Content uploaded by Ingo Wolff
Author content
All content in this area was uploaded by Ingo Wolff
Content may be subject to copyright.
INVESTIGATION ON INTEGRATED ANTENNAS FOR GSM MOBILE PHONES
DIRK MANTEUFFEL(1), ACHIM BAHR(1),
INGO WOLFF (1)
(1)IMST GmbH
Carl-Friedrich-Gauss-Str. 2
47475 Kamp-Lintfort
Germany
Email: manteuffel@imst.de
ABSTRACT
The objective of the study was to compare different types of integrated antennas for mobile phones working at the
GSM 900 standard on the same realistic platform. To realise a small integrated antenna resonant at 925 MHz two
different concepts to handle the lack in length were investigated: a rectangular patch antenna which was capacitively
loaded at the end of the radiator, and a folded antenna in the shape of a c-patch. Achievable impedance bandwidth and
worst case SAR (Specific Absorption Rate) of both types are compared. All antennas were situated in the same position
on the same model of a modern mobile phone. Two variations of the antenna volume were investigated in all cases. As
a results it can be stated that for both configurations the impedance bandwidth of the c-patch antenna is larger than the
one of the capacitively loaded patch antenna. Contrary to this the worst case SAR, assessed by a Dosimetric Assessment
System (DASY), of the capacitively loaded patch is lower compared to the c-patch. In a further step the impedance
bandwidth of the antenna was investigated as a function of the physical length of the mobile by means of numerical
simulations based on the Method of Moments (MoM). The study showed that the impedance bandwidth of the antenna
system is strongly influenced by this length as the metallic part of the mobile acts as the counterpart for the quarter
wave antenna module.
INTRODUCTION
There is a remarkable trend towards the development of integrated antennas for mobile phones. Although this trend is
mainly forced by marketing and design aspects there are also some physical advantages of integrated antennas
compared to externally mounted antennas. Using the principle of patch antennas the interaction with the user’s head in
talk position is less than using e.g. short helical antennas and thus results in lower SAR [3]. On the other hand there are
many problems to cope with. One main problem is that the bandwidth of an integrated patch antenna is usually to small
to cover the whole GSM 900 frequency band. Therefore, additional switching circuits are often used that provide
matching in rx- or tx-mode but introduce additional losses in the switching components. The bandwidth problem
becomes more critical because the allowed antenna volume is strongly limited as future mobiles become smaller and
smaller.
As most concepts for integrated antennas are based on resonant quarter wave structures the physical antenna length for
GSM 900 is too large (~8 cm) for the implementation in a small mobile. Because of this it is necessary to reduce the
size of the antenna module. From basic antenna theory [1][2] there are at least three main principles to reduce the
overall antenna size: 1) the use of a dielectric substrate, 2) capacitive loading at the end of the radiator or 3) folding of
the antenna structure to achieve a resonant path on a smaller area. Today’s mobile phones with integrated antennas
make use of these principles. For example, the Hagenuk Global Handy uses a patch antenna on a dielectric material, the
Sony CMD-C1 uses a capacitively loaded patch and the Nokia 8810 provides a folded C-patch antenna structure. Fig. 1
shows the implementation of the integrated antennas in the mobile phone. It can be noted that there are more differences
than the antenna type only. The different implementations vary also in size, position and orientation of the
antenna-module in the mobile. Furthermore, the size of the mobile itself varies. These different implementations make it
hard to compare the antenna concepts itself. In this study we investigated integrated antennas for GSM mobile phones
based on the concepts of a capacitively loaded quarter wave patch antenna and a c-patch antenna structure on the same
realistic model of a modern mobile phone. This platform contains realistic restrictions to the antenna volume, the
positioning of the antenna module and interaction to other components of the mobile phone like the battery pack. This
procedure made it possible to compare the different concepts on this generalised platform and enabled the extraction of
some relevant parameters for future antenna designs with regard to small sized antenna modules. Although this study is
restricted to singleband antennas operating at the GSM 900 resonant frequency of 925 MHz, the results are useful for
future development of integrated dualband– and multiband-solutions.
a) Sony CMD-C1 b) Nokia 8810
Fig. 1: Mobile phones with integrated antennas.
FOLDED AND UNFOLDED ANTENNAS
The volume that can be allocated by the antenna in a modern mobile phone is strongly limited. The best suited location
in many phones is a position above the rf-shielding in the upper part of the phone. But even at this position especially
the height is limited. Furthermore, the antenna module is located in close vicinity to the battery pack. In many cases the
minimum distance is less than 3 mm.
In this study we deal with two different configurations for each antenna type. In both situations the antenna module is
located above the rf-shielding. The first configuration uses an antenna module with a volume of 36 x 30 x 6 mm3 which
will be called “large module” in the following. In the second configuration the volume is restricted to 36 x 20 x 6 mm3
which we call “small module” in the following. The practical constraint for this smaller volume could be e. g. the need
to implement a vibrator or a blue tooth module in addition without enlarging the overall size of the mobile. For this first
investigation the total length of the metallic parts of the mobile is 100 mm.
a) Mobile with large antenna-module (36x30x6 mm3). b) Mobile with small antenna-module (36x20x6 mm3).
Fig. 2: Mobile with two different restrictions to the antenna volume.
Fig. 2 shows both models for the numerical simulations. For simplicity the specific antenna shapes are replaced by a
small or a large antenna module in Fig. 2. Both models include a Printed Circuit Board (PCB) which defines the total
length of 100 mm, a rf-shielding that covers nearly the whole PCB, a battery pack and the antenna module. The battery
pack is modelled as a metallic block connected to the mobile at its bottom surface. The distance between the edge of the
antenna module
battery pack
PCB rf-casing
space for
accessory
antenna module and the top of the battery pack is 2 mm. For the small module there is also a certain volume allocated
by the above mentioned accessory at the bottom of the phone.
In a first investigation antennas, based on both concepts, were developed to operate at the GSM 900 center-frequency of
925 MHz, and matched to have a maximum of bandwidth in the specific configuration. The results for the impedance
bandwidth coming out of a loss-less simulation, based on the MoM, are shown in Fig. 3.
large module small module
0
1
2
3
4
5
6
rel. bandwidth
-6dB
[%]
3,14
1,95
5,19
4,24
antenna-type
cap. patch c-patch
large module small module
0
0,5
1
1,5
2
2,5
3
3,5
rel. bandwidth
-10dB
[%]
1,73
0,97
2,8
2,5
antenna-type
cap. patch c-patch
Fig. 3: Relative impedance bandwidths (925 MHz) of capacitively loaded patch and c-patch (simulation).
It can be stated that the bandwidth for the folded antenna is higher in both configurations. Furthermore, the relative
decrease in bandwidth from the large module to the small module is less strong for the c-patch.
The SAR was measured with a DASY system for the small model only. For feeding the antenna a cw-power of 0.25 W
was provided which corresponds to 2 W averaged over eight time slots according to the GSM 900 standard. Because of
the fact that no casing has been used for prototyping the assembly was located with 5 mm distance to the ear piece of
the phantom in the test positions defined by the CENELEC standard [4]. The worst case SAR was found in “touch”
position. Detailed values are given in Table 1.
Worst case SAR [W/kg]
1g 10g
C-patch 1.4 0.9
Cap. patch 10.7
Table 1: Measured worst case SAR for the small module.
In contrast to the bandwidth results the lowest SAR appears for the capacitively loaded patch. One explanation of this
result is that the shorting edge at the top of the phone is large (36 mm) for this antenna concept. This results in a lower
current density and therefore in a lower magnetic field strength that is related to the SAR coming from this portion.
RELATION BETWEEN THE LENGTH OF THE MOBILE PHONE AND THE IMPEDANCE BANDWIDTH
Due to the quarter wavelength principle of the above described concepts the antenna module that is located at the top
uses the whole metallic part of the mobile as the counterpart for radiation. In this way the whole mobile has to be
treated as the antenna system and should act more like an unsymmetrical dipole than a patch antenna above a large
ground plane. Because of this there must be a strong influence of the mobile size on the antenna parameters. The effect
of a small ground counterpart has also been studied in [5] in a different context. In this study the influence of the length
of the mobile on the impedance bandwidth of the antenna system is investigated using numerical simulations based on
the MoM. This was done by enlarging the battery pack in length to ensure that other parameters like the distance
between the antenna and the top of the battery pack are not altered. The investigation was carried out for both types of
antennas in both configurations according to Fig. 2. Fig. 4 shows the results for a length of the mobile ranging from
80 mm to 150 mm. For each configuration the antenna was tuned to 925 MHz. To achieve this it was necessary to
modify the end-capacitor of the patch and to change the shape of the c-patch slightly. The position of the feeding has to
be modified for each situation, too. According to Fig. 4 a strong effect can be observed. The impedance bandwidth of
the antennas rises until the mobile reaches a certain length that is related to the resonant frequency. Fig. 4 indicates that
this behaviour is similar for both antenna types.
80 90 100 110 120 130 140 150
length of mobile [mm]
0
1
2
3
4
5
6
7
8
rel. bandwidth
-10dB
[%]
antenna-module
large cap.patch
small cap. patch
large c-patch
small c-patch
Fig. 4: Relation between impedance bandwidth and length of the mobile (simulation).
Although the effect can be observed for both antenna types the quantitative results differ. In this respect the study shows
that the difference in bandwidth is only small for the different sized c-patch modules on a short mobile (up to a length
of 100 mm). As a general conclusion it must be stated that the problem of narrow bandwidth using integrated antennas
becomes even harder if the size of future mobiles is further reduced.
CONCLUSIONS
The study shows that a folded c-patch structure is more suited by means of impedance bandwidth than a capacitively
loaded patch integrated in the upper part of a mobile phone. In contrast to this the SAR is higher for the c-patch
configuration. A strong effect of the mobile length on the impedance bandwidth can be regarded that has to be
considered when implementing an antenna in a short mobile phone.
REFERENCES
[1] Fujimoto, K.; James J. R.: Mobile Antenna System Handbook, Artech House, Inc., 1994
[2] Hirasawa, K.; Haneishi, M.: Analysis, Design and Measurement of Small and Low-Profile Antennas, Artech
House, Inc., 1992
[3] Bahr, A.; Schneider, M.; Manteuffel, D.; Heberling, D.: Recent Trends for Mobile Phone Antennas with Special
Emphasis to the Interaction with the User; accepted paper to AP2000, Davos Switzerland, April 2000
[4] European Specification ES59005: Considerations for Evaluation of Human Exposure to Electromagnetic Fields
(EMFs) for Mobile Telecommunication Equipment (MTE) in the Frequency Range 30 MHz – 6 GHz; CENELEC,
1998
[5] Garcia, S. G.; Baggen, L.; Manteuffel, D.; Heberling, D.: Study of Coplanar Waveguide-Fed Antennas Using the
FDTD Method, Microwave and Optical Technology Letters, vol. 19, no. 3, 1998