Abstract—A new systematic methodology for the design ofT-match based UHF-RFID tags is proposed. The great majority ofcommercial UHF-RFID tags are based on dipole antennas usinga modification of a T-match network. The literature containsexamples of models that describe the T-match, but they are notsufficiently accurate to synthesize the tag geometry. Using theproposed methodology, a global band UHF-RFID tag based on afolded dipole antenna and matched to the RFID integrated circuitby means of a T-match network is designed and fabricated. Verygood agreement between the measured and simulated read rangeis achieved within the entire UHF-RFID band, which revealsthat the proposed method is amenable for accurate analysis andsynthesis of T-match based UHF-RFID tags.
Index Terms—antennas, radio frequency identification (RFID),tags, T-match network.
I. INTRODUCTION
R ADIO FREQUENCY IDENTIFICATION (RFID) is arapidly developing technology that provides wireless
identification and tracking capability. Particularly, passiveultra-high frequency (UHF) RFID systems are very attractivein comparison with passive RFID regulated systems using lowfrequency (LF) and high frequency (HF) bands, since theycan provide superior read range, fast reading and enhancedinformation storage ability [1]. The regulated UHF-RFIDbands vary in the different world regions, including frequenciesbetween 840 MHz and 960 MHz. More specifically, RFIDis operated at 840–845 MHz in China, at 866–869 MHz inEurope, at 902–928 MHz in USA and at 950–956 MHz inJapan. Therefore, the design of inlay tags able to cover thewhole regulated UHF bands (i.e., global band tags) becomes animportant challenge.A passive UHF-RFID system consists of a reader and a tag,
which includes an antenna matched to an application specificintegrated circuit (ASIC) chip. Generally, the chip impedanceis capacitive, thereby requiring the antenna impedance to beinductive in order to obtain a proper impedance matching
Manuscript received June 11, 2013; revised October 18, 2013; acceptedOctober 20, 2013. Date of publication November 05, 2013; date of currentversion December 02, 2013. This work was supported in part by Spain MICIINunder projects CONSOLIDER CSD2008-00066 and METATRANSFERTEC2010-17512, and by AGAUR (Generalitat de Catalunya) through theproject 2009SGR-421.The authors are with GEMMA/CIMITEC Departament d’Enginyeria Elec-
trònica, Universitat Autònoma de Barcelona. 08193 Barcelona, Spain (e-mail:[email protected]).Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.Digital Object Identifier 10.1109/TMTT.2013.2287856
Fig. 1. Example of a UHF-RFID commercial tag based on the T-match network(Alien ALN-9640).
(conjugate matching). Several techniques for achieving conju-gate matching can be found in the literature [2], [3]. However,most commercial UHF-RFID tags are based on dipole antennasusing some variant of a T- match network [4]–[6]. The T-matchconnection was first proposed by Uda [7], and more recentlyexplained in [8] as an effective shunt-matching technique.Although it was initially analyzed as a general form of acylindrical folded dipole, the Uda model has been applied todesign planar structures [9]–[11] and even RFID tags [2], [4],[5]. However, many approximations are assumed when planarconductor shapes are considered [8]. Moreover, by this means,only a special case of the T-match structure can be used todesign RFID tags, the embedded T-match, that is constructedby embedding the T-match structure into the antenna [6].This is the main drawback since most commercially availableT-match based tags have more complex geometries whichcannot be analyzed by means of this planar model. An exampleof these tags is depicted in Fig. 1, where the T-match networkhas a loop shape and is located at the center of the tag. In orderto overcome this problem, some efforts to deviate from Udaclassic analysis and focus on a circuit-based approach havebeen made [5], [12], [13]. In these works, equivalent-circuitmodels of a dipole antenna (only valid over a relatively smallfrequency range near resonance) matched to the chip by meansof a T-match have been developed. Nevertheless, the synthesisprocess of the tags from the obtained circuits is not fullyexplained.In this paper, a new and very simple systematic method for
the design and synthesis of global band UHF-RFID tags basedon the T-match network is presented. Such method is based ona new equivalent-circuit model approach. Moreover, the fre-quency limits related to the validity of the presented approachare studied, and the dependence of the achieved tag bandwidthwith the antenna impedance is also discussed. To illustrate thepotential of the approach, an RFID tag is designed and syn-thetized through this method and the read range of the fabricatedprototype is measured.
ZAMORA et al.: DESIGN AND SYNTHESIS METHODOLOGY FOR UHF-RFID TAGS 4091
Fig. 2. (a) Balanced equivalent-circuit model of a T-match based UHF-RFIDtag from [13], and (b) unbalanced equivalent-circuit model reported in [5].
II. EQUIVALENT CIRCUIT MODEL OF T-MATCH BASED TAGS:NEW APPROACH AND REQUIREMENTS
The T-match structure, shown in Fig. 1, is the most commonmatching network used for the efficient matching of UHF-RFIDtags. Since this network is electrically small at the UHF-RFIDregulated bands (840–960 MHz), a lumped-element equivalent-circuit model can be considered. Moreover, it is well known thatthe input impedance of a UHF-RFID chip can be modeled by aparallel combination of a resistance and a capacitance[14], [15]. Thus, a circuit model for the T-match structure cas-caded to a commercial RFID chip can be obtained, which is ofspecial interest for tag design. Some efforts to obtain an equiva-lent-circuit model for the T-match network cascaded to the chipcan be found in the literature. In [13], the circuit diagram shownin Fig. 2(a) is proposed, whereas in [5] this balanced circuit is re-duced to an unbalanced version, depicted in Fig. 2(b). Althoughthis circuit is complete and reasonably accurate, the authors in[5] transformed thematching circuit from a series-shunt connec-tion of and into a shunt-series connection with a scaledload impedance. This allows reformulating the tag antenna andmatching circuit problem into a classical two stage bandpassfilter. The main drawback of such approach is the difficulty tosynthesize the tag antenna and matching circuit, once the circuitmodel is tuned. Moreover, the authors do not give details abouthow to synthesize the presented tags from the circuit model.In this work, due to the symmetry of T-match based tags
and the differential mode excitation, forced by the chip,
Fig. 3. Equivalent-circuit model of a T-match based UHF-RFID tag using theelectric wall concept.
Fig. 4. Equivalent-circuit model of one half of a T-match based UHF tag.
Fig. 5. Equivalent-circuit model of the circuit of Fig. 4.
the electric wall concept has been used in order to obtain anequivalent-circuit model (see Fig. 3). The required values of thecircuit elements can be obtained by considering only one-halfof the network (see Fig. 4). In this equivalent-circuit, theantenna impedance, , is modeling any generalimpedance. It can be observed that the circuit cascaded betweenthe chip and the antenna is an inductive transformer [16]. Thus,the circuit of Fig. 4 can be modeled by the circuit depicted inFig. 5. To demonstrate this, the admittancesand , given by (1)–(3), shown at thebottom of the page, are forced to be equal. Comparing the
(1)
(2)
(3)
4092 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 61, NO. 12, DECEMBER 2013
real part one obtains (4), shown at the bottom of the page. If
is satisfied, then expression (4) can be approximated by
(5)
where and is the antennaconductance. The imaginary parts of the admittances lead to (6),shown at the bottom of the page, which can be approximated by
(7)
providedand
(notice that the first condition is the same than theone required for expression 5). It can be demonstrated throughsimple algebra that these two inequalities are satisfied if
(8)
(9)
where . Hence, the circuit of Fig. 4 can be approx-imated by the circuit of Fig. 5, where and are obtainedfrom (5) and (7), respectively, as long as conditions (8) and (9)are satisfied.
III. DESIGN AND SYNTHESIS OF T-MATCH BASED TAGS
A systematic methodology for the design and synthesis ofa global band T-match based tag is presented in this section.Let us consider the circuit of Fig. 5, assuming that conditions(8) and (9) are well satisfied, as a model for T-match basedtags. The central frequency of the operating band is chosen tobe the intermediate frequency of the UHF-RFID band,
MHz. Then, in (5) must be equal to andin (7) must be equal to , where , toachieve complex conjugate matching at . It is clear that, fora given chip and , the inductances and can be easilycalculated from (5) and (7). However, in spite of the possibilityof satisfying conditions (8) and (9) by means of an antenna witha complex impedance, resonant antennas designed at arevery good candidates for using the proposed circuit approach.This is because, regardless of the considered chip, condition (9)will be easily satisfied in the vicinity of the antenna resonancefrequency since approaches zero. Therefore, the proposedequivalent-circuit model will predict the frequency response ofthe designed tag in a wider bandwidth when a resonant antennais considered.
Fig. 6. Frequency increment of the tag resonance (with respect to ) as a func-tion of the antenna resistance, , in the case of considering four different com-mercial chips. Notice that goes to zero when .
Let us consider the particular case of using a resonant tagantenna designed to exhibit a purely resistive impedance at. It is important to point out that even if conditions (8) and
(9) are very well satisfied, a frequency shift of the tag resonanceand a reduction of the matching level (with respect to conjugatematching) at this frequency are expected, as long as differsfrom , since and are exactly determined by (4) and(6) rather than (5) and (7), respectively. However, this frequencyshift can be avoided by taking it into account in the T-matchdesign stage, and the matching level at the tag resonance can bepredicted. Let us see how this is possible bymeans of an analysisof the power reflection coefficient of Fig. 5, given by [17], [18]
(10)
where is the chip admittance,and is the total susceptance of the cir-
cuit given by the sum of the susceptance of and two timesthe susceptance of the chip. Let us assume a constant resistivevalue for the antenna impedance such that in thecircuit of Fig. 4, and and from Fig. 5 are given by (4)and (6), respectively. By using (6), it can be demonstrated (seeAppendix A) that the susceptance vanishes at a frequency
, since . This corre-sponds to a frequency shift of the tag resonance, ,towards higher frequencies, which depends only upon the an-tenna resistance, for a given chip (see Appendix A). Such shift isdepicted in Fig. 6 in the case of considering four different com-
(4)
(6)
ZAMORA et al.: DESIGN AND SYNTHESIS METHODOLOGY FOR UHF-RFID TAGS 4093
mercial chips [19]–[21]. It can be seen that decreases asapproaches the chip resistance, , and becomes higher asmoves away (decreasing) from . (The case whenis not necessary to be discussed, since the presented method re-quires that ). Moreover, a reduction of the matchinglevel (with respect to conjugate matching) will take place at thisfrequency, as long as differs from . From (4) it followsthat the exact expression for evaluated at the tag resonance
(11)
is always less than or equal to , and becomes lower asmoves away (decreasing) from . Then, by evaluating (10)
at the tag resonance and introducing (11) into thisexpression, this matching level reduction can be inferred.Let us now consider any general complex antenna impedance
, designed to exhibit a real impedance valueat (being the shift related to this real impedancevalue). Obviously, the same power reflection coefficient asin the previous case (where a constant antenna impedancevalue was considered) will be achieved at , since
and is given by (11) at this frequency. Furthermore,it can be demonstrated from (10) that the minimum powerreflection coefficient occurs roughly at that frequency when
, even in the case of considering as a frequencydependent function, provided is close to at thatfrequency (see Appendix B). Notice that if conditions (8)and (9) are satisfied, is approximately given by (5) and,therefore, at the antenna resonance frequency,namely, . This can also be seen from (11), sinceapproaches as goes near . Thus, by forcing (5)to be at the antenna resonance frequency, , and
in (7), the tag resonance will be locatedat , and the matching level at this frequency canbe approximately inferred by introducing (11) into (10) andforcing and .
A. Design of a T-Match based Tag using a Resonant Antenna
From the above analysis, it can be concluded that the pro-posed method should be applied by means of a resonant antennadesigned at and the T-match network designed by forcing
at in (5) and in (7),in order to obtain the tag resonance at the desired frequency,(notice that a linear approximation of with respect to hasbeen considered, since ). The shift is obtainedfrom that curve corresponding to a given chip (see Fig. 6), eval-uated at . Then, from (5), the factor can berewritten in terms of the chip resistance, , and the antennaresistance evaluated at , given
. Thus, the inductances and can be cal-culated from (5) and (7) and are obtained as
(12)
(13)
Then, the required conditions (8) and (9) lead to
(14)
(15)
where, it has been considered and ,since variations of relative to are less than 7% withinthe whole UHF-RFID band and . Notice that con-dition (15) does not play any role to achieve complex conjugatematching at , since at this frequency (this conditionwill be used in Section III.C to discuss the frequency range ofvalidity of the proposed circuit approach). However, expression(14) lead to a simple condition, in which the required antennaresistance depends only upon the RFID chip and the in-termediate frequency as . Therefore,the greater the value of , the better satisfied the requiredcondition. However, since is forced to be at , as in-dicated before, it follows from (5) that must be less thanor equal to . Hence, it can be concluded that the presentedapproach is valid at the intermediate frequency, if the antennaresistance accomplishes
(16)
Notice that condition (16) forces , which iswell satisfied by the typical values of the RFID integrated cir-cuits available on the market today, such as Impinj Monza 5,Impinj Monza X-2K Dura, Alien Higgs 3, Alien Higgs 4 andNXP UCODE G2XM [19]–[21].
B. Tag Bandwidth Related to the Antenna Impedance
Let us now demonstrate that a degradation of the maximumachievable tag bandwidth, by means of the proposed equiva-lent-circuit approach and considering conjugate matching at ,will be mainly determined by the derivative of the antenna re-sistance at . It was demonstrated in [22] that the optimumequivalent-circuit network necessary for bandwidth broadeningin single resonant UHF-RFID tags with conjugate matching isa parallel combination of an inductor and a resistor cascadedto the chip, according to the Bode’s limit [23], [24]. However,the proposed circuit approach consists of a parallel combinationof an inductor and a frequency dependent resistor, cascaded tothe chip. Hence, bandwidth degradation with respect to the op-timum will be obtained as long as the conductance differsfrom . In a first order approximation, this reduction of thetag bandwidth is determined by the frequency derivative of theantenna resistance at , and it does not depend on the frequencyderivative of the antenna reactance , since this term is can-celled. This result is deduced from the first-order Taylor expan-sion of the conductance , obtained from (5), in the vicinityof
(17)
where is the frequency derivative of the antenna resis-tance evaluated at . Notice that, for a given value of ,
4094 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 61, NO. 12, DECEMBER 2013
the further is from zero, the further is from ,and consequently a higher degradation of the tag bandwidth willbe obtained. By introducing (17) into (10) and expanding thesusceptance (using , theapproximated bandwidth at a fixed value for the power reflec-tion coefficient can be inferred, within the frequency range ofvalidity of the proposed circuit approach.
C. Frequency Range of Validity of the Proposed Approach
Let us now focus on the validity of the presented approachbeyond the tag resonance frequency. As it has been pointed out,a T-match based tag designed following the steps indicated inSection III.A will exhibit a frequency response centered at thedesired frequency . Therefore, such a response will be sim-ilar to that of the proposed equivalent-circuit of Fig. 5, using(5) and (7) and designing the T-match network at . It followsthat, in order to determine the frequency range of validity of theproposed approach, a comparison between these two frequencyresponses makes sense. Then, assuming that the tag antenna sat-isfies the required condition at the resonance frequency (see ex-pression 16), an examination of condition (14) reveals that itwill be satisfied within the whole UHF-RFID band provided
. It can be easily demonstrated that this condi-tion holds true in the case of considering a canonic RLC seriesload as antenna impedance. Thus, in this particular case, condi-tion (14) is accomplished in all frequencies and, consequently,the frequency limits around the tag resonance from which theproposed approach no longer predicts the frequency responseof the designed tag are determined by (15). Therefore, in orderto satisfy such condition within the entire UHF-RFID band, thefrequency derivative of must be small. Conversely, if the an-tenna impedance can be approximated by an RLC shunt loadaround the resonance frequency, both expressions (14) and (15)must be taken into account in order to obtain the frequencylimits from which the proposed approach becomes invalid. Thisis exactly what happens by considering any general frequencydependent complex antenna impedance. Thus, in such cases,both the frequency derivatives of and should be smallto enhance the frequency range of validity.
D. Synthesis of a T-Match Based Tag Using a ResonantAntenna
A simple method for the synthesis of T-match based tagsusing a resonant antenna is proposed in this section. In orderto synthesize the T-match network, we start by considering aclosed loop consisting of a narrow conductor strip (e.g., 0.2 mmwidth) connected to a differential port with impedance ,as depicted in Fig. 7. This loop exhibits an inductive behaviorat the UHF-RFID frequency band. To achieve the required di-mensions for the loop inductance, a sweep of the length of theloop is carried out by means of electromagnetic simulations (byusing the Agilent Momentum commercial software), until theresonance frequency appears at . Then, the inductancecorresponding to one half of the loop must be divided intoand by connecting the antenna at the adequate position (seeFig. 7). To this end, a sweep of the position of the antenna con-nection is carried out to achieve the calculated values for these
Fig. 7. Model of an UHF-RFID T-match based tag.
inductances. The final position is obtained when the tag reso-nance reaches the desired frequency .
IV. DESIGN OF A BROADBAND UHF-RFID TAG USING THEPROPOSED METHOD
Let us now consider a typical integrated circuit for the RFIDtag (the NXP UCODE G2XM chip). The impedance reportedby the manufacturer of this integrated circuit is
, at 915 MHz. As indicated in Section II, the chip canbe modeled by a parallel combination of a resistance anda capacitance . These values were calculated from the inputimpedance of the chip transformed to its equivalent RC parallelcircuit, giving and pF. From (16) itfollows that . Thus, the tag antennahas to be designed to exhibit an antenna resistance at the op-erating frequency according to (16). As a proof of concept forthe presented method, an antenna for tag implementation wasdesigned on a commercial low loss microwave substrate, theRogers RO3010 substrate with dielectric constantand thickness mm. As it was pointed out, the use ofa resonant antenna becomes appropriate to ensure the validity ofthe presented approach around the operating frequency . Theproposed antenna is a meandered coplanar strip, folded dipoleworking at the so called antenna mode [25]–[28], which allowsthe antenna impedance (at least the imaginary part) to be ap-proximated by the canonic RLC series load, around the antennaresonance frequency . The designed antenna exhibits a purelyresistive impedance at the intermediate frequency,
(see Fig. 8). Then, in order to design the T-match networkat the correct frequency , the frequency shift was in-ferred by using expression (A4) and it was found to be 8 MHz.This result perfectly agrees with the shift obtained in Fig. 9,where a simulation (by means of the Agilent ADS circuit simu-lator) of the power reflection coefficient of the circuit of Fig. 5,using (5) and (7) and designing the T-match network at , witha sweep of the antenna impedance from 20 to1385 , is depicted. As expected in this analysis, perfect con-jugate matching at the desired frequency is achieved when
ZAMORA et al.: DESIGN AND SYNTHESIS METHODOLOGY FOR UHF-RFID TAGS 4095
Fig. 8. Input impedance of the designed tag antenna.
Fig. 9. Simulated power reflection coefficient of the circuit of Fig. 4, bysweeping the antenna impedance from 20 to 1385 .
and, consequently, and .This is because in this case, approximation (7) becomes an exactexpression for and . As it was previouslypredicted, a frequency shift of the tag resonance from to-ward higher frequencies and a reduction of the matching levelat this frequency are observed as decreases from . It canbe seen in Fig. 9 that such frequency shift and matching levelreduction at the resonance frequency become more significantas the value of moves away from . Thus, the T-match net-work was designed at 890 MHz, which corresponds to 8 MHzbelow . From (12) and (13), the elements of the T-match net-work where found to be nH and nH. Then,the synthesis method explained in the previous section was ap-plied. The layout of the designed tag is depicted in Fig. 10(a).The dimensions are mm, mm, mm,
mm, mm, mm and mm. Allthe strips of the antenna have the same width (3 mm) and thewidth of the T-match network is 0.2 mm. The total length of theT-match closed loop was found to be 30 mm, and the tag an-tenna was connected to the loop at a distance of 10.2 mm fromthe chip.
Fig. 10. (a) Layout of the designed T-match based tag. (b) Electrical simulationof the return loss of the proposed equivalent-circuit (dash dot line) and electro-magnetic simulation of the return loss of the designed tag (solid line).
The power reflection coefficient of the designed RFID tag isdepicted in Fig. 10(b). It can be seen that conjugate matchingis achieved at the intermediate frequency of the UHF-RFIDfrequency band. Very good agreement is observed between thepower reflection coefficient obtained from the equivalent-circuitapproach and the electromagnetic simulation within the entireUHF-RFID band except in the low frequency region. Althoughthe tag antenna exhibits purely resistive impedance at the reso-nance frequency , the antenna resistance is not constant withfrequency, as it is shown in Fig. 8. Hence, both conditions (14)and (15)must be examined to discuss the validity of the approachapplied in this work within the UHF-RFID band, as it wasmentioned in Section III.C. For the considered chip and antennaimpedance, we obtain and . Thus, thecondition for the antenna reactance is satisfied between 867MHzand 930 MHz. However, although the condition for the absolutevalue of the antenna impedance is well satisfied at frequencieshigher than and in the vicinity, there is a frequency regionwithin the low UHF-RFID band where such condition is lesssatisfied. Therefore, good agreement is expected between thepower reflection coefficients of the equivalent-circuit approachand the electromagnetic response of the designed tag within thewhole UHF-RFID band, except in the low frequency region.As indicated in Section III.B, by introducing (17) into (10),
the approximated bandwidth at a fixed value for the power re-flection coefficient can be predicted, within the frequency rangeof validity of the proposed circuit approach. Thus, the dBbandwidth was found to be 35 MHz, which is similar to the ob-tained by means of the electromagnetic simulation of the de-signed tag (39 MHz). The simulated gain reaches the value of1.8 dBi at the operating frequency and the radiation patternis similar to that of a conventional dipole in the wholeUHF-RFID band, as it is shown in Fig. 11.
4096 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 61, NO. 12, DECEMBER 2013
Fig. 11. (a) Electric plane radiation pattern of the designed T-match based tagand (b) magnetic plane. The proposed tag has a radiation pattern similar to thatof a conventional dipole in the whole UHF-RFID band.
V. FABRICATION AND MEASUREMENT
A. Theoretical Read Range
To determine the performance of the tags, it is necessary toobtain the read range [29], which can be calculated using theFriis free space formula as
(18)
where is the wavelength and EIRP is the equivalent isotrop-ically radiated power, determined by local country regulations(e.g., 3.3 W in Europe and 4 W in USA). is the minimumthreshold power necessary to activate the RFID chip, is thegain of the receiving tag antenna, and is the power transmis-sion coefficient, which is related to the power reflection coef-ficient by . The tag gain and the powertransmission coefficient are inferred from simulation, using asport impedance that of the chip. Despite the minimum powerlevel necessary to activate the chip used in this work reportedby the manufacturer is dBm, a different thresholdpower was obtained in [30]. In such work it was found thatexhibits a lower and frequency dependent value. Taken it intoaccount and from electromagnetic and circuit co-simulation re-sults, an evaluation of the theoretical read range was obtainedand depicted in Fig. 12(b).
Fig. 12. (a) Photograph of the fabricated RFID tag and (b) simulated and mea-sured read range. The measured read ranges of the UPM Web tag (that uses thetag chip of our prototype), and the Alien ALN-9640 tag of Fig. 1, are also shownfor comparison purposes.
B. Measured Read Range and Experimental Setup
The RFID setup available in our laboratory has anvector signal generator, which creates
RFID frames and plays the role of a reader with variable fre-quency and variable output power. Such generator is connectedto a TEM cell by means of a circulator. The tag under test is lo-cated inside the TEM cell and it is excited by the frame createdby the generator. Then the tag sends a backscatter signal to anAgilent N9020A signal analyzer through the circulator. At eachfrequency, the minimum power at the input of the TEM cellrequired to communicate with the tag is recorded. Finally, anelectric probe is placed into the TEM cell to determine the rootmean square of the incident electric field, , correspondingto the minimum power at each frequency. This electric field isrelated to the power delivered to the chip according to
(19)
where is the incident power density, is the effective areaof the tag antenna, and is the wave impedance of free space.The measured read range can be inferred by introducing (19)into (18), resulting the following expression
(20)
C. Experimental Results
The proposed RFID tag was fabricated and the read range wasmeasured (see Fig. 12) through the procedure explained above.Very good agreement between the theoretical andmeasured readranges can be observed. The fabricated tag exhibits a significantread range within the whole UHF-RFID band (840–960 MHz),with a peak value of 11 m at 898 MHz. The read range of
ZAMORA et al.: DESIGN AND SYNTHESIS METHODOLOGY FOR UHF-RFID TAGS 4097
(A4)
a commercially available tag (UPM Web) that uses the NXPUCODE G2XM chip and the T-match network is also shownin Fig. 12(b) for comparison purposes. It can be seen that ourfabricated tag exhibits a substantially superior read range inthe whole UHF-RFID band. Despite the fact that comparingRFID tags only makes sense if they use the same chip (thechip impedance and are key parameters in determiningthe read range), we have also included in Fig. 12(b) the readrange of the commercial tag Alien ALN-9640 (Fig. 1). Such taguses a chip (the Alien Higgs 3) with dBm, whereas
dBm for the NXP UCODE G2XM chip (accordingto the manufacturer specifications). In spite of this significantdifference in the threshold power, the read range at 898 MHzis comparable in both the ALN-9640 tag and our proposed tag.Therefore, the proposed approach for the design of global bandUHF-RFID tags is simple and competitive in terms of the mainfigure of merit: the read range.
VI. CONCLUSION
In this paper, a systematic and simple method for the de-sign of UHF-RFID tags, based on the T-match network, hasbeen introduced. This method is based on a new equivalent-circuit model for the RFID tag that includes the tag antenna,the chip and the matching network. The main advantage of thismethod, in comparison with the methods reported in the lit-erature, is the simplicity in synthesizing the T-match networkrequired to achieve a broad band frequency response with con-jugate matching between the chip and the antenna. Neverthe-less, such antenna must be previously designed to exhibit aself-resonance at . As a proof of concept, a global band taghas been designed using this method, and the read range ofthe fabricated prototype has been measured and compared tothose of commercially available tags. The results reveal that thedesigned tag is very competitive, and point out that the pro-posed circuit-based approach is very useful for the synthesis ofT-match based tags.
APPENDIX A
A. Calculation of the Frequency shift of the Tag Resonance
Let us consider the circuit of Fig. 4, where the tag antenna hasa purely resistive impedance such that . Thiscircuit can be exactly modeled by the one depicted in Fig. 5,where and are given by (4) and (6), respectively. Now,if we use (5) and (7) in the circuit of Fig. 5 rather than (4) and(6), one can obtain the inductances and
(A1)
(A2)
It follows that the exact expression for the inductance in thecircuit of Fig. 5 can be expressed as
(A3)
where . Hence, it is clear from (A3) thatand, therefore, the susceptance vanishes ata frequency . This frequency can be inferred by forcing
and using (A3), giving a frequency increment, , ofthe tag resonant with respect to which can be written as (A4),shown at the top of the page.
B. Minimum Power Reflection Coefficient
Let us now consider the circuit depicted in Fig. 4, where thetag antenna has any general complex impedanceand being an arbitrary frequency dependent function. Thiscircuit can be exactly modeled by the one depicted in Fig. 5,where and are given by (4) and (6), respectively.Then, also results in an arbitrary frequencydependent function. Let us demonstrate that in the case when
, the minimum power reflection coefficientoccurs roughly at that frequency when the total susceptance
vanishes. The power reflection coefficient,, from the circuit of Fig. 5 is given by expression (10). By
forcing the frequency derivative of (10) to be zero, it is foundthat
(B1)
where and are the frequency derivatives of the suscep-tance and the conductance , respectively. Notice that allthe parameters in (B1) are frequency dependent functions, ex-cept . It can be deduced that (B1) is satisfied at that fre-quency when , provided that .
REFERENCES[1] K. Finkenzeller, RFID Handbook: Radio-Frequency Identification
Fundamentals and Applications, 2nd ed. New York, NY, USA:Wiley, 2004.
[2] G. Marrocco, “The art of UHF-RFID antenna design: impedancematching and size-reduction techniques,” IEEE Antennas Propag.Mag., vol. 50, no. 1, pp. 66–79, Feb. 2008.
[3] F. Paredes, G. Zamora, J. Bonache, and F. Martin, “Dual-bandimpedance-matching networks based on split-ring resonators for ap-plications in RF identification (RFID),” IEEE Trans. Microw. TheoryTech., vol. 58, no. 4, pp. 1159–1166, Apr. 2010.
[4] J. Choo, J. Ryoo, J. Hong, H. Jeon, C. Choi, and M. M. Tentzeris,“T-matching networks for the efficient matching of practical RFIDtags,” in Proc. Eur. Microw. Conf., 2009, pp. 5–8.
[5] D. D. Deavours, “Analysis and design of wideband passive UHF-RFIDtags using a circuit model,” in Proc. IEEE Int. Conf. RFID, May 2009,pp. 283–290.
[6] N. A. Mohamed, K. R. Demarest, and D. D. Deavours, “Analysis andsynthesis of UHF RFID antennas using the embedded T-match,” inProc. IEEE Int. Conf. RFID, Apr. 2010, pp. 230–236.
4098 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 61, NO. 12, DECEMBER 2013
[7] S. Uda and Y. Mushiake, Yagi-Uda Antenna. Tohoku University: Re-search Institute of Electrical Communication 1954.
[8] C. A. Balanis, Antenna Theory: Analysis and Design, 3rd ed. NewYork, NY, USA: Wiley, 2005.
[9] R. W. Lampe, “Design formulas for an asymmetric coplanar stripfolded dipole,” IEEE Trans. Antennas Propag., vol. AP-33, no. 9, pp.1028–1031, 1985.
[10] W. Lampe, “Design formulas for an asymmetric coplanar strip foldeddipole,” IEEE Trans. Antennas Propag., vol. AP-34, no. 4, p. 611,1986, Correction to.
[11] H. J. Visser, “Improved design equation for asymmetric coplanar stripfolded dipoles on a dielectric slab,” in Proc. Antennas Propag. Int.Symp., 2007, pp. 1–6.
[12] D. M. Dobkin and S. Weigand, “Environmental effects on RFID tagantennas,” in Proc. IEEE MTT-S Int. Microw. Symp., Jun. 2005, pp.135–138.
[13] D. M. Dobkin, “The RF in RFID: Passive UHF RFID in Practice,”.Newnes, 2007.
[14] E. Bergeret, J. Gaubert, P. Pannier, and J. M. Gaultier, “Modeling anddesign of CMOSUHF voltage multiplier for RFID in a EEPROM com-patible process,” IEEE Trans. Circuits Systems I, Reg. Papers, vol. 54,pp. 833–837, Oct. 2007.
[15] G. De Vita and G. Iannaccome, “Design criteria for the RF section ofUHF and microwave passive RFID transponders,” IEEE Trans. Mi-crowave Theory Tech., vol. 53, no. 9, pp. 2978–2990, Sep. 2005.
[16] P. H. Young, Electronic Communication Techniques, 3rd ed. NewYork, NY, USA: Macmillan Publishing Company, 1994.
[17] K. Kurokawa, “Power waves and the scattering matrix,” IEEE Trans.Microw. Theory Tech., vol. MTT-13, no. 3, pp. 194–202, Mar. 1965.
[18] P. V. Nikitin, K. V. S. Rao, S. F. Lam, V. Pillai, R. Martinez, and H.Heinrich, “Power reflection coefficient analysis for complex imped-ances in RFID tag design,” IEEE Trans. Antennas Propagation, vol.53, pp. 2721–2725, Sep. 2005.
[19] Impinj RFID chips [Online]. Available: http://www.impinj.com[20] Alien Technology RFID ICs [Online]. Available: http://www.alien-
technology.com[21] NXP UCODE smart label ICs [Online]. Available: www.nxp.com[22] G. Zamora, F. Paredes, F. J. Herraiz-Martinez, F. Martin, and J.
Bonache, “Bandwidth limitations of ultra high frequency-radio fre-quency identification tags,” IET Microwaves Antennas Propag., vol.7, no. 10, pp. 788–794, 2013.
[23] H. W. Bode, Network Analysis and Feedback Amplifier Design. NewYork, NY, USA: Van Nostrand, 1945, pp. 360–371.
[24] R. M. Fano, “Theoretical limitations on the broadband matching ofarbitrary impedances,” J. Frank1in Inst., vol. 249, pp. 57–154, 1950.
[25] W. L. Weeks, Antenna Engineering. New York, NY, USA: McGraw-Hill, 1968.
[26] , H. Jasik, Ed., Antenna Engineering Handbook. New York, NY,USA: McGraw-Hill, 1961.
[27] G. A. Thiele, E. P. Ekelman, and L. W. Henderson, “On the accuracyof the transmission line model of the folded dipole,” IEEE Trans. An-tennas Propag., vol. AP-28, pp. 700–703, Sept. 1980.
[28] A. R. Clark and A. P. C. Fourie, “An improvement to the transmissionline model of folded dipole,” IEE Proc. Microwave Antennas Propag.,vol. 138, no. 6, pp. 577–579, Dec. 1991.
[29] K. V. S. Rao, P. V. Nikitin, and S. F. Lam, “Antenna design for UHF-RFID tags: A review and a practical application,” IEEE Trans. An-tennas Propag., vol. 53, pp. 3870–3876, Dec. 2005.
[30] P. V. Nikitin, K. V. S. Rao, R.Martinez, and S. F. Lam, “Sensitivity andimpedance measurements of UHF-RFID chips,” IEEE Trans. Microw.Theory Tech., vol. 57, no. 5, pp. 1297–1302, May 2009.
Gerard Zamora Gonzalez was born in 1984in Barcelona, Spain. He received the Telecom-munications Engineering Diploma, specializing inelectronics and the Telecommunications Engineeringdegree from the Universitat Autònoma de Barcelona,Barcelona, Spain, in 2005 and 2008, respectively,where he is currently working toward the Ph.D.degree.His research interests include passive microwave
devices based on metamaterial concepts and antennadesign for RFID systems.
Simone Zuffanelli was born in Prato, Italy, in 1983.He received the Electronics Engineering Diploma in2008 at the Università Degli Studi di Firenze, Italy.He received the M.S. degree in micro and nanoelec-tronics engineering from the Universitat Autònomade Barcelona, Barcelona, Spain, in 2011.He he is currently working as a researcher in the
field of metamaterial inspired antennas and RFIDtags. His previous experiences include electronicdesign in the context of European projects “Persona”and “NOMS.”
Ferran Paredes was born in Barcelona, Spain,in 1983. He received the TelecommunicationsEngineering Diploma (specializing in electronics)and the Telecommunications Engineering degreethe Ph.D. degree in electronics engineering fromthe Universitat Autònoma de Barcelona, Barcelona,Spain, in 2004, 2006, and 2012, respectively.He was Assistant Professor from 2006 to 2008 at
the Universitat Autònoma de Barcelona, where he iscurrently working as a Research Assistant. His re-search interests include metamaterial concepts, pas-
sive microwaves devices, antennas and RFID.
Ferran Martı́n (M’04–SM’08–F’12) was born inVizcaya, Spain, in 1965. He received the B.S. degreein physics and the Ph.D. degree from the UniversitatAutònoma de Barcelona (UAB), Barcelona, Spain,in 1988 and 1992, respectively.From 1994 to 2006, he has been Associate
Professor in Electronics at the Departament d’En-ginyeria Electrònica (Universitat Autònoma deBarcelona), and since 2007 he has been Full Pro-fessor of Electronics. In recent years, he has beeninvolved in different research activities including
modelling and simulation of electron devices for high frequency applications,millimeter wave and THz generation systems, and the application of electro-magnetic bandgaps to microwave and millimeter wave circuits. He is now veryactive in the field of metamaterials and their application to the miniaturizationand optimization of microwave circuits and antennas.Dr. Martı́n is the head of the Microwave and Millimeter Wave Engineering
Group (GEMMA Group) at UAB, and director of CIMITEC, a research Centeron Metamaterials supported by TECNIO (Generalitat de Catalunya). He hasorganized several international events related to metamaterials, includingWorkshops at the IEEE International Microwave Symposium (2005 and 2007)and European Microwave Conference (2009). He has acted as Guest Editorfor three Special Issues on Metamaterials in three International Journals.He has authored and co-authored over 350 technical conference, letter andjournal papers and he is co-author of the monograph onMetamaterials entitledMetamaterials with Negative Parameters: Theory, Design and MicrowaveApplications (Wiley, 2013). He has filed several patents on metamaterials andhas headed several Development Contracts. Among his distinctions, he hasreceived the 2006 Duran Farell Prize for Technological Research, he holds theParc de Recerca UAB—Santander Technology Transfer Chair, and he has beenthe recipient of an ICREA ACADEMIA Award.
Jordi Bonache (S’05–M’07) was born in 1976in Barcelona, Spain. He received the Physics andElectronics Engineering degrees and the Ph.D. de-gree in electronics engineering from the UniversitatAutònoma de Barcelona, Barcelona, Spain, in 1999,2001, and 2007, respectively.In 2000, he joined the “High Energy Physics In-
stitute” of Barcelona (IFAE), where he was involvedin the design and implementation of the control andmonitoring system of the MAGIC telescope. In 2001,he joined the Department of Electronics Engineering
of the Universitat Autònoma de Barcelona where he is currently Lecturer. From2006 to 2009 he worked as executive manager of CIMITEC. Currently he isleading the research in RFID and antennas in CIMITEC. His research inter-ests include active and passive microwave devices, metamaterials, antennas andRFID.
![The Commissar Vanishes by Isaak Zelensky [R]](https://static.documents.pub/doc/80x56/577cc0e11a28aba711917251/the-commissar-vanishes-by-isaak-zelensky-r.jpg)
