Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02356879 2001-06-22
WO OOI39943 PCT/SE99/02347
-1-
MULTI-TRANSMITTER SYSTEM
BACKGROUND
This invention relates generally to electrical communication and particularly
to
cellular radio communication.
Modem communication systems, such as cellular and satellite radio telephone
systems, employ various modes of operation (analog, digital, and hybrids) and
access
techniques such as frequency division multiple access .(FDMA), time division
multiple
access (TDMA), code division multiple access (CDMA), and hybrids of these
7 0 techniques.
In North America, a digital cellular radiotelephone system using TDNI~A is
called
the digital advanced mobile phone service (D-AMPS), some of the
characteristics of .
which are specified in the TIAlEIA/IS-136 standard published by the
Telecommunications Industry Association and Electronic Industries Association
(T1AIEIA). Another digital communication system using direct-sequence CDMA is
specified by the TIA/EIA/iS-95 standard, and a frequency-hopping CDMA
communication.system is specified by the EIA SP 3389 standard (PCS 9900). The
PCS 1900 standard is an implementation of the Global System for Mobile
communication (GSM) that is common outside Norlfi America and that has been
introduced in North America for personal communication services (PCS) systems.
In these communication systems, communication channels are implemented by
frequency modulating radio carrier signals, which have frequencies near 800
megahertz (MHz), 900 MHz, 1800 MHz, and/or 1900 MHz. In TDMA systems and
even to varying extents in CDMA systems, each radio channel is divided into a
series
of time slots, each of which contains a block of information for a user. The
time slots
are grouped into successive frames that each have a predetermined duration,
and
successive frames may be grouped info a succession of what are usually called
superframes. The kind of access technique (e.g., TDMA or CDMA) used by a
communication system affects how user information is represented in the slots
and
frames, but current access techniques all use a slotlframe structure.
Efforts to create better antenna systems for these cellular radio
communication
systems are getting more and more attention. !n particular, mufti-lobe
antennas are
frequently mentioned as the next revolution of cell planning, and examples of
systems
using such antennas are being developed for GSM systems and wideband CDMA
CA 02356879 2001-06-22
WO 00!39943 PCT/SE99/02347
-2- _
systems. Some reasons for using multi-lobe antennas will become apparent from
the
following discussion.
F1G. 1 depicts an example of a conventional cellular radio communication
system 100, which includes a plurality of radio base stations (RBSs) 170a-170n
connected to a plurality of corresponding antennas 130a-130n. The radio
coverages of
the RBS/antenna cort~binations are depicted by corresponding cells 110a-110n,
which
are illustrated as hexagons merely for convenience. The RBSs 170a-170n, in
conjunction with the antennas 130a-130n, communicate with a plurality of
remote
terminals (RTs) 120a-120r (e.g., mobile stations) located in the cells 110a-
110n. The
communication link for signals sent from a RBS to a RT is usually called the
downlink,
and the communication link for signals sent from a RT to a RBS is usually
referred to
as the uplink.
The RBSs 170a-170n are connected to a mobile telephone switching office
(MTSO) 150, which among other tasks coordinates the activities of the RBSs,
such as
during a handoff of a communication link from one RBS or cell to another. The
MTSO
can be connected to another communication nefiniork, such as a public switched
telephone network 160 that serves users through various communication devices
like a
telephone 180a, a computer 180b, and a facsimile machine 180c.
A typical GSM-type communication system is depicted in more detail in FIG. 2.
Each RBS 170 typically handles a plurality of control and traffic channels,
which may
carry voice, facsimile, video, and other information, through a plurality of
transceivers
(TRXs) 172 that are monitored by a controller 174. Each TRX 172 may be
considered
as having a baseband part 176 that, for the downlink, receives information
signal
frames from a base station controller (BSC) 140 and a radio frequency (RF)
part 178
that appropriately modulates a RF carrier signal with portions (e.g., a slot
or slots) of
the information signal frames. Also shown in the exemplary cellular
communication
system of FIG. 2 is a RT 120 that includes a transceiver 122 for communicating
with
the RBSs on the traffic and control channels and a processor 124 for
controlling
operation of the RT.
The TRXs send and receive signals through the antenna 130, which is shown
connected separately to individual TRXs simply for convenience of explanation.
Since
a GSM system can use different carriers for different portions (e.g., slots,
or bursts) in
each information signal frame directed to a particular RT, the RF part of each
TRX may
implement carrier frequency hopping. The BSC, which typically controls a
plurality of
CA 02356879 2001-06-22
WO 00/39943 PCT/SE99/02347
-3-
RBSs, "knows° how many TRXs there are in each RBS and allocates a
number of
carriers (e.g., eight) to each TRX. Thus, information signal frames or bursts
received
by the BSC from the MTSO 150 can be directed to particular TRXs 172 in
particular
RBSs 170 based on information elements included in the signal frames; it is
left to the
RBS to handle distribution of the frames and the carrier-frequency-hopping
slots to the
appropriate TRXs. Each base band part 176 includes a processor that checks the
information received from the BSC, and based on an indication of the carrier
desired
for each slot, the appropriate base band part passes the slot information to
its
associated RF part 178 for transmission to a particular RT.
The TRX handling the control channel broadcasts control messages (such as
call- or session-setup messages exchanged by RBSs and RTs and synchronization
messages used by RTs for synchronizing their transceivers to the
framelslotlbit
structures of the RBSs) to RTs cocked to that control channel. It will be
understood that
the TRXs 172 can be implemented as a unitary device for control and traffic
channels
that share the same carrier signal.
The downlink in a given cell can be degraded by signals transmitted by other
RBSs or transmitters operating in the same frequency band. A frequency re-use
plan
can mitigate such interference by locating potentially interfering cells as
far from each
other as possible. Transmission power control can also reduce interference by
ensuring that transmitters use the minimal power needed for adequate
communication.
Interterence can be reduced still further by using a plurality of narrow beam
antennas to communicate with terminals in a cell. A narrow-beam antenna
exchanges
signals with a respective, limited geographic portion of the cell, thereby
reducing the
interterence experienced by terminals outside that portion. Today's
communication
systems typically are partitioned either into three 120°-sectors, each
serviced by a
respective one of three sector antennas, or into six 60°-sectors, each
serviced by a
respective orie of six sector antennas. It will be understood that even
smaller "sectors"
can be realized by a phased array antenna having a plurality of narrow beams
that may
be fixed or independen~y steered.
FIG. 3, for instance, illustrates a cell in a communication system 200 that
has a
RBS 220 employing an antenna (not shown) having a plurality of narrow beams
(B,, B2,
B3, B4, etc.) that extend radially from the RBS 220. Preferably, the beams
overlap to
cover the cell completely. Although not shown, the antenna may include three
separate phased array sector antennas, each of which communicates with a
120°
CA 02356879 2001-06-22
WO 00/39943 PCT/SE99/07,347
_.
swath e~ctending from the RBS 220. FIG. 3 also shows a RT 210 located within
the
beam B,, and communication proceeds between the RBS 220 and the terminal 210
using the beam B,, or perhaps in addition, one or more adjacent beams.
In such communication systems, signals can be selectively transmitted and
'5 received in particular directions by narrow-beam (high-gain) antennas,
decreasing the
iriterterence experienced by the terminals. This results in an improvement in
a
communication link's carrier to-interference (C/I) ratio and an increased
system
capacity. For example, as described in Garg et al., Analications of CDMA in
WirelesslPersonal Communications, pp. 332-334, Prentice Hall (1997), an
idealized
eight-beam antenna might provide a three-fold increase in capacity over' some
cell-
sectorization schemes.
In general, use of a multi-lobe antenna for an application like a RBS in a
cellular
telephone system implies use of at least one multi-carrier power amplfier
(MCPA),
which is a device for amplifying an RF signal consisting of a plurality of
distinct carrier
signals. Two main approaches to a RBS using MCPAs can be understood from
FIGS. 4A and 4B.
FIG. 4A depicts a RBS that reflects what might be called a MCR/MCPA
approach. Digital signals produced by a plurality of base band units (i.e.,
signals just
before digital to analog (D/A) conversion) that are intended for the same
antenna beam
are summed and.processed by a common wide-dynamic-range D/A converter. As
indicated in FIG. 4A, digital signals produced by base band parts 1, 2, 3 are
summed in
a multi-carrier radio (MCR) 1 and a MCR2. The name "multi-carrier radio" or
MCR
derives from the fact that all following analog components, such as the
modulator,
filters, and RF amplifiers, are common for the combined signals. The resulting
signal
produced by each MCR is then typically amplified, by an MCPA in FIG. 4A. Each
base
band unit chooses which antenna beam to transmit on, and sends its digital
signal to
the corresponding MCR. Since many base band units can simultaneously transmit
on
the same antenna beam, the radio and the power amplifier must be multi-carrier
ones,
with one MCR and one MCPA for each antenna as indicated in FIG. 4A.
FIG. 4B depicts a RBS that reflects what might be called a SCR/MCPA
approach. Rather than combining signals digitally at base band as in the
MCRIMCPA
approach, each base band unit is handled by a separate radio (hence the name
"single
carrier radio" or SCR), and the separate modulated carrier signals are
switched and
combined at low RF power. Final amplification to a power level suitable for
CA 02356879 2001-06-22
WO 00/39943 PCT/SE99/OZ347
-5- _.
transmission requires one MCPA for each antenna beam. In this approach,
switching
and combining are usually done using hybrid combiners since their losses are
acceptable because they occur before the final amplification.
Nevertheless, as the number of antenna beams and the number of signals to be
combined increases, the complexity and physical size of the.switching and
combining
network increases. Indeed, a system having eight base band units and eight
antenna
lobes might have as many as sixty-four RF switches and fifty-six hybrid
combiners.
European Patent Publication EP 0 593 822 to Searle et al. describes a base
station
antenna arrangement that has a plurality of antenna arrays, beam-forming
devices,
and transceivers, with a switching matrix for ccinnecting each transceiver
with one or
another of the antenna arrays via the beam-forming devices. This system
employs a
plurality of hybrid combiners and then a multi-carrier power amplifier.
Another system
is described in U.S. Patent No. 5,048,116 to Schaeffer that describes a signal
routing
system for use with eight transmitters and a plurality of frequency responsive
devices.
Schaeffer's arrangement requires a large number, sixty four, of such frequency
responsive devices; making the system physically large and difficult to scale
up.
European Patent Publication EP 0 439 939 to Davis describes a signal
divider/combiner array that divideslcombines antennas among a plurality of
transceivers. The divider/combiner array includes a plurality of controllable
switches.
It is desirable for a RBS to be compatible with the other components of
today's
typical cellular communication systems and to require no or minimal
modification of
those components. One aspect of such compatibility is that the RBS should not
require any modification of the system that decides which transceiver to use
for
communicating with a remote terminal. Thus from the point of view of the l3SC
or
MTSO, the RBSs should look substantially alike. If the BSC is to be unaware of
the
internal structure of the RBS, each transceiver in the RBS must be able to
communicate with every remote terminal in the cell. Also, atI remote terminals
in a cell
may be in the same area, i.e, in the same antenna lobe, and thus each
transceiver
must be able to transmit in each antenna lobe.
While systems depicted in FIGS. 4A, 4B or described in the documents cited
above may be capable of connecting different transceivers to different antenna
lobes,
these systems have a number of disadvantages. An important typical
disadvantage is
the physical volume needed to accommodate the large number of switches and
hybrid
combiners. Another important disadvantage is the relative inefficiency of the
multi-
CA 02356879 2001-06-22
WO 00/39943 PCT/SE99/02347
-6-
carrier power amplifiers compared to single-carrier power amplifiers. In tact,
the typical
electrical efficiency of a mufti-carrier RF power amplifier may be only about
three
percent (e.g., a hundred watts of electrical power in produces only about
three watts of
RF power out), which is markedly lower than the approximately thirty per cent
electrical
efficiency typical of a single-carrier power ampffier.
Also important is the loss through the switching and combining network, a loss
that can be greater than 10 dB. Part of.this loss may be recovered by using an
antenna having higher gain than the antennas usually used in cellular systems,
but it is
common for such systems still to have poor link budgets. This might be
acceptable in
some environments, such as high density networks of geographically small
cells, but in
general it is desirable to use a mufti-lobed antenna system to enable bigger
cells.
SUMMARY
Applicant's invention satisfies the desire for a RBS that is compatible with
the
other components of a typical cellular communication system, and can easily be
implemented in current RBSs, such as the RBS 2000 family of devices available
from
Ericsson Radio Systems AB. Moreover, Applicant's invention provides a RBS that
avoids the problems of prior devices that either do not use mufti-beam
antennas or use
mufti-carrier power amplifiers. In one useful embodiment of Applicant's
invention,
signals from six transceivers can be selectively directed to one antenna lobe
with only
about 3.5-dB loss.
in accordance with one aspect of Applicant's invention, a radio transceiver
station includes an antenna system having a plurality of beams, a plurality of
radio
transmitters, each radio transmitter including a base band part and a
radio~frequency
part having a single-carrier radio frequency power amplifier, a plurality of
filters that are
respectively connected to the plurality of radio transmitters and that are
tunable to
respective carrier signals generated by respective radio transmitters, and a
switch
matrix for connecting selected ones of the filters to the antenna system such
that
carrier signals generated by the transmitters and amplified by the single-
carrier radio
frequency power amplfiers can be radiated through selected beams. Each
transmitter
can be connected to each antenna beam independently of each other transmitter.
In other aspects of the invention, the selected filters are connected to the
antenna systems according to switch control signals based on information
elements in
information signal frames to be transmitted, the tunable filters include at
least one
CA 02356879 2001-06-22
WO 00/39943 PCT/SE99/02347
-7-
tunable resonant cavity, and the switch matrix comprises a plurality of p-i-n
diodes.
The radio transceiver station may further include a signal bus connected
between the
base band parts of the radio transmitters for selectively distributing
information signal
frames to the radio frequency parts of the radio transmitters.
BRIEF DESCRIPT10N OF THE DRAWINGS
FIG. 1 illustrates a conventional cellular radio communication system;
FIG. 2 illustrates a radio base station and a remote terminal;
FIG. 3 illustrates a radio base station having a multi-lobed antenna;
FIGS. 4A, 4B illustrate two approaches to switching and combining signals in a
radio base station;
FIGS. ~SA, 5B illustrate a radio base station in accordance with Applicant's
invention; and
FIGS. 6A, 6B illustrate switch matrices for a radio base station in accordance
with Applicant's invention.
DETAILED DESCRIPT10N
This application describes the invention in a context of a cellular radio
communication system, but it will be understood that this is just an example
and that
the invention can be applied in many other contexts.
As may be known, combining a plurality of RF carrier signals and providing the
combined signal to one antenna or antenna beam can be achieved using either
hybrid
combiners or tunable cavities. In accordance with one aspect of Applicant's
invention,
signal combining is accomplished with tunable cavities rather than hybrid
comblners~
because it is desirable to minimize losses in order to maximize range. In
another
aspect of the invention, the antenna lobe is selected after the tunable
cavities, and in
yet another aspect, single-carrier power amplifiers can be used in the
transceivers.
These aspects can be seen in the system illustrated in overview by FIG. 5A and
in
detail by FIG. 5B.
Referring to FIG. 5A, a plurality of base band parts 1, 2, . . . , N feed
respective
SCRs 1, 2, . ~. . , N and SCPAs 1, 2, . . . , N. The high-RF-power outputs of
the SCPAs
are provided to a switch and combiner matrix that directs the proper signals
to the
proper antennas or antenna beams, only two of which are indicated in FIG. 5A.
Since
switching and combining is carried out with high-power RF signals, SCPAs can
be
CA 02356879 2001-06-22
WO 00/39943 PCT/SE99/02347
-$- _
used rather than inefficient MCPAs. An important aspect of Applicant's
invention is a
switching and combining matrix that minimizes losses.
FIG. 5B shows the RF parts 11-14 of four transceivers that are coupled to
respective tunable cavfies TC11-TC14. The RF parts correspond to the SCRs
depicted in FIG. 5A. Additional transceivers may be provided but are not shown
in
FIG. 5B for clarity. Signals passed by the tunable cavities are provided to a
switch
matrix SM, only two portions A, B of which are shown in FIG. 5B for clarity.
The switch
matrix connects signals from selected ones of the tunable cavities to
respective ones of
several antennas or antenna lobes. fn FIG. 5B, only two antennas ANT11, ANT12
are
shown for clarity, It will be~understood that Applicant's invention can be
embodied in
systems having more or fewer transceivers andlor antennas.
The tunable cavities TC may be conventional, selected as appropriate for the
signal frequencies used. For some GSM cellular telephone systems, carrier
signals
are separated by about 800 KHz at about 900 MHz. It is only necessary that the
tunable cavities have bandwidths suitable for sufficiently isolating one
carrier from
another. Applicant has found that an isolation of about 10 dB currently seems
to be
sufficient. If the cavities are tunable slowly, e.g., on the order of a few
seconds, the
cavities would typically be pre-tuned to the carriers used by the RBS. Since
the RBS
would be more flexible if it could change carriers more rapidly, say on the
order of
microseconds (a fraction of a slot duration), it is desirable for the cavities
to be more
quickly tunable, but this is not necessary.
In one aspect of Applicant's invention, the transceivers are interconnected by
a
bi-directional signal bus Xbus that distributes information signal frames or
'portions of
those frames among the transceivers. !n the embodiment illustrated in FIG. 5B,
the
Xbus is disposed between the base band parts (not shown in FIG. 5B) and the RF
parts of the transceivers. In this way, bursts or frames that call for
transmission by a
particular carrier can be provided to the appropriate RF part and tunable
cavity TC,
without requiring on-the fly tuning. The information distributed by the Xbus
preferably
includes the information signal frames from the BSC and may include other
appropriate
messages, e.g., an information element or elements identifying the antenna
lobe to be
used for the bursts) or frame(s). The base band units may also "know" which
antenna
lobes to use for communicating with particular remote terminals based on well
known
beam selection techniques, many of which use information derived by the
receiving
parts of the transceivers from signals received from the remote stations.
CA 02356879 2001-06-22
WO 00!39943 PCT/SE99/OZ347
_g_
It will be appreciated that the distribution of the information signal frames
or
bursts may be carried out in other ways and still be within the scope of
Applicant's.
invention. For example, it is not necessary for an RBS to use a signal bus
Xbus for
distributing. the frames. In such an embodiment, the BSC would provide the
frames to
a controller that would distribute the frames or portions of frames to the
appropriate
transceivers (i.e., to the appropriate base band parts).
Based on the information signal frames and Xbus information, the proper RF
part causes the switch matrix SM to set its switches so that the signal can be
transmitted through the chosen antenna lobe. This arrangement enables carrier-
frequency hopping by suitably distributing information at base band, and is
compatible
with currently available RBSs.
The switch matrix SM is set so that the proper antennas or antenna lobes are
connected to which RF parts of the transceivers in accordance with switch
control
signals provided by the RF parts of the transceivers. It will be recognized
that the
switch control signals generated by the RF parts can be based on information
in the
information signal frames arriving from the BSC and/or on the Xbus
information. The
particular form of these contra! signals is not critical; it is only necessary
that they
operate the switch matrix SM in a suitable manner. With the distribution of
signals by
the Xbus, each transceiver to-antenna connection is made independent of other
such
connections in this way. It will be appreciated that although FIG. 5B
indicates an
arrangement of eight transceivers and eight antennas, the numbers of
transceivers and
antennas are easily independently scalable. It will be understood that the
more
antennas or antenna beams there are in a system, the narrower each beam is,
but also
the larger the physical size of the base station becomes.
The switch matrix SM preferably comprises a plurality of p-i-n diode switches
connected in a manner that permits any antenna lobe to be connected to any
transceiver according to the switch control signals from the RF parts. By
using p-i-n
diodes as switches, the selection of antenna port for a transmitter can be
changed
rapidly, even on a slot-by-slot (a microsecond timescale) basis. It will be
appreciated,
however, that devices other than p-i-n diodes, such as transistors (e.g.,
FETs), can be
used in the switch matrix SM.
FIG. 6A depicts switches A1, A2 that are part of portion A of the switch
matrix
SM and switches B1, B2 that are part of portion B of the switch matrix SM.
FIG. 6A is
based on an assumption that the tunable cavities TC have adequately high RF
CA 02356879 2001-06-22
WO 00!39943 PCT/SE99/02347
-10-
impedances at frequencies other than the frequencies to which they are tuned.
In case
of an open switch, both sides of the switch should have high impedance (open
circuit).
The switches are configured to ground (short) the connections, with the result
that the
switches do not handle the high RF power from the transceivers. The switches
and
conductors between the antennas and tunable cavities are preferably
implemented on
a single substrate, such as a printed circuit board (PCB), to minimize the
effects of
conductor tolerances and to maximize performance.
By setting their switch control signals so that boxh switches A1, A2 are
closed,
the RF parts 11, 12 may both be connected to the same antenna (antenna A in
FIG.
6A), enabling simultaneous transmission of different carriers through the same
antenna
or antenna lobe. The other paths through the switch matrix have high
impedances due
to the open switches B1, B2, and the length-112 conductors, where J~ is the RF
signal's
wavelength in the conductor. By closing the switches A1, B2, the transceiver
TRX11
transmits alone on antenna A and the transceiver TRX12 transmits alone on
antenna
B. It should be noted that other combinations of switch characteristics and
conductor
lengths can give the same kind of functionality.
FIG. 6B illustrates an embodiment using p-i-n diodes as the switches in the
switch matrix SM that con-esponds to the more general structure depicted in
FIG. 6A.
There is one control signal for each pin-diode switch, only four of which are
shown, with
a positive control signal causing the p-i-n diode to enter a high RF impedance
state
and a negative control signal causing the p-i-n diode to enter a low RF
impedance
state. As in FIG. 6A, the tunable cavity filters, only two of which (TC11,
TC12) are
shown, are assumed to have high RF impedance at frequencies other than the
ones
they are tuned to. Seen from the output of a tunable cavity filter, a low RF
impedance
state of a p-i-n diode is seen as a high RF impedance due to the quarter
wavelength
transformer. T~o direct the RF signal from the RF part 12 to the antenna B,
the switch
A2 should have a low RF impedance and the switch B2 should have a high RF
impedance. If more than one transmitter is transmitting on the same antenna,
the high
RF impedances of the tunable cavity filters for frequencies other than the
ones they are
tuned to makes the RF signals combine in the antenna. If the RF part 11 is
transmitting on antenna~A and the RF part 12 is transmitting antenna B, the
switch B1
is set to a low RF irrapedance, which is seen as a high RF impedance at the
antenna
port of antenna B. Therefore the signal from the RF part 12 enters the antenna
B, and
vice versa.
CA 02356879 2001-06-22
WO 00/39943 PCT/SE99/02347
-11-
From FIGS. 5A, 5B, 6A, and 6B, it can be seen that Applicant's system enjoys
an advantage of having only about 3.5 dB~loss between the power amplifier and
the
antenna (2.5 dB for the tunable cavity and filter + 1 dB for each p-i-n diode
switch).
With a 5-dB improvement due to increased antenna gain (narrower beams),
Applicant's
system thus provides about 3 dB more emitted effective isotropic radiated
power
. (EIRP).than an ordinary cellular system. Moreover, Applicant's invention
enables, in
addition to better C/I performance (i.e., about 5 dB better than a
conventional system),
frequency reuse 3 in most installations and the possibility of frequency reuse
1 in some
installations, and 6 dB better sensitivity than conventional base stations.
Another significant advantage of Applicant's invention is that the
transceivers
can use single-carrier RF power amplifiers that are considerably more energy
and
space efficient than multi-carrier RF power amplifiers used in other systems.
As noted
above, the electrical efficiency of a single-carrier power ampfr~er can be a
factor of ten
greater than the electrical efficiency of a multi-carrier power amplifier,
enabling an RBS
to broadcast at higher power or with less electrical power consumption or with
a
combination of these.
It will be appreciated by those of ordinary skill in the art that this
invention can be
embodied in other specific forms without departing from its essential
character. The
embodiments described above should therefore be considered in all respects to
be
illustrative and not restrictive. The scope of Applicant's invention is
determined by the
following claims, and all modifications that fall within that scope are
intended to be
included therein.
