Magnet Incoming Inspection & Acceptance Planning

The five risks a receiving plan must catch

Design the plan around failure modes, not around a generic inspection checklist. Each risk below has a different detection method, and a plan that catches four of five is a plan with a known blind spot.

Documents first, magnetics second. Both run before anything reaches lot-controlled stock.

Risk	How it reaches you	Where it surfaces if missed	Detection
Wrong or substandard grade	Substitution upstream, mixed lots, or a mill running to a lower specification	Field failure at temperature; gradual performance loss	Material certificate review, traceable to lot
Under-magnetization	Insufficient magnetizing field, fixture fault, or partially saturated parts	Low signal, weak holding, out-of-tolerance calibration at assembly	Helmholtz coil total moment
Wrong pattern or polarity	Fixture error, mixed orientation within a lot, reversed loading	Assembly that fits and does not work; polarity-blind assemblies fail in the field	Pole indicator; rotary Hall scan for multipole
Dimensional nonconformance	Tool wear, coating thickness variation, wrong measurement basis	Press-fit failures, bond line problems, air gap variation	Standard metrology to the drawing
Coating deficiency or damage	Thin plating, poor edge coverage, handling damage in transit	Corrosion months later, in service	Coating certificate plus visual; periodic thickness audit

Why the first two matter most

Dimensional and visual defects are caught by ordinary receiving practice on almost any part. Grade and magnetization are the two risks unique to magnets, and they are exactly the two that ordinary receiving practice misses entirely — because there is nothing to see. If your plan does nothing else, it should verify the material certificate against the lot and measure magnetic moment on a sample.

The document check at the dock

Cheapest control in the plan, performed before anything is unpacked, and it catches the highest-severity failure mode.

lot identityThe lot number on the material certificate appears on the packaging. If it does not, the certificate is not evidence about these parts — it is evidence about some parts.

measured values presentBr, Hcb, Hcj and BHmax reported as measurements against grade minimums. Values identical to the published datasheet, with no measurement date, indicate a datasheet reproduced as a certificate.

grade matches the orderIncluding the temperature suffix. N42 and N42H are different parts and the difference is invisible.

coating certificateCoating system and thickness per layer, not a single total. For nickel-copper-nickel, a thin outer nickel layer hides inside an acceptable total figure.

drawing revisionThe certificate of conformance references the revision you ordered against, not a superseded one.

quantity and packaging integritySeals intact, desiccant present where specified, no evidence of repacking in transit.

What to do when documents are missing

Receive to a quarantine location rather than rejecting outright, and do not release to production. Most missing-document events are administrative and resolve in a day. What must not happen is release into stock pending paperwork — because once parts are in a bin and the documents never arrive, you have inventory of unknown provenance and no clean way to segregate it. Physical quarantine costs nothing; a retrospective containment costs a great deal.

Choosing the primary magnetic acceptance test

Pick one magnetic property as the accept/reject criterion. Multiple overlapping magnetic specifications guarantee that good parts fail one of them and produce disputes rather than quality.

Method	Measures	Suitability as primary acceptance
Helmholtz coil + fluxmeter	Total magnetic moment of the whole part	The default choice. Fast, position-tolerant, correlates well between supplier and customer, and captures grade, volume and saturation state in one number
Gaussmeter surface field	Flux density at one point	Poor as a criterion unless fully fixtured. Extremely sensitive to standoff, probe model and position — a reliable source of disputes on conforming parts
Pull force test	Force to detach from a steel plate	A system test, not a magnet test. Depends on plate thickness, alloy, finish and alignment. Use for design verification, not acceptance
Rotary Hall pole scan	Field versus angle around the part	The correct primary test for multipole and true radial parts, where pattern integrity is the thing that matters
Hysteresisgraph	Material B-H curve	Lab instrument, destructive of time rather than parts. Belongs at the mill, reported via certificate — not a receiving-dock test

Why moment beats surface field

Surface field falls steeply with distance from the magnet face, so a few tenths of a millimetre of probe lift-off shifts the reading by several percent — and the Hall element sits recessed inside the probe tip by an amount that varies between probe models. Two competent people with different gaussmeters will get different numbers on the same conforming part. A helmholtz measurement is nearly insensitive to position within the coil’s uniform region, which is precisely what makes it usable as a shared accept/reject basis. The full instrument-by-instrument treatment is in how magnets are tested.

Writing the criterion

A usable drawing note states the property, the method, the limit, the sampling and the conditions:

example acceptance note

Total magnetic moment measured per helmholtz coil and fluxmeter shall be not less than X.XX µWb·cm, measured at 23 ±3 °C. Sampling per ANSI/ASQ Z1.4, general inspection level II, AQL 1.0. Limit anchored to first-article hardware measured [date], report reference [number].

Ask your supplier to calculate the expected moment for your part’s grade and dimensions and to supply first-article parts with measured values, so the limit is anchored to real hardware rather than to an idealised calculation. Calculators assume perfect geometry and full saturation; production parts do not.

Sampling plans

ANSI/ASQ Z1.4 is the standard framework and is what most customer flow-downs reference. General inspection level II is the normal default; the acceptable quality limit is the risk decision.

Single sampling plans, normal inspection, general inspection level II. Ac = accept number, Re = reject number.
Lot size	Code	Sample size	AQL 1.0 — Ac / Re	AQL 2.5 — Ac / Re
91–150	F	20	0 / 1	1 / 2
151–280	G	32	1 / 2	2 / 3
281–500	H	50	1 / 2	3 / 4
501–1,200	J	80	2 / 3	5 / 6
1,201–3,200	K	125	3 / 4	7 / 8
3,201–10,000	L	200	5 / 6	10 / 11
10,001–35,000	M	315	7 / 8	14 / 15

Choosing the AQL

AQL 0.65 or tighter — safety-relevant characteristics, or defects that escape final test and reach the field.
AQL 1.0 — a sensible default for magnetic and critical dimensional characteristics on production parts.
AQL 2.5 — cosmetic and non-functional characteristics, minor visual defects.
AQL 4.0 — incidental appearance issues on parts hidden inside an assembly.

When to go to 100%

Polarity, where the assembly is polarity-blind. If a reversed magnet installs invisibly and fails later, sampling is not sufficient — and a polarity check takes under a second per part.
During containment following a nonconformance, until confidence is re-established.
New supplier or new tooling, for the first two or three lots.
Safety-critical characteristics where the customer or regulation requires it.

Reducing inspection over time

The point of a receiving plan is to earn the right to stop inspecting. Skip-lot and reduced inspection schemes are legitimate and are what a mature supplier relationship should converge toward: after a defined run of conforming lots, move to reduced inspection, then to skip-lot, with a defined trigger returning to normal — a nonconformance, a supplier process change, or a period without deliveries. The change-control clause in your supply agreement is what makes this safe, because it obliges the supplier to notify a mill, coating or process change rather than letting it arrive unannounced in a skipped lot. See commercial terms.

Dimensional and coating verification

Measuring magnetized parts

Use non-magnetic tooling — ceramic or carbide anvils, non-ferrous fixtures. A steel micrometer anvil is pulled onto the part and reads short, and the error is systematic rather than random.
Optical comparators and vision systems avoid contact entirely and are the better method for small magnetized parts.
Keep magnetized parts away from CMM scales and electronic gauges where the field can interfere with the instrument.
Confirm the measurement basis matches the drawing — before or after coating. This single ambiguity accounts for a large share of dimensional disputes on plated magnets.

Workmanship limits on a brittle material

Sintered magnet material is brittle and ground with diamond tooling. Small edge chips are inherent to the process, not evidence of mishandling. A drawing that says “no chips” is unenforceable and will be ignored; a drawing that defines allowable chip size and location is a real specification.

what to defineMaximum chip dimension, maximum number per part, and the zones where chips are not acceptable — typically sealing faces, bond faces and pole faces at the working gap.

what to excludeNon-functional edges away from any interface. Rejecting on these costs money and buys nothing.

how to communicate itA boundary sample set — physical accept and reject examples retained at receiving. Faster and more consistent than written descriptions, and it eliminates inspector-to-inspector variation.

Coating verification

Thickness by X-ray fluorescence, which resolves individual layers in a nickel-copper-nickel system. Magnetic induction gauges give a total only.
Edge coverage on visual inspection. Edges and corners are where coatings thin and where corrosion initiates, so that is where to look.
Adhesion by cross-hatch tape test where the part will be bonded — a periodic audit rather than a per-lot check.
Salt spray is a qualification test, performed once on a coating system per ASTM B117. It is not a receiving test, and a supplier producing fresh salt spray data with every shipment is reproducing an old report.

Setting limits that hold up

A limit that has not been validated against real hardware and a real measurement system will eventually reject conforming parts, and the resulting argument costs more than the inspection ever saved.

Anchor to first-article hardware

Receive first-article parts and measure them on your equipment, in your environment.
Compare against the supplier’s measurements of the same parts. A systematic offset between the two is normal and must be understood before limits are set — coil constants, calibration and temperature all contribute.
Set the limit from measured performance plus a tolerance analysis, not from a calculator’s ideal value.
Record the first-article report reference on the drawing, so the origin of the number survives staff turnover.

Do the gauge R&R

A measurement system analysis on the actual fixture, instrument and operators is what makes a limit defensible. It is also expected content in a PPAP submission. For magnet inspection specifically:

Temperature affects the reading. NdFeB remanence drifts around −0.12%/°C, so a part measured in a cold receiving bay reads differently from one measured on a warm bench. State the measurement temperature in the criterion.
Fixturing dominates repeatability for anything other than helmholtz measurement. If the study shows poor repeatability, the fixture is usually the cause, not the operator.
Calibrate against a reference standard and re-verify on a schedule. Coil constants drift.

The mistake that generates false rejects

Setting a magnetic acceptance limit at the grade’s nominal value. Grade specifications define minimums, and production parts scatter around a mean above the minimum with a spread of several percent between lots, plus dimensional tolerance contributing its own variation to total moment. A limit set at nominal rejects roughly half of a conforming population. Set the limit from the measured first-article mean less an allowance covering material spread and dimensional tolerance — typically several percent, and calculable from your own tolerance stack.

Handling nonconformance

Immediate actions

Segregate physically to a marked quarantine location. Not a label on a bin in the normal rack — magnets migrate between containers in ways other components do not.
Contain by lot. Identify every location holding parts from the affected lot, including work in process and finished goods. Lot-controlled storage is what makes this a one-hour task instead of a one-week task.
Record the measurement, including the method, equipment and conditions. A nonconformance report saying “magnets weak” cannot be investigated.
Notify the supplier within the agreed response window, with the lot number, the measured data and photographs.
Retain samples before returning anything. If the parts go back and the argument continues, you need evidence.

Disposition

Disposition	Use when	Requires
Return to supplier	Clear nonconformance, replacement available in time	RMA, and shipping arrangements for magnetized material
Sort	Mixed lot with a detectable characteristic and schedule pressure	Agreement on who performs and pays for the sort; 100% inspection of the sorted population
Use as is, under deviation	Nonconformance is genuinely non-functional for this application	Documented engineering assessment and approval; time-limited and quantity-limited
Rework	Re-magnetization or re-plating can restore conformance	Verification after rework; note that re-magnetization does not recover loss caused by over-temperature or corrosion
Scrap	No viable path and return is uneconomic	Demagnetization before disposal; recovery credit where the value justifies it

Two things that go wrong at this stage

Open-ended deviations. A deviation is a one-time, quantity-limited, documented engineering decision. Deviations that get renewed indefinitely are a specification that nobody updated — if the condition is genuinely acceptable, change the drawing.

Returning magnetized material without preparation. Magnets going back by air are still regulated cargo subject to magnetic field limits, and a rejected shipment at the airport turns a quality problem into a logistics problem. Agree the return method with the supplier before shipping.

Corrective action

Request a structured response — 8D or an equivalent format — on anything recurring or severe. For magnets the containment question that matters most is lot boundaries: which other lots share the affected material batch, magnetizing fixture or plating run. A supplier who can answer that quickly has real traceability; one who cannot has just told you something important about their process control.

A receiving plan you can lift

A starting point for a production magnet on a qualified program. Adjust the AQLs and frequencies to your risk profile and customer flow-downs.

Sample incoming inspection plan — sintered NdFeB production part.
Step	Characteristic	Method	Frequency	Acceptance
1	Packaging integrity, quantity	Visual, count	Every lot	Seals intact, desiccant present, count matches
2	Certificate of conformance	Document review	Every lot	References correct PO and drawing revision
3	Material certificate	Document review	Every lot	Lot number matches packaging; measured values meet grade minimums
4	Coating certificate	Document review	Every lot	System and per-layer thickness within specification
5	Total magnetic moment	Helmholtz coil and fluxmeter	Z1.4 level II, AQL 1.0	≥ limit from first-article report, at 23 ±3 °C
6	Polarity / orientation	Pole indicator	100% where assembly is polarity-blind	Marked face is north
7	Critical dimensions	Non-magnetic metrology per ballooned drawing	Z1.4 level II, AQL 1.0	Within drawing tolerance, stated coating basis
8	Visual / workmanship	Comparison against boundary samples	Z1.4 level II, AQL 2.5	Within defined chip and surface limits
9	Coating thickness audit	XRF	Quarterly or on supplier change	Per-layer thickness within specification
10	Storage transfer	Lot-controlled location, FIFO	Every lot	No commingling of lots

Right-sizing it

For a non-critical industrial part, steps 1–3 and 8 alone are a proportionate plan. For an automotive or medical program, add pole scanning on patterned parts and retain records to the customer’s required period. The objective is not maximum inspection — it is catching the two risks that only magnets have, at the one point where catching them is still cheap.

Incoming Inspection & Acceptance Planning

Contents