How Does a Fiber Optic Sensor Work and What Does It Detect?
A fiber optic sensor used for factory object detection sends light through a fiber head and measures the light reaching its amplifier. A target either blocks a through-beam path or changes the reflected light. The amplifier then switches an electrical output. Reliability depends on a distinct light change—not simply on fitting a small sensing tip.
What happens between the fiber tip and the PLC?
A factory fiber optic sensor detects an object by monitoring light at a remote sensing point. Its compact head provides the optical access; the powered amplifier turns the received light into a switching decision for a programmable logic controller (PLC).
The fiber carries light; the amplifier processes it
The amplifier contains the light source, usually an LED, and the photodetector. An emitting fiber guides light to the head. A receiving fiber brings light back to the detector. The amplifier compares the received-light signal with a threshold and applies the selected output logic.
In the fiber, a higher-refractive-index core is surrounded by lower-index cladding. Light entering within the fiber's acceptance conditions is guided by total internal reflection. This lets the optical path bend around machine features, within the fiber's routing limits. Transmission is not loss-free: tight bends, damage and poorly prepared ends can reduce the light reaching the receiver.
The passive sensing head normally needs no electrical supply. The amplifier does need power, and its electrical output must match the receiving input. A fiber head is therefore not a complete, standalone PLC sensor.
A light-level display is not a distance measurement
On a digital amplifier, the received-light value and switching threshold are different quantities. The displayed signal may use a model-specific or normalized scale; it is not automatically millimeters, lux or a value comparable across brands. Likewise, “sensing distance” describes an operating arrangement, not a continuous measured-distance output.
A different meaning of “fiber optic sensor.” Fiber Bragg grating (FBG) and distributed sensing systems use changes in optical signals to measure quantities such as temperature, strain or vibration. They require the corresponding interrogation system. This guide concerns photoelectric object detection in machines, not those measurement systems.
How do through-beam and reflective fiber sensors differ?
A through-beam pair looks for a reduction in light when an object enters the beam. A diffuse-reflective head looks for light returned by the object. That difference determines the head arrangement, useful signal states and sensitivity to the target surface.
Through-beam: interrupt the path
Diffuse-reflective: receive a return
Through-beam suits a target that reliably interrupts light
With access to both sides, through-beam sensing is a useful starting point for small opaque parts or parts with changing surface colors. It avoids relying on a bright reflection from the part. However, the smallest feature must still interrupt enough of the effective beam throughout its allowed path. A tiny head does not guarantee an equally tiny detectable object.
Reflective sensing solves one-sided access
A reflective head combines transmitting and receiving fibers at the sensing end. Dark finishes may return less light; a glossy or curved face can direct the reflection away from the receiver. A bright fixture behind the target can also imitate a part. Choose the working gap and head geometry around those competing returns.
Reflective does not mean retro-reflective. A retro-reflective arrangement uses a separate reflector to return light. A limited-reflective or focused head has its own sensing geometry. These are specific head choices, not modes that every ordinary fiber can provide.
Light-ON and Dark-ON describe output logic
Light-ON activates the output in the higher-light condition; Dark-ON activates it in the lower-light condition. For the two simple arrangements above, use the following interpretation:
| Arrangement | What the target does | Output ON for target present |
|---|---|---|
| Through-beam, opaque target | Reduces received light by blocking the path. | Dark-ON, when the blocked state crosses the configured switching boundary. |
| Diffuse-reflective, darker empty background | Increases received light by returning it to the head. | Light-ON, when the return crosses the configured switching boundary. |
PNP or NPN describes the electrical output circuit, not which optical state means “part present.” Teach method, output polarity and receiving-input compatibility need separate checks.
What can a fiber optic sensor actually detect?
It can detect a local presence, absence or contrast change when the selected optical arrangement makes the two required states distinguishable. The target need not be metal, but its material name alone cannot establish detection reliability.
| Machine question | Useful optical feature | What could defeat it? |
|---|---|---|
| Is a pin or small part present? | A repeatable shadow across a fine through-beam path, or a distinct return at a close-range head. | The smallest part misses the beam or interrupts too little light at its worst position. |
| Is an edge or hole aligned? | An edge crosses the beam; an aligned hole opens a light path. | Path wander, burrs or the fixture create the same light state as the intended feature. |
| Is a label or printed mark present? | A label edge or a contrast change at the emitted wavelength. | Artwork, clear backing, gloss or position changes erase the contrast. |
| Is clear glass or film present? | Attenuation, redirected light or reflection in a compatible transparent-object arrangement. | The clearest sample produces too little change, or another surface returns unwanted light. |
| Has liquid reached a point? | A dedicated tube-mounted or contact liquid-level fiber unit. | The tube, liquid, bubbles or wetting conditions differ from the head's specified application. |
These are detection tasks, not automatic inspection capabilities. Detecting one pin does not prove every pin is present. Seeing a printed mark does not read a barcode. Checking one edge does not measure the whole part. Define exactly what one output is meant to confirm before selecting the head.
Why can a taught sensor still miss a part?
Teach records the light conditions presented during setup. It cannot ensure that every later part, background or cable position produces the same signal. A dependable threshold needs separation between the full expected high-light and low-light ranges, not just two convenient readings.
For through-beam interruption, compare the lowest clear-path reading with the highest blocked-path reading. For diffuse reflection, compare the weakest required target return with the strongest unwanted background return. Those are different physical states even though both comparisons look for a usable gap.
A pin is present, but the cable route changes the answer
Illustrative example—not measured product data. Suppose a through-beam pin check produces clear-path readings of 1,600–2,200 and blocked-path readings of 100–250 on one amplifier's display. A conceptual threshold of 800 lies between those ranges.
Now suppose cable movement or contamination reduces the clear-path readings to 180–500 while the blocked readings remain 100–250. The ranges overlap from 180 to 250. At a reading of 200, the amplifier could be seeing either a clear path or a blocked path.
Lowering the threshold cannot remove that ambiguity. First restore a repeatable optical path, then repeat setup. For this simple threshold example, actual operate/reset boundaries and hysteresis must fit inside the separated ranges, with allowance for expected variation. The numbers are not recommended settings or a universal signal-margin requirement.
Hysteresis gives different operate and reset boundaries to reduce chatter near a switching point. It does not make overlapping steady-state signals uniquely identifiable. Similarly, increasing emitted power may lift both the wanted return and an unwanted background; more received light alone is not the objective.
For step-by-step setup rather than the principle, see the separate teach mode and sensitivity guide.
When is a fiber sensor a better fit than a standard photoelectric sensor?
Choose fiber sensing when a small remote head solves a real access, target-feature or environmental constraint. If a self-contained photoelectric sensor fits and provides the required optical result, it may simplify installation by keeping the optics and electronics in one housing.
Small sensing point, accessible electronics
A slim fiber tip can reach into a fixture while the amplifier remains where its display and controls are accessible. Threaded, sleeved, side-view and focused heads address different mounting problems. The advantage is the placement of the optical path—not an automatic improvement in accuracy, range or smallest-object capability.
For a short-range metal-only task, compare an inductive sensor. For a guided part passing a fixed gap, compare a slot sensor. If the task requires reading codes or checking several independent features, a single optical switch may not answer the question; evaluate an appropriate inspection system.
The complete fiber assembly still has limits
Plastic and glass fiber assemblies are available, including special heat-resistant and repeated-flex designs. The limits belong to the whole assembly: core, sheath, tip, adhesives, termination and installation. A passive tip does not make the powered amplifier heatproof, waterproof, suitable for an explosive atmosphere or immune to electrical interference.
Keep a model-rated head in the sensing environment and the amplifier within its own ratings. Distinguish a permitted static bend from repeated movement, and follow the specified unbendable section near the tip. Cut a fiber only when its instructions permit cutting; use the specified tool and preparation method. Looking flexible is not a flex-life specification.
How should you check the system before relying on its output?
Check the optical signal first, then the electrical output and finally the controller's interpretation. That order helps distinguish a target the sensor never detected from a valid signal the machine did not capture.
Prevent unintended motion and follow the machine's isolation procedure before moving heads or changing connections. These checks concern process detection; an ordinary fiber amplifier is not a substitute for a personnel-protection safety device.
Establish a stable optical baseline
- Define one event. Identify the smallest required feature, allowed position, material variants and background—not simply “detect the product.”
- Verify the head–amplifier combination. Use the correct ports, insertion depth, compatible fiber termination and any specified lens. For coaxial heads, follow the identification of emitting and receiving fibers.
- Use the final bracket and cable route. Compare both light states across position limits and allowed machine movement. If readings change when the route changes, investigate routing or damage before re-teaching.
- Check configuration in RUN mode. Confirm the active response mode, teach method, threshold, hysteresis, Light-ON/Dark-ON setting and timers. Record the settings and the observed light ranges.
A faster amplifier mode changes the specification
A real manufacturer example shows why the mode belongs in the record. For the OMRON E3NX-FA11 standard one-output amplifier, the standard-model specification lists operate/reset response times of 250 μs in HS, 1 ms in Stnd and 16 ms in GIGA. These are OMRON model figures, not xsz sensor ratings or a performance comparison between brands.
A part that produces a short optical change at production speed must be checked against the selected mode, not another mode's headline response. Also verify the interval between parts so the system can reset. Catalog range and minimum-object figures have their own head, target and mode conditions; do not combine unrelated best-case values.
Confirm what reaches the controller
If the received-light state changes correctly but the PLC does not register it, check the amplifier's actual output, PNP/NPN compatibility, input common, channel assignment and timer settings. A visible indicator is not proof that every short pulse reached the input.
When a slow test passes but production misses parts, compare the electrical pulse at the input with the controller's required HIGH/LOW durations, filtering and acquisition method. A pulse stretcher can hold an event longer for some applications, but can also merge closely spaced events; it cannot recreate a target the optical system missed.
The useful result is a traceable detection event: a stated target feature creates a distinct light change, the amplifier switches with the intended logic, and the controller captures it throughout the required operating conditions. Keep that baseline with the exact head and amplifier models so later maintenance has something concrete to compare.
Sources and method references
- OMRON — Fiber Sensors: introduction and operating principles. Amplifier/fiber separation, sensing arrangements and specialized head types.
- ifm — Fiber optic amplifier technology. Core/cladding guidance, through-beam and diffuse sensing, and routing effects.
- Banner Engineering — Photoelectric sensor training. Light-operate/dark-operate logic and optical arrangement distinctions.
- Banner Engineering — R55F fiber-optic color-mark sensor manual. Relative signal indication, teach methods and output timing. Its model-specific procedures are not universal setup instructions.
- OMRON — E3NX-FA specifications, standard models. Source for the E3NX-FA11 response-mode example.
- OMRON — Fiber Sensor precautions. Fiber insertion, allowable bends, unbendable sections, alignment and fault checks.
- Banner Engineering — Plastic Fibers datasheet. Individual and bifurcated construction, cut-to-length variants and amplifier-dependent performance.
- VIAVI — What is fiber optic sensing?. The distinction between object-detection fibers and point/distributed measurement systems.
The optical drawings are simplified original explanations. The signal-range example uses assumed values; it is not a customer test or a product specification.