Circuits and Systems, 2011, 2, 293-296
doi:10.4236/cs.2011.24041 Published Online October 2011 (http://www.scirp.org/journal/cs)
Study and Enhanced Design of RF Dual Band Bandpass
Filter Validation and Confirmation of Experimental
Measurements
Mohamed Mabrouk, Leila Bousbia
CIRTACOM and ISETCOM of Tunis, University of Carthage, Tunis, Tunisia
Received July 22, 201 1; revised August 12, 2011; accepted August 21, 2011
Abstract
Dual band bandpass filter is designed and optimized for RF wireless applications. The performances of that
RF dual band filter are improved especially parameters describing the insertion loss, return losses and rejec-
tion. Dual band bandpass filter using stub loaded resonators is designed and characterized. Theoretical re-
sults are compared with experimental data. This comparison shows that the magnitude of reflection coeffi-
cient S11 from ADSTM simulation is better than 28.0 dB, and the insertion loss S21 is less than 0.5 dB. The
two rejections are also better than 32.0 dB. The simulated results also show that two central frequencies are
located at desired 1.82 and 2.95 GHz. Comparison of measured and simulated results shows frequency drift.
The main reason for this frequency shifting is due to some uncertainties. These are obviously due to geomet-
rical and physical parameters H and r respectively of Duroid substrate used during design and measure-
ments.
Keywords: RF Filters, Dual Bandpass, Effective Permittivity, Physical and Geometrical Parameters
1. Introduction
RF circuits with a dual pass band operation are required
by modern wireless communication systems [1]. For
example, the RF transceiver length (TX/RX) for second
generation GSM and third-generation WCDMA mobile
communications must be able to receive and transmit
900 MHz and 1900 MHz signals. Also 2.4 GHz and 5.2
GHz bands are two bands operated by high-speed
wireless LANs [1,2]. The band pass filter is necessary to
generate two or more transmission frequency bands, in
particular dual-band filters, as they hav e an essential role
in transmit-receive systems [2]. RF filters have par-
ticularly an important role for signals filtering, rejection
and isolation between parts in transceivers systems
(TX/RX). Filters have the characteristic to be frequency
selective devices for transmitting and attenuating signal
in desired frequency ranges [3]
Many different ways are considered to design dual-
band bandpass filters such combination of two filters
operating on two different bandwidths [4] and using dual
band stepped impedances resonator [5] or dual band stubs
resonators [6]. Quasi lumped with open loop band pass
filters operating at different frequency bands [7] and
using square loop dual mode resonators [8] are also used
for designing this type of filters. Small insertion loss, low
return loss and high rejection band are the desired
characteristics of a good filter. The design of this kind of
filter is considered using electromagnetic (EM) simu-
lators. The studied filter was simulated usin g IE3DTM [6],
and in this paper we report our ADSTM simulation
optimized results which are very close to the experimental
measurements. ADSTM software has been used succe-
ssfully many times for simulation design filter so th at the
resulting performances meet the microwave filter speci-
fications. Simulated results confirm that two central
frequencies are located at desired values 1.82 GHz and
2.95 GHz.
2. Design Description of Band Pass Filter
The studied filter is composed of two ring-resonators
loaded with two open stubs. Figure 1 shows the physical
layout of the dual band pass filter using uniform
microstrip lines. This filter consists of two microstrip
open loops. Two open circuited stubs are attached [6] at

M. MABROUK ET AL. 294
the center of the respective microstrip lines. The total
length of each resonator is around half wave length
(

g/2).
The design parameters dimensions are chosen [6] as
following: W = 1.2 mm; W1 = 1.5 mm; W2 = 0.9 mm;
L=11.2 mm; L1 = 6.7 mm; g1 = g2 = 0.3 mm. The struc-
ture of open-loop resonators filter with center frequency
of 1.82 GHz and 2.94 GHz is de signed on an RT Duroid
6006 substrate with a thickness H = 0.635 mm ± 0.0254
and a relative dielectric constant

r = 6.15 ± 0.15 [9].
3. Analysis and Comparison of Simulation
with Measurement Results
Figures 2 and 3 illustrate simulated and measured S21
and S11 parameters respectively of our studied dual band
filter.
Simulated results show that two central frequencies
are located at the desired values 1.82 GHz and 2.95 GHz,
and a small frequency shifting is observed. Thus, the
ADSTM simulation results are shifted of –5.0 MHz,
(1.825 GHz instead 1.830 GHz), and of +13.0 MHz
(2.953 GHz instead 2.940 GHz). To understand the main
reason for this frequency shifted aberration which is
certainly due to the uncertainties on the guided wave
length

g, we have highlighted the following rela-
tionships:
g
eff
c
F
 ε 1
where F is the center frequency of the filter and c is the
light velocity. The effective dielectric constant is de-
pending on relative permittivity

r and ratio WH be-
tween the width W of the transmission line and the
thickness H of the substrate which is given in the fol-
lowing forms. The RT Duroid 6006 laminate substrate
is available with relative permittivity

r value

r = 6.15
± 0.15.
It’s known that the Equation (2) shows below the
effective permittivity

eff of the microstrip structure, used
Figure 1. Layout of dual band filter.
0.5 11.5 22.5 33.5 44.5
-70
-60
-50
-40
-30
-20
-10
0
Simulation
Measurement
Figure 2. Simulated and measured S21 parameters of dual
band filter.
0.5 11.5 22.5 33.5 44.5
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
Simulation
Mea surem ent
Figure 3. Simulated and measured S11 parameters of dual
band filter.
in our design, depends on the relative dielectric constant

r, which is one of physical parameters of Duroid
substrate, and the guided wave length

g also depends on

eff, so the resonant frequencies of our filter are closely
depending on g according to the Equation (1).
For WH  1, the empirical relationship of the
effective permittivity is summarized below [10-12]:
1
2
rr
eff 11 H
112
22 W





 


ε

2
Because of the uncertainty on the dielectric relative
permittivity

r of Duroid substrate recalled previously
(in §2), we can evaluate

eff as following :
eff
4.476 4.553


 3

M. MABROUK ET AL.295
g
76 mm78 mm

 4
Thus, we can deduce that the first resonant fre-
quency FC1 is the following
C1
1.79 GHzF1.84 GHz 5
For the second frequency FC2, we can also do the
same line of argument as following
g
47mm48 mm  6
That means FC2 is:
C2
2.94 GHzF3.06 GHz 7
These equations confirm the first explicit reason for
this frequency shifting is due to the uncertainty on the
first physical parameter, i.e. the dielectric relative per-
mittivity 
r of Duroid substrate.
In the same line of argument, we can see that the
uncertainty on the substrate thickness H can lead the
following limits for the first frequency FC1:
eff
4.47 4.507

 8
g
77mm78mm

 9
C1
1.810 GHzF1.850 GHz 10
For the second frequency FC2:
eff
4.47 4.507

 11
g
47mm48 mm

 12
C2
2.910 GHzF3.0 GHz 13
Here again, it’s confirmed that the uncertainty on
thickness H can obviously contribute to the frequency
shifting that we observed, and the differences between
center frequencies are surely due to the cumulative
effects of both uncertainties on 
r and H. We have
made a comparative study of our simulation results and
experimental measurements obtained and provided by
our partner Zhang. Nevertheless, these differences that
we observed on the frequency values, between our
simulations and the measurements of our partner
Zhang, are remaining very small. Thus, the difference
of 5.0 MHz on the first resonance frequency (1.830
GHz) between our simulation results and the mea-
surements of Zhang is about 0.27%, while the dif-
ference of 150.0 MHz observed (1.680 GHz instead
1.830 GHz) between the first simulation results of
Zhang and the experimental measurements is about
8.0%. The difference of 13.0 MHz on the second
resonance frequency (2.94 GHz) between our simu-
lation results and the measurements of Zhang is about
0.44%, and the difference of 130.0 MHz observed by
Zhang (2.81 GHz instead 2.94 GHz) is about 4.4%.
The little disagreement we have obtained with the
experimental measurements confirms our predictions,
and our indepth analysis shows that we can notably
improve some characteristics of the studied filter with
going in detail and making further development, con-
sequently of the global performances become en-
hanced.
Regarding the insertion and return losses, Figure 4
shows simulated and measured insertion loss S21 and
return loss S11 parameters respectively of our dual band
filter. We obtained S21 significantly lower than 0.28 dB
instead of 0.9 dB measured at FC1, and 0.4 dB instead
of 1.1 dB measured at FC2 respectively. Our simulated
S21 results depicted on Figure 2 are up to 10.0 dB
better at 4.0 GHz than the measurements. Figure 3 also
shows that our simulated results of return loss S11 are
up to 8.0 dB better at 4.0 GHz than the measurements,
and the obtained return loss S11 is better than 28.0 dB.
Moreover, the simulated filter using two open loop
ring resonators shows better rejections than the meas-
ured ones, 31.0 dB instead of 28.0 dB measured at 0.5
GHz and 37.0 dB instead of 31.0 dB measured at 4.5
GHz. Table 1 shows comparative results.
4. Conclusions
We have studied and enhanced the design of dual band
bandpass filter for RF and wireless applications. The
performances of RF dual-band filter are improved
especially parameters describing the insertion loss, return
losses and rejections. We have obtained a good agree-
ment between our simulations and experimental results.
Insertion loss S21 lower than 0.5 dB and return loss S11
better than 28.0 dB were obtained from ADS simulations.
The rejections are also better than 32.0 dB. Simulated
results show that two central frequencies are located at
desired 1.820 and 2.950 GHz. Comparison of measured
and simulated results shows frequency shift. This is
obviously due to the uncertainties on the geometrical and
Table 1. Comparison of Zhang measurements [5] and our
simulations.
Measurement Results
of Zhang [5] Our results
Resonance Frequency FC1 1. 830 GHz 1.825 GHz
Resonance frequency FC2 2.940 GHz 2.953 GHz
Band Pass 1 9.4% 6.7%
Band Pass 2 7.5% 4.37%
Insertion losses at FC1 0.9 dB 0.28 dB
Insertion losses2 at FC2 1.1 dB 0.4 dB
Return loss at FC1 24.0 dB 29.25 dB
Return loss at FC2 20.0 dB 28.81 dB
Rejection at 0.5 GHz 28.0 dB 31.64 dB
Rejection at 4.0 GHz 31.0 dB 38.64 dB

M. MABROUK ET AL. 296
0.5 11.5 22.5 33.5 44.5
-70
-60
-50
-40
-30
-20
-10
0
Frequency [GHz]
Magnitude [dB]
S
11
Simulation
S
21
Simulation
S
11
Meas ure me nt
S
21
Meas ure me nt
Figure 4. Simulated and measured S11, S21 parameters of dual-band filter.
physical parameters H and 
r respectively of Duroid
substrate used during design and measurements.
5. Acknowledgements
The authors would like to acknowledge and extend their
gratitude to our partner Xiu Yin Zhang, Associate Pro-
fessor with the School of Electronic and Information
Engineering (South China University of Technology,
Guangzhou, 510641, China) who has carried out and
provided the experimental measurements of studied pro-
totype of filter.
6. References
[1] C. Y. Chen, C. Y. Hsu and H. R. Chuong, “Design of
Miniature Planar Dual Band Filter Using Dual-Feeding
Structures and Embedded Resonators,” IEEE Microwave
and Wireless Components Letters, Vol. 16, No. 12, 2006,
pp. 669-671. doi:10.1109/LMWC.2006.885621
[2] S. F. Chang, Y. H. Jeng and J. L. Chen, “Dual-Band Step
Impedance Band Pass Filter for Multimode Wireless
LANs,” Electronic Letters, Vol. 40, No. 1, 2004, pp. 38-39.
doi:10.1049/el:20040065
[3] I. C. Hunter, L. Billonet, B. Jarry and P. Guillon, “Micro-
wave Filters—Applications and T echnology,” IEEE Micro-
wave Theory and Techniques Transactions, Vol. 50, No.
3, March 2002, pp. 794-805.
doi:10.1109/22.989963
[4] H. Miyake, S. Kitazawa, T. Ishizaki, T. Ymanda and Y.
Nagatomi, “A Miniaturized Monolithic Dual-band Filter
Using Ceramic Lamination Technique for Dual Mode
Portable Telephones,” IEEE MTT-S International Micro-
wave Symposium Digest, 8-13 June 1997, Vol. 2, pp.
789-792.
[5] J. T. Kuo, T. H. Yeh and C. Yeh, “Design of Microstrip
Bandpass Filters with a Dual Pass Band Response,” IEEE
Microwave Theory and Techniques Transactions, Vol. 53,
No. 4, April 2005, pp. 1331-1337.
doi:10.1109/TMTT.2005.845765
[6] X. Y. Zhang, J. X. Chen and Q. Xue, “Dual-band Band
Pass Filters Using Stub-Loaded Resonators,” IEEE Micro-
wave and Wireless Components Letters, Vol. 17, No. 8,
2007, pp. 583-585. doi:10.1109/LMWC.2007.901768
[7] A. Balalem, W. Menzel and A. Omar, “Quasi Lumped
Open Loop Suspended Stripline Bandpass Filters,” 36th
European Microwave Conference, Manchester, 10-15 Sep-
tember 2006, pp. 568-571.
[8] A. Balalem, J. Machac and A. Omar, “Dual-band Bandpass
Filter by Using Square Loop Dual Mode Resonator,”
International Journal of Microwave and Optical Tech-
nology, Vol. 50, No. 6, June 2008, pp. 1567-1570.
doi:10.1002/mop.23427
[9] “RT/duroid® 6006/6010LM High Frequency Laminates,”
1998.
http://www.rogerscorp.com/documents/612/acm/RT-duro
id-6006-6010-laminate-data-sheet
[10] M. V. Schneider, “Microstrip Lines for Microwave Inte-
grated Circuits,” The Bell System Technical Journal, Vol.
48, No. 5, 1969, pp. 1421-1444.
[11] H. A. Wheeler, “Transmission Line Properties of a Strip on
a Dielectric Sheet on a Plane,” IEEE Microwave Theory
and Techniques Transactions, Vol. 25, No. 8, 1977, pp.
631-647. doi:10.1109/TMTT.1977.1129179
[12] K. C. Gupta, et al., “Computer-Aided Design of Microwave
Circuits ,” A rtech H ouse, D edh am , 1 981 , pp. 13 1-134 .