SBI Bus Serializer ASSP Telecom Standard Product Data Sheet
Released
Proprietary and Confidential to PMC-Sierra, Inc., and for its customers’ internal use.
Document No.: PMC-2000168, Issue 5
322
13.10.2
Synchronization
Before being able to monitor the correctness of the PRBS payload, the monitor must
synchronize to the incoming PRBS. The process of synchronization involves synchronizing the
monitoring LFSR to the transmitting LFSR. Once the two are synchronized the monitoring
LFSR is able to generate the next expected PRBS bytes. When receiving sequential PRBS
bytes (STS-12c/VC-4-4c), the LFSR state is determined after receiving 3 PRBS bytes (24 bits
of the sequence). The last 23 of 24 bits (excluding MSB of first received byte) would give the
complete LFSR state. The 8 newly generated LFSR bits after a shift by 8 (last 8 XOR products)
will produce the next expected PRBS byte.
The implemented algorithm requires four PRBS bytes of the same payload to ascertain the
LFSR state. From this recovered LFSR state the next expected PRBS byte is calculated.
Out of Synchronization and Synchronized states are defined for the monitor. While in progress
of synchronizing to the incoming PRBS stream, the monitor is out of synchronization and
remains in this state until the LFSR state is recovered and the state has been verified by
receiving 4 consecutive PRBS bytes without error. The monitor will then change to the
Synchronized State and remains in that state until forced to resynchronize via the RESYNC
register bit or upon receiving 3 consecutive bytes with errors. When forced to resynchronize,
the monitor changes to the Out of Synchronization State and tries to regain synchronization. It
is important to note however that the monitor can falsely synchronize to an all zero pattern. If
inverted PRBS is selected, the monitor can falsely synchronize to an all 1 pattern. It is therefore
recommended that users poll the monitor’s LFSR value after synchronization has been declared,
to confirm that the value is neither all 1s or all 0s.
Upon detecting 3 consecutive PRBS byte errors, the monitor will enter the Out of
Synchronization State and automatically try to resynchronize to the incoming PRBS stream.
Once synchronized to the incoming stream, it will take 4 consecutive non-erred PRBS bytes to
change back into the Synchronized State. The auto synchronization is useful when the input
frame alignment of the monitored stream changes. The realignment will affect the PRBS
sequence causing all input PRBS bytes to mismatch and forcing the need for a
resynchronization of the monitor. The auto resynchronization does this, detecting a burst of
errors and automatically re-synchronizing.
13.10.3
Error Detection and Accumulation
By comparing the received PRBS byte with the calculated PRBS byte, the monitor is able to
detect byte errors in the payload. A byte error is detected on a comparison mismatch of the two
bytes. Only a single byte error is counted regardless of the number of erroneous bits in the byte.
All byte errors are accumulated in a 16 bit byte error counter. The error counter will saturate at
its maximum value of FFFFh, i.e. it will not wrap around to 0000h if further PRBS byte errors
are encountered. The counter is readable via the WPP Monitor Error Count Register or the PPP
Monitor Error Count Register. The error counter is cleared when transferred into the registers
and the accumulation restarts at zero. When reading error counts for concatenated payloads of
STS-3c /STM-1c or STS-12c/STM-3c sizes, it is necessary to read the error count in all slices
(all associated STS-1/STM-0s). For each independent STS-1/STM-0 monitored by a PRGM,
the error count register for each individual STS-1/STM-0 must be read.