What should I actually check on a sample before I approve a bulk order?
Check the four things the bulk order can drift on: capacity, speed, physical build, and behaviour on your own equipment. Capacity by full write-and-read-back, not a folder copy. Speed by a 4GB single-file write, not a benchmark. Physical build by looking for mould flash, print alignment, cap fit and colour against a reference. And behaviour in the actual device that will use it — the terminal, camera, scanner or dashboard — because that is the only test that predicts what your customers will experience. Sign and keep the approved sample: it is the specification for the reorder.
Why the sample test is different from a receiving inspection
A receiving inspection checks that what arrived matches what you ordered. A sample approval defines what you ordered in the first place.
That makes the sample test broader. It is not only "is this unit good", it is "is this the product I want ten thousand of, and will I be able to tell whether the ten-thousandth one is the same". Anything you do not pin down on the sample becomes an ambiguity on every reorder.
The four checks that matter
Each one targets a failure mode that the bulk order can develop independently.
| Check | How | What it protects against |
|---|---|---|
| Capacity | Full write-and-read-back (H2testw / f3) | Under-capacity and counterfeit memory |
| Speed | Single 4GB+ file, read the steady rate | Memory grade lower than quoted |
| Build | Mould flash, print alignment, cap fit, colour vs reference | Cosmetic drift in bulk |
| Device behaviour | Plug into your actual terminal / camera / scanner | Host compatibility failures at scale |
The device test is the one people skip
It is also the one that causes the most expensive surprises. A drive can pass every generic test on a laptop and still fail in a POS terminal that mounts removable storage read-only, or a camera that only accepts a specific file system, or a dashboard unit that rejects the controller.
Test it in the real device, in the real workflow — export the files, import them, eject, repeat twice. If the workflow involves a specific software package, open the files in that package. A sample that passes on your desk has been tested on your desk.
- Plug into the actual target device, not an equivalent.
- Run the real workflow end to end, not a generic file copy.
- Repeat the cycle at least twice, including an eject and re-insert.
- Record the device model and firmware version you tested against.
Then sign it, and keep it
An approved sample that exists only in a chat message is not a specification. Sign and date the unit, keep one, and ask the supplier to keep the other.
Write the reference into the purchase order: "to match approved sample dated [date], retained by both parties". That single line converts "the colour looks different" from an argument into a comparison against a physical object.
Sources
Figures quoted on this page are from the published reporting listed above. Where a number is an estimate rather than a published figure, it is described as approximate.
Frequently asked questions
Should I pay for a sample?
Usually yes, and the fee is not the interesting part. A sample fee typically covers the setup, materials and shipping of a one-off unit, and many suppliers credit it against the bulk order. What matters more is what the fee buys you: a unit made on the same line, with the same memory and controller as the bulk, in the final finish and packaging. A sample assembled by hand from different parts tells you nothing about the batch.
How many samples should I approve?
One is the minimum for the physical reference, but two or three from the same small run is more informative, because it lets you see natural variation within a batch rather than one point. If the two units differ noticeably in colour or fit, the process is not yet under control and bulk will amplify that.
What if the sample is fine but the bulk order is different?
That is the reorder and batch-consistency problem, and the defence is contractual rather than technical: a locked bill of materials including controller and memory grade, a written clause that substitution requires your approval, and a per-batch test report on the same method you used on the sample. The signed sample is the reference that makes those clauses meaningful.
Working on a specification?
Send us the product type, capacity, quantity and target date. We reply to business enquiries within 1 business day.
Request a quote