11th International Radar Symposium India - 2017 (IRSI-17)

Full U-Band Rectangular Waveguide-to-Microstrip

Transition Using E-Plane Probe
Sadhana Kumari#1 and Priyanka Mondal#2
Department of Electronics and Communication Engineering
#
National Institute of Technology Patna
Ashok Rajpath, Patna 800005, India
1
sdhn2337@gmail.com
2
pmondal@ieee.org

Abstract- A full U-Band rectangular waveguide-to-

microstrip transition using E-plane probe is presented. II. PROPOSED TRANSITION AND a
Simulations show that |S11| is below -20 dB over the SIMULATIONS
frequency band 40 GHz to 60 GHz and |S21| is below -0.27
dB in this band. Simulations have been performed using A. Structure
Monopole radiator
the 3D electromagnetic full-wave simulator HFSS from WR-19
ANSYS. For verification, HFSS simulated results are Backshort Sleeve
compared with CST Microwave Studio. Good agreements
are found over the aforementioned band. b
Index terms- Low loss, sleeve, U-band, waveguide-to-
microstrip transition.
I. INTRODUCTION Microstrip line

In radar systems, microstrip antennas are

recommended where low profile, light weight antenna
subsystem [1], [2] is in demand. Often rectangular
waveguides are used as interconnect elements due to its
low loss characteristics compared to other interconnects
at millimeter-wave and submillimeter-wave
frequencies. Transitions are essential to transform Fig. 1. Perspective view of the proposed transition.
efficiently electromagnetic energy from waveguide to
microstrip or vice versa. It requires coupling between A. Structure
TE10 mode of rectangular waveguide to the quasi-TEM As is illustrated in Fig. 1, the proposed waveguide-to-
mode of microstrip line. microstrip transition is made up of an input waveguide,
In recent years, many transitions have been proposed a waveguide backshort, a microstrip line (characteristic
using a quarter-wavelength impedance transformer impedance of 50 Ω) in a metallic enclosure, ground
along with a high-impedance inductive line [3]. plane with symmetric sleeves, and a strip probe of length
According to impedance match theory, one major approximately λ/4 (acting as a monopole radiator). The
drawback of the quarter wave transformer is its narrow corresponding probe configuration is shown in Fig.2.
bandwidth, to yield a broadband RF matching between For this type of transition, the aperture should be
the high impedance waveguide, and microstrip line one taken as large as possible, to suppress the waveguide
alternative idea on the transition probe design is the modes [5]. Assuming a 1.6-mm wide substrate for this
concept of planar monopole antenna [4]. U-Band transition, an aperture size of 1.6 mm × 0.5 mm
In this paper, a U-band low loss broadband waveguide has been used. The corresponding cutoff frequency of
to microstrip E-plane probe using symmetric ground waveguide mode is around 93.7 GHz, which is far
sleeves. Parametric studies are performed to analyze beyond the operation frequencies of the U-Band
dimensional sensitivity of the proposed structure. transition.

NIMHANS Convention Centre, Bangalore INDIA 1 12-16 December, 2017

 11th International Radar Symposium India - 2017 (IRSI-17)

The design concept of this transition is adapted from fundamental resonant frequency, and it affects second
the monopole antenna with the ground plane sleeve [6]- resonant frequency significantly. Distance between
[8]. Sleeve can be treated as an extension of ground
0
probe length (L_PR) 1.42 mm
-10 1.4 mm
1.44 mm
-20

|S11| (dB)
d
W_L -30
W_SL

L_PR
-40
S_L D

-50
37 42 47 52 57 62

Frequency (GHz)
L_gnd
(a)

sleeve length (S_L) 810 µm

-10 800 µm
790 µm
SW
|S11| (dB)

Fig. 2. The configuration of the transition probe. -20

plane inside waveguide [7], and generates an additional

-30
resonant mode above the fundamentally generated
mode, which helps to attain a wider bandwidth. In [7], it
is found the widest bandwidth can be achieved while -40
keeping the length of the sleeve (S_L) one third of the
f0R = 46 GHz
monopole probe radiator length (L_PR). So, S_L = λ/12 -50
and L_PR = λ/4 have been taken as initial value. 37 42 47 52 57 62
However, fine tuning is required as the transition probe Frequency (GHz)
is enclosed inside metallic enclosure. (b)
Although the input impedance at the microstrip seen
at the microstrip port is independent of the probe width,
0
a narrow probe might limit the transition’s matching
bandwidth and results in excessive inductive loss [5], spacing distance (D) 560 µm
[9]. To simplify the design, here it is taken as 50 Ω. -10 550 µm
570 µm
B. Simulations
|S11| (dB)

-20
3-D EM simulator HFSS from ANSYS is used for
simulations. The transition probe has been implemented
-30
using a WR-19 rectangular waveguide and a 5-mil thick
5880 RT/duroid substrate from Roger having
permittivity 2.2 and loss tangent 0.0009. The conducting -40
material is copper of conductivity 5.8×107Siemens/m.
The broad wall and narrow wall dimensions of the -50
waveguide are a = 4.7752 mm and b = 2.3876 mm, 37 42 47 52 57 62
respectively. The width of the 50 Ω microstrip line for Frequency (GHz)
the present substrate is ≈ 0.39 mm and the length of the
(c)
radiator inside waveguide is ≈ 1.37 mm at 40 GHz.
Fig. 3(a), (b) clearly indicate the probe length affects
the fundamental resonant frequency (around 46 GHz)
but the sleeve length does not have significant effect on

NIMHANS Convention Centre, Bangalore INDIA 2 12-16 December, 2017

 11th International Radar Symposium India - 2017 (IRSI-17)

0 The configuration of the back-to-back transition is

shown in Fig. 5. Simulation results with the microstrip
-10 line width (L_W) 390 µm line length variation 2λ and 4λ, λ @ 50 GHz are shown
410 µm
400 µm 0 0
|S11| (dB)

-20

-30 -10 -0.2

-40 -20 -0.4

|S21| (dB)
|S11| (dB)
-50 -30 -0.6
37 42 47 52 57 62 included
Rq
excluded
-40 -0.8
Frequency (GHz) included
excluded
-50 -1
(d)
37 42 47 52 57 62
Fig. 3. Parametric study of the designed transition with variation of
(a) probe length (b) sleeve length (c) spacing distance between probe Frequency (GHz)
and sleeve (d) probe width which is equal to microstrip line width.
Fig. 4. Optimized S-parameters of the designed transition with and
Table I without surface roughness.
Optimized dimension of the parameters
Variable Value in mm
W_L 0.4
L_PR 1.4
S_L 0.79
D 0.57
d 0.16
SW 1.6
L_gnd 4.86
Aperture height 0.509

monopole and sleeve (D) kept fixed as 560µm for the

above simulations. Simulated results with variation of D
are shown in Fig. 3(c). The figure indicates that D has
slight effect on both fundamental as well as second Fig. 5. A back-to-back transition
resonant frequencies, and can be used to improve the
matching. Fig. 3(d) indicates that line width does not 0
have significant effect on resonant frequency but
improves the matching around these frequencies. line_length = 4λ
-10
Surface roughness increases the conductor loss as line_length = 2λ
frequency increases especially when the signal skin -20
depth is comparable or smaller than the conductor
|S11| (dB)

-30
roughness [10]. Calculated skin depth at high end of
frequency band is 0.27 µm, and the surface roughness -40
for the chosen substrate given in the datasheet is 0.4 µm
on the dielectric side. Optimized values of S-parameters -50
with the effect of surface roughness are shown in Fig. 4.
Final optimized structural parameters are listed in Table -60
I.
-70
Fig. 3 has clear evidence that dimensional variations
37 42 47 52 57 62
of ± 10 µm do not have any significant effect on the S-
Frequency (GHz)
parameters in frequency range of interest. Thus, the
design can sustain fabrication tolerance ± 10 µm given (a)
by Rutherford Appleton Laboratory, UK.

NIMHANS Convention Centre, Bangalore INDIA 3 12-16 December, 2017

 11th International Radar Symposium India - 2017 (IRSI-17)

0 used to design excellent probe transition at other

-0.2 frequency band.
-0.4 REFERENCE
-0.6
40 GHz [1] K.S.Beenamole, "Microstrip Antenna Designs for Radar
-0.8 60 GHz Applications," DRDO Science Spectrum, pp. 84-86, March 2009,.
|S21| (dB)

-1 [2] Iizuka, H., Watanabe, T., Sato, K., & Nishikawa, K. (2002).
-1.2 “Millimeter-wave microstrip line to waveguide transition
fabricated on a single layer dielectric substrate.” IEICE
-1.4 line_length 4λ Transactions on Communications, vol. E 85 (B), no. 6, 1169-
-1.6 2λ 1177, June 2002.
[3] Leong, Yoke-Choy, and Sander Weinreb. "Full band waveguide
-1.8 to-microstrip probe transitions." IEEE MTT-S International
-2 Microwave Symposium Digest, vol. 4, 1999.
37 42 47 52 57 62 [4] Li, Zengrui, and Junhong Wang. "A novel broad band antenna of
Frequency (GHz) planar monopole." IEEE Asia-Pacific Microwave Conference
Proceedings. vol. 4, 2005.
[5] Shi, S. C., and J. Inatani. "A waveguide-to-microstrip transition
Fig. 6. Simulation (a) |S11| and (b) |S21| of the back-to-back transition with a DC-IF return path and an offset probe." IEEE
configuration. Ttransactions on Microwave Theory and Techniques, vol. 45, no.
0 0.1 3, pp. 442-446, March 1997.
[6] Shen, Zhongxiang, and Cheng Qian. "A broad-band double-sleeve
-6E-16 monopole antenna," IEEE Asia-Pacific Microwave Conference
-10 Proceedings. vol. 4, 2005.
-0.1 [7] Lin, C-C., K-Y. Kan, and H-R. Chuang. "A 3-8-GHz broadband
-20 planar triangular sleeve monopole antenna for UWB
|S11| (dB)

|S21| (dB)

-0.2 communication." IEEE Antennas and Propagation Society

-30 International Symposium, 2007.
-0.3 [8] Zhongxiang Shen and R. H. MacPhie, "Rigorous evaluation of the
EM simulator used CST
-40 input impedance of a sleeve monopole by modal-expansion
HFSS -0.4 method," in IEEE Transactions on Antennas and Propagation,
CST vol. 44, no. 12, pp. 1584-1591, Dec 1996.
-50 -0.5
HFSS [9] Shih, Y-C., T-N. Ton, and Long Q. Bui. "Waveguide-to-microstrip
-60 -0.6 transitions for millimeter-wave applications." IEEE MTT-S
37 42 47 52 57 62 International Microwave Symposium Digest, 1988.
[10] Horn III, Allen F., et al. "Effect of conductor profile on the
Frequency (GHz) insertion loss, phase constant, and dispersion in thin high
frequency transmission lines." Design Con, 2010.
Fig. 7. S-parameters obtained from two different EM simulators.

BIO DATA OF AUTHOR(S)

in Fig. 6. It indicates a good transition design over the
desired frequency range i.e. 40-60 GHz. Sadhana kumari received the
It is appropriate to verify the simulation accuracy and B.Tech degree in Electronics and
reliability by comparing the results obtained from Communication Engineering from the
M.J.P Rohilkhand University,
another simulator. Fig. 7 shows that S-parameters Bareilly, Uttar Pradesh, India, and the
obtained from HFSS are in a good agreement with that M.tech degree in VLSI and embedded
obtained from CST Microwave Studio. system from the NIT Rourkela,
Rourkela, Odisha, India in years 2009
and 2011 respectively. Currently she
III. CONCLUSION is pursuing Ph.D degree in the area of
A U-band low-loss waveguide to microstrip probe microwave passive components. Her
transition has been designed. The design is simulated in research interests include passive mixer, microstrip to waveguide
transition, microstrip line filter.
two different 3-D EM simulators, and results are found
to be in a good agreement. Parametric studies have been
performed to obtain the dimensional sensitivity of the
structure. It is seen that the design can sustain
fabrication tolerance ± 10 µm which is within the limit
given by Rutherford Appleton Laboratory, UK.
The transition could be utilized in millimeter wave
communication systems in U-band, which has a
promising future in astrophysics, or atmospheric
sounding and remote sensing. In addition, the improved
design approach proposed in this paper also could be

NIMHANS Convention Centre, Bangalore INDIA 4 12-16 December, 2017