Production Sample Retention Policy

Sample retention is the practice of keeping a unit from a build so that a question asked months later can be answered by examining the product rather than by rebuilding it. It is inexpensive, it is frequently omitted and it is disproportionately valuable when a customer reports a failure that cannot be reproduced. A policy that states what to keep, for how long and where is all that is required.

Why Retain Samples

The main reason is comparison. When a field failure arrives, the most informative object is an identical unit from the same build, because differences between the two point at the cause. Without a retained sample, the investigation relies on the customer’s description and on the memory of the people who built it, both of which are unreliable after a year.

The second reason is evidence. Where a dispute arises about whether a product met its specification at the time of shipment, the retained unit is the physical record of what was shipped. That is useful in a commercial discussion as well as a technical one, and it is impossible to reconstruct later. <a href="https://www.gopcba.com/pcb-traceability-guide/” title=”Traceability practice”>Traceability practice and sample retention are complementary: one records the data, the other preserves the object.

What to Retain

The typical policy keeps one unit per build, per product, or per a defined interval for a continuous production run. Where the product is serialised, the retained unit should have a serial number recorded so that the link to the build is explicit. Where the build produced units with a known difference, such as a mid-build change, a sample from each side of the split is worth keeping.

Retention should also cover the material that cannot be recovered later. A coupon from the same panel, a piece of the board from the same lot and, where relevant, a sample of the paste or the laminate used, all make the retained unit more useful. Where the product is coated, an uncoated sample in addition to the finished unit allows the surface to be examined. Batch records should say what was retained and where it is stored.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/2-2.png" alt="Archive label showing lot and date” />

How Long to Keep

The retention period should be driven by the product’s warranty and by any regulatory requirement, with a margin. A product with a two-year warranty needs samples that survive at least that period, and a product with a longer service life or a safety function may need longer. Where a customer contract specifies a period, that requirement should be recorded in the policy rather than remembered.

The period should also consider the rate at which the product changes. A product that is revised frequently benefits from keeping samples at each revision, because the differences between revisions are exactly what an investigation needs to see. Where storage space is limited, the samples for the oldest revisions can be condensed to a photograph plus the key measurements, though that loses the ability to re-test.

Storage Conditions

The storage conditions should preserve the sample’s condition: dry, at a stable temperature, protected from light where the product is sensitive, and in packaging that matches the shipped product. A sample that has corroded in storage is of little use for comparison, and a sample that has absorbed moisture will behave differently from the units in the field.

The storage area should also be secure and organised. A shelf of unlabelled boards in a store room is not a retention system, because the sample cannot be found when it is needed and may be discarded by somebody clearing space. Dedicated shelving with a visible index is enough, and it costs very little. Document control should hold the index rather than the store room.

Labelling and Access

Each sample should carry a label with the product, the revision, the work order, the date and the serial number where applicable. The label should be durable and attached so that it cannot be separated from the unit in storage, and it should not obscure the features that might need to be examined under magnification. Where the sample is bagged, the label belongs on the outside as well as inside.

Access should be controlled but not difficult. The people who need a sample are usually engineers investigating a problem, and a process that requires three approvals will result in samples being cut open in the store room. A simple sign-out record, with the sample’s condition recorded on return or removal, keeps the archive intact while allowing the work to happen.

Using the Archive in an Investigation

When a failure is reported, the retained sample provides a baseline. Measuring the same characteristics on the retained unit and the failed unit shows whether the failure is a build issue, an application issue or a defect that developed in service. Where the retained unit shows the same condition, the build is implicated; where it does not, the investigation moves towards the application or the environment.

The sample also allows a test to be repeated. A functional test, a thermal cycle or a cross section performed on the retained unit can reproduce the failure mode, which is far more informative than a description. Where the retained unit passes the test that the failed unit failed, the difference between the two becomes the focus. Assembly records provide the context for that comparison.

Disposition and Disposal

Samples that reach the end of their retention period should be disposed of deliberately, with the disposition recorded. Disposal is also an opportunity: a sample that is about to be discarded can be sectioned and examined, and the findings recorded, which converts an expiring asset into information.

Where the product contains regulated materials or a battery, disposal has to follow the applicable route, and the record should reflect it. The same applies to samples that are returned to the customer or scrapped after an investigation, since the archive should reflect what is actually held rather than what was intended to be held.

Cost and Space

Retention consumes space, and the space grows with the product range. The cost is manageable if the policy is proportionate: one sample per build for a small product is a shelf, while one per unit for a large assembly is a warehouse. The policy should state the rule in terms of the product’s value and criticality rather than applying one rule to everything.

Where space is genuinely constrained, the compromise is to retain fewer samples and to retain more data: photographs, measurements, test results and material certificates. That preserves much of the investigative value at a fraction of the space, provided the data is captured at the time of build rather than reconstructed later.

FAQ

How many samples should be retained? Normally one per build or per defined production interval, plus additional units where the build contains a change.

How long should samples be kept? At least the warranty period plus a margin, and longer where a contract, a regulation or a long product life requires it.

Where should samples be stored? Dry, stable and secure, in packaging that matches the shipped product, with a visible index so they can be found.

What if there is no space? Retain fewer units and more data: photographs, measurements and certificates captured at the time of build.