Table of Contents
An evaluation board that works with five sample ICs has answered a useful engineering question. It has not established that the volume order will contain the same package, grade, silicon revision or approved source.
Electronic component samples should therefore be purchased against a defined learning objective and an identity record. The result must state what was demonstrated, what remains untested and which production configuration the evidence supports.
What should be specified when ordering electronic component samples?
Specify the exact manufacturer orderable code, package, temperature grade, quantity, packing and intended evaluation. Ask the supplier to identify any difference between the offered samples and the production configuration before accepting the order.
Texas Instruments explains how to find sample-eligible devices through its product sampling process. Eligibility and request conditions are separate from the engineering evidence needed to approve a part. A sample listing should not be interpreted as a promise of volume allocation.
A useful request includes the board revision, tests to be performed and any destructive work. Record whether the parts are standard production devices, engineering samples or another explicitly identified status. Do not infer status from an unmarked bag or the word “sample” in an email.
Build the sample record before opening the packaging
Assign an internal receipt identifier while the manufacturer and supplier labels are still available. Link the order, package photographs, quantity, lot or date-code information and handling instructions to that identifier. Then assign individual unit identifiers where test history needs to follow a specific device.
This record separates three decisions:
| Stage | Evidence to retain | Decision supported |
|---|---|---|
| Receipt and identity | Exact code, source, labels, condition and quantity | These are the samples requested |
| Engineering evaluation | Unit ID, board revision, setup, results and exceptions | This configuration met the stated evaluation objectives |
| Production transfer | Approved code and source, unresolved differences, release scope | Procurement may buy the defined production configuration |
These are proposed control points, not a universal qualification standard. The required depth depends on the application and customer requirements.
How many samples should the evaluation order include?
Allocate samples by test use, including destructive tests, retained references and expected assembly losses. The resulting quantity supports the evaluation plan; it is not automatically a statistically justified qualification sample size.
For example, suppose an engineering plan needs eight populated boards, four separate devices for destructive analysis, two untouched reference units and two spares for assembly loss. The initial request is 16 devices. If destructive analysis uses devices already counted in the eight boards, the plan must explicitly account for that reuse rather than double-counting them.
Track the allocation as testing progresses. A damaged device does not become a valid electrical failure observation merely because it failed after handling. Conversely, consuming the spare allowance should not erase a genuine assembly or compatibility issue.
If the objective is estimating a population defect rate, a separate statistical plan is needed. A small development lot is usually selected for learning and troubleshooting, not random lot acceptance.
Which sample results transfer to a production purchase?
Results transfer only to the configuration and conditions that the evaluation actually covered. A change in orderable code, revision, package, programmed state, source or assembly process must be assessed for its effect on the evidence.
Compare the sample record with the production quotation line by line. A suffix that changes packing alone may have different consequences from one that changes temperature grade or package. Obtain the manufacturer’s interpretation rather than assuming all suffix differences are administrative.
For programmed devices, link the evaluation result to the approved firmware image and programming record. The separate preprogrammed microcontroller release controls address that additional identity layer.
Production sourcing also needs commercial evidence that samples cannot supply: required order quantities, current delivery terms, source approval and any agreed change notifications. These should be checked when the production order is placed.
Close the evaluation with explicit limits
The release record should name the approved configuration, summarize the actual tests and list excluded operating conditions. If testing covered room-temperature functionality but not environmental performance, say so.
Unresolved deviations should have a disposition: reject the candidate, perform a defined additional test, or approve a limited use that the organization can control. Avoid an unrestricted “qualified” flag when the evidence supports only prototype use. That distinction gives procurement a usable buying instruction and keeps the engineering result attached to the material it describes.
Frequently Asked Questions (FAQ)
Does a free sample establish a production price?
No. Treat sample availability and production pricing as separate offers. Request a dated volume quotation with quantity, packing, delivery and commercial terms.
Can a different package be used for early firmware development?
It can support a limited development task if the relevant functions match, but it does not validate the production footprint, assembly process, thermal path or package-specific electrical behavior.
Should unused samples be returned to production inventory?
Only after identity, handling and eligibility are verified against the production material rules. Parts exposed to experiments, rework or uncontrolled storage should remain segregated.
