A false sensor trip is a stop where a sensor reports a problem that is not there, so the operator finds nothing, resets and the line runs again. Find them by counting alarms per sensor and how many ended with no fault found, then test the usual causes in order: dirt, alignment, wiring, product and timing. If the fault will not clear after a reset, the sensor or its wiring may really have failed, so test it and log it as a breakdown.
Safety devices first
This article is about process sensors: photo eyes, proximity switches and level sensors that watch product and machine position. Light curtains, door switches, e-stops and safety scanners have their own rules.
Never bypass, tape or delay a safety device to stop nuisance trips. Bring those to a qualified person, and check the code first with pages like SICK safety scanner error codes or Pilz PNOZ error codes.
How to spot them in your data
False trips look like short stops with a reset and no repair. Often they get logged as a jam, as other or unknown, or not at all, so you need to go hunting.
The best source is the alarm history in the HMI or PLC. Export a week, then count alarms by sensor, and pick the top few.
Then stand at that sensor, or ask operators to mark each alarm for a shift or two. Add one simple question to each trip: was anything actually wrong?
Those 27.3 minutes may not reach your downtime report at all. Short stops like these need the micro-stops treatment: count them, because each one looks too small to matter.
Six causes, in the order to check them
Start with the cheapest test. Most of these take minutes and need no parts, which is why you should try them before replacing the sensor.
| Cause | The clue | The test | The fix |
|---|---|---|---|
| Dirt, dust, film | Worse late in the shift, better after cleaning | Clean it, then count trips | Add to the routine, shield or move it |
| Knocked out of line | Started after a clean or a jam | Check signal strength and bracket | Rigid bracket, marked position |
| Loose wire or plug | Trips when the machine moves or vibrates | With the machine made safe, wiggle the cable while the input is watched | New cable, strain relief |
| Wrong sensor for the product | Only on clear, shiny or dark products | Run the products side by side | A sensor type made for that surface |
| Electrical noise | Trips when a drive or motor starts | Match trips with drive starts | Separate cable routes, proper shielding |
| Timing or settings | Trips at normal gaps or speeds | Trend the input, compare with the fault timer | Adjust timer or sensitivity |
When the sensor has really failed
A dead or stuck sensor behaves differently from a false trip. The fault returns straight after a reset, or the machine will not start, and nothing at the line clears it.
A short test separates the sensor from its wiring. With the machine made safe and a trained person doing the work, block and clear the sensor by hand, and watch the sensor's output LED and the PLC input.
| What you see | Where to look next |
|---|---|
| No power light on the sensor | Fuse, supply and cable |
| Power on, output LED never changes | Sensor, alignment or teach setting |
| LED changes, PLC input does not | Cable, connector or input card |
| Both change, fault stays on | Fault logic, so call the controls owner |
Log this as a breakdown, because a part or a repair was needed, and keep it apart from false trips. Replace like with like, including the output type (PNP or NPN) and connector, then set it up again and check that it switches on real product.
For steps on a specific machine or device, start from the troubleshooting directory.
How to catch one in the act
Intermittent faults hide when someone is watching, so make the machine tell you. Three methods work well on the floor.
- Trend the input. Have your controls person log the sensor bit and the fault together, so you can see the signal just before the trip.
- Film it. A phone on a tripod over the sensor for one shift can show a tipped pack, a loose flap, or a fork truck reflection. Computer vision for OEE covers fixed cameras.
- Watch the sensor's light. Most sensors have an output LED. A flicker while product runs past suggests a marginal setting.
Also check what the fault logic demands. If a jam alarm fires after the eye has been covered for 1.5 seconds, a slower belt or longer pack can trip it with nothing wrong, which is why a new product or speed may start the problem.
One photo eye, start to finish
Example: the labeller infeed eye above trips 64 times a week, 41 of them false, and only with a clear bottle. Cleaning changes little, and the trips do not line up with drive starts.
The team runs clear and amber bottles side by side. Amber never trips, and the output LED flickers on clear ones, so they fit a sensor designed for clear objects.
They then count for a full week before calling it fixed. Trips falling in the first shift prove little, as the problem was intermittent to begin with.
Setting a fault delay safely
A short delay, or filter, can stop a process sensor reacting to a passing gap. It is a fair tool for a process sensor, and it comes with a risk, since a real jam then runs for longer before the line stops.
- Use it only after you have ruled out dirt, alignment and wiring.
- Start with the smallest delay that stops the false trips.
- Check that a real jam still stops the line before it causes damage.
- Have the controls owner make the change, and write it down.
Keep the fix in place
- Mark the good position on the bracket with paint, so people can see when it moves.
- Write the signal level down where a sensor shows one, and add it to a routine check.
- Add a clean-the-eye step to the shift routine, using an autonomous maintenance checklist.
- Keep the same sensor type in stores, so a swap does not bring new behaviour.
When one type keeps failing on several machines, take it to a root cause analysis. Then use mean time between stops to show the gain.
Moves that make it worse
- Replacing the sensor first. A new sensor with the same dirt, the same bracket and the same cable will trip the same way.
- Turning sensitivity up to the maximum. This can hide a dirty lens for a while, then trip on background objects.
- Treating "reset and run" as fine. Each reset is a stop, and they add up to a real number.
- Assuming a real fault is a false one. A sensor that trips because product truly tips is doing its job, so fix the conveyor transfer.
A false trip is over before anyone can look at it. Where cameras are fitted, Fabrico shows video of each stop, streamed on request next to its machine data, so a reset with nothing wrong is easier to see. See how Fabrico uses cameras for OEE.
What people ask next
What is a nuisance trip on a production line?
A nuisance trip is a stop raised by a sensor or device when no real fault exists. The operator finds nothing and resets. It is one common kind of micro-stop.
Why does my photo eye trip for no reason?
The usual causes are a dirty lens, a knocked bracket, a loose connector, a clear or shiny product, or a fault timer set too tight. Check them in that order, watching the sensor's LED as product passes.
How do I know if a stop was a false trip or a real jam?
Ask the operator to mark whether anything was wrong at each reset for a shift or two. If you can, film the sensor as well. Trips with no fault found are your false ones.
Can I add a delay to stop false trips?
On a process sensor, yes, once you have ruled out the physical causes, and with the controls owner making the change. Check that a real jam still stops the line in time. Never do this on a safety device.
Do false trips count in OEE?
Yes. They count as time lost, either as availability or as performance, depending on your stop threshold. Keep the rule consistent across lines.
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: First Response to a Line Stop: Operator Guide