Chapter 17: Testing and Type
Approval of the Mobile Stations
Re
´
mi Thomas and David Barnes
1
17.1 Introduction
In this chapter we aim to explain the principles and evolutions of the testing and type approval
of the GSM Mobile Equipment (ME). After having recalled the context at the end of the
1980s (paragraph 17.2), we outline the technical background (paragraph 17.3) and we explain
the first step of the type approval, i.e. the ‘‘interim type approval’’(paragraph 17.4). Then we
describe the TBR regime (paragraph 17.5) and how it evolved to cope with phase 2 (para-
graph 17.6) and phase 21 (paragraph 17.7). We explain the work of two committees, namely
ETSI/SMG7 and TAAB (paragraph 17.8), we give some hints regarding the test tools (para-
graph 17.9). Finally in the conclusion (paragraph 17.10) we consider the achievements of the
GSM type approval and we rapidly explain how the principles of the GSM conformity testing
are presently reused by 3GPP for the conformity testing of the UMTS terminals.
17.2 The First Steps in 1988–1990
To understand the ideas underlying the first actions for GSM Mobile Stations (GSM MS)
testing, it is necessary to recall some elements which were agreed by the GSM community
already at the end of the 1980s:
†
ME and SIM together form the GSM MS; the Subscriber Identity Module (SIM) is a chip
card provided by the network operator to the mobile subscriber, this card contains in
particular all the subscription related data;
†
Unlike to the SIM card, the ME is not under the control of the network operator, therefore
the network operators have to ensure through some testing process that the MEs will be
able to provide service and will not cause any damage to the network.
There were therefore two overriding criteria. Firstly the MSs from different manufacturers
should interwork with all networks. Secondly the GSM community required that roaming

should work from the beginning of the commercial service. Today this is seen as a straight-
forward feature but this was not the case in those years and this resulted in some strict
conformity requirements on the MEs.
1
The views expressed in this chapter are those of the authors and do not necessarily reflect the views of their
affiliation entities.
GSM and UMTS: The Creation of Global Mobile Communication
Edited by Friedhelm Hillebrand
Copyright q 2001 John Wiley & Sons Ltd
ISBNs: 0-470-84322-5 (Hardback); 0-470-845546 (Electronic)
To cope with these two requirements the GSM community set up two processes:
†
A group was established under the leadership of the ‘‘ Permanent Nucleus’’ (PN) of the
CEPT/GSM committee. The mandate of this group, the ‘‘ Eleven Series Drafting Group’’
(ESDG), was to specify conformity tests of the ME. The tests were produced on a volun-
tary basis by experts provided by companies contributing to the GSM standardisation;
these tests were included in a GSM specification, namely GSM 11.10;
†
A process for the development, purchase and deployment of a test tool implementing the
ME tests defined by the ESDG, this testing tool was called the ‘‘ System Simulator’’ as the
purpose of this tool was to completely simulate a GSM network.
These two tasks were in fact quite ambitious. In order to understand the situation, it is
useful to consider a general description of GSM 11.10 and the complexities of developing a
system simulator.
17.3 The Technical Context
17.3.1 Some Insight into GSM 11.10
We give some insights in tests domains which are defined in TS GSM 11.10 the purpose of
which is quite ambitious as it aims to cover all the aspects of the ME conformity, both at the
air interface and at the SIM/ME interface.
17.3.1.1 Radio Tests

This part of the testing checks that the ME is compliant with the radio performance defined in
the 05 series of the GSM recommendations. Thus a compliant unit will behave in a foresee-
able manner. This is particularly important in order to enable a GSM network operator to
make a radio design providing all the advantages of the GSM radio performance.
The main domains of testing are listed hereafter: transceiver, transmitter, receiver, refer-
ence sensitivity, usable receiver input level range, co-channel rejection, adjacent channel
rejection, intermodulation rejection, blocking and spurious response, timing advance and
absolute delay, access times during handover, temporary reception gaps, channel release
after unrecoverable errors.
17.3.1.2 Radio Link Management Tests
This comprises the cell selection and reselection tests and the received signal measurements
tests.
17.3.1.3 Signalling Tests
All the signalling protocols of the radio interface are tested. First, this comprises the tests of
the layer 2 signalling functions which verify the conformity with the LAPDm protocol
defined in GSM 04.06.
This comprises as well the layer 3 testing which verifies the conformity with TS GSM
04.08, we indicate rapidly the main domains of testing:
GSM and UMTS: The Creation of Global Mobile Communication432
†
Handling of unknown, unforeseen, and erroneous protocol data: the purpose of such tests
is to cope with future evolutions of the protocol;
†
Test of the elementary procedures for radio resource management: Immediate assignment,
normal paging, measurement report, channel assignment procedure, handover, ciphering
mode setting, channel release;
†
Test of the elementary procedures for mobility management: TMSI reallocation, authen-
tication, identification, MM connection;
†

Test of the procedures for call control;
†
- Finally testing of structured procedures is performed in order to verify that the ME under
test is able to perform call set-ups.
17.3.1.4 Testing of the SIM/ME Interface
The purpose of these tests is to check the conformity of the SIM/ME interface of the ME as
defined in TS GSM 11.11, testing of the SIM itself is not part of TS GSM 11.10.
It is tested that the ME is able to retrieve information from the SIM and to update them, for
example this comprises the following tests: forbidden PLMNs, location updating and unde-
fined cipher key; MS updating forbidden PLMNs, MS updating the PLMN selector list.
Some physical characteristics of the SIM/ME interface are further verified: mechanical
tests, and electrical tests.
The test of some services (Short Message Service (SMS), some supplementary services)
and of some MS features can be found as well in GSM 11.10.
Last but not least, GSM 11.10 comprises tests of speech teleservices and tests of speech
transcoding functions.
17.3.2 Phased Development of GSM – the 1800 MHz Band
In October 1989 the GSM community chose to develop the GSM standard in phases and to
produce a complete set of standards for each phase.
Phase 1 of GSM was functionally frozen in January 1990 and it is worth recalling that it
only encompassed the 900 MHz band thus excluding the operations in the 1800 MHz band
which were called ‘‘ DCS 1800’’ . In order to specify the DCS 1800, ETSI/SMG created delta-
specifications which were to be understood together with the corresponding phase 1 speci-
fications. For instance there was a specification named ‘‘ GSM 11.10-DCS’’ the first version of
which was approved by ETSI/SMG in June 1992.
The basic phase 2 of GSM was frozen in October 1995; among other decisions it was
decided to include DCS 1800 in basic GSM phase 2. This implied that 900/1800 operations
(mainly 900/1800 handover and 900/1800 selection/reselection processes) were part of basic
phase 2.
Then the concept of GSM phase 21 was introduced in order to add continuously new

features to the GSM standard. It is the choice of each network operator, each network
infrastructure manufacturer and each MS manufacturer to implement or not to implement
any of the GSM phase 21 features.
To achieve this, new sets of functions are added without changing the protocols for the
existing ones and without causing incompatibility problems with already existing equipment.
Chapter 17: Testing and Type Approval of the Mobile Stations 433
In other words, there is an unchanged base which is basic phase 2 as approved in October
1995 and then the phase 21 features are build on it. Based of these rules, new releases of the
GSM standard were produced: Releases 96, 97, and 98.
As a consequence a new version of GSM 11.10 was written for each phase of the GSM
standard: a limited number of tests were modified to take into account evolutions of already
existing features, and new tests were added in order to test new features.
17.3.3 Availability of System Simulators
The key to the testing of GSM MEs for type approval purposes was the availability of a
suitable System Simulator (SS). As already explained the operators required a test specifi-
cation and an SS that was sufficiently comprehensive to ensure interworking between
different manufacturers’ equipment and enable international roaming. It was apparent that
such a development would be extremly expensive and the burning question was, who should
pay?
There was no precedent for regulators to fund such system simulators where the require-
ments to ensure interworking were above and beyond those normally required for regulatory
purposes. A similar situation had existed prior to this in the UK with the introduction of the
TACS system and in this case a precedent had been set where the operators concerned had
funded the development and purchase of a system simulator.
In the UK situation the operators (there were two in the UK at that time, Cellnet and
Vodafone) funded the purchase and development of an SS. There was then a factor built into
the type approval fee for the operators to recover their outlay over a set number of type
approvals. At that stage of course no-one was sure just how many different terminals would
be produced and eventually type approved. Needless to say all estimates were considerably
less than the actual number and the financing arrangements worked successfully.

A similar approach was therefore considered for GSM and a small number of operators
who were members of the GSM MoU undertook to underwrite the development of the SS for
GSM type approval. This group of operators was known as the ‘‘ buyers club’’ and was a vital
initiative to get the development and procurement work underway for the delivery of an SS.
After the delivery of the initial SS it became apparent that it would be unfair for the original
buyers club to continue to take the full financial burden of the continued development of the
SS and the necessary test cases. The GSM MoU however provided the necessary finance and
resources by establishing a separtae company within the MoU called GSM Facilities Limited
(GSMFL). It was the purpose of GSMFL to see the development of the SS through to the
complete implementation of GSM phase 2.
Throughout the lifetime of the buyers club and the GSMFL, Peter Zollman (Vodafone)
worked as the project manager for the development of the SS and the associated test cases. In
this respect his efforts were particularly appreciated in maintaining the project on track and
keeping the GSM community informed of the current status.
The GSM 1800 (then known as DCS 1800) MEs were not catered for in the GSM SS. This
was because at that time DCS 1800 was still a separate part of the GSM family. However, the
DCS 1800 operators faced exactly the same problem as the GSM 900 operators in terms of
their requirements for a suitable SS. At that time the only networks approaching commercial
service were in the UK and One-2-One and Orange jointly funded the initial development and
procurement of an SS. In actual fact they purchased two simulators.
GSM and UMTS: The Creation of Global Mobile Communication434
At a later stage it was seen that there was a continuing process of validation of test cases for
the SSs and a DCS validation group was set up under the PCN TAEG. This group was chaired
by David Nelson (Orange).
17.4 The Interim Type Approval Procedure
17.4.1 The Technical Process
In 1987, the operators who signed the GSM MoU committed themselves to start the GSM
service in 1990. Unfortunately the delays in the different developments resulted in delaying
the fulfilment of this demanding commitment. In fact the phase 1 of the GSM standard was
functionally frozen in January 1990. As a straightforward consequence, it was not concei-

vable to have GSM network infrastructure on the field in 1990.
Nevertheless the situation for the GSM terminals was even more worrying. As a matter of
fact it appeared in 1990 that it was impossible to foresee when the test of the first GSM ME
would be completed. Type approval and testing are inevitably blamed as a source of delay in
the availability of MSs for a system such as GSM. One of the reasons for this is that by the
very nature of the standards development process the test specifications can only be produced
and finalised once the contents of the core specifications is clearly defined. Furthermore, one
has to be careful to ensure that the test specifications only reflect what is in the core specifica-
tions and that no additional requirements are inadvertently added. Finally it is difficult to
place a detailed order for test equipment without a comprehensive test specification.
However, this is effectively what had to happen with the test equipment manufacturer having
to track the changes to the test specifications as they were refined. This was further compli-
cated by the fact that the test specifications were still trying to track corrections that were
made to the core specifications.
It was therefore somewhat inevitable that delays were experienced by the extremely
ambitious processes described in the previous paragraphs (namely the drafting of GSM
11.10 and the development of the GSM SS):
†
significant work was already accomplished for the radio transmission tests but there
remained some important work to produce a sufficient amount of signalling tests;
†
the delays in the development of the SS were even more worrying, it became obvious that
the SS was on the critical path of the GSM commercial opening, in fact it was not possible
to foresee precisely when the SS would be able to test a GSM ME.
To cope with this difficult situation two types of actions were undertaken:
†
on the standardisation side more efforts were put into the production of the tests, in
particular during the first months of 1990 more signalling and protocols experts were
involved in ESDG and a ‘‘ signalling subgroup’’ was set up;
†

regarding the test tools it became obvious that it was necessary to start testing with a tool
including less features than the SS.
As a result of this the regulators and the GSM operators decided at the end of 1990 to set up
the ‘‘ Interim Type Approval’’ (ITA) in order to speed up the start of the GSM terminals
testing. At that time there was considerable discussion between the operators on the minimum
level of testing that would be acceptable for ITA. One view called for a high level of testing to
Chapter 17: Testing and Type Approval of the Mobile Stations 435