Design consideration for integrated mobile phone antennas

Abstract

Based on the investigation of the board effect on the bandwidth of
integrated antennas for mobile phones a concept for a tripleband antenna
suitable for GSM/DCS/PCS is developed. The antenna concept comprises a
parasitic element to enlarge the bandwidth. The performance of the
antenna is analyzed by means of numerical simulations based on the FDTD
method

Full text

DESIGN CONSIDERATIONS FOR INTEGRATED MOBILE PHONE ANTENNAS
D. Manteuffel, A. Bahr, D. Heberling, I. Wolff
IMST GmbH, Germany, e-mail: manteuffel@imst.de
Abstract – Based on the investigation of the board effect
on the bandwidth of integrated antennas for mobile phones
a concept for a tripleband antenna suitable for
GSM/DCS/PCS is developed. The antenna concept
comprises a parasitic element to enlarge the bandwidth.
The performance of the antenna is analyzed by means of
numerical simulations based on the FDTD method.
Introduction
Aesthetical design and marketing aspects have a large
impact on the development of mobile phones. Therefore
there is a remarkable trend to fully integrate the antenna in
the mobile phone. Compared to traditional antennas it is
more complicated to master the typical parameters like
bandwidth, efficiency and influence of the user within the
limited antenna volume. This becomes even more critical
with respect to multiband functionality which is an
essential feature of modern mobile phones. To enable a
rapid and efficient design some general relations
concerning the physical behavior of an antenna in a small
mobile have to be considered.
In this paper some of these relations will be discussed.
Based on these design guidelines a concept for a
GSM/DCS/PCS antenna suitable for the integration into a
mobile will be presented. The investigation is carried out
using EMPIRE
TM
which is a commercial FDTD software
from IMST GmbH.
Interaction of the antenna with the board
Due to the ongoing miniaturization of components and for
marketing requirements the size of modern mobile phones
decreases. Typical lengths vary between 80 mm and
130 mm. Assuming a quarter-wave antenna on a small
PCB, it is obvious that there is a strong interaction
between the antenna-module and the board [1]. In [2] the
effect of varying PCB lengths on the impedance
bandwidth at 900 MHz is investigated for a c-patch
antenna, a capacitively loaded patch antenna and a
dielectrically loaded patch antenna. An equivalent circuit
model of the interactions of the antenna module with the
board is derived in [3]. It describes the antenna module as
a high-Q resonator and the board as a low-Q resonator.
Both are coupled to each other by a certain coupling
coefficient.
The board effect for 900 MHz and 1800 MHz
With respect to multiband antennas the same investigation
has to be carried out for the 1800 MHz frequency band
additionally. For multiband functionality folded patch
antennas derived from the c-patch concept with one or
more resonators are commonly used. By folding the
antenna path it is possible to place two or more resonant
radiators on a small antenna-module. Coupling between
the radiators provides an additional parameter to tune the
resonances to the right frequency and adjust the matching
in all frequency bands with one feeding point.
In order to get a fair comparison between 900 MHz and
1800 MHz antennas the investigation presented in the
following is restricted to single resonant patches. The
investigation for 900 MHz is performed with a c-patch
antenna module of 4.3 cm³ on a realistic model of a
mobile phone. For the 1800 MHz frequency band some
parts of this antenna are removed in order to get the right
frequency. This results in a L-shaped antenna. Please note
that in this case not all the available surface on the module
is used for the antenna.
Figure 1: Folded patch antennas on a realistic model of a
mobile phone for 900 MHz and 1800 MHz.
The upper part of the mobile and the antennas are given in
Figure 1. The mobile consists of a PCB, a battery and a
RF-shielding where the antenna is mounted on. The
mobile is covered by a plastic casing of 1 mm thickness
with a relative permittivity of
3
=
r
ε
. In the simulation all
remaining parts of the mobile are treated as perfect
conductors. In the following investigation the length of the
mobile is enlarged from 80 mm to 150 mm without
changing any dimension in the upper part of the phone
near the antenna. The antenna itself is retuned to the right
frequency for every board.
80 90 100 110 120 130 140 150
length of PCB [mm]
0
2
4
6
8
10
12
14
r
e
l
.
b
a
n
d
w
i
d
t
h
-
6
d
B
[
%
]
Frequency
900 MHz
1800 MHz
Figure 2: Interaction of antenna-module and PCB: Influ-
ence on the bandwidth.

It can be observed from Figure 2 that the resonance effect
of the board can have a large impact on the impedance
bandwidth of the resulting antenna system [2], [3].
Especially for 900 MHz the bandwidth becomes very low
for a small mobile.
Consequence for a dualband application
Based on the analysis of the single resonant antennas in
the prior investigation a dualband antenna is developed by
using both paths in one module.
Figure 3: Dualband antenna composed out of two folded
resonant paths on a mobile of 110 mm length.
The coupling of the combined elements has an additional
effect on the resonances especially in the upper frequency
range. With some modifications it is possible to succeed in
a proper matching for a single feed solution in both
frequency bands.
In order to investigate the worst case scenario for the
upper frequency band this antenna concept is investigated
with a board length of 110 mm. At this length the coupling
effect with the board is only small for the DCS resonance
as shown in Figure 2. This results in a narrow bandwidth
at this frequency range. The size of the antenna-module is
5 cm³ and it is situated with some distance to the edges of
the PCB in the upper part of the mobile. The distance to
the battery, which is simulated as a metal block is 2 mm.
0,7 0,9 1,1 1,3 1,5 1,7 1,9 2,1
f [GHz]
-20
-15
-10
-5
0
|
s
1
1
|
[
d
B
]
Mobile
with casing
without casing
Figure 4: Matching of the dualband antenna with/without
plastic casing.
Figure 4 shows the matching in both frequency bands and
the influence of the plastic casing on the tuning of the
resonance frequencies. Due to the dielectric properties of
the plastic casing the GSM resonance is reduced by 4.5 %
and the DCS resonance is reduced by 5.6 %. This shift has
to be taken into account in the design of the antenna.
Figure 4 demonstrates that good matching is provided at
both center frequencies. In accordance to the above
investigation the bandwidth of the DCS mode (110 MHz
at a matching of –6 db) is quite narrow. To get a wider
matching in DCS the antenna module could be enlarged,
but in most cases the size is strongly limited.
A dualband solution with parasitic element
Another possibility to enlarge the bandwidth of an antenna
is the use of parasitic elements. This principle is widely
reported in the literature [4] for microstrip patch antennas.
An application for mobile phone antennas has recently
been published in [5].
Figure 5: Integrated antenna with parasitic element
situated on a simplified mobile.
Figure 5 shows the two radiator dualband antenna in
combination with a parasitic element. The parasitic
element is situated above the original antenna at the top of
the phone. As the height of the parasitic element is lower
than the original antenna-module this additional space
(0.4 cm³) is available at this position taking into account a
rounded casing at the top of the phone.
In Figure 5 the resonators for the upper frequencies are
placed orthogonal to each other. This provides less
coupling from the electric nearfield of the resonators
compared to having the parasitic resonator in parallel to
the DCS resonator. Nevertheless the other configuration
works also, as it is reported in [5].
Using a parasitic element the bandwidth can be increased
by a large amount. Therefore it is interesting especially for
future applications where more frequency bands have to
be covered by only one antenna. On the other hand with a
parasitic element matching becomes more complicated
because more parameters compared to the original antenna
have to be considered.

0,6 0,8 1 1,2 1,4 1,6 1,8 2 2,2
f [GHz]
-15
-10
-5
0
|
s
1
1
|
[
d
B
]
0
50
100
150
200
250
300
Z
[
Ω
]
s11
Re(Z)
Im(Z)
a) Matching and impedance behavior.
0
0,2
0,5
1
2
5
10
180
1
7
0
1
6
0
1
5
0
1
4
0
1
3
0
1
2
0
1
1
0
1
0
0
90
8
0
7
0
6
0
5
0
4
0
3
0
2
0
1
0
0
-
1
0
-
2
0
-
3
0
-
4
0
-
5
0
-
6
0
-
7
0
-
8
0
-90
-
1
0
0
-
1
1
0
-
1
2
0
-
1
3
0
-
1
4
0
-
1
5
0
-
1
6
0
-
1
7
0
-6dB
b) Smith Chart.
Figure 6: Matching and impedance behavior using a
parasitic element.
The point of most interest is the efficient coupling of the
upper modes. As it can be observed from Figure 6 the
impedance behavior of the upper resonances is different
compared to an ordinary first order resonance matching.
The lower resonance (at 1.6 GHz) is tuned to have a quite
high impedance maximum. This generates a flat curve of
the radiation resistance until the third resonance appears.
Sufficient matching is provided in-between. With respect
to the Smith Chart in Figure 6b it may still be possible to
optimize the antenna, but even with the current solution a
bandwidth of 225 MHz according to a matching of –6 dB
is obtained. For the GSM resonance the bandwidth is
80 MHz. This is sufficient to cover the whole frequency
band in this standard. With regard to the above
investigation on the board effect the bandwidth at GSM is
dependent of the dimensions of the board, too.
Parameters for tuning the coupling
In order to optimize the coupling a parameter variation is
performed.
Figure 7: Parameters for tuning the coupling.
For this purpose at least three parameters can be found that
influence the coupled resonances.
0,7 0,9 1,1 1,3 1,5 1,7 1,9 2,1
f [GHz]
-20
-15
-10
-5
0
|
s
1
1
|
[
d
B
]
DCS Resonator
- 1 mm
0 mm
+1 mm
a) Variation of the DCS resonator.
0,7 0,9 1,1 1,3 1,5 1,7 1,9 2,1
f [GHz]
-20
-15
-10
-5
0
|
s
1
1
|
[
d
B
]
Parasitic Element
- 1 mm
0 mm
+1 mm
b) Variation of the parasitic element.
0,7 0,9 1,1 1,3 1,5 1,7 1,9 2,1
f [GHz]
-20
-15
-10
-5
0
|
s
1
1
|
[
d
B
]
Feeding-Position
- 2 mm
-1 mm
0 mm
+1 mm
+2 mm
c) Variation of feeding point.
Figure 8: Parameter variation for tuning the coupling.

The first variation is performed by varying the length of
the DCS resonator. It can be seen from Figure 8a that this
variation does not affect the resonance of the parasitic
element very much. On the other hand it can be observed
from Figure 8b that a change in the capacitive load of the
parasitic element results mainly in a change of the
resonance according to this resonator. Taking the results
from both analysis it can be stated that the position of both
resonances can be tuned nearly independently from each
other by the lengths of the related elements. The reason for
this independent behavior has its origin in the orientation
of both radiators relative to each other. The distant
orthogonal placement provides decoupling by means of
the electric nearfield generated at the end of both
resonators. A galvanic coupling is only present at the
common shorting pin.
Figure 8c shows that matching can be adjusted by varying
the feeding point. Please note that the DCS resonance is
not independent from this parameter. Finally it can be
observed that there is only a minor effect on the GSM
resonance in all variations.
A GSM/DCS/PCS antenna on a realistic mobile
Based on the concept derived above an integrated
tripleband antenna for the GSM, DCS and PCS is
developed.
Figure 9: Tripleband antenna in a realistic configuration.
The size of the antenna module in Figure 9 is 6.3 cm³. The
height above the PCB is 7 mm. The parasitic element
needs additional space of 0.66 cm³ with a reduced height
of 4 mm. This results in a total volume of less than 7 cm³.
The shape of the antenna has been modified in order to fit
closer to the shape of a realistic mobile. In accordance to
the prior investigation the length of the PCB is 110 mm.
The relative permittivity of the plastic casing is
3
=
r
ε
and its thickness is 1 mm.
FDTD simulations are performed for the model of the
mobile in free-space and in the vicinity of a model of the
human head. The model of the human head is presented in
Figure 10. It consists of 16 different tissues. The spatial
resolution of the head phantom is 1 mm.
Figure 10: Realistic model of the human head.
The material parameters of theses tissues are taken from
Gabriel [6] for two different frequencies under
investigation (900 MHz, 1900 MHz). The mobile is
attached to the human head in the “intended use” position
which is defined in [7].
0,7 0,9 1,1 1,3 1,5 1,7 1,9 2,1
f [GHz]
-20
-15
-10
-5
0
|
s
1
1
|
[
d
B
]
Simulation Model
without head
with head
Figure 11: Matching of the antenna with and without the
user influence.
Figure 11 shows the matching of the antenna in free-space
and attached to the human head in “intended use” position.
The bandwidth according to a matching of –6 dB is
85 MHz in GSM and 280 MHz in DCS/PCS. Therefore
the antenna is well suited for this tripleband application.
Comparing the result with/without the influence of the
human head it can be observed that only a minor shift in
the resonance frequency occurs due to the dielectric
properties of the human head. The bandwidth is increased
due to losses in the tissue.

To complete the analysis the radiation efficiency is
calculated. As the mobile itself is simulated without any
losses the total radiation efficiency is affected by the
mismatch of the antenna and the power-loss in the tissue.
Therefore the total radiation efficiency can be defined as:
in
rad
total
P
P
=
η
(1)
According to the different standards GSM, DCS, and PCS
the mean input power is 250 mW, 125 mW and 125 mW
respectively.
Neglecting the influence of the mismatch of the antenna a
different definition of an radiation efficiency can be found
that only accounts for the losses in the tissue. It can be
defined as:
antenna
rad
absorption
P
P
=
η
(2)
η
total
[%] η
absorption
[%]
900 MHz 30 31.2
1800 MHz 48.6 57.4
1900 MHz 46.4 58.2
2000 MHz 55.8 59.4
Table 1: Radiation efficiency of the mobile in
“intended use” position.
Table 1 shows the calculated values for the radiation
efficiency taking into account the mismatch of the antenna
or neglecting it. It can be observed that there is a
difference of the amount of power absorbed by the tissue
comparing the lower and the upper frequency range. This
may also be due to the more omni directional
characteristic of the antenna in GSM where the interaction
with the PCB is larger. For this explanation is only a first
assumption this has to be investigated in more detail in the
future work.
Conclusion
The interaction of the antenna with the PCB of a mobile
has been investigated in terms of bandwidth for 900 MHz
and 1800 MHz using folded patch antenna. Based on this
study a dualband antenna has been developed. Within the
given constraints regarding the size of the antenna-module
and the dimension of the board the bandwidth for DCS
was too small. To improve the bandwidth in DCS a
parasitic element was attached to the antenna module
which needed only minor additional space. To adjust the
coupling a parametric investigation was performed that
provides guidelines for matching the antenna. Based on
the developed concept a tripleband antenna for the
standards GSM/DCS and PCS has been designed on a
realistic platform. The performance of the antenna has
been investigated taking into account the influence of the
users head.
References
[1] S. Gonzalez Garcia, L. Baggen, D. Manteuffel, D.
Heberling: Study of Coplanar Waveguide-Fed
Antennas Using the FDTD Method. Microwave and
Optical Technology Letters, vol.19, no. 3, S. 173-
176, 1998.
[2] D. Manteuffel, A. Bahr, I. Wolff: Investigation on
Integrated Antennas for GSM Mobile Phones.
AP2000 – Conference on Antennas an Propagation,
April 2000, Davos, Switzerland.
[3] P. Vainikainen, J. Ollikainen, O. Kivekäs, I.
Kelander: Performance Analysis of Small Antennas
Mounted on Mobile Handsets. 11
th
COST 259
Management Committee Meeting, April 2000,
Bergen, Norway.
[4] K. F. Lee, W. Chen (Eds.): Advances in Microstrip
and Printed Antennas. John Wiley & Sons, New
York, 1997.
[5] J. Ollikainen, O. Kivekäs, A. Toropainen, P.
Vainikainen: Internal Dual-Band Patch Antenna for
Mobile Phones. AP2000 – Conference on Antennas
an Propagation, April 2000, Davos, Switzerland.
[6] S. Gabriel, R. W. Lau, C. Gabriel: The Dielectric
Properties of Biological Tissue: III. Parametric
models for the dielectric spectrum of tissues. Phys.
Med. Biol., Vol. 41, 2271-2293, 1996.
[7] European Specification (ES 59005): Considerations
for the Evaluation of Human Exposure to
Electromagnetic Fields (EMFs) from Mobile
Telecommunication Equipment (MTE) in the
Frequency Range 30 MHz – 6 GHz. CENELEC,
1998.