диафрагмированные волноводные фильтры / ab728622-fe41-46dc-b175-0c93f3028f49

Abstract— In this letter, a new method for rectangular waveguides bandpass filters with improved stopband performance is presented. By setting the position of the coupling window, the coupling of the high-order mode can be attenuated, which makes it difficult to form parasitic passbands. It can significantly improve the stopband rejection of the filter without introducing additional elements. Furthermore, the whole structure can be fabricated using the conventional computer numerical control (CNC) milling technology. A Ka-band waveguide filter is designed and tested. The stopband rejection of the design is increased from 1.4 times to 1.94 times the center frequency of the filter with the insertion loss of 0.5 dB. Finally, the measurement results validated the proposed design.

Index Terms— High-order mode suppression, rectangular waveguides, stopband improvement of filters, waveguide filters.

I. INTRODUCTION

FAST-DEVELOPING communication systems and limited spectrum resources have driven a growing demand for bandpass filters with wide-stopband and high rejection level. For example, in the passive intermodulation (PIM) test to reduce the measurement error [1] and in satellite communication systems to suppress clutter in the wireless transceiver

process [2].

In addition, a rectangular waveguide filter has been widely used in communication systems, due to its low insertion loss and high power handling capabilities. However, its poor stopband performance may limit certain applications. As is known, distributed elements in microwave bandpass filters usually cause undesired parasitic passband, due to the resonance of high-order mode [3]. The stopband rejection could be improved by cascading a low-pass filter [4]. Although it can effectively achieve high stopband isolation, it adds complexity and leads to bulky structures.

To improve the stopband performance without a low-pass filter, variety of designs were presented in [3] and [5]–[8]. Vahldieck and Hoefer [5] and Budimir [6] used ridged waveguide resonator or multiple inserts to provide attenuation at 1.5 times the center frequency, but they added manufacturing complexities, resulting in Qu reduction. Morelli et al. [3] and Riblet [7] propose to make the resonators with different widths, so that their harmonic frequencies are separated, which also achieve good stopband performance. Morelli et al. [8]

Manuscript received August 14, 2019; accepted September 2, 2019.

(Corresponding author: Q. Wu.)

Q. Wu, F. Zhu, and X. Shi are with the School of Electronic Engineering, Xidian University, Xi’an 710071, China (e-mail:qiuy.wu@gmail.com).

Y. Yang is with the China Academy of Space Technology, Beijing 100048, China, and also with The Chinese University of Hong Kong, Hong Kong (e-mail: yeemeen.yang@gmail.com).

Color versions of one or more of the figures in this article are available online at http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/LMWC.2019.2939887

had effectively controlled the second resonance frequency by changing the impedance ratio and length of the resonator. However, these designs increased the manufacturing cost. Ma et al. [9] and Zhou et al. [10] propose a substrate integrated waveguide (SIW) design methods. By introducing a slot on the metal plate to cut off the current path, higherorder mode can be destroyed to improve the stopband performance [9]. However, it cannot be directly used in the waveguide filter design. Zhou et al. [10] suppressed undesired spurious responses by staggering harmonics, which is similar to the method used in [3].

So far, the idea for the improvement of stopband performance is mainly based on the control of the spurious or harmonic modes. This letter proposed a new option. For example, to control the coupling mechanism rather than the frequencies of the undesired modes (maintaining the coupling of the fundamental mode). Consequently, although these modes still exist, it cannot form the parasitic bands and the stopband performance is naturally improved. A Ka-band filter is designed and tested as a verification model. By investigating the field distribution and the coupling mechanism, the coupling window is carefully arranged to achieve the design goals. The simulated and measured results suggest it has an excellent performance within 1.9 times the center frequency. It can be fabricated using conventional computer numerical control (CNC) milling processes and maintains excellent insertion loss performance.

II. DESIGN METHOD AND THEORY

Generally, the only fundamental mode is considered in the most rectangular waveguide filter design process. However, each resonator supports a variety of resonant modes. Fig. 1 shows the equivalent circuit with other modes. Usually, the waveguide cross section is setting the height as half of its width. Therefore, the first two parasitic passbands of a filter designed from the fundamental mode of TE101 mode (as indicated by the black nodes in Fig. 1) are caused by

TE102 mode and TE201 mode (as indicated by the gray nodes in Fig. 1). The design goal is to modify the gray path

(high-order modes) to reduce parasitic passbands. For this purpose, in addition to stagger the high-order mode resonant frequency [3], one can also break the coupling path of the higher-order modes.

It is well known that the coupling coefficient between adjacent cavities depends on the strength of their tangential magnetic field or the normal electric field [11]. The magnetic field of rectangular resonators can be described as

Fig. 3. Relationship between K value and iris width of the two modes.

where Hx0 and Hy0 are the maximum amplitudes of the magnetic fields in the x- and y-directions, respectively. m andn indicate the index of different modes. That is, for the

TE101 mode, both Hx and Hy become the strongest at the center of the four edges. However, TE102 and TE201 only

have maximum values on the edge of x- or y-directions, and minimum values on the other edge. The magnetic field distribution of three resonance modes is shown in Fig. 2.

Consequently, an iris arranged at the center of x will almost

have no coupling for TE201 mode, and similarly for TE102 at the center of y. This can be confirmed by the simulated results

of the model shown in Fig. 3. The corresponding K inverter values can be derived from the simulated S-parameter S21 by

Set the fundamental frequency of the cavity to 22.7 GHz. When the resonator length is equal to a-side, the resonant frequencies of TE102 mode and TE201 mode are the same, which is 34 GHz. Fig. 3 shows the calculated K value at 34 GHz versus the iris width.

III. DESIGN EXAMPLE AND SIMULATION RESULT

To verify the approach in improving the stopband performance, sixth-order filters with a center frequency f0 = 22.7 GHz and a passband of 550 MHz are designed as examples. Based on the above parameters, the coupling coefficients of the new filter are synthesized and shown in Table I.

The final dimensions shown in Fig. 5 can be derived from the well-known space mapping technique [12]. Note that the minimum coupling positions need to be slightly shifted from the center to compensate for the loading effect of the adjacent resonators. The simulated results of typical and new filters are depicted in Fig. 6. It can be seen from the simulation results that the (a) filter can provide attenuation (30 dB) up to 1.4 times the center frequency of the filter, while the suppression of the (b) filter can reach 1.9 times the center frequency of the filter. The parasitic passband of (a) filter is generated at 34 GHz and (b) filter exhibits better stopband performance without introducing additional structures.

The fabricated filter is shown in Fig. 7. The lid and structural part are silver-plated and finally joined by soldering to achieve better electrical contact. It is tested with the R&S ZVA67 vector network analyzer. The measurement compared with the simulated results shown in Fig. 7 verified the effectiveness of the proposed method. The filter has 30 dB of attenuation up to 44 GHz.

A comparison with other reported works is given in Table II. Morelli et al. [3], [8], and Zhou et al. [10] stagger

Fig. 6. S-parameter response of (a) typical filter and (b) new filter. Results are obtained with the finite-element-method solver Ansys HFSS.

Fig. 7. Simulation (dashed line) and measurement (solid line) of the new filter.

harmonics by changing the resonator structure. This method can provide good stopband performance, but it reduces design flexibility [3] and increases insertion loss and cost [8]. Budimir [6] pushes the parasitic passband away by ridged waveguide resonator, but did not provide enough attenuation bandwidth. Ma et al. [9] reported a relative good stopband performance. However, it cannot be implemented in hollow waveguide structures. Unlike the above-mentioned published methods, this letter improves the stopband performance as 1.94 times the center frequency of the filter by reducing the coupling of higher-order modes. Moreover, since the ordinary

TE101 resonators are employed, this letter shows outstanding insertion loss performance as 0.5 dB. Besides, it also has advantages in power handling capability and costeffectiveness.

IV. CONCLUSION

This letter proposes a new method to improve the stopband performance of rectangular waveguide bandpass filters. The coupling path of the high-order mode is blocked by carefully arranging the coupling windows. The high-order mode still exists but cannot form a parasitic passband. With this design, the stopband can be effectively extended without affecting the insertion loss. The simulation and measurement results verify the effectiveness of the proposed method.

REFERENCES

[1]D. Smacchia et al., “Advanced compact setups for passive intermodulation measurements of satellite hardware,” IEEE Trans. Microw. Theory Techn., vol. 66, no. 2, pp. 700–710, Feb. 2018.

[2]F. Teberio et al., “High-power waveguide low-pass filter with all- higher-order mode suppression over a wide-band for Ka-band satellite applications,” IEEE Microw. Wireless Compon. Lett., vol. 25, no. 8,

pp.511–513, Aug. 2015.

[3]M. Morelli, I. Hunter, R. Parry, and V. Postoyalko, “Stop-band improvement of rectangular waveguide filters using different width resonators: Selection of resonator widths,” in IEEE MTT-S Int. Microw. Symp. Dig., Phoenix, AZ, May 2001, pp. 1623–1626.

[4]R. J. Cameron, M. Yu, and Y. Wang, “Direct-coupled microwave filters with single and dual stopbands,” IEEE Trans. Microw. Theory Techn., vol. 53, no. 11, pp. 3288–3297, Nov. 2005.

[5]R. Vahldieck and W. J. R. Hoefer, “Finline and metal insert filters with improved passband separation and increased stopband attenuation,”

IEEE Trans. Microw. Theory Techn., vol. 33, no. 12, pp. 1333–1339, Dec. 1985.

[6]D. Budimir, “Optimized E-plane bandpass filters with improved stopband performance,” IEEE Trans. Microw. Theory Techn., vol. 45, no. 2,

pp.212–220, Feb. 1997.

[7]H. J. Riblet, “Waveguide filter having nonidentical sections resonant at same fundamental frequency and different harmonic frequencies,” U.S. Patent 3 153 208, Oct. 13, 1964.

[8]M. Morelli, I. Hunter, R. Parry, and V. Postoyalko, “Stopband performance improvement of rectangular waveguide filters using steppedimpedance resonators,” IEEE Trans. Microw. Theory Techn., vol. 50, no. 7, pp. 1657–1664, Jul. 2002.

[9]L. Ma, J. Li, Y. Zhou, X. Sun, and H. Zhu, “Substrate integrated waveguide filter with higher-order modes suppression,” in Proc. Int. Conf. Microw. Millim. Wave Technol. (ICMMT), Chengdu, China, May 2018, pp. 1–3.

[10]K. Zhou, C.-X. Zhou, and W. Wu, “Resonance characteristics of substrate-integrated rectangular cavity and their applications to dualband and wide-stopband bandpass filters design,” IEEE Trans. Microw. Theory Techn., vol. 65, no. 5, pp. 1511–1524, May 2017.

[11]G. Matthaei, E. M. T. Jones, and L. Young, Microwave Filters, Impedance-Matching Networks, and Coupling Structures. New York, NY, USA: McGraw-Hill, 1964.

[12]M. A. Ismail and M. Yu, “Advanced design of large scale microwave devices for space applications using space mapping optimization,” in

IEEE MTT-S Int. Microw. Symp. Dig., Honololu, HI, USA, Jun. 2017,

pp.1515–1516.