Un test d'acceptation en usine n'est pas une brève démonstration de la machine. Pour un système de lissage, FAT doit vérifier la famille de pièces convenue, dessin contrôlé, méthode de mesure, voie de correction, acceptation à l'état libéré, sécurité, enregistrements et processus d'ouverture des articles avant expédition. Il ne remplace pas l'acceptation du site, vérifications de l'utilitaire/de l'interface client, validation finale de l'installation ou preuve de production soutenue.
Ce guide est un cadre contractuel et qualité. Il ne prétend pas que StraighteningTech utilise un protocole FAT particulier, fournit tous les documents ou peut accepter une machine sans accord spécifique du client. Le contrat, URS, revue des risques et plan de réception contrôlée définissent le périmètre final.
*Illustration du concept d'ingénierie. Il représente un contexte de processus FAT candidat, pas la preuve qu'une machine spécifique, la fonction de sécurité ou le résultat de performance a été accepté.*
Geler les entrées FAT avant de tester
Confirmer la configuration de la machine, échantillons de pièces contrôlés, révisions de dessins, étape du matériau/du processus, données, jauge client, caractéristiques d'acceptation, zones de contact autorisées, conditions de test cibles, champs de données, exigences de sécurité et signataires responsables. Un test ne peut pas prouver une exigence indéfinie.
| Entrée FAT | Pourquoi c'est important |
|---|---|
| Échantillons représentatifs et limites | Démontre la famille convenue, non seulement une partie de démonstration facile |
| Jauge client/approuvée | Connecte les lectures de ligne à la méthode d'acceptation prévue |
| Dessin et référence | Définit la caractéristique réellement vérifiée |
| Révision du processus/recette | Empêche un test d'utiliser des paramètres non approuvés |
| Portée de sécurité et de manipulation | Garantit que le test couvre les protections, verrouillages et mouvement de la pièce |
What a FAT Can and Cannot Certify
The honest scope of a factory acceptance test is narrow and valuable: it certifies that, on the agreed day, with the agreed parts, gauges and witnesses, the system performed the defined sequence and produced the defined records. It does not certify sustained capability — that accrues from weeks of production data after installation, which is what site acceptance and early-production reviews sample. It does not certify every future part variant — each new family carries its own qualification. And it does not certify the customer’s own processes — staffing, material flow, gauge maintenance — which determine much of real-world performance, comme discuté dans manual versus automatic straightening. Writing this boundary into the FAT plan protects both sides: the buyer knows what evidence they are buying, and the supplier is not held to an open-ended demonstration.
Designing the Challenge Parts
The weakest FATs demonstrate the machine on the easiest available parts; the strongest exercise it on parts chosen to probe the edges of the agreed scope. Designing those challenge parts is a joint exercise with a short checklist of its own. Geometric edge: include at least one part with the worst incoming distortion the production contract expects, because a machine that only proves itself on moderate bends has not demonstrated its working range. Family edge: if the contract covers a range of diameters or sections, the FAT samples both ends of the range, not the comfortable middle. Event edge: deliberately include a part that must be routed hold or NOK, to prove the exception path — many machines that pass every good part have never demonstrated what happens with a bad one until it matters. Condition edge: reseating, a changeover within the agreed scope, and a recovery from an interrupted cycle each test the state machine rather than the press. None of these is a trap; every one is a scenario production will deliver within the first month, and the factory floor is the cheapest place to see the response for the first time.
Vérifier la mesure et l'acceptation à l'état libéré
FAT devrait enregistrer la géométrie entrante, configuration des données, stations de mesure, événements de correction, réévaluation à l'état libéré, inspection de la surface/des caractéristiques et comparaison avec la méthode de référence approuvée. A reading under force or temporary restraint is not a final acceptance result unless explicitly defined by the agreement.
*Illustration du concept d'ingénierie. It emphasizes released-state verification; it is not a customer FAT record or proof of a universal acceptance method.*
Utiliser mesure de rectitude en charge ou en libération and machine gauge versus customer gauge correlation to define the measurement boundary.
Test the Process, Not Only the Best Part
Include representative normal parts and agreed challenge/boundary parts. Challenge conditions may include a declared geometry condition, variation in the allowed product family, a valid changeover, a controlled hold/NOK event, a reseat check or another customer-approved scenario. The purpose is not to force a failure; it is to verify the defined response and records.
*Illustration du concept d'ingénierie. It shows workpiece-specific method selection, not an assertion that every candidate variant belongs in the FAT scope.*
Check Safety, Routes de données et d'exceptions
Vérifier le gardiennage convenu, fonctions d'arrêt, manipulation des garanties, alarmes, enregistrements de données, identification de la recette/version, fonctions de traçabilité là où elles sont sous-traitées, et routage PASS/NOK/hold. Un affichage indiquant un état ne suffit pas: le FAT devrait définir ce qui a été contesté, la réponse observée, les preuves retenues et qui les a approuvées.
Pour les contrôles associés, voir Limites de tri et de reprise NOK and straightening-line traceability and DMC.
Enregistrer les écarts et les éléments non soldés
Chaque écart doit identifier l’exigence, preuve, impact, propriétaire, action corrective, date d'échéance et méthode de clôture. L'achèvement du FAT ne signifie pas que tous les écarts disparaissent; cela signifie que la disposition convenue est documentée. Éléments ouverts qui affectent la sécurité, la réception ou le plan d'installation du site doit être reporté au SAT et à la remise finale.
The Measurement Segment Deserves Its Own Agenda Item
Between the mechanical demonstration and the paperwork, the measurement segment of a straightening FAT is where acceptance substance lives, and it deserves explicit time on the agenda rather than leftovers. Three demonstrations earn their place. The reseat demonstration: a witness part measured, unloaded, reloaded and remeasured, with the spread on record — this is the machine’s measurement floor, and every later agreement about tolerances silently rests on it. The released-state demonstration: parts remeasured after correction force and restraint are removed, par loaded versus released discipline, so the report values are the values that ship. The correlation demonstration: the same parts read on the machine and on the customer’s reference method, with the comparison on record. Whether a full jauge R&étude R belongs inside the FAT or follows as a site activity is a scoping decision — but a FAT that measures nothing, measures only under force, or measures only with the supplier’s own gauge has documented a demonstration, not an acceptance.
Recognizing a Weak FAT Before You Sign It
Weak acceptance tests share a recognizable signature, and naming it helps buyers push back and suppliers self-correct. The golden-part pattern: the same easy workpiece, prepared in advance, demonstrated repeatedly while nothing from the contract’s boundary conditions appears. The undefined-condition pattern: readings quoted without stating support, restraint or gauge — numbers floating free of method. The exception-path silence: the test ends without ever routing a hold or NOK part, so the sorting logic and its records are untested at handover. The retrospective paperwork: records assembled after the test from memory, which fails the first audit that asks for contemporaneous evidence. And the disappearing open items: deviations acknowledged verbally at the test that surface in no document, no owner and no SAT carry-forward. Any one of these is a reason to extend the FAT, not a reason to renegotiate its meaning afterward.
Liste de contrôle FAT
- échantillons contrôlés, dessins, jauges et méthode d'acceptation disponibles;
- configuration, montage/outillage et révision de la recette identifiés;
- démonstration de la méthode de mesure/donnée et de l'état relâché;
- parties de limites/défis normales et convenues enregistrées;
- protection des surfaces/éléments et vérifications indépendantes requises traitées;
- sécurité, manutention, alarmes et réponse d'exception testées;
- enregistrements, traçabilité/exportations là où les exigences de contrat et de sauvegarde ont été examinées;
- écarts, éléments ouverts, propriétaires et liste de report SAT documentée;
- les autorités qualité du client/fournisseur signent le rapport convenu.
Pour une installation sur site et un fonctionnement continu, use straightening machine SAT checklist after FAT. Pour des exemples de preuves, voir Test d'échantillon de redressage et acceptation.
FAQ
FAT prouve-t-il la capacité de production finale?
Non. FAT vérifie la portée des tests convenus en usine. Conditions du chantier, interfaces client, installation, les jauges et la production soutenue nécessitent un SAT et des preuves supplémentaires.
Un court essai à sec peut-il remplacer l’acceptation de la pièce?
Non. FAT a besoin d’échantillons contrôlés, datum/gauge method, released-state verification and documented response to the agreed conditions.
Can an open item be ignored after FAT?
Non. It must have a documented owner, risk/disposition and closure route, often carried into SAT.
How many parts should a straightening FAT process?
Enough to cover the agreed matrix, not a fixed count: representative normal parts from the family, the geometric and family-edge challenge parts, and at least one part routed to hold or NOK to demonstrate the exception path. The contract scope defines the matrix; the test plan should list each part and what it is there to demonstrate.
Should a gage R&R study be part of the FAT?
It depends on scope and schedule. The FAT should at minimum demonstrate reseating repeatability, released-state measurement and correlation to the customer gauge. A full repeatability-and-reproducibility study is often better executed at site with the customer’s own operators and environment — but then it must appear in the SAT plan, not disappear between the two documents.
Who should witness the FAT?
The buyer’s quality function, not only procurement. The people who will own acceptance records, maintain the gauges and answer for the line’s disposition logic are the right witnesses for reseat spreads, released-state readings and exception-path behavior — and their signature on the report is what makes it an acceptance rather than a demonstration.