How to Detect Shiny Metal or Reflective Objects with Photoelectric Sensors
For opaque shiny parts, start with through-beam sensing when both sides are accessible, or polarized retro-reflective sensing when a reflector fits opposite. If only one side is available, evaluate background suppression or distance sensing on the actual surface. The key is a dependable change at the receiver—not how bright the object looks.
Which sensing method should you try first?
For opaque-part presence detection, try through-beam sensing first if both sides are accessible. If only one side can be wired but a reflector fits opposite, start with polarized retro-reflective sensing. Both methods let the part interrupt an established light path; the table below covers the other arrangements.
| Available arrangement | Start with | Check the limiting condition |
|---|---|---|
| Emitter and receiver can face each other | Through-beam sensing. The opaque part interrupts light from a separate emitter. | The smallest feature must block enough of the beam. Check holes, edge positions and reflected paths around the part. |
| One wired side; a reflector fits opposite | Polarized retro-reflective sensing with the manufacturer-specified reflector. | Target-reflection rejection, reflector size, minimum range and any glossy wrap or protective film. |
| No receiver or reflector can fit opposite | Background-suppression (BGS) or time-of-flight (ToF) sensing, qualified on the actual part. | Usable return at every allowed angle, target/background separation, spot position and response settings. |
| Fixed position and a repeatable surface return | Ordinary diffuse sensing may be sufficient. | The weakest target return must remain distinguishable from the strongest no-target return. |
These are geometry choices, not interchangeable product names. OMRON’s sensing-method overview explains the underlying paths; the through-beam, retro-reflective and diffuse comparison covers their broader installation trade-offs.
If the task is strictly metal presence at short range, also consider inductive sensing. It uses an electromagnetic interaction rather than reflected light. Metal type, target size and thickness, distance and mounting metal still matter; ifm’s inductive technology overview explains why material response differs between designs.
Why do shiny surfaces cause unstable detection?
Shiny surfaces concentrate reflected light into a narrow direction, so a small change in angle can remove the useful return or send a strong glare return into the receiver. This is specular reflection: visible brightness alone does not tell you whether the sensor can distinguish part-present from part-absent.
A moving reflection can miss the receiver
On a can, tube or curved fitting, the reflecting angle changes as the part moves. A strong return at the center may disappear near a radius or seam. The sensor may switch only briefly or produce several edges from one part.
Record where the signal changes across the actual part. Brushed grain, oil and protective film are separate test conditions, not cosmetic details.
A strong reflection can overwhelm the receiving optics
More light is not always better. A direct glare return can disturb a triangulation sensor’s near/far comparison. A drop in stability near one angle therefore does not automatically mean the signal is too weak.
Pepperl+Fuchs’ BGS guide illustrates this glare mechanism. Check the model’s signal or quality indication before increasing sensitivity.
A third failure is a misleading return. A standard retro-reflective sensor may receive light from the shiny part itself and mistake that for an unobstructed reflector path. This is often called proxing. It requires a different check from a weak diffuse return.
When should you use polarized retro-reflective sensing?
Use polarized retro-reflective sensing when an opaque shiny part can interrupt the sensor-to-reflector path and the specified reflector can fit opposite. It is especially useful when the target’s mirror-like return fools a standard retro-reflective sensor.
With the specified compatible reflector, the returning polarization can pass the receiver filter; many unwanted first-surface reflections are largely rejected. The part can then be detected by interruption. OMRON explains this as mirror-surface rejection (MSR).
Conceptual paths, not mounting or polarization-vector drawings. Filter rejection depends on the actual sensor, reflector and target surface. These are optical states, not electrical ON/OFF labels.
Check the reflector and the finished surface together
Do not substitute a mirror or arbitrary reflective tape. Use the accessory types and distances listed for the sensor; the reflector selection and setup guide explains what to compare.
Polarization is not immunity to every reflective package. Rockwell’s technical guide warns that clear film over a shiny surface can alter the return so a polarized receiver still detects it. Test the finished wrapped part, not just bare metal. See Rockwell’s photoelectric sensing guide, pages 6–7. OMRON also flags instability on some high-gloss and glossy-film targets in its MSR application cautions.
A matte card is detected, but a shiny can is missed
Suppose a standard retro-reflective station gives these observations in a secured, stationary test:
- With no target, the receiver indicates its normal reflector return.
- A matte opaque card fully covering the beam gives a blocked-path indication.
- A shiny can, also fully covering the beam, gives a clear-path indication.
The contrast points toward unwanted target reflection, once incomplete coverage and reflected bypass paths have been excluded. Simply reversing Light-ON/Dark-ON would reverse the logic without making the two optical conditions distinguishable.
Next test: compare a polarized sensor with its specified reflector on the same can positions, including any film or seams. This is a candidate solution—not evidence of production reliability until the complete part travel is tested.
What works when only one-sided sensing is possible?
Evaluate a background-suppression (BGS) or time-of-flight (ToF) sensor on the actual part when neither a receiver nor a reflector can fit opposite. Distance discrimination can separate the part from nearby machinery, but neither label guarantees a usable return from polished metal at every permitted angle.
BGS separates positions, not every kind of reflection
Triangulation-based background suppression determines near/far position from where returned light lands on a receiver. A cutoff can distinguish the intended part from a farther background. OMRON’s distance-setting explanation shows this receiver-position principle.
The practical question is whether the entire part-present range remains separate from the no-part background. Check minimum sensing distance, required target/background separation and behavior on the real surface. A neat cutoff setting cannot rescue a return that disappears at an allowed angle.
ToF is a different measurement method, not a universal upgrade
Time-of-flight sensing derives distance from light’s travel timing, including phase-based implementations. For example, ifm’s OGD technology description explains its phase-based approach and reflective-target applications. Those capabilities belong to the specified product design; they do not establish a guarantee for every ToF sensor.
Request sample results at the required distance and angle limits. Check usable signal quality, the selected response mode and what the output does with a weak, missing or out-of-range measurement. An invalid distance must not silently be interpreted as an acceptable part.
Some models can monitor a taught background or combine distance and return intensity. Use that only as a documented operating mode with a stable reference. Losing a background return can also mean misalignment or contamination, so “no return” is not automatically proof of a part.
How should you set the angle and spot position?
Start with the exact model’s mounting guidance and place the spot on material that remains present throughout the required detection interval. Check the permitted part positions and angles, then keep the mounting position that separates all required states—not simply the brightest setup indication.
There is no universal five-degree mounting rule
For some diffuse or BGS arrangements, a small tilt directs glare away from the receiving optics. For another target, extra tilt may remove the useful return. Follow the sensing head’s optical arrangement, especially if it is designed to collect regular reflection.
Pepperl+Fuchs illustrates roughly 5° off perpendicular to reduce BGS glare in its selection guide, page 2. Banner’s Q4X manual, p/n 181483 Rev. P, notes that 15° tilt is often sufficient for reflective targets, while making direction and magnitude application-dependent. See the Q4X installation guidance, page 15.
These are examples from different guidance and sensor contexts, not competing universal limits. Start from the exact model’s drawing, vary the allowed part angle, then lock the bracket where all required states remain distinguishable.
Before repositioning hardware or changing connections, secure the equipment against unintended motion and follow its isolation procedures. Use the prescribed setup mode and laser precautions where applicable. For more on geometry, see how installation angle changes detection.
Put the spot on a feature that remains present
A fine spot can find a small feature, but it can also cross a hole, seam or highly curved rim. A larger or line-shaped spot may average a surface change while also admitting unwanted background. Choose from the actual spot diagram at the working distance—not just a “laser” label.
For a can, compare a stable sidewall region with the shoulder and open rim. For a ring, aim at material rather than its center opening. Repeat across sideways movement and rotation; the correct spot must still fit the intended feature at the worst position.
Is the sensor failing, or is the PLC missing the pulse?
Compare the measured sensor output with the PLC’s captured signal. If the expected pulse is missing or unstable at the sensor, investigate sensing conditions, response and configuration; if a valid pulse is present but not captured, investigate the interface and controller. An indicator LED alone may not reveal a short interruption.
At 1 m/s, travelling across a 50 mm part width takes 50 ms. But suppose a reflective setup provides a stable return over only 2 mm of that travel: the usable optical window is just 2 ms (2 mm ÷ 1,000 mm/s).
The physical part width therefore does not establish the electrical pulse length. First obtain a continuous, reliable optical window; then check the sensor’s ON/OFF response, filtering and the PLC’s minimum captured HIGH/LOW durations. A pulse stretcher cannot recover a part the sensor never detected, and may hide the gap between consecutive parts.
| Observation | Check next | What the result tells you |
|---|---|---|
| A small tilt changes detection | Compare signal quality across allowed angles, with distance and target position controlled. | Weak return or glare is plausible. Identify which before changing gain. |
| Matte card works; shiny part does not | On a retro-reflective setup, check complete beam coverage, target return and reflected bypass paths. | The receiver may be accepting unwanted light. Evaluate polarization and the specified reflector. |
| One part produces several counts | Correlate output edges with seams, holes, radii and part rotation. | Repeated optical transitions call for a better sensing feature or geometry before adding debounce. |
| Measured output is correct; PLC count is wrong | Check electrical compatibility, input filtering, pulse duration and the actual capture method. | The remaining problem may be in the interface or controller, not reflectivity. |
| Fault follows a lamp, nearby sensor or drive | Under a safe controlled procedure, change one suspected source while holding the optical geometry fixed. | Investigate optical interference or electrical disturbance; do not assume every intermittent fault is glare. |
Check both the part pulse and the gap that resets the count. A headline switching-frequency rating is not the complete production-speed limit; see why high-speed machines miss sensor signals.
What must a production trial prove before approval?
The actual installed combination must distinguish the required part-present, part-absent and gap conditions at production speed, across the permitted finishes and positions. It must also deliver signals the controller captures correctly and meet the process’s defined acceptance criteria. A successful hand-held demonstration only shows that one sample worked in one position.
- Fix the hardware and operating mode.Record the complete model, reflector or fiber head, distance, bracket orientation, teach settings and response configuration. Do not transfer results to a different suffix or accessory without checking its specification.
- Represent the difficult parts and positions.Include the smallest feature, maximum offset and rotation, seams, brushed or polished finishes, oil and any permitted film. Test the real no-part background as well.
- Observe the complete path at production speed.Check entry, the required detection interval, exit and the smallest gap. Record optical diagnostics where available, sensor output and PLC result under the actual lighting and adjacent-sensor conditions.
- Use the machine’s acceptance criteria.For a counting station, confirm the intended count per part, no false counts with no part and reliable reset between parts. Define the trial size, operating conditions and permitted errors from the process requirements; one fault-free pass is not a reliability rate.
The selection is not ready if target-present and target-absent conditions still overlap, valid pulses are not captured, or required surface conditions have not been tested. Change the geometry or sensing method, or obtain exact-model application evidence before approving the station.
For a review with xsz sensor, share photos of the actual part and background, a dimensioned mounting view, speed and minimum gap, and the current model and observed fault. Include shiny, oily or film-covered samples where relevant. Those details support a useful sample trial more than a request for “a sensor for metal.”
This guide concerns process detection. An ordinary photoelectric sensor is not a personnel-protection safeguard; machine safety requires an appropriate safety system and validation.