Why Does a Sensor Miss Small Parts on a High-Speed Conveyor?

A sensor misses small parts at high speed when any link in the timing chain is slower or larger than the part’s passage demands: the beam or spot is wider than the part (diffuse modes) or the part is smaller than the mode can reliably see; the sensor’s response time or switching frequency cannot complete one ON/OFF cycle within the part’s block time; the PLC input filter or scan time swallows pulses the sensor correctly delivered; or the gap between parts is too short for the chain to reset. The first safe check is arithmetic — compute the part’s block time and gap time from its length, the gap and the conveyor speed, and compare against the whole chain: sensor response, input filter, scan. The math is safe; the machine stays running; only verification near the belt is hands-off.

Key Takeaways

• The chain is only as fast as its slowest link. Sensor response time, PLC input filter and scan time add up; a fast sensor behind a slow input still misses parts. Budget the whole chain, not the sensor alone.
• Block time and gap time are the two numbers. Part length ÷ speed = how long the sensor must stay ON; gap ÷ speed = how long it has to reset. Both must fit inside the chain’s capabilities with margin.
• Small parts can be optically invisible before they are temporally missed. A part smaller than the effective beam of a diffuse sensor, or one that fits inside the switching hysteresis zone, never generates a solid signal at any speed — that is a geometry problem, not a speed problem.
• Speed problems are repeatable, geometry problems are positional. Misses that scale with speed point to the timing chain; misses that depend on part position or orientation point to optics and geometry.
• Faster sensing exists when the chain is the limit. Fiber-optic sensors with fine spots serve very small parts; ToF families with high switching frequencies serve speed — the technology mapping is in the selection guides (object-detection selection guide, fiber-optic selection guide).

Symptom / Likely Cause / Safe Check Table

Safety Boundary

Safe observations and desk work: the timing-budget worksheet (all from dimensions, speeds and specifications); watching indicators from outside the guarded area; reviewing PLC settings from the engineering station.

Qualified-person / procedural work: any measurement or adjustment requiring access at the conveyor (stopped and locked out); input-filter/scan configuration changes under change control; optical cleaning. Never reach over a running conveyor to “test” detection.

Inputs to Collect Before Changing Settings

1. Part dimensions (smallest dimension in beam direction), material, surface, orientation range;
2. Minimum gap between parts (worst case), and whether parts can touch;
3. Conveyor speed range (minimum and maximum to be supported);
4. Sensor model, mode, response time / switching frequency from the data sheet;
5. PLC input filter setting and scan time for the channel;
6. Commissioning record of the same worksheet, if it exists;
7. Miss pattern: speed-correlated, position-correlated or random.

The Timing-Budget Worksheet

Work through five lines; every later decision hangs on them:

1. Block time = smallest part length (in beam direction) ÷ maximum conveyor speed.
2. Gap time = smallest gap between parts ÷ maximum conveyor speed.
3. Sensor capability = response time (or period from switching frequency) for one complete ON/OFF cycle, from the data sheet.
4. Input chain = PLC input filter time + scan time (both from the controller documentation).
5. Verdict = block time must exceed the total of (sensor + input chain) with margin, and gap time must exceed the chain’s reset requirement. If either fails, the chain — not the part — is the cause.

Keep the worksheet with the line’s commissioning records: speed increases and part changes are re-verified in minutes against it.

Diagnostic Sequence (Least Invasive to Specialized)

1. Observation: misses reported at high speed. Hypothesis: timing chain limit. Check: complete the worksheet with current part, gap and speed data. Expected result: block or gap time below the chain total, or with no margin. Next action: identify the failing element (sensor, filter or scan) and address that element specifically.
2. Observation: sensor indicator fires on every part; controller misses some. Hypothesis: input chain (filter/scan) is the failing element. Check: indicator-vs-controller comparison at fault speed; confirm filter/scan settings. Expected result: divergence — indicator complete, tag missing. Next action: correct the input configuration within the logic design; the chain diagnostics are in intermittent-PLC-input troubleshooting guide.
3. Observation: sensor indicator itself misses parts. Hypothesis: sensor response limit or optical geometry. Check: slow the line (if process permits) and observe; if misses vanish at low speed, it is response; if they persist at all speeds at the same positions, it is geometry. Expected result: separation of the two families. Next action: response → faster model (below); geometry → step 4.
4. Observation: misses at all speeds, position-dependent. Hypothesis: part smaller than effective beam, or riding outside it. Check: compare part dimensions with beam/spot size from the data sheet; observe part presentation at slow jog. Expected result: part at or below the beam’s reliable detection size, or offset presentation. Next action: geometry fix (beam position, guides) or sensing choice with a finer spot — fiber optics serve very small targets (fiber-optic selection guide).
5. Observation: problems began after a part or speed change. Hypothesis: new dimensions/speed broke a previously marginal budget. Check: rerun the worksheet with new values vs. commissioning values. Expected result: margin consumed by the change. Next action: restore margin — speed, part presentation or sensing upgrade.
6. Observation: random misses unrelated to speed or position. Hypothesis: contamination, alignment or vibration. Check: standard false-signal sequence (photoelectric false-signal troubleshooting guide). Expected result: environmental cause found. Next action: per that guide.
7. Observation: worksheet passes with margin, misses persist. Hypothesis: intermittent electrical path or noise. Check: qualified-personnel verification; signal-path diagnosis (intermittent-PLC-input troubleshooting guide). Expected result: path fault isolated. Next action: correct the path; the replacement-accessory equivalence discipline applies if cables were touched (cable-or-bracket replacement topic).

Conditional Corrective Actions

• Input chain is the limit → shorten filter time / adjust scan within the logic design; keep change control — faster filters admit more noise, so verify with the real line.
• Sensor response is the limit → select a model with switching frequency matching the worksheet with margin; KJT Sensors’ ToF photoelectric family states switching frequencies up to 1,000 Hz (manufacturer-stated) — verify the exact model.
• Geometry is the limit → finer spot (fiber optic, fiber-optic selection guide), through-beam arrangement for smallest cross-sections, or presentation guides.
• Gap time is the limit → the chain cannot reset between touching/nearly touching parts: increase gap at the infeed, or move to a counting method tolerant of merged parts.
• Marginal after part change → treat as a new application: rerun the selection logic (object-detection selection guide) with the new parts.

Replacement and Reselection Criteria

Reselect when the worksheet shows a structural gap between demand and capability (no settings can close it); when parts are smaller than the current mode’s reliable detection geometry; or when line-speed plans will exceed the current margin soon. Technology mapping: very small targets and confined spaces point to fiber optics (fiber-optic selection guide); mode trade-offs (through-beam vs. retro vs. diffuse) are in photoelectric sensing-mode comparison; moving-target distance sensing is moving-target distance guide; the full selection entry point is object-detection selection guide.

Information to Send KJT Sensors Support

• The completed worksheet: part dimensions, minimum gap, speed range, chain elements with settings;
• Part photos and material/surface description, worst-case orientation;
• Sensor model and mode; PLC input type, filter and scan;
• Miss pattern (speed/position/random) and commissioning history;
• What changed most recently (speed, parts, controller program).

Limitations

• Response times, switching frequencies, beam/spot geometry and filter effects are model-specific: the worksheet compares your numbers against specifications, and no specification values are invented here.
• Merged/touching parts defeat any discrete counting chain at the sensor level — that is an infeed engineering problem before it is a sensor problem.

Frequently Asked Questions

Why does a photoelectric sensor miss small parts at high speed?

Either the part is not seen (geometry: smaller than the effective beam, or riding outside it) or the sighting is not registered (timing: block time shorter than the sensor-plus-input chain needs, or gap time too short to reset). Separate them by slowing the line: misses that vanish at low speed are timing; misses that persist at all speeds at the same positions are geometry. The worksheet then identifies which chain element fails.

How do response time and PLC scan time affect detection?

They add. The sensor must complete its switch cycle, the input filter must pass the edge, and the controller must scan the input within the part’s block time — the slowest element defines the chain. A 1 ms sensor behind a 10 ms filter misses what the sensor cleanly saw. That is why the worksheet totals all elements: optimizing the sensor alone is the classic wrong fix.

What part length, gap and conveyor speed data are required?

Three minimum numbers and their worst cases: smallest part dimension in the beam direction, smallest gap between parts (including touching, if possible), and the maximum conveyor speed to be supported. From these, block time and gap time are computed directly. Add part material/surface and orientation range, because they decide the sensing mode’s reliability once the timing passes.

Conclusion

Small-part misses at speed are solved by the worksheet: block time and gap time versus the whole chain — sensor, filter, scan. Speed-correlated misses are timing; position-correlated misses are geometry; the indicator-vs-controller comparison splits sensor from logic. Restore margin at the failing element, and reselect — fiber optics for size, fast switching for speed — when the budget shows a structural gap.

Sources

• KJT Sensors, Photoelectric Sensors category (ToF family stating switching frequencies up to 1,000 Hz and transparent-object capability; fiber amplifier and fiber-optic series listed; verified 2026-10-10) — https://www.kjt-sensors.com/list-photoelectric.html
• KJT Sensors, Service & Support — https://www.kjt-sensors.com/list-services.html (verified 2026-10-10)
• KJT Sensors internal knowledge base: photoelectric sensors (ToF 1,000 Hz; fiber-optic small-target capability — manufacturer-stated) — version current at drafting

https://www.kjt-sensors.com/
KJT Sensors