Why Tiny Metal Parts Need a Different Ring Sensor
A ring sensor can make tiny screws, pins and springs easier to detect by surrounding their path. For reliable counting, match the sensing mode and aperture to the real part—then make sure each passage produces a separate signal the controller can capture.
When should you try a ring sensor instead of a standard proximity sensor?
Try a ring sensor when small metal parts travel through a guide and cannot stay close to one sensing face. A screw tumbling against the far wall of a tube may pass outside a side-mounted sensor’s usable detection zone. Increasing the advertised sensing distance does not, by itself, prove that the smallest part will be detected.
An inductive ring sensor concentrates an alternating electromagnetic field inside its opening. A metal target changes the evaluated signal as it enters that field. Passing the guide through the ring makes the sensing arrangement surround the travel path, rather than watch it from one side.
A ring is not automatically necessary for every small part. If a fixture presents the target repeatably at a short, controlled gap, a suitable miniature proximity sensor may already solve the problem. For non-metal parts, evaluate an appropriate optical method instead.
Check where the information is lost. No distinct sensor output for a passing part points toward the target, geometry or sensing mode. Distinct outputs that never become PLC counts point toward the electrical interface or acquisition settings. Changing the sensor will not necessarily fix the second problem.
Do you need to detect a passing part or a part that has stopped?
Choose the operating mode from the event you need to know about. Static operation evaluates metal presence; dynamic operation responds to a passing target. Neither mode should be selected from the housing shape alone.
| Required information | Starting point | Important limit |
|---|---|---|
| Did a small part pass? | Dynamic sensing is worth evaluating for small, low-mass moving targets. | Verify the permitted speed range, target sensitivity and recovery between events. |
| Is metal still present here? | Static sensing, with a target large enough for that mode. | The target must occupy the actual detection zone; a blockage elsewhere is not directly observed. |
| Has feeding failed? | Compare expected passages with the machine’s feed command and timing. | No pulse alone does not distinguish an empty feeder, a jam, a missed target or a signal fault. |
Can one ring prove that a screw reached the assembly position?
Not just from an upstream passage signal. It confirms an event at the ring, not successful arrival, seating or tightening farther along the machine. If arrival matters, use evidence at the destination or a validated sequence that covers the intervening transfer.
Likewise, one ordinary binary passage signal does not establish travel direction. A part that falls back and passes again can create another event. Feed mechanics and counting logic must account for that possibility.
How small a metal part can the sensor actually detect?
There is no category-wide minimum. Read the stated reference target, operating mode and settings before comparing a sensor with your part. “Detects a 1 mm steel ball” does not mean “detects every component with a 1 mm outside dimension.”
Use the reference target as a comparison, not an approval
A thin spring, hollow rivet and solid ball can have similar outside dimensions but interact differently with the sensing field. Alloy, wire thickness, shape and orientation belong in the comparison.
Test the least detectable accepted part in its least favourable allowed position. Do not rely only on the largest screw or the easiest hand-fed sample.
A documented example: one aperture, two target limits
The ifm I7R205 datasheet, revision I7R205-02 dated 30 March 2026, specifies a 15.1 mm inside diameter. Its listed steel-ball resolution is 0.8 mm in dynamic mode and 2 mm in static mode. The document identifies the variant as I7R3015-FRKG/IO/US-104.
This illustrates why bore size alone is not a sensitivity specification. It does not qualify an irregular spring or establish an xsz sensor rating. Match the delivered variant and configured mode to the referenced manufacturer datasheet, then test the production part.
How do you choose the aperture and position the feed tube?
Choose an opening that accommodates the real guide and its movement without compromising target detection. The sensor’s inside diameter, the tube’s outside diameter and the part’s dimensions answer three different questions.
- Ring apertureFit the complete guide, including its dimensional tolerance and permitted movement.
- Guide outside diameterAccount for the wall, ovality and replacement tube—not just its nominal inner bore.
- Real targetCheck the whole allowed path, including near-wall and tumbling positions.
Keep installation changes in the test
Use a suitable non-metallic tube. A reinforced hose, metal coupling or nearby clamp is not equivalent to a plain polymer guide. Follow the selected model’s metal-clearance and adjacent-sensor instructions; do not assume that dynamic operation makes nearby metal harmless.
Check whether a closed ring can be fitted over the tube’s ends, or whether a separable housing is needed for maintenance. Secure the guide as intended and repeat detection checks after tube or bracket replacement. Do not oversize the opening simply for convenience without confirming the weakest target.
Why can the PLC miss parts that the sensor detects?
The sensor must detect a part, produce a usable output and return to a state that lets the next event be distinguished. The controller must recognise both the active interval and the gap between pulses. Fast response time, long output pulses and high counting capacity are not the same specification.
Longer pulses help capture—but can erase separation
Pulse stretching holds the output long enough for a receiving input to register it. If the hold time overlaps the next passage, the output may remain active or the new event may be suppressed, depending on the device’s retrigger and recovery behaviour. Verify that behaviour in the timing diagram; do not assume that every passage produces a new edge.
For a genuinely periodic output, the inactive interval is the period minus the active interval. Both intervals must meet the receiving input’s requirements. Use the shortest interval in a burst, not only the average parts-per-minute rate.
Illustrative example: two screws, one count
Assume a feeder averages 600 parts per minute: an average interval of 100 ms. In one burst, however, two screws arrive only 30 ms apart. This example uses an active-HIGH output and rising-edge counting.
- The first detection occurs at 0 ms and starts a 40 ms output hold.
- For this example only, the second detection at 30 ms restarts that hold. The output stays active until 70 ms.
- A rising-edge counter receives one uninterrupted active signal, so it has only one rising edge to count.
A faster PLC cannot reconstruct a missing second edge. The next test is to establish a compatible pulse setting, sufficient part separation or a supported internal event counter. A different sensor may handle retriggering differently. These assumed timings demonstrate the mechanism, not a recommended setting or a measured customer result.
Check the input filter as well as the PLC scan
A clean short pulse can be rejected before the program reads it. Siemens’ S7-1200 documentation explains that input filtering also applies to high-speed-counter inputs; selecting an HSC does not remove the need for a suitable filter.
Verify minimum HIGH and LOW durations, electrical compatibility, filtering and the actual acquisition method. Use a supported high-speed input or event-counting function when ordinary cyclic sampling cannot capture the required events. An LED flash is not a measurement of pulse width. Have qualified personnel make wiring and configuration changes under the machine’s safe commissioning procedure.
What should you check first when counts are wrong?
Compare a known quantity of parts with the sensor output and the controller’s recorded events under the same test conditions. Establish a safe test arrangement before inspecting the feed path; isolate the machine before moving hardware or changing wiring.
| Observation | Next check |
|---|---|
| Only the smallest or tumbling parts are missed | Compare those samples across the permitted path, with the final guide and mounting. Recheck the selected mode and sensitivity. |
| Separated parts work; close pairs do not | Check mechanical separation, output hold time and recovery. Confirm whether two distinct output events exist before altering PLC logic. |
| Two distinct pulses exist, but only one count is recorded | Check input filter, minimum HIGH/LOW durations, acquisition rate and edge-counting logic. |
| One part produces multiple counts | Look for reverse travel or bouncing, multiple output transitions, or logic counting both edges. Do not hide the cause with an arbitrary long delay. |
| Performance changes after maintenance | Compare tube material and position, bracket clearances, connections and restored settings against the accepted configuration. |
What must a production trial prove before you approve the sensor?
Approve the part, sensor, guide, settings and controller combination, not just the ring’s catalogue description. A sensible trial records where an error occurs and which operating conditions were actually covered.
- Target coverage: the smallest accepted parts, material variants, thin sections and allowed orientations—not just one convenient reference sample.
- Feed coverage: minimum and maximum operating speed, the closest permitted spacing, burst release, stops and restarts.
- Installed condition: production tube, supports, nearby metal, neighbouring sensors and expected contamination.
- Signal and count agreement: independently known part quantities reconciled with output events and controller counts, including missed and extra events.
- Recovery: documented behaviour after interruption, maintenance and settings restoration, without unintended extra counts.
Set the permitted error level and trial duration from the process requirements before testing. “No errors observed” should be accompanied by the number of parts and conditions tested; it is not a guarantee for every future batch. Two touching parts should not be assumed individually countable from a single binary signal.
For a supplier review, send the part drawing and representative samples, guide dimensions, travel speed and shortest spacing, presence-versus-passage requirement, and controller input details. Ask for the proposed model, its mounting drawing, reference-target specification and timing behaviour. That gives a ring proximity sensor discussion a useful technical starting point.
Sources and method references
- Pepperl+Fuchs: inductive sensor housing designs — sensing-face arrangements and the ring’s internal field.
- ifm: inductive ring sensor overview — static and dynamic operation in feeding applications.
- ifm I7R205-02 datasheet — the specific aperture and steel-ball values used in the documented example.
- wenglor IR5D002 operating instructions — model-specific installation, sensitivity and pulse-length behaviour; not universal settings.
- Siemens S7-1200: digital input filter times — filtering of HIGH/LOW transitions and high-speed-counter inputs.
- OMRON: proximity sensor terminology — reference targets, response time and response frequency.
The timing example and cross-section are explanatory illustrations, not test results. The hero is an AI-generated concept scene, not a photograph of a specified product or an approved mounting arrangement.