Slave Control; Slave Status Manager; Features Common To All States; Data Poll Response Handling - CleaverBrooks CFC Operation, Service & Parts Manual

CB FALCON LEAD/LAG

Slave Control

The LL master sends each active slave a command and also
performs a slave status read for each known slave device. It
also sends a Master status broadcast that is heard by all
slaves.
There are 5 commands that might be sent:
• All slaves are commanded to turn off and remain off.
• The LL master sends message to slaves that are off, to
turn their fans on.
• The LL master suspends operation which request a burner
to recycle and remain in Standby if it has not yet opened its
main valve (e.g. it is in Prepurge or PFEP) but to keep
firing if it has reached MFEP or Run. This suspend may be
for the fan to be on or off in standby.
This message is used to abort the startup of a slave that is
not yet firing (because demand went away just before it
was firing), but to keep it on if it actually is firing (the LL
master will discover what happened in a subsequent status
response).
The LL master also sends this message to a slave that is
OnLeave. (This ensures that if the slave is firing when it
returns to LL master control, it will stay that way until the
master has decided whether to use it; or conversely, if the
slave stops firing for some reason that it will not start up
again until the LL master has requested this.
In either case, the command will be to turn on the off cycle
fan if any other slave burners are firing, or to turn the fan off
if the slave is the only slave that might (or might not) be
firing.
• The LL master sends message to turn the burner on and to
assign the burner's firing rate.
If the commanded modulation rate is less than the burner's
minimum modulation rate, then the burner should always
operate at its minimum rate.

Slave Status Manager

The LL master keeps track of slave status for each unit that is
enabled as a slave device. The slave status manager
operates internally for each slave device (up to 8).
There is a table entry for each device containing the following
data:
• SlaveState:
— Unknown - indicates the table entry is unused and
empty
— Available - indicates the slave is OK and ready to use,
but is not
— currently firing as a slave
— AddStage - stage is getting ready to fire
— SuspendStage - stage was getting ready but is not
needed
— Firing - indicates the slave is currently firing
— OnLeave - indicates the slave is operating for some
other demand source within it that has higher priority
than slave demand.
— Disabled - indicates the slave is locked out or disabled
in some way
— Recovering - indicates the slave is in a time delay to
ensure that it is
OK before it is again considered to be available.
• RecoveryTime: Saves how long the slave must be OK to
recover.
• RecoveryTimer: Used to measure the slaves recovery
time
• RecoveryLimitTimer: Enforces a maximum slave
recovery time
• DataPollFaultCounter: Used to tolerate momentary
communication problems and to act on these if they are
excessive.
• StatusReadFaultCounter: Used to tolerate momentary
communication problems and to act on these if they are
excessive.
• AbnormalFaultCounter: Used to tolerate momentary
abnormality
• StagingOrder: Used to record the stage-on order, for use
by the sequencer when it needs to drop a stage.
• Storage for each item described in the Periodic data polling
section
• Storage for each item described in the Slave status read
response section
• Slave Command - the command word from the master to
the slave.

Features common to all states

• Whenever a slave device is not in an expected condition
then a recovery function is used to set up timers to give a
faulty slave:
— minimum time that it must appear to be OK, and
— limit how long a slave has to recover from any error.
• If the slave status read was bad then the slave's
FaultCounter is incremented and if it to reaches the fault
value tries, then a recovery action is invoked.
This action does nothing else if the status read was Bad.
If the slave status read was OK then the status function
puts the slave read response data in a slave status table.
If a transition to another state is indicated then the SlaveS-
tate is simply set to the indicated state.

Data poll response handling

VALID RESPONSE MESSAGE
When a slave responds with a properly formatted message
it is examined to see if Slave enable value is "Enable for
Master".
• If the "Enable for Master" value is not present then the
slave status table is checked and if the slave is not in
the table then the message is ignored (this is normal).
However if the slave is in the table then the message is
stored as usual and the slave will invoke the action as a
disabled slave and cause recovery action to occur.
• If the "Enable for Master" value is present then the slave
status table is checked and if the slave is not in the table
then the slave data is stored in an empty position in the
table. However if the slave is in the table then the
message is stored as usual (this is the normal case).
INVALID RESPONSE OR NO RESPONSE
When a CB Falcon responds to a data poll with an improp-
erly formatted message or it does not respond then the
slave status table is checked and:
If the polled slave device is in the table then the Data Poll
Fault is noted. If this causes a fault counter to exceed the
fault value then the SetRecovering handling is invoked.
44
750-322