Production Consistency in High-Voltage Cable Assembly Programs

Qualification establishes that a specific design, built from specific materials through specific processes, meets the program's performance requirements. It does not guarantee that the assembly shipped two years later is identical. In high-voltage programs — where performance is determined by controlled geometry, material properties, surface condition, void content, and process stability — production drift is a credible and preventable path to reduced dielectric margin and field failure. The mechanisms are understood, and a controlled production system reduces the risk before it reaches the customer.

Process Control & Configuration Management · 1 kV – 250 kV · AS9100 · IPC-trained workforce · Aerospace · Defense · Scientific · Industrial

Why Qualification Is Necessary but Not Sufficient

Qualification tests a specific assembly — a specific article, built at a specific time, from specific material lots, through specific process conditions. When that assembly passes, the qualification establishes one thing: that this design, as built at this moment, meets the performance requirements.

What qualification does not establish is that the next assembly will be identical. Or the one after that. Or the lot shipped against a follow-on contract 30 months later. The materials may have changed. The process parameters may have drifted. The operators may have turned over. The suppliers may have revised their formulations. None of these changes necessarily produce a discrepant acceptance test result. All of them can degrade the dielectric performance, the service life, or the reliability profile of the finished assembly.

For a standard cable assembly, production drift is often a workmanship or quality concern. For a high-voltage cable assembly, it can become a dielectric reliability concern. The characteristics that most strongly determine whether an assembly will perform reliably — corona inception voltage margin, void content at the termination, insulation geometry through the mold, interface adhesion — are often not visible in the finished part and may not be fully captured by a generic end-of-line test plan unless the correct electrical screens, sensitivity thresholds, and environmental conditions are specified.

Why HV assemblies are more sensitive to production variability

A high-voltage cable assembly's critical performance characteristics are determined by controlled geometry, material properties, surface condition, void content, and process stability — insulation thickness variation of a few thousandths of an inch, a small void at a material interface, a process parameter that drifted by 5% over six months. These variations are often inconsequential in a low-voltage assembly. In a high-voltage assembly, they can change the field distribution, void content, and long-term dielectric reliability in ways that acceptance testing may not detect — but field failure eventually will.

How Production Drift Happens — The Four Primary Mechanisms

Material substitution without engineering review

A supplier changes a raw material — insulation compound, overmold elastomer, potting material — because the original is on extended lead time, a better price is available, or a purchasing decision is made without engineering involvement. The substituted material meets the nominal specification on paper. Its processing characteristics are different: cure shrinkage, melt viscosity, Shore hardness after cure. The finished assembly has a slightly different insulation geometry, void content, or interface adhesion than the qualified article. The change is not documented as an engineering change. The assemblies ship.

This is not a hypothetical failure path. In long-duration programs, unreviewed material substitution is a recurring risk when purchasing, engineering, and quality controls are not tightly connected — and is often identified when a field failure investigation reveals that the current production configuration diverged from the qualified configuration without a corresponding re-qualification event.

Process parameter drift

Overmolding and insert molding processes are controlled by temperature, pressure, and cycle time — and by the lot-to-lot characteristics of the material being processed. Over time, machine parameters drift as tooling wears, calibration intervals lapse, or operators develop informal adjustments that diverge from the documented work instruction. The divergence is gradual and may not produce a detectable change in any single assembly's acceptance test result. The dielectric properties of the molded structure change incrementally. The change becomes detectable only when a failure occurs and the build record is reconstructed.

Personnel change without captured standard work

In low-volume, high-mix manufacturing environments, experienced assemblers carry substantial process knowledge that is not always encoded in work instructions — judgment calls about conductor preparation, tooling setup, when an overmold shot looks right, how much pressure to apply during a manual operation. When a key assembler transfers, retires, or is on extended leave, that knowledge does not transfer automatically. If work instructions do not capture the critical decisions in sufficient detail, a replacement assembler makes reasonable decisions that are different from the qualified process. The resulting assembly may pass acceptance testing and still have a different reliability profile.

Supplier formulation changes without notification

A qualified material supplier may revise their formulation, manufacturing process, or raw material sourcing. For specialty elastomers, silicone compounds, potting materials, and high-voltage wire, these changes can affect cure behavior, adhesion, viscosity, dielectric performance, or long-term aging — sometimes without producing an obvious change in nominal datasheet properties. Supplier change-notification requirements must be flowed down explicitly in purchase agreements and verified through supplier qualification, incoming inspection, and change-control review. AS9100 Clause 8.4 requires that organizations control externally provided products and services and communicate relevant requirements to external providers — including notification of changes. Whether a specific supplier's QMS actually implements this requirement is a matter of verification, not assumption.

AS9100 Certification Is Not a Guarantee of Adequate Controls

AS9100 certification confirms that a supplier maintains a documented quality management system within the scope of its certification and most recent audit. It does not, by itself, prove that the specific configuration management, material traceability, change control, and process records required for your high-voltage program are functioning at the necessary level. These must be verified directly — by reviewing documentation, asking for specific build records, and auditing the specific controls that govern the processes that matter for your assembly. For aerospace and defense programs, an AS9100-certified QMS is often a baseline expectation; it is not, by itself, sufficient evidence that the specific controls required for a high-voltage program are functioning effectively.

What the Standards Require — and Where Practice Often Falls Short

AS9100 Rev D Clause 8.5.1.3 requires production process verification: the organization must verify that production processes can produce product meeting requirements. First Article Inspection is a common implementation method, and AS9102 Rev C provides a widely accepted aerospace framework for performing and documenting FAI — including part number accountability, product accountability, and characteristic accountability, documented in a First Article Inspection Report (FAIR). AS9102 Rev C includes updated expectations around digital product definition and modern documentation practices.

The FAI establishes a production baseline. What AS9100 and AS9102 do not automatically ensure is that the baseline is maintained. The controls that protect the baseline — configuration management, material traceability, engineering change control, and process records — must be implemented as ongoing operational disciplines, not as activities completed once at qualification and then left to decay.

In practice, the gap between what the standard requires and what is functioning often shows up in the same places: purchasing changes to approved materials without engineering sign-off, process adjustments made informally on the production floor, work instructions that describe the nominal process but not the critical parameters, and build records that are retained but not specific enough to reconstruct what actually happened on a given assembly.

What a Controlled Production System Actually Requires

Preventing production drift in a high-voltage cable assembly program requires a connected set of controls operating from design release through long-term production. No single element is sufficient alone — the controls must work as a system.

• Locked configuration at design release. Every material, process parameter, tooling specification, and work instruction that defines the qualified assembly must be formally documented and frozen at design release. Changes after this point require engineering change control — review, impact assessment, and approval before implementation. The configuration baseline is not a document filing exercise; it is the reference against which all future production is verified.
• Formal engineering change control. Any deviation from the locked configuration — materials, process parameters, tooling, work instructions — must go through a formal engineering change process with documented review and approval. Purchasing changes, production floor adjustments, and material substitutions driven by supply chain constraints are all subject to this control. The change control process exists precisely because informal changes are the primary mechanism through which production drift occurs.
• Full material traceability to each assembly. For high-reliability high-voltage programs, each assembly should carry traceability for the materials and critical processes that affect dielectric performance — cable lot, insulation compound lot, overmold elastomer lot, connector body lot, potting compound lot — linked to supplier certifications and incoming inspection records. If a material is later found to be nonconforming, traceability determines which assemblies are affected. Without it, the response to a material nonconformity is either a blanket recall or an assumption that in-service assemblies are unaffected — neither of which is a defensible position on a high-voltage aerospace program.
• Process records for every critical operation. Critical manufacturing operations — particularly overmolding, insert molding, and vacuum potting — must generate process records for each production lot: actual temperature, actual pressure, actual cycle time, material lot, and operator identification. These records create a traceable link between the assembly's performance and the actual process conditions under which it was built. Without them, failure investigation cannot determine whether a field failure resulted from a design deficiency or a process deviation.
• Standard work at sufficient process depth — to certified workmanship criteria. Work instructions must capture the critical decisions in each step — not just the nominal procedure, but the specific parameters, acceptance criteria at each step, and the judgment calls that distinguish a correctly executed operation from one that will produce a marginal assembly. Instructions too general in scope transfer decision authority to the operator, whose judgment may deviate from the qualified process without producing an immediately detectable result. For soldering and cable/harness assembly operations, IPC/J-STD-001 and IPC/WHMA-A-620 provide the industry workmanship framework — defined acceptance standards that are independent of operator judgment and auditable against a published reference. Internal requirements that reference these standards, taught by certified trainers, are what translate that framework into consistent shop-floor practice.
• 100% electrical acceptance testing on applicable production units. Applicable electrical acceptance screens should be defined at design release and performed on every production unit unless a customer-approved sampling plan is explicitly justified. Sampling provides population-level statistics; it does not screen the specific unit that will fail in the field. For high-voltage assemblies, hi-pot / dielectric withstand is typically a baseline screen. Corona / partial discharge testing should be required where field concentration, voids, altitude, or long-term dielectric reliability are design risks. Testing on every applicable unit is also the mechanism that detects process drift that periodic audits may miss between audit intervals.

Quality at the Source — Why End-of-Line Testing Is Not Enough

End-of-line testing is a detection mechanism. It identifies assemblies that have already been built incorrectly. In a high-voltage cable assembly, the process steps that most strongly determine long-term reliability — conductor preparation, positioning in the mold, overmold shot conditions — are not visible in the finished assembly and are not recoverable after the fact.

An assembly with an improperly prepared conductor in the termination region, or with a mold shot that ran 10°C low, may pass a hi-pot test. The dielectric margin is reduced, but not below the test threshold. It will pass all acceptance criteria and still have a different reliability profile than the qualified article — one that may manifest as a field failure after a year of service under the actual operating conditions.

Quality at the source means that the controls preventing defects operate at the point where the critical decisions are made — during preparation, during assembly, during the mold cycle — not after the part is complete. This requires work instructions specific enough to define the correct process unambiguously, operators who understand the purpose of each step, and engineering involvement on the production floor to identify and correct deviation before it becomes a finished assembly. Quality at the end of the line confirms an already-built assembly meets a threshold. It cannot recover process decisions made incorrectly during build.

How Caton Addresses It — AS9100 Process Discipline from First Article Through Long-Term Production

• AS9100-certified QMS applied across production operations. Caton's AS9100-certified quality management system applies across production operations within the scope of its certification. The QMS provides the framework for configuration management, change control, material approval, process control, and corrective action. It is audited by an accredited third-party registrar on a defined schedule and continuously by Caton's internal audit program.
• Locked configuration with formal engineering change control. Qualified assemblies are released with a locked configuration that covers materials, process parameters, tooling, and work instructions. Changes to any element require engineering review and documented approval. Purchasing cannot substitute materials and production cannot adjust process parameters outside the approved window without a formal change action. This is the specific control that prevents the most common mechanism of production drift.
• Build records for applicable high-voltage production assemblies — material certifications, process records, and test data. For applicable high-voltage production assemblies, Caton maintains build records that include material certifications traceable to the specific lots used, process records for critical operations including overmolding and insert molding, and electrical test data defined by the program requirements. When a question arises after shipment — about a specific serial number, material lot, or process condition — the build record supports containment, root cause analysis, and corrective action.
• Engineering support at the production floor for critical operations. Caton's engineers work directly with the production team on critical operations — first builds, process changes, overmolding setup, insert molding, vacuum potting, and failure investigation. In a 45-person organization, engineering and production are physically and operationally adjacent, which is the practical mechanism for catching process drift before it becomes a finished assembly. A deviation visible to an experienced engineer during the mold cycle is correctable; the same deviation, invisible in the finished part, may not be detectable until a field failure occurs.
• Quality at the source — Lean and Shingo operating practices. Caton applies quality-at-the-source practices consistent with Lean and Shingo principles: standard work, in-process verification, error-proofing where practical, operator ownership of quality, and escalation when process conditions deviate from the approved standard. End-of-line testing confirms that process control worked; it is not a substitute for process control itself.
• 100% hi-pot testing on applicable production assemblies; CIV measurement on applicable high-voltage assemblies. For high-voltage production assemblies, Caton performs hi-pot / dielectric withstand testing on 100% of applicable units before shipment according to the acceptance criteria defined at design release. For assemblies rated for corona testing, Caton additionally ramps to measured corona inception voltage (CIV) and verifies adequate margin above the rated operating voltage — under defined voltage, fixture, environmental, and detection-threshold conditions. These screens are complementary: hi-pot verifies withstand integrity; corona testing verifies no detectable discharge at or below the specified acceptance condition. Which screens apply is a function of program requirements, specified at design release and locked with the configuration.

What to Ask Your Current or Prospective Supplier

These questions reveal whether a supplier's production control system is actually functioning — not whether their quality documentation says it should be. They distinguish suppliers with functioning production controls from suppliers whose controls exist primarily in documentation. Ask for specific evidence. General statements about quality culture or system coverage are not answers.

1. Show me the configuration documentation for an assembly qualified two or more years ago. What is the exact current production configuration, what changes have been made since qualification, and how was each change reviewed, assessed, and approved?
2. If I asked you to pull the complete build record for a specific serial number shipped 18 months ago — material certs, lot traceability, process records, and test data — what could you produce, and in what timeframe?
3. What triggers an engineering change action? Who has authority to approve a material substitution, and what documentation is generated? Can I see an example of a completed change action from a production program?
4. What process records are generated for overmolding and insert molding operations? Do those records capture actual temperature, actual pressure, and actual cycle time — or only a pass/fail check against a tolerance band?
5. Where are your quality controls in the process — at the end of the line, or at the point where critical fabrication decisions are made? What is the mechanism for catching a process deviation during an overmold cycle, before the part is complete?
6. Do you test every assembly electrically before shipment, or do you sample test? If sampling, what is the sample rationale per your quality plan — and what happens when a sample fails? Which electrical screens — hi-pot, corona, continuity — are specified in the acceptance test plan, and at what point in the design process were those screens defined and locked?

Technical References

1. AS9100 Rev D:2016. Quality Management Systems — Requirements for Aviation, Space, and Defense Organizations. SAE International / IAQG, 2016. The quality management system standard widely used by aerospace, defense, and space suppliers. Clause 8.5.1.3 addresses production process verification including first article inspection. Clause 8.4 addresses control of externally provided products including supplier change notification requirements. The standard against which Caton's QMS is certified and audited.
2. AS9102 Rev C:2023. Aerospace First Article Inspection Requirement. SAE International / IAQG, 2023. A widely accepted aerospace framework for performing and documenting first article inspection referenced in AS9100 Clause 8.5.1.3, covering part number accountability, product accountability, and characteristic accountability. Establishes the documentation requirements for a First Article Inspection Report (FAIR) that creates the production baseline for long-term program control. Rev C includes updated expectations around digital product definition and modern documentation practices.
3. IPC/J-STD-001J (2024) & IPC/WHMA-A-620E (2022). Requirements for Soldered Electrical and Electronic Assemblies (J-STD-001) and Requirements and Acceptance for Cable and Wire Harness Assemblies (WHMA-A-620). IPC / Global Electronics Association. Widely used workmanship standards for soldering and cable/harness assembly operations. Caton maintains in-house IPC Certified Trainers for both standards and has developed internal soldering and crimp requirements that reference these standards. All Caton assemblers are trained against the internal requirements by the in-house IPC Certified Trainers — providing defined, auditable criteria for the operations where undocumented process variation most directly affects dielectric integrity in high-voltage assemblies.
4. ISO 9001:2015, Clause 8.5. Production and Service Provision — Control of Production and Service Provision. ISO, 2015. Establishes the ISO baseline requirements for production control, including documented information, monitoring and measurement, and traceability. Relevant as the foundational standard underlying AS9100 and directly applicable to the configuration management and process control requirements discussed in this note.
5. IEC 60270:2000+AMD1:2015 CSV. High-voltage test techniques — Partial discharge measurements. International Electrotechnical Commission, consolidated edition. The international standard for partial discharge / corona measurement, covering measurement methods, quantities, calibration, and test circuit guidance. Provides the metrological basis for corona acceptance testing when specified by design for a high-voltage program.
6. Shingo, S. (1986). Zero Quality Control: Source Inspection and the Poka-Yoke System. Productivity Press. The foundational text establishing the principle that quality is created at the source of production, not detected at the end of the line. Shingo's analysis of the difference between judgment inspection (end-of-line detection) and source inspection (process-embedded prevention) directly informs Caton's operating approach to quality control in critical manufacturing operations.

References are provided to support the technical claims in this application note and to allow engineers to verify source material independently. Where standards are cited, the edition current at time of publication applies. Contact our engineering team with questions about how any reference applies to a specific application.