A controller that trips silently helps no one at 3 a.m. FactoryTalk Alarms & Events turns conditions into named, prioritised alarms with a full lifecycle — raise → acknowledge → return-to-normal — and a time-stamped audit trail of exactly what tripped and when. Drive the cell below and work the alarms the way an operator would.
Two alarms watch this cell: an ALMA on the analog Tank_Level (trips at the HI limit) and an ALMD on the digital Motor_01.Fault. Fill the tank past the limit, trip the motor, then acknowledge — and watch each alarm walk the ISA-18.2 states while the event log writes an audit record for every transition.
| Alarm | Type | State | Ack |
|---|
Notice the four states: Normal (mint), Active · unacked (red, blinking on the banner), Active · acked (it's still in alarm, but someone has seen it), and RTN · unacked (amber — the condition cleared on its own but still needs an acknowledgement so it isn't missed). An alarm only goes back to Normal once it is both clear and acknowledged.
🔔 ALMD — digital alarm. One instruction per discrete condition (a fault bit, an E-stop, a door). It carries severity, the in-alarm/acked state, and latches the time it tripped — drop it on the rung next to the logic that detects the condition.
📈 ALMA — analog alarm. Watches a real value against LL / L / H / HH limits with a deadband so a noisy signal near the limit doesn't chatter. One instruction guards a level, a pressure, a temperature across its whole range.
🖥️ Tag-based vs device-based. Modern Logix alarms live in the controller (the ALMD/ALMA tags) and FactoryTalk subscribes to them — no separate alarm database to keep in sync. Older setups defined alarms up in the HMI's FactoryTalk Alarms & Events server instead.
🗂️ Events & the historian. Every raise, ack and return is logged with a timestamp, the operator, and any comment — to the FactoryTalk Alarm & Event log and on to a historian. That record is your audit trail and your post-mortem.
Severity (1–1000) drives sort order and colour on the banner and summary; shelving lets an operator silence a known-nuisance alarm for a set time without disabling it; and latched alarms hold until acknowledged even if the condition flickers. Good alarm design is mostly fewer, better alarms — rationalised so every one means "do something now."
An alarm's condition clears on its own before anyone acknowledges it. In an ISA-18.2 lifecycle, what state is it in?