Sensor selection decides more machine uptime than almost any other component choice — a sensor in the wrong technology for its environment fails weekly, while the right one runs for a decade. Here is the practical decision guide across the three workhorse technologies, with the environmental factors that actually kill sensors in the field.
The three technologies in one view
| Inductive | Capacitive | Photoelectric | |
|---|---|---|---|
| Detects | Metals only | Almost anything (incl. liquids, granules, through walls) | Almost anything with optical contrast |
| Typical range | mm-scale (to ~60 mm specials) | mm-scale, adjustable | cm to tens of metres |
| Immune to dirt? | Excellent | Moderate — buildup can false-trigger | Weakest — optics must stay clean |
| Cost | Lowest | Low–mid | Mid–high |
| Classic use | Metal position, speed, end-stops | Level in vessels, non-metal presence | Long-range detection, small objects, colour/contrast |
Choose inductive when you can
If the target is metal and the range is short, inductive wins on every axis: sealed, no optics, indifferent to dust, oil, and light. Selection details that matter: flush vs non-flush mounting (flush allows embedding in metal at reduced range), correction factors for non-ferrous targets (aluminium and brass reduce effective range on standard sensors), switching frequency for fast targets, and output type (PNP dominates EU panels; verify before ordering).
Capacitive for what inductive cannot see
Level detection through plastic vessel walls, powder and granule presence, non-metal targets — capacitive territory. The sensitivity adjustment is both the strength and the weakness: set it marginal and humidity, condensation, or product buildup will false-trigger. Mount away from wash-down splash, set sensitivity with real product at real conditions, and prefer models with visible switching-state LEDs — diagnosing an intermittent capacitive sensor without one is misery.
Photoelectric: capability with housekeeping
- Through-beam — longest range, most reliable detection, but two devices to mount, wire, and align.
- Retro-reflective — one device plus a reflector; polarised versions ignore shiny targets pretending to be the reflector.
- Diffuse — easiest install, shortest range; background-suppression variants solve the classic “sees the conveyor frame behind the product” failure.
Photoelectrics fail dirty: dust films, fog, and vibration-induced misalignment. In dirty zones, either specify sensors with excess gain reserve and mount them protected, or step up to technologies that do not care — including ultrasonic for hostile optical environments.
The environmental checklist that predicts lifespan
- Wash-down or coolant → IP67/IP69K rating and chemically compatible housing (stainless bodies for aggressive media).
- Welding zones → weld-field-immune inductives with spatter-resistant coatings.
- Temperature extremes and vibration → check rated ranges and use proper mounting brackets, not cable-tie engineering.
- EMC-noisy panels (drives!) → shielded cable, separation from motor cables — sensor “ghost switching” is usually coupling, not the sensor.
PLC North stocks tested sensors from SICK, ifm, wenglor, and other major brands across all three technologies. Send us the failed sensor’s type code — we will match it, or suggest the variant that will stop it failing. Browse our tested sensors.