AIRSPRAYTECH ACADEMY
Finishing Quality - From Spec to Sign-Off | Article 23 of 28
Sampling Plans and Inspection Frequency
Measuring one convenient part is not the same as understanding the production lot.
A finishing operation may process ten parts, ten thousand parts, one large fabricated structure, or a continuous stream of product. Inspecting every characteristic on every part may be unnecessary, impractical, or destructive.
That is why the quality plan must define what will be inspected, when it will be inspected, where measurements will be taken, how samples will be selected, and what happens when a result is unacceptable.
A sampling plan is not a way to avoid inspection. It is a documented method for using selected observations to understand and control a larger body of work.
Begin by Defining the Production Lot
A sampling plan cannot work until the facility defines what the sample represents. The production lot may be based on:
- One customer order
- One part number
- One coating-material batch
- One pretreatment or process batch
- One shift or production period
- One rack, carrier, load, or oven batch
- A defined quantity produced without a significant process change
A lot should contain product made under reasonably consistent conditions. If the coating batch, substrate, line, pretreatment, oven setting, or application process changes, the original lot may need to be divided.
The lot definition should appear in the control plan and inspection record.
Identify the Characteristics to Be Inspected
Not every characteristic requires the same sampling frequency. Some conditions are monitored continuously. Others are checked at startup, by production lot, at scheduled intervals, or through periodic laboratory testing.
Finishing characteristics may include:
- Incoming material identity and condition
- Process-water quality
- Pretreatment chemistry
- Coating mix ratio and viscosity
- Temperature, humidity, and dew point
- Application-equipment settings
- Wet-film thickness
- Dry-film thickness
- Cure or oven profile
- Color, gloss, texture, and appearance
- Adhesion and mechanical properties
- Holiday or continuity testing
- Final packaging and identification
The inspection frequency should reflect the importance, variability, detectability, and consequence of failure for each characteristic.
Shop-floor rule:
Sample where the risk is, not where the part is easiest to reach.
When One Hundred Percent Inspection Is Appropriate
Some characteristics may require inspection of every part or the complete coated area.
One hundred percent inspection may be appropriate when:
- The customer or specification requires it
- Failure creates a significant safety or service risk
- The characteristic can be checked quickly and nondestructively
- The process has recently produced failures
- The operation is new or has undergone major change
- The product has high value or low production volume
- Every exposed area must be continuous or defect-free
One hundred percent inspection does not guarantee that every defect will be found. Inspector fatigue, poor lighting, inadequate test coverage, instrument limitations, and unclear standards can still allow defects to pass.
When Sampling Is Appropriate
Sampling may be appropriate when the process is stable, the lot is clearly defined, the characteristic can be measured reliably, and the applicable specification permits it.
Sampling is especially useful when:
- Production volume is high
- Testing is destructive
- Testing is time-consuming or costly
- Process measurements provide continuing evidence of control
- Historical data demonstrates consistent performance
The sampling plan must still provide enough evidence to make the required decision. A small sample from a highly variable or unstable process may provide false confidence.
Use Random Selection
A random sample gives each eligible part a reasonable opportunity to be selected. This reduces the chance that the inspector will unconsciously choose the cleanest, easiest, or most convenient pieces.
Selection methods may include random numbers, predetermined time intervals, systematic selection with a random starting point, or another documented method.
Sampling the top part from every finished rack may not be random. If rack position affects cleaning, application, or cure, that practice may repeatedly miss the weakest locations.
Use Stratified Sampling When Position Matters
Stratified sampling divides production into meaningful groups and takes samples from each group. This is useful when the process may perform differently across locations or conditions.
Finishing examples include:
- Top, middle, and bottom rack positions
- Left, center, and right conveyor positions
- Beginning, middle, and end of a production run
- Different shifts or applicators
- Different spray guns or automatic zones
- Light and heavy parts in the same oven load
- Flat surfaces, recesses, edges, and weld areas
This method helps reveal variation that could disappear inside one overall average.
Inspect at Startup
Startup inspection confirms that the process is ready before a large quantity of product is completed.
Startup checks may include:
- Material and batch verification
- Pretreatment and rinse-water conditions
- Environmental conditions
- Equipment setup and spray pattern
- Grounding and powder delivery
- Wet-film or powder-film application
- Dry-film thickness
- Appearance, color, gloss, and cure
The first acceptable part or group should be approved before unrestricted production continues when the control plan requires first-article approval.
Inspect During Production
A passing first article does not prove that the rest of the run will remain acceptable. Tips wear, filters load, pretreatment changes, oven loading varies, powder characteristics shift, and operators make adjustments.
In-process inspection should be frequent enough to detect meaningful change before a large quantity is affected.
Frequency may be based on:
- Elapsed time
- Number of parts or racks
- Each material batch
- Each oven load
- Each shift or operator change
- Each equipment adjustment
- Each line interruption or restart
The chosen frequency should be documented and supported by process knowledge and performance history.
Inspect After Change
Change increases uncertainty. Additional inspection should be considered after:
- A new coating, color, or material batch
- A new substrate supplier or part revision
- A pretreatment or water-treatment adjustment
- A spray-gun, pump, nozzle, or powder-system change
- Maintenance or equipment repair
- A conveyor-speed or oven-setting change
- A new operator, robot program, or rack configuration
- A process interruption or abnormal event
Change-point inspection helps separate production made before the change from production made afterward.
Include the Last Part
End-of-run inspection helps confirm that the process remained acceptable through completion. It is particularly useful when equipment wear, material condition, filter loading, pot life, or bath drift could affect later production.
When both the first and last samples are acceptable, the facility has stronger evidence that the process remained stable. When the last sample fails, the team must determine when the process left control and which product may be affected.
Plan Destructive Testing Carefully
Adhesion, impact, flexibility, abrasion, chemical resistance, and environmental exposure tests may damage or consume the specimen.
The control plan should define:
- Whether production parts or test panels are used
- How specimens are selected
- How frequently destructive tests are performed
- Whether damaged production areas may be repaired
- How representative panels are processed
- How specimens and results are retained
A test panel should follow the same cleaning, pretreatment, coating, and cure process as the production represented by the panel.
Consider the Cost of a Missed Failure
Inspection frequency should reflect risk. A missed defect in a decorative indoor bracket does not carry the same consequence as a missed discontinuity in a chemical-service lining.
Consider:
- Safety and regulatory consequences
- Product-service severity
- Cost of field failure
- Ability to detect the defect later
- Ease of repair before shipment
- Historical process performance
- Customer-specific requirements
Higher risk generally requires stronger prevention, more frequent process monitoring, or more extensive product verification.
Tightened and Reduced Inspection
A mature quality system may change inspection frequency according to documented performance rules.
Tightened inspection may be appropriate after:
- A failed inspection or customer complaint
- A significant process change
- Repeated rework or unstable results
- An equipment or calibration problem
- Loss of process data or traceability
Reduced inspection may be considered only when permitted and supported by stable process data, capable measurements, consistent supplier performance, and a documented approval process.
Reduced inspection should never mean no control. Process monitoring and reaction plans remain necessary.
Define Acceptance Before Looking at the Results
The sampling plan should identify the acceptance number, rejection number, averaging method, individual reading limits, and any special rules before inspection begins.
Do not change the rule after seeing the data. Do not average an unacceptable area into an acceptable result unless the specification expressly permits that calculation.
When the customer supplies a sampling standard or inspection level, ensure the correct lot size, sample size, classification, and acceptance criteria are used.
Separate Defect Categories
Not every defect has the same consequence. The quality plan may classify defects according to risk, function, appearance, safety, or customer requirements.
Examples may include:
- Critical defects affecting safety or required performance
- Major defects affecting function, durability, or customer acceptance
- Minor defects affecting appearance within defined limits
Definitions and acceptance rules should be documented. Inspectors should not create defect classifications while examining the product.
Use a Clear Reaction Plan
A failed sample raises an immediate question: How much production could be affected?
The reaction plan should define:
- When production must stop
- How the affected lot is identified and held
- Whether expanded or one hundred percent inspection is required
- How the last known acceptable point is determined
- Who investigates the process
- Who may authorize sorting, repair, rework, or disposition
- What evidence is required before production restarts
A failed sample should never be replaced quietly with another sample in hopes of obtaining a passing result.
Record What Was Actually Sampled
An inspection report should make clear what the sample represented and how it was selected.
Useful records may include:
- Customer, job, part, and lot identification
- Lot size and lot definition
- Sampling plan and inspection level
- Sample size and selection method
- Part, rack, time, or location sampled
- Characteristic and acceptance criteria
- Instrument and test-method identification
- Individual results and calculated values
- Acceptance or rejection decision
- Inspector, date, and time
- Corrective action and final disposition
Use the Data to Improve the Plan
Sampling results should be reviewed for patterns. Compare results by shift, applicator, rack position, booth, gun, coating batch, pretreatment condition, oven load, and part geometry.
If defects repeatedly appear in one location, increase attention there and investigate the cause. If a stable process consistently produces acceptable results, review whether the current inspection effort is focused on the highest risks.
The sampling plan should improve as the quality team learns more about the process.
Include Applicators and Operators
The people running the line often know where variation is most likely. Applicators understand difficult spray areas. Rackers know which positions are crowded. Oven operators recognize loading differences. Pretreatment operators see changes in parts and bath behavior.
Ask these employees where samples should be taken and what process changes should trigger additional inspection. Their experience can make the sampling plan more representative and more useful.
The Takeaway
A sound sampling plan defines the production lot, selects representative parts and locations, establishes inspection frequency, sets acceptance rules, and explains what happens when a sample fails.
The objective is not to inspect the easiest part or collect the smallest number of readings. The objective is to gather enough trustworthy evidence to control the process and make a responsible product decision.
Measure It. Track It. Test It. Prove It.
Copyright 2026 Azimuth Spray Systems, LLC. All rights reserved.
