How-tos Concepts Training paths

How to find your top alarms

Wiki: Floor concepts, Downtime

When to use this

You want to know which alarm is costing you the most time, whether it comes from a simple on/off alarm signal, a numbered fault code, or a machine’s own condition reporting, one ranked list either way.

Steps

  1. Click Performance, then Downtime (/downtime). The Top alarms by lost minutes board sits above the reason Pareto, ranked longest total lost time first.
  2. A machine’s own detail page has the same board. Open the machine from Home, Facility Map, or the Downtime scope picker, then look under More evidence.
  3. Each row shows how many times the alarm fired and how many minutes it was active in total. A rare, long alarm and a frequent, brief one can both land near the top, the minute total is what ranks them, not the count.
  4. A chattering tag reads frequently and briefly, the classic won’t stay quiet nuisance alarm. A severity badge (warning or critical) appears when the machine reports one.

Notes

  • The board covers every alarm signal your machine sends, a plain on/off alarm bit, a numbered fault code, or a condition reading a CNC or other controller reports on its own. Nothing to configure per protocol, the platform reads whatever the machine gives it.
  • If a machine has no alarm signal assigned yet, the board says so and points at Tag Discovery, it never guesses which reading might be an alarm.
  • If alarm signals are assigned but quiet in the selected window, the board says so instead of showing a blank chart.
  • The window shown matches the date range picker at the top of the app, same as every other timed board.
  • The autoML pipeline also watches for an alarm that reliably shows up shortly before a particular kind of stop, or before scrap. When it finds a real pattern it surfaces it as an insight on the Insights page, with the pattern (which alarm, how many times, how far ahead) spelled out.