e011020.PDF
(
783 KB
)
Pobierz
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
GENERAL
INTEREST
Universal Mobile
Telephone Service
Part 2: Third generation mobile technology
By G. Kleine
The introduction of Third Generation mobile phones will offer
vastly increased data capacity ensuring the proliferation of
mobile Internet use and new multimedia uses.
rent GSM systems use FDD with
respect to the uplink and downlink
capacity. An alternative to FDD is
the Time Division Duplexing (TDD)
System that has been standardised
by Bosch, Siemens and Alcatel. A
major advantage of this standard is
the use of non-paired frequency
bands thereby allowing unsymmet-
rical channel capacity between the
uplink and downlink channels. This
is much more suited to the ‘bursty’
nature of TCP/IP data exchange that
occurs when accessing Internet
information. In this case uplink data
could consist of little more than a
web page address which results in a
downlink data flow of many pages of
data. For FDD a Duplex bandpass fil-
ter with relatively sharp cut off char-
acteristics is used to prevent cou-
pling of send and receive data. With
TDD equipment it is necessary to
strictly control the synchronisation
between the base stations so that
adjacent cells can function correctly
together and that the entire network
can operate synchronously.
The two UMTS frequency ranges
each 60 MHz wide split into 12 chan-
nels each with a 5 MHz bandwidth
are at the moment available for use
in European countries and in the
majority of Asian countries. In the
USA the situation is not quite so
clear. The USA currently uses the
To ensure that global roaming would be pos-
sible for third generation mobile phones the
World Radiocommunications Conference
(WRC) in 1992 defined the operating fre-
quencies for UMTS to be 1885 MHz to
2025 MHz (uplink) and 2110 MHz to
2200 MHz (downlink). Uplink is defined as
information passing from the subscriber to
the base station while downlink is informa-
tion passing from the base station to the sub-
scriber.
Figure 1
shows UMTS frequency allo-
cation. The upper band of these frequencies
is reserved for future Mobile Satellite
Service (MSS). The core frequencies
for Frequency Division Duplexing
(FDD) with W-CDMA modulation are
1920 MHz to 1980 MHz for the uplink
and 2110 MHz to 2170 MHz for the
downlink. These maintain a duplex
distance of 190 MHz between the
uplink and downlink. FDD is
favoured by Nokia and Ericsson and
provides identical capacity for the
uplink and downlink channels. Cur-
20
Elektor Electronics
1/2001
GENERAL
INTEREST
in the mobile phone world: Vodafone, BT3G,
one2one (owned by Deutsch Telecom) and
Orange (owned by Vodafone but currently on
the transfer list). The fifth licence was
awarded to TIW, a relative newcomer com-
prising Canada Telesystem International and
Hutchison Whampoa of Hong Kong. The U.K.
auction for 3G licences raised £22.48 Billion
for the Government.
More details of the licencing can be found
in UMTS forums on the Internet.
GSM900
GSM1800 DECT
PCS
(USA)
880
915 925
960
1710 1785 1805 1880 1900
1930
1990
Uplink
Downlink
Uplink
Downlink
paired FDD/WCDMA
Uplink
Downlink
MSS
MSS
1920
1980
2010
2025
2170
806
960
1700
1885
2110
2200
2500
2690
frequency range for UMTS (12 FDD blocks of 5 MHz each)
A refined modulation method
additional frequency range according to WARC 2000
000183 - 1 - 11
The classical methods of providing multiple
communication channels by using FDMA and
TDMA were described in part 1 of this article
in the previous issue of
Elektor Electronics.
Figure 2
shows diagrammatically how the
channel is subdivided in frequency and time
when either FDMA or TDMA is used. Another
method that we have not looked at yet is also
shown. UMTS will use a Code Division Multi-
ple Access (CDMA) method. CDMA allows all
users to send simultaneously using the same
frequency. At first glance this may seem a
very poor method: If everyone is using the
same frequency how is it possible to select
just one transmitted signal at the receiver? A
simple filter would attenuate all channels
equally because they are all sending on the
same frequency. The key to this method is
that each transmit and receive channel is
assigned a unique code or ‘Pseudo Noise’
(PN) sequence. Transmitted data is combined
with this sequence before transmission. At
the receiver, we detect what just seems like
noise from these many transmitters all send-
ing together however when this received
‘noise’ is passed through a correlator which
uses the same code sequence as the trans-
mitter, the original transmitted signal is mag-
ically recovered.
This concept may be difficult to grasp but
as an analogy imagine that you are having a
conversation with a friend in a lift. The lift
stops and two people enter conversing in
Japanese, next two people enter conversing
in Russian. If we assume (crucially) that
everyone in the lift can only understand their
own language then you can see that these
three conversations can occur simultaneously
without any mix-up of information. Some-
where deep in your brain the equivalence of
correlating the sounds at your ears with the
code sequences of the English language is
occurring.
The process of spreading can be seen as
sacrificing a narrow bandwidth signal for a
signal that can operate in a high noise envi-
ronment and is more immune to noise. A sim-
plified example is given here in
Figure 3
.
Figure 1.UMTS frequency allocation.
1.9 GHz band for its second genera-
tion phone network called Personal
Communications System or PCS. The
plan here is to free-up some of these
frequencies for UMTS use in the
coming year.
It is also anticipated by the WRC-
2000 that the 806 MHz to 960 MHz
band will be used for UMTS when
the current GSM era comes to a
close. The 1710 MHz to 1885 MHz
will also become available when the
DCS 1800 service comes to an end.
Apart from these frequencies there
is also a part of the spectrum which
is currently used by DECT cordless
phones. Lastly 2500
2690 MHz is also eventually antici-
pated to be used for future capacity
increase of the UMTS network in.
UMTS frequency licences
The allocation of these highly desir-
able 5 MHz frquency pairs has been
handled differently in different coun-
tries. In Spain, Norway and Sweden
for example, these licences have
been distributed on a practically cost
free basis whereas in most other
countries they have been auctioned
off to the highest bidder. Here in the
U.K. five licences were available and
four of these went to the big hitters
MHz to
t
FDMA
P
1
12345
f
f
1
f
2
t
TDMA
A
L
P
C
N
B
M
A
L
12345
f
f
1
f
2
t
CDMA
P
A, B, C, ....
L, M, N, ....
f
f
1
f
2
000183 - 2 - 12
Figure 2. FDMA, TDMA and CDMA access methods.
1/2001
Elektor Electronics
21
GENERAL
INTEREST
a
1
0
1
0
0
1
b
0
0
1
0
1
1
+1
-1
+1
-1
+1
-1
+1
-1
+1
-1
foreign Data
Data
foreign
Chip Code
Chip Code
+1
-1
Spread Signal
foreign
Spread Signal
+1
-1
+1
-1
Chip Code
+1
-1
own Chip Code
Spread Signal
*
Chip Code
Spread Signal
*
own Chip Code
+1
-1
+4
+ Threshold = +3
+4
+ Threshold = +3
Integration
0
Integration
0
-4
- Threshold = -3
Sampling
-4
- Threshold = -3
Sampling
+1
-1
Received Data
0 no signal received
+1
-1
Received Data
Received Data
1
0
0
1
0
1
000183 - 2 - 13
000183 - 2 - 14
Figure 3.The principle of CDMA communication. a) Reception of the desired signal b) Reception of an unwanted signal.
Encoding of the transmission signal and
reception of the desired signal with the cor-
rect Chip sequence is shown in
Figure 3a
.
The Chip sequence data rate is four times the
actual data rate in this example. Processing
gain is a measure of by how much the band-
width of the data signal is increased by the
spreading process. It is defined as the ratio of
the chip rate to the data rate. Each input data
bit is multiplied by a four bit long chip
sequence. A high is +1, a low is –1 and no
signal is 0. The resultant transmitted data
signal after multiplication with the spreading
signal has a bandwidth of four times the data
rate.
By signal reception the input data integra-
tor is initially at set to zero. The integrator
amplitude is given by the processing gain for
the system. The sample time of the waveform
occurs at the end of each integration period.
The sampling threshold is here set to +3 and
–3. At the output of this process we have
recovered the actual data stream which was
input to the transmitter with an additional
time delay.
Figure 3b
shows the same process but
this time with an unwanted signal with an
unknown chip sequence. The first two lines
show the unwanted data signal and its chip
sequence. Line three is the spread signal. As
before line four is the chip sequence used by
the receiver here we are using the same chip
sequence as in
Figure 3a
. The next line
shows the product of the above two
lines and this is input to the integra-
tor. We can see that at no point dur-
ing the integration process does the
signal step above or below the +3
and –3 threshold so the output
remains at zero. Zero indicates that
no signal has been detected. From
this we can see that the chip
sequence should be carefully chosen
so that it has a sharp auto-correla-
tion peak i.e. that it does not allow
any other unwanted code to be mis-
taken for the desired code. Cross-
correlation between two chosen chip
sequences should be as small as
possible (optimally zero).
quency it is crucial that each
received signal must reach the
receiver with approximately equal
signal strength in order that the inte-
gration process at the receiver func-
tions correctly (see The near/far
problem of W-CDMA). Good quality
receivers must be employed in the
base stations and in the mobile
handsets along with transmitter
stages with good linearity otherwise
intermodulation products will have
adverse effects on the signal quality.
A W-CDMA
communication system
A simplified block diagram of a W-
CDMA communications system is
shown in
Figure 4.
This example
uses a data stream of 1,024 Mbits/s
and is passed through an encoder
and interleaver. The interleaver pro-
vides error protection by shifting bit
positions so that the effects of inter-
ference bursts are distributed
throughout the data stream and can
be corrected once the signal is de-
interleaved at the receiver. The resul-
tant data stream is now multiplied
by the chip sequence (12.288 MB/s
in this example). This processing is
equivalent to performing an XOR
The Wideband variant
Originally CDMA was used in a mil-
itary environment and employed a
relatively narrow signal bandwidth
of 100 kHz. UMTS signals require a
bandwidth of 5 MHz in order to sup-
port a data rate of 2 Mbit/s and a
chip rate of 4,096 Mbit/s. The corre-
spondance of the chip rate to the
data rate is relatively small (2 Mbit/s
or 2 MHz) yielding a relatively small
processing gain. This low processing
gain means that when several sub-
scribers are sending on the same fre-
22
Elektor Electronics
1/2001
GENERAL
INTEREST
operation between the data and the
spreading code (also known as ‘chip-
ping’). A bandpass filter limits the
signal bandwidth to that of the chip
rate. The original bandwidth of
1,024 MHz is spread to 12,288 MHz
and the processing gain is 12.
The signal is now passed through
the power amp (PA), to the antenna.
As the signal travels through the
ether to the receiving antenna it is
subject to interference and noise. It
also mixes with transmissions from
other mobiles using the same fre-
quency. The resulting noise-like sig-
nal is picked up by the receiver
antenna, amplified by the input
amplifier (LNA) and filtered through
the chip rate bandwidth filter. The
signal is now passed to the correla-
tor this will mix the signal with a
locally generated chip sequence that
must be synchronised with the
transmitter chip sequence. At the
output of the correlator the actual
transmitted data stream is reconsti-
tuted. Now after the de-interleaving
and decoding the DATA OUT signal
will correspond to the original DATA
IN signal at the input.
Figure 5
shows the comparison
between the present day radio cell
structure compared to the radio cells
that will be used in the W-CDMA
system. Uplink and downlink capac-
ity can be adjusted according to
needs.
1.024 MHz
1.024 MHz
P
P
P
12.288 MHz
P
12.288 MHz
Noise
&
Interference
f
f
carrier
f
f
carrier
f
f
DATA
IN
1.024 Mbit/s
Encoder
&
Interleaver
Deinterleaver
&
Decoder
DAT
A OUT
1.024 Mbit/s
Digital
Correlator
PA
BW =
12.288 MHz
BW =
12.288 MHz
Chip Code
12.288 Mbit/s
Chip Code
12.288 Mbit/s
Other Users
f
LO
f carrier
Synchro-
nization
VCO
000183 - 2 - 15
Figure 4. W-CDMA Communication system.
dynamically control the power output
of each mobile so that the received
signal strength at the base station
from each mobile is equal. The out-
put power of a UMTS mobile will
need to be adjustable over a range of
70 dB i.e. 1:10 million! The power
amplifier must also be capable of a
switching speed of 1500 gain steps
per second. In contrast current GSM
mobiles only need to alter power out-
put a few steps per second.
with Vodafone to provide pilot implementa-
tion of a UMTS system in the UK.
Alcatel and Fujitsu are jointly working on
development of GPRS, EDGE and UMTS tech-
nology and are in the process of promoting
this on a bus tour of Europe demonstrating
the principles and possibilities of new UMTS
technology. Many companies are already pro-
ducing dedicated UMTS chipsets. Infineon
Technologies of Austria have a purpose built
UMTS development centre This facility goes
under the name of Danube Integrated Circuit
Engineering (DICE) and it has already in April
2000 delivered the first UMTS chip for the
Japanese market. Toshiba Corporation have
also set up a small Telecommunications
Research Laboratory in Bristol with the aim
of developing UMTS chipsets.
As for the implementation timescale for
UMTS it is anticipated that by 2001 we can
expect the first field trials of a UMTS network
and by the end of 2002 the system should be
partially implemented leading to total cover-
age by 2005.
By 2010 GSM will probably still have many
users, especially using GPRS and EDGE
enhancements.
The future looks bright
Many companies are currently co-
operating to develop UMTS base sta-
tions and prototypes of UMTS hand-
sets. Key players in the production of
net equipment and base stations are
Ericsson (Sweden) and Nokia (Fin-
land). Both companies have collabo-
rated closely with research facilities
and institutions to produce the first
prototype of a UMTS system. Nokia
has also won many contracts in
China and Ericsson is collaborating
The near/far problem of
W-CDMA
The principle that many subscribers
in the same cell can use the same
frequency relies on the necessity
that each signal arriving at the base
station must have approximately
equal signal strength so that the de-
correlation process can function cor-
rectly. This applies to the mobile
5 km away from the base station as
well as the user just across the road
from the base station. This is the so
called near/far problem and going
back to our analogy, it is equivalent
to some people in the lift shouting,
this is sure to upset everyone else’s
ability to carry on their conversation.
Unwanted signals cannot be
removed with filters because they
are using the same frequency as the
desired signal. The method used in
W-CDMA is for the base station to
(000183-2)
A
ABC
C
C
ABC
ABC
B
ABC
B
B
ABC
ABC
A
A
ABC
ABC
C
ABC
ABC
GSM and similar
W-CDMA
f
1
f
2
00018 - 2 - 16
Figure 5. Radio cells and frequency use
1/2001
Elektor Electronics
23
Plik z chomika:
TirNaNog
Inne pliki z tego folderu:
bge.jpg
(26 KB)
detail11.htm
(6 KB)
detail1.htm
(5 KB)
detail10.htm
(4 KB)
e011016.PDF
(1901 KB)
Inne foldery tego chomika:
1974
1975
1976
1977
1978
Zgłoś jeśli
naruszono regulamin