In the first minutes of a line stop, make it safe, look before you touch anything, note the alarm and what you saw, try only the checks you are approved to do, then call for help at a set time. The order matters, because evidence disappears at the first reset and waiting for the right person can outlast the repair.
Where the minutes really go
People tend to think a long stop means a long repair. Time one and you will often see something else.
A good first response works on the red bars. It takes no technical skill, just a known routine that everybody on the shift has practised.
The routine, step by step
Is anyone hurt, is there smoke, a leak or a guard open? Use the e-stop if needed, and follow your site's lockout rules.
Read the alarm on the screen. Look at where the product stopped and what looks different.
Note the time, the alarm text and what you saw. Take a phone photo if your site allows it.
Trigger the andon or call the team lead, so people know the line is down.
Clear what you are trained to clear, then try one reset.
If it is not running by your limit, or the same alarm returns, call the next person.
Why you look before you touch
The reset wipes out the evidence. Alarms clear, the jammed part gets pulled out, and the person who arrives later sees a machine with nothing wrong.
Thirty seconds of looking is what makes later root cause analysis possible. Ask three questions: what stopped, where was the product, and what changed just before this?
What you must not do
- Reach into a machine that has not been made safe. A stopped machine can start again by itself, so follow your lockout rules every time.
- Bypass a guard, interlock or sensor. This is never part of a first response.
- Keep resetting. One reset is a test, and repeat resets can damage the machine or bury the alarm history.
- Change settings you do not own. Speed and timer changes hide the fault, and the next shift inherits them.
- Restart without checking product. Anything made during the fault may need to be held for quality.
The clock: who to call, and when
Decide this in advance, in a calm moment, and write it on the card. The operator should not have to judge whether a stop is serious enough to bother someone.
| Time stopped | Who acts | What they do |
|---|---|---|
| 0-2 min | Operator | Safe, look, record, signal, approved checks |
| 2 min | Team lead | Comes to the line, helps decide, warns nearby machines |
| 5 min | Technician | Is called with the alarm text and what was seen |
| 15 min | Shift manager | Reviews the plan, the people and the customer orders |
A safety event skips the clock, so call at once. The same applies to a repeat alarm: if the same fault returns after one reset, go straight to the next step.
The idea of stopping and calling at once comes from jidoka, and andon is the usual way to signal it.
Not every stop needs a technician
The routine above fits a machine fault. Some stops have a different first call, and the card should say so.
| What stopped it | First call | Say or do |
|---|---|---|
| Material ran out | Internal logistics | The item and how long until it arrives. See waiting for material. |
| Nobody at a station | Team lead | Which station and which skill. See waiting for operator. |
| Suspect product or a failed check | Quality | Hold the product made since the last good check. |
| Power or air lost on the whole line | Maintenance or utilities | Tell the shift manager too, and keep clear until it is safe. |
Also write down what happens if nobody answers. Name a second person for each role, and give the night shift its own list, since the day list often fails there.
What to tell the person you call
A good call saves a walk back for tools. Four short facts are enough, and they fit on the card.
- Which machine, and which part of it.
- The alarm text or code, read exactly. A troubleshooting page or code list may also help the technician prepare.
- What you saw, and what you already tried.
- What it is doing now: safe, locked out, or waiting.
Tell the neighbours
A stop rarely stays on one machine. Let the machines either side know, because a conveyor full or empty in 3 minutes gives them time to prepare, and upstream and downstream stops explains how it spreads.
If the stop looks long, use the time. A crew standing idle could do cleaning, a material check or a planned job, as long as they can come back at once.
Restarting well
- Confirm that guards are closed and everyone is clear.
- Check product on the line and hold anything suspect.
- Start at low speed where you can, and watch the first few parts or packs.
- Check the first good product, then bring the line up to speed.
- Tag the reason before you move on. Operator downtime logging keeps this short.
Measure the response
Two time stamps show whether your routine works: when the stop started, and when a person who can fix it arrived. A shrinking gap means the first response is improving.
A single repair average can hide this, depending on what you include. How to calculate MTTR splits a stop into stages, so you can see the waiting.
Make it stick
- Practise it. Run a short walk-through at one station in a gemba walk, using a recent stop.
- Train new starters on day one. OEE training for operators shows where it fits.
- Thank people for calling early. If early calls get complaints, people will stop making them.
- Pass it on. Put open stops on the shift handover checklist.
The first response starts with knowing the line has stopped. Fabrico records stops live from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line, so OEE per line and shift shows where the minutes went. See how Fabrico tracks stops.
What people ask next
What should an operator do first when a machine stops?
Check that everyone and everything is safe. Then read the alarm and look at the machine before touching it, note what you saw, and signal the stop. Only after that try the checks you are trained to do.
How many times can an operator reset a machine?
One reset is a sensible rule, then call for help if the same alarm returns. Your site's rules come first. Repeat resets can damage the machine and clear the alarm history.
When should an operator call maintenance?
At a time limit set in advance, such as 5 minutes, and at once for a safety event or a repeat alarm. Base your limit on your own stop data. Shorten it for the bottleneck machine.
What should be recorded during a line stop?
The start time, the machine, the alarm text, what you saw and what you tried. Add the reason before restart or shortly after. The downtime log template has a simple layout.
Should the whole line stop when one machine stops?
On most lines, the other machines keep running until they become blocked or starved, which uses up the buffer. Read blocked and starved to see what happens next. If there is a safety concern, stop whatever your site's rules say to stop, and leave output for later.
Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.
Related guides: Shop Floor Visual Management: A Working Guide · MTTR Reduction Plan: A 12-Week Template to Use