A very-low-profile, six-antenna MIMO system aimed at operating in the concurrent 2.4 and 5 GHz bands for WLAN access-point applications is proposed. The MIMO system consists mainly of an antenna ground plane and six short-circuited monopole antennas, among which the three antennas are designated for 2.4 and 5 GHz operation respectively. The antennas are set in a sequential, rotating arrangement on the ground plane, and the 2.4 and 5 GHz antennas are facing each other one by one. The results show that well port isolation can be obtained together with good radiation characteristics. With a low profile of 6 mm in height, the proposed design can easily fit into wireless access points or routers and allow the 2.4- and 5-GHz band signals to be simultaneously received and transmitted with no need of external diplexer.
2. and at the same time, short-circuited to the ground for ease of
mounting [together with foam rubber as shown in Fig. 1(c)]. The
near optimal dimensions for the 2.4 and 5 GHz antennas are
detailed in Figure 2. The antenna height in this study is 6 and 5 mm
for the 2.4 and 5 GHz monopoles, respectively, making it possible
for the design to be integrated into a wireless AP or router as
internal MIMO antennas. Notice that the thickness of the proposed
design is merely about 4.8% free-space wavelength at 2442 MHz,
the center frequency of the 2.4-GHz WLAN band. Moreover, one
can easily alter the center operating frequency (fc) of the 2.4 GHz
antenna by fine-tuning parameter d. With an increase in d, the fc
increases too. As for the 5 GHz antennas, the center frequency fc
can be adjusted by fine-tuning parameter g with all the other
dimensions untouched and in general, goes up from lower to
higher frequencies as g decreases. These parameters are substan-
tially useful, especially when the antennas are installed inside the
housing of a wireless AP, because the operating frequencies of
both 2.4- and 5-GHz antennas are affected by dielectric loading
(device housing) [9] and usually shifted to lower frequency band.
3. RESULTS AND DISCUSSION
On the basis of design dimensions given in Figure 2, the proposed,
MIMO AP antennas was constructed, studied, and tested. Figures
3(a) and 3(b) show the measured reflection coefficients and isola-
tion between antennas. The reflection coefficients are plotted by
the curves of S11, S22, S33 for the 2.4 GHz antennas and of S44, S55,
S66 for the 5 GHz antennas. The isolation between any two of the
six antennas is only presented by the curves of S21, S31, S41, S51,
S61, S54, S64 due to symmetrical structure of the proposed antenna
system. It can be first seen that all measured impedance bandwidth
of the 2.4 and 5 GHz antennas satisfy the required bandwidth
Figure 1 (a) Configuration of the proposed, low-profile monopole an-
tennas for concurrent, dual-WLAN-band operation for MIMO access-point
applications. (b) Top view of the proposed six-antenna MIMO system. (c)
Photograph of a design prototype. [Color figure can be viewed in the online
issue, which is available at www.interscience.wiley.com]
(denoted as antenna 1, 2, 3), and three 5 GHz antennas (denoted as
antenna 4, 5, 6). Each antenna is situated next to another which
operates in different frequency band so as to mitigate mutual
coupling therein between. The six antennas are set 15 mm away
from the vertex and mounted on the ground with an equal incli-
nation angle (formed by two adjacent vertices and the center) of
60°. In this case, the proposed antennas are in a sequential, rotating
arrangement and of a symmetrical structure. Figure 1(b) gives a
comprehensible drawing of the design described earlier. A photo-
graph of the design prototype is shown in Figure 1(c), too, for
better understanding. To feed each antenna, six 50- mini-coaxial
cables with I-PEX connectors are utilized [see Fig. 3(c)]. The inner
conductors of the coaxial cables are connected to the feed points,
and the outer braided shielding are connected to the hexagonal
ground plane. Figure 2 Dimensions of the 2.4- and 5-GHz monopole antennas in
In order to realize the proposed AP antennas that have a very detail. [Color figure can be viewed in the online issue, which is available
low profile, the designed monopole antennas are bent two times at www.interscience.wiley.com]
DOI 10.1002/mop MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 11, November 2009 2615
3. (a)
Figure 4 Measured 2D radiation patterns at 2442 MHz for antenna 1
studied in Figure 3(a). [Color figure can be viewed in the online issue,
which is available at www.interscience.wiley.com]
conical radiation patterns in the y-z plane and omnidirectional
radiation patterns in the x-y plane. For the 5 GHz antennas, similar
conical-pattern (in the y-z plane) and omnidirectional (in the x-y
plane) radiation has been also observed but with less backward
radiation (below the x-y plane) obtained. Figure 6 gives the 3D
radiation patterns at 2442 and 5490 MHz for antennas 1 and 6.
Other frequencies in the bands of interest were also measured, and
no appreciable difference in radiation patterns was found. It can be
seen that both 2.4 and 5 GHz antennas have maximum field
strength in the lateral directions instead of horizontal directions in
elevation planes. The proposed design is favorable to ceiling-
mount AP applications in this case.
(b) Figure 7 presents the measured peak antenna gain and radiation
efficiency against frequency. The peak gain over the 2.4 GHz band
Figure 3 Measured S-parameters for the antennas of a constructed is seen at a constant level of about 2.1 dBi; the radiation efficiency
prototype; d 1.7 mm, g 1 mm: (a) reflection coefficients (S11, S22, S33
exceeds about 75%. For the 5 GHz band, the peak gain varies from
for the 2.4 GHz antennas, S44, S55, S66 for the 5 GHz antennas); (b)
3.7 to 4.6 dBi with radiation efficiency larger than 79%. Notice
isolation (S21, S31, S41, S51, S61, S54, S64) between any two of the six
antennas. [Color figure can be viewed in the online issue, which is avail- that the radiation efficiency was obtained by calculating the total
able at www.interscience.wiley.com] radiated power of the antenna under test (AUT) over the 3D
spherical radiation first and then dividing the total amount by the
input power (default value is 0 dBm) given to the AUT. Finally,
specification for 2.4 and 5 GHz WLAN operation with reflection the studies on substituting a circular ground of diameter 120 mm
coefficient below 9.6 dB (or VSWR of 2). Second, the isolation for the hexagonal ground were also conducted. The simulation
between any two antennas is found to be below 15 and 20 dB
over the 2.4 and 5 GHz bands, respectively. In general, poor
isolation occurs between the two antennas that operate in the same
frequency band, as can be observed in Figure 3(b). The variation
between S21 and S31 (or between S54 and S64) is very small largely
due to the symmetrical multiantenna structure. Notice that the
decoupling between antennas 4 and 1 is better than that between
antennas 5 and 1 despite the fact that the two antennas (4, 1 and 5,
1) are spaced the same distance apart. This behavior is probably
because the shorting portion of antenna 1 faces antenna 4. In this
case, the shorting portion acts as a shield [4, 10, 11] against nearby
fringing field from antenna 4 and also suppressing coupling effect
on antenna 4.
Figures 4 and 5 plot the far-field, 2D radiation patterns at 2442
and 5490 MHz, the center operating frequencies of the 2.4 and 5
GHz bands, in E and E fields. Because of the sequential, rotating
arrangement of the six antennas, it is only needed to analyze the
radiation of one 2.4 GHz antenna and one 5 GHz antenna. Thus, Figure 5 Measured 2D radiation patterns at 5490 MHz for antenna 6
antennas 1 and 6 are chosen to suit the convenience of defining the studied in Figure 3(a). [Color figure can be viewed in the online issue,
antenna coordinates. For the 2.4 GHz antennas, the antenna yields which is available at www.interscience.wiley.com]
2616 MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 51, No. 11, November 2009 DOI 10.1002/mop