From First Article To Full Production: 100 Parts, 8 Hours And 10 Days

Overview

A cleaning machine is accepted on measurements, but it is validated on repetition. A single clean part proves that the process can work; it does not prove that it will work on the thousandth part, after a shift of tool wear, with a tank that has been running for eight hours. The difference between those two statements is the difference between a demonstration and an acceptance test.

Production specifications solve this by escalating the test: a first article, then a run of consecutive parts, then a full shift, then a multi-day campaign. Each stage catches a class of failure the previous one cannot, and each has a defined sampling plan so that the result is a number rather than an impression.

Big Bird Industrial (formerly Shimada Big Bird Industrial) has been building cleaning and deburring systems for automotive powertrain components since 2001, and the sequence described here follows the acceptance structure used on cylinder head and block lines.

1. What Validation Is Not

The most common mistake is to treat a sample part as the acceptance test. A part cleaned with fresh chemistry, clean filters and a freshly filled tank describes the best case the machine can produce, not the case production will experience.

The Fresh-Tank Fallacy

Cleanliness degrades as the bath loads with contamination, and the rate of degradation depends on the filtration chain and the drag-out volume. A machine that meets a 2 mg limit on day one and fails it on day five has a filtration or bath-management problem, and only a multi-day test exposes it.

Why Visual Checks Do Not Count

A visual check finds gross contamination and misses everything the specification actually measures. The measurement has to extract residue from a defined region of the part, using a defined method, and report mass and particle size separately. A part that looks clean can carry several milligrams of residue in a gallery that no one can see.

2. The Gate Before Acceptance

Acceptance does not begin on the day the machine is switched on. It begins with documentation delivered before the pre-acceptance run, and the value of that gate is that it moves disputes earlier.

Documentation Delivered In Advance

Four weeks before the pre-acceptance, the supplier is typically required to deliver a static accuracy report for the machine and a cleanliness report on sample parts, both measured by an agreed method. The buyer reviews those before agreeing to schedule the run, which prevents a situation where the equipment is installed and the method is still under discussion.

Agreeing The Sampling Method First

The sampling method, the extraction region and the laboratory that performs the measurement should be fixed before any part is cleaned. When they are agreed afterwards, the measurement becomes negotiable, and a machine that fails one method can appear to pass another.

3. Three Escalating Tests

The progression from 100 parts to 8 hours to 10 days tests three different things: repeatability, sustained operation, and stability over a production rhythm.

What Each Stage Catches

A run of 100 consecutive parts tests whether the machine repeats the same result without intervention, and it is also where cycle time, dryness, exit temperature, noise and automatic changeover are verified together. An 8-hour loaded run tests whether performance holds through a shift, including the effect of rising bath contamination and of thermal drift in the machine. A 10-day campaign with a daily sample tests whether the process stays inside the criterion across tank changes, filter changes and weekend shutdowns.

None of these stages can be compressed by running the machine harder. The failures they detect are time-dependent in a way that speed does not reproduce: a filter loads over a shift, a bath degrades over days, and thermal drift appears only after the machine has been at temperature for hours. A supplier who proposes to demonstrate the 10-day result inside one afternoon is describing a capability test, not a stability test, and the difference is exactly what the buyer is trying to establish.

Cleaning a clean part proves very little. Acceptance tests therefore contaminate the test parts deliberately, using the same contaminants the production process produces, in a defined quantity.

4. Deliberately Contaminated Parts

Preparing The Test Parts

A typical protocol takes five parts of each variant and contaminates them with a mixture of cutting fluid and machining chips, supplied by the buyer so that the mixture reflects the actual process rather than a laboratory approximation. The parts are then cleaned and measured to the specified criterion, with a minimum number of parts measured and photographs retained for comparison.

Marker Substances For Access Checks

Where the question is whether the fluid reaches a region rather than what mass it leaves, marker substances answer it directly. A visible indicator painted on the surface must be fully removed; toothpaste or plasticine packed into a bore must not remain; grease applied to a specific area must be gone after the high-pressure stage. These checks test access, which cleanliness measurement alone cannot distinguish from insufficient energy.

5. What The Ten-Day Run Actually Catches

The failures that appear after installation are rarely mechanical. They are the slow ones: a bag filter that loads faster than expected because the swarf separation stage is undersized, a bath that needs dumping at a shorter interval than the plan assumed, a nozzle that clogs at a rate that becomes visible only after a week, or a robot path that drifts by a millimetre and starts missing the edge that mattered.

A ten-day run with one sample per day does not prove the machine will run for a year without intervention. It does establish the maintenance interval, because the day the sample starts to drift is the day the filter or the bath needs attention. That number is worth more to the operating team than the acceptance certificate.

6. Documenting The Result

The output of validation is not a pass mark. It is a set of numbers that the operating team uses afterwards: the cleanliness result per region at each stage, the cycle time under load, the exit temperature, the bath condition at which the sample started to drift, and the maintenance actions that followed.

What The Report Should Contain

A usable acceptance report records the part identification for every sample, the measurement method and the laboratory, the measured value against the limit, and the machine settings in force at the time. Photographs of the sampled regions are worth keeping for the first article and for any failing sample, because a number without a location cannot be reproduced.

Turning It Into A Maintenance Baseline

The same data becomes the baseline for the maintenance plan. The day the sample drifted, the bath concentration at that point, and the filter that was changed define the interval that the plant should use going forward. A machine handed over with a certificate alone leaves that interval to be discovered again through failures.

This is the practical value of an escalating validation sequence: it produces a record of how the process behaves over time, not just a statement that it was acceptable on the day of handover.

Frequently Asked Questions

How Long Does Full Validation Take?

The test sequence itself is roughly two weeks if the pre-acceptance run, the 8-hour shift and the 10-day campaign are scheduled back to back. The elapsed time including documentation review and laboratory turnaround is typically longer.

Who Supplies The Contamination For The Test?

The buyer, normally, because the mixture has to match the actual machining process. A supplier-provided mixture removes the buyer from the part of the test that most often produces disagreement.

What If The Machine Passes 100 Parts But Fails At Day Seven?

That pattern points to bath management or filtration rather than to the cleaning process. The corrective action is usually a change in filter grade, filter capacity or dump interval rather than a change in pressure.

Is A 10-Day Run Required On Every Project?

Not always. It is most valuable where the cleanliness limit is tight, where the part has internal features, and where the line runs continuously. On simpler applications the 100-part and 8-hour stages may be sufficient, but the decision should be written down rather than assumed.

Conclusion

Validation is a sequence, and each stage exists because the previous one cannot see a particular failure. The first article proves capability, 100 parts prove repeatability, 8 hours proves stability under load, and 10 days establishes the maintenance interval.

A specification that names all four stages, with sampling plans and measurement methods, gives both sides the same definition of success. It also produces the one output that a certificate does not: a number that tells the operator when to change the filter.