Quality Assurance and Testing Technology

01Purpose and scope

• This manual supports the course Quality Assurance and Testing Technology. It is written as a working reference for people who plan, perform, or review tests and inspections on manufactured parts, assemblies, process outputs, and industrial equipment.
• Quality assurance is the system that keeps products and processes meeting stated requirements. Testing technology is the set of methods, instruments, and software used to generate evidence that those requirements are met. In practice the two are inseparable, because a test result means little unless the test method, the equipment, the sampling approach, and the record keeping are all controlled.
• This manual covers incoming material checks, in-process inspection, final verification, functional and destructive testing, measurement system control, and the documentation that links them. It does not cover product design theory, detailed metallurgy, or software test automation beyond the points where they touch industrial release decisions. Practice varies widely between a small job shop, a regulated medical device plant, and a high volume food packaging line, and the notes here are meant to be adapted to those settings.

02Core principles

• Requirements come first. Every test should trace back to a specification, drawing, customer requirement, or internal procedure. If a characteristic has no requirement behind it, the first task is to find out why someone is measuring it before deciding how.
• Test methods must be repeatable and documented well enough that a different technician would get comparable results. A method description should name the instrument, the fixture, the sample preparation, the environmental conditions if they matter, the acceptance limits, and the way results are recorded.
• Measurement uncertainty is real and must be considered. A gauge reading of 10.02 millimeters against a limit of 10.00 plus or minus 0.05 is not the same as a reading that sits safely inside the band. Gauge repeatability and reproducibility studies, or equivalent measurement system analysis, tell you how much of the observed variation belongs to the instrument and the operator rather than the part.
• Inspection does not create quality. A final inspection can catch defects, but it cannot make defects go away. Preventing nonconformance by controlling the process is almost always cheaper than sorting good parts from bad after they are made.
• Sampling is a decision tool, not a guarantee. A sample of 20 parts from a lot of 2,000 can give useful confidence about the lot, but only if the sampling plan is chosen deliberately and the residual risk is understood by the people who accept the lot.

03Planning and preparation

• Start by defining the product or process characteristics that matter to the customer and to safe operation. Classify them by criticality if your organization uses that approach. Safety related or function critical characteristics usually need tighter controls, more frequent checks, and stronger evidence than cosmetic features.
• Build a test plan that lists each characteristic, the method used to verify it, the equipment and fixture, the sample size or frequency, the acceptance criteria, the responsible role, and the record that will be produced. A plan that exists only in someone's head does not count as a plan for audit or troubleshooting purposes.
• Confirm that every instrument is fit for its use. Check calibration status, calibration range, resolution, and any environmental limits such as temperature stability. A dial indicator with 0.01 millimeter resolution cannot reliably verify a tolerance of 0.005 millimeter, no matter how carefully it is handled. Calibration should be traceable to recognized reference standards, and the calibration interval should reflect how the instrument is used and how it has performed.
• Prepare the test environment and the people. Operators and inspectors need training on the specific method, not just general familiarity with the equipment. Write down who is qualified to perform which tests. Where the work involves hazardous energy, pressure, high voltage, or chemicals, the plan must identify the safety controls and any permits required before testing starts.
• Decide early how nonconforming results will be handled. Define who can disposition a nonconforming item, what evidence is needed, and how suspect product is segregated so it cannot be shipped by mistake.

04Recommended workflow

1. 1
   : Review the requirement set and the customer or internal specification for the item under test. Confirm the revision level in use, because a common failure is testing against an outdated drawing after an engineering change has been released.
2. 2
   : Select or confirm the test method for each characteristic. Where a standard test method exists, such as a published tensile test procedure or a recognized electrical test sequence, follow it and note any deviations with justification. Where no standard applies, write a method and have it reviewed before use.
3. 3
   : Verify the measurement system. Confirm instrument calibration, check the fixture for wear or damage, run a reference or master part if one is available, and record the result. For attribute gauges such as go and no-go plugs, check that the gauge itself is inside its wear limits.
4. 4
   : Perform the inspection or test under the defined conditions. Record raw values, not just pass or fail, whenever the data may later be used for trending. Time stamp entries, identify the operator and instrument, and avoid rounding during recording.
5. 5
   : Evaluate results against acceptance criteria. Where the measured value falls close to a limit, consider whether measurement uncertainty could change the decision. Apply the agreed rule for guard banding or retesting, and do not invent a retest simply to get a passing result.
6. 6
   : Handle nonconforming items. Tag and segregate the affected material, record the nonconformance with its description and quantity, and notify the responsible engineer or quality representative. Decide disposition through the documented authority, such as use as is, rework, repair, return to supplier, or scrap.
7. 7
   : Release or hold the lot, batch, or shipment based on the completed record. The release decision should be traceable to the specific results and to the person who authorized it. Where sampling was used, record the plan applied and the acceptance number.
8. 8
   : Feed results back into the process. Review trends, open corrective actions where a pattern appears, and update the test plan if the data shows that a method, limit, or frequency is not doing its job.

05Quality checks, risk controls and compliance

• Quality checks run at several levels. Operator self checks catch obvious problems quickly. Independent inspection verifies the operator's work and the critical characteristics. Periodic audits confirm that the whole system is being followed. Each level should have its own records so that gaps are visible.
• Risk controls should be proportional to consequence. A failed fastener torque check on a structural joint carries a different level of risk than a label misprint. Failure mode and effects analysis, control plans, and process capability studies are common tools for deciding where controls belong and how strong they need to be. The depth of analysis should match the potential harm.
• Compliance requirements vary by industry, product, and location. Many industrial manufacturers use ISO 9001 as the basis for a quality management system, and some sectors add requirements such as automotive, aerospace, or medical device rules. Certification and accreditation requirements are set by the relevant certifying bodies and regulators, so check the current versions and the requirements that apply to your site. Where laws or regulations govern product safety, testing, or records, the current authoritative source in your jurisdiction controls.
• This guidance does not replace qualified professional judgment, approved procedures, or required authorization. Testing that affects safety, pressure systems, electrical equipment, or regulated products should be reviewed and released by people with the proper qualifications and authority under your organization's procedures.
• Test records may themselves be subject to retention rules. Know how long records must be kept, who can access them, and how they are protected from alteration. Electronic records should have audit trails where the system supports them.

06Collaboration and documentation

• Quality assurance sits between design, production, maintenance, purchasing, and the customer. Good collaboration means that engineers receive usable failure data, production gets clear acceptance criteria, and purchasing knows which supplier results matter most. Regular short reviews of open nonconformances and test results usually do more good than long periodic meetings.
• Documentation should be controlled. Work instructions, test methods, gauge lists, calibration certificates, inspection records, and nonconformance reports each need a clear owner, a revision status, and a way to retire obsolete versions. The most common documentation failure is an operator working from a printed copy of a method that has since been revised.
• Suppliers are part of the testing chain. Incoming inspection results, certificates of conformance, and supplier corrective action reports should be linked to the parts they cover. Where a supplier has a long record of stable results, the level of incoming inspection may be reduced, but that decision should be documented and reviewed.
• Use a single record system where possible, or make sure separate systems can be cross referenced by serial number, lot number, or work order. Investigations slow down badly when a test result is in a spreadsheet, a calibration record is on paper, and the nonconformance is in an email thread.

07Measures and continuous improvement

• Track indicators that reflect both process health and the reliability of testing itself. Useful measures include first pass yield, escape rate to customers, number of nonconformances by cause, percentage of gauges calibrated on time, gauge repeatability and reproducibility results, and average time to disposition a hold. Choose a small set and review them regularly rather than building a large dashboard nobody reads.
• Process capability indices such as Cp and Cpk are common measures for characteristics with a defined tolerance. They are only meaningful when the process is stable and the data is representative, so check for stability before drawing conclusions. A high index from a short run of parts may not hold over a full production cycle.
• Continuous improvement works best when each finding leads to a specific action with an owner and a verification step. A corrective action should confirm that the root cause was addressed and that the defect has not returned. Closing an action because the paperwork is complete is not the same as closing it because the result has changed.
• Be honest about the limits of testing. Some failure modes cannot be detected by a final test without destroying the part, and some tests are too slow to run on every unit. In those cases, the improvement may be a better process control, a sampling plan that matches the risk, or a change in the design that makes the characteristic easier to verify.

08Common mistakes and how to prevent them

• Testing against the wrong revision of a specification is one of the most expensive mistakes. Prevent it by controlling the link between the test plan and the released drawing, and by checking revision status at the start of each lot or shift.
• Using an instrument that is out of calibration or outside its range leads to results nobody can defend. Prevent it with a visible calibration status tag, a check of calibration due dates before use, and a rule that equipment past due is removed from service until recertified.
• Recording only pass or fail hides trends that would have shown a drifting process. Prevent it by recording actual values for characteristics that are measured, and reviewing those values, not just the pass rate.
• Accepting marginal results without considering uncertainty causes escapes. Prevent it by applying guard banding rules where the measurement uncertainty is significant compared with the tolerance, and by requiring a documented justification for any exception.
• Allowing operators to retest until a part passes is a serious control failure. Prevent it by defining retest rules in the test method, limiting retests to specific causes such as a confirmed fixture error, and reviewing all retest occurrences.
• Treating corrective actions as paperwork exercises means the same problem returns. Prevent it by requiring evidence of effectiveness, such as a follow up capability study or a defined period with no recurrence, before the action is closed.

09Practical checklist

• Confirm the revision of the requirement document and the test method before testing begins.
• Verify that every instrument and fixture is calibrated, in range, and in good condition.
• Check the measurement system with a master part, reference standard, or gauge study result.
• Record raw measured values for characteristics that may be trended later.
• Apply the agreed acceptance criteria and guard band rules, and note any measurement uncertainty concerns.
• Segregate and tag nonconforming material, then route disposition through the documented authority.
• Link test results, certificates, and nonconformance reports to the lot, serial number, or work order.
• Review trends and open corrective actions for repeated causes at a set interval.
• Confirm that qualified personnel and required authorizations are in place for regulated or safety related tests.

★Learning sources

• Micro Degree: Quality Assurance and Testing Technology
• Micro Degree: Solar Power Systems
• Micro Degree: Energy Management and Audit
• Micro Degree: Industry 4.0 Technologies