A quality engineer’s first meeting with a wireless return is often a photograph of a coin-shaped scorch, or a customer who says the pad “got warm” and wants a replacement. Those two pictures are not the same failure. Foreign-object detection (FOD) on a foreign object detection wireless charger production line exists to stop a metal object in the magnetic field from becoming a heat source; a line that only records “it got warm” has not yet diagnosed whether FOD missed, tripped early, or was never the mechanism at work.

This page is a production-line verification path for quality and wireless design engineers. It is not a consumer safety blog, and it will not invent unofficial Wireless Power Consortium pass/fail thresholds. Use current WPC/Qi documentation and the test partner’s method for any limit. Category and certification context are on wireless charging, wireless charger OEM, the Qi2 certification and selection article, and OEM and ODM. Magnet-and-yield production is on the Qi2 25W tolerances article in this series; this page does not repeat it.

What foreign-object detection is for, in one accurate paragraph, then into test method

Foreign-object detection identifies a metallic object in the charging field and reduces or stops power so that object does not become a heat source. A production test then checks that this function behaves as the agreed method requires on units that will ship — not that “it got warm” on a bench phone.

The method belongs to the current Qi or Qi2 specification and to the laboratory or factory procedure written against it. This article describes how a line can implement a verification of that method: sample objects, a defined pass/fail behaviour, a place in incoming and outgoing sampling, and a record. It does not replace the WPC test, and it does not offer a shorthand temperature that can stand in for FOD.

WPC/Qi test concepts a production line can actually implement

A production line cannot run a full certification campaign on every unit, and it should not pretend to. What it can implement are Qi test concepts in a production form: a controlled set of sample objects, a defined placement, a defined source and receiver state, and a defined response — power reduced or stopped, a signal, a time window — taken from the current WPC method or from a laboratory procedure that references it. The Wireless Power Consortium’s site, Qi wireless charging page, knowledge base and certified products database are the official entry points; open the current documents there, and do not copy a number from memory. The WPC’s Qi2 25W news release is relevant only when the SKU is a Qi2 25W product; it is not an FOD threshold document.

The production version of that method must use a named fixture, a named object set, and a record. A technician waving a house key over a pad is not a Qi test concept, even if it is well intentioned.

Sample objects, pass/fail behaviour, and what “it got warm” is not

The object set should be the set the agreed method requires, identified by part number, and stored so that a worn or substituted object cannot quietly enter the line. Pass/fail behaviour is the response the method requires: typically that power is reduced or stopped when the object is present, and that a normal session can still start when the object is absent. Those two sides both matter.

“It got warm” is not a pass/fail result. A pad and a phone will warm in normal charging. A metallic object that FOD missed may become much hotter, and that is a different event, which should be recorded with temperature only if the method calls for a temperature measurement — and then with the method’s own limit, not with an unofficial figure. A return described as “warm” needs a teardown and a log, not a reflex FOD fail stamp.

White foldable 3-in-1 wireless charger standing open, with a phone, a watch and an earbud case on three separate pads
On a multi-coil pad, FOD has to be verified at each position a user can occupy, with a controlled object set, not with a pocketful of keys.

False trips and missed detections as two different return drivers

False trips and missed detections drive two different returns, and they should not be averaged into one “FOD yield.” A false trip is FOD that stops or reduces a legitimate session — a phone in a thin case, a magnetic mount, a ring on a finger — and the customer says “it won’t charge.” A missed detection is FOD that allows a metallic object to remain in the field, and the customer says “it burned a coin” or worse. Tightening a threshold without a method may reduce misses and increase false trips, or the reverse. That is why unofficial thresholds do not belong on this page, and why a change to FOD firmware is a change-controlled event.

The acceptance specification should state both failure modes, with example objects for each, and should ask the factory to report them separately.

Incoming versus outgoing FOD sampling on a wireless OEM line

Incoming sampling is for bought-in pads, coils or finished modules: a sample from a lot, run against the agreed method, before the lot is released into assembly or into a private-label carton. Outgoing sampling is for finished units leaving the line, including each coil on a 3-in-1 or 5-in-1 dock. The two samples answer different questions. Incoming catches a vendor or a lot. Outgoing catches assembly, firmware load and a housing that may have moved a coil.

SampleWhen it runsWhat it is forWhat it is not
IncomingOn received coils, pads or modules, by lotVendor and lot releaseA substitute for outgoing finished-unit checks
OutgoingOn finished units, including each user-accessible coilAssembly, firmware and housing stackA full WPC certification re-run
After a changeAfter firmware, coil-vendor or construction changesRe-establishing the method on the new constructionA visual inspection only

Sample sizes are a contract item. This article will not invent an AQL. A plan that only runs FOD when a customer has already returned a unit is not an incoming or outgoing plan.

Firmware and coil-vendor changes that require FOD to be run again

Firmware that touches coil control, object detection, or thermal fold-back should reopen FOD sampling, and may require the certification or construction file to be revisited. A coil-vendor swap, a coil-turn change, or a housing change that moves the coil relative to the surface is in the same class. These are not colour changes. The contract should say that production may not resume on the changed construction until the agreed FOD method has been run again and the record filed.

A “small firmware update” after sample lock, loaded on a Friday to “improve charging,” is a classic way a false-trip or a miss appears in the next retail lot.

How to write an FOD clause into an OEM acceptance specification

An FOD clause a factory can quote should name: the method and the document it references (current WPC/Qi, or a laboratory procedure that cites it); the object set by part number; the placement and the source/receiver state; the required response; the incoming and outgoing sample; the two failure modes (false trip and miss) as separate reports; the record (video, log, fail photograph, teardown note); and the change-control trigger. It should not contain a temperature or power number copied from a blog, and it should not say “FOD as standard” without a method.

If the product is a multi-coil dock, the clause applies to each coil. If the product is Qi2 25W, the clause sits next to the magnet-performance sheet, it does not replace it.

White foldable wireless charger shown disassembled into its charging pads and folded body, the kind of construction a failed-unit teardown would examine
A failed FOD unit should be teardowned with a photograph and a log; a warmth complaint without that record is not a closed diagnosis.

Evidence a factory should keep: videos, logs, and failed-unit teardown notes

A useful FOD record is reproducible. Keep a short video or a timed log of the object placement and the pad’s response, the firmware version, the fixture and object identities, and, for fails, a teardown note: where the object sat, what the coil and the surface look like, and whether the failure is a miss, a false trip, or something else (a damaged surface, a wrong firmware load). Store those artefacts with the lot.

Without them, a debate about a returned unit becomes a memory contest. With them, a resident inspector can fill the same sheet the line used.

Linking FOD to thermal and Qi2 production articles without repeating them

Thermal gates and Qi2 magnet-array production are companion topics. A thermal gate measures temperature under a stated load; it is not FOD. A magnet-alignment check measures whether a phone seats; it is not FOD. This page links to those ideas and stops. Do not copy a Qi2 25W yield table into an FOD clause, and do not treat a passing thermal sample as evidence that a coin would have been detected. The three gates can share a visit and a lot record; they cannot share a result.

What this article will not invent: unofficial WPC thresholds

This article will not publish unofficial WPC or Qi temperature, power, time or object-size thresholds, will not invent a “typical” FOD pass rate, and will not claim that a named WECENT model has a particular FOD performance or a particular certification. Any pass/fail number used in a program should be traceable to a current official document or to a laboratory method that cites one. If a factory quotes a number and cannot show that source, leave the cell on the acceptance sheet blank until the source arrives. A blank cell is more honest than a memorable figure.

Conclusion

Foreign-object detection on a wireless production line is a method, an object set, two failure modes, and a record. “It got warm” is not that method. Write incoming and outgoing samples, reopen FOD after firmware or coil-vendor changes, and take every pass/fail number from current WPC/Qi documents — never from an unofficial threshold on a blog or on this page.

If you are writing a foreign-object detection clause for a wireless charger OEM line, you can send the intended SKU and method scope to WECENT through the wireless charger OEM page. This is a request for a sourcing discussion, a quotation or a sample plan, not an order.

FAQ

What is foreign-object detection for on a wireless charger?

Foreign-object detection is a function that identifies a metallic object in the wireless charging field and reduces or stops power so that object does not become a heat source. It is a Qi / Qi2 safety-related function defined in the current Wireless Power Consortium documentation, not a marketing feature. A production test then checks that finished units, and incoming lots, behave as the agreed method requires. A report that a pad “got warm” during ordinary charging is not, by itself, an FOD result, because a normal session also produces heat. Diagnosis requires the method, the object and a record.

Why should false trips and missed detections be reported as two different figures?

Because they produce two different customer returns and they move in opposite directions if a detection threshold is changed without a method. A false trip stops a legitimate session and arrives as “it will not charge.” A missed detection leaves a metallic object in the field and arrives as heat damage or a safety complaint. Averaging them into one “FOD yield” hides which way the process has drifted. An OEM acceptance clause should name both modes, with example objects, and should require the factory to report them separately, especially after a firmware or coil-vendor change.

When must FOD be run again after a change on the line?

When firmware that touches coil control, object detection or thermal fold-back is changed, and when a coil vendor, a coil construction, or a housing that moves the coil relative to the surface is changed. Those events should reopen the agreed FOD method and may require the certification or construction file to be revisited. They are not equivalent to a colour change. Production on the new construction should not resume until the method has been run and the record filed. A verbal assurance that “FOD is still the same” is not that record.

Will this article provide the WPC temperature or object-size limits to put in a specification?

No. Official pass/fail numbers belong in the current WPC Qi or Qi2 documentation, or in a laboratory procedure that cites that documentation. This article will not invent unofficial thresholds, typical pass rates, or model-specific FOD performance. An OEM clause should reference the current official method, name the object set and the required response, and leave any numeric limit to a source the factory can show. A blank cell waiting for that source is preferable to a memorable figure with no clause behind it.

{
“@context”: “https://schema.org”,
“@type”: “FAQPage”,
“mainEntity”: [
{
“@type”: “Question”,
“name”: “What is foreign-object detection for on a wireless charger?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Foreign-object detection is a function that identifies a metallic object in the wireless charging field and reduces or stops power so that object does not become a heat source. It is a Qi / Qi2 safety-related function defined in the current Wireless Power Consortium documentation, not a marketing feature. A production test then checks that finished units, and incoming lots, behave as the agreed method requires. A report that a pad “got warm” during ordinary charging is not, by itself, an FOD result, because a normal session also produces heat. Diagnosis requires the method, the object and a record.”
}
},
{
“@type”: “Question”,
“name”: “Why should false trips and missed detections be reported as two different figures?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Because they produce two different customer returns and they move in opposite directions if a detection threshold is changed without a method. A false trip stops a legitimate session and arrives as “it will not charge.” A missed detection leaves a metallic object in the field and arrives as heat damage or a safety complaint. Averaging them into one “FOD yield” hides which way the process has drifted. An OEM acceptance clause should name both modes, with example objects, and should require the factory to report them separately, especially after a firmware or coil-vendor change.”
}
},
{
“@type”: “Question”,
“name”: “When must FOD be run again after a change on the line?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “When firmware that touches coil control, object detection or thermal fold-back is changed, and when a coil vendor, a coil construction, or a housing that moves the coil relative to the surface is changed. Those events should reopen the agreed FOD method and may require the certification or construction file to be revisited. They are not equivalent to a colour change. Production on the new construction should not resume until the method has been run and the record filed. A verbal assurance that “FOD is still the same” is not that record.”
}
},
{
“@type”: “Question”,
“name”: “Will this article provide the WPC temperature or object-size limits to put in a specification?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “No. Official pass/fail numbers belong in the current WPC Qi or Qi2 documentation, or in a laboratory procedure that cites that documentation. This article will not invent unofficial thresholds, typical pass rates, or model-specific FOD performance. An OEM clause should reference the current official method, name the object set and the required response, and leave any numeric limit to a source the factory can show. A blank cell waiting for that source is preferable to a memorable figure with no clause behind it.”
}
}
]
}

Related Posts