ADAptive Multicarrier Access System

Deliverables

Currently, there are a number of factors that prohibit the development of broadband Fixed Wireless Access (FWA) systems such as cost, lack of licenses and system unreliability. There have been numerous typical unsuccessful examples of FWA systems with poor business planning that led to unobtainable targets and unsatisfied customers. Also, new entrants want alternative access routes but they have not been satisfied since the majority of the frequency bands are reserved. Additionally, there is an increasing need for delivery of broadband services to individuals, households and businesses of all sizes. The ADAMAS project result is a powerful, adaptive, point-to-multipoint FWA system that can overcome the above difficulties by providing a low cost, flexible service-delivery mechanism, and the possibility for new entrants to apply alternative routes for new services deployment. In addition, ADAMAS can support a system deployment architecture that enables cost reduction and efficient utilisation of frequencies, since the total cost per line of a broadband fixed wireless system is still prohibitively high. There are two sources of cost: licensing and equipment. Thus, a system such as ADAMAS that uses the available spectrum efficiently and at the same time reduces the total cost per line can be very useful for such applications. The results of the ADAMAS project demonstrate that cost reduction and efficient utilisation of frequencies can be obtained with the selection of a Time Division Duplex (TDD) scheme accompanied with adaptation issues in DLC and PHY layers. In wireless access systems, two techniques are considered for the medium access scheme: the Frequency Division Duplex (FDD) and the Time Division Duplex (TDD). TDD requires only a single unpaired frequency band, which is divided in a downlink and an uplink period while FDD requires two distinct paired traffic frequency bands for uplink and downlink period, acting on separate frequencies. Although the use of TDD schemes allows the reduction of the cost, the system design is still based on the worst-case scenario, so there is a need for adaptation in order to achieve spectrum efficiency. The adaptation can be either in DLC or in PHY layer and is very important enhancing its performance. Adaptive systems offer a number of advantages such as the mitigation of the traffic variation, the control of QoS, the fading compensation and the increase of system capacity. The project addressed the adaptation issues in order to increase the transmission bit rate assuming constant transmitted power. Moreover, an objective of this project was to determine the range of frequencies where the OFDM might be not beneficial in terms of complexity and performance when compared to single carrier modulation schemes. OFDM systems provide a robust behaviour in the presence of multipath fading. It is very tolerant to ghost effects (provided that a suitable cyclic prefix is used) eliminating the Intersymbol and Interchannel Interference and providing a good protection against co-channel interference and impulsive noise. In mm waves, the multi-path phenomenon decreases while the line of sight seems to be mandatory. Additionally, the ADAMAS project addressed an integrated digital receiver with two-engine architectures (state-space and likelihood-metric engines) that was optimized from the perspective of the performance/complexity trade-off. This architecture was parallelised and it now has inherent reconfigurable features that can be reconfigured adaptively (''on the fly") by autonomous receiver decisions as the conditions, or even the basic traditional design frameworks, change. In conclusion, the ADAMAS results have yielded a highly efficient wireless access system that can contribute considerable to the growth of broadband access.

RUNCOM's BENEFITS FROM ADAMAS RUNCOM as a commercial partner regards collaboration in ADAMAS with other partners as a unique opportunity for future commercial exploitation of the OFDM technology which we believe will serve as the basis for a large scale broadband access system. The implications of the adaptive OFDM are profound in opening new business opportunities for broadcasters, broadband communication service providers and telecommunication companies in introducing new broadband access technology (OFDMA). Appreciable part of the research work carried out by the partners and especially by the universities have been studied in RUNCOM and as a result the technical team had better view that was used to achieve more scores on the standard committees. OVERVIEWRUNCOM major participation in the development of the adaptive OFDM baseband and the Wireless Data Link Control was to exploit its professional expertise and in-depth know-how, gained in previous development of advanced military communication systems and commercial WLL systems (Narrow-band and Wide-band Multi-Code-CDMA), in developing a highly survivable and affordable two way broadband communication system capable of offering a wide range of services (Video, data and telephony) to business and private customers. RUNCOM has used in the OFDM modem that has been developed within ADAMAS project VHDL cores that were a prior knowledge and IP (Intellectual property) of RUNCOM. Amendments have been done to the cores in order to meet the specification of ADAMAS modem. As the leader of the OFDMA technology in the world RUNCOM has not created any new IP (Intellectual property) in the OFDM technology. Some general ideas in ADAMAS project have helped RUNCOM especially in the standard bodies (IEEE 802.16a, ETSI/HIPERMAN). EXPLOITABLE RESULTS RUNCOM has exploited the tangible results of ADAMAS and invest further efforts beyond the scope of ADAMAS project to fully commercialize products in the following areas: I. To exploit the results of the ADAMAS system as an enhancement to other systems using in particular, OFDM modelling and carry out full-scale system development with the aim to reach a mature commercial system. II. Marketing of versatile FPGA board with Library of cores as a powerful tool for developing various broadband communication applications in fast and efficient way. III. The collaborative nature of IST programme and in ADAMAS consortium was very important in influencing standard bodies to adopt the OFDM uplink and the interactive concept underlying ADAMAS work plan. A list of a such contributions and comments to the standards are presented within this document. RUNCOM's GENERIC BOARDRUNCOM has developed an FPGA Generic platform in order to test the modem base-band VHDL cores. The base band board was developed as a generic board that can be used in ADAMAS project for the TDD OFDM modem. The generic board is consisting of interfaces to network on the base station and compatible interfaces to the customer unit as well. Full interfaces to RF module, PC, and JTAG are allocated on the board. The generic board is fully programmable hardware (FPGAS) for the implementation of the BASE BAND TDD OFDM algorithms and for the interface with the DLC-PHY layer.

The ADAMAS baseband prototype platform is based on high-programmable components able to accommodate the full fixed wireless access baseband modem functionality, including the control for the IF and RF sections. The platform consists of two boards: a digital board (baseband modem board) and a hybrid Digital/Analog board. The digital board has a Compact PCI (cPCI) interface and can be fit in a backplane of a cPCI rack, as in the ADAMAS demonstrator case. It hosts the baseband modem functionality including the HW part of the MAC layer, the control of the analogue converters (residing in the hybrid board) and performs the IF and RF synthesizers programming in addition to the switching and control commands for TDD operation. The digital board consists of fully programmable components: A Xilinx Virtex II 6 million gates FPGA and a Texas Instruments TMS320C6416 fixed point DSP. An efficient HW (FPGA) and SW (DSP) partitioning has been specified and implemented in ADAMAS in order to assure flexibility in design, debug and implementation and robustness in the real-time operation of the system. The main responsibility of the FPGA in the transmit direction is the HW MAC Tx part, the full baseband modem transmitter and the interface (control & data towards DAC) with the hybrid board. In the receive direction the FPGA interfaces with the hybrid board (control & data towards ADC), performs the digital IQ demodulation and then interfaces with the DSP. After the DSP processing, it performs the last part of the baseband receiver chain consisting of the QAM decoder, de-interleaver, Viterbi, de-randomizer and the Rx part of the HW MAC. The DSP in the Tx direction plays the bridging role between the DLC and SW MAC functions residing on another cPCI board via its PCI interface, and the HW MAC functions in the FPGA. At the receive direction the DSP has the responsibility to do the main baseband receiver algorithms such as Time & Frequency Synchronization, Cyclic Prefix extraction and FFT, Channel Estimation and Equalisation, Phase Noise tracking and correction etc., and extracts the metrics used for the ADAMAS adaptive mechanism. Finally it forwards the packets from the HW MAC Rx part to the DLC via the PCI interface. In addition to the above, the DSP plays the role of the overall modem controller since it handles all the events coming from the FPGA and the DLC and sequences the needed actions, and also controls the switching & control of the ADC and DAC and IF and RF subsections by giving appropriate commands to the specific interfacing FPGA modules. Controlled functions include the opening/closing of the analogue converters, programming the IF & RF synthesizers, switching (Tx/Rx) the IF and RF subsections, RSSI, AGC, APC, and AFC. The hybrid Digital/Analog board hosts the interfaces to/from the Digital board and the IF indoor unit, the Analog to Digital and Digital to Analog converters including the corresponding anti-aliasing filters, and finally the master reference clock source that drives the converters, the digital board and the IF/RF section. The hybrid board also forwards the digital control words towards the IF and RF subsections. At the transmit path it outputs a Low-IF IQ modulated signal at 36.125 MHz towards the IF IDU, while at the receive path it gets the signal at the same frequency from the IF IDU. ADAMAS modem platform can be also utilized outside the ADAMAS project as a generic OFDM modem prototype platform that can be used in many other OFDM/Multicarrier modem systems (such as powerline or wireline systems and wireless LANs) thus extending the exploitable potential of the ADAMAS results.

1. RF/IF Technology in ADAMAS CRL�s role in ADAMAS was the development of the 5.8GHz and 10.5GHz wireless subsystems. There were other contributions to the project in other areas, but these were the key efforts.The RF subsystems are described in deliverable 4D5. The approach taken is to interface with the A/D and D/A converters at 36MHz and up convert the signals to 5.8GHz via an IF of 415MHz. For 10.5GHz a second IF at 1960MHz is used. The transmission and reception paths have a similar structure and the units operate in a TDD manner, switching between transmit and receive modes on a single channel. For ADAMAS, the base-station and subscriber units are identical. The prototype wireless subsystems are designed to be representative in terms of performance to commercial units but have not been engineered to the cost and size requirements of production units. 2 REUSE OF ADAMAS TECHNOLOGYDuring the ADAMAS project CRL looked for opportunities to use the ADAMAS technology in commercial projects. Generally these opportunities concern cases where there is a requirement to develop RF sections for another companies' product. CRL is not a manufacturer so we did not embark on developments without a commercial partner. Generally we did not achieve as much success as hoped due to the large number of existing manufacturers in the fixed wireless access market, which was generally not meeting market expectations. However we did win some contracts and developed a 5.8GHz OFDM radio module for a wireless news gathering product and more recently have developed a 3.6GHz/4.2GHz FDD outdoor transceiver for OFDM E1 telecoms links. The ADAMAS synthesiser board was adapted for the 3.6GHz/4.2GHz FDD transceiver mentioned above. We are currently developing a 5.8GHz OFDM transceiver for outdoor point to multi-point applications. It is intended to launch this product in early 2004 as part of Wavelength Solutions product line up. A lot of the radio circuitry is derived from ADAMAS although it is a new product with a different PCB architecture.

ADAMAS and STANDARDISATION BODIES. ADAMAS members monitored and contributed to the emergent standardisation bodies in both US and Europe (IEEE 802.16a and ETSI/BRAN HIPERMAN). In June 2000, INTRACOM and RUNCOM promoted the establishment of a new Working Group (WG) in Europe that targeted to study the functional requirements of the F-BWA systems below 11 GHz. This WG was the preparatory phase of the HIPERMAN standardisation body that emerged some months later. Towards this direction, both companies contributed and a subset of contributions´ list is given below: - A. Andritsou, P. Dallas, BRAN22-HMd082: Contribution to Functional Requirements for HIPERMAN, ETSI/BRAN#22, 27-30 Jan. 2001. - RUNCOM, BRAN29d072r1, OFDM sub-channelization proposal, ETSI/BRAN HIPERMAN, Sophia Antipolis, France, 2-5 July 2002. - RUNCOM, BRAN30d119, OFDM versus OFDMA comparison - comments, ETSI/BRAN HIPERMAN, Sophia Antipolis, 1-4 Oct. 2002. Additionally to that, INTRACOM and OTE were among the leaders for the creation of a new Working Item (W.I) to study the feasibility and sharing issues of license exempt F-BWA in 5 GHz,band C (5.725-5.875 MHz). This W.I. is still in progress and the final decision for the creation will be made within 2004. ADAMAS partners expanded the contribution activities also in US. A subset of the contributions is given below. - Kitrozer et al, Frame Duration for 802.16a, C802.16a-02/65, IEEE802.16a, Session#19, 20-24 May 2002, Calgary, Alberta, Canada. - Nico Van Waes (NOKIA), I. Kitrozer (RUNCOM), Editorial rewrite of P802.16a/D4, C802.16a-02/67, IEEE802.16a, Session#20, Vancouver, British Columbia, Canada, 8-12 July 2002. CONCLUSIONS Hence, ADAMAS project had a great impact on the standardisation bodies and participated actively in all the decisions of both sides of the Atlantic. The result of this effort is that many of the ADAMAS internal contributions have adopted by the standards and therefore, ADAMAS is very similar to the relative standards.

1. MMIC OBJECTIVES AND TECHNOLOGY ETH�s task within the ADAMAS project was to develop linear power amplifiers (PA) for both ADAMAS demonstrator systems. Due to implementation and performance constraints, monolithic microwave integrated circuit (MMIC) linear power amplifier were considered. Three different process technologies have been selected i.e. Triquint�s TQTRx GaAs MESFET process, Fraunhofer-Gesellschaft�s IAF V42 AlGaAs/InGaAs HEMT technology, and the BiCMOS 6HP process from IBM. After evaluating the characteristics of these processes, it was found that due to its high breakdown voltage, high quality passives and faire gain performance around 5.8 GHz, the TQTRx process should be well suited for the unlicensed system scenario units. Both IBM�s BiCMOS and the IAF V42 HEMT process from Fraunhofer Gesellschaft (FHG) appeared less attractive than the GaAs MESFET technology because the breakdown voltage performance combined with the quality of the available passive components makes it difficult to realise competitive medium power amplifiers in the lower microwave range despite of higher intrinsic-gain transistors. At higher frequencies, the competition between the two GaAs-based technologies gets tougher since gain of MESFET-based amplifiers tends to become low and in addition, the standard design-kit model accuracy starts to worsen considerably. Therefore, ETH decided to use both technologies in parallel for the unlicensed 10.5 GHz-band units. 2. RESULTS Two design notes and a conference paper, presented at the IEEE International Microwave and Optoelectronic Conference 2003 (IMOC03) document the progress made during the project period in the area of linear power amplifiers. A short summary of the most important results and findings will be given next. From the linearization method assessment made within the ADAMAS project, the pre-distortion in general and the RF pre-distortion in particular were considered best suited for the ADAMAS system user terminals of both frequency bands. The relative simplicity, the low impact on power or efficiency, and the small additional chip area required, if monolithically integrated, are the motivations for this choice. 2.a. Unlicensed 5.8 GHz system Within the three design runs performed, several power amplifiers targeting linear output power capabilities between 24 dBm and 32 dBm were designed, manufactured, and tested. After the first design cycle, measurements showed that grounding issues were the reasons for the lower than expected output power capability of the circuits. Since no substrate vias could be used to improve performance, the output power requirement of the following designs was lowered to 27 dBm. To counteract to a certain degree the power capability reduction, special effort was put in making the core amplifiers as linear as possible. Bias circuitry investigation and matching network analyses were performed to assure that both amplitude (AM-AM) and phase (AM-PM) characteristics of the MMIC PAs were optimised for linear transmission. Power amplifiers with 1 dB compression points of almost 29 dBm, power added efficiency (PAE) above 35% in class-A operation, and relative phase changes from small-signal operation up to the P1dB of less than 6 °C were achieved. In addition, different RF predistorters have been designed for monolithic integration with the power amplifier. Measured results showed a slight IMD improvement near the 1 dB compression point. 2.b. Licensed 10.5 GHz system Using Triquint�s TQTRx process four different MMIC power amplifiers have been produced. After the excellent experience from the second design run with the 10.5 GHz driver amplifier - showing more than 8 dB of small-signal gain, a PAE of 34 %, and a P1dB of more than 24 dBm - a two-stage version as well as the power-stage alone have been manufactured and tested for the X-band ADAMAS demonstrator. The performance of these chips was not as good as simulated due to parasitic underestimation. An additional design-cycle would be needed to correct for this. Designs carried out relying on FHG�s V42 HEMT process were not very successful. Problems with the breakdown voltage performance during manufacturing resulted in poor output power capability for the two-stage power amplifier module. The small-signal performance of the chips was nevertheless in good agreement with simulations. CONCLUSION The main achievement for ETH during the ADAMAS project period was the design of linear medium power amplifiers for the X-band frequency range by using a low-cost technology such as Triquint�s TQTRx GaAs MESFET process. There are many possibilities in how this knowledge can be of use in future projects, i.e. fully monolithic integration of the RF front-end for 10 GHz applications, design of wideband power amplifiers covering different communication standards, and so on.

Searching for OpenAIRE data...

Deliverables

Share this page

Download