SiteIt's peak. The induct on the sorter keeps stopping with a jam that isn't there. The operator clears it, it runs for two minutes, then it stops again. Totes are stacking up behind it. The shift manager has walked over twice.
Nine times out of ten in our experience, the fault is a photocell. And very often the photocell itself is fine. It's out of line, it's dirty, or something shiny is fooling it.
Work out which before anyone orders a new sensor. A new one fitted to the same bent bracket does exactly the same thing.
Know which type you're looking at
There are three kinds you'll meet on most lines, and each fails in its own way.
A through-beam has a sender on one side and a receiver on the other. The product breaks the beam. They see a long way and they're hard to fool, but both halves have to line up.
A retro-reflective has the sender and receiver in one body, aimed at a reflector. The product breaks the beam between the two. They're easy to fit, and they're the ones that shiny product fools.
A diffuse sensor bounces light straight off the product. No reflector. They're handy, but they see the background too, and they care a lot about colour and distance.
If you don't know which one it is, look at the label or the part number. It changes where you look next.
Lock the motion out first
To test a photocell, people wave a hand through the beam. On a running conveyor that hand is a signal. It can start a zone, fire a pusher or release a gate.
HSE's code of practice for PUWER lists sensors and "photoelectric guards" among the inputs to a control system (L22, paragraph 192). HSE's own maintenance guidance is blunt: "Most maintenance should be carried out with the power off" (HSE, maintenance). Isolate the drives and the air, keep the control supply on so you can still see the LEDs, and then test.
Alignment: check the bracket, not the sensor
Most photocells sit on brackets right where totes, cases and pallets hit things. A knock of a few degrees is enough to put the beam on the edge of the reflector. It works when it's cold, then drops out as things warm up or vibrate.
Many sensors have a stability or margin LED. If yours does, use it. Steady means good margin. Flickering or off means the sensor is only just seeing.
To line up a retro-reflective, aim at the middle of the reflector, then move it slowly each way until the LED drops out. Set it halfway between the two points. Do the same up and down. Then lock the bracket and mark it with paint, so the next person knows if it's moved.
Through-beams are the same job, but with two ends to get right. Do the receiver last.
Dirt: clean it before you blame it
Dust, cardboard fibre, shrink-wrap film, condensation, a greasy thumbprint. All of them cut the light. On a dusty line, a lens that looked clean a month ago won't be.
Clean the lens and the reflector with a soft cloth. If it's damp, find out why. A sensor near a dock door or a chiller will mist up when the air changes. A hood or a different spot may cure it for good.
Then check the sensitivity pot. If it's been turned right up to "fix" a dirty lens on another shift, the sensor may now see through thin product, or see the background. Set it back with a clean lens and real product.
Reflections: when shiny things lie
Shrink-wrapped packs, foil, glossy cartons and stainless steel all bounce light. A retro-reflective sensor can see a reflection off the product and think the beam is still clear. So the product goes past and the PLC never sees it.
If the product is shiny, check the sensor is a polarised retro-reflective with the right reflector. A cheap reflector, or a strip of reflective tape stuck on as a repair, can stop the polarising filter doing its job.
Diffuse sensors have the opposite problem. A white wall, a shiny guard or a bright tote behind the product can hold them on. Background suppression types are made for this. If a diffuse sensor keeps seeing things that aren't there, look at what's behind the product.
Watch for sunlight too. A low sun through a dock door can blind a receiver for an hour a day. If your fault only happens late afternoon, that's a clue.
Light-on or dark-on
Most photocells can switch when they see light or when they don't. Swap one and get that setting wrong, and the PLC sees the opposite of what's there. The sensor LED looks fine. The machine still stops. If you've fitted a new one and the fault has flipped, check that setting first. Then check the input at the card, as we cover in finding the missing signal at the PLC.
Light curtains are a different job
A light curtain or a safety photocell is a safety device. It protects people. Don't treat it like the photocell on the conveyor.
Guards and protection devices must be designed and installed so they can't easily be bypassed or disabled (L22). Never move, mute or link out a safety beam to keep production going. If it keeps tripping, fault-find it properly, as we set out in safety relays and light curtains.
The cost of a sensor that "nearly" works
A photocell that drops out once every ten minutes doesn't look like a breakdown. Nobody logs it. But on a sorter or a palletiser it can cost more output than one big stop. Those are the micro-stops nobody logs, and they add up over a shift.
They also wear people down. An operator who resets the same jam forty times a night stops calling it in. So it's worth having someone on shift who can line up a sensor and set a pot properly. If that skill only lives on days, the night shift just resets and hopes. For sorter faults in general, see finding the real cause of sorter jams.
What to do on Monday
- Paint-mark every photocell bracket on your worst conveyor, so a knocked one shows.
- Add lens and reflector cleaning to the weekly PPM on dusty lines.
- Keep spare sensors and the right reflectors together, with the type and setting on the bag.
- Ask each shift which jams they reset most. Start with those sensors.
Paragraph numbers refer to L22, fourth edition (2014, amended 2018).