Selecting a laser displacement sensor by the smallest resolution number in a catalogue is a poor way to specify an inline measurement. A factory in Thailand measuring one point of height, step, runout or position needs to define the workpiece surface, working distance, installation geometry, measurement time, fixture, PLC interface and acceptance test together. This guide turns a component comparison into a practical decision about whether a repeatable measurement point can be built on a production machine.
For the wider choice among CMM, vision and inline methods, see our dimensional inspection automation guide. This article is about single-point, triangulation-based laser displacement measurement. The discipline of testing candidate sensors against actual media and installation conditions also appears in our capacitive level sensor selection guide, although the physics and application are different.
Define what the number means before choosing a model
Is the requirement to monitor the bow of stamped sheet, the step between two moulded surfaces, the height of a part on a conveyor, or the runout of a rotating shaft? These are different measurements. A height needs a defined datum and repeatable seating. A step is the difference between two positions or two sensors. Runout needs a rotation phase and a reference axis. A moving-part position needs a trigger and a time window. The displayed distance from the head is not automatically a drawing dimension.
Write one measurement brief covering material, colour, finish, roughness, curvature, tolerance and decision threshold; whether the part stops; conveyor speed and takt; allowable mounting envelope; heat, vibration, oil and dust; and the machine or system that receives the result. Separate the product acceptance tolerance from a process-control alarm. If the part does not seat in the same place, a stable sensor reading may still fail to represent the intended dimension.
| Application | Raw result | Additional reference required |
|---|---|---|
| Part height | Distance to top face | Datum, fixture seating and zero |
| Step | Difference between two readings | Locations and timing |
| Rotating runout | Range or other feature over a turn | Rotation phase, axis and fixture |
| Moving-part position | Displacement over a time window | Trigger, speed and invalid-sample handling |
| Sheet bow | Difference across locations | Support method, gravity and temperature |
Measuring thickness from one side requires the underside to be fixed to a known datum. If that face can move, consider two-sided measurement or a different method. The datum design determines whether the sensor reading can be converted into the specified product characteristic.
Understand the optical principle and failure modes
A typical triangulation head projects a laser spot onto the workpiece and estimates distance from where returned light reaches the receiver. The visible spot does not prove that a valid measurement has been obtained. A dark surface can return too little light. A mirrorlike surface can reflect light away from the receiver. Transparent or layered materials may return light from several interfaces. Curved parts, deep grooves and edges can change the return path as the part moves.
OMRON’s displacement sensor technical information and KEYENCE’s measurement-system selection guide help frame the method. Neither substitutes for testing the actual parts. Make a sample set containing the hardest colours and finishes, coating variations, oil-film conditions, upper and lower geometry, and more than one production lot. Evaluate both stationary and moving parts. If blue laser is proposed for a challenging metal surface, use the Micro-Epsilon blue-laser information to understand the intended applications, then decide from a controlled comparison on your workpiece. Blue light is not a universal solution.
Select working distance, range and spot geometry as a set
A catalogue reference distance is not simply the desired bracket dimension. The full spread of product tolerance, fixture seating, conveyor variation and adjustment must fit inside the selected head’s measurement range. Mounting at the extreme edge leaves little room for drift or realignment. Draw both the projected beam and receiver path; a clamp, cover or adjacent part must not block either. Include a protective window, air purge and maintenance clearance where needed.
OMRON’s ZP-L specifications list head-specific reference distances, ranges, spot shapes, linearity and repeatability. The 25 mm family, for example, lists a 20–30 mm measurement distance; that figure cannot be transferred to another head. Longer-distance heads have different ranges and spots. Every comparison row should therefore include the exact model, usable distance, target surface and settings.
A small round spot can target a narrow feature but is sensitive to local texture and reflection. A line spot can average some surface detail yet may bridge a groove or edge and combine the wrong surfaces. Draw the illuminated region on the part, enlarged for workpiece positioning tolerance. Check that the spot remains on the intended face throughout the actual motion.
The mounting drawing should show the head datum, optical axis, upper and lower part position, bracket, fixture, cover, adjustment direction and neighbouring equipment. Document any required tilt. Add a repeatable reference dimension for head replacement and a written zero-setting procedure. An informal production-floor adjustment is not a controlled geometry.
Test reflectance, transparent layers and curvature on real parts
Manufacturer specifications are often obtained under particular reference conditions. The notes in the OMRON ZP-L data sheet define linearity against a white diffuse standard target and warn that it can vary by workpiece. Its repeatability entry also specifies a standard target at the reference distance, a 1 ms measurement cycle and 128 averages. These conditions cannot be silently converted into an accuracy guarantee for black plastic or polished metal.
Record the mean, spread, invalid-reading rate, returned-light state, outliers and recovery after reset for each real sample. Include good parts near the middle of the tolerance, boundary samples and the hardest optical condition. Treat later changes to plating, paint or finish as controlled changes. Ask a supplier claiming compatibility to provide the model, setup and raw workpiece evidence.
The OMRON ZX2 features and KEYENCE 1D sensor portfolio can help identify candidates. Application photographs on product pages are starting points, not acceptance evidence for a different factory.
Do not equate resolution with accuracy
Resolution or repeatability usually describes the spread of repeated readings under specified conditions. Linearity concerns deviation from an ideal response along a measurement range. Absolute error of an installed measurement system also includes the reference standard, calibration, bracket, fixture, thermal effects, workpiece surface, signal conversion and positioning. Record each vendor’s exact definition and test conditions next to its figure.
Start from the product tolerance and the guard band needed around the pass/fail boundary. Observe separately the calibration reference, stationary repeatability, variation between parts, moving-part presentation, temperature and electrical noise. A blanket “sensor resolution must be one tenth of tolerance” rule can conceal a surface or fixture failure. Agree the measurement uncertainty budget with the quality team for the actual use. Ask whether a quoted figure belongs to the head alone, the head plus controller, or the installed machine.
| Property | Question for the supplier | Verification |
|---|---|---|
| Resolution/repeatability | Which target, averaging, cycle and distance? | Repeated raw readings at fixed conditions |
| Linearity | Across the whole range or the usable zone? | Compare several heights with a reference |
| Real-part behaviour | Does colour, tilt or lot change it? | Boundary parts and multiple lots |
| Dynamic measurement | What changes with speed and vibration? | Test at production speed |
| Invalid result | How is lost light or out-of-range signalled? | Deliberate shading and range excursions |
A good FAT bench result may change at SAT when machine vibration or lighting is added. Diagnose the source against this table before labelling the sensor itself defective.
Balance sampling and averaging against takt
Fast sensor updates do not automatically produce a valid high-speed inspection. Put on one timeline the duration for which the spot is over the intended surface, trigger jitter, sensor measurement cycle, averaging window, PLC input scan, decision computation and reject-actuator response. If a moving part crosses into another surface during a long averaging window, the smooth result may describe neither surface correctly. Averaging can reduce stationary scatter but costs time resolution.
The OMRON ZP-L specification lists selectable measurement cycles and averaging functions. Choose settings for the workpiece reflectance, tilt and receiver signal rather than copying only the fastest available cycle into a takt calculation. If OMRON ZX-L-N is another candidate, compare its exact model and specified output conditions on the same basis.
Run dynamic trials at minimum, normal and maximum line speeds. Test the start and end of the measurement region, invalid-reading rate and decision latency. If pitch or part length varies, include those combinations. Define whether edge samples are discarded and which central window is used. After changing an average count, recheck both thresholds and timing. A late result attached to the next part ID is a traceability failure as well as a measurement problem.
Design the bracket and machine as part of the measurement system
A thin bracket can move more than the sensor’s catalogue repeatability. A head fixed to a separate frame may measure relative frame vibration rather than only the part. Consider air-conditioning flow, warm-up of nearby machinery, part temperature after washing, and thermal expansion of a long support arm. A sensor’s temperature coefficient does not describe the entire installation. Secure cables so they cannot pull the head; allow access for cleaning and replacement of a protective window.
Introduce a reference part for checks of zero and the relevant operating span. Control its own dimension, storage and calibration. Specify locating features for the bracket and a settings backup so a replacement returns to the same measurement point. Check seating and stiffness before using software compensation to hide a mechanical problem.
Laser class varies by exact model. The OMRON ZP-L specifications, for example, include both Class 1 and Class 2 variants. Check the chosen product’s labelling, instructions, local requirements and beam path during maintenance. Define controls for the installed arrangement rather than making a blanket assumption from the word “laser.”
Define PLC and data I/O before ordering
Does the machine need only discrete pass/fail, a continuous distance, or also validity and signal-state information? Specify the PLC’s NPN or PNP input, analog current or voltage range, communication option, cable length, shielding, scaling, sample timestamp, zero reset and authority for changing product recipes. The available I/O depends on the exact amplifier and option.
OMRON’s ZP-L amplifier specifications distinguish NPN and PNP variants, analog-output variants and a separate IO-Link compatible amplifier. An RFP saying merely “ZP-L output” is therefore incomplete. Map each desired signal to a selected model and PLC channel. For analog signals, define the conversion to millimetres, overrange, broken-wire behaviour and where calibration coefficients live. For digital communication, define status bits, time stamping, loss of connection and recovery.
A stored measurement should connect part ID, time, product recipe, sensor and settings version, measurement point, unit, raw value, decision and fault state. Do not record “measurement unavailable” as a product NG without distinction. Define whether an invalid reading causes a retry, stop or quarantine. Test that a product change cannot leave the previous threshold active.
| Interface item | Contract detail |
|---|---|
| Trigger | Source sensor and PLC event |
| Value | Unit, sign, scaling, rounding and selected instant |
| State | Valid, no reading, out of range, low signal, communication fault |
| Decision | Threshold, guard band, boundary rule, recipe version |
| Identity | Part ID, timestamp, machine and measurement-point IDs |
| Reaction | Stop, quarantine, repeat, alarm, log |
| Recovery | Restart, settings restoration, communication interruption |
Compare sensors with a common workpiece matrix
Demos often use a favourable sample, angle or averaging setting. Supply every candidate with the same workpiece collection, location, movement and decision definition. A vendor may use its recommended settings, but record them and compare their effect on takt and maintenance. Keep failed samples and fault events in the comparison rather than removing them from the average.
Score real-part valid-reading rate, error and scatter over the working range, speed performance, mounting clearance, I/O and local service separately. Ask suppliers about spare-head lead time and support in Thailand; do not invent prices or delivery guarantees. Put catalogue standard-target numbers and your real-workpiece trial in different columns. Two nominal “1 µm” figures may have different test conditions.
| Criterion | Candidate A | Candidate B | Acceptance basis |
|---|---|---|---|
| Validity on real parts | Enter trial data | Enter trial data | Every specified finish and lot |
| Error and spread | Enter trial data | Enter trial data | Margin needed for the decision |
| Invalid readings by speed | Enter trial data | Enter trial data | Fastest production condition |
| Mounting | Check drawing | Check drawing | Clear optics and maintenance access |
| I/O and faults | Enter exact model | Enter exact model | Match the PLC contract |
| Service | Record answer | Record answer | Replacement and recalibration practical |
If neither candidate meets the requirement, revisit the fixture, measurement point or measurement method. Buying the better of two unsuitable heads does not resolve a system-level failure.
What to put in a displacement sensor RFP
Avoid a line saying only “guarantee ±X µm with a laser sensor.” Specify what is measured and under which conditions. List the sample plan, confidentiality, production lots, speed, environment, reference gauge, point location, data format, pass/fail boundary and repeat rule. Define whether the quotation includes the head, amplifier, bracket, cables, protection, air purge, PLC changes, HMI, data storage, training, local tuning, acceptance tests and spares. Compare cost and delivery using actual supplier responses.
A useful RFP clause is: “The supplier shall measure the attached workpieces and limit samples within the stated mounting envelope, line speed and environment; deliver per-part raw data, invalid and fault counts, settings, exact model, range, reference equipment and decision results; and separate standard-target catalogue specifications from real-part results. Deliver drawings, PLC signal list, settings backup and recovery procedure that reproduce the decision after installation.” Fill in project-specific quantities and limits before issue.
Ask how the system handles material, colour and finish changes; what happens on a black or mirrorlike surface with no reading; how head position is recovered after replacement; who may alter thresholds; and who owns the bracket, wiring, PLC logic and quality decision. A single-word “possible” answer should be converted into a model, test and evidence. Clear responsibility matters when SAT exposes a problem spanning sensor maker, integrator and plant.
Laser displacement sensor FAT and SAT
FAT checks the system before shipment at the supplier. SAT checks the installed system in the production environment. FAT should cover real workpieces, fixture, I/O and fault states, not just a catalogue review. SAT adds plant vibration, lighting, temperature, conveyor variation, actual PLC and quality-data connection. Freeze the samples, test methods and acceptance thresholds before the test; the values are project-specific and must be agreed from the part tolerance and risk.
At FAT, check multiple heights across the working range with a reference, repeatability on standard and real targets, invalid readings on difficult surfaces, upper and lower boundary parts, product change, zero reset, settings after power cycle and out-of-range indication. Simulate the trigger and verify that a result connects to the right part ID. Deliver raw readings, settings, test report and a list of failures and retest actions.
At SAT, run the line at low, normal and high speeds. Include stop and restart, nearby-machine vibration, lighting changes, cleaning, temperature shifts and presentation variation. Confirm that the PLC and HMI distinguish product NG from measurement unavailable and that no value shifts to the next workpiece. Simulate head replacement, restore settings and repeat the reference check. Record the post-start monitoring period, owner and threshold-change approval.
Do not test only a continuous stream of good parts. Remove a part, present a dark surface, tilt a reflective sample, go out of range, interrupt communication and recover after power loss where safe. The machine must not release an unmeasured part as good. Use a safe simulation for any injection that would create a hazard on the running line.
Maintain the measurement after commissioning
Lens contamination, loose fasteners, a new nearby machine or a new surface treatment can change results. Record a reference-part check and signal state at shift start. Put approved cleaning materials, post-clean checks, spare-part models, settings backups and replacement steps into work instructions. Set calibration intervals from risk, manufacturer advice and the plant quality system rather than an unexplained default.
Control changes to material, colour, finish, conveyor fixture, bracket, head, amplifier settings, PLC scaling, threshold and averaging. Define which changes trigger a partial FAT or SAT repeat. Measurement cycle and average count can alter both values and latency, so prevent informal adjustment. Review invalid-reading trends to detect contamination and failure early.
Frequently asked questions
Should laser displacement sensors be ranked by the smallest resolution number?
No. Check its test conditions and separately test real-part linearity, spread, invalid results and dynamic latency. Standard-target data cannot establish performance on black or mirrorlike workpieces.
Can a single-point sensor directly measure a drawing dimension?
Only when the datum, seating, temperature and workpiece position establish a valid relationship between distance and dimension. Multiple locations or whole-profile inspection may require another method.
Is a blue laser essential for noncontact measurement?
No. Some difficult surfaces favour particular optical approaches, but compare exact red and blue models on the same material, distance, speed and mounting geometry.
Should an RFP name a model from the start?
Specify a model when compatibility requires it. Otherwise define the workpiece, decision, space, I/O and FAT/SAT evidence, then ask suppliers to propose a supported model.
What evidence should FAT/SAT deliver?
An acceptance matrix, raw data, sample list, exact model and settings, installation and PLC drawings, calibration information, fault-injection results, settings backup, recovery procedure and open-item list.
Conclusion: procure measurement conditions, not only a sensor
A robust selection binds the meaning of the measurement and its datum, actual optical surfaces, working distance and geometry, the definitions behind resolution and linearity, timing and averaging, fixture and environment, I/O and fault handling, and FAT/SAT evidence. Catalogue data narrows candidates. Reproducible results on real workpieces and the installed line make the purchasing decision.
If your Thailand plant is comparing heads, preparing a bracket drawing and PLC signal list, or drafting an RFP and acceptance matrix, you can contact TOMAS TECH. A discussion can start with one troublesome measurement point and a sample of the actual part.
Primary technical references
- OMRON ZP-L Specifications — model-specific range, linearity, repeatability, timing, I/O and conditions.
- OMRON ZX-L-N Specifications — specifications for another series.
- OMRON ZX2 Features — application and series overview.
- OMRON Displacement Sensor Further Information — technical background.
- Micro-Epsilon Blue Laser Sensors — blue-laser applications.
- KEYENCE Laser Displacement Sensors — 1D product families.
- KEYENCE Selecting a Displacement Sensor and Measuring System — method selection.
*Published specifications depend on model, settings and test targets. This guide does not guarantee product performance or prescribe an acceptance limit for any project. Confirm final requirements against current manufacturer documents, real-part tests and the purchase contract.*
