Bit Error Rate Test
SMIQ
1125.5555.03
2.376
E-9
PN generator
N
a1
a2
a3
Output
Acc. to standard
PN9
9
4
-
-
non-inverted
ITU-T Rec. O.153 Fascicle IV.4
PN11
11
2
-
-
non-inverted
ITU-T Rec. O.152 Fascicle IV.4
PN15
15
1
-
-
inverted
ITU-T Rec. O.151 Fascicle IV.4
PN16
16
5
3
2
non-inverted
--
PN20
20
3
-
-
non-inverted
ITU-T Rec. O.153 Fascicle IV.4
PN21
21
2
-
-
non-inverted
--
PN23
23
5
-
-
inverted
ITU-T Rec. O.151 Fascicle IV.4
Measurement Result, Accuracy, Measurement Time
Value range
Measurement results for the bit error rate (ie the quotient of error bits and total
bits) are usually found between 10
-2
and 10
-9
. This means that a large number of
bits may have to be checked before a faulty bit is detected. Because of the large
number of bits involved the measurement time is usually very long.
Since 32-bit-wide counters are used for the total bits and the error bits, the
maximum measurement time is 4.29 x 10
9
bits.
Statistics
The BER measurement measures statistical bit errors, ie errors which do not
occur at regular intervals but at random. Although a single measurement
determines the exact number of errors in the measured interval, a statistically
reliable BER can only be obtained when a sufficient number of errors occurs in the
observed interval. Only this ensures that the single BER measurement result
approaches the true error rate with high probability .
Termination
criteria
To keep the measurement time short with low and high bit error rates, two
termination criteria have been defined in the SMIQ for the BER measurement.
•
Criterion 1: Total number of bits
The measurement is terminated when the specific number of data bits is
reached. Due to this criterion the BER measurement is reliably stopped after
the specified number of bits even if no error or only a few errors were detected;
the measurement result is not statistically reliable (few bit errors).
•
Criterion 2: Number of errors
The measurement is terminated when the specified number of bit errors is
detected. With this criterion, the measurement is rapidly terminated when high
bit error rates occur. Since a great number of errors is counted, the
measurement result is statistically reliable.
The two criteria are used together. The criterion which finally yields a valid result is
indicated in the results screen ("Measurement terminated by...").
Interruption of
measurement
At the end of a measurement, the restart of a new one is delayed until the first
measurement result has been queried with
:BERT:RES?
. The resulting brief
measurement interruption is irrelevant because the subsequent measurement will
be synchronized within 24 data bits.
