Rohde and Schwarz CMU200v10 Data Sheet

MOBILE RADIO

Test methods

Universal Radio Communication Tester CMU200

Measuring bit error rate
on GSM mobiles

The test modes defined by GSM stan-

dardization committees prescribe how

to measure the receiver characteris-

tics of GSM mobile phones. This is the

kind of application that CMU200 [1]

with its advanced concept excels in.

BER – a measure of receiver
sensitivity

The transmitter characteristics of GSM
mobiles are relatively simple to measure,
since the physical effects can be
checked directly on the tester. But when
it comes to receiver characteristics, the
physical effects appear in the tested
device itself, so no direct measurement
is possible. GSM standardization commit­tees therefore defined test modes for
measuring the receiver characteristics of
GSM mobiles.

The major feature of a receiver is its
sensitivity. In digital systems, this is
determined through the bit error rate
(BER). The receiver is fed a test signal
with pseudo-random bit sequence and
defined level, and the number of bit
errors is measured at its output.

In development and conformance testing
of GSM mobiles the receiver character­istics have to be tested under various
aspects like fading, multipath reception
or intermodulation. But in production it
is sufficient to stimulate the receiver
with a low-level GSM signal. Usually,
either reference sensitivity or absolute
receiver sensitivity is measured on GSM
mobiles.

Reference or absolute
sensitivity?

To check the reference sensitivity, a
signal with defined level (e.g. –102 dBm
or –104 dBm for GSM900) is applied
to the receiver. If the measured BER is
below the specified limit, the receiver is
ok. To determine the absolute receiver
sensitivity on the other hand, the level of

the test signal is varied until a defined
BER is obtained.

Obviously, absolute receiver sensitivity
takes more time to measure than refer­ence sensitivity. So in production, where
you are interested in maximum through­put, measurement of the reference sensi­tivity is naturally often preferred.

BER test modes

The basic principle of the BER test modes
is simple: the radiocommunication tester
sends a data stream to the mobile,
which then sends it back to the tester
(loop). The tester compares sent and
received data streams to determine the
number of bit errors (FIG 1).

Various test modes (loop types) are
defined. With types A, B, D, E and F
the tester generates a pseudo-random
bit stream, which is channel-coded and
applied via the RF interface to the
receiver of the mobile. There the data
stream passes through the channel
decoder and – via channel coder, RF
interface and channel decoder – is sent
back to the tester. What precisely the
mobile sends back depends on the type
of loop. With loop B it returns exactly
what it has received. With loop A, how­ever, received voice frames with non-cor­rectable class 1a errors are not returned
but marked as erased frames. This is pos­sible because GSM voice transmission
is protected by bits so that bit errors
can be corrected. Depending on their sig­nificance, the protection bits are divided
into the following classes:

• Class 1a bits: very good protection

• Class 1b bits: little protection

• Class 2 bits: no protection

News from Rohde&Schwarz Number 169 (2000/IV)

11

o-

t

g

r

C

r

or

r

C

r

or

C

r

C

r

or

or

00

e

MOBILE RADIO

Test methods

MU2

Pseud

random bi

enerato

Bit error

analyze

FIG 1 The GSM mobile can send the received data stream back to the tester via different loops:
before (loop C) or after the channel decoder (loops A, B, D, E and F). The diagram shows switch
positions for loops A, B, D, E and F. All switches are set opposite for loop type C

In the case of erased frames the mobile
sends back a voice frame consisting
entirely of zeroes. On receiving such a
voice frame, the tester increments the
FER (frame error rate) counter. With
this type of loop therefore only voice
frames with a certain minimum quality
are considered in the BER. This explains
the singular effect occurring with this
loop type, namely that with decreasing
receive level the BER suddenly improves.

The lower the level, the more erased

frames occur. So only voice frames with

the smallest number of bit errors will be

considered in the BER measurement.

Loop types D, E and F are used for half­rate connections and are of minor signifi-

cance in production. Like with loop type
A, certain frames are not considered in
the BER measurement (unreliable frames,

erased SID frames and erased valid SID

frames). Such frames are marked by
zeroes.

The larger the numbers of mobile

phones produced, the more important it

becomes to cut testing times. This was
the reason for introducing loop type C.
The mobile sends back the received data

stream without taking it through the

channel decoder. The advantage here is
that, for the same transmission period,

about five times as many bits are avail-

hannel code

hannel decode

Modulat

Demodulat

able for determining bit errors. This type
of loop is not supported by all mobile
phones however.

Further loop types and test modes
have meanwhile been defined for the
new and upcoming mobiles that support
HSCSD (high-speed circuit-switched
data) and GPRS (general packet radio
services).

Measuring GSM BER with
CMU200

Outstanding convenience

The advanced concept of Universal
Radio Communication Tester CMU200
excels not only in transmitter measure­ments (see [2] and [3]) but also in
receiver measurements. BER measure­ment is coupled to a special transmitter
level setting for instance. Using a high
transmitter level it ensures reliable call
setup with the mobile. As soon as the
BER measurement is active, the tester
automatically selects a low transmitter
level and, after completing the measure­ment, returns to the high level.

Data stream transfer from CMU200 to
the mobile is also very convenient for
the user: during the BER measurement,
CMU200 automatically selects a pseudo-

Demodulat

Modulat

Mobile phon

hannel decode

hannel code

random bit sequence. And of course it
opens and closes the test loops in the
mobile automatically. All these features
enable straightforward operation of the
tester.

CMU200 also classifies the bits and pro­vides limit values for each class (FIG 2).

Reference sensitivity

Ten test setups are available for fast and

convenient checking of reference sensi­tivity. Different transmitter levels, test
sequence lengths, BER limits and loop
types can be preset. The setups can then
be called up as test routines, avoiding
tiresome reconfiguration between differ­ent BER measurements.

CMU200 also reduces test time for faulty
mobiles by prematurely terminating the
BER measurement if the required refer­ence sensitivity cannot be achieved.

Absolute receiver sensitivity

For determining the absolute receiver
sensitivity the tester provides an opti­mized routine that allows presetting of
the desired averaging depth for BER
measurement. During the ongoing mea­surement the sliding BER average is mea­sured with the aid of this window. The
user can at the same time directly vary
the transmitter level by means of numeri-

News from Rohde&Schwarz Number 169 (2000/IV)

12