Long before software HMIs, plants used panels of illuminated windows and a horn to tell operators something was wrong, and those panels followed a strict choreography. An annunciation sequence is that choreography: the defined series of visual and audible states a point passes through as it goes into alarm, is acknowledged, and eventually returns to normal. The classic sequences described in ISA-18.1 style still shape how alarms behave today, even on modern screens. This guide walks through the states of a typical sequence, explains variations like ringback and manual reset, and connects the logic to how alarms are annunciated in modern SCADA.
Annunciation sequence in one line: An alarm annunciation sequence is the defined set of states a point moves through as it is annunciated, historically standardised in ISA-18.1 style for hardware annunciators. In the basic sequence a point sits normal until it goes abnormal, at which moment it flashes and sounds an audible horn to draw attention; the operator acknowledges, which silences the horn and changes the flashing light to steady on; and when the condition clears the point returns to normal, often after a manual reset. Variations add behaviours such as ringback, which flashes and sounds again when the condition clears to make the operator notice the return to normal.
Every annunciation sequence is built from a handful of states that a point moves between. It begins in the normal state, where the window is dark or unlit and the horn is silent, signalling that the condition it represents is healthy. When the process condition goes abnormal, the point enters the alarm state, and the sequence prescribes a distinct, attention-getting presentation: the window flashes rather than glowing steadily, and an audible horn sounds. The flashing plus horn combination is deliberate, because a new, unacknowledged alarm must be impossible to miss amid a panel of many windows.
The operator's response drives the next transition. Acknowledging the alarm tells the sequence that a person is now aware of it, and the presentation changes to reflect that: the horn is silenced and the flashing light becomes steady on. A steady-on window means the condition is still abnormal but has been seen, which is a fundamentally different message from a flashing one that has not. This is the heart of why annunciation sequences exist, to encode in light and sound the difference between an alarm that is new and one that has been acknowledged, so an operator scanning the panel can instantly tell them apart.
The sequence completes when the condition returns to normal. In the simplest form the window goes dark again once both the condition has cleared and the operator has acknowledged, closing the loop. The exact order of clearing and acknowledging, and whether a separate reset is required, is what distinguishes the standard variations from one another. The underlying idea is constant: a point occupies a well-defined state at every moment, and the flash, the horn, the steady light, and the dark window each carry a specific, unambiguous meaning.
The standard sequences differ mainly in how they handle the moment a condition returns to normal and how much operator action is required. A ringback sequence adds a distinct alert when the condition clears: instead of the window quietly going dark, it flashes and sounds again in a recognisably different way, so the operator is actively told that a point they were watching has returned to normal. Ringback exists because the return to normal can be as significant as the onset, and an operator who does not notice it may keep responding to a problem that has already resolved.
Manual reset variations require an explicit operator action to clear a point even after the condition has returned to normal. Rather than the window extinguishing itself automatically, it holds a state until the operator presses reset, which forces acknowledgment that the excursion happened and was seen through to its end. This is valuable for conditions important enough that a plant does not want them to vanish silently, since the held state and the required reset guarantee the event leaves a trace in the operator's awareness rather than flickering by unnoticed.
These variations are why the classic references describe several lettered sequences rather than one. Some acknowledge and clear automatically once conditions allow, others latch and demand a manual reset, and others add ringback on return to normal, so a plant can match the behaviour to how critical the point is. What they all share is a precise state diagram: for every combination of condition and operator action there is one defined presentation, so the panel never leaves an operator guessing about whether a point is new, acknowledged, cleared, or awaiting reset.
Although these sequences were codified for panels of physical windows, their logic did not disappear when alarms moved onto screens; it was carried into software. A modern HMI or SCADA alarm still flashes when it is new and unacknowledged, still changes to a steady presentation once acknowledged, and still distinguishes an active condition from one that has cleared. The audible horn became a configurable tone, and the illuminated window became a coloured entry in an alarm list or a symbol on a graphic, but the state model underneath is the same annunciation sequence the hardware standard described.
Understanding the sequence explains behaviours operators rely on every day without thinking about their origin. The reason acknowledging an alarm silences the tone and stops the flashing but does not remove the alarm is precisely the acknowledge-to-steady transition of the classic sequence. The reason some systems make a point of announcing a return to normal is a software echo of ringback. The reason certain critical alarms will not disappear from the list until someone deliberately clears them is a manual-reset variation preserved in software. The vocabulary of flash, acknowledge, steady, ringback, and reset remains the accurate way to describe what the screen is doing.
In a cloud SCADA platform such as Merobix, the same sequence logic governs how alarms are presented across screens and how they are pushed to responders in the field, even though there is no physical annunciator anywhere. A new alarm is presented in its attention-getting state and driven out as a notification; acknowledgment, whether at a screen or remotely, moves it to the acknowledged state and stops the reminders; and the return to normal is reflected consistently everywhere the alarm appears. The old annunciator diagram, in other words, is still the backbone of how an alarm moves through its life, whether it lights a window on a panel or a symbol on a phone.
Acknowledgment tells the sequence that a person is now aware of the alarm. It silences the audible horn and changes the flashing light to steady on, marking the point as abnormal-but-seen rather than new-and-unnoticed. It does not clear the alarm; the point stays in the acknowledged state until the underlying condition returns to normal, and in some sequences a manual reset is then required before it goes fully dark.
Ringback is a sequence variation that alerts the operator when a condition returns to normal rather than simply letting the window go dark. When the point clears, it flashes and sounds again in a distinct way so the operator actively notices the return to normal. It exists because the moment a problem resolves can be as important to know about as the moment it began, and a silent clearing can be missed.
Yes. The physical annunciator gave way to screens, but its state logic was carried into modern HMI and SCADA alarming. New alarms still flash, acknowledgment still moves them to a steady acknowledged state, returns to normal are still signalled, and critical alarms can still require a deliberate reset. The flash, acknowledge, steady, ringback, and reset vocabulary remains an accurate description of how software alarms behave.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.