Ka-band waveguide-to-microstrip transition design and implementation

A low-loss broadband waveguide-to-microstrip transition has been developed successfully. The design procedure is based on the achievement of a matched impedance probe through the use of a high impedance line and a quarter wavelength transformer. This device is used in Ka-band systems. INTRODUCTION RF subsystems of the new Ka-bands systems (e.g., LMDS: Local to Multipoint Distribution System) require low-loss waveguide as passive components (e.g., RF filters and diplexers) technology. However the active subsystems are usually realized in MMIC (Monolithic Microwave Integrated Circuits) or hybrid MIC (Microwave Integrated Circuits) technologies. In many cases microstrip lines are the preferred energy transmission structures between these subsystems and the RF subsystem. Therefore a low-loss waveguide-to-microstrip transition is required. Several design approaches exist [1],[2]. In this paper a compact design leading to broadband low-loss waveguide-to-microstrip transition is presented. The method is based on a probe that extends naturally from a microstrip line [I]. The design approach is developed in the context of a particular design for a WR28 waveguide and SOR-microstrip line supported by CuClad substrate (~ ,=2 .17 , H=0.254 mm) for a frequency band from 27 GHz up to 35 GHz. Two transitions have been constructed and measured showing excellent performances. DESIGN The transition consists of an extension of a printed microstrip circuit through an aperture in the broad wall of a short-circuited waveguide as depicted in figure l(a). The.meta1 strip supported by the substrate couples the energy of the TElo mode of the waveguide and the energy of the quasi-TEM mode of the microstrip line in a similar way as a coaxial probe in a waveguide-to-coaxial transition. The aperture of the waveguide must be as small as possible in order not to perturb the field distribution in the waveguide. The maximum power transmission is reached when the microstrip line is matched at the feed point of the probe. 0-7803-733@8/02 $ 1 7 . 0 0 0 2 ~ I E E E 404 The electromagnetic behavior of the probe has been analyzed with a finite element electromagnetic analysis tool (Maxwell 3D Electromagnetic Simulator). The frequency variation of the reflection coefficient at the feed point depends on,the probe length (L), the distance of the waveguide short-circuit from the probe (D), and the probe width (W). In order to obtain a broadband design, i.e., to minimize the mentioned frequency variation, a parametric analysis of the reflection coefficient at the feed point with respect to the above-mentioned parameters has been realized. The optimized reflection coefficient is calculated through the analysis of the structure depicted in figure I(a), followed by a translation of the reference plane of the microstrip line port to the probe feed point. This analysis was performed for the WR28 waveguide and using CuClad substrate. The optimum values for L, D and W were thus obtained. The reflection coefficient of the optimized geometry is shown in figure 2. In order to obtain a matched structure a high-impedance transmission line (that simulates an inductor in series with the probe) is placed after the probe as shown in figure I(b). This structure was analyzed and the reflection coefficient obtained is presented in figure 3. It can be seen that the reflection coefficient is almost real, and can be easily matched to a 5OR-microstrip line with a quarter wavelength transformer. The final structure is shown in figure l(c). The HFSS simulation of the final transition is presented in figure 4. MEASUREMENTS Two transitions were fabricated and connected back-to-back through a 14 mm long microstrip line in a test-fixture. Figure 5 shows the measured insertion-loss and returnloss. The maximum insertion-loss is 0.95 dB and the minimum return-loss is 18 dB from 27 GHz up to 35 GHz. The estimated total insertion-loss for the WR28 waveguides of the test fixture was 0.2 dB, and the estimated insertion-loss for the 14 mm long microstrip line was 0.15 dB. Thus, after accounting for these losses, the obtained insertion-loss per transition is 0.3 dB approximately. CONCLUSION A Ka-band low-loss broadband waveguide-to-microstrip transition has been presented. The measured results of two transitions show excellent performance. The same design approach can be used to design optimum transitions at other frequencies and using other substrates. ACKNOWLEDGEMENTS The authors would like to thank Josd Mellado-Bema1 and Josd-Maria MonteroSerrano (both with Grupo de Microondas y Radar: GMR) for the construction and the assembly of the transitions. This work has been done under a contract between IKUSl S.A. (Spain) and GMR.