How Long Does Certification Testing Take?

← All guides

Ask three laboratories how long testing will take and you will often get three different numbers, from "about a week" to "six to eight weeks". All three can be honest, because they are answering slightly different questions. One is quoting bench time. Another is quoting bench time plus its current queue. The third has quietly allowed for the fact that your manual is still a draft.

So the useful answer is not a number. It is a breakdown of the calendar, with the parts you control separated from the parts you do not.

Bench time is one segment of five

A normal certification project moves through five stages. Only one of them is the laboratory actually testing your product.

Stage	What it involves	Typical duration
1. Scope and quotation	Deciding which standards apply to your configuration and target markets, then pricing them	1–3 working days (longer for a multi-band radio product where modes and power levels have to be pinned down first)
2. Samples and documents ready	Final samples, final manual, label artwork, BOM and any module certificates in the laboratory's hands	Days to weeks. This is where most schedules actually break
3. Bench time	The measurements themselves	A few days to a few weeks, depending on the standard
4. Retest, if the product fails	Diagnosis, a change to the product, and re-test of the affected clauses	Commonly one to three weeks
5. Report and any scheme review	Writing and issuing the report; plus certification body, notified body or TCB review where the market requires it	3–7 working days for the report, and one to three weeks more where the scheme runs its own review

Add those up and a clean single-configuration project is often three to five weeks from the day the laboratory has everything. One failure loop plus a scheme review pushes it to six to nine. What rarely explains the difference between a three-week project and a nine-week project is the measuring.

What bench time looks like by test type

These ranges are for a single, well-defined configuration, tested without incident. They are planning numbers, not promises.

Test	Typical bench time	What drives the length
EMC — emissions and immunity	3–7 working days	Number of ports and cables, and how many operating modes the product has to be exercised in
Safety (IEC/EN 62368-1 or the applicable standard)	5–10 working days	Construction review, abnormal-operation and fault-condition tests, and how many component alternatives are declared
Radio (RED / FCC ID)	3–10 working days	Bands, modes and data rates. Each combination is a separate set of measurements
SAR / RF exposure	Several days	How many positions have to be measured and at which frequencies
RoHS / REACH screening	1–3 working days for XRF screening	Wet chemistry, needed when screening is inconclusive or phthalates are in scope, runs to 5–10 working days
UN38.3 (lithium batteries)	Around two to three weeks	Eight tests, several of which include rest periods that cannot be shortened

The gap between a two-mode and a six-mode wireless product is measured in days, not hours. That is why a laboratory that quotes you a timeline before asking how many modes you ship is guessing.

The clock starts when the lab has everything, not when you send the PO

This is the single most common misunderstanding about testing schedules, and it is worth stating plainly: a project is not "in progress" because the purchase order is signed. It starts when the laboratory has three things at once.

• The final samples, not an engineering build that still needs a firmware change or a substitute part
• The final documents — manual, label artwork, BOM, ratings, and any module certificates or previous reports
• A booked test slot. Most laboratories run a bench queue. If you send samples without reserving capacity, they wait their turn

Shipping is part of the calendar too. Express freight from a factory to a Shenzhen laboratory is usually a few days, but it is a few days you can remove from the plan by shipping earlier rather than by asking the laboratory to test faster.

If you are still choosing who to test with, the questions in how to choose a testing laboratory are the ones that decide whether the schedule you are quoted is the schedule you get.

Where the weeks really go

In practice, five things account for most of the slippage on projects that were expected to be quick.

1. Samples that are not the final build. This is the classic one. A pilot build goes to the laboratory, testing passes, and then the production tooling changes a ground plane, an antenna or a substitute component. The report describes a configuration that no longer exists. If your samples are still pre-production, that has to be declared before the bench slot, not after — the trade-offs are set out in can I use engineering samples for certification testing.

2. Documents that arrive late, or change. A manual that is still in draft, or a rating label that gets redesigned mid-project, means the report has to reference something that does not exist yet. This sounds minor and is not: the document set is what the report and the Declaration of Conformity point back to.

3. Radio modes that were not declared. A product is described as "Bluetooth", the engineer discovers it also ships with a 2.4 GHz proprietary mode enabled, and that mode has never been measured. Every additional mode is another session on the bench plus more report content. Declaring the full mode list at the start costs nothing and saves days.

4. A change made while the project is open. A quieter adapter, a firmware update that raises transmit power, a PCB revision to fix an unrelated issue. Each change is assessed against the report. Some are covered by a documented change assessment; others restart part of the measurement.

5. A failure. One failed test is not the end of the world, but it has a cost in calendar time: the failure has to be diagnosed, a fix designed and implemented, the affected clauses re-tested, and where the fix touches an already-passed area, that area may have to be re-checked as well. The sequence, and what you should ask for at each step, is in what happens if my product fails a certification test.

What can be compressed, and what cannot

Customers often ask whether paying more shortens the schedule. The honest split is this.

Can usually be compressed	Cannot be compressed
Queue position, by reserving a bench slot in advance or using a priority slot where the laboratory offers one	Rest, conditioning and stabilisation periods written into the standard. A test with a specified rest period takes the time it takes
Calendar time lost to document and sample rounds, by sending everything together and freezing the configuration	Abnormal-operation, fault-condition and endurance tests, which run to their own duration
Total wall-clock time for multiple standards, by running compatible tests in parallel — which usually needs more samples	A certification body, notified body or TCB review of the file. That runs on the scheme's clock
Report writing, where the file is complete on the day testing finishes	Physical shipping time of samples or equipment between sites

The pattern is consistent: money and planning can move your project up a queue, but they cannot shorten a measurement.

Multi-market projects are not multiples of one market

A frequent fear is that testing for three countries means three times the work. It usually does not. A single EMC test campaign very often produces the data that goes into more than one market's report set, so several markets can be covered by one round of measurements.

What does add calendar time is any market whose route requires a third-party review on top of the test report — a Technical Certification Body for an FCC ID, a notified body where an EU type examination certificate is needed, or a national scheme in markets that operate their own approval systems. Each of those is a queue you join after the testing is done, and each should be shown as a separate line in your schedule rather than buried in one headline number.

Documentation for each market is separate as well, but it is paperwork rather than bench time, and it can run alongside the testing if the file is organised early.

Two dates on the Chinese calendar to plan around

If your laboratory is in China, two periods each year affect every schedule, and neither is a surprise.

The National Day holiday in early October and Chinese New Year in January or February both see laboratories running a reduced or closed schedule for around a week. The days immediately before and after are heavier than normal, because everyone is trying to finish or start around the break. A project that would take two weeks in June can take noticeably longer if the bench slot falls on either side of one of these periods.

For a launch in October or in February, work backwards from the holiday rather than forwards from the purchase order. That single change of direction prevents most holiday-driven delays we see.

What a useful timeline from a laboratory looks like

A schedule you can plan a launch against has four things in it:

• A stage-by-stage breakdown, with working days against each stage rather than one total
• The configuration the estimate assumes — models, radio modes, adapters, and how many samples you will send
• A named owner for each stage: the laboratory, you, or a third party such as a certification body
• The stated trigger for the clock — normally receipt of final samples, final documents and a booked slot

If a laboratory gives you one number with no breakdown, that number is a bench estimate wearing a calendar's clothes.

Where we come in

Scheduling is part of quoting here, not an afterthought. When you send a product and its target markets, the estimate you get back is written as day-counts against the stages above: scope, sample and document readiness, bench time per test, any scheme review, and report issue. Where a market needs a notified body, a TCB or a national certification body, that queue is listed as its own line so you can see what the delay is waiting on.

We also flag the items that sit on your side and the date each one is needed by, because those are the dates that actually decide whether a launch date holds. If your configuration is not frozen yet, we will say so and show you what testing early would and would not lock in. Our testing and certification scope lists what we run in-house and what we handle through partner laboratories.