Substrate Integrated
Substrate Integrated
Substrate Integrated
Introduction
One particular challenge facing SIW circuits is their integration with active components that
require DC biasing. An inherent DC short-circuit is formed between conventional SIWs’ top
and bottom waveguide plates due to the conducting vias. Several schemes have been
proposed to address this issue. They include the use of discrete capacitors on microstrip lines
leading up to the SIW circuit [5], embedding an interdigitated capacitor into the SIW-to-
microstrip transition [6], and using AC coupling on half mode SIW [7]. None of the existing
methods allow direct shunt connection of active components, such as Gunn diodes, into a
waveguide. Compromises are usually made to accommodate the need of DC bias; and
consequently, the advantages of SIW structure are not fully utilized and the solution is
suboptimal.
In this work we propose a new SIW structure, called the corrugated SIW (CSIW), which does
not require conducting vias to achieve TE10 type boundary conditions at the side walls.
Instead, the vias are replaced by quarter wavelength microstrip stubs arranged in a corrugated
pattern on the edges of the waveguide. This, along with series interdigitated capacitors [6],
results in a waveguide section comprising two separate conductors, which facilitates shunt
connection of active components such as Gunn diodes [8]. The concept of this arrangement is
shown in Fig. 1.
The layout of a CSIW and its critical dimensions are given in Fig. 2. The waveguide width (a),
and 50 microstrip to SIW transition are designed according to previously described
methods [3, 9]. They can be fine tuned and optimized separately by way of EM simulations.
The length of the corrugated wall stubs (t) is approximately one quarter of the TEM guided
wavelength (g) at the centre frequency. This estimation is refined by EM simulation which
accounts for fringing effects and transition from TE10 mode to quasi-TEM mode at the side
wall.
Different combinations of stub spacing (u) and width (v) have been investigated. We
conclude that they have very little impact on the overall performance of the CSIW. Rules for
choosing equivalent dimensions in conventional SIWs can be applied here and values suitable
for fabrication should be adopted for the final design.
Authorized licensed use limited to: University of Canterbury. Downloaded on November 29, 2009 at 21:01 from IEEE Xplore. Restrictions apply.
Application of CSIW in X-band
Fig. 3 shows a set of back-to-back SIWs, corrugated and conventional, fabricated on 60 mil
thick Taconic RF-30 (r = 3.0) PCB with 35 μm copper layer on both sides. Two sets of SIWs
were made with different waveguide section lengths: one 81 mm and another 46.5 mm.
The SIWs were designed for X-band (8.2 GHz to 12.4 GHz), and the resulting structure
parameters are given in Table 1. The internal width (a) and the microstrip transitions are the
same for both SIW and CSIW. The parameters d and s refer to the via diameter and spacing
for the conventional SIW respectively. The substrate thickness, and hence waveguide height,
is denoted as b. The surface current density plot for both SIW and CSIW at 10 GHz is shown
in Fig. 4 and this is consistent with the TE10 mode.
a b r d s ml mw w l u v t
13.5 1.524 3.0 1.0 2.0 4.0 4.0 7.0 17.5 1.0 1.0 5.5
Table 1. Optimized dimensions for conventional and corrugated SIW in mm.
S-parameter measurements were performed using an Agilent E8362B 20 GHz vector network
analyzer. Results for the 46.5 mm waveguides are given in Fig. 5. Measured return loss is
better than 10.3 dB for CSIW and 8.6 dB for SIW; global insertion loss is lower than 3.5 dB
for CSIW and 2.9 dB for SIW.
Based on the multiline calibration method [10], the propagation constants (Fig. 6) were
extracted from the measurements of the two lengths of CSIW and SIW. The attenuation
constant is less than 0.48 dB/cm for CSIW and 0.45 dB/cm for SIW.
The discrepancies between measurement and simulation are largely attributed by the
irregularities in the fabrication process; nonetheless, the results are sufficient to confirm that
the CSIW offers performance comparable to that of the conventional SIW.
The interdigitated capacitor on the DC decoupled CSIW (Fig. 7) was designed following a
method similar to [6]. It uses 13 fingers of 3.2 mm length, and a gap of 0.3 mm. Measured
results (Fig. 8) suggest that the two capacitors increased insertion loss by approximately 0.5
dB overall when compared to CSIW results in Fig. 5.
Conclusion
A novel corrugated SIW has been shown to have comparable performance to the
conventional SIW. It has the additional advantage of complete DC isolation between its top
and bottom plates. This permits shunt connection of active components in the waveguide for
a variety of applications.
Acknowledgement
The authors would like to thank Nick Smith, David Healy, and Michael Cusdin for their
assistance in making the PCBs and test fixtures.
Authorized licensed use limited to: University of Canterbury. Downloaded on November 29, 2009 at 21:01 from IEEE Xplore. Restrictions apply.
References
Authorized licensed use limited to: University of Canterbury. Downloaded on November 29, 2009 at 21:01 from IEEE Xplore. Restrictions apply.
Figure 3. SIW and CSIW with 46.5mm waveguide Figure 4. Surface current density plot of SIW and
length. CSIW at 10 GHz.
(a) (b)
(a) (b)
Figure 6. Propagation constant of the waveguides. (a) Attenuation constant. (b) Phase constant.
Authorized licensed use limited to: University of Canterbury. Downloaded on November 29, 2009 at 21:01 from IEEE Xplore. Restrictions apply.