MUTUAL COUPLING BETWEEN PLANAR INVERTED-F ANTENNAS

Transcription

1 MUTUAL COUPLING BETWEEN PLANAR INVERTED-F ANTENNAS H. Carrasco, H. D. Hristov, R. Feick an D. Cofré Departamento e Electrónica Universia Técnica Feerico Santa María Av. España 68 Valparaíso, Chile ABSTRACT: Mutual coupling etween two planar inverte-f antennas (PIFA) ase on a groun plane has een stuie numerically an experimentally. Several arrangements of collinear, orthogonal an parallel PIFA elements with interelement spacing ranging from.λ to.9λ have een examine at the esign frequency of.45 GHz, an in the frequency an. 3. GHz. Key wors: inverte-f antenna, antenna arrays, mutual coupling. INTRODUCTION Along with the monopole, patch an slot antennas, the inverte-f antenna (IFA) has ecome of primary importance for portale an hanhel wireless communication units. It is known as a high efficiency quasi-omniirectional antenna with a height of aout.5λ.λ, i. e..5 5 times shorter than the quarter-wave monopole []-[3]. The inverte-f antenna, wire or planar (PIFA), is a low-profile moification of the quarter-wave monopole, an thus elongs to the group of unalance antennas. There is a huge amount of research an evelopment work on classical an novel single IFA configurations []-[7], an on their iversity arrangements with monopole or patch antennas [8]-[]. Surprisingly, few results are known on the theory an practice of IFA arraying. Very little has een pulishe, in particular, for the mutual coupling etween IFA array elements []-[], which epens not only on the interelement spacing, ut also on their mutual orientation [3]. The same oservation is vali for the statistical receive characteristics of two-element or multi-element IFA arrays. Normally, the mutual electromagnetic coupling aversely affects the antenna array input an raiation characteristics. This is typical for the single-port arrays an aaptive antennas, for example. In the multi-port antenna arrays however, exploite in MIMO communication systems, the mutual coupling may prouce a positive effect: an improvement of system correlation an ata capacity [4]. In this paper, the results of numerical an experimental stuy of single PIFA an mutual coupling in a oule PIFA array are presente an iscusse. Several asic array arrangements with ifferent PIFA joint orientations (collinear, orthogonal an parallel) have een examine numerically an experimentally for a spacing ranging from aout.λ to.9λ, at the esign frequency of.45 GHz an for a frequency an. 3. GHz. For comparison, arrays consisting of PIFA an monopole, an of two quarter-wave monopoles have also een stuie.. PIFA AND MONOPOLE DESIGNS The PIFA element, rawn in Fig., is locate on an infinite groun plane x-y (not shown in the figure). It is fe y a coaxial cale C an consists of two sections. h ' e' w' '' ' e'' a x s s φ m C '' Figure Planar inverte-f antenna ase on infinite groun plane The first is a low-profile inverte-l planar section, mae of a thin metal sheet of with w an thickness t. It has a horizontal ranch cc - of length s +s, an a vertical ranch -ee of height h, with the ege ee groune. The secon section is a short cylinrical monopole of height h an iameter m fe at point a, an connecte to the L-section at point. Functionally the PIFA can e viewe as an inverte-l antenna cc - -awith a match circuit - -ee ae. The single PIFA an PIFA arrays were simulate an optimize numerically y means of Ansoft HFSS an Optimetrics software package [5]. The measurements were mae y use of a Rohe & Schwarz Vector Network Analyzer, moel ZVRE [6]. At the esign frequency of.45 GHz the final PIFA element use for the mutual coupling stuy, ha the following optimize imensions: h =.mm, w = 7.4 mm, s = 9. mm, s = 6.7 mm, t =. mm an m =.5 mm. It was foun that at.45 GHz the single PIFA has a value of the S-parameter S, or the reflection coefficient, equal to 38, compute, an 4, t y c' c''

2 measure. Its numerical input anwith at S = is.7% an the corresponing measure value is.%. The simulations have shown that the PIFA antenna, in contrast with the monopole can e tune for a very goo match not only to the stanar 5-ohm cale as is the case here, ut also for a range of other impeance values. The PIFA was compare to a.45-ghz quarter-wave monopole with a wire iameter of.5 mm an a resonant height of 9.5 mm. The monopole is.54 times higher than the PIFA an its match at the esign frequency is inferior: S = 7.3, compute, an 8.5, measure. On the other han the monopole has etter input anwith: 6.5 %, compute, an 9.6 %, measure. It has to e note here that the computer simulation of PIFA an monopole antennas was mae uner the assumption of lossless metal structures in air an ase on a lossless infinite groun plane. The measure antenna moels were actually faricate in copper, an were positione in the mile of a square ronze plate of size 55 cm, which eing much larger than the antenna arrays provies a goo approximation of the infinite groun plane. As was shown, the PIFA iffers significantly from the monopole in height an input characteristics. There is also some ifference in their raiation patterns. The groune vertical monopole is an ieal omniirectional antenna in the azimuth plane, while the PIFA s horizontal pattern slightly eviates from the circular shape. The stuie PIFA esign has a gain pattern G( φ ), which for four values of the azimuth angle φ is escrie y: G ( ) = 4.i, G (9 ) = 4.4 i, G(8 ) = G( ) an G (7 ) = 3.7 i. Thus, there is a.7 front to ack gain ifference ue to the non-symmetrical an more complex current an near-fiel istriution in the PIFA. 3. PIFA ARRAY ARRANGEMENTS Fig. is a view from aove of a two-element PIFA array efine y the interelement spacing an rotating angles α anα. The array axis passes through the PIFA element s fee points. The rotation of each PIFA is mae aroun its fee-monopole line a-. Because the array characteristics are relate also to the groun plate or container size an shape, in orer to isolate the epenency of mutual coupling on the element position an orientation, the PIFA arrays were situate on an infinite groun plane. Changing the angles α an α y a step of 9 seven arrangements of the two-element PIFA array were efine an stuie, with: (a) collinear elements: C- ( α = 8, α = ), C- ( α = α = ) an C-3 ( α =, α = 8 ); () orthogonal elements: O- ( α = 9, α = ) an O- ( α =, α = 9 ); (c) parallel elements: P- ( α = α = 9 ) an P- ( α = 9, α = 7 ). α α Figure Geometry of two-element PIFA array A' In aition, arrangements involving monopoles were examine: (i) mixe-antenna arrangement PM, where the PIFA # is replace y a monopole an PIFA # is turne at α = 9, an (ii) two-monopole arrangement MM. The arrangements C-, C-3, P-, P- an MM are electromagnetically symmetric in reference to the line AA, while the rest are nonsymmetrical, ecause of specific PIFA orientation (C-, O- an O-) or ifferent antenna elements (PM). 4. MUTUAL COUPLING OF PLANAR INVERTED-F ANTENNAS The stuie two-element PIFA arrays were computersimulate on an ieal infinite groun plane. The faricate array elements were fixe in the mile of a large finite groun plate. The single PIFA an the groun plate imensions were specifie in the previous section. Figs. 3, 4 an 5 illustrate the simulate (soli line) an measure (circles on a otte line) S-parameter S, or the mutual coupling, in eciels, etween collinear (C), orthogonal (O) an parallel (P) PIFA elements, respectively, as a function of spacing in wavelengths. Similarly, Fig. 6 shows the coupling vs. spacing of a PIFA positione next to a monopole (PM), an the coupling of two quarter-wave monopoles (MM). The spacing is efine y the istance etween fee points of the two antennas. Accoring to Fig. 3, among all collinear arrangements the smallest coupling is prouce y C-, while C-3 has the iggest coupling for all spacing values. A simplifie explanation of this ehavior follows from A

3 the current/charge istriution on a PIFA. On its open ege cc (Fig. ) the current is zero an the Coupling is much larger when the open en of one antenna faces the fee point of the other (O-). Coupling, () - C- C- C-3 Coupling, () - P- P Figure 3 Coupling etween collinear PIFA elements vs. spacing in wavelengths for arrangements C-, C- an C Figure 5 Coupling etween parallel PIFA elements vs. spacing in wavelengths for arrangements P- an P- Coupling, () - - O- O Figure 4 Coupling etween orthogonal PIFA elements vs. spacing in wavelengths for arrangements O- an O- voltage an charge are maximum. If the charge eges of the two PIFA elements are very close, as in the case C-3, there will e a strong electromagnetic tie or coupling etween them, an vice versa (case C-). By a similar argument the coupling ifference etween the parallel arrangements P- an P- can e justifie. This explanation woul suggest that the arrangement C- shoul have coupling values intermeiate to those exhiite y C- an C-3, as was actually foun oth through simulation an measurements. From the coupling graphs for O- an O- (Fig. 4) it is seen that for an orthogonal arrangement, esies spacing an important role is again playe y the irection (orientation) of the open en of the PIFA. At a spacing.3λ (Figs. 3) the mutual coupling has small variations, ±.75 aroun a central value of for all array arrangements, except for C-3 an O-, where it has much igger values: 5.5 an 8.5, respectively. For spacing greater than aout.4λ the comparison etween all PIFA cominations, stuie in this paper, reveals that C- an O- act as minimum-coupling arrangements, with almost equal measure coupling values, which average 9.8, 6.5 an 8.5, for / λ =.5,.5 an.75, respectively. The mixe array (PM) comprising PIFA an monopole, an the two-monopole array (MM) have very similar coupling ehavior for the whole range of / λ (Fig. 6). More exactly, at / λ =.5,.5 an.75, the arrangements PM (an MM, respectively) have the following measure coupling values: 9.6 ( 9.), 3.9 ( 3.) an 6.4 ( 5.7 ). Hence, in contrast to PIFA arrangements C- an O-, for spacing greater than.5λ the arrays PM an MM have stronger coupling, aout 3 4 larger. In most cases, C-3, O-, P-, PM an MM, there is a very goo agreement etween the graphs of simulate an measure coupling values for the complete range of spacing values. Exceptions occur for arrangements C-, C-, O- an P-, where for spacing larger than.5λ the ifference etween simulations an measurements ecomes significant. At this point we have to consier that the simulations were carrie out for an infinite groun plane, while practical 3

4 consierations limite the groun plate for the measure antenna moels to a square of size 4.5λ. Coupling, () - - PM MM Figure 6 Coupling etween PIFA an monopole (PM), an etween two monopoles (MM) vs. spacing in wavelengths Fig. 7 is a color contour presentation of the measure two-element PIFA array S-parameters, S, S an S = S, as functions of spacing, in wavelengths, an frequency, in gigahertz, for five array arrangements: (a) collinear PIFA arrangement C-, () orthogonal PIFA arrangement O-, (c) parallel PIFA arrangement P-, () array PM of PIFA an monopole, an (e) two-monopole array MM. While the graphs in Figs. 3-6 are limite only to the coupling epenency on spacing, at the esign frequency of.45 GHz, the pictures in Fig. 7 contain aunant information aout the array coupling an match performance in the spacing omain..9λ an frequency omain -3GHz. Several oservations follow from this figure: (i) As is expecte S is equal to S for the symmetrical arrays C-, P-, an MM, while for the nonsymmetrical arrays O- an PM S iffers from S ; (ii) All antenna array elements have practically preserve their esign resonant frequency, anwith an match performance for spacing greater than.4.5λ ; (iii) For smaller spacing (say less than.4λ ) the resonant (match) frequency of oth PIFAs ecomes somewhat shifte from the esign frequency; (iv) The two-monopole array has etter frequency anwith ut worse match performance than the twotwo-pifa array, a ehavior similar to what was escrie in the comparison etween the single monopole an PIFA (Section ). S a) ) c) ) e) Spacing, (wavelengths) S S Frequency, (GHz) Figure 7 Scattering parameters S, S an S vs. spacing an frequency of: (a) collinear PIFA array C-, () orthogonal PIFA array O-, (c) parallel PIFA array P-, () array of PIFA an monopole an (e) two-monopole array 5. CONCLUSIONS The numerical an experimental stuy of mutual coupling an match performance in several two-pifa arrays as function of spacing an joint orientation has resulte in a large amount of numerical an experimental ata. In the spacing omain it was oserve that the coupling epens mainly on the istance etween the PIFA open-ene sies. For a constant spacing etween the fee points the coupling experiences consierale changes with the relative angular orientation of the PIFA elements. Also, the stuie mixe-antenna array (PM) an two-monopole array (MM) have stronger coupling, aout 3 4 greater than the majority of two-pifa arrangements. These ifferences are especially pronounce for larger array spacing. 4

5 In the frequency omain has een foun that for spacing greater than.4.5λ all PIFA array elements have practically preserve their singleelement esign resonant frequency, anwith an match performance. The two-monopole array has etter frequency anwith ut worse input match than the two-pifa array. The finings for mutual coupling in two-pifa arrays will e of practical value for multi-pifa array esign an optimization. In contrast with the monopole array, where the coupling epens only on the interelement spacing, the relative element orientation in the PIFA array provies an option for controlling the egree of mutual coupling an the shape of the element raiation pattern. This option can e useful, for example, in multi-port MIMO communication systems for further reuction of space correlation y aing the effects of mutual coupling an pattern (angle) iversity. y hea, han, an shouler effects at 9 MHz: Parts I an II, IEEE Trans. Veh. Technology, 5, (), M.A. Jensen, an Y. Rahmat-Samii, FDTD analysis of PIFA iversity antennas on han-hel transceiver unit, IEEE Int. Antennas Propagat. Symp. Digest,, (993), K. Tsunekawa, Diversity Antennas for Portale Telephones, IEEE 39 th Veh. Technology Conference, Vol., (989), J. Thaysen, Mutual Coupling Between Two Ientical Planar Inverte-F Antennas, IEEE AP-S Int. Symp. Digest, Vol. 4, (), T. Svantensson, an A. Ranheim, Mutual Coupling effects on the capacity of Multielement Antenna Systems, Proc. IEEE Int. Conf. Acoustics, Speech, an Signal Processing, Vol. 4, (), Ansoft HFSS 9. an Optimetrics, Ansoft Corporation,. 6. Rohe & Schwarz Catalog, Test & Measurement Proucts, / ACKNOWLEDGMENTS: The authors wish to acknowlege the support for this work mae y the Chilean Agency CONICYT (Fonecyt project #9) an y the UTFSM project DGIP-33. REFERENCES. R.J.F. Guertler, Isotropic transmission-line antenna an its toroi-pattern moification, IEEE Trans. Antennas Propagat., AP, (977), K. Fujimoto, A. Henerson, K. Hirasawa, an J.R. James, Small antennas, Research Stuies Press, C.R. Rowell an R.D. Murch, A capacitively loae PIFA for compact moile telephone hansets, IEEE Trans. Antennas Propagat., AP- 45, (997), K. Fujimoto an J.R. James (Eitors), n e., Moile antenna systems hanook, Artech House, Boston-Lonon,. 5. A.K. Skrivervik, J.-F. Zürcher, O. Stau an J.R. Mosig, PCS antenna esign: the challenge of miniaturization, IEEE Antennas Propagat. Mag., 43, (), C. Soras, M. Karaoikis, G. Tsachtsiris an V. Makios, Analysis an esign of an inverte-f antenna printe on PCMCIA car for.4 GHz, ISM an, IEEE Antennas Propagat. Mag., 44, (), M. Olmos, H.D. Hristov an R. Feick, Inverte-F antennas with wiean match performance, Electronic Letters, 38, (), T. Taga, Analysis of planar inverte-f antennas an antenna esign for portale raio equipment, in K. Hirasawa an M. Haneishi (E.), Analysis, esign an measurement of small an low-profile Antennas, Artech House, Norwoo, MA, J.S. Colurn, Y. Rahmat-Samii, M.A. Jensen, an G.J. Pottie, Evaluation of personal ual-antenna hanset iversity performance, IEEE Trans. Veh. Technology, 47, (998), K. Ogawa, T. Matsuyoshi, an K. Monma, An analysis of the performance of a hanset iversity antenna influence 5