Transition Tackles MM-Wave Frequencies

Some complex approaches have been developed in pursuit of waveguide-to-CPW transitions.^3,4 In a continuous transformer, the waveguide ridge acts as the center conductor of the CPW, and such a configuration can gradually transform the CPW’s electric fields into the transverse electromagnetic (TE₁₀) electric fields of rectangular waveguide. A waveguide-to-CPW transition with enhanced performance at X-band employed a discrete three-step Chebyshev transformer.⁵ No bond wires are needed to suppress the CPW odd mode in those transitions; however, their assembly becomes very difficult at millimeter-wave frequencies, incurring high fabrication costs.

For simpler fabrication, transition circuits of the planar variety were developed.^6-16 Among these waveguide-to-CPW transitions, an open-ended planar probe fed by asymmetric CPW was employed to excite the rectangular-waveguide TE₁₀ mode.⁶ Although the bandwidth of this transition is extremely wide, air bridges are needed to suppress the CPW odd mode resulting from the asymmetric discontinuity.

Some broadband transitions have been achieved by using a tapered finline.^7-10 In ref. 7, the air bridge is eliminated by means of a special CPW housing design. In ref. 10, there is likewise no need for an air bridge, but the bandwidth is limited by the design’s phase shifter, and an intermediate transition of CPW to microstrip is needed.⁹ Those transition circuits may occupy an electrically large area due to the use of a long tapered finline.^7-10

Some researchers have presented waveguide-to-CPW transitions using an open-ended tapered strip.^11,12 Although the circuit size for this approach is smaller than finline-based transitions, there is the need for an intermediate transition from CPW-to-slotline and air bridges are also required. The transition with conductor-backed CPW¹² has a narrower bandwidth than the other configurations and also requires viaholes.

In ref. 13, an end-fire antenna was applied in the design of a waveguide-to-CPW transition. A truncated bow-tie antenna was used for broadband performance,¹⁴ but it still requires an intermediate transition and air bridges.^13,14 In one design, a right-angle transition between waveguide and CPW was achieved by using slot antennas.¹⁶ The approach doesn’t require an air bridge, but it is narrowband. More recently, a miniaturize waveguide-to-CPW transition was realized by using inductance-compensated slotline technology, which involves multiple-section slotline.¹⁵ But this multiple-section design has some problems at millimeter-wave frequencies, and also requires an air bridge.

Novel Transition

In pursuit of simplicity, a new type of waveguide-to-CPW transition was developed based on coupling probes; it requires neither air bridges nor an intermediate transition. An experimental version was fabricated using WR-28 rectangular waveguide and RT/Duroid 5880 printed-circuit-board (PCB) material from Rogers Corp. A back-to-back transition prototype was fabricated; the scattering parameters were measured to verify the performance of the design. The transition, with good performance at Ka-band frequencies, offers advantages of ease of fabrication, broad bandwidth, and low loss.

Figure 1 shows the configuration for the proposed transition between rectangular waveguide and CPW. The transition consists of rectangular waveguide, antisymmetric tapered probes, symmetric radial stubs, and CPW. The rectangular waveguide is standard WR-28 with dimensions of 7.112 × 3.556 mm, which will ensure single-mode TE₁₀ propagation at Ka-band frequencies.

RT/Duroid 5880 PCB material was used as the circuit substrate material for the transition; the PCB material has relative permittivity of 2.22, loss tangent of 0.009, and thickness of 0.254 mm. The waveguide-to-CPW transition is a planar, double-sided circuit. CPW is fabricated on the top of the substrate, symmetric radial stubs are fabricated on the bottom of the substrate, and antisymmetric tapered probes are fabricated on both sides of the substrate.

Figure 2 shows the cross-sectional E-field distribution for the proposed transition. The E-field is parallel to the X/Y plane in the CPW section. The radial stubs serve to change the direction of the E-field from parallel to vertical. The antisymmetric tapered probes then transform the z-polarized E-fields energy in the input port of the antisymmetric tapered probes to the y-polarized E-field energy of the TE₁₀ mode in the rectangular waveguide. In this way, the CPW E-field can smoothly change to the TE₁₀ E-field mode in rectangular waveguide.

Figure 3(a) shows the schematic circuit for the proposed transition, while Fig. 3(b) presents the circuit layout on the top side of the substrate. The characteristic impedance of the CPW was set to 75 Ω, with the widths of the input ports of the antisymmetric probes chosen accordingly. Ground planes are circularly tapered to minimize reflections, with radius R₁ set to 1 mm. The two antisymmetric probes are also tapered in a smooth outline to obtain a good reflection coefficient.

Figure 3(c) depicts the circuit layout on bottom sides of the substrate. The end of the bottom probe is directly connected to the two symmetric radial stubs. The radial stubs are rotated with an optimal angle of 16 deg. to smoothly change the electric field from a y-axis direction to a z-axis direction. The overlap area between the radial stubs and the CPW ground planes should be optimized to provide sufficient electromagnetic (EM) coupling and to minimize reflections.

The radial stub is optimized with a radius, R₂, of 1.85 mm and an angle of 31 deg. The length of the entire transition circuit is approximately 4.6 mm. No air bridge is needed for this transition due to the symmetric structure of the radial stubs. Figure 4 shows the fabricated substrate of the transition circuit.

The proposed transition was simulated and optimized using the High-Frequency Structure Simulator (HFSS) EM simulation software from Ansoft Corp. The dimensions of the transition circuit are shown in Fig. 3. The final value of the parameters are W_cpw = 0.55 mm, W_sg = 0.15 mm, W₁ = 1.4 mm, W₂ = 1 mm, W₃ = 0.26 mm, W₄ = 0.4 mm, W₅ = 0.38 mm, W₆ = 0.5 mm, L₁ = 1.38 mm, L₂ = 2.13 mm, and L₃ = 2.74 mm. To experimentally evaluate the performance of the proposed transition, two identical transitions were cascaded back to back.Figure 5 shows the fabricated back-to-back transitions. Measurements were made with a commercial vector network analyzer (VNA) and a pair of coaxial-line-to-waveguide adapters. A thru-reflect-line (TRL) calibration was performed at the WR28 access points. Figure 6 shows the simulated and measured results of the back-to-back transitions, which are in close agreement. The measured results show that the insertion loss of back-to-back transition is less than 0.7 dB, with better than 15-dB return loss from 26.5 to 40 GHz. This implies that the insertion loss of a single transition is less than 0.35 dB across Ka-band frequencies.

In short, a broadband waveguide-to-CPW transition has been proposed based on tapered probes and symmetric radial stubs. Neither an air bridge nor an intermediate transition is needed for this transition, making it relatively easy to fabricate. To demonstrate the effectiveness of the design approach, a Ka-band back-to-back transition was fabricated and measured, with close agreement between simulated and measured results Given its advantages of ease of fabrication, compact size, and low loss, such a broadband transition should be attractive for a wide range of millimeter-wave applications.

Acknowledgment

This work was supported by the Fundamental Research Funds for the Central Universities of China under Grant No. ZYGX2012J030.

Jun Dong, Doctoral Candidate

Ziqiang Yang, Associate Professor

Yu Liu, Associate Professor

Tao Yang, Professor

School of Electronic Engineering, University of Electronic Science and Technology of China, No.2006, Xiyuan Ave, West Hi-Tech Zone, Chengdu 611731, People’s Republic of China