Multiport Technology: New Perspectives and …downloads.hindawi.com/journals/specialissues/961060.pdfJournal of Electrical and Computer Engineering Multiport Technology: New Perspectives

Journal of Electrical and Computer Engineering

Multiport Technology: New Perspectives and Applications

Guest Editors: Serioja Ovidiu Tatu, Adriana Serban, Alexander Koelpin, and Mohamed Helaoui

Multiport Technology: New Perspectives andApplications

Journal of Electrical and Computer Engineering

Multiport Technology: New Perspectives andApplications

Guest Editors: Serioja Ovidiu Tatu, Adriana Serban,Alexander Koelpin, and Mohamed Helaoui

Copyright © 2014 Hindawi Publishing Corporation. All rights reserved.

This is a special issue published in “Journal of Electrical and Computer Engineering.” All articles are open access articles distributed underthe Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, providedthe original work is properly cited.

Editorial BoardThe editorial board of the journal is organized into sections that correspond tothe subject areas covered by the journal.

Circuits and Systems

M. T. Abuelma’atti, Saudi ArabiaIshfaq Ahmad, USADhamin Al-Khalili, CanadaWael M. Badawy, CanadaIvo Barbi, BrazilMartin A. Brooke, USAY. W. Chang, TaiwanTian-Sheuan Chang, TaiwanChip Hong Chang, SingaporeTzi-Dar Chiueh, TaiwanM. Jamal Deen, CanadaA. El Wakil, UAEDenis Flandre, BelgiumP. Franzon, USAAndre Ivanov, CanadaEbroul Izquierdo, UKWen-Ben Jone, USAYong-Bin Kim, USA

H. Kuntman, TurkeyParag K. Lala, USAShen-Iuan Liu, TaiwanBin-Da Liu, TaiwanJoafko Antonio Martino, BrazilPianki Mazumder, USAMichel Nakhla, CanadaSing Kiong Nguang, New ZealandShun-ichiro Ohmi, JapanMohamed A. Osman, USAPing Feng Pai, TaiwanMarcelo Antonio Pavanello, BrazilMarco Platzner, GermanyMassimo Poncino, ItalyDhiraj K. Pradhan, UKF. Ren, USAGabriel Robins, USAMohamad Sawan, Canada

Raj Senani, IndiaGianluca Setti, ItalyJose Silva-Martinez, USANicolas Sklavos, GreeceAhmed M. Soliman, EgyptDimitrios Soudris, GreeceCharles E. Stroud, USAEphraim Suhir, USAHannu Tenhunen, SwedenGeorge S. Tombras, GreeceSpyros Tragoudas, USAChi Kong Tse, Hong KongChi-Ying Tsui, Hong KongJan Van der Spiegel, USAChin-Long Wey, USAFei Yuan, Canada

Communications

Sofiène Affes, CanadaDharma Agrawal, USAH. Arslan, USAEdward Au, ChinaEnzo Baccarelli, ItalyStefano Basagni, USAJun Bi, ChinaZ. Chen, SingaporeRené Cumplido, MexicoLuca De Nardis, ItalyM.-Gabriella Di Benedetto, ItalyJ. Fiorina, FranceLijia Ge, ChinaZ. Ghassemlooy, UKK. Giridhar, India

Amoakoh Gyasi-Agyei, GhanaYaohui Jin, ChinaMandeep Jit Singh, MalaysiaPeter Jung, GermanyAdnan Kavak, TurkeyRajesh Khanna, IndiaKiseon Kim, Republic of KoreaD. I. Laurenson, UKTho Le-Ngoc, CanadaC. Leung, CanadaPetri Mähönen, GermanyMohammad A. Matin, BangladeshM. Nájar, SpainM. S. Obaidat, USAAdam Panagos, USA

Samuel Pierre, CanadaNikos C. Sagias, GreeceJohn N. Sahalos, GreeceChristian Schlegel, CanadaVinod Sharma, IndiaIickho Song, KoreaIoannis Tomkos, GreeceChien Cheng Tseng, TaiwanGeorge Tsoulos, GreeceLaura Vanzago, ItalyRoberto Verdone, ItalyGuosen Yue, USAJian-Kang Zhang, Canada

Signal Processing

S. S. Agaian, USAPanajotis Agathoklis, CanadaJaakko Astola, FinlandTamal Bose, USAA. G. Constantinides, UKPaul Dan Cristea, Romania

Petar M. Djuric, USAIgor Djurović, MontenegroKaren Egiazarian, FinlandW.-S. Gan, SingaporeZ. F. Ghassemlooy, UKLing Guan, Canada

Martin Haardt, GermanyPeter Handel, SwedenAndreas Jakobsson, SwedenJiri Jan, Czech RepublicS. Jensen, DenmarkChi Chung Ko, Singapore

M. A. Lagunas, SpainJ. Lam, Hong KongD. I. Laurenson, UKRiccardo Leonardi, ItalyS. Marshall, UKAntonio Napolitano, ItalySven Nordholm, AustraliaS. Panchanathan, USAPeriasamy K. Rajan, USA

Cédric Richard, FranceW. Sandham, UKRavi Sankar, USADan Schonfeld, USALing Shao, UKJohn J. Shynk, USAAndreas Spanias, USAYannis Stylianou, GreeceIoan Tabus, Finland

Jarmo Henrik Takala, FinlandClark N. Taylor, USAA. H. Tewfik, USAJitendra Kumar Tugnait, USAVesa Valimaki, FinlandLuc Vandendorpe, BelgiumAri J. Visa, FinlandJar Ferr Yang, Taiwan

Contents

Multiport Technology: New Perspectives and Applications, Serioja Ovidiu Tatu, Adriana Serban,Alexander Koelpin, and Mohamed HelaouiVolume 2014, Article ID 194649, 4 pages

Performance Driven Six-Port Receiver and Its Advantages over Low-IF Receiver Architecture,Abul Hasan and Mohamed HelaouiVolume 2014, Article ID 198120, 8 pages

Performance of 2–3.6GHz Five-Port/Three-Phase Demodulators with Baseband Analog 𝐼/𝑄Regeneration Circuit in Direct-Conversion Receivers, Kaissoine Abdou, Antoine Khy, Kais Mabrouk,and Bernard HuyartVolume 2014, Article ID 631867, 11 pages

Six-Port-Based Architecture for Phase Noise Measurement in the UWB Band, J. M. Ávila-Ruiz,A. Moscoso-Mártir, E. Durán-Valdeiglesias, L. Moreno-Pozas, J. de-Oliva-Rubio, and I. Molina-FernándezVolume 2014, Article ID 646738, 10 pages

A Compact, Versatile Six-Port Radar Module for Industrial and Medical Applications, Sarah Linz,Gabor Vinci, Sebastian Mann, Stefan Lindner, Francesco Barbon, R. Weigel, and Alexander KoelpinVolume 2013, Article ID 382913, 10 pages

Complete Characterization of Novel MHMICs for V-Band Communication Systems, C. Hannachi,D. Hammou, T. Djerafi, Z. Ouardirhi, and S. O. TatuVolume 2013, Article ID 686708, 9 pages

EditorialMultiport Technology: New Perspectives and Applications

Serioja Ovidiu Tatu,1 Adriana Serban,2 Alexander Koelpin,3 and Mohamed Helaoui4

1 Institut National de la Recherche Scientifique-Energie, Matériaux et Télécommunications,800 rue de la Gauchetière Ouest, Montréal, Québec, Canada

2Department of Science and Technology, Linköping University, Bredgatan 34, SE-601, 74 Norrkoping, Sweden3 Institute for Electronics Engineering, University of Erlangen-Nuremberg, Cauerstr. 9, 91058 Erlangen, Germany4 Intelligent RF Radio Technology Laboratory (iRadio Lab), Department of Electrical and Computer Engineering,Schulich School of Engineering, University of Calgary, Calgary, AB, 91058 Erlangen, Canada T2N 1N4

Correspondence should be addressed to Serioja Ovidiu Tatu; [email protected]

Received 13 April 2014; Accepted 13 April 2014; Published 7 July 2014

Copyright © 2014 Serioja Ovidiu Tatu et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

1. Introduction

The multiport circuit theory was initially developed in the1970s by scientists for accurate automatized measurementsof the complex reflection coefficients, in microwave networkanalysis [1–3]. These multiport pioneers highlighted its use-fulness in microwave low-cost circuit characterizations (S-parameters).

Since 1994, the multiport techniques were further devel-oped for microwave and millimeter-wave radios [4–6]. Untiltoday, several multiport architectures for specific applica-tions, such as communication transceivers [7–17], radarsensing [18–23], direction of arrival estimation [24–26], orphase noise measurements [27], have been developed andimplemented.

Basically, the multiport is a passive circuit composed ofseveral couplers interconnected by transmission lines andphase shifters. Its specific architecture and design are stronglyrelated to the target application and the operating fre-quency. The multiport acts as an interferometer; its outputsignals are linear combinations of phase shifted input signals.By using the appropriate circuit design and appropriate devi-ces connected to the output ports, this circuit can pro-vide specific parameters, such as reflection coefficient, dis-tance or modal measurements, phase and frequency anal-ysis, quadrature down-conversion, or direct modulation ofmicrowave/millimeter-wave frequencies.

As originally designed for automated measurements ofthe complex reflection coefficient, the multiport has a local

oscillator input, a measurement port, and four outputs [3].One of the outputs is used as a reference power level andpowers measured at the other ones are function of thecomplex coefficient of the device under test connected tothe measurement port. There are three different reflectioncoefficient values named 𝑞

𝑖points, whichminimize the power

at the corresponding 𝑖 output.The ideal architecture requeststhat 𝑞

𝑖points are to be spaced by 120∘ and located equidistant

from the origin of the complex plane.The new application fields require a different architecture

of the circuit and specificmodules to be connected at its ports.The S-parameter matrix of the multiport circuit reveals thatthere are two clusters of ports, 1 to 4 and 5 and 6. Inside eachcluster, all the ports are perfectly matched and isolated, oneversus the others [15]. In all applications they play separatefunctions, such as four outputs and, respectively, two RFinputs for down-conversion or four control inputs and RFoutput/input for direct modulators. If the S matrix is furtheranalyzed, then it is straightforward that if four matched loadsare connected to the first group of ports (1 to 4) and twoRF signals are applied to other pair of ports (5 and 6), alloutput signals at first group of ports are function of both inputsignals of the second group.This is a fundamental difference,if compared to the multiport used in reflection coefficientmeasurements, where one of the outputs is used as a powerreference [3]. The multiport has now four 𝑞

𝑖points spaced by

90∘ multiples and located equidistant from the origin of thecomplex plane. The phase difference between the pair of odd

Hindawi Publishing CorporationJournal of Electrical and Computer EngineeringVolume 2014, Article ID 194649, 4 pageshttp://dx.doi.org/10.1155/2014/194649

http://dx.doi.org/10.1155/2014/194649

2 Journal of Electrical and Computer Engineering

𝑞𝑖points is 180∘. The same result is obtained for the pair of

even points [17].The use of multiport technology in RF design is a good

choice, especially if the operating frequency is in the highmicrowave or millimeter-wave range. The dimensions of themultiport circuit fabricated in miniature hybrid microwaveintegrated circuit (MHMIC) technology, usually around1.5𝜆𝑔× 1.5𝜆

𝑔, where 𝜆

𝑔represents the guided wavelength,

become small enough to be integrated on the same substratewith antennas [20]. Even if the multiport is further minia-turized, the antenna or array antenna size will determinatethe final dimensions of a front end. The multiport circuitcan be also used in the front-end design to operate at thefrequencies where active components are not yet available inthemarket. In order to operate as demodulator or modulator,it requires only the use of power detectors or switches[28–32]. Therefore, research activities can be validated byfront-end prototyping measurements, years before standardtechnologies become available.

This special issue highlights, through several examplesand multiple references, some of the modern applicationsof the multiport technology and significant advances infabrication procedures, in the recent years.

2. Special Issue Papers

This special issue dedicated to multiport technology hostsseveral papers covering last advances in six-port receivers,demodulators, radar sensing, and ultrawide band (UWB)phase noise measurements.

An interesting question in multiport technology is howmany ports should be used to fulfill a given system speci-fication. In their paper entitled “Performance of 2–3.6 GHzfive-port/three-phase demodulators with baseband analog 𝐼/𝑄regeneration circuit in direct-conversion receivers,” K. Abdouet al. compare the performance of a five-port (FPD), athree-phase (TPD), and a quadrature demodulator. First,the authors describe the basic principles of FPD and TPD.Unlike the FPD, that uses detectors for the down-conversion,the TPD multiplies the radio frequency (RF) and the localoscillator signal with the help of mixers. A baseband circuitfor the analog 𝐼/𝑄 regeneration is designed to reduce thenumber of analog-to-digital converters from three to twoand allows suppression of DC offset and second orderintermodulation distortion (IMD2).

Finally, the implementation of all architectures is demon-strated; furthermore, detailed measurement results are pre-sented. These results indicate that TPD outperforms FPDin terms of residual DC offset, IMD2, noise figure, andsensitivity as well as error vector magnitude. Moreover, TPDhas advantages over quadrature demodulators in the case ofRF carrier aggregation. Thus, the authors conclude that thepresented TPD architecture is a potential candidate for futurelong term evolution (LTE-A) standard.

The use of the 60-GHz band has attracted a great dealof interest over the last few decade, especially for its use infuture compact transceivers dedicated to high-speed wirelessapplications in indoor environments (57–64GHz).

In this context, C.Hannachi et al., the authors of the paper“Complete characterization of novel MHMICs for V-bandcommunication systems” show the characterization results ofseveral new passive millimeter-wave circuits integrated onvery thin ceramic substrate (125microns, relative permittivityequal to 9.9). These components are designed for a completeintegrated V-band six-port receiver front end, hosting anantenna array, low-noise amplifier, and quadrature demod-ulator.

The work is focused on the design and characterizationof a novel rounded Wilkinson power divider, a 90∘ hybridcoupler, a rat-race coupler, and a novel six-port circuit. Theauthors describe in detail the equipment, the calibrationtechnique, and the multiple challenges in millimeter-wavemeasurements. A series of microphotographs show a typicalfabricated ceramic die and details of all circuits, as preparedfor probe station testing.

Measurement results prove that the proposed circuitsare UWB components. The supplementary insertion losses,amplitude, and phase unbalancements are considered morethan acceptable to build modulators/demodulators for mod-ulation schemes having up to 16 symbols (BPSK to 16 QAM,PSK, or dual star). Keeping in account the 7GHz bandwidthallowed for V-band communication systems, the data ratescan reach quasioptical values.

Apart from its ease of use, simplicity of implementation,and low power requirements, which are due to the passivecircuit implementation nature of six-port based receiver,other benefits drove interests in using six-port technologyin receivers’ design for applications using lower frequenciesthanmillimetre wave. Indeed, the wide frequency bandwidthcoverage, configurability, and low cost are some of theadditional features that make this architecture very attractivefor the implementation of the software defined radio (SDR)concept in the wireless communication frequency bands.The utilization of six-port concept as a RF front end incommunication receivers was investigated by A. Hasan andM. Helaoui; in their paper “Performance driven six-portreceiver and its advantages over low-IF receiver architec-ture” the authors provide a comprehensive analysis of theperformance of six-port-based receiver in terms of signalquality, dynamic range, noise figure, isolation, bandwidth,and cost. They compare these metrics to a conventionallow-IF architecture performance. On one side, the wide RFfrequency bandwidth available in six-port circuitry is a majoradvantage that allows for use in ultrawideband andmultibandconfiguration, reconfigurability, and, therefore, suitability forSDR applications. On the other side, signal quality measuredin terms of receiver noise figure, sensitivity, and dynamicrange is the main limitation in such topology. The maincomponent limiting the signal quality is the dynamic rangeof power detectors. By driving the power detectors beyondtheir square law region, the dynamic range can be extended ata cost of linearity degradation. The use of postcompensationand calibration technique can compensate this degradation,which results in configurable receivers with good linearityand dynamic range performance. A. Hasan and M. Helaouishowed that by using such postcompensation techniques,error vectormagnitudes (EVMs) less than 1% can be obtained

Journal of Electrical and Computer Engineering 3

for complex modulated signals suggesting that such topologycan be a very good candidate for an SDR receiver front end.

Besides the broad range of applications already intro-duced in this paper, another very promising topic is rangingand direction finding. In particular for industry grade dis-tance measurements, the six-port operated as a microwaveinterferometer shows extraordinary performance, comparedto common radar architectures.

In the paper entitled “A compact, versatile six-port radarmodule for industrial and medical applications,” S. Linz et al.demonstrate that the six-port receiver used as continuouswave radar is suitable for a variety of application scenar-ios ranging from distance and vibration measurements inindustrial environments to heartbeat detection for medicalapplications.

The authors present a compact 24GHz six-port radarsensor with reliable high-integration of all subcomponentsin one single module. This sensor has been developed incooperation with industry. After a general system overviewand the theoretical description of the implemented six-portnetwork, a short introduction to the digital signal processingstrategy is given. Simulation results of the subcomponents arefurthermore validated by measurements. The demonstratedresults of the system’s performance show a maximum relativeerror of ±400𝜇mand high precision in themicrometer range(±40 𝜇m).

J. M. Ávila-Ruiz et al. propose in their paper new six-port for low-power oscillator phase noise measurements.This paper entitled “Six-port based architecture for phasenoise measurement in the UWB band” demonstrates, onone side, the advances of the multiport technology thatcontinues to be used as a reliable, accurate, low-power coreof automated measurement setups. On the other side, itreflects the advances in development of the analytical theory,supporting and explaining the multiport technology and itslarge field of applications.

The presented phase noise measurement system com-bines two phase noise measurement approaches, that is, thedelay line discriminator method and the phase noise meas-urement method based on the quadrature 𝐼/𝑄 six-portdemodulator. The proposed solution takes advantage of theability of the multiport demodulator to accurately handlelow-power input signals and to detect useful signal mag-nitude and phase information. In addition, the delay linediscriminator eliminates the need for expensive referenceoscillator, as opposed to traditional phase noisemeasurementsystems.

After briefly introducing the state-of-the-art phase noisemeasurements, the authors provide a detailed theoreticalanalysis that demonstrates the complex processing of sig-nals performed by the multiport system. It is shown thatthe required phase noise information is embedded in thequadrature 𝐼/𝑄 components of the down-converted signal.To assess the intrinsic limitations of the system, internalnoise sources are analysed and solutions are proposed toincrease measurement sensitivity. Finally, phase noise mea-surements performed with the prototype are compared tothose obtained with a commercial signal source analyzer.

The results presented by J. M. Ávila-Ruiz et al. reconfirmthe multiport technology in its original application proposedin the 1970s as an accurate, flexiblemeasurement set-up basedon interferometric signal processing.

3. Conclusion

In the last 40 years, since the multiport (six-port) circuitis used, dozens of architectures have been proposed. Theoperating frequencies, fabrication technologies, and applica-tions do not cease to evolve. Through several well-chosenexamples, the special issue highlights some of the multiporttechnology modern applications.

These papers and related references demonstrate thatmultiport technology is a good competitor of the conven-tional approaches, especially in the microwave and milli-meter-wave frequencies, where circuit dimensions, closelyrelated to the guided wavelength, are in the cm range orsmaller. Broadband circuits can be designed to cover awide frequency spectrum and to allow quasioptical datarates for wireless communication. Impressive resolution andhigh accuracy can be obtained in range or angle-of-arrivalmeasurements. Calibration and impairment mitigation tech-niques can also be used to improve demodulation results,especially in low microwave spectrum, where analog todigital convertors (ADC) and digital signal processing (DSP)techniques are available today.

Future work targets, among others, improvements inmultiport circuit design for each specific application, theincreasing of the operating frequencies, and, why not, the useof multiport systems in THz frequency range.

Serioja Ovidiu TatuAdriana Serban

Alexander KoelpinMohamed Helaoui

References

[1] S. B. Cohn and N. P. Weinhouse, “An automatic microwavephase measurement system,” Microwave Journal, vol. 7, no. 2,pp. 49–56, 1964.

[2] G. F. Engen and C. A. Hoer, “Application of an arbitrary 6-portjunction to power-measurement problems,” IEEE Transactionson Instrumentation and Measurement, vol. 21, no. 4, pp. 470–474, 1972.

[3] G. F. Engen, “The six-port reflectometer: an alternative networkanalyzer,” IEEE Transactions on Microwave Theory and Tech-niques, vol. 25, no. 12, pp. 1075–1080, 1977.

[4] T. Hentschel, “The six-port as a communications receiver,” IEEETransactions on Microwave Theory and Techniques, vol. 53, no.3, pp. 1039–1047, 2005.

[5] A. Koelpin, G. Vinci, B. Laemmle, D. Kissinger, and R. Weigel,“The six-port in modern society,” IEEE Microwave Magazine,vol. 11, no. 7, pp. 35–43, 2010.

[6] G. Vinci, S. Lindner, F. Barbon, R. Weigel, and A. Koelpin,“Promise of a better position,” IEEE Microwave Magazine, vol.13, no. 7, pp. S41–S49, 2012.

[7] J. Li, R. G. Bosisio, and K. Wu, “Computer and measurementsimulation of a new digital receiver operating directly at


millimeter-wave frequencies,” IEEE Transactions on MicrowaveTheory and Techniques, vol. 43, no. 12, pp. 2766–2772, 1995.

[8] S. O. Tatu, E. Moldovan, K. Wu, and R. G. Bosisio, “A newdirect millimeter-wave six-port receiver,” IEEE Transactions onMicrowaveTheory and Techniques, vol. 49, no. 12, pp. 2517–2522,2001.

[9] S. O. Tatu, E. Moldovan, K. Wu, R. G. Bosisio, and T. A.Denidni, “Ka-band analog front-end for software-defined directconversion receiver,” IEEE Transactions on Microwave Theoryand Techniques, vol. 53, no. 9, pp. 2768–2776, 2005.

[10] Y. Zhao, J.-F. Frigon, K. Wu, and R. G. Bosisio, “Multi(six)-port impulse radio for ultra-wideband,” IEEE Transactions onMicrowave Theory and Techniques, vol. 54, no. 4, pp. 1707–1711,2006.

[11] S. O. Tatu and E. Moldovan, “V-band multiport heterodynereceiver for high-speed communication systems,” EURASIPJournal on Wireless Communications and Networking, vol. 2007,Article ID 34358, 7 pages, 2007.

[12] R. G. Bosisio, Y. Y. Zhao, X. Y. Xu et al., “New wave radio,” IEEEMicrowave Magazine, vol. 9, no. 1, pp. 89–100, 2008.

[13] A. Serban, J. Östh, O. M. Karlsson, S. Gong, J. Haartsen, andP. Karlsson, “Six-port transceiver for 6-9GHz ultrawidebandsystems,”Microwave and Optical Technology Letters, vol. 52, no.3, pp. 740–746, 2010.

[14] C. de la Morena-Álvarez-Palencia and M. Burgos-Garcia,“Four-octave six-port receiver and its calibration for broadbandcommunications and software defined radios,” Progress inElectromagnetics Research, vol. 116, pp. 1–21, 2011.

[15] S. O. Tatu, E. Moldovan, and S. Affes, “Multi-port front-endand transceivers for V-band multi-gigabit/s communicationsystems,” in Digital Front-End in Wireless Communicationsand Broadcasting Circuits and Signal Processing, pp. 707–732,Cambridge University Press, Cambridge, UK, 2011.

[16] A. Hasan and M. Helaoui, “Novel modeling and calibrationapproach for multi-port receivers mitigating system imper-fections and hardware impairments,” IEEE Transactions onMicrowaveTheory and Techniques, vol. 60, no. 8, pp. 2644–2653,2012.

[17] T. Jiang, D. Hammou, C. Hannachi et al., “Six-port technologyformillimeter-waveMIMO systems,” in Proceedings of the Euro-pean Radar Conference (EuRAD ’13), pp. 279–282, Nuremberg,Germany, October 2013.

[18] E. Moldovan, S.-O. Tatu, T. Gaman, K. Wu, and R. G. Bosisio,“Anew94-GHz six-port collision-avoidance radar sensor,” IEEETransactions on Microwave Theory and Techniques, vol. 52, no.3, pp. 751–759, 2004.

[19] B. Boukari, E. Moldovan, R. I. Cojocaru et al., “A 77-GHz six-port fmcw collision avoidance radar sensor with baseband ana-lytical calibration,” Microwave and Optical Technology Letters,vol. 51, no. 3, pp. 720–725, 2009.

[20] S. O. Tatu, B. Boukari, E. Moldovan, R. G. Bosisio, and K.Wu, “Millimeter-wave multi-port radar sensor with integratedreceiver front-end for automotive applications,” in Proceedingsof the IEEE MTT-S International Microwave Symposium Digest(MTT ’12), pp. 1–3, Montreal, Canada, June 2012.

[21] F. Barbon, G. Vinci, S. Lindner, R. Weigel, and A. Koelpin, “Asix-port interferometer based micrometer-accuracy displace-ment and vibration measurement radar,” in Proceedings of theIEEE MTT-S International Microwave Symposium Digest (MTT’12), pp. 1–3, Montreal, Canada, June 2012.

[22] G. Vinci, S. Lindner, F. Barbon et al., “Six-port radar sensor forremote respiration rate and heartbeat vital-sign monitoring,”

IEEE Transactions onMicrowaveTheory and Techniques, vol. 61,no. 5, pp. 2093–2100, 2013.

[23] G. Vinci, S. Lindner, F. Barbon, S. Mann, R. Weigel, and A.Koelpin, “Six-Portmicrowave interferometer radar formechan-ical vibration analysis,” in Proceedings of the European RadarConference (EuRAD ’13), pp. 287–290, Nuremberg, Germany,October 2013.

[24] S. O. Tatu, K. Wu, and T. A. Denidni, “Direction-of-arrival esti-mation method based on six-port technology,” IEE Proceedings:Microwaves, Antennas and Propagation, vol. 153, no. 3, pp. 263–269, 2006.

[25] G. Vinci, A. Koelpin, F. Barbon, and R. Weigel, “Six-port-based direction-of-arrival detection system,” in Proceedings ofthe Asia-Pacific Microwave Conference (APMC ’10), pp. 1817–1820, December 2010.

[26] G. Vinci, F. Barbon, R. Weigel, and A. Koelpin, “A novel,wide angle, high resolution direction-of-arrival detector,” inProceedings of the 8th European Radar Conference (EuRAD ’11),pp. 265–268, October 2011.

[27] S. O. Tatu, E. Moldovan, S. Affes, B. Boukari, K. Wu, andR. G. Bosisio, “Six-port interferometric technique for accurate𝑊-band phase-noise measurements,” IEEE Transactions onMicrowave Theory and Techniques, vol. 56, no. 6, pp. 1372–1379,2008.

[28] J. Östh, O. Owais, M. Karlsson, A. Serban, and S. Gong, “Schot-tky diode as high-speed variable impedance load in six-portmodulators,” inProceedings of the IEEE International Conferenceon Ultra-Wideband (ICUWB ’11), pp. 68–71, Bologna, Italy,September 2011.

[29] J. Östh, A. Serban, M. Karlsson, and S. Gong, “LO leakagein six-port modulators and demodulators and its suppressiontechniques,” in Proceedings of the IEEE MTT-S InternationalMicrowave Symposium Digest (MTT ’12), pp. 1–3, Montreal,Canada, June 2012.

[30] A. Serban, M. Karlsson, J. Östh, Owais, and S. Gong, “Differen-tial circuit technique for six-port modulator and demodulator,”in Proceedings of the IEEE/MTT-S International MicrowaveSymposium Digest (MTT ’12), pp. 1–3, Montreal, Canada, June2012.

[31] J. Östh, M. Karlsson, A. Serban, and S. Gong, “Carrier leakagesuppression and EVM dependence on phase shifting networkin six-port modulator,” in Proceedings of the International Con-ference onMicrowave andMillimeterWave Technology (ICMMT’12), pp. 1–4, Shenzhen, China, May 2012.

[32] J. Östh,M. Karlsson, A. Serban, and S. Gong, “M-QAM six-portmodulator using only binary baseband data, electrical or opti-cal,” IEEE Transactions on Microwave Theory and Techniques,vol. 61, no. 6, pp. 2506–2513, 2013.

Research ArticlePerformance Driven Six-Port Receiver and Its Advantages overLow-IF Receiver Architecture

Abul Hasan and Mohamed Helaoui

The Intelligent RF Radio Technology Laboratory (iRadio Lab), Department of Electrical and Computer Engineering,Schulich School of Engineering, University of Calgary, Calgary, AB, Canada T2N 1N4

Correspondence should be addressed to Abul Hasan; [email protected]

Received 5 October 2013; Revised 22 November 2013; Accepted 30 December 2013; Published 23 February 2014

Academic Editor: Adriana Serban

Copyright © 2014 A. Hasan and M. Helaoui. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

This paper provides an extensive analysis of the performance of a six-port based direct conversion receiver (SPR) in terms of signalquality, dynamic range, noise figure, ports matching, isolation, bandwidth, and cost. Calibration technique using multimemorypolynomials has been adopted in order to improve the signal quality of the six-port receiver. The performances of the calibratedreceiver are then compared with the performances of a commercially available I-Q demodulator used as a low-IF receiver.Themainadvantages and disadvantages of the SPR compared to the low-IF receiver are highlighted.Themajor advantages of the SPR come interms of its available input frequency bandwidth and the low power requirement. The SPR system requires no external bias supplybut suffers in terms of the available conversion gain. A better port matching of the SPR can be guaranteed over a wide frequencybandwidth, which mixer based receiver systems lack.Themain component limiting the performance of a SPR is the diode detector.A faster and a better diode detector will alleviate some of the problems highlighted in this paper.The SPR system is calibratable andits error-vector-magnitude performance can be made better than the I-Q demodulator used as a low-IF receiver.

1. Introduction

The Six-Port technique was first introduced in the 1970s foraccurate power measurement and reflectometer applications[1, 2]. This technique for receiver applications, first intro-duced as a digital direct conversion receiver in 1994 [3], hascaught considerable attention from the RF and microwaveresearch community. Six-Port based receiver (SPR) archi-tecture has been identified as a promising architecture forimplementing the software defined radio (SDR) receiverconcept [4, 5]. It offers various advantages over conventionalreceiver architectures as it avoids the use of costly andactive nonlinear mixer based circuits to achieve frequencydown-conversion. Wide frequency bandwidth coverage, lowpower consumption, configurability, linearity, passive circuitimplementation, low cost, and higher integration are someof the features that make this architecture very attractive forimplementation of the SDR concept. The six-port concept,as a radio frequency (RF) front-end in communicationreceivers, is usually implemented using power dividers and

quadrature hybrid circuits [6, 7]. Figure 1 shows the mostcommon implementation of the six-port technique as anRF front-end in receiver architectures. This architecture usesthree quadrature hybrids (𝑄

1–𝑄3) and one power divider

(𝐷) connected in such a fashion that appreciable amount ofsignal powers from both the RF and the local oscillator (LO)ports reaches the four power detectors with premeditatedattenuations and phase shifts. The concepts and reasoningbehind various configurations for implementing this six-portjunction circuit or wave-correlator is beyond the scope of thepresent paper.

The receiver front-end shown in Figure 1 is designedusing passive microwave or discrete components. Thoughthese passive circuit elements can be designed to covera very wide frequency spectrum, their performances areguaranteed best at the design (or center) frequency. Also, thepower detectors used in this architecture are usually imple-mented using diodes, which have limited linear dynamicrange. Any deviation in the six-port wave-correlator circuit(𝑄1–𝑄3and 𝐷) performance or the diode power detectors


http://dx.doi.org/10.1155/2014/198120


aRF = IRF + jQRF

aLOLO

D4

D3

D5

D6

Q3

Q2

Q1 D

Digital

Processing

Sign

al

IEST

QEST

Figure 1: A typical architecture of receiver front-end using six-porttechnique.

(𝐷3–𝐷6) behaviors results in degraded receiver performance

reflected in the error vector magnitude (EVM) between thetransmitted constellation points and the received constella-tion points. In order to account for the circuit and systemimperfections, various calibration techniques [8, 9] have beenproposed for the SPR systems. The most common approachfor the SPR systems is the linear modeling and calibrationapproach. In this linear modeling and calibration approachfor the SPR systems, the in-phase (𝐼) and the quadrature (𝑄)components of the received RF signal (𝑎RF) are written aslinear combinations of the four detectors powers, the detailsof which are being provided in Section 2 of this paper.

In this paper the performance of SPR system has beenthoroughly analyzed. A calibration algorithm using multiplememory polynomials has been used in order to improve thelinearity performance of the six-port receiver. This linearityperformance along with other criteria is compared to theperformance of a commercially available 𝐼-𝑄 demodulator,and the advantages and the drawbacks of the SPR systemare highlighted. ADL5382 [10] from Analog Devices, Inc.has been chosen for this purpose. This demodulator is abroadband quadrature 𝐼-𝑄 demodulator and it covers anRF input frequency range from 700MHz to 2.7GHz. Theevaluation board [10] for this 𝐼-𝑄 demodulator has a dc-block at the baseband outputs 𝐼- and 𝑄-ports. To avoid anydistortion in the received signal due to this narrowbandfrequency filtering near zero frequency (dc), the presentevaluation board has been used as a low-IF (intermediatefrequency) receiver.Theblock diagram to implement the low-IF receiver concept using an 𝐼-𝑄 demodulator is shown inFigure 2.The further details of this implementation are beingprovided in Sections 3 and 4.

The remainder of this paper is organized as follows.Section 2 formulates the linear calibration problem andthe used algorithm to boost the linearity of six-port-basedreceiver systems, Section 3 provides the technical details ofthe implemented low-IF receiver using 𝐼-𝑄 demodulator,Section 4 details the measurement setup used for evaluatingthe performances of the implemented and calibrated six-port based receiver and the commercial low-IF receiver,Section 5 provides the results and comparisons, and Section 6

M2

M1

VBIAS

aRF = IRF + jQRF

aLO LO

Digital

Processing

IEST

QEST

Sign

al

90∘

Figure 2: Block diagram of a low-IF receiver using 𝐼-𝑄 demodula-tor.

concludes the paper and provides some final comments andobservations.

2. Linear Calibration of Six-Port Receiver

Modern communication systems use complex representationfor the baseband signals. A real-world message is digitizedand coded in terms of a bit sequence {𝑑

𝑖} which is further

modulated digitally and mapped to a stream of messagesymbols with the help of two orthogonal signals, namely, thein-phase component signal and the quadrature componentsignal. If 𝐼RF(𝑡) and 𝑄RF(𝑡) represent the in-phase and thequadrature components, respectively, of a digitallymodulatedbaseband signal, then the complex baseband signal 𝑎RF(𝑡) iswritten as in

𝑎RF (𝑡) = 𝐼RF (𝑡) + 𝑗𝑄RF (𝑡) ; 𝑗 = √−1. (1)

The complex baseband signal 𝑎RF(𝑡) is frequency upconvertedto a desired carrier frequency 𝑓

𝑐in the transmitter using a

complex mixer. The bandpass representation of the commu-nication signal transmitted by the transmitter is denoted by𝑠(𝑡):

𝑠 (𝑡) = Re {𝑎RF (𝑡) 𝑒𝑗2𝜋𝑓𝑐𝑡}

= 𝐼RF (𝑡) cos (2𝜋𝑓𝑐𝑡) − 𝑄RF (𝑡) sin (2𝜋𝑓𝑐𝑡) .(2)

The fundamental task of a communication receiver is tofaithfully recover the complex baseband signal 𝑎RF(𝑡), thatwill in turn provide the originally transmitted informationbits {𝑑

𝑖}. The following subsections describe the approach

throughwhich a six-port based receiver recovers the complexbaseband signal in an efficient and accurate way.

2.1. Six-Port Junction Calibration. Let 𝑎RF = 𝐼RF + 𝑗𝑄RF =|𝑎RF|𝑒

𝑗𝜑RF be the phasor representation of the original trans-mitted signal reached at the receiver antenna and let 𝑎LO =|𝑎LO|𝑒

𝑗𝜑LO be the phasor representation of local oscillator


(LO) signal; then the power 𝑃𝑘detected by the 𝑘th power

detector can be written as in

𝑃𝑘=𝑆𝑘2𝑎RF+𝑆𝑘1𝑎LO

2, 𝑘 = 3, 4, . . . , 6, (3)

where 𝑆𝑖𝑗, 𝑖, 𝑗 = 1, 2, . . . , 6, are the 𝑆-parameters of the six-

port junction or wave-correlator circuit. Equation (3) can beexpanded and rewritten in a matrix form as in

[[[

[

𝑃3

𝑃4

𝑃5

𝑃6

]]]

]

=[[[

[

𝑇31𝑇32𝑇33𝑇34

𝑇41𝑇42𝑇43𝑇44

𝑇51𝑇52𝑇53𝑇54

𝑇61𝑇62𝑇63𝑇64

]]]

]

[[[

[

1

𝐼2

RF + 𝑄2

RF𝐼RF𝑄RF

]]]

]

, (4)

where

𝑇𝑘1=𝑆𝑘12𝑎LO

2; 𝑇𝑘2=𝑆𝑘22; for 𝑘 = 3, . . . , 6,

𝑇𝑘3= 2𝑆𝑘1𝑆𝑘2𝑎LO cos (∠𝑆𝑘2 − ∠𝑆𝑘1 − 0LO) ,

𝑇𝑘4= −2

𝑆𝑘1𝑆𝑘2𝑎LO sin (∠𝑆𝑘2 − ∠𝑆𝑘1 − 0LO) .

(5)

Matrix [𝑇]4×4

in (4) is dependent only on the 𝑆-parameters ofthe six-port junction and the LO signal. From (4), 𝐼RF and𝑄RFcan be written as a linear combination of 𝑃

3, 𝑃4, 𝑃5, and 𝑃

6

by inverting [𝑇] as in

𝐼RF =6

∑

𝑘=3

𝛼𝑘𝑃𝑘, (6)

𝑄RF =6

∑

𝑘=3

𝛽𝑘𝑃𝑘. (7)

Here, 𝛼𝑘and 𝛽

𝑘are the entries in the third and the fourth

row of [𝑇]−14×4

matrix, respectively. We call this approachof modeling and calibrating a six-port receiver as linearmodeling and calibration. These eight constants (𝛼

𝑘and 𝛽

𝑘;

𝑘 = 3, . . . , 6) are called calibration constants and theymust beobtained through a suitable calibration means for successfulrecovery of the 𝐼- and 𝑄-components of the received signal.These parameters are independent of the RF signal and arefunctions of the 𝑆-parameters of the six-port junction and theLO signal.

2.2. Power Detector Linearization. It is obvious from (6) and(7) that this equation uses four powers readings (𝑃

3, . . . , 𝑃

6)

from the four power detectors. The outputs from the powerdetectors are voltages which shall be proportional to theircorresponding input powers, if the detectors are operating intheir ideal linear regions. In the linear regions of the diodepower detectors, the relationship between the input powersand the output voltages can be written as in

V𝑘= 𝑘𝑘𝑃𝑘; 𝑘 = 3, 4, . . . , 6, (8)

where 𝑘𝑘is a parameter dependent on the characteristic of the

𝑘th power detector.Diode-based power detectors are widely used nowadays

because of their low power and simplified support cir-cuitry requirement. The dynamic range of the diode powerdetectors, where the output voltage from the detector isproportional to the input microwave power, is quite limited.This limits the useful range of operations of the diode powerdetectors formany applications. For example, when the diodepower detectors are used in a SPR system, the dynamic rangesof the diode power detectors limit the dynamic range of thewhole receiver system. To extend this dynamic range, thepower detectors can be operated beyond their linear regions,where they exhibit nonlinear characteristic. Linearization ofthe diode power detectors characteristics is performed toopen up this limitation and to extend their dynamic rangewithout affecting the received signal quality. The outputvoltage of the diode power detector is used to predict itscharacteristic and compensate for its nonlinear behavior.In nonlinear region of a diode power detector, the outputdetector voltage can be thought of as a nonlinear function ofthe detector input power as shown in

V𝑘= 𝑓 (𝑃

𝑘) ; 𝑘 = 3, 4, . . . , 6, (9)

where 𝑓(⋅) is a nonlinear function governed by the nonlinearcharacteristic of the 𝑘th power detector. In order to obtainthe linear power corresponding to any output voltage, aninverse function 𝑓−1 must be applied to the detector outputvoltage (V

𝑘). There are different ways [9, 11, 12] to implement

this inverse function and linearize the diode power detectorcharacteristic, but we choose here a wideband linearizationapproach [9] that results in a single step linearization andcalibration for the whole six-port based receiver system.

2.3. Wideband Linearization and Calibration. It has beenreported in [9] that the diode detectors display memoryeffect, which is a frequency dependent phenomenon. Whenthe inputs to the diode detectors arewidebandmodulated sig-nals, the relationship between the instantaneous inputs powerand the corresponding output voltage is quite different fromthe usual continuous wave (CW) characteristics modeled in(8). This phenomenon has been modeled more accurately in[9] using a modified memory polynomial:

𝑃𝑘 [𝑛] =

𝑀𝑘

∑

𝑞=0

𝑁𝑘

∑

𝑝=1

𝑐𝑝𝑞𝑘

V𝑝𝑘[𝑛 − 𝑞] . (10)

The detector linearizer model in (10) is applied to (6)and (7) to obtain a wideband single step in situ linearizationand calibration model for the whole six-port based receiversystem. This wideband model [9] to linearize and calibrate


the SPR system and to obtain the 𝐼- and 𝑄-components ofthe received signal is written as in

𝐼EST [𝑛] =6

∑

𝑘=3

𝑀𝑘

∑

𝑞=0

𝑁𝑘

∑

𝑝=1

𝛼𝑝𝑞𝑘

V𝑝𝑘[𝑛 − 𝑞] , (11)

𝑄EST [𝑛] =6

∑

𝑘=3

𝑀𝑘

∑

𝑞=0

𝑁𝑘

∑

𝑝=1

𝛽𝑝𝑞𝑘

V𝑝𝑘[𝑛 − 𝑞] . (12)

Here, 𝛼𝑝𝑞𝑘

(=𝑐𝑝𝑞𝑘𝛼𝑘) and 𝛽

𝑝𝑞𝑘(=𝑐𝑝𝑞𝑘𝛽𝑘) are the wideband

calibration parameters thatmust be estimated using a suitablecalibration means. Once these parameters are calibrated forthe SPR system, the 𝐼- and 𝑄-components of the receivedsignal can be recovered from the four detectors voltages using(11) and (12).

3. Low-IF Receiver Using 𝐼-𝑄 Demodulator

The commercially available 𝐼-𝑄 demodulator used forcomparison in this paper is an evaluation board (EVB) ofADL5382 [10] from Analog Devices, Inc. This demodulatorcovers anRF input frequency range from700MHz to 2.7 GHzand requires a 5V supply to generate internal biases [10].The reason for using this EVB as a low-IF receiver insteadof direct conversion receiver (DCR or zero-IF receiver) hasbeen explained in Section 1. To reinforce the point, Figure 3shows the spectrum plots for a transmitted and received 16-QAM modulated signal when 𝐼-𝑄 demodulator is used asa DCR. The removal of frequency components around zerofrequency from the received signal by the EVB results indistortions and the increased EVM between the transmittedand the received constellation points.

The 𝐼-𝑄 demodulator for the low-IF receiver is operatedin single ended mode as is the case with the six-port-based receiver system discussed in Section 2. To operate the𝐼-𝑄 demodulator as a low-IF receiver, the LO frequency(𝑓LO) supplied to the demodulator has been offset by anintermediate frequency (𝑓IF) from the carrier frequency (𝑓𝑐)of the RF signal sent by the transmitter.The 𝐼-𝑄 demodulatoroutput in this case can be written as in

𝑠IF (𝑡) = Re {𝑎RF (𝑡) 𝑒𝑗2𝜋𝑓𝑐𝑡𝑒−𝑗(2𝜋𝑓LO𝑡+𝜑)}

= Re {𝑎RF (𝑡) 𝑒𝑗[2𝜋(𝑓

𝑐−𝑓LO)𝑡+𝜑]}

= Re {𝑎RF (𝑡) 𝑒𝑗(2𝜋𝑓IF𝑡+𝜑)} .

(13)

Here, 𝜑 is the initial phase difference between the transmitterLO signal (𝑓

𝑐) and the receiver LO signal (𝑓LO). The real

signal 𝑠IF(𝑡) is the signal which is captured and processedin the digital signal processing block shown in Figure 2.The digitized signal 𝑠IF(𝑡) is digitally demodulated fromintermediate frequency 𝑓IF to the baseband and corrected forthe phase error 𝜑. The demodulated signal approximates theoriginal transmitted baseband signal 𝑎RF(𝑡).

In order to assess the performance of both receivers,a performance metric in terms of error vector magnitude(EVM) between the transmitted and the received signal is

−4 −3 −2 −1 0 1 2 3 4−110−100

−90−80−70−60−50−40−30−20−10

Frequency (MHz)

PSD

(dB)

Input signalReceived signal

Figure 3: Input and output signal spectrums of a 16-QAM modu-lated signal having 1MHz frequency bandwidth when 𝐼-𝑄 demod-ulator is used as a direct conversion receiver (DCR).

Six-port DSPLO

RF

E8247C

N5182AMSO9404A

10MHz REF.

Figure 4: Complete test setup for the six-port based communicationreceiver.

used.TheEVMbetween the transmitted signal (𝑇[𝑛]) and thereceived signal (𝑅[𝑛]) of symbol lengths𝑁 is defined as in

EVM = √(1/𝑁)∑

𝑁

𝑛=1|𝑅 [𝑛] − 𝑇 [𝑛]|

2

(1/𝑁)∑𝑁

𝑛=1|𝑇 [𝑛]|

2× 100%. (14)

4. Implementation and Measurement Setup

TheSPR and the low-IF receivers test setups are implementedas shown in Figures 4 and 5, respectively.

The baseband signals for the test setups are generatedin MATLAB and sent to the MXG vector signal genera-tors (N5182A from Agilent Technologies, Inc.) for signalmodulation and frequency upconversion to a desired carrierfrequency (𝑓

𝑐). The output of this signal generator is con-

nected to the RF ports of the receivers to simulate anactual RF received signal. The LO signal for the receivers


ADL5382 EVB

I

Q

DSPLO

RF

VBIAS

GND

E8247C

E3631A

N5182A

MSO9404A

10MHz REF.

Figure 5: Test setup for the low-IF receiver.

is supplied by a PSG CW signal generator (E8247C fromAgilent Technologies, Inc.) and the desired receiver LOfrequency is set here. Off-the-shelf 3 dB quadrature hybridsand power divider are used to configure the six-port junctionstructure for the SPR system. Four zero bias Schottky diodedetectors (8472B from Agilent Technologies, Inc.) are usedas port power detectors as shown in Figure 4. The EVB forADL5382 is used as 𝐼-𝑄 demodulator for the low-IF receiveras shown in Figure 5. DC power supply (E3631A fromAgilentTechnologies, Inc.) provides the bias voltage required forADL5382EVB. The four detectors output voltages (in case ofSPR) or the demodulator 𝐼- and 𝑄-signals (in case of low-IF receiver) are captured using synchronized four-channelmixed signal scope (MSO9404A from Agilent Technologies,Inc.) with the VSA software.The captured voltages, in case ofSPR, are used to calibrate the receiver system and estimatethe received symbols using (11) and (12) as explained inSection 2. The captured voltages, in case of low-IF receiver,are used to bring down the low-IF signal to baseband usingan appropriate demodulation process explained in Section 3.All the calibration, estimation, and demodulation are carriedout in MATLAB on a desktop computer.

5. Results and Observations

The 16-QAM and the 64-QAM signals with different fre-quency bandwidths are sent at a carrier frequency of 2.0GHzto both the SPR and the low-IF receivers. The basebandsymbols are upsampled by a factor of 8 and passed througha raised cosine filter of roll-off factor 0.3 and a delay of 3 taps.The mean power of the RF signal is set to be −3 dBm. TheLO signal is supplied at the carrier frequency, in case of SPR,and at a frequency offset by 𝑓IF from the carrier frequency,in case of low-IF receiver. The LO power is set to be 0 dBm.The RF and the LO powers are kept constants for both setupsthroughout the experiment.The low-IF receiver bias circuitrydraws a current of 0.268A at constant 5.0 V from the DC

Table 1: LO frequencies for the low-IF receiver.

Modulationscheme

Bandwidth(MHz)

Samplingfrequency(MHz)

ReceiverLO frequency

(GHz)

𝑓IF(MHz)

16/64-QAM 1 8 1.9985 1.516/64-QAM 2 16 1.9970 3.016/64-QAM 10 80 1.9850 15

power supply. Table 1 summarizes the LO frequencies usedfor different signals in the low-IF receiver.

The baseband signal sampling frequency for a specificsignal type is kept the same for both receiver architectures forfair comparison.

5.1. Architectural Comparison. Table 2 summarizes the mainarchitectural differences between the SPR systemand the low-IF receiver used for study in this paper. The demodulationbandwidth in case of SPR system is lower as compared to low-IF receiver because the diode detectors used here are limitedin their speeds.

5.2. EVM Performance Comparison. With the settings inTable 1 and the transmitted signal carrier frequency at2.0GHz, the 𝐼- and 𝑄-data for low-IF receiver are obtainedthrough the process explained in Section 3. Similarly, the𝐼- and 𝑄-data for the SPR are obtained according to themethod explained in Section 2 after boosting the linearity ofthe SPR receiver using the calibration technique shown in (11)and (12).

Table 3 summarizes the EVM performances of bothreceivers under identical signal conditions. For the sakeof completeness, the EVMs of the received signals for the𝐼-𝑄 demodulator used as DCR are also listed in Table 3.Conventional SPR performance is based on polynomialdetector linearizer [13] which corresponds to the linear andmemoryless calibrationmodel of [9].The in situ linearizationand calibration model improves the overall performanceof the SPR system. Figures 6 and 7 show the transmittedand the received constellation points of a 64-QAM signalhaving 2MHz bandwidth for the low-IF receiver and theSPR system, respectively. Similarly, Figures 8 and 9 show thetransmitted and the received signals frequency spectrums of a64-QAMmodulated signal having 2MHz bandwidth for thelow-IF receiver and the SPR system, respectively. The EVMperformance of the SPR system is better in comparison tothe low-IF receiver. The high EVM for low-IF receiver comesmainly from image frequency and IMDs (intermodulationdistortions).We have not compensated for these impairmentsin the low-IF receiver, while the SPR system is calibrated forany nonlinearities and memories present. The spectrum forthe low-IF receiver shown in Figure 8 is filtered to remove theimage and IMDs coming from the IF-receiver.

5.3. Power Requirements Comparison. Table 4 compares thepowers of the RF signal and LO signal and the total powerrequirement for both receiver systems.The SPR requires four


Table 2: Architectural comparison between SPR and low-IF receiver.

Architecture Six-port based receiver Low-IF receiver

Key components 4 × Diode detectors + 3 × quadrature hybrids + 1× power divider

2 × 4 transistors Gilbert cell mixer + 1 × phaseshifter + 1 × power divider + 2 × low pass filters

Bias requirement No YesReceiver LOFrequencyPlanning

No, same as carrier frequency Yes

Input RF bandwidth 2–18GHz 0.7–2.7 GHz3 dB demodulation bandwidth ∼29MHz ∼185 MHz

Table 3: EVM (%) performance of different receivers under test.

Signal type Conventionalsix-port receiver [9, 13]Calibrated six-port

receiver [9] Low-IF receiverDirect conversion

receiver16-QAM 1MHz 4.19 0.64 6.88 21.2516-QAM 2MHz 6.42 0.80 4.99 15.0616-QAM 10MHz 7.35 2.51 4.69 7.0264-QAM 1MHz 4.50 0.73 6.90 22.2064-QAM 2MHz 6.58 0.77 5.38 16.2364-QAM 10MHz 7.40 2.48 4.59 8.65

Table 4: Power requirements of different receivers under test.

Power type six-port receiver Low-IF receiverRF power 0.0005W (−3 dBm) 0.0005W (−3 dBm)LO power 0.001W (0 dBm) 0.001W (0 dBm)DC power 0W 1.34W (31.27 dBm)ADC power 4 × 250mW (30 dBm) 2 × 250mW (27 dBm)Total power 1.0015W (30.01 dBm) 1.8415W (32.65 dBm)

ADCs for sampling the baseband signals, while the low-IFreceiver requires only two. A 12-bit ADC (AD9626 fromAnalog Devices, Inc.) [14] has been chosen for the powerconsumption comparison provided in Table 4.

The SPR system has an advantage here as it does notrequire any external bias power supply. This drasticallyreduces the power consumption of the whole receiver system.For the example case considered here, the total powerrequirement of the SPR system (1.0015W/30.01 dBm) is abouthalf (∼54%) of the total power requirement of the low-IFreceiver (1.8415W/32.65 dBm). It is true, from the datasheet[10] of the EVB, that the low-IF receiver can be operatedat much lower power levels of the RF and the LO signals,still the bias power requirement dominates the total powerrequirement of the low-IF receiver.

The abovementioned levels of the RF and the LO powershave been chosen such that the signal levels at the inputof the diode detectors are above their sensitivity levels andthe detectors generate detectable voltages at their outputterminals.

5.4. Ports Matching Comparison. The RF and the LO portsmatching and the isolation between them are important

Table 5: Port matching of different receivers under test.

Port data Six-portreceiverLow-IFreceiver

RF port return loss ∼30 dB ∼10 dBLO port return loss ∼30 dB ∼10 dBIsolation between RF and LO port ∼40 dB ∼65 dBc

parameters for characterizing a receiver system. Table 5 sum-marizes these characteristics for both receivers under study.Because of passive quadrature hybrids and power dividersinvolvement better port matching over a wide frequencybandwidth can be achieved in case of SPR system, while low-IF receiver offers better isolation between the RF and the LOports. The isolation between the RF and the LO ports thatresults in dc-offset in the output is an important parameterfor direct conversion receivers. The performance loss due tothis imperfection can be accurately modeled and calibratedfor in the SPR systems.

5.5. Conversion Gain Comparison. The outputs of the low-IF receiver EVB are the single ended 𝐼- and 𝑄-signals whilethose of the SPR system are four voltages generated by thefour power detectors. The peak value of the 𝐼/𝑄-signals incase of low-IF is about 250mV, while the peak value of thedetectors output voltages is about 22mV. The RF and the LOpowers for the two test cases are identical. Considering theconnectors and cable losses for both the test setups identical,the low-IF receiver has at least 18 dB higher conversiongain as compared to the six-port based receiver. The higherconversion gain of the low-IF receiver comes from the activemixers used in the demodulator circuitry.


Transmitted symbolsReceived symbols

−0.8

−0.6

−0.4

−0.2

0

0.2

0.4

0.6

0.8

−0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8

In-phase component (I)

Qua

drat

ure c

ompo

nent

(Q)

Constellation-64QAM-2MHz-2.0GHz

Figure 6: Transmitted and received constellation points of a 64-QAM signal having 2MHz bandwidth for low-IF receiver.

−0.8

−0.6

−0.4

−0.2

0

0.2

0.4

0.6

0.8

−0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8

In-phase component (I)

Qua

drat

ure c

ompo

nent

(Q)

Transmitted symbolsReceived symbols

Constellation-64QAM-2MHz-2.0GHz

Figure 7: Transmitted and received constellation points of a 64-QAM signal having 2MHz bandwidth for six-port-based receiver(SPR).

5.6. Noise Figure, Sensitivity, and Dynamic Range. The noisefigure of the SPR RF front-end circuit is defined here in termsof the loss of the wave-correlator circuit and the noise figureof the diode detector using Friis equation:

𝐹 = 𝐿 + 𝐿 (𝐹𝐷− 1) . (15)

Here, 𝐿 is the loss of the wave-correlator circuit. This lossfactor 𝐿 is about 7 dB for passive wave-correlator consistingof quadrature hybrids and power divider.The noise factor 𝐹

𝐷

of the diode detector is defined as in

𝐹𝐷=

SNR𝑖

SNR𝑜

=𝑆𝑖

𝑆𝑜

𝑁𝑜

𝑁𝑖

, (16)

where 𝑆𝑖, 𝑁𝑖and 𝑆

𝑜, 𝑁𝑜are the detector input and output

signal and noise powers, respectively. Typical noise for single-tone continuous wave (CW) input for the used diode detectoris less than 50 𝜇V peak-to-peak [15]. The detector outputsare more than 300mV for an input CW signal of power20 dBm. Assuming the input noise to the diode detector to be

−3 −2 −1 0 1 2 3 4

−90−80−70−60−50−40−30−20−10

Frequency (MHz)

PSD

(dB)


Figure 8: Transmitted and received frequency spectrums of a 64-QAM signal having 2MHz bandwidth for low-IF receiver.

−3 −2 −1 0 1 2 3 4

−90−80−70−60−50−40−30−20

Frequency (MHz)

PSD

(dB)


Figure 9: Transmitted and received frequency spectrums of a 64-QAM signal having 2MHz bandwidth for six-port-based receiver(SPR).

at the thermal level, the resulting noise figure of the six-portreceiver is much higher as compared to the low-IF receivernoise figure of about 16 dB [10].

At about 2GHz frequency, the minimum RF input levelfor acceptable EVM for the low-IF receiver is close to−40 dBm [10]. The minimum RF level for the SPR is limitedby the minimum detectable level of the diode detectors used.Given the noise level at the output of the detectors (50𝜇Vpeak-to-peak), the minimum RF signal level that can beaccepted by the SPR system is about −10 dBm [15].

The maximum power handling capability of the useddiode detector is about 20 dBm. This decides the dynamicrange of the diode detectors and hence the dynamic rangeof the whole receiver system. Though the linear region ofthe diode detector used is quite limited, its dynamic rangecan be improved to about 30 dB using the linearization andcalibrationmethod used in [9].Thedynamic range of the low-IF receiver evaluated here is more than 50 dB [10]. Table 6summarizes the results of the analysis carried out.


Table 6: Summary of the comparisons.

Architecture Calibrated six-portreceiverLow-IFreceiver

Input RF bandwidth ++++ ++3 dB demodulation bandwidth ++ ++++Circuit complexity +++ +Port matching +++ −Ports isolation ++ +++Power requirement ++ −−EVM performance +++ ++Noise figure − − − ++Sensitivity − − − +++Dynamic range −− +++Intermodulation +++ −−Cost ++ +Conversion gain −− +++

6. Conclusions

In this paper the performance of a six-port based receiver(SPR) system has been analyzed. Multimemory polynomialcalibration technique has been applied to the detected voltagereadings from the four power detectors in order to boost thelinearity of the receiver.This linearity performance and otherperformance metrics such as dynamic range, power con-sumption, port matching, and isolation are compared to theperformance of a commercially available 𝐼-𝑄 demodulator.The calibrated six-port based receiver offers a wide availableRF bandwidth and better port matching as compared to the𝐼-𝑄 demodulator but suffers heavily in terms of noise figure,sensitivity, and dynamic range. The major limitations of theSPR come from the diode power detectors. Better powerdetectors will alleviate some of the issues highlighted here.The error-vector-magnitude (EVM) performance of the SPRsystem can be improved by linearizing the diode detectorsand calibrating the whole system in situ. The other advantageof the SPR system comes in terms of its power requirement.SPR system requires no external bias supply and the LOpower requirement is comparable to the RF power level. Theexternal supply helps improve the conversion gain of the 𝐼-𝑄 demodulator. The 3 dB demodulation bandwidth of the𝐼-𝑄 demodulator is almost a decade higher than the SPRsystem. The demodulation bandwidth of the SPR system canbe improved by using faster diode detectors in the system.

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper.

Acknowledgments

This work was supported by the Alberta Innovates, Technol-ogy Futures (AITF), the Natural Sciences and Engineering

Research Council of Canada (NSERC), and the CanadaResearch Chair (CRC) Program.

References

[1] G. F. Engen and C. A. Hoer, “Application of an arbitrary 6-portjunction to power-measurement problems,” IEEE Transactionson Instrumentation and Measurement, vol. 21, no. 4, pp. 470–474, 1972.

[2] G. F. Engen, “The six-port reflectometer: an alternative networkanalyzer,” IEEE Transactions on Microwave Theory and Tech-niques, vol. 25, no. 12, pp. 1075–1080, 1977.

[3] J. Li, R. G. Bosisio, andK.Wu, “Six-port direct digitalmillimeterwave receiver,” in Proceedings of the IEEE MTT-S InternationalMicrowave Symposium, pp. 1659–1662, May 1994.

[4] R. G. Bosisio, Y. Y. Zhao, X. Y. Xu et al., “New wave radio,” IEEEMicrowave Magazine, vol. 9, no. 1, pp. 89–100, 2008.

[5] A. Koelpin, G. Vinci, B. Laemmle, D. Kissinger, and R. Weigel,“The six-port in modern society,” IEEE Microwave Magazine,vol. 11, no. 7, pp. S35–S43, 2010.

[6] A. Serban, M. Karlsson, J. Osth, Owais, and S. Gong, “Differen-tial circuit technique for six-port modulator and demodulator,”in Proceedings of the IEEE MTT-S International MicrowaveSymposium, 2012.

[7] C. de La Morena-Alvarez-Palencia, M. Burgos-Garcia, and J.Gismero-Menoyo, “Contribution of LTCC technology to theminiaturization of six-port networks,” in Proceedings of the41st European Microwave Conference (EuMC ’11), pp. 659–662,October 2011.

[8] C. de laMorena-Álvarez Palencia andM. Burgos-Garćıa, “Four-octave six-port receiver and its calibration for broadbandcommunications and software defined radios,” Progress inElectromagnetics Research, vol. 116, pp. 1–21, 2011.

[9] A. Hasan and M. Helaoui, “Novel modeling and calibrationapproach for multiport receivers mitigating system imper-fections and hardware impairments,” IEEE Transactions onMicrowaveTheory and Techniques, vol. 60, no. 8, pp. 2644–2653,2012.

[10] “700 MHz to 2700 MHz Quadrature Demodulator,” AnalogDevices, Norwood, Mass, USA, ADL5382 datasheet, 2012.

[11] G. F. Luff, P. J. Probert, and J. E. Carroll, “Real-time six-port reflectometer,” IEE Proceedings H: Microwaves Optics andAntennas, vol. 131, no. 3, pp. 186–190, 1984.

[12] F. Lan, C. Akyel, F. M. Ghannouchi, J. Gauthier, and S. Khouaja,“Six-port based on-line measurement system using specialprobe with conical open end to determine relative complexpermittivity at radio andmicrowave frequencies,” inProceedingsof the 16th IEEE Instrumentation and Measurement TechnologyConference (IMTC ’99), vol. 1, pp. 42–47, June 1999.

[13] A. Hasan, M. Helaoui, and F. M. Ghannouchi, “Dynamiclinearization of diodes for high speed and peak power detectionapplications,” in Proceedings of the IEEE MTT-S InternationalMicrowave Symposium Digest (MTT ’12), pp. 1–3, June 2012.

[14] “AD9626: 12-Bit Analog-to-digital converter,” Analog Devices,Norwood, Mass, USA, AD9626 datasheet, 2007.

[15] “Agilent low barrier Schottky diode detectors,” Agilent Tech-nologies, Santa Clara, Calif, USA, 8472B datasheet, 2009.

Research ArticlePerformance of 2–3.6 GHz Five-Port/Three-PhaseDemodulators with Baseband Analog 𝐼/𝑄RegenerationCircuit in Direct-Conversion Receivers

Kaissoine Abdou,1 Antoine Khy,1 Kais Mabrouk,2 and Bernard Huyart1

1 Communications and Electronics Department, TELECOM ParisTech, 46 rue Barrault, 75013 Paris, France2 Laboratoire de Recherche et d’Innovation Technologique, Ecole Polytechnique de Sousse, rue Khalifa Karoui Sahloul,4000 Sousse, Tunisia

Correspondence should be addressed to Antoine Khy; [email protected]

Received 4 October 2013; Revised 24 December 2013; Accepted 26 December 2013; Published 20 February 2014

Academic Editor: Serioja Tatu

Copyright © 2014 Kaissoine Abdou et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

A comparison between performance of a five-port demodulator (FPD) and a three-phase demodulator (TPD) with botharchitectures connected to a so-called baseband 𝐼/𝑄 regeneration circuit is carried out. In order to compare these two receiversto a classical architecture, the performance of an 𝐼/𝑄 demodulator (IQD) is also presented. Measured results show the superiorityof TPD and IQD over FPD in terms of residual DC offsets and 2nd order intermodulation distortion (IMD2) products, noisefigure (NF), and sensitivity due to the use of active balanced mixers instead of diode power detectors. Lastly, demodulation of threenoncontiguous RF carriers shows the stronger potential of the three-phase demodulator (TPD) for applications in future long termevolution-advanced (LTE-A) receivers through EVM and constellation diagram measurements.

1. Introduction

When it was first introduced in the beginning of the 70’s, thesix-port technique was originally used in metrology to accu-rately measure reflection coefficients of complex impedances[1] like a vectorial network analyzer (VNA). Since then, thefield of applications of six-port systems has been extendedto measurement of direction of arrival (DoA) of multiplesignals [2, 3], phase/frequency discriminator for automobileradars [4, 5], microwave and millimeter-wave sensing [6],and demodulators in direct-conversion receivers [7–20]. Thesix-port technique is based on an RF circuit with 2 inputsand 4 outputs that performs the linear combination of the2 input signals, 4 power detectors connected to each outputof the RF circuit for quadratic detection, 4 analog-to-digitalconverters (ADCs), and a digital calibration procedure. Themain advantage of the six-port technique is due to theinherent redundancy of the output data that allows alleviatingthe performance of the RF components of the circuit. On theother hand, the downside of this lies in the use of 4 ADCs

at the outputs of the system. However, in direct-conversionreceiver applications, the number of ADCs can be reducedto 3 by using the so-called five-port technique [21–26]. Inthis particular case, the power level of the local oscillator(LO) is supposed to be constant or prone to very smallvariations, which accordingly permits the suppression of oneADC. In order to further reduce the number of ADCs toonly 2 likewise conventional 𝐼/𝑄 architectures, the 3 outputsof the five-port circuit have to be designed so that one ofthem is an amplitude and phase symmetry axis in relationto the two others [26–28]. When these two conditions arefulfilled, a single analog OP-amp circuit can be connectedto the 3 outputs of the five-port to perform basic operations(sum and difference) between them, resulting in two 90∘quadrature 𝐼 and 𝑄 output components of the demodulatedRF signal. For this reason, we call this analog OP-amp circuita baseband 𝐼/𝑄 regeneration circuit [26]. Other valuable ben-efits of the use of this circuit are DC offsets suppression and2nd order intermodulation distortion (IMD2) cancellation,which are serious issues in direct-conversion receivers [29].


http://dx.doi.org/10.1155/2014/631867


RF

LO

B3

B4

B5

A3 A4 A5

𝛾3

𝛾4

𝛾5

�3(t)

�4(t)

�5(t)+

+

+ 2

2

2

𝜑3

𝜑4

𝜑5

(a)

RF

LO

B3

B4

B5

A3 A4 A5

𝛾3

𝛾4

𝛾5

�3(t)

�4(t)

�5(t)

𝜑3

𝜑4

𝜑5

(b)

Figure 1: General model of the (a) FPD and (b) TPD.

Lastly, it should be pointed out that the use of the 𝐼/𝑄regeneration circuit allows to avoid the tedious calibrationprocedure mentioned above for a five-port circuit alone [21].However, in practice the two outputs of the 𝐼/𝑄 regenerationcircuit are two scaled versions of the actual 𝐼 and 𝑄 com-ponents, showing amplitude and phase mismatch. Thus, theactual 𝐼 and 𝑄 values are determined by means of simpleamplitude/phase equalization. In [26], a 2.1 GHz microstripring-based five-port system used with the 𝐼/𝑄 regenerationcircuit exhibits 90 ± 5∘ phase shifted 𝐼 and 𝑄 componentsover more than 2 octaves, 10 dB DC offsets suppression, andup to 23 dB IMD2 rejection compared to microstrip five-port circuit alone. However, the main drawback of five- orsix-port architectures used in direct-conversion receivers istheir low sensitivity due to the very high conversion loss(30–40 dB) of diode power detectors. In [30], for a 2.1 GHzmicrostrip five-port circuit that uses Agilent HSMS-2850diodes for detection, measured conversion loss is 39 dB andnoise figure is 36 dB. Simulation results in [12] show thatconversion loss can be reduced to 16 dB with the use ofAgilentHSCH-9161 diodes. In this paper, the FPDuses tunneldiodes instead of Schottky diodes that exhibit less than 11 dBconversion loss, which still remains unacceptable. To improvesensitivity of five- or six-port circuits, mixers can be usedinstead of diodes since they show lower conversion loss(even conversion gain for active mixers) and lower noisefigure and generate less DC offsets and IMD2 products.Obviously, since mixers are three-port devices, the RF circuitthat performs the linear combination of theRF andLO signalshas to be slightly modified compared to the case of two-portpower detectors. For this reason, the naming “three-phasedemodulator” (TPD) is used for this type of architectureincluding mixers while the “five-port demodulator” (FPD)term is dedicated to architectures with power detectors, only.

In this paper, the performances of an FPDand aTPDbothconnected to the baseband 𝐼/𝑄 regeneration circuit discussedabove are studied and compared to those obtained with aclassical 𝐼/𝑄 demodulator (IQD). These 3 receivers are dedi-cated to long term evolution-advanced (LTE-A) applicationsand operate between 2GHz and 3.6GHz.While the spectrumallocated to future LTE-A systems is expected to span from

700MHz to 3.6GHz, our study was limited to the 2–3.6GHzfrequency range due to the limited bandwidth of the 3-waypower dividers used in the implemented FPD and TPD.First, the operating principles of the FPD, TPD and the 𝐼/𝑄regeneration circuit are explained and the implementationof the 2 demodulators is presented. Then, experimentalresults are presented for 𝐼/𝑄 amplitude/phase imbalance, DCoffsets and IMD2 products suppression, noise figure, andsensitivity for each architecture. Also, baseband spectrum ofthree noncontiguous RF signals after demodulation is shownas well as error vector magnitude (EVM) measurementsand constellation diagrams. These results demonstrate thesuperiority of the TPD architecture in overall performanceexcept for residual DC offsets that are higher in TPD thanin IQD. To the best of our knowledge, this is the first studydealing with performance comparison between FPD, TPD,and IQD.

2. Five-Port Demodulator (FPD)and Three-Phase Demodulator (TPD)Basic Principles

Figure 1 shows the general models of the FPD and TPDthat are used in direct-conversion receiver architectures. Thedifference between FPD and TPD lies in the active devicesused for downconversion, which are detectors andmixers forFPD and TPD, respectively. The circuits have 2 inputs forLO and RF signals and 3 low frequency outputs V

3(𝑡), V4(𝑡),

and V5(𝑡). After LO and RF signals are both split by 3-way

power dividers with different attenuation coefficients 𝐴𝑖and

𝐵𝑖and phase shifts 𝛾

𝑖, 𝜑𝑖(𝑖 = 3, 4, 5), the signals are combined

together by means of adders. Then, square-law devices per-form envelope detection as well as downconversion, low-passfilters suppress any high frequency components, and lastly,three ADCs digitize the output voltages V

3(𝑡), V4(𝑡), and V

5(𝑡).

In the case of the TPD, the adders and envelope detectors arereplaced by mixers that multiply LO and RF signals [23, 28].Although, devices used for downconversion differ in the FPDand TPD, the basic operating principle is the same for both.In depth analysis with equations derivations can be found in[21, 26, 27].


It can be shown that the 3 low frequency voltages of theFPD or TPD can be expressed as follows:

V𝑖 (𝑡) =

𝐾𝑖

2𝐴2

𝑖𝑉2

LO +𝐾𝑖

2𝐵2

𝑖[𝐼2(𝑡) + 𝑄

2(𝑡)] + 𝐾𝑖𝑉LO𝐴 𝑖𝐵𝑖

× [cos(𝛾𝑖−𝜑𝑖) 𝐼 (𝑡)+sin(𝛾𝑖−𝜑𝑖) 𝑄 (𝑡)] 𝑖 = {3, 4, 5} ,

(1)

where𝐾𝑖is the second-order coefficient of the diode transfer

characteristic (resp., conversion loss of the mixer) at port𝑖 in the case of the FPD (resp., TPD) while 𝐼(𝑡) and 𝑄(𝑡)are the in-phase and quadrature-phase components of theinput modulated RF signal. The first term in (1), that is aDC component, is due to LO signal self-mixing. It is themain contributor to DC offsets. The second term, whichcomprises a DC and low frequency time variant components,results from self-mixing of the RF signal. This low frequencycomponent is the IMD2 product. As for the last term,it contains the desired actual 𝐼(𝑡) and 𝑄(𝑡) components.Accordingly, (1) can be rewritten as follows:

V𝑖 (𝑡) = DC𝑖 + IMD2𝑖 (𝑡)

+ 𝐾𝑖𝑉LO𝐴 𝑖𝐵𝑖 [cos (Φ𝑖) 𝐼 (𝑡) + sin (Φ𝑖) 𝑄 (𝑡)]

𝑖 = {3, 4, 5} ,

(2)

where DC𝑖is the sum of the DC components due to LO and

RF signals self-mixing; IMD2𝑖(𝑡) is the low frequency time

variant component due to 2nd order nonlinearity and Φ𝑖=

𝛾𝑖− 𝜑𝑖.

3. Baseband Analog 𝐼/𝑄 Regeneration Circuit

As mentioned before, if one of the 3 low frequency voltagesV𝑖(𝑡) is an amplitude and phase symmetry axis in relation to

the 2 others, the number of ADCs in Figure 1 can be reducedto only two by adding a single 𝐼/𝑄 regeneration circuit at thethree outputs of the FPD or TPD. If we arbitrarily chooseV4(𝑡) as the symmetry axis, these 2 conditions can be stated

as follows:

(1) amplitude symmetry condition:

𝐴3= 𝐴5= 𝐴, 𝐵

3= 𝐵5= 𝐵, 𝐾

3= 𝐾5= 𝐾, (3)

(2) phase symmetry condition:

𝜙3= 𝐶0+ 𝛼, 𝜙

4= 𝐶0, 𝜙

5= 𝐶0− 𝛼, (4)

where 𝐶0is a phase shift of V

4(𝑡) whose effect is a rotation of

the 𝐼-𝑄 phase diagram of an angle equal to𝐶0. For simplicity,

we will consider that 𝐶0= 0.

Now, if the amplitude symmetry condition is extended toV4(𝑡), that is,

𝐴3= 𝐴4= 𝐴5= 𝐴, 𝐵

3= 𝐵4= 𝐵5= 𝐵,

𝐾3= 𝐾4= 𝐾5= 𝐾,

(5)

−−

−

++

+

�3(t)

�4(t)

�5(t)

IOUT (t)

QOUT(t)

R1

R1

R1

R1 R1

R1

R2

R2

R2

R2/2

Figure 2: Architecture of the baseband analog 𝐼/𝑄 regenerationcircuit.

then, it can be shown that the actual 𝐼(𝑡) and𝑄(𝑡) componentsare [26, 27]

𝐼 (𝑡) = 𝜇𝐼 [−V3 (𝑡) + 2V4 (𝑡) − V5 (𝑡)] = 𝜇𝐼𝐼OUT (𝑡) ,

𝑄 (𝑡) = 𝜇𝑄 [V3 (𝑡) − V5 (𝑡)] = 𝜇𝑄𝑄OUT (𝑡)(6)

with

𝜇𝐼=

1

{2𝐾𝑉LO𝐴𝐵 [1 − cos (𝛼)]},

𝜇𝑄= tan(𝛼

2) ⋅ 𝜇𝐼.

(7)

The 𝐼/𝑄 regeneration circuit whose architecture is shownin Figure 2 realizes the sums and differences of the outputvoltages V

𝑖(𝑡) to compute 𝐼OUT(𝑡) and 𝑄OUT(𝑡) in (6). These

are scaled versions of the actual 𝐼(𝑡) and𝑄(𝑡) components, soan equalization procedure is needed.

However, it is shown in [26] that the quadrature of 𝐼(𝑡)and 𝑄(𝑡) is ensured as long as the two symmetry conditionsstated in (3) and (4) remain fulfilled, independently of thevalue of 𝛼, so that no phase compensation has to be applied.On the other hand, the amplitude of 𝐼(𝑡) and𝑄(𝑡) is functionof 𝛼 and they differ from each other except when 𝛼 =90∘ and (5) are fulfilled. In this particular case, 𝐼OUT(𝑡) and

𝑄OUT(𝑡) have the same amplitude and since quadrature isalways satisfied, the actual 𝐼(𝑡) and 𝑄(𝑡) components aredirectly recovered from 𝐼OUT(𝑡) and 𝑄OUT(𝑡) without anyamplitude/phase compensation, and therefore, no equal-ization procedure is required if the transmitter is directlyconnected to the receiver. However, in a real communicationlink, equalization is still needed due to the degradation ofRF propagation channel. Furthermore, amplitude symmetrycondition in (5) allows the 𝐼/𝑄 regeneration circuit tosuppress DC offsets and IMD2 products in both 𝐼 and 𝑄paths. Indeed, if the 3 paths of the FPD or TPD are perfectlyidentical, the DC

𝑖and IMD2

𝑖(𝑡) in (2) are canceled when

V𝑖(𝑡) is substituted into (6). The Analog Devices AD8056A

OP-amp circuit was used in the fabricated prototype and thevalues of the resistors were chosen in order to obtain a unityvoltage gain in each of the 𝐼 and 𝑄 paths, as in [26].


RF

LO

Herotek DT1018P tunnel diode

detectors

Microstrippower dividers

Anaren 430203-way dividers

I/Q regenerationcircuit

Baseband

IOUT (t)

QOUT(t)

(a)

RF

LO

MeuroMMB0050050F0020

active mixers

Anaren 430203-way dividers

I/Q regenerationcircuit

Baseband

IOUT(t)

QOUT(t)

(b)

RF

MeuroMMB0050050F0020

active mixers

MinicircuitsZAPD-4

power divider

LO

MinicircuitsZAPDQ-4

divider 90

∘ power

IOUT (t)

QOUT (t)

(c)

Figure 3: Schematic of (a) FPD and (b) TPD architectures with baseband 𝐼/𝑄 regeneration circuit (c) IQD.

4. Implementation ofFive-Port (FPD), Three-Phase (TPD),and 𝐼/𝑄 Demodulators (IQD)

Figure 3 shows a schematic of FPD and TPD with basebandanalog OP-amp circuit for 𝐼/𝑄 regeneration as well as IQDarchitecture. For FPD and TPD, RF and LO signals are splitby means of Anaren 43020 3-way power dividers showinginsertion loss below 0.4 dB in each branch and isolation betterthan 23 dB between 2GHz and 3.6GHz. FPD architectureimplies the use of 1 GHz to 4GHz microstrip power dividersfabricated in our lab on printed circuit boards (PCBs) in orderto combine the RF and LO signals to deliver to the powerdetectors.These are based onHerotekDT1018P tunnel diodesshowing tangential signal sensitivity (TSS) of −50 dBm andgain flatness of ±1 dB from 1GHz to 18GHz. For the TPD andIQD, the downconversion operation is performed by meansof MeuroMMB0050050F0020 active mixers exhibiting 10 dBconversion gain and above 30 dB LO-to-RF isolation from0.45GHz to 3.6GHz.

In the IQD, RF signal is split by a Mini-Circuits ZAPD-4power divider while LO signal is split and 90∘ phase shiftedby aMini-Circuits ZAPDQ-4 showing less than 7∘ offset from90∘ ideal quadrature in the 2–3.6GHz frequency range.

5. Measurement Results

5.1. 𝐼/𝑄 Amplitude and Phase Imbalance. First, in order toverify the symmetry properties of the implemented FPD andTPD without 𝐼/𝑄 regeneration circuit, amplitude and phaseof the three low frequency voltages V

𝑖(𝑡)weremeasured at the

outputs of the two architectures. Two continuous-wave (CW)signals with Δ𝑓 = 10 kHz separation were applied to the RFand LO input ports of the circuits. Agilent EXG N5172B andE4432B signal generators were used for this purpose with RFpower equal to −25 dBm and LO power level set to 5 dBmwhile the three output voltages V

𝑖(𝑡)were measured bymeans

of Agilent 54622D oscilloscope. Figure 4 shows themeasuredoutput voltage amplitudes for FPD and TPD architectures inthe 2–3.6GHz frequency range.

As expected, the output voltage amplitudes are higher inthe case of TPD due to the use of active mixers instead ofdiode detectors, in FPD. Amplitude balance between outputvoltages V

3(𝑡), V4(𝑡), and V

5(𝑡) is better for TPD than FPD,

especially for frequencies above 3.2GHz due to mismatchbetween diode detectors used in FPD architecture. Indeed,amplitude imbalance for TPD does not exceed 1 dB from2GHz to 3.6GHz showing slight performance degradationbetween 3GHz and 3.4GHz. On the other hand, amplitude


2 2.42.2 2.6 2.8 3 3.2 3.4 3.60

20

40

60

80

100

120

140

160

180

200

Frequency (GHz)

Am

plitu

de (m

V)

(FPD)(FPD)(FPD)

(TPD)(TPD)(TPD)

�3

�4

�5

�3

�4

�5

Figure 4: Measured output voltage amplitudes for FPD and TPDarchitectures without 𝐼/𝑄 regeneration circuit (𝑃LO = 5 dBm, PRF =−25 dBm, and Δ𝑓 = 10 kHz).

imbalance for FPD remains below 2.5 dB from 2GHz to3.2GHz but is up to 5 dB from 3.2GHz to 3.6GHz. Never-theless, amplitude imbalance between V

3(𝑡) and V

5(𝑡) is kept

below 3.2 dB for FPD. Figure 5 shows phase shift of V3(𝑡) and

V5(𝑡) with respect to V

4(𝑡) (Φ

3− Φ4and Φ

4− Φ5) and phase

difference between 𝐼(𝑡) and 𝑄(𝑡) when the 𝐼/𝑄 regenerationcircuit is connected to FPD or TPD. For comparison, 𝐼/𝑄phase difference is also plotted for the IQD. Phase imbalancebetween V

3(𝑡) and V

5(𝑡) with respect to V

4(𝑡) is maintained

below 10∘ and 4∘ for FPD and TPD, respectively, exceptbetween 3GHz and 3.3GHz for the latter. In this frequencyrange, the high discrepancy betweenΦ

3−Φ4andΦ

4−Φ5(up

to 25∘) originates from mismatch among the active mixers inTPD.This is quite consistentwithwhat is observed in Figure 4from 3GHz to 3.4GHz in terms of components mismatch.As for the absolute value of Φ

3− Φ4and Φ

4− Φ5, which

corresponds to the parameter 𝛼 introduced in Section 3, itvaries approximately from 60∘ to 125∘ between 2GHz and3.6GHz for both architectures with ideal value of 90∘ at thecenter frequency of 2.8GHz. This results in 90∘ ± 10∘ phasedifference between 𝐼OUT(𝑡) and 𝑄OUT(𝑡) signals at the outputof the 𝐼/𝑄 regeneration circuit with perfect 90∘ quadratureat 2.8GHz, in accordance with theoretical results. As for theIQD, phase difference between 𝐼(𝑡) and 𝑄(𝑡) is 90∘ ± 7∘,showing lower fluctuations around ideal 90∘ value than FPDand TPD. This is probably due to the fact that the IQD wasmade with 2 mixers showing better matching between eachother among the 3 available mixers.

To demonstrate the sensitivity of the IQD to imbalancebetween its 𝐼 and 𝑄 paths, the 𝐼/𝑄 phase difference is alsoplotted when the IQD is built with a mixer pair showinghigher mismatch. In this configuration, phase differencebetween 𝐼(𝑡) and 𝑄(𝑡) is 100∘ ± 15∘, which features an

important degradation of phase quadrature; while, close toideal 90∘, it can still be achieved with the TPD despite theuse of onemixer showing poormatching with the two others,hence making evidence of the robustness of the TPD leadingto imbalance.

Measured 𝐼OUT(𝑡) and 𝑄OUT(𝑡) signal amplitudes at theoutput of the 𝐼/𝑄 regeneration circuit and the IQD areplotted in Figure 6. Amplitude balance better than 1 dB (12%amplitude imbalance) is obtained for the IQD across the 2–3.6GHz frequency range as mixer pair exhibits good match-ing. Amplitude imbalance between 𝐼OUT(𝑡) and 𝑄OUT(𝑡)signals remains below 1 dB from 2.8GHz to 3.4GHz forFPD and from 2.6GHz to 3GHz for TPD. Accordingly, verysmall distortion of the constellation should be observed forthese frequencies when a modulated RF signal is applied atthe input of the FPD or TPD and equalization procedurecould then be avoided. This is consistent with theoreticalresults in the case of TPD since 𝛼 = 90∘ and amplitudesymmetry stated in (5) are almost verified between 2.7GHzand 2.9GHz but this is quite unexpected in the case ofthe FPD. Indeed, good 𝐼OUT(𝑡)/𝑄OUT(𝑡) amplitude balanceis obtained from 3.2GHz to 3.4GHz while 𝛼 is beyond 110∘and high amplitude imbalance is measured between V

3(𝑡),

V4(𝑡), and V

5(𝑡) as shown in Figure 4. This point is still under

investigation as correct operation of the 𝐼/𝑄 regenerationcircuit was verified by applying 3 low frequency voltages atits inputs and measuring quasi null output voltages at 𝐼OUTand 𝑄OUT (cf. Figure 2).

5.2. DC Offsets and IMD2 Suppression. As was mentionedin Section 3, the use of the 𝐼/𝑄 regeneration circuit allowsDC offsets and IMD2 products to be suppressed if voltagesV3(𝑡), V4(𝑡), and V

5(𝑡) have same amplitude which is stated in

(5). In order to verify this property, we measured the ratiobetween the DC offsets component and the desired signalat the three outputs of the FPD and TPD and then at theoutputs of the 𝐼/𝑄 regeneration circuit when connected toboth architectures. The measurement setup is the same asfor the 𝐼/𝑄 amplitude and phase imbalance determination;that is, two CW signals with Δ𝑓 = 10 kHz separationwere applied to the circuit with RF and LO power equal to−25 dBm and 5 dBm, respectively, while the output voltageswere measured bymeans of an oscilloscope.

Documents

Multiport Technology: New Perspectives and …downloads.hindawi.com/journals/specialissues/961060.pdfJournal of Electrical and Computer Engineering Multiport Technology: New Perspectives