Laser displacement and distance sensors: triangulation, confocal and time-of-flight

What displacement and distance sensors actually measure

A laser displacement or distance sensor reports one number: how far a point on the target is from the sensor's reference plane. From that single measurement you derive everything else: the height of a feature, a step or coplanarity between two surfaces, thickness (with one sensor against a known backstop, or two sensors facing each other), runout, vibration, sag, fill level, or deviation from a nominal setpoint. The sensor does not see the part the way a camera does; it samples distance at a spot, and your application logic turns that distance into a pass/fail or a process value.

Vocabulary matters when you spec a unit. 'Distance' usually implies absolute range to a target that may be far away, from centimetres to many metres. 'Displacement' implies high-resolution relative change within a narrow window close to the sensor, typically micrometres over a few millimetres or centimetres. The same physics appears in both, but the optics, electronics and price are tuned very differently, so a long-range distance sensor and a short-range displacement sensor can both be 'laser sensors' and yet share almost nothing in their performance envelope.

Three terms anchor every datasheet comparison, and you should read them together rather than in isolation. Measuring range is the span over which the sensor is valid, usually offset by a standoff (the 'start of measuring range') you must respect mechanically. Resolution is the smallest change the sensor can detect, almost always quoted at a stated measuring rate. Linearity (or linearity error) is how far the reported value can deviate from truth across the range, which sets the practical accuracy limit. A sensor with excellent resolution but poor linearity is a precise liar: it flags tiny relative changes well, but you cannot trust it for absolute dimensions.

Laser triangulation: micron resolution at short range

Triangulation is the workhorse of high-resolution displacement sensing. A laser diode projects a small spot onto the target; the diffusely reflected light is imaged through a receiver lens onto a position-sensitive detector, historically a PSD and now more commonly a CMOS line or array. As the target moves nearer or farther, the imaged spot shifts position on the detector, and the fixed angle between the emitted beam and the receiver optics converts that lateral shift into a distance. Because a small change in distance produces a comparatively large, easily measured shift on the detector, triangulation reaches very high resolution; sub-micron to single-digit microns is realistic in the short-range, high-end class.

The catch is geometric. Resolution and range are coupled through the triangulation angle and the optics, so you cannot have both micron resolution and a large range in one device. Short-range models, spanning a few millimetres to a few tens of millimetres, deliver the finest resolution; longer-range models, spanning hundreds of millimetres, trade resolution away. There is also a mechanical standoff: the part must sit within a defined measuring window, and you must mount the sensor so the receiver's line of sight to the spot is never blocked. That occlusion constraint is real around walls, bores and recesses, where the reflected beam can be shadowed even though the laser still reaches the target.

Triangulation is sensitive to the target because it assumes diffuse reflection. Matte surfaces behave well. Highly specular (mirror-like) surfaces reflect the beam in a single direction, so depending on tilt the detector is either starved of light or hit by a strong glint that saturates it. Steep tilt angles, sharp edges, color and brightness changes, and translucent materials all distort the spot shape and shift its apparent centre, biasing the reading. Good sensors counter this with real-time exposure control, but you should still validate on the actual material and orientation. The classic failure mode is a flawless result on a calibration block followed by drift and noise on a glossy, curved or dark production part.

Use triangulation for height and step checks, coplanarity, thickness, glue-bead height, weld-seam tracking, electronics and connector inspection, vibration and runout: anywhere you need micron-class precision over a short reach and the surface is reasonably diffuse and optically accessible.

• Strengths: excellent resolution-to-cost ratio at short range, fast measuring rates, compact and mature.
• Watch-outs: specular and very dark surfaces, steep tilt, edges and bores that occlude the receiver, and the range-versus-resolution coupling.
• Typical envelope: spans from a few mm to a few hundred mm; resolution from sub-micron at short range to tens of microns at long range.

Confocal chromatic sensing: precision on shiny and transparent targets

Confocal chromatic sensing handles the cases triangulation struggles with. White light is passed through an optic that deliberately introduces strong chromatic aberration, so each wavelength comes to focus at a slightly different distance. Only the wavelength in focus on the target surface returns efficiently through a confocal aperture to a spectrometer; the sensor reads which wavelength came back and maps that color directly to distance. The emitting and receiving paths are coaxial, so there is no triangulation angle and no separate line of sight that can be occluded.

Two properties make confocal special. First, the coaxial geometry and very small spot tolerate shiny, polished and steeply curved surfaces far better than triangulation, so mirror-finish metal, glass, ceramics, lenses and wafers are natural fits. Second, on transparent and multilayer materials the spectrometer can resolve several return peaks at once, letting a confocal sensor measure the thickness of glass, film, adhesive or coating layers from one side by reading the front and back surface distances simultaneously. Resolution is very high, reaching the sub-micron and, on short-range models, nanometre region, with a small spot that resolves fine features.

The trade-offs are range, speed and price. Confocal measuring ranges are short, typically from fractions of a millimetre up to tens of millimetres, so it is a precision short-reach technology, not a distance sensor. Many systems split into a passive optical pen connected by fibre to a separate controller, which is excellent for tight or hot mounting locations but adds cost and a fixed calibration relationship between pen and controller. Spectrometer readout also tends to cap the measuring rate below the fastest triangulation sensors. Heavily absorbing, very rough or strongly scattering surfaces weaken the return and can defeat it. When the application is shiny, transparent, multilayer, or simply demands the last micron where triangulation noise is unacceptable, confocal is usually the answer.

• Strengths: handles specular, transparent and tilted surfaces, sub-micron to nanometre resolution, coaxial (no occlusion), single-side layer-thickness measurement.
• Watch-outs: short range only, higher cost, pen-plus-controller architecture, lower measuring rate, weaker on very rough or strongly absorbing surfaces.
• Best for: glass, polished metal, optics, wafers, films, coatings, micro-step and flatness metrology.

Time-of-flight: long range, coarser resolution

Time-of-flight (ToF) sensors measure distance by timing light. Direct (pulsed) ToF clocks the round-trip travel time of a laser pulse; indirect ToF measures the phase shift of an amplitude-modulated beam. Either way the principle is distance = speed of light x time / 2, and because light is fast, ToF naturally addresses long distances, from decimetres to many metres, and far beyond in dedicated laser distance meters, all from one compact head. There is no triangulation angle to constrain the geometry, so range is set by optical power, the modulation or timing scheme and the receiver rather than by a fixed angle.

The price of long range is resolution. Timing a few picoseconds of flight is hard, so general-purpose industrial ToF sensors resolve at the millimetre to sub-millimetre level rather than microns. That is ample for positioning, presence, anti-collision, stack and pallet height, level in large vessels, and long-stroke web or crane applications, but it is the wrong tool for micron tolerances. Accuracy and noise also depend on target reflectivity and the amount of returned light: a dark, distant or steeply angled target returns fewer photons and degrades the reading, which is why datasheets quote different ranges for white versus black targets. Strong ambient light, and highly reflective or transparent obstacles in the beam, can also corrupt a ToF measurement.

Note that interferometric methods reach extreme resolution, but they measure relative displacement rather than absolute range and are specialist metrology instruments, not the rugged, affordable distance sensors most integrators deploy. For everyday automation, treat ToF as the long-reach, robust, modest-resolution option that complements, rather than replaces, triangulation and confocal at the precision end.

• Strengths: long measuring range from one compact head, flexible target distance, good for positioning, level and anti-collision.
• Watch-outs: millimetre-class resolution, reflectivity-dependent range (white versus black), sensitivity to ambient light and specular or transparent targets.
• Best for: long-stroke positioning, pallet and stack height, level in tall tanks, collision avoidance, coarse distance gating.

Resolution vs range vs target: reading the trade-offs

The single most useful mental model is a triangle of resolution, range and target dependence. Push range up and resolution falls, whether through triangulation geometry or ToF timing. Push resolution to the limit and you accept a short range and a small spot, as with confocal or short-range triangulation. And every technology carries a target tax: the published numbers assume a cooperative surface, so your real part, whether dark, glossy, transparent, rough, tilted or moving, shifts the operating point. The honest way to spec is to start from your worst-case target and the tolerance you must hold, then find the technology and range that still leave comfortable margin.

Measuring rate is the fourth axis, and it interacts with the other three. Resolution is normally quoted at a specific sampling rate, and pushing the rate higher gives each sample less light and less averaging, so noise rises and effective resolution falls. A figure that looks impressive at a slow rate can degrade sharply at the rate your line actually needs, so always read resolution and measuring rate together and check the value at your real cycle time. Faster targets also force shorter exposures, which compounds the light-starvation problem on dark or specular surfaces.

Put the axes together and a method usually selects itself. Need microns on a diffuse part over a few millimetres or centimetres: triangulation. Need the same precision on glass, polished metal or a layered film, or need single-side thickness: confocal chromatic. Need metres of reach with millimetre-class accuracy for positioning or level: time-of-flight. The hard cases are the boundaries (a fast, dark, tilted part at the edge of a range), and that is exactly where validating on real material, not a spec line, earns its keep.

Outputs and integration: analog, IO-Link and fieldbus

A measurement is only useful once the controller can read it, so match the output to the architecture rather than defaulting to whatever the sensor ships with. Analog outputs (4-20 mA current or 0-10 V voltage) are the simplest path into a PLC analog input and suit straightforward closed loops and setpoint comparisons. The 4-20 mA current loop is the more robust choice over long cable runs and electrically noisy environments, because the current is immune to voltage drop and a reading of 0 mA cleanly signals a broken wire. The cost is finite resolution: the analog value is reconstructed through a DAC and ADC, and long runs add noise, so you can lose some of the sensor's native precision before the value reaches the controller.

IO-Link carries the measured value as a noise-free digital number over a standard unscreened three-wire cable, and it adds parameters, remote configuration, and diagnostics such as signal strength or contamination warnings. It also makes device replacement painless, because the master can restore settings to a new unit automatically. For higher-rate, deterministic measurement and tight motion synchronisation, look to a native industrial fieldbus or industrial Ethernet (EtherCAT, PROFINET, EtherNet/IP and similar), which preserves full digital resolution and timing. High-end displacement and profile controllers frequently expose Ethernet or a fieldbus precisely so the value never has to survive an analog round trip.

Whatever the output, plan triggering and synchronisation deliberately. Distance and profile measurements are often gated to a part's arrival by an external trigger, and for any moving or rotating measurement the sensor's sampling must be locked to motion, typically via an encoder, so each sample maps to a known position. Getting that timing right is what turns a stream of distance values into a faithful height map or thickness profile.

• Analog 4-20 mA / 0-10 V: simplest PLC integration; prefer 4-20 mA for long runs and noise immunity; finite, noise-limited resolution.
• IO-Link: noise-free digital value over standard cable, plus parameters, diagnostics and automatic device replacement.
• Fieldbus / industrial Ethernet (EtherCAT, PROFINET, EtherNet/IP): full resolution, deterministic timing, best for high-rate and motion-synced measurement.
• Plan external triggering and encoder synchronisation so each sample maps to a known part position.

When a point sensor beats a vision system, and when it does not

A point or line laser sensor and a camera-based vision system solve overlapping but distinct problems, and choosing wrongly wastes money on both ends. A dedicated laser sensor is a purpose-built 1D (or, for a profile sensor, 2D) measuring instrument: it delivers a calibrated distance or profile at high rate, with low latency and a simple analog or digital output, and it needs no image processing pipeline. For fast height, step, gap, runout, vibration, thickness or presence-by-height checks, it is smaller, cheaper, faster to commission and more robust than standing up a full vision station, and its accuracy is specified directly in microns.

A vision system wins when the decision depends on appearance or two-dimensional geometry that a single distance value cannot capture: presence and absence of features across a field, dimensional gauging in X and Y, pattern matching, OCR and code reading, surface and cosmetic defect detection, and any task that benefits from classification rather than a threshold on one number. If you need to know what and where across an area, not just how far at a point, a camera is the right tool, and a deep-learning layer can handle defects that resist rule-based logic.

Two intermediate technologies bridge the gap. A laser line-profile (2D) sensor projects a line, images its deformation and returns a full cross-sectional profile, so you get true height and shape across a width in one shot, and stacking profiles as the part moves builds a 3D surface. Chromatic confocal line sensors extend confocal precision from a point to a line for high-resolution topography on demanding surfaces. The practical rule: if one calibrated distance answers the question, use a point sensor; if you need a profile or surface, use a line or 3D laser sensor; if you need appearance, geometry over an area, or classification, use a camera. The three are complementary, and many cells run a laser sensor for fast metrology alongside a camera for inspection.

• Point or line laser sensor: fast, low-latency, calibrated 1D/2D height, gap, thickness, runout and presence-by-height.
• Vision system: 2D geometry, presence, gauging in X-Y, OCR/codes, and cosmetic or surface defect classification.
• Bridge technologies: laser line-profile (2D) sensors for true cross-sectional profile and 3D; chromatic confocal line sensors for high-resolution topography.
• Common pattern: a laser sensor for metrology and a camera for inspection, working side by side in the same cell.