Installation guide

Estimating Network Performance
890 USE 100 00
99
join and does not do so, approximately 2 ms must elapse before its
absence can be assumed and the token passed to the next node.
The worst-case event occurs when node 64 wants to join at the moment
the node that currently holds the token has a message count of 1 and a
next-try address of 1 also. A minimum of 64 x 64 messages must pass
before node 64 can be invited to join.
The average latency for this case is approximately 6 ... 7 s. Worst-case
time is approximately 15 s. Because the actual timing is beyond the
direct control of the application, the network planner should provide for
worst-case timing to handle the event of a new node joining the ring.
3.19.1 Adding or Deleting Nodes
When you plan your network application you should provide adequate
safeguards for the effects of nodes dropping out and rejoining the
network ring. You can provide programming in each node’s application
that will safely suspend the node’s activity, or that causes an orderly
shutdown of the processes controlled by the node.
You should consider the inclusion of a dedicated node that monitors the
network activity and reports on the status of your application. This can
assist you in determining the origin of a condition in which several
nodes are programmed to shut down if proper data is not received from
one or more other nodes. With the monitoring node, you can more
easily identify the node which started the shutdown sequence.
If you use one or more spare (open) drop cables and connectors at
selected sites as points for temporarily connecting a network
monitoring device, you should be aware of the effects on network timing
when you add or delete that device on an active network. Adding the
new node causes an increase in the network’s token rotation time,
reducing the overall data throughput. Deleting the node causes
dropout latency times.
Caution: Before you connect or disconnect any device on an
active network, you should be aware of its effect on network
timing. Use the formulas in this chapter to estimate your
network’ s token rotation time, data throughput, and node
dropout latency time. Apply sufficient margin to your timing
to provide for worst-case conditions, such as multiple nodes
leaving or rejoining the network due to area power faults.