Sensor for Position Detection: Industrial Options and Selection Rules

Do you need an arrival signal or a position value?

Use a switching sensor when the controller needs to know whether a target has reached a defined point. Use a measuring sensor when it needs the target’s position throughout a working range. An encoder can provide rotary or linear position feedback; it is not simply another kind of end-position switch.

A checkpoint answers “has it arrived?”

A metal flag passing an inductive sensor, a part breaking a light beam and a piston passing a cylinder switch can each produce a discrete event. The output tells the controller that a condition has changed. It does not normally describe how far the target has moved beyond that point.

A measurement answers “where is it now?”

A displacement sensor or encoder reports a position-dependent value. This is useful for variable recipe positions, travel monitoring, dimensional checks or closed-loop motion control. “Continuous” means feedback across the measuring range, not infinitely fast or infinitely fine measurement.

Write the requirement as a machine decision: “confirm this metal stop has arrived,” “measure the jaw opening,” or “report shaft angle after restart.” These lead to different specifications. A measuring sensor with a switching output can also be justified for a single checkpoint if its threshold performance or changeover flexibility is needed.

Which sensing method should you try first?

Start with the physical target and the access available around it. Shortlist a principle that can reliably see the required feature, then compare the exact sensor’s range, repeatability and interface.

When a different principle earns its place

A mechanical limit switch is worth considering when intentional contact is acceptable; allow for actuation force, travel and wear. Capacitive switches can detect nonmetal targets, but material, moisture and buildup can change the signal. They should not be assumed to provide precision coordinates simply because they detect presence.

For continuous measurement, distinguish the measuring span from the distance between the sensor and target. Ultrasonic devices have a near blind zone and target-dependent detection regions. Optical measurement needs a suitable return from the actual surface. A contact probe introduces contact force. Magnetostrictive transducers use a position magnet along a measuring element and can suit linear travel where that arrangement is practical.

Use this page to choose the feedback approach. For an application that only needs to distinguish an occupied station from an empty one, the part-presence selection guide covers that narrower task.

Does a cylinder switch prove that the part is clamped?

No. A magnetic cylinder switch detects the piston magnet near its sensing location. It does not directly measure jaw contact, workpiece seating or clamping force. Select the observed feature around the condition the machine actually needs to establish.

When you need position between the endpoints

Continuous magnetic cylinder transmitters are a separate option from ordinary cylinder switches. Match the transmitter’s usable measuring interval and compatibility to the actuator; fitting the same slot is not enough. An external transmitter may cover only part of a long stroke.

For hydraulic or longer-travel designs, consider a compatible linear transducer and its installation requirements. Do not assume an externally mounted pneumatic switch can see through every cylinder wall or replace an in-cylinder measuring system.

Which specification tells you whether positioning is accurate enough?

No single number does. Compare the required position tolerance with the relevant sensor specifications, their test conditions and the rest of the measurement chain.

• Resolution: the smallest reported or detectable increment, as defined for the device. Fine output steps do not establish small total error.
• Repeatability: the spread when the same position or transition is checked repeatedly under defined conditions. Separate same-direction results from approach-direction effects.
• Accuracy or specified measurement error: how far the indicated position can differ from the reference. Read the manufacturer’s definition and range of applicability.
• Linearity: deviation from a specified straight-line relationship. It is not automatically the complete accuracy specification.
• Hysteresis: for a switching sensor, the separation between operation and reset positions. It is not random variation between repeated approaches.

Allocate the application’s allowable error across measurement, mounting, mechanics and acquisition. Do not simply add every catalogue number: some specifications overlap, while others use different reference conditions. For a tight tolerance, request the specified maximum error and its conditions rather than choosing from resolution alone.

Should you measure the motor, the slide or the workpiece?

Measure the position that matters to the process, or establish how another measured position maps to it. Motor rotation can be converted into slide travel, but that conversion depends on the transmission between them.

Backlash, coupling slip or screw expansion can make slide position differ from the value inferred from motor rotation. A linear encoder at the slide makes that travel observable instead of relying only on the transmission relationship. It does not remove every machine error: movement between the slide, fixture and workpiece can still matter.

For example, a mechanism that returns consistently from one direction but shifts after reversal needs a bidirectional check. Increasing the motor encoder’s count resolution alone will not correct mechanical lost motion. Choose feedback location before paying for finer feedback increments.

Do you need position immediately after a power interruption?

If the controller must recover position without rebuilding a relative count, consider absolute feedback. Decide what “absolute” needs to cover: one revolution, multiple turns or the full linear travel.

Incremental feedback needs a reference strategy

An incremental encoder provides pulses as motion occurs. Quadrature A/B signals can identify direction; an index signal, where provided, identifies a reference within a revolution. The controller needs a valid count and machine reference. Saving a count is not sufficient if movement can occur while the system cannot count it.

Absolute feedback still needs machine-level checks

A single-turn absolute encoder distinguishes positions within one revolution; it does not by itself identify which revolution a multi-turn axis occupies. For multi-turn devices, confirm how turns are retained or detected during power loss, along with battery/service requirements where applicable.

The encoder coordinate must also match the machine zero and mechanical relationship. Replacement, a slipped coupling or changed offsets can invalidate that relationship. Absolute feedback can remove a particular homing requirement; it is not permission to restart motion without the machine’s required checks.

Will the PLC receive the right position at the right time?

Check the complete signal path, not only the sensor connector. A valid sensor output can still be incompatible with the input, scaled incorrectly or acquired too late for the process.

• Switching outputs: match PNP/NPN or other output circuitry to the input diagram and common. NO/NC describes state behavior, not the same thing as source/sink wiring.
• Analog values: match current or voltage input, electrical loading, assigned measuring endpoints and PLC scaling. Define how out-of-range and fault signals are handled.
• Encoder pulses: match signal levels, channel format, counter capability and maximum input frequency. State whether the receiver counts one, two or four edges per cycle.
• Digital position data: verify supported interface, process-data format, update/cycle timing and invalid-data behavior. An IO-Link label alone does not specify the complete timing chain.

Illustrative calculation: detection delay becomes travel

Assume a target travels at a constant 0.5 m/s and the combined delay from physical detection to the controller recording the event is 4 ms. The target travels 0.5 × 0.004 = 0.002 m, or 2 mm, during that delay.

This is distance travelled before registration—not sensor accuracy, random repeatability or the machine’s final stopping distance. Delay variation creates position variation; a fixed offset may only be compensable under stable, verified conditions. Check sensor response, input filtering and acquisition timing separately. For repeated events, the input must also capture the shortest required ON and OFF intervals.

What should a selection trial prove before you commit?

Prove that the chosen arrangement reports the required position condition across the intended operating envelope. A bench LED demonstration or a catalogue range does not establish that.

1. Define the observable feature.Name the target, reference surface, travel direction, required positions and tolerance. State whether you need a checkpoint, a window or a position value.
2. Use the intended geometry.Include the final bracket, target material and size, nearby metal or optical background, and actual cylinder/encoder mounting. Cover allowable misalignment and target variation.
3. Compare against an appropriate reference.Record transitions or readings over repeated cycles and required positions. Include reverse approaches when used. A sensor-to-PLC comparison alone cannot establish true mechanical position.
4. Check operating speed and environment.Include relevant warm-up, load, vibration, contamination and cable movement. Test at the configured response/filter settings, not only a favourable slow mode.
5. Verify controller interpretation and recovery.Confirm polarity, scaling, valid-data handling and restart/reference behaviour. Record the exact model, configuration and acceptance result so a replacement can be checked against the same basis.

Before ordering a sample, send a dimensioned sketch, target details, travel or measuring window, allowable position error, maximum speed, controller/input model and environmental conditions. Those facts usually narrow the choice more effectively than asking for “a high-precision position sensor.”

The right sensor is the one that observes the correct feature and delivers usable evidence to the controller. Start with the required decision, choose the sensing method and reference location, then prove the performance in the installation.