NO 1,1996 - ericssonhistory.com

RE v IE W
THE TELECOMMUNICATIONS TECHNOLOGY JOURNAL
NO 1,1996
Fixed cellular systems - an alternative way to provide basic telephony
New power distribution and earthing for AXE
Telecool Aero - A tailor-made cooling system for small telephone exchanges and containers
Ericsson echo cancellers - a key to improved speech quality
TEMS - A system for testing and monitoring air interfaces

CONTENTS
No. 11996 • Vol.73
Cover:
Fixed cellular systems offer an attractive alternative
means of extending telephony services in both
developing and developed regions of the world.
Fixed cellular systems - an alternative
way to provide basic telephony
New power distribution
and earthing for AXE
Telecool Aero - A tailor-made
cooling system for small telephone
exchanges and containers
Ericsson echo cancellers - a key
to improved speech quality
TEMS - A system for testing and
monitoring air interfaces
4
14
19
25
34
CONTENTS
Previous issues
No. 1, 1995
APZ 21220 - The New High-end
Processor
Measuring Quality of Service in
Public Telecommunications Network
AXE 10 System Processing Capacity
Using Predictions to Improve
Software Reliability
Test Marketing of Mobile
Intelligent Network Services
AXE 10 Dependability
No. 2, 1995
A Swedish Airborne Early Warning
System Based on the Ericsson
Erieye Radar
High-performance Packaging for a
RISC Processor Application
D-AMPS 1900 - The Dual-Band
Personal Communications System
The Apollo Demonstrator - New Low-
Cost Technologies for Optical
Interconnects
Development of AXE for New and
Very Demanding Transit/Tandem
Switching Application
No. 3, 1995
System T - A Modular System
for Wide Area Paging
Using DECT for Radio
in the Local Loop
A System for Flexible Service-Independent
Access Network Solutions
Handling Overload in AXE 10
UniSwitch - A New Flexible STM
Switch Fabric Concept
No. 4,1995
The Introduction of UniSwitch in
AXE 10
Copper Enhancement
MINI-LINK E-ANew Link for
Flexible Transmission in Cellular
Networks
Dynamic Routing in Circuit-Switched
Networks
Ericsson Review - 70 Years Young
Ericsson Review © Telefonaktiebolaget L M Ericsson • Stockholm 1996. • Responsible publisher Hakan Jansson • Editor Steve Banner
• Editorial staff Eva Karlstein • Layout Paues Media • Address Telefonaktiebolaget L M Ericsson S-126 25 Stockholm, Sweden
• Fax +46 8 681 27 10 'Published in English and Spanish with four issues per year • Subscription one year USD 45
Ericsson Review No. 1. 1996 1

CONTRIBUTORS
in this issue
Cecilia Brandstrom is manager of sales and marketing support,
Cellular Systems American Standards, Ericsson Radio Systems AB.
She received her MSc in Electronics and Industrial Marketing from
the Linkoping Institute of Technology in 1991 and joined Ericsson
the same year.
Hakan Andersson is manager of customer technology partnership,
Cellular Systems American Standards, Ericsson Radio Systems AB.
He holds an MSc in Electrical Engineering and a PhD in Physics
from the Royal Institute of Technology in Stockholm.
Daniel Dufour has been working in the CMS88 Systems Design &
Management Department at Ericsson Research Canada since
1984. He received a BSc in Electrical Engineering from the University
of Montreal, Ecole Polytechnique de Montreal.
Karl Forsselius is manager of strategic business development in
the Radio Communications business area. He received his MSc in
Electrical Engineering from the Royal Institute of Technology in
Stockholm in 1988 and joined Ericsson the same year. He is currently
working on radio access strategies and the evolution of
mobile systems.
Mats Barkell holds a position as Expert in power technology at
Ericsson Telecom. He received his MSc in Electrical Engineering
from the Lund Institute of Technology in 1973.
Per Haging holds a position as Expert in power supply systems at
Ericsson Telecom. He joined Ericsson in 1960 and has been working
in the field of power supply and earthing ever since.
Dag Heiistenius joined Ericsson Components AB in 1994 as a
designer of cooling equipment. Currently, he is manager for CE marking
projects in the field of EMC. He earned his MSc in Mechanical
Engineering from the Royal Institute of Technology in 1993, specialising
in refrigeration technology and applied thermodynamics.
Anders Eriksson is working in the Echo Cancellation and Noise
Reduction Research group at the Radio Systems & Technology core
unit of Ericsson Radio Systems AB. He received an MSc in Electrical
Engineering from the Linkoping Institute of Technology in 1989,
and a PhD in Automatic Control from Uppsala University in 1994.
Gunnar Eriksson is manager of the signal processing section at
the Echo Canceller Competence Center of Ericsson Radio Systems
AB. He received his MSc in Electrical Engineering from the Royal
Institute of Technology in Stockholm in 1970.
Johnny Karlsen is manager for the Echo Cancellation and Noise
Reduction Research group at the Radio Systems & Technology core
unit of Ericsson Radio Systems AB. He received his MSc in Electrical
Engineering from the Linkoping Institute of Technology in 1990.
Anders Roxstrom received his MSc in Electrical Engineering from
the Royal Institute of Technology in Stockholm in 1992. He joined
Ericsson Radio Systems AB in 1994 and is currently working on
the speech quality of echo cancellers at the Echo Canceller Competence
Center.
Teresa Vallon Hulth is product manager at the Echo Canceller Competence
Center of Ericsson Radio Systems AB and responsible for
the ECP303 requirements specifications. She holds an MSc in
Physics and a PhD in Theoretical Physics from the International
School for Advanced Studies in Trieste, Italy.
Rikard Lundqvist, is manager of market and product strategy at
Erisoft AB, Skelleftea, in the Mobile Phone Applications business
area. He joined the company in 1991 and is currently working on
the definition of new products, analyses of customer needs and
the building of strategic customer relations. He earned his MSc
in Computer Science and Engineering from the University of Lulea
in 1992.
2
Ericsson Review No. 1. 1996

FROM THE EDITOR
Steve Banner
In today's fast-moving world of telecommunications
it's easy for industry
observers and insiders alike to become
caught up in the whirl of technology. For
example the remarkable explosion in the
use of the mobile telephone over the past
decade has seen its level of social acceptance
evolve from one of almost novelty
value to the practically indispensable tool
of business and social life which it has
become today. New and improved applications
such as messaging and fax capabilities
are constantly adding to the usefulness
of the mobile telephone.
The bold new trend of the 90s
The accessability of tremendous quantities
of information through the Internet
appears set to revolutionise social and
business life in the 90s and has thus
emerged as the bold new mass market
trend in telecomms. Some would argue
that in the world of the Internet, "content
is king". But technology is not far behind
and already it is possible to combine
mobility with Internet access through a laptop
computer connected to a mobile telephone.
In such a dynamic world of communications
can Dick Tracy wristwatch
radios be far away
But let us not forget that none of the
above could have occurred without the
continuous development of the infrastructure
upon which the systems and features
are built. Somewhat akin to a visit
to Disney World, it is easy to marvel at the
rides, buildings and other above-ground
structures while being unaware of the vast
underground network of tunnels, wiring
and hydraulic equipment without which
the world's most popular theme park
would be nothing more than a static display.
This issue of Ericsson Review is thus
devoted largely to the less visible technology
without which the consumer products
could not function.
While it is a given fact that telecomms
equipment must be powered by stable
electrical sources, the rapid trend
towards miniaturisation of equipment has
resulted in a growing susceptibility of that
equipment to the potential damage that
can be caused during operation by unreliable
power and earthing systems. The
continuous refinement of Ericsson's highohmic
distribution system for power and
earth is described in this issue, as it has
culminated in a new and improved version.
However because of the heat produced
by power supply systems and the telecomms
equipment they support, the need
for an effective air cooling system arises.
In small or remote equipment installations
it is often difficult to provide cooling
which is effective both in terms of size
and performance. Ericsson's Telecool
Aero provides such a system, which is reliable
and effective in a wide range of conditions
- even during a mains supply failure.
Testing the air interface
While considerations of power and cooling
are vital to the performance of a network,
the network's grade of service under
operating conditions is also a subject
under continuous scrutiny. Thus apart
from monitoring switch performance,
mobile telephone network operators must
continuously measure the performance of
their link to their customers - the air interface.
As the air interface has evolved and
increased in complexity with a number of
digital cellular standards, the challenge of
monitoring the interface has also become
more complex. The TEMS (test mobile system)
was therefore developed to allow the
close scrutiny of the performance of the
air interface in each of the digital cellular
standards supported by Ericsson.
Another closely-watched aspect of
mobile telephone system performance is
that of speech quality. As networks
become more complex, an undesireable
side-effects has been an increase in echo
on some types of connections. The echo
canceller device has thus been widely
employed to eliminate this problem and
Ericsson's latest version is described in
this issue.
The application of technology in innovative
ways has served to change the way
we live. As an example, cellular telephony
systems are almost always thought of
as being used exclusively by mobile customers
- yet there is no reason that this
should be the case. Ericsson offers its
fixed cellular systems as a means by
which telephone service can be rapidly
offered to the home in a wide range of
locations. Considering that according to
the ITU-T some 50% of the world's population
has never made a phone call, such
systems have vast potential to change
many lives. But let us not forget, to return
to my earlier point, that such highly-visible
systems would cease to function without
their supporting structures "underground".
Steve Banner
Editor
Ericsson Review No. 1, 1996 3

Fixed cellular systems - an alternative way to
provide basic telephony
Cecilia Brandstrom, Hakan Andersson, Daniel Dufour and Karl Forsselius
In wired networks, the local loop represent around 50% of telecom operators'
total investments. The problems and costs associated with providing
the local loop are also two key obstacles to the expansion of telecommunications
in developing countries. But with the success and rapid development
of cellular telephone systems, the benefits of wireless technologies
are becoming increasingly obvious, in terms of rapid deployment, flexibility,
high reliability and cost-effectiveness. Accordingly, Ericsson has developed
fixed cellular systems conforming to all current major standards.
The authors describe the fixed cellular solution based on the D-AMPS standard,
which offers an attractive means of extending telephony services, in
particular in developing countries and for operators planning to combine
fixed and mobile telephony in the same network.
Fig. 1
Wireless access may be the solution to provide
basic telephony to a large part of the world's population,
in a quick and cost-effective manner.
Box A
Abbreviations and acronyms
AC Authentication centre
ACA Adaptive channel allocation
BTS Base transceiver station
BS Base station
BSC Base station controller
CAS Channel-associated signalling
COPD Cellular digital packet data
DECT Digital Enhanced cordless telecommunications
E-TACS Extended total access communications
system
GSM Global system for mobile telecommunications
HCS Hierarchical cell structure
HLP./SCP Home location register/Service control
point
IN Intelligent network
ISDN Integrated services digital network
ITU The international telecommunication
union
LE Local exchange
MC Message centre
MLT Multi-line terminal
MSC Mobile switching centre
NMT Nordic mobile telephony
OSS Operation and support system
PBX Private branch exchange
PCS Personal communication services
PDC Personal Digital Cellular
POTS Plain old telephone services
PSTN Public switched telephone network
PTO Public telecom operator
RLL Radio in the local loop
SLT Single-line terminal
TDMA Time division multiple access
VLR Visitor location register
Telecommunications networks have
undergone massive changes over the
past two decades due to the introduction
of digital transmission and switching and
the evolution of fibre optic technology.
The benefits of such technologies have
been slow to appear in the local loop -
the final connection from the local
exchange to the subscriber's premises -
but that is now changing, with radio representing
a new and attractive alternative.
For the operators, investments in the
local loop represent a very large percentage
of their costs, typically around 50%,
due to the extensive and labour-intensive
civil engineering work involved. The local
loop also represents a large portion of
operational costs, because maintenance
and repair in this part of the network are
costly and time-consuming.
Over the last decades, the telecom
infrastructure has been recognised as
one of the key factors that affect economic,
social and cultural development in both
developing and industrialised countries.
Almost universally, governments are trying
to speed up the provisioning of highquality,
low-cost telecommunication to all
parts of society.
Deregulation of the telecommunication
sector and privatisation of public telecom
operators (PTO) have become global
trends. Clearly, the resulting growth has
been fastest in market segments where
competition is in place, e.g. mobile
telephony, long-distance traffic, private
networks and, in some countries, in traditional
telephony for fixed subscribers.
In this latter case the local loop is of
increasingly strategic importance to the
operators as a means of controlling the
delivery of service to the subscribers and
because of its impact on operators'
costs. At the same time, the size of the
investment required to match the incumbent
operator means that introducing real
competition in the local loop poses considerable
problems.
However, here too, changes are taking
place. New fibre-based solutions, flexible
multiplexers, compact remote concentrators
and new solutions for telecom
services over cable-TV networks are
examples of technologies now used by
operators to build more effective access
networks. And with the success and rapid
development of cellular (and cordless)
telephone systems, the benefits of wireless
technologies have become more
obvious both for operators and regulators,
and have enabled the development
of reliable and cost-effective equipment
that can be used as an alternative and
complement to wired systems.
Why wireless
Utilising wireless technologies instead of
traditional copper wire offers a number of
benefits to the network operator in several
respects:
Cost
Deployment of wireless systems is costeffective.
Once the initial wireless coverage
has been established, the size of
a network can be matched precisely to
demand, allowing the operator to expand
his network in line with subscribergrowth,
thus reducing the upfront investment. The
cost of building a radio system, which
requires considerably less civil engineering,
is in many cases lower than the cost
of building the wired equivalent.
Operation and maintenance costs for
radio systems are relatively low. There are
4 Ericsson Review No. 1, 1996

no wires in the ground that can be accidentally
cut off, or poles that break during
storms, etc. One of the basic characteristics
of the radio part of a wireless
system is that it serves simultaneously
as a trunk, a multiplexer and a concentrator
of traffic.
Time
Wireless systems can be rapidly
deployed. Today, complete cellular systems
are often installed in a
few months, and service can start as
soon as the first radio base station is
installed and connected to a switch (a
mobile switching centre, MSC, or local
exchange, LE).
Revenue
The rapid deployment of wireless
systems enables the operator to offer services
over a wide area in a very short time.
Service revenues are maximised at early
stages of network roll-out, and these revenues
can be used for expansions.
Reliability
Wireless systems have often proved to
be very reliable in the case of natural disasters,
such as earthquakes and hurricanes.
Even in the case of failures, the
mean time to restoration is very short
compared with the traditional copperwired
networks.
Wireless systems also provide redundancy-
a failure in one channel at a radio
base station will only reduce capacity (or
increase blocking) but no subscribers will
lose their service.
Flexibility
Wireless systems are highly flexible and
can be deployed to provide service in
every type of environment, from dense
urban to remote rural areas. The precise
location of subscribers is irrelevant in a
wireless system, provided that they are
within the coverage area, and such a
system feature is vital for the operator,
who is uncertain of which residents will
sign up for his service. This is often the
case in markets where not every household
can afford a telephone or where several
local access providers compete for
potential subscribers.
Wireless equipment can easily be relocated
to other areas, if there are changes
in traffic behaviour or population density.
With wireless, the operator has the
option of providing local and wide area
mobility (cellular) depending on his
licence.
Wireless technologies -
An overview
There are two basic solutions in the case
of wireless access: radio in the local loop
(RLL) and fixed cellular systems, see
Fig. 2.
An RLL system replaces the copperwire
between a local exchange and the subscriber
premises. It can be connected to
Fig. 2
There are two basic solutions for wireless access:
A radio-in-the-local-loop (RLL) system is a dedicated
access system that connects to any manufacturer's
local exchange (LE). A fixed - or mixed -
cellular system is based on a complete cellular
system with mobile switching centres (MSC), etc.
For fixed-cellular users, special terminals that interface
standard telephone sets are used, but ordinary
cellular terminals can also be used In the same network,
with individual restrictions on mobility or full
cellular mobility (mixed cellular).
Ericsson Review No. 1, 1996 5

Box B
D-AMPS Fixed Cellular Services
Basic Services
Call waiting
Call transfer
Three-party service
Data and fax, 9.6 kbit/s
Do not disturb service
Intelligent network services
Business groups
Flexible call forwarding
Outgoing call restriction
Incoming call acceptance
Selective call rejection
Private numbering plan
Optional features
Calling number identity
Voice mail with notification
Priority access
Short message service (alphanumeric messaging)
CDPD (cellular digital packet data)
Location-based charging
BoxC
GSM, DCS1800 and PCS 1900
fixed cellular systems
Since its launch in Europe, the digital GSM standard
has evolved into a global standard, first at
900 Mhz and later at higher frequency bands - in
Europe and Asia as the DCS 1800 standard, and
in North America as the PCS 1900 standard. By
the end of 1995, GSM/DCS and PCS systems
had around 12 million subscribers in 75 countries
on four continents, and more than 100 networks
had been taken into commercial operations.
Fixed cellular systems from Ericsson are available
for all these frequency bands and standards,
while infrastructure such as base transceiver stations
(BTS), base station controllers (BSC) and
mobile switching centres is available from several
suppliers in a fiercely competitive worldwide
market-place.
any manufacturer's LE through a standard
interface, e.g. two-wire, CAS (channelassociated
signalling) or the ETSI standard
V5.1 or V5.2. An RLL system is transparent
to the services offered by the LE,
depending on the capabilities of the RLL
system selected. Standard telephone
sets are used on the subscriber side.
On the other hand, a fixed cellular
system is a complete cellular system
including MSCs, radio base stations, IN
nodes, etc. In addition, it is provided with
a subscriber terminal allowing standard
telephone equipment to be connected.
The services offered are the same as
those of the corresponding mobile cellular
system, which typically includes a
number of advanced services, see Box B.
For the operator, the key benefit of a fixed
cellular system is its ability to offer both
fixed and mobile services in the same network.
Ericsson has two dedicated RLL products:
RAS 1000 which is based on cellular
technologies and DRA 1900 which is
based on the cordless DECT standard 1 .
A third product, ACTRAN, is a flexible
point-to-multipoint system that can be
applied as an RLL-system to provide
direct customer access or be used as a
transport medium in the access network.
Fixed cellular systems from Ericsson
are available for all major standards on
the market:
- NMT 450/900
- TACS/E-TACS
- AMPS/DAMPS 800/1900
- PDC 1500
- GSM/DCS1800/PCS 1900, Box C.
This article gives a more detailed description
of Ericsson's fixed cellular solution
for the D-AMPS standard, but the principles
described can be applied to all cellular
products.
Choosing the solution
There is a great demand for telecommunications
in the world. According to the
ITU, more than two-thirds of households
around the globe still do not have a telephone
2 . The distribution of lines is also
very uneven: an average of 45 lines per
100 people in the Western world compared
with three lines per 100 people in
Asia. One encouraging sign is that the
teledensity in developing countries
increases more rapidly than was the case
in the industrialised world. But, using
wireless, it can be made even faster.
The decision to choose RLL or fixed cellular
will vary, depending on specific circumstances
in each individual case. Several
factors will affect this decision:
- Spectrum allocation. This is a key issue
for all operators who are considering
wireless solutions.
- Whether the operator is or expects to
be allowed to provide mobile sen/ice.
This consideration is especially important
not only to a new, competing operator
but also to the traditional PTO, who
often has both mobile and fixed operations
and thus needs to consider which
is the most cost-effective solution for
certain parts of his service area.
- Whether the operator has specific service
obligations imposed by regulatory/licensing
conditions. For example,
the licence may require that the operator
should provide a certain number of
lines in a short time. Another case may
be that a specific number of lines must
be provided to rural areas in exchange
for a licence in the more profitable
industrialised urban areas.
- Whether the operator can get leverage
from - or has investments in - a fixed
network or some other type of infrastructure.
- Whether the operator is new or incumbent.
- Potential subscriber base and penetration.
- The type of area or combinations of
areas to be served (urban/suburban/rural).
- What service level - basic/advanced
POTS, ISDN, mobile, etc - the operator
is planning to offer.
- What kind of end-customers - residential,
small/large business, etc - the
operator will address.
The fixed cellular concept
A mixed network
In Ericsson's fixed cellular network, radio
transmission takes the place of wires,
connecting the subscribers to the network.
The system has exactly the same
architecture as that of ordinary mobile
networks and uses the same type of
equipment. The only difference is that the
subscribers will be connected to the network
with ordinary two-wire terminals,
equipped with a radio interface, and that
their mobility can be limited to a single
cell or cell sector.
6 Ericsson Review No. 1, 1996

Using the cellular system as a base
when providing fixed service offers the
operator a high degree of flexibility.
Through the use of exchange data, the
Ericsson system will be able to differentiate
between a mobile and a fixed subscriber,
allowing both types of subscribers
to coexist within the same system at
the same time. An existing mobile network
can be expanded with fixed subscribers,
which will increase the use of the
installed infrastructure. And, conversely,
a cellular system providing fixed service
can be supplemented with mobile subscribers.
In an existing mobile network, the
opportunity to serve fixed users is very
attractive since the peak usage hours for
mobile subscribers are different from
those of fixed subscribers. Mobile subscribers'usage
tends to peak in the morning
and early evening (commuting time),
while fixed usage tends to be higher during
the day (for business users) and in the
evening (for residential users). By combining
fixed and mobile terminals in a
"mixed cellular" network, the operator
can achieve optimal utilisation of his network.
If so desired, mobile-originated
calls, which generate higher revenues,
can be given priority over fixed cellular
calls at peak hours. In cases where fixed
rural subscriber service is provided, the
resulting increased coverage for mobile
subscribers will in turn make the mobile
service offering more attractive.
The cellular technology will also support
concepts which fall somewhere between
pure fixed and mobile applications. The
degree of mobility can be set by system
parameters on an individual subscriber
basis, thereby providing restricted mobility
within just one or several cells in the
network. By means of different charging
schemes based on location, the operator
may create subscriptions with lower
charge rates in specified home cells and
a higher rate whenever the subscriber
initiates or receives a call outside this
area.
Value-added services
An extensive set of services will increase
the revenue for the operator. In addition
to the standard range of PSTN voice, data
and fax services, cellular-based systems
also support value-added services, such
as voice-mail, e-mail, fax, alphanumeric
messaging and IN services.
Capacity and coverage
An important feature of a cellular-based
solution is the high capacity and wide
coverage area that can be obtained. By
means of digital technology and features
such as hierarchical cell structures
(HCS), wideband frequency hopping, and
adaptive channel allocation (ACA), the
capacity will be virtually unlimited. The
fixed cellular concept will therefore not
only be an application for rural areas but
may also be a cost-effective solution for
urban areas.
Another feature is the flexibility of a eel-
Figure 3
A fixed, or mixed, cellular system has the capability
to provide telephony service to fixed subscribers
in private homes, apartment buildings and offices,
as well as to users requiring mobility.
Ericsson Review No. 1, 1996 7

Fig. 4
A fixed cellular system has the same components
as an ordinary cellular system except for the subscriber
unit, which allows the connection of ordinary
telephones.
lular system. A cellular network can grow
with demand, limiting the upfront investment
for the operator. Equipment can be
added to a base station at any time, in
order to serve a larger number of subscribers.
Adding more cells will increase
capacity, and in dense traffic areas,
implementation of microcells will provide
both extended capacity and improved
coverage.
In summary, the operator can respond
quickly to a demand for telecommunication,
with high voice quality and with a
competitive set of services. And, more
importantly, this can be done with great
flexibility and at a competitive price. The
ability to mix the subscriber base by providing
both fixed and mobile services within
the same network means that the operator
can address different market
segments and thereby obtain higher utilisation
of the installed infrastructure.
The D-AMPS standards
In North America, digital cellular systems
built according to the IS-54 D-AMPS standard
have been in operation since 1992.
The DAMPS standard has also been very
successful in other countries. By the end
of 1995, D-AMPS was serving more than
two million subscribers, in 17 countries.
In order to ensure continuous work on
evolutions of the IS-54 standard, a TDMA
forum was formed by North American operators
and vendors in 1993. Its first main
task was to specify a new control channel
for D-AMPS, and by the end of 1994, IS-
136 rev. 0 was published. Subsequently,
IS-136 was also specified for the new
1900 MHz frequency bands in the US. IS-
136 rev. A, which unifies the 800 and
1900 MHz standards into one common
standard, will be published in early 1996.
IS-136 rev. A also includes an
enhanced full-rate speech coder. In
speech quality evaluations performed by
independent research laboratories, this
new speech coder has proved to give
mean opinion scores (MOS) that equal
the MOS rating of the 32 kbit/s ADPCM
speech coder. (MOS is the method most
frequently used to evaluate the speech
quality of speech coders.) The 32 kbit/s
ADPCM speech coding is in general use
in wireline telephony and is also the
speech coder used in typical cordless
systems, e.g. DECT, CT2, PHS.
Through the open inter-switch protocol,
IS-41, D-AMPS supports a multi-vendor
environment in which systems from different
vendors can be interconnected.
IS-41 also allows international roaming
between different AMPS/D-AMPS operators.
The D-AMPS fixed cellular
systems
Ericsson has implemented D-AMPS in
their CMS 8800 system. Since afixed cellular
system is based on the ordinary cellular
concept, the architectures of the
CMS 8800 fixed cellular and the CMS
8800 mobile cellular systems are identical.
Therefore, only a brief overview of the
architecture will be shown here.
Switching system
MSC
The mobile switching centre (MSC) is the
heart of the CMS 8800 system. It performs
the telephony switching functions
for the network. It handles calls to and
from other telephone and data communication
systems, such as public switched
telephone networks (PSTN), integrated
services digital networks (ISDN), public
land mobile networks (PLMN), public data
networks and various private networks.
HLR/SCP
The home location register (HLR) contains
data about all subscribers, their services
and location. In large networks with
high subscriber density, the HLRs are
separate nodes. In small networks, the
8 Ericsson Review No. 1, 1996

HLR functionality can be integrated into
the MSC. The HLR can also be complemented
by a service control point (SCP),
providing IN services.
VLR
The visitor location register (VLR) database
contains all temporary subscriber
information necessary for the MSC to
serve visiting subscribers. No VLR is
needed in a pure fixed cellular system.
AC
The authentication centre (AC) provides
authentication and encryption parameters
that verify the user's identity and
ensure the confidentiality of each call.
Besides protecting against eavesdropping
of conversation, this functionality
also protects network operators from different
types of fraud found in the cellular
industry today.
MC
The message centre (MC) supports
numerous types of messaging service; for
example, voice mail, fax mail and E-mail.
Base station
The base station (BS) contains the radio
equipment needed for radio communication
with the subscriber units in a cell. In
order to increase capacity, a cell can be
divided into several sectors. Ericsson can
offer a wide range of different base stations,
from small, low-power pico base
stations (300x250x100 mm) to highpower
macro base stations, aimed to
cater for different scenarios.
OSS
The cellular operation and support
system (OSS) consists of a number of
tools that provide major support for
system management, operation and engineering.
The user-friendly interface offers
easy on-line fault handling, consistency
checks and network configuration
improvements. The OSS is of modular
design and uses an open non-proprietary
computer environment, enabling flexible
configurations according to customerspecific
needs.
Subscriber units
Ericsson has developed two different terminals
- a single-line and a multi-line terminal
- in order to allow ordinary PSTN
equipment to be used in a fixed cellular
system. The subscriber units, or terminals,
basically provide the radio connection
to the base stations and a socket for
the telephone set.
Single-line terminal
The single-line terminal (SLT) is the equivalent
of one PSTN telephone line. Up to
five telephones, all sharing the same telephone
number, can be connected to one
SLT - a small unit which can either be
placed on a table or mounted on the wall.
Fax machines and data modems can also
be connected to the SLT. It is powered by
a standard AC/DC adapter connected to
an ordinary wall socket. It can also be
equipped with an optional battery power
backup unit, which will supply power in
case of a mains failure.
The main parts of the single-line terminal
are:
- Radio unit: The radio unit consists of a
transmitter, a receiver and a duplex filter.
The duplex filter makes is possible
to use a single antenna for simultaneous
transmission and reception. The
radio unit supports the dual-mode
AMPS/D-AMPS standard.
- Control unit: The control unit decodes
and manages information transmitted
over the air interface as well as information
from and to the line interface
unit.
- Line interface unit: The line interface
unit provides a two-wire interface to an
ordinary PSTN telephone set. The line
interface can be programmed to emulate
different national PSTN standards
for ringing tones, dialling tones, line
voltages, etc.
Multi-line terminal
With the multi-line terminal (MLT), the
radio transmission equipment on the subscriber
side can be used more efficiently.
This will allow a more cost-effective
solution in cases where groups of sub-
Fig. 5
The single-line terminal is the equivalent of one PSTN telephone line. It serves as an Interworking unit
between an ordinary telephone/fax/ modem and the cellular system.
Ericsson Review No. 1, 1996 9

Figure 6
Siingle-line terminal - an attractive and simple solution for single subscribers. The terminal can be mounted
on an interior wall or may be placed next to a standard telephone. It can be equipped with an external
antenna to extend coverage to remote portions of the cell.
Fig. 7
The multi-line terminal allows up to 95 subscribers
to share 16 radio channels, each subscriber having
an individual telephone number.
scribers are located close to each other.
Up to 95 users can be trunked together
onto 16 radio units in the MLT. Each of
these 95 users will have a unique identity
in the fixed cellular network and be
treated as an individual subscriber.
Two possible applications of this type
may be mentioned: Using the multi-line
terminal for connection of the different
telephones in a block of flats to one common
MLT; or, as a rural application, connecting
all the telephones in a small village
to one or several MLTs.
The main parts of the multi-line terminal
are:
- RF unit
Handles the antenna and the antenna
interface.
- Terminal unit
Consists of the radio units needed for
communication with the cellular
system. A maximum of 16 radio units
can be connected to the terminal unit.
- Control unit
Contains all logic for signalling between
the different units in the MLT. It also
handles tasks such as radio channel
allocation, A-number transfer, etc.
- Subscriber switch
Contains the interface to the subscriber
lines.
Various types of antenna can be used for
both types of terminal, depending on signal
strength requirements, practical
installation considerations, and aesthetics.
The simplest choice is an antenna
mounted directly on the terminal but, as
an alternative, a small, omnidirectional
antenna can be placed almost anywhere
indoors, preferably close to a window. In
rural areas, an outdoor omnidirectional or
directional antenna can be used to provide
the desired coverage.
Pay phones
A variety of pay phones adapted for cellular
networks are available on the market.
The most common solution for tariff calculations
is to implement - in the pay
phone -tariff tables that can be updated
remotely via modem connections.
Standard cellular terminals
In addition to the SLT and the MLT, any
cellular phone complying with the
IS-54/IS-136 standards will work in a
D-AMPS fixed cellular system. These
phones can be used to offer subscribers
cordless service, but with a wider coverage
than that of an ordinary cordless telephone.
In the CMS 8800 fixed cellular
system, geographically-based charging
can be offered; for example, a low-charge
cordless service for calls initiated in the
cells that cover the subscriber's house,
whereas a medium or high charge would
apply in other cells.
Software functionality for fixed-cellular
services
Fixed-cellular functionality is based on
the concept of subscription areas. A subscriber
to a fixed-cellular service will only
have access to it if he is located within
his own subscription area (which may
consist of one or more cells) and, consequently,
cannot make or receive calls
in cells that are not part of this area. The
operator can also activate a function that
releases a call in progress if the subscriber
moves outside his subscription
area.
Predictions of radio wave propagation
10 Ericsson Review No. 1, 1996

are always subject to a certain degree of
uncertainty. The propagation conditions
may also vary from time to time due to
moving objects, for example. Therefore,
at the boundary between two cells, it can
be difficult to determine which of the cells
is part of a subscriber's subscription
area. This difficulty can be overcome in
two ways. One way is to assign more than
one cell to each subscriber in a boundary
area, i.e. those cells that have similar signal
strength.
The other way of solving this problem
is to use a function called redirected
access. If a terminal accesses a cell that
is outside its subscription area, the
mobile system will check whether any
neighbouring cell belongs within it. If so,
the system will orderthe terminal to make
a directed retry to that cell.
From a system point of view, all the different
concepts of charging areas, subscription
areas and private systems can
be considered as geographically dependent
service offerings. In the Ericsson
CMS 8800 system, all these concepts
have been merged into a functionality
called location-based services. This functionality
allows individual settings of a
subscriber profile, containing system
access rights (e.g. the possibility of making
and receiving calls) and differentiated
charging, depending on in which cell a
user makes his call.
For example, with location-based services,
a fixed-cellular subscriber can be
treated as any other subscriber except
that he will only have access to full telephone
service inside his own subscription
area. Outside this area he can still have
access to the short-message service and
have the possibility of making calls, but
at a higher rate. The system enables an
operator to define a wide range of profiles
of location-based service accesses and
tailor his service offerings to his customers'
needs.
system. Atwofold increase of capacity will
be achieved by applying adaptive channel
allocation.
Coverage
The maximum distance between the base
station and a terminal is dependent on
many factors, such as terrain profile,
antenna gain, cable loss, etc. The cell
radius must therefore be calculated for
each specific deployment. If the basic
Okumura-Hata model is used to estimate
some typical coverage scenarios, it
should be noted that this model is valid
for distances up to about 20 km. In the
table below it has been modified to cover
larger distances too.
In today's cellular systems, most people
use hand-held terminals. Table 2
shows that cells with more than 10 km
radius can be obtained in these cases.
But it also shows that if the system is
designed only for fixed services, the
coverage can be quite significantly
increased. In favourable cases, cells with
up to 90 km radius can be achieved if the
fixed cellular terminals are equipped with
directional terminals mounted a few
metres above ground.
A terminal output power of 0.6 W has
been assumed. It is also assumed that
Table 1
Spectrum allocation (each way) 5 MHz 10 MHz 20 MHz
Cells/site 3 3 3
Frequency reuse 7/21 7/21 7/21
Number of frequencies/site 20.8 44.6 92.2
Number of traffic channels/site 62 134 276
Erlangs/site 38 100 230
Number of subscribers/site 760 2000 4600
Capacity
The D-AMPS system offers high capacity.
An operator can serve more than 700 subscribers
per site with spectrum allocations
as small as two bands of 5 MHz (5
MHz for each link). Further examples are
given in Table 1, where 1% blocking and
50 mErlangs/subscriber has been assumed.
Capacity enhancement techniques for
the D-AMPS cellular system may also be
applied in the D-AMPS fixed cellular
Table 2
Environment Small city Small city Suburban Suburban Open area
BS antenna height (m) 30 50 50 50 100....
BS antenna gain (dBi) 20 20 20 20 20
Terminal antenna height (m) 1.5 1.5 1.5 5.0 5.0
Terminal antenna gain (dBi) 0 0 0 6 6
Max cell radius (km) 7 9 18 36 >50
Ericsson Review No. 1, 1996 11

Fig 8.
Multi-line terminal - suitable for hotels, office buildings,
apartments complexes, housing estates, and
small villages. It connects up to 95 subscribers
through 16 radio channels.
12
the uplink is the limiting link due to lower
power in the terminals than in the base
stations.
Hierarchical cell structures (HCS)
It has been found that in high-traffic downtown
areas, the most cost-effective way
to provide high-capacity systems is
through hierarchical cell structures. A
microcell layer employed as a complement
to the existing layer of macrocells
can add considerable capacity to a
system. Since the overlaying macrocell
will also serve as an alternative in case
of blocking in a microcell, the microcells
can be designed for high channel utilisation
without causing high blocking in the
system. This allows more effective use of
installed equipment.
The concept of hierarchical cell structures
can also be extended to areas outside
a city. Small villages needing high
capacity may be interspersed with
sparsely populated areas in between. In
such a scenario, the hierarchical cell lay-
out is essential for a system to provide
both capacity and coverage in a costeffective
manner. High-rise towers are
installed first, in order to create the
initial wide areas of coverage (macro-
cells). Next, the higher capacity needed
in small villages is provided by installing
smaller, lower-cost cells (microcells) in
a layer below the macrocell. Subse-
quently, if some particular area is found
to have insufficient coverage, this prob-
lem can be solved by installing small
microcells.
The HCS building practice will ensure
cost-effective deployment in terms of
coverage and capacity,
Services
Since the D-AMPS fixed cellular system is
based on the ordinary cellular concept,
there is immediate access to all function-
ality provided by the cellular D-AMPS
system. In addition to the basic telepho-
Ericsson Review No. 1, 1996

ny services - such as voice, fax and data
- a variety of more advanced services is
available, Box B.
Fixed cellular in the
marketplace
Ericsson's largest contract for a fixed cellular
network is for Malaysia. This order
was received from a newly-licensed operator,
Syrikat Telefon Wireless Sdn. Bhd.,
in September 1994. The system in question
uses Ericsson's CMS 8800 infrastructure
operating with the AMPS/D­
AMPS standard. The first installations,
with a capacity of 5,000 subscribers in
the fast-developing resort island of Lankawi,
were in place at the end of 1994,
and capacity for 30,000-40,000 subscribers
was available at year-end of 1995.
Cellular-based systems -
a look ahead
In some operating environments, it is difficult
to differentiate clearly between fixed
and mobile telephony. In offices, the use
of business cordless systems-e.g. DECT
- is becoming increasingly popular. These
systems offer wireless access and mobility
within the office premises. And by connecting
a private cordless base station to
the PSTN telephone jack, a user can benefit
from the convenience of having wireless
access in his home. Both these applications
are associated with fixed
telephony providers.
For a cellular operator, one obvious
solution is to use subscription areas and
charging areas to create a wireless office
system. The employees of a company can
use cellular telephones within the office
premises at a call charge that differs from
the normal cellular tariffs (in the office,
unlimited air time for a fixed monthly fee
could apply, for example). Together with
private numbering plans and other IN services,
a very competitive wireless office
concept will be achieved. Even when an
employee leaves his office, he will have
access to his wireless office services, but
then at ordinary cellular tariffs. The possibility
of using the terminal outside the
office can be provided on a per subscriber
basis. In fact, wireless office systems
based on D-AMPS cellular technology are
already in use.
With the introduction of additional competition
to the cellular services such as
PCS in the United States and PCN in
Europe, we find that the boundaries
between the fixed and the cellular environments
are becoming increasingly
blurred. Many PCS/PCN operators are
already targeting the residential market
segment using cellular technologies,
complementing the wide-area mobility
service with a low-tariff service for usage
of the mobile phone in the "home cell",
and ordinary fixed-telephony operators
will have a limited capability to compete
in such situations. But if an existing operator
has offered fixed services by using a
fixed cellular system, he will be prepared
to compete on the mobile market as well,
because he will be able to offer both
mobility and personalised services to his
subscribers.
Conclusions
Using wireless technologies offers a number
of benefits to the network operator.
Wireless systems are reliable, cost-effective
and highly flexible; they can be
deployed to provide service in every type
of environment, from dense urban to
remote rural areas.
One of the greatest advantages of the
fixed cellular system described - a complete
cellular system including MSCs,
radio base stations, IN nodes, etc - is
its ability to offer both fixed and mobile
services in the same network. The
system enables the operator to define a
wide range of profiles of what is called
location-based service accesses. He
can tailor his service offerings to his customers'
needs and address different
market segments, and thereby obtain
optimal utilisation of the installed infrastructure.
References
1 Jager, J.: Using DECT for Radio in
the Local Loop. Ericsson Review
No. 3, 1995 pp 111-117. .17.
2 Dr. Tarjanne, P, Secretary-General,
ITU: A new era in telecommunications
and information.
Ericsson Review No. 1. 1996 13

New power distribution
and earthing for AXE
Mats Barkell and Per Haging
The constant trend towards increased miniaturisation of electronic equipment
and improved reliability has resulted in a continuous evolution of the
hardware structure and packaging of AXE equipment. The methods used for
distribution of equipment power have also undergone steady development
in order to accommodate the changing requirements of the AXE system.
The authors describe Ericsson's new two-step high-ohmic power distribution
system (TS-HOD) and the advantages of using it in public switches,
mobile telephony base stations and PBXs.
Box A Abbreviations
CSP Central system power
DC-C Common DC return
DC-I Isolated DC return
ETSI European Telecommunications
Standards Institute
HOD High-ohmic distribution
IEC International Electrotechnical Commission
The new two-step high-ohmic power distribution
system (TS-HOD) developed by
Ericsson optimises power distribution for
small-size equipment and small installations.
It also offers generalised power
interfaces to all sizes of physical units -
at the cabinet, subrack and board levels.
The TS-HOD lends itself to use in public
switches, mobile telephony base stations
and PABXs, in applications ranging in size
from telecom centres down to small
remote access installations.
System voltage
The system voltage is -48V DC for all
types of equipment, except the smallest
types of remote equipment. The use of
the TS-HOD in old installations with a different
battery voltage will require adapters.
The reasons for using system voltages
other than -48V have been either national
demands or technical optimisation of
the equipment. However, national
demands for system voltages other than
-48V will become extremely rare once
ETSI's prETS 300 132 a is given its status
as standard within the European
Union.
In radio base stations, +24V has been
commonly used in order to optimise the
efficiency and cost-effectiveness of the
transmitteroutput stages. But, since optimum
performance of this type of equip-
ISP Integrated system power
MESH-BN Meshed bonding network
MESH-IBN Meshed isolated bonding network
PBA Printed board assembly
PIU Plug-in unit
SELV Safety extra tow voltage
TS-HOD Two-step high-ohmic distribution
UPS Uninterruptible power supply
ment requires DC/DC converters as it is,
there are options open to the users as
regards system voltage. Some markets
have already changed to -48V; hopefully,
others will follow.
In view of what has just been said, the
dominant role of-48V as system voltage
for telecom equipment will become even
stronger than it is today.
The recent increase of the safety extra
low voltage (SELV) limit from 42 to 60V
in some countries will also allow the use
of -48V system voltage in other types of
non-telecom, battery backed-up equipment.
This development further supports
the use of-48V in telecom applications,
since some of the units are common to
both types of equipment.
Two-step high-ohmic
distribution (TS-HOD)
System reliability structure
From the point of view of reliability, telecom
systems are built up as a number of
"disablement units", defined as "the
maximum amount of equipment that is
allowed to be affected by a single fault".
The size of a disablement unit is normally
expressed in terms of functionality,
such as the number of communication
lines concerned, but this criterion has
generally a direct correlation with parts of
the physical equipment. A reliability structure
designed in this way requires that
single faults should always be isolated so
as to affect only the disablement unit
within which they occur.
The power distribution for such a
system uses a branch-out (star-type)
topology, where each branch uses a separate
fuse and separate DC/DC converters
to supply the tertiary voltages (such
as +3.3V, +5V...). There are different
solutions (low-ohmic or high-ohmic distribution)
to the problem of possible transient
undervoltages before such a fuse
14 Ericsson Review No. 1, 1996

opens in case of short circuit or overload.
These solutions are described in the following.
The physical size of a disablement unit
is shrinking over time, for two reasons:
- increased reliability requirements,
which lead to a reduction of the number
of communication lines to and from
each disablement unit;
- the general trend towards increased
miniaturisation of electronic equipment.
As a result, the disablement unit will be
a plug-in unit (PIU), and in many applications
the type most frequently used: a single
printed board assembly (PBA). In order
to maintain the reliability structure, the
power distribution branches must be
made proportionally finer.
Fig. 1
General principle of high-ohmic distribution (HOD).
The figure shows that shortcircuits in one distribution
branch cannot affect the system voltage
thanks to resistance ratio Rc/ RD.
The principle of high-ohmic distribution
The general principle of a high-ohmic
power distribution system - regardless of
whether it is a single-step or two-step
system - is that of employing current limitation
in each branch-out point. Proper
design ensures that an overcurrent situation
in one of the distribution branches
will not affect the operation of equipment
powered from other distribution branches.
Such overcurrent situations could be
caused by an equipment failure or by the
connection of new loads.
The current limitation required is normally
achieved by series distribution
resistors. In previous versions of the AXE
power distribution system, these resistors
have been 50mQ. Even during a
short circuit in one of the distribution
branches, the voltage in all other distribution
branches will stay within its
-40.5 - -60V limits if a central battery of
approximately six times lower internal
resistance (c. 8m£2) is used, Fig. 1.
In Fig. 1, the central power is represented
by a battery, with an internal resistance
R c . R D represents the distribution
resistance mentioned, and F lt ..F n are distribution
fuses. R D and F^.-F,, have so far
been located in the fuse panels of the
normally centralised power plant. The
DC/DC converter in the uppermost distribution
branch represents one piece of
powered equipment, and V s is its supply
voltage. The switch, S, represents a short
circuit caused by a failure in another distribution
branch and occurring at time tj_.
During the fuse blowing time, t ± to t 2 , V s
for all other equipment stays within its
allowed range. The transient at t 2 , which
occurs when the fuse releases (due to the
inductance in the distribution branches),
has a short duration in a HOD system and
is compensated by small filters in the powered
equipment. The fuse blowing time is
normally in the range 5-20 ms.
Low-ohmic distribution
In a low-ohmic distribution system - i.e. a
system without R D - the supply voltage,
V s , would drop outside its allowed range
(normally below half its nominal value) during
the fuse blowing time. Blocking diodes
combined with large capacitors need to
be used in all equipment of the other
branches to provide energy to the load during
the period of undervoltage.
Since the power distribution branches
are getting more "fine-meshed", the
increased number of energy storage
capacitors of a low-ohmic distribution
system would be very space-consuming.
The reason for using high-ohmic distribution
has therefore become even more
obvious.
The evolution of HOD into TS-HOD
In two-step high-ohmic distribution, the
branch-out of power distribution takes
place in two separate steps. The purpose
of this arrangement is to allow the finer
distribution grid mentioned above.
As in earlier systems, the first branchout
occurs in the fuse panels. However,
the distribution resistance value has been
increased by a factor of three. This change
in resistance optimises the efficiency and
cost-effectiveness of power distribution in
Ericsson Review No. 1, 1996 15

Fig. 2
Two-step high ohmic distribution (TS-HOD) using
single (a) and double feed (b).
Subrack
Fuse panel
Power supply
smaller installations, since it allows the
proportionally higher internal resistance
that is characteristic of lower battery
capacities.
The general parameters of the fuse panels
are:
- distribution resistance 0.15A
- fuse 10A
- max. nominal load 5A
This means that each distribution
branch can supply up to 250W, and that
the worst-case short-circuit current transient
is limited to 350A.
The second branch-out occurs in the
subracks. Each PIU has a small (1Q)
fusible resistor in series with its-48V supply
from the backplane power "bus", serving
both as a current limiter and a fuse.
The physical location of this second-step
distribution resistor has two major advantages:
- the resistor limits any current transients
during "hot insertion" of PIUs;
- replacement will automatically occur
together with the faulty PIU.
Single or double feed of PIUs
The availability of the -48V supplied by
the subrack backplane to the PIUs is very
high. Due to conservative rules for insulation
distances in the backplane and the
use of double-insulated distribution cable,
the major contributor to the supply
voltage's failure intensity is the fuse in
the fuse panel. For this reason, single
feed as described so far is satisfactory
in many applications.
However, a PIU can be double-fed by
adding an extra fusible resistor and insulation
diodes to it. The additional feed
may be connected to the same fuse panel
or handled separately all the way back to
the central battery. If the central power
plant is duplicated, the entire power feed
can be separated. Fig. 2 shows a diagram
of single and duplicated power feed.
Distribution efficiency
One obvious objection against the use of
series resistors in a distribution system
is the effect of these resistors on power
efficiency. But, due to the low resistance
values used, this is not a serious drawback.
The following table shows overall
power efficiency values, as a function of
PIU power requirements, for a typical distribution
branch carrying 200W from the
fuse panel:
200W distribution branch:
- Forty 5W PIUs 98.6%
- Twenty 10W PIUs 98.4%
- Ten 20W PIUs 98.0%
- Four 50W PIUs 96.9%
The calculations assume a nominal
voltage of -50V from the battery, and
that the losses of the distribution resis-
16 Ericsson Review No. 1, 1996

Fig. 3
Four examples of TS-HOD installations of different
sizes. Interfaces "a" and "b" use common specifications
in all the examples.
Subrack
Fuse panel
Power supply
tors in both branch-out points are included.
For the typical high-volume PIUs,
which require 10W, the efficiency must
be considered acceptable. For 50W
PIUs, the efficiency starts to decrease,
but these latter represent a very small
portion of the system, which means that
the overall system power efficiency is
still high.
Generalised interfaces
The use of the TS-HOD in four typical
examples of installations is schematically
outlined in Fig. 3.
The first example, A, shows a small
remote unit. Power can be fed from the
mains, possibly via a UPS.
The second installation, B, is a central
office using integrated system power,
ISP, where the central power is located
in the same room as the system, possibly
within the same rack.
The third example, C, illustrates a traditional
installation, where the central
system power, CSP, is located in a separate
power room.
Example D, shows an installation
where the equipment is used together
with a low-ohmic central power plant and
where low-ohmic distribution is applied.
The interface denoted by "a" in these
four examples is the same for all installations.
This means that the same PIUs
can be used in a variety of applications.
As appears from the examples,
interface "b" is the same for all applications
powered with -48V. All specifications
of this interface are in conformity
with the preliminary ETSI specification
prETS 300 132 1 .
If an adaptation has to be made, a distribution
unit as indicated in example D
of Fig. 3 will be used. It is essentially a
fuse panel, combined with the adaptation
circuitry. If the adaptation is made for low-
Ericsson Review No. 1, 1996 17

ohmic central power, the adaptation circuitry
will consist of the diode and backup
capacitors that are required to bridge
transient undervoltages. This type of
adaptation is very space-saving and costeffective,
compared with a solution that
involes including the adaptation circuitry
in all powered units.
Other adaptations - such as other battery
voltages or earthing systems - may
be made for requirements specified in
non-ETSI standards.
Earthing
A new ETS standard 2 for the earthing of
telecommunications equipment was
issued in 1995. The new AXE earthing
system, based on the two major principles
of "multi-point" (MESH-BN) earthing
and "two-wire" (DC-C) power distribution,
conforms to this standard.
Multi-point earthing means that telecom
signal earth is deliberately connected
to the building and to mains protective
earth, in order to achieve one,
extremely stable earthing system. It follows
that the telecom earthing system
must be capable of handling high fault
currents. The new AXE earthing system
was verified using 3000A current transients
(according to IEC801-5) in order to
provide internal design rules for signal
connections.
Two-wire power distribution means that
the power return wire ("OV") is connected
to telecom signal earth. This solution
is not new; it was also used in previous
power distribution systems for AXE. The
parallelling of all power return wires and
signal earth simplifies the fulfilment of
the requirements for very stable multipoint
earthing.
Of course, new AXE equipment can also
be used with "single point" (MESH-IBN)
earthing, since the requirements for
equipment current handling capacity are
then lower. However, "three-wire" (DC-I)
distribution will require use of adapters.
Compatibility
As described in the foregoing, today's
AXE equipment will generate lower - but
accept higher - fault currents than would
previous versions. It also accepts higher
central power (battery) internal impedances.
From the point of view of power distribution
and earthing, the new AXE power
distribution systems can therefore be
used together with previous versions of
AXE equipment as well as most other
installations without extra measures having
to be taken.
Standardisation
The prETS 300 132 standard referred to
is preliminary (denoted by its "pr" prefix),
but it is unlikely that its main points will
be revised. Hopefully, it will reach the status
of standard during 1996. The ETS standards
(Ref. 1,2) are intended for members
of the European Union, but lack of
other standards will probably widen their
use to many other countries.
The new AXE power distribution system
complies fully with the prETS 300 132
standard.
Conclusions
The new two-step high-ohmic power distribution
system (TS-HOD) developed by
Ericsson offers many advantages in all
areas of telecommunication: public
switches, mobile telephony base stations
and PBXs - in applications ranging in size
from telephone centres down to small
remote installations.
One of the driving forces behind the
development of TS-HOD has been the constant
trend towards miniaturisation of
electronic equipment. In the TS-HOD, the
branch-out of power distribution takes
place in two separate steps, which allows
a fine and space-saving distribution grid.
The TS-HOD complies with the prETS
300 132 standard and is fully compatible
with previous AXE power distribution
systems as well as most other system
installations.
References:
1 ETSI standard prETS 300 132-2
September 1995: Power supply
interface at the input to telecommunications
equipment.
2 ETSI standard ETS 300 253, January
1995: Earthing and bonding of
telecommunication equipment in
telecommunication centres.
Ericsson Review No. 1, 1996

Telecool Aero - A tailor-made cooling system
for small telephone exchanges and containers
Dag Hellstenius
The spread of economic development around the world has been largely
accompanied by a corresponding spread of telecommunications networks
and infrastructure. Telecom installations have been deployed in increasing
numbers by an increasing number of suppliers. A common consideration in
all these installations is the need for air-conditioning or some other means
of dealing with the heat generated by the electric and electronic equipment.
The ideal cooling system in a small installation is one that is not only effective
but also reliable and robust in a wide range of circumstances. In this
article, the author describes such a system - Telecool Aero.
© 1995 IEEE, Reprinted, with permission, from
Proceedings of tntelec' 95, The Hague,
The Netherlands. October 29-November 1,1995,
pp 394-396.
The idea of developing a small, compact
cooling unit for use in containers and
small buildings was conceived in February
1993 at Ericsson Components. The
concept was based on a new technology,
known as cooling with displacement air
flow. Basic performance criteria demanded
that the cooling unit should be tailormade
to meet the strict indoor-climate
requirements of telecommunications
equipment, possess high reliability, and
be able to cool the equipment even in the
case of a mains outage. The result was
the Telecool Aero, a series of products
whose characteristics are described in
the following.
Convection and displacement
air flow
The Telecool Aero functions according to
the principle of cooling with displacement
air flow. This involves the combination of
two phenomena: natural convection and
displacement flow.
Transfer of heat by natural convection
occurs when a temperature difference is
present in a fluid, which may be in the
form of either a gas or a liquid. Fluids
expand when they are heated, and this
results in a lowering of their density.
Water is the only exception: it exhibits its
highest density at +4°C.
In a gravitational field, such as we have
here on earth, fluids with a lower density
rise upwards and initiate movement in the
fluid. This movement - known as convection
- permits the fluid to absorb and give
up heat. When a movement in a fluid is
caused only by differences in density, the
process is referred to as natural convection;
when the movement is caused by
pumps, fans or the like, it is referred to
as forced convection.
Fig. 1
A pre series 4.5 kW Telecool Aero. After the photo was taken, the evaporator grid has ben reinforced to
improve shock resistance and allow increased air flow.
Ericsson Review No. 1, 1996 19

Fig. 2
The present design of an 8.5 kW Telecool Aero.
20
A fluid exhibits different densities at
different temperatures. If two quantities
of the same fluid are allowed to distribute
themselves in a space under the
effect of gravity, the quantity with the
higher density - and thus at the lower
temperature -will occupy the lower level.
This law of physics causes the colder and
heavier fluid to distribute itself in a space
by "flowing out", just as water runs
across a floor. When gases exhibit the
same behaviour, the phenomenon is
referred to as displacement flow, and
when this process is combined with natural
convection, we obtain cooling with
displacement air flow.
Principles of operation
For the sake of simplicity, the following
description deals with the cooling of electronic
equipment in a container. The
same principle applies to rooms in a building,
of course.
In the case of Telecool Aero, the cooling
process based on displacement flow
is as follows:
Hot air, which occupies a space immediately
under the roof of the container, is
sucked into the cooling unit by fans. When
cooled, the air is distributed overthe floor
at a low speed - approximately 1.2 m/s.
It is then sucked into the cabinets
(through the cabinet door, which is perforated
for this purpose) where it cools the
electronic components. The air takes up
heat from the components, assumes a
lower density and, consequently, rises
towards the roof according to the law of
natural convection. This completes the
cooling cycle; another cycle starts when
the cooling air is again sucked into the
cooling unit.
In order to reduce wear on the electronic
components, and thus the number
of potential sources of faults in the
telecommunications equipment, it is
essential that the temperature should
be low, and stable. The requirement for
stable temperature was met by designing
and testing a hot-gas bypass valve
(serving as a capacity control) to guarantee
that the variation in temperature
is less than 0.5°C/minute. In order to
provide an optimal climate for the components,
the air temperature should not
deviate significantly from 20°C. Different
components exhibit different sensitivity
to temperature, of course, and the
failure intensity also increases with rising
temperature. This increase - which
is dramatic for some components and
less pronounced for others - will inevitably
have repercussions for the operator,
in terms of higher maintenance
costs and shorter service life of the
equipment.
Description of the refrigerant
circuit
The refrigerant circuit in the Telecool Aero
is a tailor-made adaptation of the familiar
compressor-driven refrigerant circuit.
Its principal function involves absorbing
heat in one space and transferring this
heat to another space, where it is given
Ericsson Review No. 1, 1996

up to the surroundings. This functions in
the following way:
Refrigerant in liquid form that has been
compressed to a high pressure (in this
case approximately 12 bar) is allowed to
expand to a lower pressure, by means of
a valve. A drop in pressure is accompanied
by a drop in the temperature of the
refrigerant, which starts boiling and, in
this process, absorbs energy in the form
of heat from its surroundings.
The refrigerant, which passes into gaseous
form when it is boiling, is then
sucked into a compressor and compressed
into a highly-pressurised gas.
The temperature increases as the gas is
compressed and, when sufficiently high,
makes the gas give up heat to the surroundings.
As this takes place, the refrigerant
condenses to form a highly-pressurised
liquid, which means that it reverts to
the starting point. The system is now
ready to expand the refrigerant once more
and so perform another cooling cycle.
The special adaptation to telecom
applications involves accurate capacity
control and the technique of utilising the
principle of displacement air flow.
Description of the cooling
unit at component level
The number listed with each of the cooling
unit components refers to the system
diagram, Fig. 4.
Compressor 1
The compressor is of the hermeticallysealed
type, without lubrication points. It
is equipped with an oil heater and a regulator
for high- and low-pressure to protect
the system and its surroundings. All
moving parts in the compressor are
mounted on vibration dampers, and the
compressor itself is suspended on rubber
feet to eliminate the effect of vibration.
The fact that the compressor is hermetically
sealed means that the motor in
this case is cooled by the refrigerant gas
as it is sucked in from the evaporator.
Filter drier 121
A filter drier is installed in the fluid line to
ensure that the system is free from moisture.
Thermostatic expansion valve 131
The purpose of the expansion valve, a
high-precision component, is to measure
very accurately the right amount of refrig-
Fig. 3
Backup unit. The upper part contains the inverter
that is used in the case of a mains outage, when
the cooling system is battery-operated.
Fig. 4
Schematic diagram of the refrigerant circuit in
Telecool Aero.
Compressor 1
Filter drier 2
Thermostatic expansion valve 3
Suction gas cooling unit 4
Evaporator 5
Condenser 6
By-pass regulator valve 7
Ericsson Review No. 1, 1996 21

Fig. 5
Graphs showing how the backup unit of the Telecool
Aero cools the air during a mains outage.
Without cooling, the air temperature will reach
55°C within 15 minutes, which means that a safety
thermostat will switch off all equipment. If no provision
for cooling backup time has been made, the
exchange will have no backup.
^__-
—^—
^ _ _
Air temperature to unit
Water temperature to cooler
Cooling capacity
erant to the evaporator. The valve is controlled
partly by the pressure in the evaporator,
and partly by a bulb which senses
the temperature of the refrigerant. This
bulb is mounted on the outlet pipe after
the evaporator.
Suction gas cooling unit 14 '
This unit, too, is in the form of an expansion
valve. It cools the refrigerant gas,
under sometimes extremely arduous conditions,
before the gas is sucked into the
compressor. This cooling is necessary;
for example, when a cooling system is
started aftera mains supply failure intropical
countries, where the air temperature
in the telecommunications building can
be very high.
Evaporator* 51
The evaporator is the component that
cools the air. It consists of tubes with soldered-on
fins through which the refrigerant
flows, so as to transfer heat from the
air to the refrigerant. The refrigerant boils
inside the tubes of the evaporator, at a
temperature below that of the surroundings,
and "sucks up" heat from the air
that is blown through the outside of the
evaporator.
Condenser 16 '
The condenser functions as an evaporator,
but in reverse. The refrigerant gas,
which has also become very hot after compression,
cools down in the condenser
and changes into liquid form. When the
refrigerant condenses, it gives up heat to
the tube walls and fins of the condenser.
These parts are cooled by air that is
sucked through the condenser by fans.
Bypass regulator valve' 7 '
It is the hot-gas bypass valve that controls
the capacity of the cooling equipment
and ensures that the air retains a
uniform temperature, regardless of variations
in outdoor temperature or thermal
load. The capacity control function utilises
an artificial thermal load in the form
of hot refrigerant which the bypass valve
allows to pass directly into the evaporator.
Very high reliability
In order to produce a cooling unit with
exceptionally high reliability, the most
sensitive components were duplicated
and the design was tested with great
care. The unit was designed so that two
fans are always operating, both on the
evaporator side and on the condenser
side. This means that one condenser fan
and one evaporator fan can fail without
causing the cooling function to cease
completely. If one condenser fan or evaporator
fan fails, a damper closes in order
to prevent air circulation in the unit; the
result will be a reduction of the cooling
effect by about 30%. If both an evaporator
fan and a condenser fan fail, the cooling
effect will be reduced by about 50%.
The unit has two alarm outputs. Alarm
is issued at high temperatures and in the
event of a component failure.
Cooling in the event of
mains supply failure
For the system to be able to provide cooling
in spite of a voltage drop, a solution
involving the "storage of cold" in a backup
unit was developed. This function is
achieved by connecting the cooling unit
such that it not only cools its surrounding
air but also a liquid stored in a water
tank. The liquid may be water, of course,
although it is often a solution consisting
of 70% water and 30% glycol, which will
not freeze before the temperature drops
below -30°C. The reason for using a glycol
solution is to prevent the tank from
freezing and bursting if the temperature
falls below zero during a lengthy loss of
voltage in the winter.
Mains supply failure tests
In the event of a mains supply failure, batteries
supply an inverter which, in turn,
supplies the evaporator fans so that the
22 Ericsson Review No. 1,1996

Fig. 6
Displacement air flow - cold air "flowing out"
across the floor of the container. The air is then
sucked into the racks and heated by the electronic
components; caused to rise towards the roof
and again sucked into the cooling unit.
air flow remains unchanged. The inverter
also supplies a pump, which then begins
to circulate the cooled liquid through the
evaporator coils. In this way, the cold
water or glycol-water solution cools the air
for a certain length of time. The cooling
effect is at its greatest at the start of the
backup period and then slowly decreases,
Fig. 5.
The test of the backup unit was designed
so as to maintain the temperature of the
supply air at the normal value of 35°C. The
glycol-water solution was cooled to a constant
value by the cooling unit. This means
that the unit was working just above its
nominal capacity, since the condensation
temperature was around 50°C.
The mains voltage to the cooling unit
was then disconnected, at the same time
as the thermal load from the cabinets was
maintained with its power unchanged.
Temperature-rise tests
Test of Telecool Aero with backup unit
were performed, to determine the cooling
capacity and the rise in temperature of
the air supplied to the cooling unit. The
Fig. 7
Sectional drawing of a container. The air flow in
normal operation (middle) and cooling with a backup
unit during a mains outage (bottom).
Ericsson Review No. 1. 1996 23

esults of these tests - shown in Box A -
confirm that a backup period of one and
a half hours can be guaranteed for applications
with a 8.5 kW unit connected to
one backup unit, and a backup period of
three hours for applications with a 4.5 kW
unit so connected.
Climatic and manual tests
In order to satisfy ourselves of its reliability,
the Telecool Aero was tested according
to IEC 68-2 environmental specifications.
The cooling unit was tested with
respect to both climatic and mechanical
requirements.
The climatic tests were as follows:
- IEC 68-2-1 cold -30°C, 100 hours
- IEC 68-2-2 dry heat +60°C, 100 hours
- IEC 68-2-30 moist heat +25°C/+40°C,
six cycles of 24 hours each at a relative
humidity of between 90% and 100%.
The mechanical tests were as follows:
- IEC 68-2-6 vibration 0.5 g in the frequency
band 10-150 Hz with the unit
operating
- IEC 68-2-6 vibration 2.0 g in the frequency
band 10-150 Hz with the unit
not operating
- IEC 68-2-27 shock test, three shocks
in all directions, 15 g, with the unit not
operating
- IEC 68-2-57 seismic test 0.3 g in the
frequency band 1-40 Hz, with the unit
operating
The department also carried out performance-verifying
tests, such as:
- Performance at -20°C, ±0°C, +20°C,
+40°C and +55°C.
Starting tests were also carried out.
The voltage applied varied from 198 to
253 V, and the frequency from 45 to 55
Hz. These values were changed in such
a way that all combinations were tested.
The system performed satisfactorily in
all cases.
Conclusion
Through a stringent series of tests and
field applications, the Telecool Aero unit
has proved itself notonlycapable but also
very suitable for cooling and air-conditioning
of small telephone exchanges and
containers.
With its robust design and use of few
moving parts, Telecool Aero is reliable
and effective in a wide range of conditions,
including periods of mains supply
failure.
Box A
Testing with Telecool Aero 4.5/8.5 kW and backup unit
To protect the equipment from excessively high temperatures, it is switched oft automatically when the air
temperature in the container (or room) exceeds 55°C.
Testing with Telecool Aero 8.5 kW and backup unit
Start
After In 30mln
Temperature of supply air 35°C 54°C
Cooling effect 5.0 kW 4.0 kW
Temperature of supply water 17°C 40°C
Testing with Telecool Aero 4.5 kW and backup unit
Start
After 3 h
Temperature of supply air 37°C 48°C
Cooling effect 2.1 kW 1.3 kW
Temperature of supply water 20°C 38°C
24
Ericsson Review No. 1, 1996

Ericsson echo cancellers - a key
to improved speech quality
Anders Eriksson, Gunnar Eriksson, Johnny Karlsen, Anders Roxstrom and Teresa Vallon Hulth
Speech quality is becoming a competitive factor in today's telephony systems.
A number of connections have an inherent transmission delay that
makes echo control necessary. Echo cancelling, which is the modern way
of handling the echo problem, is used extensively both in long-distance terrestrial
or satellite networks and in digital-cellular-to-PSTN circuits.
The design of the echo canceller plays a vital role in the overall speech
quality of telephony systems. Ericsson, as a system supplier, has responded
to this fact by designing the ECP series of system-integrated, high
speech quality echo cancellers for all AXE 10 network applications including
the digital cellular standards GSM, PDC and D-AMPS.
The authors describe the causes of echo, the design and evaluation of
Ericsson's echo cancellers and the advantages of the highly integrative
pool-configuration concept that has been introduced.
4-wire
2-wire
transmission Hybrid subscriber
network
line
Fig. 1
Echo originating from the PSTN subscriber line
Interface.
Echo in telephony
systems
The main cause of echo in telephony circuits
is the imperfect impedance matching
in networks with both 2-wire and
4-wire transmission sections. The normal
location of the interface between these
sections is in the access part of the public
switched telephone network (PSTN). In
this interface, the hybrid is responsible
forthe conversion between the 4-wire and
2-wire sections. Impedance mismatches
cause some of the incoming speech energy
from the 4-wire section to be reflected
back to the talker as a distorted and
delayed replica of the incoming speech
from the far end, Fig. 1.
As long as the delay of the transmission
path is short, the reflected speech
energy is not a problem; it will then only
be added to the normal talker sidetone
present in all handsets. But when the
delay is increased and the reflected
speech energy becomes separated from
the sidetone, it will be perceived as an
echo, so-called talker echo.
Another cause of echo is acoustic
crosstalk, i.e. crosstalk between the loudspeaker
and the microphone in a handset
or in a loudspeaking telephone set.
Delay is introduced in the transmission
path as a consequence of propagation
time over long distances and/or the coding
of the transmitted signals. The most
obvious cause of delay is the use of satellites
for intercontinental calls: a geostationary
satellite causes a one-way transmission
delay of 260 ms, and the total
echo path delay will consequently exceed
520 ms. Another telecom application suffering
from delay is digital cellular systems,
where it is caused by the speech
and channel coding necessary for radio
transmission. This coding results in a
one-way transmission delay of typically
100 ms when blocks of speech samples
are transmitted over the air interface. The
delay of both these applications is well
above the value - 25 ms - at which
ITU-T G. 131 recommends the use of echo
control devices.
The modern way of handling the phenomenon
of echo is to employ echo cancellers.
In digital cellular applications,
these devices are located in a mobile
Abbreviations and definitions
ADPCM Adaptive differential pulse code EPD Echo path delay, the delay of the
modulation
echo
APZ The control system of AXE 10 Far end Definition used in echo cancelling
Comfort noise Noise inserted in the signal path in theory: The end of the connection
order to minimise the audible
where the echo is heard (see Fig.
effect of the NLP, when the echo 5)
canceller is used in a connection G.165 ITU-T Recommendation on echo
with background noise
cancellers
Dispersion Frequency-dependent delay caus- ITU-T International Telecommunication
ing a spread in the impulse
Union - Telecommunication Stanresponse
dardization Section
DP Microcontroller (device processor) Near end The end of the connection where
DSP Digital signal processor echo reflection takes place (see
EC Echo canceller Fig. 5)
ECAN3 Echo canceller board in the NLP Non-linear processor, used to sup-
ECP303, performs the necessary
press residual echo
signal processing O&M Operation and maintenance
ECON3 Echo control and interface board in PDC Personal Digital Cellular; Japanese
the ECP303
standard
ECP101 Echo Canceller in Pool, model 101 PSTN Public switched telephone network
ECP303 Echo Canceller in Pool, model 303 TERL Total echo return loss, the total
EM Extension module. In AXE 10, the attenuation of echo in a connecequipment
controlled by a pair of
tion; measured from the talker's
regional processors is arranged in
mouth, over the transmission path
groups called EMs.
to the echo-reflection point and
ERL Echo return loss, the attenuation back to the talker's ear
in the echo path as seen from the TSS Trunk and signalling subsystem,
echo canceller one of the subsystems of AXE 10
Ericsson Review No. 1, 1996 25

Fig. 2
Location of echo canceller in a digital cellular application.
MSC
EC
PSTN
LE
H
Mobile switching center
Echo canceller
Public switched telephony network
Local exchange
Hybrid
Fig. 3
Location of echo canceller in a long-distance network
application.
ISC
EC
LE
H
International switching center
Echo canceller
Local exchange
Hybrid
switching centre (MSC), Fig. 2. In long-distance
telephony circuits, they are usually
located in an international switching
centre (ISC), Fig. 3.
Echo canceller requirements
The requirements of an echo canceller
are defined by three different transmission
characteristics.
- One-way transmission delay between
the two parties engaged in conversation
The one-way transmission delay determines
the necessary total echo return
loss (TERL) for the connection. This loss
is defined as the total level loss between
the talker's mouth and his ear. Recommended
minimum values for TERL are
specified in ITU-T's Recommendation
G.131; for the two applications mentioned,
the requirement is in the range
46-54 dB. This is a very stringent TERL
requirement, considering the signal-tonoise
ratio of the speech transmission
used, < 37 dB.
- Echo path delay
The echo path delay determines the necessary
length of the echo canceller filter.
The delay is defined as two times the
transmission delay between the echo
canceller and the hybrid, plus the dispersion
of the hybrid, Figs. 2 and 3.
- Type of transmission and end-user
equipment in the echo path
The performance of the echo canceller is
affected by non-linearities in the echo
path. Examples of devices that cause
non-linearity are reduced bit-rate coders
and end-user equipment with acoustic
crosstalk.
The total echo return loss requirement
must be met even when this type of equipment
is included in the network.
The above requirements are the same
for the two major applications, long-distance
telephony circuits and digital cellular
mobile applications.
- In both applications the echo canceller
is facing the same PSTN. It should
therefore be able to handle the same
echo path delay, type of transmission
impairment and type of end-user equipment
in both cases.
- The necessary total echo return loss,
in the proximity of 50 dB, cannot be
achieved solely with linear techniques.
Special non-linear techniques must be
employed to eliminate the residual
echo.
The requirements discussed are caused
by the transmission path and the enduser
equipment. The type of end-user
equipment and transmission impairment
will vary from connection to connection,
of course. In order to achieve the best
possible speech quality, it is extremely
important that the echo canceller adapt
to the specific situation on a per call
basis.
The echo cancellers employed must
also be capable of handling voice-band
data generated by modems and faxes,
which are widely used today. In the design
of echo cancellers, careful consideration
must be given to the type of signals used
by these devices.
Use of echo cancellers
For connections with long transmission
delay between PSTN subscribers, two
echo cancellers are necessary - one on
either side of the delay. For digital cellu-
26 Ericsson Review No. 1, 1996

lar applications, only one echo canceller
facing the PSTN is needed. The digital
mobile station is equivalent to a 4-wire
phone and must be carefully designed to
handle its own acoustic crosstalk. A standard
network echo canceller is not
designed to handle the phenomenon of
acoustic echo in digital cellularterminals.
This type of echo, which is characterised
by a longer echo path delay, is typically
non-linear because of the digital speech
coding and because of transmission
errors in the air interface.
Ericsson's echo
cancellers
Echo cancellers have existed for some
time as necessary components in longdistance
circuits (in satellite connections,
for example). Ericsson used to rely
on external products, commercially available
on the market, but things changed
when digital cellular telephony became a
reality. The echo canceller turned out to
be an essential component in the new
digital networks too. Every digital-cellularto-PSTN
connection needs its own echo
canceller device.
Ericsson realised that echo cancellers
are indispensable to be able to keep full
control of the speech quality of digital
cellular systems. It was therefore decided
that an Ericsson echo canceller
should be developed, as a general product
for all types of AXE 10 network applications.
The first product, ECP101, was released
in large volumes in Japan in 1994.
The idea was to create the best speech
quality conditions for the new personal
digital cellular (PDC) standard in Japan.
The ECP101 was a success and has now
been sold in many countries, for both
long-distance circuits and cellular applications.
The latest echo canceller in the ECP
series - called ECP303 - was released
in December 1995. The ECP303 is characterised
by outstanding speech quality
performance, thanks to a new adaptive
algorithm which is the fruit of close collaboration
between algorithm designers
and speech quality experts.
POOL configuration
ECP stands for Echo Canceller in Pool.
The possibility of designing a new product
from scratch led to the development
of an echo canceller completely integrated
into AXE 10. The ECP products form
part of the trunk and signalling subsystem
(TSS) and are placed in direct connection
with the group switch in a pool
configuration. The traditional way of handling
echo cancellers in the network has
been to connect a device to each trunk,
Fig. 4.
When an echo canceller is needed, AXE
10 picks up one of the devices in the pool
and routes the connection through it. In
this way it is possible to concentrate the
traffic and thus reduce the total number
of echo cancellers, compared with what
would have been needed in a trunk configuration.
At the same time, increased
Fig. 4
Echo cancellers in pool and trunk configuration.
ETC
EC
ECP
PSTN
MSC
Exchange terminal circuit
Echo canceller
Ericsson echo canceller in pool
Public switched telephone network
Mobile switching centre
Ericsson Review No. 1, 1996 27

Rg. 5
Echo canceller principles. A replica of the echo is
obtained via a linear filter and subtracted from the
input signal (S,N). The residual echo signal is further
suppressed using a non-linear processor.
reliability is obtained through the pool
configuration: there are always echo canceller
devices that can take over in case
of a fault.
The pool configuration not only increases
the flexibility of a system - an echo
canceller can always be connected without
affecting the traffic in the trunk - but
also allows more intelligent control of the
echo canceller devices. In the rerouting
of calls, or in a conditional call forward,
the unwanted doubling of the number of
devices involved in the connection can be
avoided.
Operation and maintenance
Another important principle in the development
of ECP products is that of integrating
the echo cancellers into the overall
operation and maintenance (O&M) of
the switch. In this way all the usual
AXE 10 maintenance requirements are
met, which means that the following functions
are implemented:
- parameter control
- continuous supervision
- fault detection and isolation
- restoration
- alarm
- manual test
- command-controlled tests.
In the latest ECP product (ECP303) the
platform of the echo canceller device may
be transformed into a test instrument. An
ECP303subrack in this configuration may
perform automatic or command-controlled
tests of the echo canceller function
of the pool, according to the ITU-T
0.27 standard.
The integration into the AXE 10 O&M
also offers the possibility of extracting
statistics, for example on the number of
echo canceller devices in use.
All these features assure the highest
level of service, reliability and availability.
Echo canceller algorithms
Themainpartofanechocanceller-which
makes it differ from previous echo control
strategies, such as echo suppressors
- is a linear filter, Fig. 5. This filter
makes a replica, or estimate, of the echo
path. Passing the input signal through the
filter generates an estimated echo signal
that is subtracted from the received signal,
in order to reduce the echo. Since
the echo path remains approximately the
same throughout the conversation, echo
reduction can also be obtained during
periods of simultaneous talk from both
parties.
In many situations, the echo reduction
obtained by the linear filter is not sufficient.
The output from the linear filtering
part is therefore passed to a non-linear
processor (NLP), which further reduces
the echo by blocking the signal, completely
or partially, when it is dominated
by non-cancelled (residual) echo.
The linear filter and the NLP are the two
basic blocks in most echo cancellers. For
good echo canceller performance to be
achieved, the operation of these two
blocks must be governed by control logic.
The control logic for the linear filter
assures that a good estimate of the echo
path is obtained. The control logic for the
NLP is responsible for detecting the presence
of a large amount of non-cancelled
echo to be suppressed by the NLP.
Adaptive filtering
The echo canceller can be used towards
any of a multitude of subscriber lines,
each resulting in a different echo path.
Hence, the echo canceller needs to adapt
its filter coefficients to the current echo
path. A simple solution, when a call is
established, would be to apply a test
pulse to the echo path, measure the
response and use it to determine the filter
coefficients. Unfortunately, there are
at least two drawbacks to this solution.
Firstly, the echo path may vary during the
call, which would require several recomputations
of the filter coefficients in the
course of the call. Secondly, due to measurement
noise, a single measurement of
28 Ericsson Review No. 1, 1996

the echo path may not yield a sufficiently
good estimate of the echo path.
Most of today's echo cancellers use an
adaptive linear filter with continuous
updating of the filter coefficients. The
speech from the far end (Fig. 5) is used
in place of the test signal, and the filter
coefficients are updated in response to
the correlation between the echo and the
speech. Because of complexity constraints
on the echo canceller, the most
widely used algorithm for adjusting the filter
coefficients is the normalised least
mean squares (NLMS) method. The
NLMS method adjusts the filter coefficients
so as to minimise the power of the
error between the true and the estimated
echo.
Non-linear processor
Depending on how well the echo path can
be modelled with a linear filter, the filter
gives an echo reduction of 10-35 dB,
where the higher reduction typically is
obtained for a linear hybrid and well-balanced
signal levels in the network. The
lower reduction is what can be obtained
if there are non-linearities in the echo
path, such as adaptive differential PCM
(ADPCM) links or acoustic echo at the
near end. It should be noted that -35 dB
is a limit. A linear echo canceller cannot
exceed this limit because of the non-linearity
in the signal caused by PCM compression
according to the A (or u) law. For
long delays, this obtainable echo reduction
is not sufficient. The echo is therefore
further suppressed by the non-linear
processor, NLP.
The function of this device is similar to
that of the previously used echo suppressors;
the signal is blocked, completely
or partially. This can be done with
a so-called centre clipper or with a controlled
attenuator. However, due to the
echo reduction of the adaptive filter, the
residual echo on which the NLP operates
has a much lower level than that on which
an echo suppressor operates. Hence, the
NLP can be tuned to make a much less
intrusive operation on the signal than an
echo suppressor would.
The NLP suppresses the output signal
if the residual power is below a predefined
threshold level. In the presence
of background sound (or noise) from the
near end (Fig. 5), the operation of the NLP
may lead to undesirable modulation of
this background sound. The perceptible
effect of such modulation is reduced by
adding generated noise to the output
when the NLP is blocking the signal - socalled
comfort noise.
Fig. 6 shows the signals at positions
a)-d) of Fig. 5 for an example where the
far-end party is speaking and where there
is a substantial amount of background
noise from the near end.
Control logic
The design of the control logic for the
adaptive filter and the NLP has a large impact
on the overall performance of an
echo canceller. The most important part
of the control logic of an echo canceller
is the part which secures good echo attenuation
throughout the call. The two parties
are normally talking one at a time,
which means that it is clear when to
update the filter in order to have a good
echo path estimate. But there are also
situations when the parties are talking at
the same time, referred to as double-talk
situations. The adaptation of the filter
should then be inhibited, otherwise an
erroneous estimate of the echo path is
obtained, which results in poor echo cancellation.
The control logic function that
inhibits adaptation in a doubletalk situation
has to allow the echo canceller to
Fig. 6
Echo canceller operation in a high background
noise environment. The signal diagrams a)-d) represent
the signal at points a)-d) in Fig. 5.
Ericsson Review No. 1, 1996 29

Fig. 7
System structure of the echo canceller in pool ECP
303. The three software blocks ECDHU, ECPU and
ECPR are part of the TSS (trunk and signalling subsystem)
of AXE 10. The echo cancellers in pool are
completely integrated in AXE 10, in direct connection
with the time switch module of the group
switch.
STS Statistics subsystem
TCS Traffic control subsystem
TSS Trunk and signalling subsystem
GSS Group switch subsystem
OMS Operation and maintenance subsystem
ECP303 Ericsson echo canceller In pool 303
ECPU Echo canceller in pool HW owner, central
program
ECPR Echo canceller in pool HW owner,
regional program
ECDH Echo canceller device handler
TSM Time switch module
converge, i.e. adjust its filter coefficients,
in the (much less frequent) situation of a
change in the echo path.
A fundamental problem of echo cancellation
is that these two situations,
which demand different actions, show the
same signal pattern. Double talk is shown
as a sudden increase of the power of the
residual signal at the same time as the
power of the far-end signal is high. At a
sudden change in the echo path, the linear
filter will not describe the echo path,
and the output of the filter will not be a
good replica of the echo. Hence, the
power of the residual echo will increase.
These two cases can be discriminated
by implementing a double-talk detector,
which is a state machine based on comparisons
of the power of the measurable
signals. The decisions on which action to
take concerning filter adaptation are
based on preset threshold values. These
values pose a problem, however. The
echo canceller is used in a variety of situations,
characterised by the signal levels,
the amount of background noise, the
echo path attenuation, and the possibility
of modelling the echo path with a linear
filter. If the situation differs from the
nominal case, for which the threshold values
of the double-talk detector were set,
the performance of the detector will be
degraded, which will affect the echo
reduction of the echo canceller.
Another important part of the control
logic of an echo canceller is the part which
controls when to activate the NLP. The
control of the NLP is based on comparisons
between estimates of the residual
echo and the signal from the near end. If
the power of the residual echo is expected
to dominate, then the residual echo
signal should be suppressed.
With the increased use of mobile and
cordless phones, the range of characteristics
of the background environment has
broadened. The phone is often used in
situations with a relatively high background
sound, and the operation of the
NLP is particularly noticeable in these
cases. Just like an echo suppressor, a
poorly tuned NLP would block this background
sound, whereas a well-tuned NLP
would allow it to pass - given sufficient
echo reduction through the adaptive filter.
The ECP303
echo canceller
The echo canceller has a great impact on
the overall speech quality performance of
telephony systems. However, it is not
easy to analyse how the speech quality
is affected by the adaptive algorithms
used. At Ericsson, collaboration between
speech quality and algorithm expertise
has made it possible to tailor a completely
new adaptive algorithm for the latest
product in the ECP series: the
ECP303.
The ECP303 echo canceller has been
designed to handle echo path delays up
to 128 ms. Normally, the echo path delay
in the PSTN network is shorter than
64 ms. If the operator knows that it is significantly
shorter, a filter length or 16 or
32 ms may be chosen. If the echo path
delay is longer than 64 ms, the filter may
be preceded by a pure delay of up to
64 ms, corresponding to a total echo path
delay of 128 ms.
ECP303 algorithm
The design of the ECP303 echo canceller
has focused on control logic capable of
handling a wide range of signal and echo
situations. The control logic of the
ECP303 is based on parameters that
have a direct impact on the actual performance
of the echo canceller. The logic
is not tailored to any particular situation,
in terms of signal level and echo path
attenuation. The ECP303 gives high echo
attenuation, without any noticeable
degradation of the near-end signal, for a
wide variety of situations.
30 Ericsson Review No. 1, 1996

The speech signal levels may vary
depending on the network characteristics
and the talker. The ECP303 can handle
these variations, including considerable
level differences between the
transmission directions. The attenuation
of the echo path - the echo return loss
(ERL) - has a great impact on echo canceller
performance. Normally, the operator
does not know the ERL value with any
accuracy. The ECP303 has therefore
been designed to obtain good echo cancellation
for all ERL values greater than
0 dB, without any need for parameter
adjustment.
The NLP in the ECP303 has an operation
threshold that is continuously adapted
to the characteristics of the situation.
This well-tuned NLP threshold, combined
with careful matching of the level and
spectral characteristic of the comfort
noise, greatly reduces the noticeable
effects of the NLP.
ECP303 characteristics
Apart from the outstanding speech quality,
the ECP303 has many other interesting
features. It is designed as a hardware
platform, containing a high-capacity digital
signal processor (DSP) and a microcontroller
(DP). The software of these
units can be loaded from the AXE 10 central
processor, APZ. This facilitates the
implementation of enhancements to the
product, such as customised functions or
modifications caused by changes in product
requirements.
The ECP303 can also be set in a "virtual
trunk" configuration. Some functional
properties (for example an extra pure
delay) can be assigned to a route, which
means that the echo canceller chosen
from the pool will automatically assume
the specific properties of that route. In
this way, the ECP303 can have access to
all the advantages of a trunk echo canceller
without suffering from its disadvantages.
Fig. 8
Fully equipped ECP 303 subrack, providing 256 echo cancellers organised in 8 extension modules (EMs) of
32 channels.
ised in the digital signal processor on the
ECAN3 board. There is one signal processor
for each echo canceller channel.
All boards have their own power module,
which simplifies power distribution in
the subrack: only -48 V is needed.
The ECAN3 boards use only surfacemounted
3.3 V components.
Echo canceller
speech quality
The most convenient way of evaluating
speech quality would be to use some
recognised objective measurements.
Fig. 9
An EM of 32 echo canceller channels consists of i
control board (EC0N3, right) and 2 digital signal
processing boards (ECAN3, left).
ECP303 building blocks
The ECP303 product consists of hardware
controlled by a number of software
blocks, Figs. 7 and 8.
A fully equipped ECP303 subrack provides
256 echo cancellers, organised in
eight so-called extension modules (EMs)
of 32 channels. Each EM consists of a
control board, EC0N3, and two digital signal
processing boards, ECAN3, Fig. 9.
The echo canceller algorithm is real-
Ericsson Review No. 1, 1996 31

Fig. 10
The desired goal Is an echo canceller with no clipping
and no echo. However, these requirements are
contradictory to some extent.
ITU-T's Recommendation G.165 specifies
a number of such measurements,
although it does not guarantee high
speech quality. Many echo cancellers
comply with G.165, yet behave totally different
in live conversations. Manufacturers
therefore have to rely on carefully
designed methods of statistically evaluating
subjective measurements in order
to obtain data for speech quality evaluation.
This situation calls upon the manufacturers
of echo cancellers to present
true and meaningful speech quality
evaluations to the operators, who can
also play an important role in requesting
this type of measurement. It is necessary
to make the evaluations in a
broad spectrum of network and environmental
situations in order to analyse
the various optimisations performed in
the design.
Subjective measurements of the
speech quality of echo cancellers cannot
be referred to traditional, subjective
speech quality evaluation of transmission
channels, as in the case of speech
coders, for example. Due to the duplex
nature of an echo canceller, evaluations
must involve simultaneous speech from
the far end and the near end. It is almost
only when a signal from the far end is present
that speech quality problems occur
on the near-end signal. Methods that take
these effects into account are therefore
needed.
Classes of speech quality impairment
To be able to evaluate speech quality, different
types of impairment have to be
classified. Today, four classes can be
identified:
- Echo
This is the residual echo that is not handled
by the echo canceller. It may be
the initial echo before convergence,
echo during or soon after a period of
double talk, etc.
- Clipping
This is the loss of speech from the nearend
caused by the NLP. It may be clipping
during a period of double talk or
clipping of a weak near-end signal.
Echo and clipping are to some extent contradictory
of each other, Fig. 10. If the
echo canceller has a high NLP threshold,
it will clip and suppress weak near-end
signals, and if the NLP has too low a
threshold, the echo canceller will not suppress
all the echo.
- Distortion
This is the distortion of the near-end
speech. It may be caused by erroneous
cancellation or by NLP problems which
do not result in a complete loss of
speech, as in clipping.
- Other impairments
This class includes all speech quality
problems that cannot be classified as
echo, clipping or distortion. Echo canceller
performance during the presence
of background noise from the near end,
e.g. comfort noise, belongs here.
ECP303: Technical data
Subjective speech quality: Tests have proved excel-
lent speech quality, better than or equal to other
products available on the market.
The ECP303 is an ITU-T type C digital echo canceller
conforming to Recommendation G.165. It provides
adaptive echo control.
Mechanical characteristics Echo path delay 16,32 or 64 ms
Subrack type BM12
selectable
Dimensions: Height: 244 mm Pure delay 0 to 64 ms selec-
Width: 488 mm
table in steps of 8 ms
Depth: 223 mm NLP modes NLP ON, with corn-
Weight (fully equipped): 11.8 kg fort noise
NLP ON, without
Electrical characteristics
comfort noise
Power feeding: -48 V DC NLP OFF, without
Power consumption:
comfort noise
fully equipped subrack, busy 110 W A/u, law: selectable
fully equipped subrack, idle 45 W Tone disabler in
accordance with
Performance ITU-T G.164 or G.165 selectable
Objective speech quality according to ITU-T G.165.
32 Ericsson Review No. 1, 1996

Subjective evaluation methods
Compromises are inevitable in the development
of algorithms. Especially in nonideal
situations - initial convergence or
non-linearities in the echo path, for example
- compromises between different
classes of speech quality impairment
have to be made. Because of these compromises,
evaluation of echo cancellers
against a fixed reference, i.e. an ideal
speech channel, is very difficult. The evaluation
should therefore be conducted
against real echo cancellers, since all
impairments may occur at the same time.
The listener has to judge which impairment
classes are the most disturbing in
each case and which echo canceller
makes the best compromises among the
different classes.
The most convenient way of performing
subjective speech quality measurements
is through listening tests. A group of test
persons are given the task of listening to
a recorded, simulated conversation after
echo cancellation (at point d in Fig. 5).
The advantage of these tests is that all
impairments created by the echo canceller
can be evaluated in a controlled
way, since all the echo cancellers are
exposed to the same signals. One disadvantage
is that a recorded conversation
is not a real-life situation. Live conversation
tests have to complement the
evaluation to include all effects.
The speech quality is evaluated in three
steps.
- During algorithm design, development
tests are performed. These tests underlie
the decisions on which compromises
to make.
- After implementation, verifying tests
are made to ensure that the desired
speech quality has been obtained. (The
group of randomly chosen test persons
must not include any of those who have
developed the echo canceller.) The
results of this test are statistically
analysed.
- In the marketing of echo cancellers,
tests are conducted to decide which
device has the best overall performance.
This third step is performed by
non-Ericsson companies, including customers.
Speech quality of the ECP303
Ericsson has developed new methods for
subjective evaluation of echo cancellers.
A large number of different test cases,
involving different echo paths, different
Ericsson Review No. 1, 1996
speech material, etc, have been chosen,
and test results for a number of echo cancellers
available on the market have been
recorded.
The best-performing echo canceller in
each test case was declared the "winner".
All the winners were then put together
to form a reference, called The Snake,
Fig. 11. An echo canceller whose performance
equals that of The Snake would
be the best echo canceller in the world.
In the development of the ECP303, The
Snake was the speech quality reference,
and the ECP303 was required to match
up to or outperform The Snake in every
single test case. This methodology has
ensured an outstanding speech quality in
the ECP303.
Conclusions
The ECP303, the latest product in
Ericsson's EPC series of echo cancellers,
is proof of the designers' commitment
to developing the best echo
cancellers that are used in today's telephony
systems.
The ECP303 has a compact subrack
completely integrated into AXE 10 and its
comprehensive operation and maintenance
system.
The echo cancellers have been implemented
in a pool configuration to ensure
the best economy, flexibility and reliability.
The outstanding speech quality performance
of the ECP303 has been achieved
thanks to a new sophisticated and robust
adaptive algorithm, which is the fruit of
close collaboration between algorithm
designers and speech quality experts.
w
Fig. li
The Snake Is defined as the ensemble of all the
winners in a contest of 62 test cases.
33

TEMS - A system for testing and
monitoring air interfaces
Rikard Lundqvist
The air interface is often a weak link in the chain that connects wireless
networks with mobile users. Efficient and easy-to-use tools for testing and
verifying the functionality of the interface are crucial to secure, high-quality
operation. These tools can also support the introduction of new features
and the associated standards.
With Ericsson's test mobile system, TEMS, and its supporting applications
FICS and GIMS, ongoing air-interface signalling and parameters of the radio
environment can be visualised and verified. The use of TEMS will help both
network operators and Ericsson's system developers to achieve higher
quality and reduce development time.
The author describes the continuous development of test tools based on
commercial hand-held mobile stations and the applications used to postprocess
gathered information.
In 1989, a group of engineers at Erisoft
- an Ericsson subsidiary based in Northern
Sweden - took part in a pre-GSM project
and later in the development of large
parts of Ericsson's GSM mobile station
software. They soon realised the need for
efficient, easy-to-use test tools to support
the development, verification, installation
and maintenance of cellular networks.
Fig. 1
In-door coverage measurements are easily carried
out, using TEMS Light.
34 Ericsson Review No. 1, 1996

Fig. 2
Data stored and replayed by TEMS and TEMS Light
may also be used in GIMS and FICS.
A test system prototype, based on the
first commercial GSM mobile - the Orbitel
TMT 900 - was displayed at the Telecom
91 exhibition in Geneva. It was very
favourably received, and by the end of
1991 the first release of a log mobile station,
LOMS, was used by the Swedish
operator Televerket Radio (nowadays
known as Telia Mobitel) and internally by
Ericsson.
After enhancement to include advanced
test and control functionality, the
LOMS was renamed TEMS (test mobile
system). It is now available in all digital
cellular markets supported by Ericsson.
The system includes two post-processing
tools: FICS (file and information converting
system) and GIMS (geographical information
mobile surveys).
System
components
One of the core components in any airinterface
test system is the data acquisition
unit. The TEMS solution utilises
the commercial platform provided by
standard Ericsson mobile stations,
supplementing it with TEMS mobile software
to form specialised test applications.
The main benefits of this set-up
are speed of development, the ability to
test like a user, and the familiar interface.
To provide message extraction
capabilities and other control functionality,
the existing software of the mobile
station has been modified and extended.
Another important part is the presentation
unit on which the data gathered is
displayed in different ways to the user. A
PC with DOS and Microsoft Windows is
the preferred control and presentation
system for most TEMS applications; the
same platform is also used for the FICS
and GIMS post-processing tools. The Pen
Windows platform has been chosen for
an application soon to be released: the
TEMS Light, which is intended for indoor
operation and walk-around measurements.
A standard RS-232 serial connection,
with a capacity of 9,600 bit/s, is used for
all communication between the data
acquisition unit (mobile station) and the
Fig. 3
Normal TEMS setup
Ericsson Review No. 1, 1996 35

Fig. 4
Modifications and additions to the mobile software
give access to control and extraction points.
36
presentation unit, and for distribution of
measurement data. Two air interfaces
can be monitored and tested simultaneously
by one PC.
TEMS applications
TEMS and its supporting applications are
versatile and capable of serving different
purposes in the development, installation
and maintenance of cellular networks.
TEMS functions range from low-level protocol
simulation to high-level testing and
data acquisition. Visual presentation
based on the commonly available Windows
graphical interface also makes for
ease of interpretation of measurement
data and reduces the need for training to
a minimum.
Several benefits can be derived from
developing a TEMS concurrently with
other system components, such as
mobiles and base stations:
- the use of one set of tools throughout
the testing process boosts productivity
through increased familiarity with the
tools
- in-process testing and debugging of
tools result in more stable and useful
commercial products
- test tools are commercially released
together with systems to be tested
- support to internal Ericsson development
units is provided
- test tools can be used both internally
and externally.
The pattern of use varies between the
different stages of development and
between different users. The TEMS environment
offers capabilities for prototyping
and rapid application development
(RAD) and for rapid try-out of new
features before they are introduced in
commercial networks or mobile stations.
During the different development
stages, the visualisation and control
functions provided are used to ensure
and monitor proper operation. When a
system has been developed and is
ready for verification, its performance
can be evaluated and analysed, and
TEMS also supports installation of
transceivers or other equipment that
require quick and easy go/no-go tests
on the spot. The fine-tuning of network
parameters during the optimisation
process - such as those related to
power and cell boundaries - takes
advantage both of TEMS' ability to present
relevant parameters and of its simulation
capabilities. Monitoring quality
of service and coverage and the troubleshooting
of problem areas are operation
and maintenance procedures supported
by TEMS.
Efficient testing
with TEMS
In addition to the functionality-oriented
reasons stated above, there are other
equally important reasons why TEMS
should be used:
- Rapid installation of new networks
requires controlled testing of newly
installed equipment.
- Deployment of new equipment in commercial
systems calls for thorough preinstallation
testing and non-intrusive
post-installation testing.
- Continuous testing provides a means
of improving quality and thereby
increasing customer satisfaction.
Increased availability also ensures efficient
use of installed equipment. Both
these factors will generate higher operator
revenue.
- Testing like a user is very important.
More and more end-customers use
hand-held mobile phones, which
means that relevant tests have to be
Ericsson Review No. 1, 1996

performed with hand-held test mobiles.
- Easy-to-use tools require less training
and fewer operating staff.
Arguments against the different types
of testing listed above count for little,
compared with the weighty arguments
for. Toughening competition among the
growing number of players in mobile
telephony makes it increasingly important
for Ericsson to safeguard its competitive
edge. After prices and services,
quality of service is the next field in
which the battle for customers will be
fought.
Even though methods of prediction are
becoming more and more reliable, there
are factors and conditions that cannot
be accurately predicted. The radio environment
does not readily lend itself to
simple and deterministic algorithms,
and the involvement of human factors,
weather conditions and other sources of
distortion and disturbance makes testing
vital to successful operation. Predictions
will always have to be validated.
The necessity of testing at the mobile
station may also be questioned, inferring
that the up-link is the crucial factor. It is
true that high quality in the up-link is vital,
but for overall high system quality to be
achieved there must be a good balance
between the up-link and the down-link.
The user must also be able to verify that
changes made at the base station were
appropriate. Since network data does not
indicate the accurate geographical position
of the mobile, the best results are
obtained when data is gathered both from
the network and from the mobile.
Mobile stations with
TEMS capability
The data acquisition units used in TEMS
applications are based on Ericsson's
mobile stations. Software modifications
and additions transform ordinary handsets
into test tools, while regular services
(placing and receiving calls, etc) are still
supported. Thus, a mobile station with
TEMS capability can be used just like a
normal mobile phone, both during a test
session and when it is not connected to
a PC.
The software modifications fall into two
categories: intrusive and non-intrusive
ones. Intrusive modifications are those
that alter specific mobile functionality so
that it will be controlled by the software,
while non-intrusive modifications normally
deal with the monitoring of messages
or the status of internal mobile parameters.
The most significant addition to the
software is the serial communication
process. As the name implies, this
process controls the data link between
the PC and the mobile, but it also distributes
messages and commands and
changes the global parameters that
define how the TEMS-specific part of the
mobile operates.
Messages between the network and
the mobile can be copied and sent to the
PC over the serial link. This is performed
in one or more extraction points. The
functions for active control are normally
implemented in conjunction with the
related original code, which results in
several control points. The most
advanced control point is that which controls
how messages received or sent
over the air interface are handled. Messages
can be modified, sent, delayed or
deleted according to pre-defined test
scenarios. Other important control and
monitoring points are implemented in the
display and keypad processes, for ease
of access to the user interface of the
mobile station.
Windows-based testing
with TEMS
TEMS consists of a modified mobile station,
as described above, a serial cable
and a PC-based software package. The
PC software utilises Microsoft Windows'
powerful graphical environment to control
and initiate events and to monitor messages
and reports generated in the
mobile - or the two mobiles - connected.
Depending on standards and configuration,
the TEMS tools support slightly different
functions. Some examples from
the GSM and D-AMPS versions of TEMS
are given in the following.
Information is monitored in open windows.
There are both status windows,
continuously updated with relevant information,
and message windows, which
mainly display the flow of messages
between the mobile station and the base
station. The latter windows also display
TEMS messages specifically created in
the mobile. These created messages
increase the clarity and usefulness of the
application by visualising parameters
that are normally internal to the mobile.
Ericsson Review No. 1, 1996 37

Status windows are created and combined
so that interrelated information elements
can be found in the same window.
For example, parameters relating to the
cell currently being camped on can be
found in the serving cell window, and
radio parameters - such as signal
strength and bit error rate - are available
in the radio environment window. Other
examples of status windows are dedicated
channel window, and windows
showing measurements of surrounding
cells and received signal strength indications
(RSSI).
Messages received from the mobile
station are presented in multiple windows.
The last received message is displayed
at the very bottom of a window.
Depending on the extraction points and
the user-defined filtering, messages from
several layers may be presented. Layer 2
and layer 3 messages are most frequent,
but layer 4 is also available for short-message
services (SMS) presentation. If a
particular message is selected, it will be
decoded and presented in detail by
decoding routines implemented in the
TEMS PC program. These routines are
very similar to those found in the base
station and in the mobile station. Decoded
information is presented in plain English
and in accordance with existing
standards, specifications and recommendations.
Storage of information
Because of the complexity of cellular networks
and the large number of messages
and parameters monitored by TEMS, the
ability to store information for later postprocessing
is important. In order to
reduce the amount of information saved,
the user can apply several types of filter
on the basis of different criteria: point in
time, message type, or a specific message.
All information stored is marked
with an exact indication of time, obtained
from the mobile station. TEMS supports
a wide range of positioning systems, e.g.
the global positioning system, GPS. If
such a system is used, the current geographical
position will also be included
in the stored information. File marks with
a user-defined text string and a unique
number indicating special events, problems
and even geographical locations,
may also be included in the stored data
files.
When stored data is replayed, the
speed at which information is updated
can be controlled, thereby enabling more
careful examination and analysis. Stored
information can also be exported into
ASCII text files for post-processing in
some user-selected application, e.g.
Microsoft Excel.
Simulated handset
Full access to the man-machine-interface
(MMI) of the mobile station is provided in
a window that simulates the handset.
Both the keypad and the display are fully
operational and function concurrently
with - and as an exact copy of - the real
handset. All keys can be used, and the
display can be closely monitored. This
functionality is achieved through TEMS
access points implemented both in the
display driver and in the keypad handler
routines of the mobile station software.
Dialling sequences can be recorded
and replayed at a later time. In addition
to the keys available on the handset,
these sequences may also contain operational
commands, e.g. commands that
request the user to wait a defined number
of seconds or to start saving information.
If an automatically controlled call
is inadvertently disconnected, the user
can have it redialled automatically.
Incoming (mobile-terminated) calls can
be automatically answered and optionally
terminated after a pre-set time if the
user so specifies.
Fig. 5
Window presenting the bit error rate and received
signal strength of a serving cell together with the
signal strength of the two strongest neighbours
and handover indications.
Mh ;L,
RSSI and BER Graph
A\k -SOdBm
30dSa
l-IOOdBm
' Sample data
I at H R974 3M7
Ion E 01757.3412
RSSI -CE.8S dBn
BER: 0072
Channel: 2G5
' Selection range
From: 15:27 59 53
Length: 00:00 01 08
38 Ericsson Review No. 1, 1996

Fig. 6
Air interface messages are decoded in detail to
make interpretation quick and easy.
Messages handling
One of the most advanced features of
TEMS is its ability to control the flow of
messages in the layer 2 - layer 3 interface
in the software of the mobile. This
ability-which is most useful at the development
and verification stages - provides
a means of altering messages after
they have been received from the network,
and modifying messages before
they are transmitted from the mobile station.
Modifications are controlled by
downloading a so-called script to the
mobile. The script serves as a short program,
defining the sequence, type and
parameters of the different operations to
be performed. There are message-oriented
operations, such as DELETE, MOD­
IFY, SEND and DELAY, and control operations,
such as WAIT, REPEAT and GOTO.
Alternating the relative sequence of messages,
sending unexpected messages,
modifying the content of internal messages
and delaying messages help to
analyse the system and to verify correct
behaviour even in abnormal situations.
Counters may be introduced into the
scripts, for automatic counting of events
and situations, e.g. assignment failures
and hypothetical success ratio statistics.
In most TEMS mobiles, a lower power
class can be simulated by temporarily
changing the mobile station class mark.
For instance, the power restraints of a
hand-held mobile can be applied to a vehicle-mounted
mobile.
In GSM-based tools, parameters can
be read from and added to the subscriber
identity module (SIM) card. No access is
given to parameters that might jeopardise
security. A similar functionality is the ability
to read number assignment module
(NAM) information in D-AMPS-based products.
Idle mode and dedicated mode
Non-standard cell selection includes
functions both for idle mode and dedicated
mode (with a dedicated channel
allocated). In idle mode, a mobile can be
locked to a particular channel or a group
of channels regardless of signal strength.
In dedicated mode, handovers can be prevented
either to all measured neighbours
or to a selected group. This is implemented
by no signal strength being
reported on prevented channels, thus
misleading the network-based handover
algorithms. In a similar way, handovers
can be introduced by increasing the signal
strength reported for a selected neighbour
to a value exceeding that of the currently
dedicated channel, while reporting
no other selectable neighbours. In GSM,
the parameter that bars a cell for initial
access can be overridden, which means
that the cell can be set up and used as
a test cell before other users are allowed
to use it. This has proved efficient when
a network operator wishes to go on-line
with new cells and features, and try out
new network settings in a commercial network
with many customers.
Fig. 7
A simulated handset
gives complete access
to all keys and continuously
updates and displays
the content of the
phone LCD.
Ericsson Review No. 1, 1996 39

Automatic channel testing
When new equipment has been installed,
automatic channel testing will quickly
indicate whether it is working properly or
not. Repeated calls are set up according
to the user's specifications, and if the
assigned channel is on the testing list,
the measured parameters are compared
with defined threshold levels. If this comparison
is favourable, the channel is
marked OK; otherwise it is marked NOT
OK and additional testing commences. All
testing is aborted if none of the channels
on the testing list has been allocated during
a user-defined number of attempts.
The channel testing functionality is also
used as a maintenance tool for quick verification
of the availability of channels.
The user can enter cell names to make
it easier to recognise used channels. The
cell name - for example, a combination
of the site identifier and a sector indication
- is then used whenever the channel
is referenced. Another function that facilitates
operation is the system's capability
to save in a set-up file the configuration
of open windows on the screen.
Fig. 8
Drive testing of cellular networks may be performed
at the cell site, cluster or network level
Frequency scanning
Frequency scanning gives the user
access to the inherent capability of the
mobile to tune the receiver to different
channels and then measure the signal
strength. By selecting the desired channels
and activating repeated scanning,
continuous measurements can be made
to verify cell planning and neighbour list
definitions and to indicate suitable channels
and/or groups of channels for additional
sites. Scanning data can be saved
for later post-processing.
Fixed testing and drive testing
TestingwithTEMScan bedivided into two
major groups: fixed testing in a laboratory
environment and drive testing in the
field. Fixed testing usually takes place
either at the network development stage
or at the verification stage. All messages
must then be carefully examined and timing
issues investigated.
When the deployment of network equipment
has started, either for field testing
of new functionality or to set up a new
commercial network, drive testing with
TEMS becomes essential. Testing is usually
performed from the cell level and
upwards in the network topology.
Separate cells are tested for channel
availability, coverage and ability to successfully
set up and terminate both
mobile-originated and mobile-terminated
calls. If intra-cell handover (handover
between channels in the same cell) is
defined, this function can also be tested.
At the next level, a whole site with all its
cells is tested. The coverage of the site
will be the aggregate coverage of the
cells, but it may be necessary to analyse
and adjust parameters that define the cell
boundaries to optimise both reselection
and handover behaviour. A group of sites,
usually without any frequency reuse, is
called a cluster. Clusters are tested much
in the same way as cells, for boundary
discrepancies, ability to maintain a call
while switching between serving cells,
etc. Testing the whole network or part of
it usually reveals any problems related to
co-channel interference.
For operation and maintenance purposes,
both fixed testing and drive testing
is used to verify quality of service,
track down elusive problems and analyse
customer complaints, after a network has
been installed.
TEMS Light
As we move from car-mounted mobile
phones towards personal communication
systems, high accessibility with virtually
complete indoor coverage
becomes a necessity. Testing in this
environment requires small, portable,
battery-powered test tool with position-
40 Ericsson Review No. 1, 1996

Fig. 9
TEMS is used world-wide to visualize cellular airinterface
signalling.
ing capability. TEMS Light is a tool that
caters for these needs.
The TEMS Light system contains a
mobile station, Pen-Windows software to
be executed on a palm-top computer, and
a serial connecting cable. The mobile station
and the cable are of the same type
as those used in the TEMS system, but
the presentation and control software
has been adapted to the walk-around
environment with standard functions and
operations automated or simplified and
optimised for pen use. The most important
information elements are quickly
viewed at a glance.
Since most 'standard' positioning systems-GPS,
for example-are inadequate
in the indoor environment, an easy, yet
very flexible positioning method has been
implemented in TEMS Light. It is simply
a matter of point and click. The blueprints
or maps of the currently surveyed area
are imported into the background, and
positioning only requires a light tap at the
current location to pinpoint collected
measurement data.
Data gathered by TEMS Light can be
saved for later examination. The file format
is TEMS-compatible, enabling closer
analysis and more detailed presentation
Fig. 10
Typical TEMS light screen with background blueprints
and measuring points.
Ericsson Review No. 1, 1996 41

Fig. 11
Using different converters, the FICS shell creates ;
number of output formats.
in TEMS. During replay of data in TEMS
Light, any previous measurement may be
selected, so that all status information
originating in that point can be displayed.
The TEMS-compatible file format also
means that data files collected by TEMS
Light can be post-processed by FICS and
GIMS. Even the background blueprints
can readily be used in GIMS.
TEMS Pocket
For a user it is impracticable always to
have a PC with a TEMS - or a palm-top
computer with a TEMS Light - hooked up
to his or her mobile. But if a call is lost
while you are driving to work, it is difficult
to pinpoint precisely what went wrong
without having access to some information
about the radio environment - channel
parameters, for example. This is an
example of situations where TEMS Pocket
comes in handy.
TEMS Pocket is a mobile station with
some integrated monitoring and test features
added. Information is presented on
the mobile's LCD, and test commands
are operated through the menu system
by using the ordinary keypad. TEMS Pocket
will always be available and, hence,
suitable for unexpected communication
problems, presenting the most important
parameters of the network. Another typical
area of use is the installation of new
equipment, when a number of parameters
have to be checked very quickly. Ideally,
all radio and network engineers in an
operator's organisation should have a
TEMS Pocket installed in their personal
mobiles.
TEMS Pocket does not support any
external logging capabilities, but a limited
number of user-activated events can
be temporarily stored in the volatile RAM
memory of the mobile. When a problem
area has been identified by TEMS Pocket,
TEMS or TEMS Light can be used to
analyse the cause of the problem and to
gather data for processing.
File and information
converting system, FICS
FICS is a Microsoft Windows-based product
for post-processing of data gathered
by TEMS. It contains a software shell, an
operating platform and multiple add-ons.
Each add-on, or tool, serves a specific
purpose and performs conversion to a
particular format.
Several conversion jobs can be simultaneously
defined for effective batch
operations. The outcome of a requested
conversion is presented in a history
window. Several input files can also be
combined into an output file. Defining a
FICS job may require the user to specify
additional information on how conversion
is to be made. After a job has
been defined, conversion is automati-
42 Ericsson Review No. 1, 1996

Fig. 12
Example output from a "call statistics" FICS converter includes a number of values in ranges defined by the
user.
cally controlled without any additional
user intervention. Lengthy operations
can therefore be executed at night or
during a break.
Call statistics is one FICS conversion
in which events and the distribution of
parameters are analysed. The success
ratio of a number of call events, such as
initiation, termination and handover, is
analysed in addition to the distribution of
system response time, transmitted
power, signal strength and power level.
Other examples of FICS post-processing
functions is the conversion to an EETreadable
format and the creation of a pure
text file for user-defined post-analysis in
another tool.
Geographical information
mobile surveys, GIMS
Like TEMS and FICS, GIMS is a Microsoft
Windows-based product for post-processing
of data gathered by TEMS and
TEMS Light. GIMS uses Maplnfo, a commercial
GIS (geographical information
system) software package to visualise
and geographically distribute relevant
data with different colours and symbols
indicative of status and operation.
Survey files created by TEMS or TEMS
Light are easily imported into GIMS. During
the import process, advanced userdefined
filtering can be applied to reduce
the number of samples, increase readability
and make the data more suitable
for map presentation.
The user can select information to be
filtered on a per message basis or on a
per event basis to suit specific needs and
to speed up the analysis when large
amounts of data are handled. If a second
mobile has been used in TEMS, data from
that mobile can also be used in GIMS.
General import of text files with positioned
data is useful for importing and
presenting other types of information,
such as base station locations on the
map background.
Applying analysis themes to the data
defines what is presented and how it is
done. Themes define symbols, colours,
ranges, etc. A set of pre-defined themes
are delivered along with GIMS, but the
user can redefine these themes or set up
completely different ones according to
needs, previous experience and corporate
standard. GIMS is based on Maplnfo,
which gives great flexibility and editing
functions.
GIMS contains map windows for presentation
of geographical data with
accompanying blueprints or background
maps. Data is presented in a tabular format
in browser windows, and graphs can
be used for all numerical values. Numerical
values can also be statistically
analysed. A layout is provided as a tool
for defining a standardised format for all
printouts.
When position information cannot be
obtained in TEMS -for example, if no GPS
signal is available due to a large number
of high-rise buildings or other obstacles
- file marks can be used as a crude positioning
method. When the data is later on
imported into GIMS, accurate positions
of the file marks have to be indicated, and
all intermediate measurements will then
be evenly distributed along the straight
line that connects two consecutive file
marks.
Ericsson Review No. 1. 1996 43

Fig. 13
The GIMS addition to Mapinfo makes it easy to present
geographical information gathered by TEMS.
Colours and symbols indicate parameter values and
specific events.
Examples of parameters presented
with geographical positions in GIMS
include signal strength, bit error rate, possible
interference, number of neighbours,
dragging or slow handovers and cell reselections.
Future challenges
Continued improvement and extension of
the TEMS family of test tools poses some
challenges:
- introducing new software into featureladen
mobiles without disturbing normal
operation
- keeping up with the continuous development
of mobile stations
- maintaining a flexible tool set that satisfies
both high-level and low-level functionality
while focusing on ease of use
and suitability for specific tasks
- utilising new and improved operating
systems and software packages, both
in the PC and in the mobile, while keeping
the need for hardware at a minimum
- implementing coding and decoding of
signalling protocols concurrently with,
or even in advance of, the finalisation
and implementation of standards in
base stations and mobiles
Combining internal and external development
and maintaining a flexible test platform
will strongly support Ericsson's
objective of achieving high-quality cellular
products and cutting development cost
and lead time.
Conclusions
Ericsson's test mobile system TEMS is a
versatile tool for testing and monitoring
air interfaces. TEMS functions range from
low-level protocol simulation to high-level
testing and data acquisition in the development,
installation and maintenance of
cellular networks.
TEMS Light is adapted to walk-around
testing in indoor environments, with standard
functions and operations simplified
and optimised for pen use.
TEMS Pocket is a mobile station with
integrated testing and monitoring features.
It is a handy tool, always available
and suitable for analysing unexpected
communication problems.
FICS and GIMS are Microsoft Windowsbased
products for post-processing of
data gathered by TEMS.
Together with these supporting applications,
TEMS can be used by both network
operators and Ericsson's system
developers, to achieve higher quality and
reduce development time.
44 Ericsson Review No. 1, 1996

Telefonaktiebolaget L M Ericsson
S-126 25 Stockholm, Sweden
Phone: +46 8 7190000
Fax: +46 8 6812710
ISSN 0014-0171
Ljungforetagen, Orebro 1996