A design is a set of promises about the parts in it: this transistor withstands that voltage, this capacitor tolerates that ripple, this part is rated for that temperature. When a part is not what its marking claims, the design's margins are gone and the failure that follows can look inexplicable — because the analysis is being done against a datasheet that does not describe the part actually fitted.
What 'counterfeit' actually covers
The term spans several quite different situations, and the distinction matters because each implies different responsibility. Recovered or salvaged parts pulled from scrap assemblies and resold as new. Re-marked parts, where a lower-grade or different device is relabelled as a higher-grade one. Parts from an unauthorised production run. Refurbished parts sold as new. And outright fakes with no functional relationship to the marking.
Adjacent to all of these is the out-of-spec but genuine part: authentic, correctly marked, and simply operated outside its ratings by the design. That is a design finding rather than a supply-chain one, and confusing the two sends an investigation in the wrong direction.
External examination comes first
A great deal is visible without destroying anything. Marking permanence, font and laser-mark quality, evidence of sanding or resurfacing beneath a re-mark, inconsistent date and lot codes across a reel, package dimension and lead coplanarity against the manufacturer's drawing, and the condition of leads — reused parts often show solder residue or reworked plating.
Solvent resistance testing of markings is a standard early step for exactly this reason: a genuine laser mark and an inkjet re-mark behave very differently. None of this is conclusive alone, but a cluster of anomalies justifies the destructive work that follows.
Inside the package
Radiography compares internal construction — die size, lead frame, bond wire count and routing — against a known-good reference of the same part number. A die visibly different in size or a bond pattern that does not match is decisive and requires no cutting.
Decapsulation exposes the die for direct inspection of die markings and manufacturer identifiers, which frequently disagree with the package marking on a re-marked part. Elemental analysis of the package compound and lead plating adds a further comparison. Throughout, the value depends on having an authentic reference part to compare against.
Electrical characterisation
Parts can also be tested against the parameters the datasheet promises, across the temperature range claimed. A device marked as an automotive or industrial grade that fails parametrically at the extremes of that grade is out of specification whatever its package says, and commercial-grade silicon re-marked to a higher grade shows up precisely here.
This is also the step that connects a suspect part to the observed failure. Establishing that a part was counterfeit is one thing; establishing that its deficiency explains what happened is the step that makes the finding matter.
The paper trail is half the case
Physical findings are considerably stronger alongside procurement records. Whether the part came from the manufacturer or an authorised distributor, or from a broker on the open market; whether certificates of conformance exist and trace to a real production lot; whether incoming inspection was performed and to what standard.
Open-market purchasing is where most counterfeit exposure originates, and it usually happens for an understandable reason — an allocation shortage, an obsolete part, a schedule that could not wait. That context is discoverable and is often the pivot of the matter.
Population, not just the failed unit
One failed part establishes very little about a population. Sampling other units from the same lot, the same reel and the same assembly date separates a single anomaly from a systemic supply problem, and it changes the scope of any recall or remediation.
It also guards against the opposite error. A genuine part that failed for an ordinary reason should not become a counterfeiting allegation because a single specimen looked unusual, and comparison across a population is what prevents that.
Chain of custody
Because these findings can support serious allegations against named suppliers, handling discipline is proportionate. Parts identified, bagged and tracked individually; the assembly documented before desoldering; the location of each part on the board recorded so a finding can be tied to a specific circuit position.
Desoldering itself alters lead condition, which is one of the diagnostic features. Documenting leads before removal preserves evidence that removal consumes.
Standards give the sequence a name
This work is not improvised. Published inspection standards, most prominently the AS6081 and AS5553 family for avionics and defence supply chains, set out a graded sequence: documentation review, external visual, marking permanence, X-ray, decapsulation, and electrical test.
Following a recognised sequence matters for two reasons. It puts the destructive steps last, so a finding is not foreclosed by an early mistake. And it makes the work reviewable by another examiner, which is what turns an observation into evidence rather than an assertion about something now destroyed.
What it usually establishes
Counterfeit findings tend to resolve to procurement practice and incoming inspection rather than to design or manufacture. The relevant questions become who authorised an open-market purchase, what verification was specified, and whether the inspection performed was capable of catching what was found.
Where no incoming inspection existed for parts bought outside the authorised channel, that absence is generally the finding, and it is a documentary one.
This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.