Threshold alarming is the oldest, simplest, and still most important form of alerting in a control system: cross a preset limit and the alarm trips. It is the mechanism that tells an operator a pressure is too high or a tank is too low, and it is the backbone of safe operation. Its very simplicity is its strength - but it needs care to avoid nuisance chatter. This guide explains how threshold alarming works, the standard alarm levels, and the tricks that keep it clean.
Threshold Alarming in one line: Threshold alarming is the technique of triggering an alarm when a measured value crosses a predefined limit. When a reading rises above a high limit or falls below a low limit, the system raises an alarm to alert the operator - a fixed, deterministic rule that fires at a set value.
A threshold alarm compares each new reading against configured limits and changes state when a limit is crossed. Most tags carry multiple levels: a high (Hi) and low (Lo) limit for values entering an abnormal range that warrants attention, and a high-high (HiHi) and low-low (LoLo) limit for a more serious excursion demanding immediate action. This tiered scheme lets a slightly-off condition and a genuinely dangerous one be distinguished and prioritized.
The logic is deterministic and transparent: the alarm depends only on the value and the limit, so it is predictable and easy to audit - qualities that matter for safety and regulatory alarms. That predictability is exactly why threshold alarming remains the foundation of alarm systems even as smarter analytics are layered on top.
A naive threshold alarm has a problem: when a noisy value hovers right at the limit, it flips in and out of alarm repeatedly, flooding the operator with chatter for a single condition. The standard fix is an alarm deadband (hysteresis) - the alarm activates at the limit but does not clear until the value has receded past the limit by a set margin. A tank alarming at 90 percent might not clear until it drops back below 88 percent, so momentary wobble around 90 does not chatter.
An on-delay (or off-delay) timer helps similarly, requiring the condition to persist for a few seconds before the alarm is declared, so brief transients are ignored. Alarming on a smoothed value such as a rolling average rather than the raw signal is another common tactic. All of these serve alarm quality - keeping alarms meaningful so operators trust them, which connects directly to alarm rationalization and flood prevention.
Threshold alarming checks against a fixed number and is blind to anything within limits, so it will not catch a value drifting abnormally while still in range. Anomaly detection and trend deviation judge behavior against history and context, catching subtler problems, but they are less transparent and harder to certify. The two are complementary tiers: fixed thresholds for hard, auditable safety and regulatory limits, and anomaly or deviation methods for earlier, softer warning.
A cloud SCADA like Merobix evaluates threshold alarms on the tags it reads from field devices over Modbus, DNP3, OPC UA, and other protocols, applying Hi/Lo and HiHi/LoLo limits with deadbands and delays, and presenting the resulting alarm state to operators in the browser. The thresholds are simple by design; the platform applies them consistently across every remote site so an out-of-limit condition anywhere surfaces immediately.
It is triggering an alarm when a measured value crosses a predefined limit - above a high limit or below a low limit. It is a fixed, deterministic rule that fires at a set value, and it is the foundational alerting mechanism in control and SCADA systems for keeping operations safe.
They are the more severe tier of threshold limits. Hi and Lo alarms flag a value entering an abnormal range that warrants attention; HiHi and LoLo flag a more serious excursion requiring immediate action. The tiered scheme lets operators distinguish a slightly-off condition from a genuinely dangerous one and prioritize accordingly.
Use an alarm deadband so the alarm does not clear until the value recedes past the limit by a set margin, and an on-delay timer so the condition must persist briefly before the alarm is declared. Alarming on a smoothed value like a rolling average also helps. These keep a single noisy condition from flooding the operator with repeated alarms.
This page references the protocol specifications published by the organizations below. Editions, product capabilities, and documentation change over time - confirm current requirements and specifications directly with the source.
Last reviewed: July 27, 2026. Merobix is not affiliated with, endorsed by, or sponsored by these organizations; their names are used only to identify the standards and products discussed.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.