Design of Parallel Coupled Microstrip Bandpass

Filter for FM Wireless Applications

Salima Seghier,Nasreddine Benahmed Nadia Benabdallah
and Fethi Tarik Bendimerad
Department of telecommunications Department of Physics
University Abou Bekr Belkaid-Tlemcen Preparatory School of Sciences and Technology
P.O.Box 119, (13000) Tlemcen, Algeria (EPST-Tlemcen)
Seghier7@yahoo.fr

Abstract— Design, analysis and optimization of a parallel-
coupled microstrip bandpass filter for FM Wireless applications
is presented in this paper. The filter is designed and optimized at
a center frequency of 6 GHz. Half wavelength long resonators
and admittance inverters are used to design the filter. A brief
description of coupled microstrip lines and immittance inverters
is also included. Design equations to compute physical
dimensions of the filter are given in the paper. The filter is
simulated using ADS (Advanced Design System) design software
and implemented on Roger 4003C substrate.
and odd- mode characteristic impedances of parallel-coupled
half-wave resonators are computed using admittance inverters.
These even and odd mode impedances are then used to
compute physical dimensions of the filter [3].
The filter is implemented on Roger 4003C substrate with
dielectric constant of 3.38, loss tangent of 0.0021 and
substrate height of 0.508 mm.

Keywords- Bandpass filter; coupled microstip lines; immittance

inverter; Method of moments.

I. INTRODUCTION
Parallel coupled transmission-line filter in microstrip and
stripline technology are very common for implementation of
bandpass and band-stop filters with required bandwidth up to a
20% of central frequency. Due to their relatively weak
coupling, this type of filter has narrow fractional bandwidth Y0  1 / Z 0
but instead has desired advantages such as low-cost
fabrication, easy integration and simple designing procedure Figure 1. General layout of parallel coupled microstrip bandpass filter.
(designing equations for the coupled line parameters such as
space-gap between lines and line widths and lengths, can be II. IMMITTANCE INVERTER
found in classical microwave books [1]. This way, following a Immittance inverters play a very important role in filter
well-defined systematic procedure, the required microstrip- design. They are used to transform a filter circuit into an
filter parameters can be easily derived for Butterworth and equivalent form that can be easily implemented using various
Chebyshev prototypes [2]. microwave structures. Immittance inverters are either
This paper presents the design of a parallel-coupled impedance or admittance inverters. Because of the inverting
microstrip bandpass filter centered at 6 GHz, bandwidth BW action, a series inductance with an inverter on each side
of 200 MHz with minimum attenuation of -15 dB at 6.2 GHz, looks like a shunt capacitance and a shunt capacitance with an
pass-band ripple of 0.5 dB. This frequency band is used by inverter on each side looks like a series inductance.
FM Wireless communication. The filter is designed using half Immittance inverters are shown in Fig.2.
wave long resonators and admittance inverters. Theory of
general immittance inverters and coupled lines is briefly
described.
General layout of a parallel coupled microstrip bandpass
filter is shown in Fig.1. The filter structure consists of open
circuited coupled microstrip lines. These coupled lines are
quarter wavelength, (  / 4 long) and are equivalent to shunt Figure 2. Immittance inverter.
resonant circuits. The coupling gaps correspond to the
admittance inverters in the low-pass prototype circuit. Even
 Making use of the properties of immittance inverters,  (2i  1)   (2i  3) 
bandpass filters may be realized by series (L-C) resonant 4 sin   sin  
1  2n   2n  (5)
circuits separated by impedance inverters (K) or shunt (L-C) gi 
g i 1  (i  1) 
parallel resonant circuits separated by admittance inverters (J).  2  sin 2  
To design a bandpass filter, first of all a low-pass prototype  n 
circuit is modified to include immittance inverters. These For i = 2, 3… n
low pass structures are then converted to bandpass circuits 1.0 for n odd
 (6)
by applying conventional low-pass to bandpass g n 1   2 
transformation [3]. coth 4 for n even

III. PARALLEL- COUPLED FILTER DESIGN Where:

The first step in designing a filter is to determine the order   Ripple 
  ln coth  
of the filter, n required. The order of the filter can be   17.37 
determined from (1), (2) and (3) [4] and [5]. From the result of   
the (3), the order of the filter can be determined from Fig.3.   sinh  
2n
Fractional bandwidth,
 2  1 The element values obtained are g0 = g4 = 1, g1 = g3=
FBW  (1)
0 1.5963, g2 = 1.0967.The low- pass prototype elements values
obtained can be represented as shown in Fig.4.
Bandpass transformation,
 2    0  (2)
  
1  0
'
FBW 

Normalized frequency,
 (3) (8)
1
1'
The low-pass prototype elements values obtained can be

Figure 4. Low-pass filter prototype.

The low-pass filter consists of series and parallel branch, J-

inverter is used to convert low-pass filter to bandpass filter
with only shunt branch as shown in Fig.5. (7), (8) and (9) are
used to obtain the J-inverters [7]-[11].

Figure 3. Attenuation versus normalized frequency for 0.5 dB

ripple low-pass filter prototype.

Based on the design specification, a 3rd order filter is Figure 5. Bandpass filter prototype
required for FM Wireless application. The element values for
3rd order using 0.5 dB equal ripple low-pass prototype are J 01  FBW (7)
 
determined from (4), (5) and (6) where g0 = 1.0 [6]. Y0 2 g 0 g1
J j , j 1  FBW 1 (8)
2      j  1 to n  1
g i  sin   (4) Y0 2 g j g j , j 1
 2n
J n,n1  FBW (9)

Y0 2 g n g n,n1
Where FBW is the fractional bandwidth of the bandpass
filter, Jj,j+1 are the characteristic admittances of the J
– inverters and Y0 is the characteristic admittance of the
terminating lines.
To realise the J-inverters, even- and odd-mode
characteristic impedances of coupled lines, are determined by
using relations (10) and (11). The calculated results are listed
in Table I.
This type of filter was studied by [8] at a center frequency
of 2.491 GHz for Satellite Receiver.
The result of the microstrip filter simulation is shown in
Fig.7. In the plot, we can find out that the center frequency of
the filter has deviated from the specified frequency 6 GHz. So
the design can’t be used in practice.
0

-10

1  J j , j 1  J j , j 1  
2

Z  (10)

dB(S(1,1))
dB(S(2,1))
0e j , j 1
 1      -20
Y0  Y0  Y0  
 -30

1   
2 -40
J J
Z  0o j , j 1
 1  j , j 1   j , j 1  
Y0  Y0
(11)
 Y0  
-50
 5.4 5.6 5.8 6.0 6.2 6.4

freq, GHz
Where j = 0 to n
Figure 7. Scattering parameters of the designed filter using
conventional method
TABLE I. PARAMETERS OF PARALLEL- COUPLED FILTER

Order of admittance Even-mode Odd-mode

the filter inverters impedance Z0e impedance Z0o Fig.8 shows the effect of changing coupled line length on
n Jj,j+1/Y0 (Ohms) (Ohms) the filter response. In this experiment, the width and space are
(Ohms) kept constant while the length is simulated for both 10%
1 0.1805 60.655 42.605
increment and 10% reduction. As the length reduces, the
2 0.0394 52.045 48.105 frequency band shifted to the right. Based on (12), as guided
3 0.0394 52.045 48.105 wave length,  g gets shorter, the center frequency, f 0 will
4 0.1805 60.655 42.605 increase and shifted towards higher frequency band.
Attenuation and insertion losses are not affected by the change
of length.
Based on the -even and odd characteristic impedance
shown in Table I, the dimensions of the coupled lines can be
obtained using ADS Line Calculator, LineCalc. The length of
0
coupled line is quarter wavelength,  / 4 long. The physical
dB(Reduced_Len..S(2,1))
dB(Increase_Len..S(2,1))

dimension of the coupled lines, which is computed by -20

LineCalc is shown in Table II. +10% -10%

dB(S(2,1))

-40

TABLE II. PHYSICAL DIMENSSION OF COUPLED LINE -60

Line description Width Length Gap -80

(mm) (mm) (mm)
-100
50 ohm-line 1.1763 7.604 - 4.5 5.0 5.5 6.0 6.5 7.0
Coupled lines 1 and 4 1.0858 7.7020 0.3083 freq, GHz
1 andlines
Coupled 4 2 and 3 1.1758 7.6372 1.4395
Figure 8. Effect of changing coupled line length

A 3rd order parallel-coupled bandpass filter as shown in 300 (12)

Fig.6 is setup using ADS to study the effect of optimizing the  g (mm) 
f 0 (GHz )  eff
different parameters of coupled line. The circuit dimensions are
based on Table II.
Fig.9 shows the effect of changing coupled line width on
the filter response. In this experiment, the length and space are
kept constant while the width is simulated for both 10%
increment and 10% reduction. As the width increases,
insertion loss increases.

Figure 6. Circuit of investigating coupled line parameters.

 We use VAR components to set the tunable parameters
such as microstrip width, gap and length. In this design, the
0 filter is symmetrical geometrically, indeed only two group
dB(Reduced_Width..S(2,1))
dB(Increase_Width..S(2,1))

different values are needed to determined during the

-20
optimization which are set to be (w1,S1, l1) and (w1,S2, l2).
dB(S(2,1))

-10% The parameters in the VAR component should set around

-40
the values which we have calculated with conventional
+10% method. As the coupled microstrip filter is sensitive to the
-60
increment and decrease of its dimension, the value should not
-80
deflect the centre too much to lessen the time exhausted in the
4.5 5.0 5.5 6.0 6.5 7.0 optimization.
freq, GHz
The plot of the amplitude versus frequency after
optimization is shown in Fig.12. In the plot, we can see clearly
Figure 9. Effect of changing coupled line width
that the centre frequency of the filter has been adjusted to 6
Fig.10 shows the effect of changing coupled lines pace on GHz and the corresponding insert loss is less than 1dB, the
the filter response. In this experiment, the width and length are reflect ratio in pass band is -19.316dB with a -34.468dB
kept constant while the space is simulated for both 10% attenuation in the alias frequency, indicating that the request
increment and 10% reduction. As the space increases, performance is well satisfied.
insertion loss increases as the coupling between the lines
become weaker. m2 m1
freq=6.034GHz freq=6.001GHz
dB(S(1,1))=-19.316 dB(S(2,1))=-0.914
0 m1
dB(Increase_Space..S(2,1))
dB(Reduce_Space..S(2,1))

-20 -10
-10% m2
dB(S(2,1))

dB(S(1,1))
dB(S(2,1))

-40 -20

-30 m3
-60
m3
+10% -40
freq=5.604GHz
-80 dB(S(2,1))=-34.468
-50
5.4 5.6 5.8 6.0 6.2 6.4
-100
freq, GHz
4.5 5.0 5.5 6.0 6.5 7.0 Figure 12. Scattering parameters of the designed and
Figure 10. Effect of changing optimized bandpass filter.
freq, GHz coupled line space
After the optimization, we update the dimentions of the
microstrip with the new parameters which are derived from
IV. OPTIMAL DESIGN OF THE COUPLED the optimization. The final refined parameters of the
MICROSTRIP FILTER dimension are listed in Table III.
To reach0 the specified request, we adopt the optimal
design ability owning by ADS, numerical software based on TABLE III. PHYSICAL DIMENSSION OF COUPLED LINE
-10
the use of method of moments (MoM) [8] and [12] (Fig.11). Line description Width Length Gap
dB(S(1,1))
dB(S(2,1))

We put the-20OPTIM controller in the schematic as well as three (mm) (mm) (mm)
GOAL controllers to specify the optimal goals which we want 50 ohm-line 1.1763 7.604 -
-30
to get after optimization. The paper sets three GOAL
controllers -40which define the band-pass insert loss, alias Coupled lines 1 and 4 1.0858 7.4475 0.2072
attenuation and reflect ratio respectively. To make the insert Coupled lines 2 and 3 1.1758 7.4181 1.3456
loss in band-50as least as possible, a 3 dB loss is set when the
5.4 5.6 5.8 6.0 6.2 6.4
frequency is between 5.9 and 6.1 GHz.
freq, GHz V. CONCLUSION
In this paper the detail procedures for designing a parallel-
coupled bandpass filter for FM Wireless applications, has been
presented. For the selected center frequency of 6 GHz and on
a substrate with a dielectric constant of 3.38, our filter is
designed and optimized. Half wavelength long resonators and
admittance inverters are used to design the bandpass filter.
The optimization function owning by ADS software is an

Figure 11. Layout of parallel-coupled microstrip bandpass filter

efficient tool to amend the drawback of conventional method
with theoretic formulas.

REFERENCES
[1] H. Karimi zarajabad and S. Nikmehr, “A Novel Fractal Geometry for
Harmonic Suppression in Parallel Coupled- Line Microstrip BandPass
Filter”, IEEE 2008.
[2] Miguel Bacaicoa, David Benito, Maria J. Garde, Mario Sorolla and
Marco Guglielmi, “New Microstrip Wiggly-Line Filters with Spurious
Pass-band Suppression”, IEEE Transactions on microwave theory and
techniques, vol. 49, NO. 9, September 2001.
[3] Rashid Ahmad Bhatti, Jahangir Khan Kayani, “Design and analysis of
parallel coupled microstrip band pass filter”, 2nd International Bhurban
Conference on Applied Sciences and Technology,Bhurban,Pakisten.
June 16-21, 2003.
[4] G. Mattaei, L. Young, and E.M.T. Jones, “Microwave Filters,
Impédance Matching Networks, and Coupling Structures”, Artech
House, Norwood, MA, 1980.
[5] Selinda Lim, Huey Pin, “Design and Modification of Parallel-coupled
Bandpass filter”, School of Electrical Engineering and Computer
Science, University of Newcastle, Callaghan, NSW 23 08, Australia.
[6] Hong, J.S., M.J, “Microstrip Filter for RF/Microwave Applications”, A
Wiley- Interscience Publication, Canada, 2001.
[7] D. M. Pozar, “Microwave Engineering”, John Wiley & Sons Inc., 1998.
[8] Yanling Hao, Bingfa Zu, and Ping Huang, “An optimal microstrip filter
design method based on Advanced Design system for satellite receiver”,
Proceedings of IEEE International Conference on Mechatronics and
Automation, 2008.
[9] Bey-Ling Su, Ray Yueh-Ming Huang, “5.8GHz Bandpass Filter Design
Using Planar Couple Microstrip Lines,” IEEE, 2004.
[10] Annapurna Das and Sisir K Das, “Microwave Engineering”, MacGraw
Hill, p305, 2001
[11] Jen-Tsai Kuo, Senior, Meshon Jiang, and Hsien-Jen Chan, “Design of
parallel- Coupled Microstrip Filters With Sppression of Spurious
Resonances Using Substrate Suspension”, IEEE Transactions on
microwave theory and techniques, vol. 52, NO.1,January 2004
[12] “Tuning, Optimization and Statistical Design”, Agilent Technologies,
May 2003.