Progress In Electromagnetics Research, PIER 77, 409–416, 2007
NOVEL DUAL-MODE DUAL-BAND BANDPASS FILTER
USING DOUBLE SQUARE-LOOP STRUCTURE
Z.-X. Chen, X.-W. Dai, and C.-H. Liang
National Key Laboratory of Antennas and Microwave Technology
Xidian University
Xi’an 710071, China
Abstract—A dual-mode dual-band bandpass microstrip filter using
double square-loop structure is proposed in this paper. Each of the
square-loop forms a dual-mode resonator with controllable respective
passband. Two tuning patches placed symmetrically at the side of the
perturbation patch are used to change the higher passband frequency,
while while keeping the lower invariable. Several attenuation poles in
the stopband are realized to improve the selectivity of the proposed
bandpass filter. The filter is evaluated by experiment and simulation
with good agreement.
1. INTRODUCTION
Recent development in wireless communication systems has created
a need of RF circuits with a dual-passband operation. Therefore, the
bandpass filter, as a key component filtering unwanted frequency in RF
systems, is necessary to generate two or more transmission frequency
bands. Recently, the synthesis theory of microwave filters present two
passbands mostly use frequency-variable transformations [1]. However,
the strong attenuation is required for practical applications; Many new
structures, such as stepped impedance resonators (SIRs) or parallel
coupling [2, 3] or equal-length coupled-serial- shunted lines [4], have
been proposed for a dual- band bandpass filter. In [5], a dual-mode
dual-band bandpass filter was initially reported. Unfortunately, this
solution suffers from high insertion loss and none transmission zeros
in the stopband. However, an extra matching network is needed to
combine them. Recently, the dual-mode resonator using patch [6] or
square loop [7–9] has attracted many attentions for its low radiation
and compact size in design of single band filter [11–15], which might
410 Chen, Dai, and Liang
become a candidate for dual-band bandpass filter design. A dual-
mode dual-band filter with stacked loop structure is proposed in [10].
However, the stacked loop structure may introduce higher cost and
difficulty in fabrication.
In this paper, a dual-mode dual-band bandpass filter using double
square-loop structure is presented. The filter provides two transmission
bands, and each of them is realized using a dual-mode square-loop
resonator. This structure can provide convenience to change the second
passband frequency, while the first keeps same. At the same time, there
is good isolation between the two passbands.
(a) (b)
Figure 1. (a) Conventional dual-mode microstrip bandpass filters.
(b) Proposed the structure of double square-loop.
2. DUAL-MODE SQUARE-LOOP RESONATORS
Dual-mode microstrip bandpass filters have been investigated by many
researchers for applications in both wired and wireless communication.
These filters are based on a variety of symmetric dual-mode resonating
structures. For dual-mode operation, a perturbation is introduced in
the resonator in order to couple its two degenerate modes. Fig. 1(a)
shows the layout of conventional model of dual-mode square-loop
bandpass filter [7, 10] . The square loop consists of four identical arms
in length. Different filter responses can be obtained with different
positions and size of the perturbation, which is analyzed in detail in [6].
The fundamental resonance occurs when a ≈ λ
g
/4, where λ
g
is the
Progress In Electromagnetics Research, PIER 77, 2007 411
guided wavelength.
λ
g
=
c
f
√
ε
eff
(1)
where c is the velocity of light in free space, and ε
eff
is the effective
dielectric constant of the substrate. According to (1), for a fixed
resonant frequency, a ≈ λ
g
/4 is decreased to realize size reduction
as ε
eff
increased. Similarly, for a fixed ε
eff
, the resonant frequency f
is decreased as a increased.
34567
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
S
21
(dB)
Frequency(GHz)
With inner loop
Without inner loop
Figure 2. Frequency responses of outer loop with and without inner
loop.
Figure 1(b) shows the proposed double square-loop structure,
where the outer and inner loops have different side-length. There are
two different pass bands for the proposed structure. Two connecting
patches at the diagonal corners are used to feed the inner loop, and the
tuning patches are located at different positions to realize the control
of the second passband. Given the side-length of the outer loop,
the frequency responses of the square-loop model with and without
inner loop were illustrated in Fig. 2, which shows the lower passband
has a slight shift to higher frequency. The waveguide wavelength
corresponding to the lower passband is
λ
g1
≈ 4a (2)
where a is the side-length of the outer loop. The center frequency of
412 Chen, Dai, and Liang
the lower passband is related by submitting (2) in (1).
f
01
=
c
λ
gl
√
ε
eff
≈
c
4a
√
ε
eff
(3)
1
w
3
w
w
x
2
w
p
i
q
t
mmh
r
1
65.2
a
b
g
ε
=
=
Figure 3. Proposed dual-mode dual-band bandpass filter.
3. PROPOSED BANDPASS FILTER DESIGN
The structure of the proposed dual-mode dual-band bandpass filter
is shown in Fig. 3, whose design parameters were determined as
follows: a = 15 mm, b =14.2 mm, g =0.25 mm, w
1
=0.3 mm,
w
2
=0.4 mm, p =0.5 mm, w
3
=0.3 mm, d =1.2 mm, t =13.6mm
and w =2.8 mm is the width of 50 Ω microstrip feedline. The
outer loop and inner loop provide two transmission paths for RF
signal. Based on the discussion mentioned above, the proposed
filter generates two separated passbands by using two square loops
resonating different frequency, the outer loop for lower passband, and
the inner loop for higher passband. Changing the place of tuning
patches, we can control the location of higher passband. Due to the
Progress In Electromagnetics Research, PIER 77, 2007 413
tuning patches, the resonating length of inner loop can be changed.
The simulation and experiment results disclose that the waveguide
wavelength corresponding to the higher passband should satisfy the
relation
λ
g2
≈ 4t − 2x (4)
where t is the side-length of the inner loop and x is distance between
the tuning patch and the outer perturbation corner. Submit (4) in (1),
the center frequency of the higher passband is deduced as
f
02
=
c
λ
gh
√
ε
eff
≈
c
(4t − 2x)
√
ε
eff
(5)
34567
-60
-50
-40
-30
-20
-10
0
x=7.3
x=8.8
x=10.3
S
21
(dB)
Frequency (GHz)
Figure 4. Different frequency responses with variation of x.
Figure 4 shows the variation of the location of the higher passband
frequency with the different x. The higher passband frequency is
increased as x is enlarged, while the lower passband keeps the same. So
the operating frequency of higher passband can be easily controlled.
In addition, each dual-mode loop generates own passband and two
attenuation poles at respective stopbands, which keeps good selectivity
of the proposed filter. In order to reduce the coupling between the
two perturbation patches and the mutual effects of the two passbands,
the two patches are located at inner corner of the inner loop and
outer corner of the outer loop, respectively. A demonstration filter is
optimized and measured when x =8.3 mm. Good agreements between
the simulation and measurement are achieved.