Key takeaways
A machine sits idle because one small sensor stopped telling the PLC the truth, and now the whole line waits. This guide is for maintenance technicians, maintenance managers, and plant engineers troubleshooting inductive and capacitive proximity sensors on production equipment: the symptoms, the causes in order of likelihood, and a testing sequence that finds the fault without guesswork.
Inductive proximity sensors generate a small electromagnetic field at the sensing face and detect metal targets through the eddy currents the target induces. They detect metal only, and they detect it at very short range.
Capacitive proximity sensors detect a change in the dielectric near the face, so they respond to almost any material: metal, plastic, glass, liquids, powders. That versatility also makes them more sensitive to contamination, moisture, and material variation.
Both types have short rated sensing distances, typically a few millimeters and always specified per sensor in the datasheet. If the failed sensor is optical rather than inductive or capacitive, follow our photoelectric sensor troubleshooting guide instead, since the failure modes differ.
| Symptom | Most likely cause | First check |
|---|---|---|
| No output, ever | No supply voltage, broken cable, wrong output type (NPN/PNP) | Power LED, then voltage at the sensor connector |
| Intermittent output | Cable damage in flex duty, marginal gap, loose connector | Wiggle the cable while watching the LED |
| Output stuck ON | Chip or coolant buildup on the face, damaged face | Clean and inspect the sensing face |
| Detection distance shrinking | Fixture wear, target misalignment, non-standard target material | Measure actual gap vs rated sensing distance |
| Works cold, fails hot | Marginal gap drifting with thermal expansion, aging electronics | Measure the gap at temperature, compare to spec |
Most proximity sensors live inside guarding, near tooling that moves. Lock out and tag out before reaching in, and verify stored energy is released: electrical, pneumatic, hydraulic, gravity-loaded axes, and anything that can still move.
Some checks below, such as reading the LED, the cable wiggle test and measuring voltage at the connector, need control power on. For those, isolate every source of motion first (dump hydraulic and pneumatic pressure, disable the drives, block gravity-loaded axes), keep your hands out of the tooling area, and follow your site's procedure for testing with the lockout temporarily removed.
Never adjust, jumper, or defeat a sensor that serves a safety function, such as a guard interlock or safety-rated position switch. Safety devices are governed by different rules than standard automation sensors: they must be repaired through your safety procedures, never bypassed.
The two LEDs settle it. If the sensor LED switches but the input card LED never does, suspect wiring, the connector, or the input channel itself. If the input LED switches but the program never reacts, suspect addressing or a failed channel, and try moving the wire to a spare input.
Controller-side diagnostics can shortcut this: on Siemens hardware, for example, module faults surface on the CPU status LEDs, and our Siemens S7 SF fault LED troubleshooting guide covers how to read them. Exact LED meanings vary by model and firmware, so confirm in the manufacturer manual.
Three-wire DC sensors come in two output types. PNP (sourcing) switches the positive supply to the input; NPN (sinking) switches the input to 0 V. A PLC input wired for one type will never read the other.
This is the classic replacement failure: the new sensor's own LED works perfectly, the machine still faults, and the part gets condemned as bad out of the box. Check the output type printed on the old sensor before ordering, and verify the wiring convention (commonly brown positive, blue 0 V, black output) against the datasheet.
A sensor that fails every few weeks is not a parts problem, it is an engineering problem: exposed mounting, chip accumulation, fixture wear, or wrong sensor selection. You only see that pattern if every occurrence is logged as a downtime event with a cause code, and if you track MTBF and MTTR per asset.
Chronic offenders then earn real fixes: recessed or bunkered mounting, an air blast on the face, flex-rated cable, or a scheduled clean-and-check on your preventive maintenance schedule. The same event data feeds availability, which is why sensor nuisance trips show up directly in OEE for manufacturing long before anyone tallies the repair hours.
Many proximity sensor faults show up as short, repeated stops that never make it into a manual downtime log. Fabrico is a cloud-based MES and OEE platform for manufacturers, with maintenance management (CMMS) built in: it records every stop with its cause and duration straight from the machine, and maintenance work orders take each loss from detection to fix. If nuisance sensor trips are eating your availability, book a Fabrico demo and see the pattern on your own lines.
A typical 3-wire DC sensor on an M12 connector uses this layout, shown here from Carlo Gavazzi's datasheet for its ICB18 series: pin 1 is brown (+V), pin 3 is blue (0 V), pin 4 is black (the normally open output), and a normally closed version puts its black output on pin 2 instead. Meter pins 1 and 3 for supply. Then, with the sensor connected to its input and while you present the target, meter pin 4 to pin 3 on a PNP sensor, or pin 1 to pin 4 on an NPN sensor. Always confirm against the datasheet of the sensor in your hand.
IEC 60947-5-2 defines the assured operating distance as no more than 0.81 times the rated distance (Sn). Inside it, the sensor is assured to switch under the conditions the standard sets, which allow for unit-to-unit tolerance, supply voltage and temperature. A sensor rated 8 mm on mild steel is therefore assured to about 6.5 mm on steel. On aluminum, with a correction factor of 0.30 to 0.45, that falls to roughly 2 to 3 mm, so a 4 mm gap that worked on a steel target is outside the assured range for an aluminum one.
Verify supply voltage between the positive and 0 V pins at the sensor connector, then measure the output pin while presenting a target: a PNP output should swing to near supply voltage, an NPN output to near 0 V. If the voltage is right and the output never switches, run a swap test with a known-good sensor.
The most common cause on inductive sensors is metal chip or coolant buildup on the sensing face, which the sensor reads as a permanent target. Clean the face and check for a damaged face or a target resting too close; a shorted output or wiring fault comes next.
PNP (sourcing) sensors switch the positive supply to the PLC input; NPN (sinking) sensors switch the input to 0 V. They are not interchangeable on the same input wiring, and installing the wrong type is a frequent cause of a "dead on arrival" replacement.
Rated sensing distance is specified for a standard mild steel target. Non-ferrous metals like aluminum and brass reduce the usable range significantly, so a gap that works for steel can be out of range for aluminum. Apply the correction factor from the datasheet or choose a sensor with more range.
Only if the sensor is a standard automation sensor, your site procedures allow a documented temporary measure, and the machine remains safe without it. Never bypass any sensor that performs a safety function, such as a guard interlock: safety devices must be repaired under your safety procedures, not defeated.
A flaky proximity sensor causes short stops that look random. With downtime tracking software every stop is logged with its cause, and the random stops turn out to be one sensor on one station.
Stop chasing this fault by hand. Fabrico logs each stop against the machine's history, so a repeating fault shows up before it becomes a hard line stop, and the team can open a prioritized work order with the stop itself captured on video wherever a camera covers the line.
Programați o întâlnire individuală cu experții noștri sau înscrieți-vă direct în planul nostru gratuit.
Nu este nevoie de card de credit!