Sensor for Label Detection: How to Choose the Right Method

Are you detecting a gap on the roll or a label on the product?

A label-gap sensor tells the machine where successive labels lie on the moving backing strip, or release liner. It compares two constructions: label face material plus adhesive plus liner, and liner alone. The useful difference may be optical transmission, capacitance or ultrasonic attenuation—not necessarily visible color.

A downstream sensor answers a different question: did the label reach the bottle, carton or other product? A successful gap pulse cannot prove that the label peeled, transferred, adhered or landed in the right position.

Write the required result first: one trigger per label, missing-label rejection, position measurement, or printed-content verification. This prevents buying a good gap sensor for an inspection task it cannot perform.

Which sensing method should you try for your label material?

For opaque labels on a sufficiently light-transmitting liner, an optical fork is a practical first candidate. When a clear label produces too little optical difference from its backing, compare dedicated capacitive and ultrasonic label sensors. Test the complete construction, including ink, coatings, adhesive and foil—not just an unprinted face-material sample.

Clear-label capability is not universal material compatibility

A capacitive label sensor responds to a change in its electric field caused by the passing material stack. An ultrasonic label sensor compares sound transmission through the stack. These methods can distinguish materials that look almost identical to an ordinary optical sensor, but each still needs sufficient signal separation.

Metal restrictions are model-specific. Banner's SLC1 manual does not recommend metallic inks, foil embossing or metal substrates. Lion Precision describes its LRD6300 differently: it can handle metallic inks but excludes solid metal-foil labels. Neither statement should be extended to every capacitive sensor.

Cavitated BOPP—polypropylene film containing internal voids—is another reason to check the construction. Leuze identifies its hybrid GSX14E approach for difficult, inhomogeneous films including cavitated BOPP. That is a reason to request a trial, not proof that any ultrasonic device will work on every film.

When does a fiber optic sensor help?

A fiber optic sensor is still a photoelectric system: a small fiber head carries light to and from an amplifier. Its main advantage here is mechanical access where a conventional housing cannot fit. It does not automatically solve a transparent-label contrast problem.

Specify the head and amplifier as a pair. Check sensing geometry, effective spot or aperture, selected response mode, fiber routing and bend limits. A compact head is useful only if the installed combination resolves the label and gap reliably.

What if there is no usable gap?

A contrast sensor may trigger from a deliberate printed registration mark. That mark must remain distinguishable from nearby artwork and correspond to the event the machine needs. Continuous or linerless stock without an exposed backing gap needs a different reference; do not assume a liner-based gap sensor can identify an absent feature.

How small and fast can the label gap be?

Check the minimum gap along the sensing lane and direction of travel. Then check the exact sensor's minimum detectable gap and label size. A fast electronic response cannot recover a gap that the sensing field cannot spatially distinguish.

Use the speed while the gap actually crosses the sensor

At approximately constant local web speed, the geometric gap-passage time is:

Gap time (ms) = gap length (mm) ÷ web speed (mm/s) × 1,000

Separate pulse capture from machine response delay

The electrical pulse may not equal the geometric gap time: sensing-field size, switching thresholds and ON/OFF delays affect it. Measure both the shortest ON pulse and shortest OFF interval at the controller input. Then check that input's filtering, update mechanism and capture function.

A sensor LED does not prove that every pulse reached the program. Conversely, capturing every pulse does not prove that the applicator acted at the correct position. At constant speed, an additional delay Δt produces an additional travel distance v × Δt. A repeatable offset and varying timing error need different corrections.

Where should the sensor sit, and how should it be taught?

Choose a stable sensing lane that crosses the intended label edge and gap consistently. Keep the label web controlled through the sensing area, following the model's specified position and clearance. Lock out hazardous motion before adjusting brackets or reaching into the machine.

Check reach, clearance and edge shape separately

Fork depth determines whether the sensing point can reach the chosen lane from the web edge. Fork opening provides physical room for the web. Neither dimension is the minimum detectable gap between successive labels.

Prefer a straight, repeatable edge crossing where the stock allows it. With a curved or angled die-cut edge, sideways web movement can change where that edge intersects the sensing lane. The trigger can shift without an electronic fault. Check the worst lateral position, not only the centered roll.

Teach the intended states, then verify the selected edge

Use the exact model's procedure: some devices need static label/gap samples, some need moving stock, and others adjust automatically. There is no universal button sequence. Where a signal display is available, compare the variation within each state as well as the separation between them.

After teaching, verify whether the output is active during the label or during the gap, and which electrical transition the controller uses. PNP/NPN describes the output interface; it does not identify the physical label edge. Record the active mode, teach settings and wiring in the machine recipe.

When do you need downstream inspection or vision?

A simple presence sensor may be sufficient if an applied label creates a stable difference from the product surface at a known point. Test the unlabeled product too: bottle reflections, packaging graphics or a seam must not imitate the label.

Use position-sensitive inspection or vision when the requirement includes skew, placement, artwork or printed information. KEYENCE's label-inspection guidance distinguishes placement and alignment issues from print legibility and product-data errors. Specify which defect matters before selecting the inspection system.

• Presence: Is a label in the inspection region?
• Placement: Are its offset and angle within the agreed limits?
• Content: Does the printed text or code match the required product or batch?

These are separate acceptance tests. A readable code does not by itself prove correct placement or complete label adhesion. Validate lighting, product orientation, line speed and the reject mechanism using representative good and defective samples.

Why are labels missed or counted twice?

Locate the first wrong event in the chain: label stock → sensor signal → electrical output → controller count → applicator action. Change one factor at a time so the test can distinguish material, mechanical and timing causes.

Illustrative failure pattern: suppose extra output transitions always occur as a foil feature passes, while the real gap still switches correctly. That repeatable location points toward material contrast within the label. Increasing the PLC filter could hide those events but also erase a short genuine gap. First test the sensor signal and material compatibility.

What should a production trial prove before approval?

Approve a documented sensor, material and machine configuration—not a successful hand-fed sample. Agree on the acceptable count error and trigger-position variation before the trial, then record them separately.

1. Challenge the material range. Include the smallest gap, full artwork, clear and metal-containing regions where applicable, different liner lots, and known splices or missing-label sections. Define whether those exceptional sections must be counted, ignored or rejected.
2. Challenge the motion. Test startup, jog, maximum local speed, acceleration, stop/restart and the permitted web wander. Keep guarding and safe operating procedures in place.
3. Capture the whole signal path. Compare known label events with sensor transitions, controller counts and final placement. Include both shortest ON and OFF intervals.
4. Make the result repeatable. Save model and suffix, amplifier/head combination where relevant, mounting dimensions, output edge, mode, teach settings, input configuration, sample identity and test duration or cycle count.

A finite trial with no observed errors is evidence for the tested conditions, not proof of zero future failures. Stop approval if a material state overlaps the gap signal, a valid pulse is lost, or placement exceeds the agreed tolerance. Correct that failure and repeat the affected checks.

For a supplier evaluation, send actual label-and-liner samples together with gap/pitch dimensions, the motion profile, mounting space, controller interface and required result. Ask the supplier to identify the exact tested configuration and limitations. Compare installed cost—including setup, changeover and troubleshooting—not sensor price alone.