Inductive proximity sensors: operation, wiring and selection

An inductive proximity sensor detects a metal target without touching it. Typical tasks include confirming a component’s presence, checking a mechanism’s position and detecting passing gear teeth. Most switching models report a target-present/target-absent state; a distance-reading application requires a measurement-capable device. Inductive sensors can also detect a stationary target—movement is not a condition for detection. ^[1]

Inside the sensor: a changing field, not a temperature measurement

In a conventional oscillator-based design, the sensing coil generates a changing electromagnetic field ahead of the active face. A conductive target develops eddy currents, drawing energy from the oscillator and reducing its oscillation amplitude. An evaluation circuit detects the threshold crossing and changes the output. Although electrical losses are involved, the sensor is not measuring the metal’s temperature. ^[1]

That description covers a common implementation, not every internal circuit. Multiple-coil and pulsed inductive designs evaluate the target’s influence differently. Their shared purpose is metal detection, but their construction and operating limits need not be identical. ^[17]

Metal targets, material factors and detection through plastic

Ordinary inductive proximity switches are intended for electrically conductive metal targets. Steel, stainless steel, aluminum, brass and copper need not produce the same operating distance. Target area, thickness and approach direction also matter. Plain, non-metallic plastic, wood or paper normally calls for a different sensing principle. ^{[2] [4]}

A suitable device may detect metal through a thin, non-metallic plastic cover. It is detecting the metal, not the cover, and the plastic thickness consumes part of the available sensor-to-target distance. Metallic coatings or conductive fillers require separate evaluation. Detection through plastic is not a promise of unlimited penetration. ^[2]

A reduction factor expresses the effect of a target material on sensing range; it is not one universal number for every model. Factor 1 sensors are designed to maintain the specified range across different metals. They still have requirements for target dimensions and installation—material compensation does not eliminate those constraints. ^[3]

The benefits of non-contact operation have limits

Unlike a mechanical limit switch, an inductive sensor does not need a target to press an actuator. That avoids contact-related mechanical wear. It does not prevent impact damage, cable fatigue, loose brackets or metal-chip buildup. Non-contact does not mean maintenance-free or indestructible. Leave a mechanical clearance rather than using the active face as a stop. ^[15]

Choose a housing and a mounting arrangement separately

Cylindrical sensors commonly use a threaded barrel and nuts; block-style sensors use model-specific fixing holes or brackets. Establish the active-face position, target approach and fixing dimensions before selecting a replacement. Include the space required by the connector and cable—the sensing face is not the entire enclosure.

A flush-mountable sensor can be embedded in surrounding metal as specified by its manufacturer. A non-flush version needs the prescribed metal-free region around its sensing head. Mounting it too deeply can cause predamping or false switching. Neighboring sensors also need the specified separation. Shielded/unshielded sensing-head construction is a different issue from shielding on the electrical cable. ^[5]

Three-wire and four-wire DC connections

A conventional three-wire DC sensor has two supply conductors and one output conductor. A common color scheme is brown for +V, blue for 0 V and black for the output, but the device diagram takes priority. Match the power supply and the receiving PLC input to the sensor’s voltage and electrical interface. ^[6]

A four-wire model may provide two outputs, often complementary NO and NC signals. That is not guaranteed: designs with two NO outputs or another additional function also exist. Likewise, wire count does not establish the presence of short-circuit or reverse-polarity protection. Confirm the complete model specification. ^[18]

PNP/NPN describes the interface; NO/NC describes the logic

With PNP and NPN outputs, the difference is the side of the load being switched. An active PNP path is +V → sensor output → load → 0 V. An active NPN path is +V → load → sensor output → 0 V. These describe conventional current flow. The circuit’s 0 V reference is not the same concept as protective earth, PE. ^[7]

For a correctly powered sensor using ordinary target-detection logic, NO means OFF without a target and ON with a target. NC reverses that relationship. Thus PNP NO, PNP NC, NPN NO and NPN NC are different combinations. An electronic NC output must not be assumed to remain connected when sensor power disappears, as a resting mechanical NC contact might. ^[8]

Why a two-wire sensor is not a dry-contact replacement

A two-wire sensor operates in series with a compatible load. Its electronics can draw OFF-state leakage current, while an ON-state voltage drop remains across the sensor. The leakage must stay below the load’s activation threshold, and the remaining ON-state voltage must be sufficient to operate that load. Both checks matter when replacing a mechanical switch. ^[9]

For two-wire AC models, also verify minimum load current, residual voltage and the permitted AC/DC operating conditions. An AC sensor is not automatically compatible with a 24 V DC PLC input. Thyristor or triac output arrangements depend on the model. Two conductors do not imply a volt-free output or permission to use any supply voltage. ^[9]

Analog inductive sensors: a measurement function, not just a switch

Analog inductive sensors provide a current or voltage related to position or distance over a stated measuring range. Suitable models may offer 0–10 V or 4–20 mA. Check the reference target, rising/falling characteristic, non-linearity and repeatability. Analog does not mean perfectly linear, and it does not establish a three-wire connection. A switching proximity sensor cannot substitute for a measurement device simply because both detect metal. ^[10]

NAMUR needs the right receiving interface

A NAMUR sensor communicates target state through a low-level current change on a two-wire interface. The interface associated with IEC 60947-5-6 requires a suitable switch amplifier or dedicated NAMUR input rather than an ordinary PNP/NPN input. An appropriate evaluation unit can monitor the current levels for line-break and short-circuit conditions. ^[11]

NAMUR is not, by itself, a hazardous-area approval. An intrinsically safe application requires the appropriate sensor approval, barrier or isolator, cable parameters and verification of the complete circuit. The receiving interface is still necessary in an ordinary environment. Do not connect a NAMUR device to an arbitrary 24 V supply or standard digital input. ^[12]

Sensing-distance terminology: Sn is not a guaranteed installation gap

Sn is the rated operating distance, without allowances for production variation, supply changes or temperature. Sr is the effective distance measured under reference conditions; Su addresses usable distance under specified working conditions; Sa defines the assured operating region. These terms, used in the context of IEC 60947-5-2, answer different questions. ^[13]

For the conventional standard-target assessment, 0 < Sa ≤ 0.81 × Sn. With Sn = 10 mm, the resulting limit is 8.1 mm, not an unconditional guarantee for every metal part. The square steel reference target is 1 mm thick, with a side equal to the larger of the active face’s inscribed-circle diameter and 3 × Sn. Verify the actual target and installation as well. ^[4]

Hysteresis is different from repeatability

Hysteresis separates the operate point on approach from the release point on withdrawal, reducing repeated switching from small movements near the threshold. Repeatability describes variation in the switching point during repeated tests under the same conditions. Consistently switching at one location is not the same as accurately reporting that distance. When a specification is expressed as a percentage, check whether its reference is Sn or Sr. ^[8]

Switching frequency and a gear-tooth example

Switching frequency states the detection cycles per second under defined test conditions; it is not the internal oscillator frequency. There is no universal 3.3 kHz ceiling for inductive sensors. Both slower and faster designs exist; for example, Turck lists options reaching 5 kHz within its inductive range. Assess counting frequency alongside target size, gaps and response times. ^[14]

Worked example: Assume each tooth produces one detection cycle. A 20-tooth wheel at 600 rpm produces f = 20 × 600 / 60 = 200 Hz. That describes event frequency only. Narrow teeth may generate short pulses, and the gaps must remain long enough to distinguish successive events.

On the controller side, check input filtering and whether a high-speed counter is required. Evaluate the appropriate signal edge to count each event once. A speed-monitoring device can interpret the arriving pulses to detect running, underspeed or standstill conditions. The sensor’s rated frequency does not prove that the complete machine will capture every pulse.

Electrical ratings worth reading before connection

Operating voltage gives the permitted supply range. No-load current consumption is the current used by the sensor itself with no output load; it is not power expressed in watts. Continuous load current rates the output during sustained operation, whereas a peak-current rating applies only for its specified duration and conditions. ^[13]

A minimum load current may be required, especially for two-wire devices. Remember the distinction: leakage current belongs to the OFF state; voltage drop belongs to the ON state. Short-circuit protection is not permission to overvoltage or continuously overload an output. Follow the model’s protection and inductive-load suppression requirements. ^{[9] [13]}

IP65, IP67 and chemical resistance are separate specifications

IP65 covers dust-tight construction and the specified water-jet test; IP67 covers dust-tight construction and the specified temporary-immersion test. Neither alone establishes compatibility with oils, solvents, coolants or permanent submersion. Manufacturers may state oil resistance separately from IP67, as seen in model-specific specifications. Check the housing, active face and cable materials against the real environment. ^[16]

Inductive, capacitive or photoelectric?

Inductive detection uses a metal target’s electromagnetic interaction. Capacitive sensors respond to capacitance changes and, with a suitable design, can detect non-metallic targets. Photoelectric sensors evaluate transmitted or reflected light. Rather than asking which is universally better, establish the target material and how it passes the detection point. Compare the sensing methods against that job. ^[2]

Commissioning and fault-finding

An example check sequence: Prevent unintended machine motion and isolate energy before changing connections. Verify the full model code, supply, output and pin assignment. A qualified person can then check the supply and switching response under appropriate test conditions using the actual target. An unloaded voltage reading on a multimeter is not proof of a healthy output; consider the loaded circuit and the state received by the controller together.

Detection with no target present suggests checking for metal chips, an over-embedded head or an incorrect NO/NC interpretation. No detection with the target present calls for checks on distance, target material, dimensions and supply. Errors only at higher speed also call for pulse-width and input-filter checks. These are diagnostic directions, not proof that each symptom has a single cause.

Numbered references open manufacturer documentation. Worked examples explain the selection logic; the selected device’s data sheet and wiring diagram remain essential.