Installation guide

98 Estimating Network Performance
890 USE 100 00
3.19 Planning for Ring Join T ime
Nodes can be connected to the network while it is active, dynamically
joining into the address sequence. A node that was previously inactive
due to a power-down state can join the active ring upon its power-up.
The network automatically senses the presence of the new node and
begins to include it in the address sequence.
In most cases a node joins under the direct control of the user  for
example, when a new node is connected to the network cable, or when a
connected node has been without power, and is then manually powered
up. This case is illustrated in Figure 40.
MODBUS PLUS NETWORK
CPU
1 20 30 40 64
CPU CPU CPU
50
CPU CPU
NEW
NODE
TO JOIN
Figure 40 Planning for ‘Ring Join’ T ime
The amount of time that is required for a node to join the address
sequence (the ring ) is a function of a message counter and next-try node
address that are maintained mutually by all nodes that are present.
The count and next-try address are passed with the token from node to
node.
For a given node, the counter increments with each message that is
sent by the node. Only when the counter reaches a value of at least 64
does the node invite the node with the next-try address to join, at which
time the message counter is reset to zero. The next-try address is
incremented with each attempt to allow a new node to join. Neither the
counter nor the ‘next-try’ address are accessible to the user application.
For example, consider the network shown in Figure 40. Node 40 has
been absent, and is ready to join now. Regardless of which node holds
the token, only when the network message counter reaches 64 and the
next-try address reaches 40 will the current node invite node 40 to join
the ring. As node 40 joins, all other nodes will add its address to their
member list. However, each time a next-try node address is invited to