论文研究-Optmization of SIW Band-pass Filter with the Wide and Sharp Stop-band Using Space Mapping.pdf

所需积分/C币:9 2019-08-16 10:46:46 632KB .PDF
26
收藏 收藏
举报

空间映射优化宽阻带特性基片集成波导带通滤波器,徐娟,李兆龙,本文提出了一种具有陡峭宽阻带特性的基片集成波导带通滤波器。这种滤波器采用了主模与高次模耦合的方式实现非物理结构的交叉耦合
山国武获论文在丝 http:/www.paper.edu.cn describes the same object as the fine model: less accurate but much faster to evaluate. Surrogate models R in(2)are constructed as follows Ro (x)=R,(x, po (3) where R is a generic SM surrogate model: R composed with suitable SM transformations.A 90 vector of SM parameters P>, is obtained using the parameter extraction procedure p= argmin >R R k=0 Coarse model is the critical component of SM algorithms. Because the parameters extraction (4)and surrogate model optimization (2) require numerous evaluations of the surrogate, R should be computationally cheap. On the other hand, the coarse model should be reasonably 95 accurate. Otherwise, the Sm optimization process may require many iterations(2)-(4)or may even fail to find a design satisfying given specification requirements. Tab. 1 compares possible coarse model types. Analytical and equivalent circuit models are computationally cheap but they may lack accuracy and they are typically not available for structures such as antennas and substrate-integrated circuits. For that reason, many SM-related papers use filters as test examples 100 [. it is relatively easy to build a reliable equivalent circuit model for a filter structure The generic type of coarse model, available for any device, can be obtained through EM simulation of the coarsely-discretized structure of interest. It is typically accurate; however, it is also rclativcly cxpcnsivc. The cost is a major bottleneck in adopting coarscly-discrctizcd EM models to SM optimization Tab. 1 Coarse models and their characteristics Model Tvpe CPU Cost Accuracy Availability Analytical Vcry low Rather limited Equivalent circuit Low 冂 ecent Limited(mostly filters) Coarsely-discretized Generic: available EM Simulation High Very good For all structure 2 Optimization of the siw Filter The siw cavity resonator is build utilizing the ton and bottom conducting plancs as horizontal walls and via fence as side walls. The size and spacing between vias are properly 110 chosen to prevent electromagnetic fields leakage. The cavity can be filled with substrate as rectangular waveguide cavity. Hence it can be designed like the conventional rectangular waveguide resonator. Fig. 1 depicts the SIw TE1o1/TE201 cavity excited by two 50 Q2 microstrip lines close to each other. The initial dimensions of the siw cavity are determined by setting the resonant of the TE mode equal to the center frequency of the desired passband using the following 115 formula e whcrc 山国武获论文在丝 http:/www.paper.edu.cn W=w LM=L 0.95p 0.95 120 where W and L are the width and length of the SiW cavity respectively, and d and pare the diameter of the via-hole and the center to center pitch between two via-holes respectively. The paramctcrs of thc cavity for thc dcfincd position of the transmission zero arc then adjusted using the models in Fig. 1. For the SIW TE1oI/TE201 cavity, the position of the transmission zero can also be tuned by changing the position of the input/output lines 125 Fig. 1. The single SIW TE101/TE201 cavity The dcsign spccification arc: a rcturn loss >20dB ovcr thc cntirc pass band(12GH7s-Wws 12.5GHz), and the insertion loss>60 db for8GHz≤w≤1 GHz and13.GHz≤W≤16.5GHz Duc to the requirements of the dcsign of the filter bandwidth Bw=0. 5GHz, allowing the band 130 within a certain fluctuation, therefore the selection of narrowband Chebyshev filter design. In modern microwave filier design theory as a foundation, according to the known parameters of filter, the filter order N=4, which can meet the performance requirements of filter rejection band The low-pass prototype filter [131 parameters are g0=1,g1=16703,g2=1.1926,g3=20361,84=0.8419,g5=1.9841 135 Then the coupling coefficients and external quality values(Hong and Lancaster 2001) could be computed by the formulas as follows( 8) FBw BW g1 FBw FBn n-112,t+ for i=l to m 8,gil Where the FBW is the fractional bandwidth, BW is the pass band width, f is the center frequency of the filter, Q and g are the external quality factors of the input and output ports respectively, ki i is the coupling coefficient between i cavity and i+ 1 cavity, n is the order of filter. Then, can be obtained the initial design value of the coarse model through the formula( 8) Based all the equations mentioned above in formula (8), the coupling matrix then can be extracted. These data can give the ensuring design a place to begin and can be optimized with the 4 山国武获论文在丝 http:/www.paper.edu.cn space mapping method 145 Using space mapping method, the coarse model is assumed to be much faster than the fine model. In the optimization, the coarse model is implemented using a circuit simulator Agilent ADS shown in Fig. 2. The LC resonant circuit represents the siw cavity and the phase-shifter rclatcs with thc coupling cocfficicnt. ADs can bc considcrcd as the primary coarse modcl evaluator in the microwave area because it is widely used and it allows convenient and 150 straightforward creation of coarse models for many microwave structures and devices. The parameters of coarse model are k,2 k23 k Firstly, optimized the coarse model in ADS and obtained the optimal solution of coarse model x=[12.24621.860.03670.02630.0367 Phase shift Phase shift Phase shift Phase shift Phase shift Phase=90 Phase--90 Phase=-90 Zref=sqrt(Qe*50)Ohm Zref(l/k12)Ohm Zref=(1/ k23)Ohm Zref-(1/k34)Ohm Zref-sqrl(Qe*50)OhIm PCI PLC PLC3 PLC4 Portl L(0ct0)pHL=100020)HL=10002x0)pHL-(1000c0) ph Port2 Z=50 Ohm C(000FC=(1000)pFC=10002x*m)FC-(1002n0)pF7=50Ohm 155 Fig. 2. The coarse model of siw bandpass filter in ADs The microstrip impedance of fccd is 50 ohms. Firstly, a microstrip to SIW transition by thc form of CPw, considering the machining precision, slit width of CPW is 0.8 mm. Using 160 microstrip tapered line impedance matching to achieve wide The fine model is simulated with full-wave method, and implemented using Ansoft HFSS shown in Fig 3. In order to reduce the number of optimized parameter, the width of each SIW cavity is fixed: W=13mm, and the feed line is Wso=50s2. The optimized parameters of fine model are 165 LL L Wi2 W23 W34] (11) W Fig. 3. The fine model of siw bandpass filter in HFSS 山国武获论文在丝 http:/www.paper.edu.cn 00 0 11.1 112113 114 11.5 11E 03.2343638404244斗64849 L(mm) W≥3mm) 170 (a)Relationship between fa and I (b) Relationship between k23and w23 s0.05 急26 25 V12Nm)写q 4363840 44464.85.3 4.24446 54 Ls(mm (c)Relationship between k12(k.34)and W12(W34) (d)Relationship Q and L Fig 4. Design curves for the proposed space mapping algorithm.(a) relationship between the center frequency of the SIW cavity fo and the width of the SIW cavity; (b)and(c) Relationship between the coupling coefficient k and the iris distance w;(d)Relationship between the external quality factor and the length L, Then, obtain the curves between the coarse model parameters and the fine model parameters in HFSS through parameters scanning shown in Fig 4 The inter-resonator coupling coefficients between two SiW cavities as shown in Fig. 4(b), 80 could be extracted by simulation to find the two charactcristic frcqucncics (f,f)that arc the frequencies of the peaks in the transmission response of the coupled structure. The coupling coefficients can be controlled and modulated by the width of inductive window between two cavities. The relationship between the coupling coefficients and the characteristic frequencies is shown as follows 185 2 +fl The external-coupling structure as shown in Fig. 4(c) can be used to extract the external quality factor. The external quality factor is controlled by the length of L. Based on simulation the external factor could be evaluated in terms of the following relationship (13) +90 Where Afn is the frequency diference belween the +90 phase responses ofS11 山国武获论文在丝 http:/www.paper.edu.cn For this problem we used simplified input space mapping with shift parameter. The fine model initial and optimized responses after 5 space mapping iterations are shown in Fig. 5. The Tab. 2 shows the line model pararneters of5 iterations Tab. 2 The parameters of the fine model Fine model(mm) L W12 23 W34 4.6l 4.42 Xf(2) 10.73 4.42 3.94 3.96 Xf(3) 10.55 4.1 4.24 3.99 10.44 4.38 4.17 4.21 Xf(5) 10.48 4.10 4.10 4.02 4.12 0 Xf(1)-S11 20 Xt(5-S11 50 13.5 14.0 Fig.5.Initial and optimized S1l and 521 versus frequency for the SIW filter The size of Slw filter is 5178mm X 30mm X0.508mm. The photo of proposed filter is 200 shown in Fig. 6. And it has been measured by using Universal Test Fixture(Anritsu 3680V) and thc Vcctor Nctwork Analyzer Systcms (Agilent N5244A). Thc mcasurcd insertion loss includes the loss of the test fixture and the transmission line. From the measured result as shown in Fig. 7, the measured results are in good agreement with the corresponding simulation results and show the good frequency selective performance. AnritsU MODEL 3680-20 205 Fig. 6. Thc photo of proposcd filter In this work, non-physical cross-coupling provided by highcr modcs in the Siw cavities is used to generate the finite transmission zeros for improved stopband performance which different from input/output topologies of the filter discussed for wide stopband applications. The designed 210 substrate integrated waveguide bandpass filter has wide and precipitous stopband shown in Fig. 7. 山国武获论文在丝 http:/www.paper.edu.cn To sum up, it exhibits an insertion loss <ldB at the center frequency(12.25GHz),a return loss>174dB over the entire pass band. The insertion loss >40dB in the stop band is from 2 GHz lo 11 GHz and 13.5 GHz Lo 17.3 GHz shown in Figure 7. Because of machining accuracy and lest methods. there are some deviations between the measured results and the simulated results Measured S11 simulation S21 Simulation s11 Measured s21 Measured 13.5 215 (a) The results of filter in widc frcqucncy (b) The results of filter in narrow frcqucncy Fig. 7. Simulation and the measurement results of the SIw band-pass filter. 3 Conclusion In this paper an optimized procedure for a siw passband filter based on the space mapping technique is studied. The proposed design approach is based on a combined use of equivalent circuit model of a filter and a space mapping technique. A reduced number of full-wave evaluations are needed, leading to a reduced optimization time. The proposed filter has good frequency selectivity and a compact size of 5178mm X 30mm X0.508mm. The simulated results 225 show this filter has an excellent frequency selectivity and return loss Acknowledgements The authors would like to thank thc support of Spccializcd Rcscarch Fund for the Doctoral Program of Higher Education, 20133219120004 and 20123219110018 References 230 [1]WUK. Integration and interconnect techniques of planar and non-planar structures for microwave and millimeter-wave circuits-Current and future trend[c]. Asia-Pacific Microw. Conf, 2001, 411-416 [2] DESLANDES, D and WUK Integrated microstrip and rectangular wave-guide in planar form[ J].IEEE Microw. Wireless Compon. Lett. 2001,11: 68-70. [3]DESLANDES, D and WUK Single-substrate integration techniques for planar circuits and waveguide 235 filter[J]. IEEE Trans. Microw. Theory Tech, 2003, 51: 593-596 [4]SHEN W, WUL.S, SUN X.W., YIN W.Y. and MAO J F Novel substrate integrated waveguide filters with mixed cross coupling(CC)[]. IEEE Microw Wireless Compon. Lett., 2009, 19: 701-703 [5LI Z, CHEN R.S. and WUK System design considerations of a generic integrated frequency-modulation continuous-wave radar front-end J Microwave and Optical Technology Letters, 2013, 55: 1907-1912 240 16JL1 7 and WU K. 24-Gil Iz Frequency-Modulation Continuous-Wave Radar Front-End System-on-SubstrateUJ IEEE Trans. Microw Thcory Tcch. 2008, 56: 278-285 7] AMARIS, ROSENBERG U and BORNEMAnn J Adaptive synthesis and design of resonator filters with source/load-multi-resonator coupling[J]. IEEE Trans. Microw Theory Tech, 2002, 50: 1969-1978 8]CHEN X.P. DROLET D and wuK Substrate integrated waveguide filters for airborne and satellite system 245 applications[C ]. Can Elect. Comput. Eng. Conf, 2007, 659-662 [9] DONG Y D. HONG W. and TANG H.J. a novel millimeter wave substrate integrated waveguide filter using TE301 mode cavities[C]. Global Millimeter Waves Symp., 2008, 91-93 [10] BANDLER J W, BIERNACKIR M. CHEN S.H., GROBELNY P A and HEMMERS R H Space mapping technique for electromagnetic optimization [J]. IEEE Transactions on Microwave Theory and Techniques, 山国武获论文在丝 http:/www.paper.edu.cn 2501994,42:2536-2544 [111 BANDLER J W, CHENG Q S, and DAKROURY S.A. Space Mapping: The State of the Art[J. IEEE Transactions on Microwave Theory and Techniques, 2004, 52: 337-361 [12]KIOZIEL S. BANDLER J.W. and MADSEN K Quality assessment of coarse models and surrogates for space mapping optimization[]. Optimization and Engineering, 2008, 9: 375-391 255 [13] CAMERON R.J. Advanccd coupling matrix synthcsis tcchniqucs for microwave filters[J]. IEEE Transactions on Microwave Theory and techniques, 2003, 51: 1-10 4]Agilent ADS, Version 2003C. Agilent Technologies[7]. 1400 Fountaingrove Parkway, Santa Rosa, CA, 2003,95:403-1799 26 空间映射优化宽阻带特性基片集成波 导带通滤波器 徐娟,李兆龙2,陈如山 1.南京理工大学电子工程与光电技术学院,南京,210094; 2.东南大学毫米波国家重点实验室,南京,210096) 摘要:本文提出了-种具有陡峭宽阻詣特性的基片集成波导带通痣波器。这种滤波器采用了 主模与高次模耦合的方式实现非物理结枃的交叉耦合,引入了传输零点,改善了滤波器的阻 带特性。在浪波器的设计中,需要用到全波巳磁仿真,使得进一步的优化工作比较费时。本 文在优化基片集成波导滤波器时采用空间映射算法,减少了设计周期提高了效率。具体的实 270施是通过优化快速但精度低的粗模型ⅹ修正并预测耗时但精度高的细模型,直到粗模型与细 模型之间的误差达到可以接受的范围,算法结束。最后,设计了一个中心频率为1225GHz, 带宽为0.5GHz的四阶基片集成波导带通滤波器。测试结果与伤真结果吻合较好,通带内回 波损耗大于174dB,阻带2-1GHz和13.5-17.5GHz范围内插入损耗都大于40dB.进一步说 明了采用空间映射算法优化基片集成滤波器的高效性 275关键词:基片集成波导;带通滤波器;非物理结构的交叉耦合;空间映射算法 中图分类号:TN61

...展开详情
试读 9P 论文研究-Optmization of SIW Band-pass Filter with the Wide and Sharp Stop-band Using Space Mapping.pdf
立即下载 身份认证VIP会员低至7折
一个资源只可评论一次,评论内容不能少于5个字
您会向同学/朋友/同事推荐我们的CSDN下载吗?
谢谢参与!您的真实评价是我们改进的动力~
  • 至尊王者

关注 私信
上传资源赚钱or赚积分
最新推荐
论文研究-Optmization of SIW Band-pass Filter with the Wide and Sharp Stop-band Using Space Mapping.pdf 9积分/C币 立即下载
1/9
论文研究-Optmization of SIW Band-pass Filter with the Wide and Sharp Stop-band Using Space Mapping.pdf第1页
论文研究-Optmization of SIW Band-pass Filter with the Wide and Sharp Stop-band Using Space Mapping.pdf第2页

试读结束, 可继续读1页

9积分/C币 立即下载