Schweitzer Engineering SEL-351P, Быстрый старт

Страница: 667 / 674

I-2 M
IRRORED
B
ITS
Date Code 990430
SEL-351P Manual Técnico
Enter the synchronization process described below.
The relay will assert ROKx only after successful synchronization as described below and two
consecutive messages pass all of the data checks described above. After ROKx is reasserted,
received data may be delayed while passing through the security counters described below.
Transfer of received data to RMB1x
-
RMB8x is supervised by eight user-programmable
pickup/dropout security counters settable from 1 (allow every occurrence to pass) to at least eight
(require eight consecutive occurrences to pass). The pickup and dropout security count settings
are separate.
A pickup/dropout security counter operates identically to a pickup/dropout timer, except that it is
set in counts of received messages instead of time. An SEL-351P talking to another SEL-351P
sends and receives M
IRRORED
B
ITS
messages four times per power system cycle. Therefore, a
security counter set to two counts will delay a bit by about 1/2 power system cycle. Things
become slightly more complicated when two relays of different processing rates are connected
via M
IRRORED
B
ITS
(for instance, an SEL-321 talking to an SEL-351P). The SEL-321 processes
power system information each 1/8 power system cycle but processes the pickup/dropout security
counters as messages are received. Since the SEL-321 is receiving messages from the SEL-351P,
it will receive a message per 1/4 cycle processing interval. So, a counter set to two will again
delay a bit by about 1/2 cycle. However, in that same example, a security counter set to two on
the SEL-351P will delay a bit by 1/4 cycle, because the SEL-351P is receiving new M
IRRORED
B
ITS
messages each 1/8 cycle from the SEL-321.
Synchronization
When a node detects a communications error, it deasserts ROKx and transmits an attention
message, which includes its TX_ID setting.
When a node receives an attention message, it checks to see if its TX_ID is included.
If its own TX_ID is included and at least one other TX_ID is included, the node transmits data.
If its own TX_ID is not included, the node deasserts ROKx, includes its TX_ID in the attention
message, and transmits the new attention message.
If its own TX_ID is the only TX_ID included, the relay assumes the message is corrupted unless
the loop back mode has been enabled. If loop back is not enabled, the node deasserts ROKx and
transmits the attention message with its TX_ID included. If loop back is enabled, the relay
transmits data.
In summary, when a node detects an error, it transmits attention until it receives an attention with
its own TX_ID included. If three or four relays are connected in a ring topology, then the
attention message will go all the way around the loop, and eventually will be received by the
originating node. It will then be killed and data transmission will resume. This method of
synchronization allows the relays to determine reliably which byte is the first byte of the
message. It also forces mis-synchronized UARTs to become re-synchronized. On the down side,
this method takes down the entire loop for a receive error at any node in the loop. This decreases
availability. It also makes one-way communications impossible.