Approving a power window switch manufacturer should not be based on a showroom tour, a supplier questionnaire, or one passing sample alone. A useful audit follows selected production lots through documents, materials, tooling, assembly, testing, packaging, and traceability. It also distinguishes an attractive factory from a process that repeatedly produces the approved configuration.
Power window switches are not all simple contact devices. A driver master switch may integrate window lock, door lock, mirror control, illumination, one-touch stages, resistance-coded outputs, a PCB, or network communication. The audit must therefore match the exact product architecture and customer requirements. The scoring model below is a buyer-controlled template, not a universal approval standard; revise weights, thresholds, and knockout rules in the audit plan before the visit.
On this page: Audit Method and Scoring | Product-Specific Manufacturing Audit | Evidence Sampling and Quality Controls | Findings, Approval and Closure
Audit Method and Scoring
Define Scope, Criteria, and the Evidence Trail
Name the legal manufacturing site, product families, customer programs, processes, shifts, outsourced operations, and sub-suppliers included. Audit criteria may include contracts, customer-specific requirements, approved drawings, process specifications, control plans, quality-system procedures, and applicable regulatory obligations. An area not sampled is not audited, not automatically conforming.
Select production or recently completed lots before walking the floor. Trace one finished switch backward to resin, stamped contacts, plating lot, springs, PCB or resistor components, assembly records, test data, and operators. Trace another material or process lot forward to all affected finished goods and shipments. Record an evidence ID for every conclusion so a reviewer can retrieve the source.
ISO 19011:2026 provides guidance for auditing management systems, managing audit programmes, conducting audits, and evaluating auditor competence. It is not a power window switch product specification, vehicle-data source, product-test method, or evidence of supplier approval. Use it to structure an evidence-based, risk-based audit; use controlled product and customer documents for technical acceptance.
Use a Defined 0–5 Scoring Scale
Score only after recording the sampled evidence and finding. Not applicable requires a documented reason and weight redistribution approved before calculating the total.
| Score | Evidence-based meaning |
|---|---|
| 0 | No control exists, required evidence is absent, or observed practice contradicts the requirement. |
| 1 | Informal or reactive practice exists, but ownership, criteria, records, or implementation are substantially missing. |
| 2 | A documented control exists but has major implementation gaps, weak sampling, or recurring ineffective reaction. |
| 3 | Requirement is implemented and sampled records generally conform; limited gaps do not threaten product release. |
| 4 | Control is consistently implemented across sampled shifts or lots, with effective reaction and verified follow-up. |
| 5 | Control is demonstrably effective, risk-linked, measured, and improved using reliable trend or prevention evidence. |
Calculate each module result as module weight × score ÷ 5. The weighted total cannot override a Critical finding or knockout condition.
Weight the Audit Modules
| Audit module | Weight |
|---|---|
| Requirements, APQP/project planning, and change control | 10% |
| Incoming materials and sub-supplier control | 12% |
| Molding, stamping, plating, and tooling | 18% |
| Switch, terminal, PCB, and final assembly | 20% |
| Product-specific inspection and end-of-line testing | 20% |
| MSA, SPC, calibration, maintenance, and competence | 10% |
| Traceability, nonconformance, packaging, and continuity planning | 10% |
| Total | 100% |
A power window switch audit should trace requirements through process planning and the records created at the production station.
Product-Specific Manufacturing Audit
Executable Audit Matrix
Use this matrix as the working core of the audit. Replace general sample quantities with the risk-based quantity defined in the approved audit plan.
| ID | Audit question | Risk | Evidence to sample | Sample-size requirement |
|---|---|---|---|---|
| PW-01 | Are customer requirements linked through project planning, design/process risks, approval status, and serial controls? | Missing or conflicting requirement | Contract review, application matrix, project plan, approval file | 2 active projects plus 1 recent change |
| PW-02 | Who can approve drawings, OE cross-references, software, process deviations, and design changes? | Unauthorized release | Authority matrix, revision history, deviation approvals | Current product plus 2 changes |
| PW-03 | Are special characteristics transferred through Process Flow → PFMEA → Control Plan → Work Instruction → Inspection Record? | Control-chain break | Matched document revisions and floor records | 3 characteristics end to end |
| PW-04 | Are resin grade, color, moisture handling, cavity, molding parameters, and molded dimensions controlled? | Warpage, poor fit, weak clips | Resin certificates, dryer logs, parameter records, cavity data | 3 lots across available cavities |
| PW-05 | Are contact material, stamping tool condition, burr, plating specification, and plating-lot evidence controlled? | High resistance or wear | Material lot, tool checks, plating report, inspection data | 3 material/plating lots |
| PW-06 | Are spring identity, force checks, lubricant identity, quantity, and application method controlled? | Low contact force or unstable feel | Incoming records, dispensing setup, assembly checks | 3 lots and 2 operators |
| PW-07 | Are rocker force-travel, first/second stage, detent, return, free play, and alignment verified? | Incorrect manual or AUTO actuation | Force-travel method, fixture study, results, reaction | 5 samples across 2 lots |
| PW-08 | Are connector terminals fully seated, oriented, retained, and protected from damage? | Backed-out terminal or open circuit | Insertion poka-yoke, retention checks, connector inspection | 5 pieces plus reject challenge |
| PW-09 | For resistance-coded designs, are every switch state, resistor value, tolerance, and test condition controlled? | Incorrect module command | Circuit revision, component lot, state-resistance results | All states on 5 samples |
| PW-10 | For PCB products, are SMT recipe, polarity, soldering, AOI, ESD, rework, and moisture-sensitive parts controlled? | Latent electronic failure | Program version, AOI records, ESD checks, rework history | 2 lots and 2 stations |
| PW-11 | Where firmware exists, are source, binary, programming, checksum, access, and revision locked to the part? | Wrong vehicle behavior | Release record, programmer log, challenge test | 3 units and 1 wrong-file challenge |
| PW-12 | Are illumination source, polarity, color, symbol orientation, brightness method, and light leakage controlled? | Wrong appearance or night visibility | Approved master, test method, artwork revision, results | 5 samples in controlled viewing setup |
| PW-13 | Does end-of-line testing cover every rocker, lock, mirror, illumination, AUTO stage, output state, and communication function? | Missing function shipped | Test specification, software version, fault insertion | All functions plus 3 defect challenges |
| PW-14 | Are pinout, continuity, isolation, resistance, current, or communication tests tied to the exact architecture? | Electrical damage or false pass | Test limits, fixture wiring, master unit, raw records | 5 records and 3 live retests |
| PW-15 | Are fixture wear, golden-unit verification, sensor checks, and test-software access controlled? | Test escape | Verification schedule, maintenance, access log, alarms | Current fixture plus last 3 checks |
| PW-16 | Do endurance and environmental reports identify sample, revision, load, sequence, conditions, failures, and acceptance source? | Untraceable qualification claim | Approved plan, raw data, report, deviations | 1 current and 1 prior report |
| PW-17 | Are startup, mold/tool change, material change, adjustment, maintenance, and restart approvals defined? | Unverified transition output | First-piece records and reaction evidence | 3 event types across recent records |
| PW-18 | Are nonconforming parts segregated, rework authorized, and reworked units fully reverified? | Mixed or incompletely tested stock | Hold area, rework instruction, retest records | 3 cases plus floor observation |
| PW-19 | Can one packed switch be traced backward and one component lot forward within the buyer’s target time? | Incomplete containment | Lot, operators, equipment, materials, test, shipment | 1 backward and 1 forward exercise |
| PW-20 | Do packaging, label verification, capacity, bottleneck, backup equipment, and contingency plans protect supply? | Mix-up, damage, or interruption | Pack specification, label challenge, capacity and contingency evidence | 2 SKUs and 1 simulated disruption |
For each row, add these working columns in the audit record: Evidence ID | Finding | Score | NC Level | Owner | Due Date | Closure Evidence | Verified By/Date. The matrix is intentionally product-specific: for a mechanical switch, PCB and firmware rows may be approved as not applicable; for an electronic master switch, omitting them would leave a major gap.
Final-function testing and packaged-sample traceability should be connected to the same approved product revision and production lot.
Evidence Sampling and Quality Controls
Verify Measurement and Process Capability
Calibration status alone does not prove that a measurement system can resolve the requirement. Review MSA or GR&R for drawing and process characteristics where measurement variation could affect decisions, such as molding dimensions, terminal position, force-travel, contact resistance, illumination, or programmed output. Confirm actual operators, fixtures, part variation, resolution, range, and analysis method match production use.
Where SPC or capability is appropriate, connect the chart to a named characteristic, stable measurement system, subgroup logic, control limits, reaction plan, and retained evidence. Review what happened after an out-of-control point. Cp or Cpk without process stability, specification source, period, and reaction history is not approval evidence. Verify first-piece and restart controls at the machine rather than reviewing only blank procedures.
Challenge Poka-Yoke and End-of-Line Tests
Ask the supplier to demonstrate approved defect challenges without damaging production equipment. Examples include a missing terminal, wrong connector orientation, incorrect resistor, reversed LED, unprogrammed PCB, wrong software file, stuck rocker, or omitted function. The expected result is a controlled reject, alarm, lockout, and traceable reaction. A green screen with no verified challenge history does not establish test coverage.
Use the Power Window Switch Buying Guide to define the functions and interfaces that belong in the audit. The OEM number cross-reference guide helps separate application evidence from manufacturing evidence, while How to Test a Power Window Switch provides context for controlled electrical diagnosis. None replaces the approved production test specification.
Audit Sub-Suppliers, Changes, and Traceability
Review controls for molding resin, stamped contacts, plating, springs, lubricant, terminals, PCBs, LEDs, and outsourced processes. Confirm incoming identification, supplier performance, acceptance method, nonconformance response, and notification obligations. Where customer or contract requirements call for APQP, PPAP, core tools, or customer-specific records, verify the applicable submissions and approval status; do not assume every project uses the same system.
Change notification should cover design, resin, contact material or plating, spring, lubricant, artwork, PCB, electronic component, firmware, tooling, process, site, sub-supplier, test method, and packaging. A controlled golden sample supports appearance and configuration, but it cannot replace drawings, software identification, test limits, or change authorization.
Findings, Approval and Closure
Classify Nonconformities and Knockouts
| Level | Definition | Required response |
|---|---|---|
| Critical | Evidence integrity failure or uncontrolled condition that creates unacceptable safety, legal, traceability, or product-release risk | Stop approval; contain affected scope and escalate immediately |
| Major | Systemic absence or failure of a required control, or evidence that nonconforming product could escape | Corrective action and verified closure before production approval |
| Minor | Isolated lapse that does not indicate systemic loss of control or immediate release risk | Defined correction, owner, due date, and effectiveness review |
Recommended buyer decision rules are: 85–100 approved only with no open Critical or Major findings; 75–84 conditional approval only for a named product, site, process, quantity, and expiry date; below 75 rejected. Any open Critical, knockout, or evidence-integrity concern overrides the score. Buyers should approve these rules before the audit rather than changing thresholds after seeing the result.
Knockout candidates include falsified or altered records; refusal to provide agreed traceability evidence; unapproved design, material, software, site, or sub-supplier change; inability to identify or contain affected lots; uncontrolled critical end-of-line testing; mixed product identity; and shipping product outside an explicit customer hold. Define the contractual list and escalation route in advance.
Close Findings and Make the Release Decision
Corrective action should identify containment, affected product and dates, verified root cause, systemic corrective action, document changes, training, implementation evidence, and effectiveness review. Photographing a revised instruction is not enough when the finding concerns actual process performance. Revisit the station, sample later production records, repeat a defect challenge, or conduct a focused on-site review according to risk.
A weighted score above the threshold with an open Major is not approved. Conditional approval must state product numbers, factory, processes, purchase quantity, duration, deviation, owner, due date, enhanced inspection, and closure evidence. Unreviewed products or processes remain outside the approval scope.
Before issuing an RFQ or sample order, use the Power Window Switch RFQ and Sample Approval Checklist to align commercial and technical evidence. The switch-versus-motor-versus-regulator guide helps prevent field complaints from being assigned to the supplier before the failed system element is identified.
Work With TONFUL
TONFUL’s automotive power window switch range presents current vehicle applications and OE reference numbers. For a supplier review or sourcing discussion, provide the exact vehicle and door position, OE reference, connector, pinout, function matrix, approved sample, required quality documents, audit criteria, forecast, packaging, and validation plan. Final supplier and product approval should remain tied to the buyer’s controlled requirements, sampled evidence, and signed release.