Disc-Style Chip Detectors and the Problem of Nuisance Chip Lights

Magnet Stability and Capture

A disc-style detector, also called a pole-piece detector, places a samarium cobalt ring magnet inside a probe, between two soft-iron pole pieces. The iron directs the magnet's flux to a narrow chip gap that runs around the probe. The gap is filled flush with an insulator. Debris that bridges the gap at any point around its circumference turns on the chip light.

This design solves the magnet problems of the two-prong detector. The pole pieces enclose the magnet, so it is protected from impact, chipping and other magnets. Samarium cobalt is a high-coercivity material, like the neodymium in the NZMS. Treated conservatively as a bare ring with no pole pieces, it takes no permanent loss from stray fields below about 763 mT. Its irreversible thermal loss begins above 505°F (263°C). None of the Alnico weaknesses described in Part 1 apply.

The disc-style detector also captures far more debris. In our analysis, the volume of oil in which its magnet pulls a 100 µm steel chip faster than 100 mm/s is 0.094 in³. For the two-prong detector it is 0.025 in³. In our pipe capture model, the disc-style detector captures 40.8% of the passing 250 µm chips, against 20.0% for the two-prong detector. Independent laboratory testing of production detectors agrees. In every matched condition, the disc-style detector captured 1.2 to 2.6 times as much debris as the two-prong detector.

On these measures, the disc-style detector appears to be a clear improvement. Its indication tells a different story.

A Chip Detector Does Not Report What It Catches

A chip detector does not report how much debris it holds. It reports only whether the gap has bridged. Debris collected before that point is not reported. Once the gap bridges, the result is a chip light.

As a result, the shape of the gap matters as much as the strength of the magnet. On a two-prong detector, chips favor the outer corners of the prongs. In independent testing, the gap often stayed open while a large amount of debris built up. The disc-style gap runs around the entire probe, where the field is strongest (about 432 mT in the gap). Debris collects on it and bridges it reliably, after about a fifth of the chips the two-prong detector needed. Its bridges were also low in resistance, typically a few ohms to a few tens of ohms.

Normal Wear Debris and Nuisance Trips

All gearboxes and engines produce debris in normal operation. Gears and bearings shed a steady stream of fine flakes, known to maintainers as “fuzz.” Fuzz indicates normal operation, not failure. A chip detector cannot distinguish fuzz from the first flakes of a spalling bearing. Both are ferrous, both are captured, and both bridge the gap once enough has collected. What separates them is size. Fuzz is fine flakes. A spalling bearing sheds chips large enough to see. A chip detector reports neither the size nor the amount of what it holds, only that the gap has bridged, so the difference is found only when the detector is removed and the debris is examined.

A two-prong detector, with its less reliable gap, needs a large amount of fuzz before it trips. The disc-style detector needs very little. The better it captures, the sooner it collects enough normal debris to bridge its gap. The chip light then turns on during normal wear. This is a nuisance trip. It says nothing about the condition of the machine. It still requires removal, inspection, cleaning and refitting, and it can ground an aircraft until the detector is signed off.

Some systems address this with a fuzz burner, which sends a pulse of current to burn away fine debris bridging the gap. Others verify an indication electronically before it reaches the crew. Both approaches reduce false alarms, but both add hardware. A fuzz burner needs a power supply, wiring and a control unit. These add weight, cost and parts to maintain, and each must be qualified for the aircraft. The burn itself is an electrical discharge inside the oil system, which must be assessed as part of that qualification. Both approaches also reduce false alarms by discarding evidence. Debris that has been burned away can no longer be examined.

The Simulated Debris Record

To show the effect over a service life, the KasperAero white paper includes a simulation. Its key inputs come from the independent laboratory tests, chiefly how much debris each detector needs before it bridges.

The simulation follows one gearbox through one debris history. It then shows how that same history would appear on each type of device, had it been the one installed. Each device sees the same debris under the same conditions and is given the same capture efficiency, so the only difference is what each one reports. The chip detectors run on a 28 V supply. Their current stays at zero until the gap bridges, and each chip light means the detector is removed, cleaned and refitted.

Normal wear sheds a steady stream of fine flakes for the whole record. At 1,150 flight hours a bearing begins to spall, and from then on it also sheds occasional large chips of about 1 to 3 mg each. The figure follows the two devices that capture best, the disc-style detector and the NZMS.

During normal wear, the disc-style detector trips nine times. Each is a nuisance trip, because the gearbox has no fault yet. A two-prong detector, given the same history, trips twice.

Both devices catch the first large chip, at 1,200 flight hours. The NZMS reading jumps by 2.0 mg in one hour, about 40 times its normal hourly capture. That one reading shows that a chip has arrived and that it is large. The disc-style detector turns on its chip light in the same hour. The light is identical to the nine before it, so the indication alone cannot say whether the gap holds fuzz or a spall chip. Only removing the detector and examining the debris tells them apart.

A single record could be favorable or unfavorable by chance. The white paper therefore ran its simulation 1,000 times, with random particle arrivals, captures and bridging amounts. In those runs wear accelerates gradually after the onset, with no single large chip, so any warning has to come from the rise in the debris rate:

Ranges are the 10th to 90th percentile. FH is flight hours. The NZMS alerted before the first two-prong chip light in 90% of the histories, and before the first disc-style chip light in 54%.

The Cost of Cleaning

Each nuisance trip has a second cost. Cleaning a chip detector erases its history. After each removal, it starts again from zero and must collect enough debris to bridge the gap before it can indicate anything. This collection time delays every warning a chip detector gives. It is also why a chip detector cannot report the measure that changes when a machine begins to fail: the rate at which debris arrives.

Repeated nuisance trips also change how the chip light is received. Maintainers still remove, inspect and clean the detector each time, as the procedure requires. After nine chip lights that turned out to be fuzz, however, the tenth is expected to be fuzz as well. Crews and maintainers become desensitized to the indication, and the one that matters is taken less seriously than it should be.

Where the NZMS Fits

The NZMS captures about as well as the disc-style detector. Its capture volume is nearly identical (0.095 in³ against 0.094 in³), and its pipe capture is 37.7% against 40.8%. It handles the debris in a different way. It has no chip gap to bridge. It measures the mass of debris on its sensing face continuously, so normal wear appears as a slow, steady line rather than a chip light. When wear accelerates, the line bends, and the bend is the warning. The sensor does not need to be removed or cleaned to keep working, so its history is preserved.

In the simulation, the NZMS warned as early as the most sensitive chip detector, with no nuisance trips and no false alarms. Less uncertainty means fewer unnecessary removals. Fewer removals mean less maintenance.

Every capture figure in this article comes from KasperAero's magnetic capture analysis tools, the same models behind the white paper. We use them to help customers place NZMS sensors in their own systems. The full white paper, KR_WP#261023, covers the methods, validation and appendices. It is available on request.

Part 1: Alnico Demagnetization in Two-Prong Chip Detectors »
Our Magnetic Capture Analysis Tools »
Request the White Paper »

References

1. KasperAero, White Paper KR_WP#261023, NZMS vs Traditional Chip Detectors (2026). Available on request.
2. Independent laboratory capture-efficiency test reports for production two-prong and disc-style chip detectors in gearbox test cavities with MIL-PRF-23699 oil (two campaigns). Used in summary form with the written permission of the data owner.
3. United States Patent 5,406,208, “Method and apparatus for eliminating chips in a chip detection circuit and for determining chip size.”
4. United States Patent 5,583,441, “Method and apparatus for automatically verifying faults and monitoring chips in a chip detection circuit.”
5. Arnold Magnetic Technologies, RECOMA Sintered Samarium Cobalt Magnets, grade datasheets.
6. SAE International, AIR1828D, Guide to Engine Lubrication System Monitoring.

The debris record is a simulation. Capture results are KasperAero analysis.

Back to Knowledge Hub »