UP FUNDING & EXPENDITURES
Auburn Semiconductor Systems
Assume 18 months of start-up work before the start of Year 1 Sales.
Start-up Cash $
  Equity Investments 6,000,000    
  Loan Proceeds -                   
  Real-Estate Loans -                   
Total Start-up Cash 6,000,000    
Start-up Expenditures
Security Deposits
  Rent (last month's) 5,000           
  Telephone Deposit 1,000           
  Utilities Deposit 1,000           
  Other Deposits 2,000           
Total Security Deposits 9,000           
Start-up Expenses
  Accounting Fees 9,000           
  Activation Fee -                   
  Corporate Fees & Taxes -                   
  Federal Tax ID 18,000         
  Fictitious Name Costs 18,000         
  Insurance 9,000           
  Legal & Consulting Fees 9,000           
  Meals & Entertainment 9,000           
  Office Supplies 9,000           
  Payroll Expenses (training/setup)
     Salaries & Wages 3,067,500    
     Payroll Taxes 920,250       
     Benefits 613,500       
  Pre-opening advertising 9,000           
  Printing (cards, stationery, brochures) 1,800           
  Sales Tax Permit 1,800           
Rent 90,000         
Total Start-up Expenses 4,784,850    
Other Costs
  Opening Inventory -                   
Capital Expenditures
  Computer Equipment 250,000       
  Equipment/Machinery -                   
  Furniture & Fixtures 40,000         
  Vehicles -                   
  Leasehold Improvements -                   
  Buildings -                   
  Land -                   
Total Start-up Capital Expenditures 290,000       
Total Start-up Expenditures 5,083,850     
                           
 
 172 
2011 2012 2013
Net Sales (less returns & allowances) 54,796,500          150,494,850        463,691,250        
Cost of Goods Sold 30,279,150          96,923,970          331,474,275        
Gross Income $ 24,517,350          $ 53,570,880          $ 132,216,975        
Operating Expenses
  Advertising 12,000                 3,000                   3,000                   
  Bad Debt Expense 2,739,825            7,524,743            23,184,563          
  Bank Charges 300                      300                      300                      
  Depreciation & Amortization 89,048                 89,048                 89,048                 
  Dues & Subscriptions 1,200                   1,200                   1,200                   
  Insurance 21,000                 12,000                 12,000                 
  Licenses & Fees 37,800                 -                          -                          
  Marketing & Promotion 61,800                 60,000                 60,000                 
  Meals & Entertainment 27,000                 18,000                 18,000                 
  Miscellaneous -                          -                          -                          
  Office Expense 1,800                   1,800                   1,800                   
  Office Supplies 12,600                 3,600                   3,600                   
  Outside Services 12,000                 12,000                 12,000                 
  Payroll Expenses
    Salaries & Wages 5,112,500            2,882,000            3,013,000            
    Payroll Taxes 1,533,750            864,600               903,900               
    Benefits 1,022,500            576,400               602,600               
  Professional Fees 18,000                 -                          -                          
  Property Taxes -                          -                          -                          
  Rent 120,000               120,000               120,000               
  Repairs & Maintenance 12,000                 12,000                 12,000                 
  Shipping & Delivery 6,000                   6,000                   6,000                   
  Telephone 36,000                 36,000                 36,000                 
  Training & Development -                          -                          -                          
  Travel 120,000               120,000               120,000               
  Utilities 24,000                 24,000                 24,000                 
  Vehicle 12,000                 12,000                 12,000                 
Other -                          -                          -                          
Other -                          -                          -                          
Other 90,000                 -                          -                          
Total Operating Expenses $ 11,123,123          $ 12,378,690          $ 28,235,010          
Operating Income $ 13,394,227          $ 41,192,190          $ 103,981,965        
Interest Expense -                          -                          -                          
Other Income (interest, royalties, etc.) -                          -                          -                          
Income Before Taxes $ 13,394,227          $ 41,192,190          $ 103,981,965        
Income Taxes (if C Corp) -                          -                          -                          
Net Income $ 13,394,227          $ 41,192,190          $ 103,981,965        
Auburn Semiconductor Systems
 Year-End
Income Statement (Projected)
 
                           
 
 173 
2011 2012 2013
Assets
Current Assets
Cash  & Equivalents 14,371,782    72,915,235    202,233,459     
Accounts Receivable 4,566,375      17,557,733    61,825,500       
Inventory 305,850         (38,654,958)   (112,779,674)   
Security Deposits 9,000             9,000             9,000                
Other Current Assets -                     -                     -                       
Total Current Assets $ 19,253,007    $ 51,827,009    $ 151,288,285     
Fixed Assets
Property, Plant & Equipment 290,000         290,000         290,000            
Less:  Accumulated Depreciation (89,048)          (178,095)        (267,143)          
Other Non-Current Assets -                     -                     -                       
Total Non-Current Assets $ 200,952         $ 111,905         $ 22,857              
Total Assets $ 19,453,959    $ 51,938,914    $ 151,311,142     
Liabilities
Current Liabilities
 Accounts Payable 4,078,000      7,728,422      34,313,275       
Line of Credit -                     -                     -                       
Other Current Liabilities -                     -                     -                       
Total Current Liabilities $ 4,078,000      $ 7,728,422      $ 34,313,275       
Long-term Liabilities
Loans -                     -                     -                       
Mortgages -                     -                     -                       
Other Non-Current Liabilities -                     -                     -                       
Total Non-Current Liabilities $ -                     $ -                     $ -                       
Total Liabilities $ 4,078,000      $ 7,728,422      $ 34,313,275       
Equity
Equity Investments 6,000,000      6,000,000      6,000,000         
Retained Earnings 13,394,227    54,586,417    158,568,382     
Less:  Owner's & Investor's Draws (4,018,268)     (16,375,925)   (47,570,515)     
Total Equity $ 15,375,959    $ 44,210,492    $ 116,997,868     
Total Liabilities and Equity $ 19,453,959    $ 51,938,914    $ 151,311,142     
Auburn Semiconductor Systems
Balance Sheet (Projected)
Year-End