Rohde & Schwarz CMU-K20, Руководство по эксплуатации

Страница: 203 / 646

CMU-K20/K21/K22/K23/K24
Signalling: Receiver Quality Measurements
1115.6088.12 4.125 E-12
Frame type Channel coder Bits per frame
MCS7 448
MCS8 544
MCS9 592
Statistical BER Tests
In a normal bit error rate test, a fixed number of bits is transmitted, leading to a fixed test time for each
BER result. The idea behind statistical testing is to apply probability theory and predict a range for the
BER at an early stage of the measurement. The prediction can be used to stop the measurement if the
probability of the DUT to pass or fail the receiver quality test is large enough (early pass or early fail
decision). Statistical testing can reduce test times considerably, especially if the exact BER of a
receiver is very low or very high.
The general rules for statistical testing are described in the Terminal Conformance Specification 3GPP
TS 34.121. The CMU parameters
(Confidence Settings)
are described on p. 4.138 . The purpose of the
present section is to explain the principle of the test and outline some typical applications.
Performing a
confidence BER
test
A confidence BER test generally involves the following steps:
1. Activate the test: In the
Receiver Quality
menu, select
Application – BER
and
use the
Stop Condition
hotkey associated to the
BER
measurement control
softkey to select the stop condition
Confidence Level.
2. Set the test parameters: Press
BER
for a second time to open the
Receiver
Quality Configuration
menu. In the
Control
tab, scroll to the
BER – Common
Settings – Confidence Settings
section to select the parameters appropriate for
your test (see remainder of this section). Close the configuration menu.
3. View results: Select the measurement control softkey
BER
again and press the
CONT/HALT
key to initiate a single-shot measurement. The result is indicated
in an output field together with the
Specified Class II limit.
Probability
distribution
Confidence BER tests are based on the statistical independence of the bit error
events: The probability of a bit error is equal for each received bit. The exact bit
error rate is given by the limit of the ratio n
e
/n
s
, where n
e
denotes the number of bit
errors and n
s
the number of received bits, and where the number of received bits
tends to infinity. Any real bit error rate test necessarily provides an approximation to
the exact bit error rate.
Given the statistical independence of the bit error events, the probability p for any
number of bit errors n
e
at a given exact BER and a given number of received bits n
s
can be described by a binomial distribution or an appropriate approximation, such
as the Chi Square distribution. Estimates of the exact BER are based on this
distribution.
Confidence
Levels
The distributions p(n
e
, n
s
, BER) provide a confidence interval [BER
low
, BER
high
] for
the exact BER at any n
e
and n
s
. The confidence interval has the following meaning:
•
The probability for the exact BER to be larger than BER
low
is larger than the
lower confidence level.
•
The probability for the exact BER to be smaller than BER
high
is smaller than the
upper confidence level.
深圳德标仪器
135-1095-0799