Rockwell Automation PLC-5 Fieldbus Solutions for Integrated Architecture User Manual - Manual (Page 94)

Publication 1757-UM006A-EN-P - May 2002
4-8 Configurating the 1757-FIM
ACSYNCINTR This is the period of time between Application Clock synchronization
messages. Application Clock synch messages are used to coordinate
the 'application clock' among the various nodes. The Application
Clock is used by each Node to begin execution of its Function Blocks
at the Scheduled Time.
This is important so that the Function Block will be Done Executing at
the time it is Scheduled to Publish the answers (outputs) of the
Function Block to other nodes on the Fieldbus. On H1, the LAS will
tell the node (through a Compel Data message) when to Publish, but
it is up to the node to schedule block execution at the right time.
If the crystals of the clocks of all the Fieldbus Nodes were exactly the
same, this message would not be necessary, but since a man with 2
watches never knows what time it is, the time must be
re-synchronized.
The clock synchronization method facilitates synchronization of
clocks which run at different speeds very well. It does not do so well
for clocks that speed up and slow down. i.e. consistently slow clocks
work ok, but sporadically slow clocks are difficult.
So the question is how long does it take until the clocks drift far enough
apart for anyone to care?
If it is set to 1 ms, then you waste all your fieldbus messages updating
the clock and never get anything done. If it is set to say 5 hours, then
a clock can drift a long way from the LAS's clock causing “Stale Data”
because the function block did not execute prior to the node being
Compelled to Publish its data (there are other causes of Stale Data
also).
5 seconds is a comprimise - reigning in deviant clocks before they get
too far out of hand, yet not wasting much network bandwidth. You
might increase it if you are in a real pinch for network speed, but
there won’t be a measurable improvement (by removing 1 ms of
traffic out of 5 seconds or 1/5000).
You can decrease it if you think that the clocks are drifting apart prior
to 5 seconds, but most of the cause of clock drift is because of the
jitter in delivery of time sync messages (theory, don't repeat it), so if
anything, spreading the transmission error out over a longer time
would help the synchronization.