Connect981 – Content Dev

RSC Topic: Equipment & System Qualification

  • first article build

    A first article build commonly refers to the initial production build of a part, assembly, or product configuration used to confirm that the released design, planned manufacturing process, tooling, materials, and work instructions can produce the intended result. It is typically associated with the transition from design or setup into controlled production.

    The term describes the build activity itself, not only the inspection records generated from it. In regulated and quality-driven manufacturing, the first article build often provides the physical basis for downstream review activities such as dimensional verification, configuration checks, process confirmation, and formal first article inspection documentation where required.

    What it includes

    • The first planned build from approved drawings, specifications, and routing
    • Use of intended materials, tools, fixtures, equipment, and manufacturing methods
    • Verification that the product can be built consistently to the defined requirements
    • Collection of production and quality evidence that may support inspection, traceability, or process validation activities

    What it does not mean

    A first article build is not the same as an engineering prototype, lab sample, or informal trial unless those items are explicitly controlled as the initial production-representative build. It is also not identical to first article inspection. The build creates the item, while first article inspection is the review and verification activity performed on that item and its associated records.

    Operational meaning

    In manufacturing systems, a first article build may appear as a flagged work order, traveler step, routing status, or quality hold point. It is often linked to document control, revision status, material lot traceability, inspection results, and approval workflows so the organization can distinguish the initial production-representative unit from routine production.

    For example, when a new aerospace component is released, the first article build may be the first serialized unit produced under the intended process, with operators, inspectors, and planners capturing evidence needed for quality review and production release decisions.

    Common confusion

    First article build vs. first article inspection: the build is the act of making the item; the inspection is the act of verifying that item against defined requirements.

    First article build vs. prototype build: a prototype is often used for design learning or testing and may not follow the final production process. A first article build is usually expected to be production-representative.

    First article build vs. pilot run: a pilot run may involve multiple units to test readiness or flow. A first article build usually refers to the initial unit or initial build event used for that confirmation.

  • Does moving production to a different machine always trigger a delta FAI?

    No.

    Moving production to a different machine does not automatically mean a delta FAI is required in every case. What matters is whether the move represents a change that could affect product characteristics, process capability, or the approved manufacturing method, and whether your customer, contract, or internal quality system treats that change as FAI-impacting.

    In practice, this connects to digital AS9102 FAI when teams need to turn the answer into repeatable execution habits.

    In practice, a machine change often can trigger a delta FAI, but not simply because the asset ID changed. The decision usually depends on factors such as machine equivalence, control software differences, fixture changes, tooling changes, CNC program transfer risk, process parameter changes, operator method changes, and whether the process is considered the same validated or qualified process after the move.

    What usually drives the decision

    • Customer or contract-specific requirements: Some customers are stricter than the baseline expectation and may require delta FAI for machine moves that another customer would not.

    • Whether the manufacturing process changed in a meaningful way: A move between truly equivalent machines with the same tooling, program revision, setup method, and demonstrated capability may be treated differently from a move to a different machine platform or control.

    • Impact on product characteristics: If the new machine could influence dimensions, surface finish, hole quality, material condition, or other characteristics, a delta FAI becomes more likely.

    • Qualification and validation status: In regulated and aerospace environments, process validation, equipment qualification, and change control can matter as much as the part drawing itself.

    • Risk classification of the part and process: Critical features, special processes, or tight capability margins raise the burden of evidence.

    When a delta FAI is more likely

    • The replacement machine is a different model, control platform, kinematic configuration, or capability class.

    • Tooling, fixturing, probing routine, offsets, or setup method changed.

    • The CNC program was reposted, modified, or adapted for the new machine.

    • Inspection results or capability on the new machine are not yet established.

    • The part has critical or tightly toleranced features sensitive to machine behavior.

    • Your internal procedure or customer flowdown explicitly lists machine changes as FAI-triggering events.

    When a delta FAI may not be required

    • The new machine is formally controlled as equivalent and uses the same approved method.

    • Tooling, program revision, setup, process parameters, and inspection method remain unchanged.

    • You have documented change assessment showing no expected effect on form, fit, function, or process capability.

    • Your quality system and customer requirements allow a justified no-delta decision with objective evidence.

    That said, a justified no-delta decision needs more than tribal knowledge. It typically requires documented review, traceable rationale, and supporting records. If the only basis is that the machines are “basically the same,” that is usually weak in an audit or customer review.

    Brownfield reality

    In mixed-vendor plants, this decision is often harder than it should be because machine identity, CNC revision, tooling records, setup instructions, inspection plans, and FAI history live in different systems or paper files. MES, ERP, QMS, PLM, and CMM software may not agree on what actually changed. That integration gap does not eliminate the requirement. It just means engineering and quality need a more disciplined change assessment and evidence trail.

    This is one reason full system replacement is rarely the practical answer. In long-lifecycle regulated environments, replacing core execution and quality systems to “clean up” FAI decisions often creates more qualification burden, validation work, downtime risk, and traceability gaps than the original problem. Coexistence with stronger change control and clearer system-of-record definitions is usually more realistic.

    Practical answer

    If production moves to a different machine, treat it as a controlled change and assess whether it could affect the part or approved process. Do not assume the answer is always yes, and do not assume it is automatically no. Review the contract, customer expectations, internal FAI procedure, machine equivalence, process risk, and available objective evidence before deciding whether a delta FAI is required.

    If there is any ambiguity, escalate the decision through quality and the responsible customer-facing authority rather than relying on an informal shop-floor judgment.

  • Can AI change qualified process limits in aerospace without re-qualification?

    No, not as a general rule.

    If the limits are part of a qualified process, validated workflow, approved manufacturing method, or controlled inspection regime, an AI system should not change them on its own and keep the process treated as still qualified. In most aerospace environments, changing those limits is a controlled change. Whether that requires full re-qualification, partial re-qualification, re-validation, engineering approval, customer approval, or internal review depends on the process, the product, contractual requirements, and how the limits are tied to product conformity.

    In practice, this connects to qms integration and evidence trails when teams need to turn the answer into repeatable execution habits.

    The key point is simple: AI can support analysis and propose changes, but autonomous modification of qualified process limits is usually not acceptable unless the operating model, controls, and approval pathway were explicitly designed, validated, and approved for that behavior.

    What AI can usually do

    • Monitor trends and detect drift earlier than manual review.

    • Recommend tighter controls, maintenance actions, or investigation triggers.

    • Simulate likely effects of a parameter change before any production use.

    • Help classify events, prioritize review, or flag out-of-family conditions for engineering or quality.

    • Operate within fixed approved guardrails if those guardrails are clearly defined, technically enforced, and covered by change control and validation.

    What usually triggers re-qualification or equivalent review

    • Changing process windows or control limits that affect fit, form, function, strength, durability, or other critical characteristics.

    • Changing inspection thresholds, acceptance logic, sampling logic, or measurement interpretation that influences disposition decisions.

    • Changing machine recipes, CNC offsets, cure cycles, coating parameters, torque ranges, or similar controlled parameters beyond approved tolerance bands.

    • Allowing a model to adapt itself in production without a locked version, documented rationale, and approved deployment record.

    • Using data of uncertain quality, incomplete lineage, or weak traceability to justify process changes.

    In practice, the more directly an AI output can alter product realization or acceptance, the stronger the expectation for review, traceability, validation evidence, and controlled release.

    Important boundary conditions

    There are narrow cases where not every change means full re-qualification. For example, some plants define pre-approved operating envelopes, adjustment rules, or advisory-only optimization logic that operators or engineers can use without re-qualifying the entire process each time. But that only works when the boundaries are explicit, justified, documented, and enforced. If AI crosses those boundaries, changes the boundaries themselves, or changes how acceptance is determined, the burden goes up quickly.

    This is also highly configuration-dependent. A model that recommends a parameter change for human approval is very different from a closed-loop controller that writes directly to equipment setpoints. The second case carries much higher validation, cybersecurity, traceability, and operational risk.

    Brownfield reality

    In aerospace plants, AI rarely operates in a clean, standalone stack. It has to coexist with MES, ERP, PLM, QMS, historian, SCADA, machine controllers, and document control systems that were not designed for adaptive models. That creates practical constraints:

    • Approved limits may exist in multiple systems, and inconsistency creates execution risk.

    • Audit trails may be fragmented unless integration is done well.

    • Legacy equipment may not support granular permissions, rollback, or modern model governance.

    • Downtime windows are limited, so even technically sound changes can be operationally hard to deploy.

    This is one reason full replacement strategies often fail in long lifecycle, regulated environments. Replacing execution and quality systems to make autonomous AI easier usually runs into qualification burden, validation cost, downtime risk, integration complexity, and the need to preserve traceability and change history across legacy assets.

    Safer implementation pattern

    A more realistic pattern is to use AI first for advisory decision support, not autonomous limit changes. That means:

    • Lock model versions and training data sources.

    • Require engineering and quality approval before parameter updates take effect.

    • Record who approved what, when, why, and against which evidence set.

    • Keep rollback mechanisms and effective dating for changed limits.

    • Separate process monitoring from process authority.

    If a company wants closed-loop adjustment, it needs much stronger governance, validation, exception handling, and technical controls than most organizations initially assume.

    So the practical answer is no: AI should not change qualified process limits in aerospace without the applicable controlled review and, where required, re-qualification or re-validation. The exact threshold depends on product criticality, process design, customer and internal requirements, and whether the AI is advisory or authoritative.

  • system and services acquisition

    System and services acquisition commonly refers to the structured process an organization uses to plan, procure, validate, and accept systems and services. In industrial and regulated environments, this includes IT systems, OT assets, software platforms such as MES or ERP, cloud services, and vendor-managed solutions.

    The focus is on making sure that what is acquired is clearly specified, evaluated, and brought into operation under defined technical, security, quality, and compliance requirements. It typically covers:

    • Defining business, technical, security, and regulatory requirements for new systems and services
    • Evaluating suppliers and proposed solutions against those requirements
    • Including appropriate clauses in contracts for support, updates, data handling, and access control
    • Performing security, quality, and interoperability reviews before approval
    • Planning deployment, validation, and acceptance testing
    • Documenting ownership, responsibilities, and lifecycle expectations

    In manufacturing and regulated operations

    In manufacturing, system and services acquisition typically applies when organizations select and onboard:

    • Industrial control systems, PLCs, HMIs, and associated network components
    • Manufacturing IT platforms such as MES, LIMS, QMS, and historian systems
    • Cloud or SaaS services used for production scheduling, maintenance, or quality management
    • Third-party services such as remote monitoring, managed OT security, or data integration services

    Operationally, the acquisition process helps ensure that new systems and services can be integrated with existing OT/IT infrastructure, support required audit trails and data retention, and respect plant safety, cybersecurity, and change control practices.

    Relation to security and supply chain controls

    In security and control catalogs such as NIST SP 800-53, system and services acquisition is treated as a control family governing how organizations specify, evaluate, and approve systems and services with security and supply chain risk in mind. This includes:

    • Requiring security and privacy capabilities in purchased products and services
    • Considering software supply chain risks when selecting vendors and components
    • Ensuring that contracts and service agreements address access, updates, and incident response
    • Requiring documentation needed for audits, traceability, and configuration management

    In brownfield manufacturing environments, these practices are often applied when upgrading legacy OT systems, adding remote connectivity, or onboarding new software suppliers that interact with production networks.

    Common confusion

    • Not the same as general procurement: System and services acquisition is a focused subset of procurement that emphasizes technical, security, lifecycle, and compliance requirements, rather than only price and commercial terms.
    • Different from vendor management: Vendor management looks at the broader relationship with a supplier over time. Acquisition focuses on the specific process of defining, selecting, and accepting particular systems or services.

  • special processes

    Core meaning

    In manufacturing and regulated industries, **special processes** are processes whose resulting output **cannot be fully verified by subsequent inspection or testing**, and therefore must be controlled primarily through:

    – prior qualification of the process,
    – qualification of equipment and facilities,
    – qualification and ongoing approval of personnel, and
    – tightly controlled parameters and documentation.

    The quality of the product is assured by demonstrating that the process is consistently capable, rather than by checking every characteristic of the finished item.

    Common examples include:

    – welding, brazing, soldering
    – heat treatment
    – non-destructive testing (NDT) and some surface treatments
    – plating, anodizing, coating and painting
    – certain composite layup and curing operations

    Use in industrial and regulated environments

    In regulated environments (such as aerospace, medical devices, and pharmaceuticals), special processes typically:

    – require **formal procedures, work instructions, and records** that prove the process was followed as qualified,
    – are often governed by **industry standards or customer specifications**, and
    – are subject to **periodic requalification or reapproval** when equipment, materials, methods, or key parameters change.

    Manufacturing execution systems (MES), quality systems, and ERP integrations may:

    – track which operations in a routing are designated as special processes,
    – enforce that only **qualified operators, certified equipment, and approved materials** are used, and
    – capture **full traceability** of process parameters (e.g., temperature profiles, torque settings, lot numbers).

    Boundaries and what it is not

    Special processes:

    – **are defined by verifiability**, not by importance or cost; an operation can be critical but not a special process if all characteristics can be fully inspected afterward.
    – are **not limited to a specific industry**; the concept appears in aerospace, automotive, energy, medical, and others.
    – are **not the same as critical-to-quality (CTQ) characteristics**, though CTQs often exist within special processes.

    A standard machining operation with measurable dimensions is usually **not** considered a special process if all relevant features can be reliably inspected and defects can be detected without destructive testing.

    Common confusion and misuse

    Special processes are commonly confused with:

    – **Critical or key processes**: Many organizations use “critical process” to mean any operation that strongly influences product performance or safety. A process may be critical but not “special” if its output is fully verifiable.
    – **Special cause variation (SPC)**: In statistical process control, “special cause” refers to a type of variation, not to the concept of special processes.

    When using the term in quality or compliance discussions, it is helpful to confirm whether the intended meaning is **“not fully verifiable by inspection”**, which aligns with how many standards, customers, and auditors use the term.

    Site-context application: aerospace and high-cost waste

    In aerospace and similar highly regulated sectors, special processes are tightly controlled because:

    – parts may rely on **invisible attributes** (e.g., material microstructure after heat treatment, weld integrity) that cannot be checked without damaging the part,
    – process failures can force **scrap or extensive rework** of high-value, previously qualified components, and
    – significant **revalidation, investigation, and documentation** effort may be triggered when a special process is suspected to be out of control.

    As a result, waste or rework involving special processes often has a cost impact far beyond the direct material value, due to lost qualified parts, additional testing, and schedule risk.

  • NDT

    Core meaning

    NDT (nondestructive testing) commonly refers to a family of inspection methods used to detect discontinuities, defects, or material property variations in parts, welds, structures, or assemblies **without** impairing their intended use.

    In industrial and regulated manufacturing environments, NDT is used to confirm product integrity, fitness for service, and compliance with specifications and standards, while leaving the item in serviceable condition.

    Typical methods in manufacturing

    Common NDT methods used in factories and industrial plants include:

    – **Visual testing (VT)** – direct or remote visual examination, often with magnification or borescopes.
    – **Liquid penetrant testing (PT)** – dye or fluorescent liquids applied to reveal surface-breaking defects.
    – **Magnetic particle testing (MT)** – magnetic fields and particles used to find surface and near-surface flaws in ferromagnetic materials.
    – **Radiographic testing (RT)** – X-rays or gamma rays used to image internal features of welds, castings, and structures.
    – **Ultrasonic testing (UT)** – high-frequency sound waves used to detect internal flaws, wall thickness, and bonding.
    – **Eddy current testing (ET)** – electromagnetic techniques for surface and near-surface defects, often on conductive alloys.
    – **Thermography and other advanced methods** – infrared, acoustic emission, phased array UT, and digital radiography, among others.

    NDT may be automated, semi-automated, or fully manual, and often produces both human-readable reports and stored digital inspection records.

    Use in regulated and aerospace manufacturing

    In regulated industries (such as aerospace, nuclear, medical devices, and oil & gas), NDT is typically classified as a **special process**, because the quality of the result cannot be fully verified by later inspection and depends strongly on:

    – Qualified procedures and validated techniques
    – Calibrated equipment and controlled parameters
    – Certified NDT personnel
    – Traceable and reviewable records

    NDT is often integrated with Manufacturing Execution Systems (MES) or quality systems to:

    – Link inspection results to specific parts, lots, or serial numbers
    – Enforce that required NDT processes are performed at defined steps
    – Capture parameter data and images (e.g., radiographs, UT data files)
    – Support electronic review, disposition, and long-term traceability

    Boundaries and exclusions

    NDT **includes** techniques that:

    – Leave the inspected item in a condition suitable for its intended use
    – Are designed to monitor material condition, integrity, or structure

    NDT **does not typically include**:

    – **Destructive testing** (e.g., tensile tests that break samples, sectioning welds, metallographic mounts)
    – **Routine in-process measurements** that alter the part (e.g., coupons sacrificed for testing, samples removed from a batch)
    – **General preventive maintenance checks** that do not use defined NDT methods (e.g., simple visual housekeeping checks)

    Common confusion and alternate uses

    – **NDT vs. NDE vs. NDI**:
    – NDT (nondestructive testing) focuses on the act of testing for defects.
    – NDE (nondestructive evaluation) is often used where quantitative assessment of material properties or remaining life is emphasized.
    – NDI (nondestructive inspection) is a closely related term, often used interchangeably with NDT in aerospace and defense.

    – **NDT vs. quality inspection**:
    – NDT is a subset of quality inspection focused on nondestructive techniques.
    – Other inspections (dimensional checks, gauging, destructive sample tests) are part of quality control but are not NDT unless they meet the nondestructive criterion and use recognized methods.

    Site context: NDT as a special process

    Within the site’s focus on industrial and regulated manufacturing systems, NDT is treated as a **special process** that:

    – Is tightly linked to product release and certification decisions
    – Requires controlled procedures, qualification, and traceable records
    – Often benefits from MES or other digital systems for routing control, data capture, image and report management, and audit-ready traceability across the product lifecycle.