MB-OFDM

download MB-OFDM

of 4

Transcript of MB-OFDM

  • 8/8/2019 MB-OFDM

    1/4

    46 www.rfdesign.com August 2005

    Consumer Electronics

    Low-power receiver architecture forinterference-robust UWB radio

    Ultra-wideband wireless communications promises unprecedented levels of seamlessconnectivity between consumer electronics devices, enabling gigabytes of data to be

    transmitted in seconds rather than hours without exhausting the batteries of hand-held

    portable devices. To meet the low-cost requirement of consumer product applications,

    the multiband-orthogonal frequency-division multiplexing (MB-OFDM) system has been

    designed to minimize transceiver architecture complexity.

    By Domine M.W. Leenaerts and Jozef R.M. Bergervoet

    Ultra-wideband (UWB) wireless commu-

    nication promises unprecedented levelsof seamless connectivity between consumerelectronics devices, enabling gigabytes of datato be transmitted in seconds rather than hourswithout exhausting the batteries of hand-heldportable devices such as digital cameras, au-dio jukeboxes and mobile phones. The FCChas opened up spectrum in the 3.1 GHz to10.6 GHz range for unlicensed UWB opera-tion, and the WiMedia Alliance (an amalgama-tion of the WiMedia and Multiband OFDMAlliance SIG groups) is proposing transmis-sion standards that will encourage worldwideacceptance of UWB by accommodating antici-pated differences in spectrum allocation fromregulatory bodies around the world.

    The alliances main standard, known asMB-OFDM (multiband-orthogonal frequency-division multiplexing) transmits data simul-taneously over multiple, accurately spaced,carrier frequencies and features high spectralflexibility plus resilience to RF interferenceand multipath effects. To meet the low-costrequirement of consumer product applications,the MB-OFDM system is designed to minimizetransceiver architecture complexity, requiringa single analog receive chain, relatively low-resolution ADCs and DACs and limited internalprecision in the digital baseband.

    It divides the allocated spectrum intoseveral QPSK-OFDM (quadrature phaseshift keying-orthogonal frequency-divisionmultiplexing) modulated sub-bands, spaced528 MHz apart, and employs a frequency-hopping scheme between these sub-bands toachieve efficient and robust communicationfor piconets operating simultaneously in closeproximity. Transmit power remains below theFCC limit of 41.25 dBm/MHz, and frequencyhops take place every 312.5 ns.

    Implementing the receiver chain is not,however, without challenges. The low RFsignal level requires the use of a low-noise

    receive chain, while the potential presence

    Figure 1. Philips UWB receive-chain architecture.

    Figure 2. Voltage and current feedback loops in the LNA correct linearity in the absence of

    balanced differential inputs.

    of strong out-of-band interferers in the

    2.4 GHz and 5 GHz bands (for example,

    Bluetooth and IEEE 802.11a/b/g links)

    requires a receiver with high linearity and

  • 8/8/2019 MB-OFDM

    2/4

    48 www.rfdesign.com August 2005

    selectivity. The transceiver must be able tohop between carrier frequencies within 9 ns,and the carriers generated by its frequencysynthesizer must be spectrally pure. Forexample, spurious tones in the 5 GHz rangemust be below 50 dBc to avoid downcon-

    version of strong out-of-band interferers intowanted signal bands.

    The receiver must be able to discrimi-nate the wanted UWB signal, which may be as weak as 70 dBm at a distance of10 meters from the transmitter, in the pres-ence of a nearby interferer such as an802.11a transmitter from which the receivedsignal level is around +23 dBm and whosefrequency is only 500 MHz away fromthe UWB carrier. This poses challenges inthe design of the receiver front-end over avery wide bandwidth with respect to noiseand distortion. Furthermore, to limit the

    dynamic range in the subsequent ADCs, theinterferer has to be filtered to a level well below the wanted signal, requiring a wide- band IF filter with high attenuation and anaccurate and steep roll-off.

    RF architectureTo meet these requirements, a receive-chain

    architecture as illustrated in Figure 1 wasdeveloped. It was test-chip implemented in anew QUBiC4G silicon-germanium BiCMOSprocess technology. QUBiC4G realizes thehigh transistor f

    Tfigures required for opera-

    tion in the 3.1 GHz to 10.6 GHz band and

    also features the passive component integra-tion capabilities required to create a highlyintegrated solution.

    The radio is based on a zero-IF architec-ture with the frequency hopping realized in amultitone frequency synthesizer. An off-chippre-filter reduces strong out-of-band interfer-ers. The single-ended low-noise amplifier(LNA) input avoids the added complexity ofan external wideband balun, thereby minimiz-ing RF signal loss and system cost. Generationof true differential signals is left until theGilbert mixer stage, where it is achievedthrough the use of an active balun under-stage.The Gilbert mixer generates quadrature (I andQ) outputs, which are filtered and amplifiedby the IF filter amplifier, after which they aredigitized in separate ADCs.

    Although the single-ended LNA configu-ration significantly reduces system cost (noexternal wideband balun required), it slightlyworsens input linearity. Linearity is correctedby applying voltage and current feedback tothe LNA, as shown in Figure 2. This feedbackmatches the input impedance to 50 over thelower three MB-OFDM sub-bands in whichthis receiver chain is designed to operate (3432MHz, 3960 MHz, and 4488 MHz), without theneed for external matching components.

    The I and Q local oscillator inputs for the

    Figure 3. Block diagram of the multitone frequency synthesizer.

    Figure 4. IF filter architecture.

    Figure 5. LNA frequency response curves in dB for gain (S21), noise figure (NF), return loss mea-sured on pcb (S11_PCB) and reverse isolation (S12).

    Gilbert mixer are generated by a multitonefrequency synthesizer, which has also beentest-chip fabricated in Philips QUBiC4Gsilicon germanium (SiGe) process technology.There are a number of ways of generatingthe three local oscillator frequencies for thesub-bands, obvious options being the use of

    a single fast-settling phase-locked loop (PLL)

    or three separate PLLs feeding a three-wayRF switch. However, the first of these wouldrequire an unrealistically high reference fre-quency to achieve the required settling speed,while the second solution would be sensitiveto inductive coupling between the three PLLsand to carrier leakage between them.

    The chosen architecture for the frequency

  • 8/8/2019 MB-OFDM

    3/4

    50 www.rfdesign.com August 2005

    synthesizer uses two quadrature I and Q PLLs,one generating a local oscillator frequencyfor the middle sub-band (3960 MHz) whilethe other generates a frequency equal to thesub-band spacing (528 MHz). The upper and

    lower sub-band local oscillator frequencies arethen generated by mixing the outputs of thesetwo PLLs in a single-sideband (SSB) mixer,inverting the sign of the 528 MHz frequencyto generate the upper and lower sub-band

    carriers (3960 MHz + 528 MHz = 4488 MHzand 3960 MHz 528 MHz = 3432 MHz). The3960 MHz signal is simply shifted by 0 Hzto generate the local oscillator frequency forthe middle sub-band. Using this architecture,the achievable frequency hopping speed is

    well below 1 ns.A block diagram of this multitone generator

    is shown in Figure 3. The filter is needed tosuppress the third harmonic that is generated inthe 528 MHz PLLs frequency divider, whichcould otherwise result in in-band spurioussignals being generated after downconversionin the presence of a strong interferer suchas an 802.11a signal. Both local oscillatorsuse inductors to minimize phase noise andpower dissipation and are tuned by digitallycontrolled MOS capacitors, all of these com- ponents being integrated on-chip using thepassive integration capabilities of QUBiC4G

    process technology.After mixing the RF input with the local

    oscillator signals in the Gilbert mixer, thequadrature IF frequency outputs are filteredin a fully differential low-pass Chebyshevactive filter with a nominal gain of 45 dBand a passband ripple of 2.8 dB. This filtercomprises multiple gain stages, filter stagesand passive switched attenuators (Figure 4)that are distributed in a way that makes theIF chain highly linear, both when the wantedsignal is weak and strong interferers arepresent, and when the wanted signal level ishigh. The filter also provides a high level of

    rejection to large out-of-band interferers,which receive only limited attenuation aheadof the LNA. The better the IF filtering, thecloser the receiver can co-exist with localinterferers such as 802.11 wireless networks.One of the most severe situations is an 802.11ainterferer at 5.15 GHz, which is only 660 MHzaway from the carrier frequency of the 4488-MHz UWB sub-band.

    To achieve the transition-band and stop-bandaccuracy, the position and quality factor of thefilter poles is digitally calibrated by switched-capacitor arrays realized using high capacitancedensity (5 fF/m2) metal-insulator-metal(MIM) capacitors that are integrated on-chip aspart of the QUBiC4G BiCMOS process tech-nology. These capacitors are used to calibrateout process-induced variations in performance.The nominal tuning range of the band edge is25 MHz in 16 steps. The high f

    Tfigures (>70

    GHz) of the transistors in the QUBiC4G pro-cess provide the filters two-stage differentialamplifiers with the high unity-gain frequencyneeded to maintain low distortion levels.The attenuator is a digitally switched passiveresistor network and achieves an attenuationrange of 0 dB to 30 dB in steps of 6 dB.

    Performance resultsMeasurement results for the naked die

    Figure 6. Transfer characteristics of the IF filter.

    Circle 34 or visit freeproductinfo.net/rfd

    Gain(dB)

    Gain(dB)

    60

    40

    20

    0

    -20

    -40

    -60

    -80

    46

    44

    42

    40

    38Gain(dB)

    45

    40

    35

    30

    5.3dB 25MHz

    660MHz, -12.4dB

    Frequency

    Frequency (MHz)Frequency (MHz)

    Bandwidth TuningQ Tuning

    57.9dBr

    0 100 200 300

    -2000 -1500 -1000 -500 0 500 1000 1500 2000

    0 100 200

    Attenuation steps

    [0, 6, 12, 18, 24, 30 dB]

    simulated transfer

    measured transfer

  • 8/8/2019 MB-OFDM

    4/4

    RF Design www.rfdesign.com 51

    Figure 7. Micrographs of the multitone synthesizer (left) and LNA/mixer/IF filter (right) clearly show thepassive component integration capabilities of Philips QUBiC4G RF BiCMOS process technology.

    LNA are shown in Figure 5. Excellent gainand noise figures are maintained up to around13 GHz, making this receive-chain designsuitable for the full 3.1 GHz to 10.6 GHzUWB frequency range. Input impedancematching was measured with the die mounted

    in a HVQFN package assembled onto an FR4laminate printed circuit board, because boththe device package and the board mountingaffect the impedance match. These test results,therefore, represent typical application condi-tions. Figure 6 shows the measured magnitudetransfer characteristics of the IF filter fordifferent gain and bandwidth settings. The gaincan be varied between 16 dB and 46 dB in6 dB steps, and the bandwidth can be tuned inthe range 232 MHz to 254 MHz. At 660 MHzoffset a 57 dBr was achieved.

    Phase noise tests on the multitone frequencysynthesizer operating in the 4488 MHz UWB

    sub-band show in-band phase noise to be below90 dBc/Hz and out-of-band phase noise to bebelow 100 dBc/Hz at 1 MHz offset. The sig-nal-to-noise ratio over a single UWB channel is,therefore, degraded by less than 0.1 dB as a resultof phase noise in the multitone frequency syn-thesizer. When operating in the same sub-band,spurious tones in the 5 GHz band are below 50dBc. With a realistic antenna filter, a worst-case+23 dBm 802.11a interferer at 20 cm distancewill be attenuated by 20 dB, sufficient to avoid asevere interference problem. A similar reason-ing holds for the 2.4 GHz ISM band where therequirement is set to 45 dBc. The measured

    spurious tone at 2.37 GHz is attenuated by46 dBc and sufficiently suppressed to allowcoexistence between UWB and 2.4 GHz ISMusers.

    The complete prototyped receiver drawsless than 80 mA from a 2.7 V supply (ap-proximately 47 mA for the receive-chain and27 mA for the multitone frequency synthe-sizer). Micrographs of the two test-chips areshown in Figure 7.

    When developing a receive-chain archi-tecture for UWB operation in the 3.1 GHz to10.6 GHz bands, a designer must meet certainrequirements. Passive component integrationcapabilities are required to create a highlyintegrated solution. RFD

    ABOUT THE AUTHORSDomine M.W. Leenaerts is a principal

    scientist at Philips Research in Eindhoven,the Netherlands, where he is involved inRF integrated transceiver design. He holdsan M.S. and Ph.D. in electrical engineer-ing from the Eindhoven University ofTechnology. He can be reached at domine.leenaerts@philips.

    Jozef R.M. Bergervoet is a principalscientist at Philips Research in Eindhoven,the Netherlands, where he focuses on thedesign of integrated transceivers. He studiedphysics, graduated cum laude and received aPh.D. in theoretical physics from NijmegenUniversity, the Netherlands.

    Circle 35 or visit freeproductinfo.net/rfd