October 30, 2026 · Technical Guide
The picture most buyers carry of a counterfeit connector is a plastic shell that crumbles in the hand. That is the least common case and the easiest to catch. The common case is a real connector, made on a real production line, whose history has been edited. It was pulled from decommissioned equipment, cleaned, re-marked and re-bagged. Or it is an overproduction run sold outside the manufacturer’s authorised channel. Or it is a commercial-grade part re-marked with a military part number and a finish letter it never earned.
None of those parts is fake. Only the paperwork is. And at 25 °C on a test bench there is no measurable difference between them and a new part, which is why every control that depends on measuring the part at goods-in is structurally incomplete.
That is the argument of this article: for connectors, the decisive evidence is provenance, and the decisive test is destructive. Our receiving inspection guide covers the AQL plan, the eight-document set and the bench electrical checks — what to do when a lot arrives. This article covers the layer underneath it: how a counterfeit connector is actually constructed, which standard binds which party, what a certificate does and does not prove, and what has to be cut open before anyone can honestly call a lot authentic.
Screening practice was built for integrated circuits, and most of it does not transfer. A connector is a mechanical assembly whose electrical performance is its easiest property to satisfy: clean gold on clean gold is clean gold, whether the shell was made last month or reclaimed from a scrapped radar rack.
“Counterfeit” is a category, not a single act. The five types below have different tells, different economics and different controls, and a programme that only looks for one of them will miss the others.
| Type | What it actually is | Tell on a connector | What only a destructive test finds |
|---|---|---|---|
| Recycled / refurbished | Recovered from decommissioned equipment, cleaned, contacts dressed, re-bagged as new | Polished coupling threads, burnished keyways, two date codes in one bag, faint or doubled marking, solder or crimp residue in a contact barrel | Plating worn through at the contact wipe zone; elastomer already compression-set beyond the specification limit |
| Re-marked / up-graded | A lower environmental class, a commercial part, or a different finish sold under a military part number | Marking font, layout or placement differing from the qualified drawing; ink where the drawing says laser; a finish colour that does not match the class letter | Plating chemistry and per-layer thickness (XRF plus cross-section); insert polymer that is not the specified compound |
| Overproduction / unauthorised channel | Genuine, fully conforming hardware sold outside the manufacturer’s authorisation | None. Every visible and measurable property is correct | Nothing. Every test passes; only the channel and custody chain can exclude it |
| Cloned / unauthorised second source | A reverse-engineered copy, often dimensionally close but made to a different process | Keyway and thread form at the edge of tolerance, coarser insert cavity finish, different contact spring rate, shallower accessory thread | Shell alloy (a casting presented as wrought 6061-T6), contact base metal, plating stack, elastomer compound |
| Tampered / up-screened | A genuine part modified to claim a rating it was never qualified for: re-plated, re-terminated, re-lubricated, re-dated | Non-original plating on threads, over-polished seal surfaces, a boot or grommet that does not match the drawing, unspecified lubricant | Plating adhesion and thickness, seal-surface geometry, lubricant identity, seal compression set |
Two rows deserve emphasis. Overproduction is undetectable by inspection, which means no test programme — however deep — substitutes for buying from a source you have qualified. And tampering is the normal case in legacy-platform support, where the original manufacturer has exited, the drawings are decades old, and the aftermarket is the only place a part number still exists. That is precisely the population where a re-plated shell circulates without anyone intending fraud.
The counterfeit-avoidance documents are frequently quoted as a single block, as if holding one implied the others. They do different jobs, and a purchase order that cites the wrong one buys the wrong assurance.
| Document | Who it binds | What it actually requires | What it does not do |
|---|---|---|---|
| AS5553 | Producers — manufacturers and their supply chains, where the organisation controls design and process | A counterfeit avoidance control plan: authorised sourcing, purchasing controls, receiving inspection, in-process detection, reporting, quarantine and disposition, and flow-down to sub-tiers | It certifies no individual part. It describes a system, and a system can be documented without being operated |
| AS6081 | Independent distributors, who do not manufacture and therefore cannot control the process | The same discipline adapted to characterising risk: supplier approval, traceability, and defined inspection and test sampling where provenance is incomplete, with the test plan scaled to the risk | It tells you how a distributor screened a part, not that the part was ever in the authorised channel |
| AS6171 | Whoever is doing the testing | A family of test methods, organised as slash sheets, each defining one method together with the defect classes it can detect and its relative sensitivity | It does not choose the method for you. Selecting methods against the specific risk is the buyer’s engineering decision, not the standard’s |
| Reporting programmes (GIDEP, ERAI) | The whole industry | Disclosure of suspect parts, suspect sources and observed failure modes so that others can act before they buy | They test nothing. They are an early-warning network, and they only work if buyers report |
| Qualified-source listings (DLA QPL/QSL, OEM approved-supplier lists) | The design authority and its procurement chain | Answers “is this a source the programme has already accepted for this specification” | It does not prove the unit in your hand came from that source, or from that source’s authorised distribution |
The distinction that costs programmes the most money: AS6081 governs how a distributor tests, not whether the part was ever authorised. A distributor can hold the certificate, run a defensible test plan, and still deliver an overproduction lot, because nothing in the standard asks where the material came from beyond the paperwork the seller provided. Conversely, a franchised distributor’s invoice answers the channel question with no test at all. Buyers routinely collect the test evidence and skip the channel question, which is exactly backwards.
Provenance is a chain of links. Each link is easy to obtain and easy to defeat, and the chain is only as strong as the hop nobody checked.
| Evidence | What it proves | How it is defeated |
|---|---|---|
| Manufacturer or franchised-distributor invoice naming the lot | The material left the manufacturer’s authorised chain | A forged invoice, or a genuine invoice for a different quantity than the one shipped |
| Certificate of conformance naming part number, lot and quantity | Someone attests conformance for that population | It is an attestation, not a measurement; the weight depends entirely on the signer’s scope |
| Date or lot code on the part and on the packaging, mutually consistent | The parts and the paper are the same population | Re-marking, or bagging several lots into one |
| Original sealed packaging: bag, desiccant, humidity indicator, ESD bag for PCB-tail parts | The parts have not been opened, handled or re-bagged | Resealing is trivial and cheap |
| Chain of custody across every hop, named | Nothing was substituted between manufacturer and dock | It breaks silently at the broker hop, and the paperwork still looks complete |
| QPL/QSL or approved-supplier status of the seller | The source is acceptable to the design authority | The unit may not be from that source at all |
| Supplier-side traceability to the plating and elastomer sub-tier | The finish and compound are what the drawing says | Only as good as the supplier’s own records, which you cannot see unless you ask to audit |
The practical rule is that the certificate answers a question about a population, while your exposure is a question about a unit. Everything in the table above narrows the gap; nothing in it closes the gap. That is not an argument for abandoning paperwork. It is an argument for treating paperwork as a filter that decides how much destructive testing a lot deserves.
No single method detects counterfeits. The methods form a ladder of increasing cost and increasing reach, and the honest way to use it is to read the “blind to” column before deciding you have covered the risk.
| Method | Detects | Blind to | Destructive |
|---|---|---|---|
| Visual and marking inspection at 10× to 30× | Marking style, placement and font against the drawing; finish colour; thread condition; insert cracks; contact tines; keyway damage; boot and grommet identity | Plating thickness and composition, elastomer compound, base alloy, internal assembly | No |
| Marking permanency (solvent rub to the applicable method) | Re-inking over old marking, ink-jet applied over a previously marked surface, cheap surface printing | A well-executed laser re-mark, which is now the common method | No (on a sample) |
| X-ray fluorescence | Plating chemistry and approximate thickness; cadmium, lead and other restricted substances; a paint film presented as a plated finish | Plating adhesion and porosity; the alloy beneath the plating; anything under a thick layer | No |
| Radiography | Internal voids, mis-seated contacts, insert assembly errors, solder voids in PCB-tail parts | Plating, elastomer, thread class, heat treatment | No |
| Electrical suite: low-level contact resistance, insulation resistance, withstand voltage, shell-to-shell resistance, contact retention | Gross defects, contamination, wrong contact fit, open cavities, a broken ground path | Anything that ages. A used part with a compressed seal and 40 % of its plating left will pass every one of these | No |
| Mechanical suite: insertion and withdrawal force, keying and polarisation check, thread gauging, accessory-thread depth | Wear, deformation, wrong thread class, a damaged key or keyway, a cloned shell at the tolerance edge | Alloy and finish; anything time-dependent | No |
| Cross-section and metallography | Plating per-layer thickness and adhesion, base-metal alloy and temper, thread form, shell wall thickness, insert cavity walls, glass-to-metal seal meniscus | Elastomer identity; anything you did not cut | Yes |
| Elastomer identification (FTIR, DSC, TGA) | Compound family and cure state against the qualified material; a substituted O-ring or grommet | How much useful life is left; that needs compression-set data, not identity | Yes (sample) |
| Seal and environmental confirmation: mated leak or pressure test, salt spray, thermal cycling | The actual function under the actual stressor — the only method that answers “will this survive the environment” | Weeks to months of calendar time, which is why it is almost never run at goods-in | Yes, for salt spray |
| Functional life sampling: mating cycles to failure, retention after vibration | Wear state and the remaining margin against the rating | Statistical power — a sample of two tells you about two parts | Yes |
Read the third column as a whole. Rungs one to six are what a competent goods-in department runs, and none of them can see plating thickness, plating adhesion or elastomer compound — the three properties that decide whether the connector survives salt spray, thermal cycling and ten years in a bay. That is the entire argument for destructive sampling, and it is why a lot with impeccable paperwork and a clean incoming inspection can still be the lot that fails in year three.
Destructive physical analysis is not a single test. On a circular connector it is a set of cuts chosen to reach the properties the non-destructive rungs cannot see. A defensible sample is small — one or two pieces per lot — but it must come from the lot being accepted, not from a “reference” piece supplied separately.
The economics are straightforward once the alternative is stated properly. DPA consumes one or two units per lot; an escaped counterfeit lot consumes a build, a retrofit or a recall. The mistake is applying it uniformly: the right pattern is to run it on the first lot from any new source, on any lot whose channel story is thin, and on a periodic sample thereafter — scaled by criticality, not applied as a ritual to every delivery.
The instinct to send a suspect lot back to the seller is the instinct that keeps counterfeits in circulation. The part goes back into a channel that has already demonstrated it will resell unverified material, and the next buyer has no warning at all.
None of the above survives contact with a commercial dispute unless the purchase order says it. These ten clauses are the practical core of a counterfeit control plan at the buyer’s end.
The question to ask a supplier changes with the family, because what a counterfeit would have substituted changes with the family:
| Family | What to ask to see | Why that is the question |
|---|---|---|
| J599 — MIL-DTL-38999 equivalent | The plating class in writing, the elastomer compound, and lot traceability from shell to plating sub-tier | The finish letter is the environmental class; it is the single most commonly upgraded claim |
| ZH83723 — MIL-DTL-83723 equivalent | Grommet compound and mating-cycle data, with the wear figure stated separately from the electrical rating | Bayonet and coupling wear is where a recycled part shows its history |
| MIL-DTL-26500 equivalents | Whether the parts are current production or legacy stock, with date codes named before order | Legacy platforms are the natural habitat of aftermarket and re-marked parts |
| M85049 backshells | Boot compound, plating on the accessory, and the adapter’s thread class | Accessories are where re-plated and re-bagged hardware circulates most freely |
| Marine and offshore | A salt-spray report naming the exact finish, not the family | The claim under attack in marine work is always the finish |
| Specialty and hermetic | Leak-rate basis and glass-seal process control, not just a leak figure | Hermeticity cannot be inspected from outside; the process record is the evidence |
| Circular connectors | The marking specification and the date-code convention used in production | So that what arrives can be checked against a documented standard rather than a habit |
Every connector we ship carries a lot code traceable to raw material, plating and test record — the traceability statement on our quality page is the specific document to test us against. If you are qualifying a source, ask for the traceability chain before you ask for the price, and ask for a DPA sample on the first lot. A supplier who finds that request unreasonable is telling you something useful.
Usually not. A recycled or up-screened connector that meets its contact-resistance, insulation-resistance and withstand-voltage limits at ambient is, electrically, indistinguishable from a new one. What has actually been lost is plating thickness, elastomer elasticity and wall thickness at the seal, and those show up as failures months or years later under salt spray, thermal cycling or vibration — not on the bench at goods-in.
Destructive physical analysis: a sample from the lot is cut, polished and measured so the plating stack, base metal, thread form, insert cavities, contact barrel and elastomer can be compared against the specification. Ask for it whenever provenance is incomplete, whenever price or lead time does not match the authorised channel, on the first lot from any new source, and on any long-life or safety-critical programme where a latent seal or plating failure cannot be recovered from.
A certificate is an attestation, not a measurement. It answers who says this conforms and part of which lot it is, but it says nothing about plating thickness or compound identity, and it is only as strong as the scope of the entity that signed it. A certificate from a trading company naming a manufacturer it does not represent carries almost no evidentiary weight.
AS5553 applies the counterfeit avoidance, detection, mitigation and disposition discipline to producers and their supply chains, where the organisation controls the design and the manufacturing process. AS6081 applies the same discipline to independent distributors, who do not manufacture and therefore have to characterise risk through inspection and test instead of process control. Buying to AS6081 tells you how a distributor screens parts; it does not tell you the parts were ever in the authorised channel.
Generally no. Returning a suspect lot to its source puts it back into circulation and is the mechanism by which bad parts move between brokers. Quarantine it in place, segregate it physically and in the records, notify your customer and the relevant reporting programme, and dispose of it by controlled scrap rather than by credit note. Return only when the seller is the original manufacturer or a franchised distributor acting on their instruction.
Yes. Overproduction and out-of-warranty stock is genuine hardware sold outside the manufacturer’s authorisation. No inspection or test can detect it, because nothing about the part is wrong. The only control is the channel and the custody chain, which is why provenance evidence — not test coverage — is the primary defence.
Qualifying a source for a MIL circular family, or holding a lot whose paperwork does not add up?
Send us the part number, the finish class, the lot or date code and whatever documentation came with the material. We will tell you what the traceability chain should contain for that family, which characteristics a DPA sample has to measure, and what the plating and elastomer records behind the part actually are.
Email: info@airoadcon.com | WeChat: 18991927716