The IEC 60601-1 test lab process is a structured, stage-by-stage engagement with an accredited laboratory that takes your device from a booking enquiry through intake, test execution, draft report, and final signed test report. For a well-prepared Class IIa or Class IIb active device, the whole process typically runs six to twelve weeks from first contact to signed report, with the on-bench portion taking two to four weeks of that window. The lab executes the clauses of EN 60601-1:2006+A1+A12+A2+A13:2024 and any applicable collateral or particular standards; it does not write your test plan, debug your design, or decide your essential performance. Founders who understand the stages, the handovers, and the decisions that belong to the manufacturer come out with a signed report on the first pass. Founders who treat the lab as a black box come out with a rebooking and an invoice.

By Tibor Zechmeister and Felix Lenhard. Last updated 10 April 2026.

TL;DR

  • The IEC 60601-1 test lab process runs in seven stages: lab selection, pre-test documentation, sample submission, test sequence execution, witness or unattended decision, report and certificate, and retest considerations.
  • A well-prepared campaign for a small active device takes six to twelve weeks end to end, with two to four weeks of bench time at the general-standard lab.
  • The lab executes clauses under EN 60601-1:2006+A1+A12+A2+A13:2024. It does not own your essential performance definition, your standards scoping, or your risk file.
  • Witness testing is slower and more expensive per day but catches problems in real time. Unattended testing is cheaper but stops cold the moment an ambiguity appears.
  • The formal test report and the subsequent certificate are separate deliverables. The report is the evidence that feeds your technical file. The certificate (where issued) is a short attestation that the report exists and concludes pass.
  • Design changes after the report is issued trigger a change-control analysis driven by the risk management file, which determines whether a partial or full retest is required.

Why the test lab process deserves its own map

A founder who has never been through a 60601-1 test campaign walks into the first booking call with a simple mental model: pay money, drop off the device, come back with a certificate. The reality has more handovers, more decisions, and more ways for a week to disappear than that model supports. Every test lab has a version of the same story about the startup that arrived on day one with an undefined essential performance and left on day four with a redesign quote.

Post 505 covers the overview of electrical safety testing and what readiness looks like. This post goes one level deeper into the process itself — the sequence of stages you will actually move through, the decisions that belong to you versus the lab, and what to expect at each handover. The aim is simple: by the end, a founder should be able to plan a test campaign on a calendar, budget for it honestly, and know which phone calls to make in which order.

One frame before we start. The MDR is the North Star. EN 60601-1:2006+A1+A12+A2+A13:2024 is the harmonised route to presumption of conformity with MDR Annex I Section 14. The test lab is the facility that executes the clauses of that standard against your device. None of those three things are interchangeable, and the process below exists to keep each of them in its proper place.

Step 1 — Lab selection and the booking enquiry

The process starts before you call anyone. Identify the shortlist of accredited laboratories whose ISO/IEC 17025 accreditation explicitly covers EN 60601-1:2006+A1+A12+A2+A13:2024 and the specific collaterals and particulars your device needs. Accreditation is not "the lab does medical testing." It is a named scope on a named standard, published on the accreditation body's website. If EN 60601-1 is not on the scope, the report from that lab will not carry the weight you need at Notified Body review.

For European startups, the shortlist typically includes two to four labs within reasonable travel distance. Distance matters because you will want an engineer from your team on site for at least the intake day, and because sample shipment internationally adds friction, customs time, and risk of transit damage.

The booking enquiry is where you hand the lab a one-page device description, the standards you believe apply, the rough sample count, and the target window. A good lab replies with a list of questions, not a quote. The questions are a quality signal — a lab that asks about applied part classification, intended use environment, and essential performance on day one is a lab that will not surprise you in week three. A lab that sends a quote without asking is a lab that will surprise you in week three.

Lead times at European medical test labs run from four to twelve weeks from booking to bench, depending on the season and the facility. Plan the booking around a device freeze date you can actually hold.

Step 2 — Pre-test documentation package

Once the slot is booked, the lab will send a documentation request. This request is not paperwork for the file. It is the raw material the lab engineer uses to build the test plan for your specific device.

The package typically includes: the risk management file under EN ISO 14971:2019+A11:2021 or the relevant extracts, the essential performance definition in writing with numerical limits, the standards scoping document naming every standard in and out of scope with justifications, schematics and PCB layouts, the applied part classification and Means of Protection derivation, the instructions for use as they will ship, declarations of conformity for critical components (mains power supply, transformer, isolation parts), and any prior pre-compliance data.

The lab reviews the package before the device arrives. A clean package means the lab engineer opens the box on intake day and starts work. An incomplete package means the engineer opens the box, finds a missing essential performance definition, and sends an email. That email can cost a week of calendar time even if it only costs an hour of lab time. Pre-test documentation is the single biggest lever on campaign duration, and it costs nothing except discipline.

Step 3 — Sample submission and intake

Ship two or three samples of the device under test, all the same hardware revision, all with the same firmware build, all configured identically, along with the accessories the device needs to run on the bench (cable sets, programming adapters, spare fuses, whatever bring-up requires). Two samples is the practical minimum because some tests — drop, impact, over-voltage — can damage a sample permanently. Three is safer.

Intake day at the lab covers a formal check-in of the samples, a review of the documentation package against the samples, and a physical inspection of construction-level features the engineer can see by eye: markings, labelling, mains connection integrity, enclosure fit, creepage and clearance on visible sections. Any ambiguity raised at intake stops the clock until the manufacturer resolves it. This is why sending an engineer on site for intake day is not optional — the ambiguities that stop the clock are almost always resolvable in ten minutes if the right person is in the room.

Step 4 — The test sequence

The bench portion of the campaign follows a recognisable shape for a Class IIa or Class IIb active device with a mains connection. Post 505 covers the individual test categories in more depth; the sequence view is what matters here.

Dielectric and insulation tests run first because they are fast if the design holds and because a dielectric failure tells the team immediately that later tests against the same barrier will also fail. Leakage current measurements — earth, touch, patient leakage, normal and single-fault — come next, typically on the same bench setup. Thermal measurements run long because steady state takes time to reach; these are often run in parallel with other clauses if the lab has multiple stations available. Mechanical tests (enclosure strength, impact, stability, drop where applicable) come after the non-destructive electrical work because they can damage the sample. Single-fault and abnormal operation tests close out the general-standard clauses. Particular-standard tests and collateral tests are interleaved where the setups allow or run as a dedicated block.

EMC testing under the 60601-1-2 collateral is almost always a separate booking at a separate facility with anechoic chambers and specific instrumentation. Post 511 covers the EMC campaign as its own topic.

Expect the general-standard bench portion to run two to four weeks for a well-prepared device. A device that triggers clarification emails or reveals unexpected failures can extend the bench portion significantly without extending the calendar time the lab has reserved — which means the campaign finishes late and gets queued behind other bookings.

Step 5 — Witness testing versus unattended testing

One decision most first-time founders do not know they have to make: witness or unattended. Witness testing means a representative from your team is on site for the test days and the engineer performs the tests in their presence. Unattended testing means the lab runs the tests on its own schedule and the manufacturer is not there.

Witness testing costs more per day because you are paying for your engineer's time and often for a dedicated lab slot. The payoff is speed on ambiguity — if a clause can be interpreted two ways, the decision happens in the room and the work continues. Witness testing also gives your engineer a first-hand view of how the standard is actually applied, which is training no course can replace. For a first campaign on a new device, witness testing is almost always the right call.

Unattended testing is cheaper and fits the rhythm of experienced teams with devices on their second or third generation. The trade-off is that any ambiguity stops the clock while the lab sends an email, and each round trip can cost a day. Unattended is a volume-play decision. It does not belong on a first campaign.

A common middle ground: attend the intake day and the first two days of bench work in person, then leave the lab to execute the remaining clauses unattended with a clear escalation channel. This captures most of the training value and most of the ambiguity resolution while keeping cost under control.

Step 6 — The test report and the certificate

The deliverable the technical file actually needs is the test report. A 60601-1 test report is a long, structured document: a clause-by-clause record of every test the lab executed, with the measured values, the acceptance criteria, the pass or fail determination, and references to the specific edition of the standard and amendment set applied. For EN 60601-1:2006+A1+A12+A2+A13:2024 the amendment set is stated explicitly on the report cover. The Notified Body will verify the edition and amendment set during technical file review.

The report typically arrives first as a draft, which the manufacturer reviews for factual errors — wrong revision number on the device under test, wrong serial number, wrong essential performance statement in the summary. Factual corrections go back to the lab. Interpretive disagreements do not — if you disagree with a pass or fail determination, the discussion is a technical one with the lab engineer, not a markup on the draft. Once the draft is agreed, the lab issues the final signed report.

Some labs also issue a certificate as a short attestation that the report exists and concludes pass. The certificate is not a substitute for the report — the technical file needs the full report, not just the attestation — but the certificate can be useful as a one-page summary for investors, distributors, or non-regulatory audiences.

Expect two to four weeks between the end of bench work and the final signed report. Draft turnaround is often one to two weeks, manufacturer review one week, final signing one week.

Step 7 — Retest considerations after design changes

The test report covers one specific hardware revision, one specific firmware build, and one specific configuration. Any change after the report is issued goes through change control in your QMS. Change control asks the risk management file what the change affects, the risk file identifies the clauses of EN 60601-1:2006+A1+A12+A2+A13:2024 that depend on the affected function, and those clauses define the scope of the retest.

A firmware change to a non-safety, non-essential-performance function — a log format, a UI translation, a cosmetic label — typically does not trigger retest, provided the change control documentation shows the analysis. A firmware change affecting alarm behaviour, dosing control, or any essential performance parameter typically triggers partial retest on the affected clauses. A change to the applied part classification, the isolation scheme, the Means of Protection count, or the mains-side power supply typically triggers a full retest.

Plan retest capacity into your release timeline. A partial retest can often be fitted into a lab's next available slot in four to six weeks. A full retest on a loaded lab can take as long as the original campaign. Post 508 covers the full change-control-to-retest decision flow in detail.

Common mistakes that cost startups real weeks

  • Booking the lab before the essential performance is defined in writing. The lab cannot close out any clause whose acceptance criterion depends on essential performance, and an essential performance definition written in the lab car park does not survive scrutiny.
  • Sending samples from a different hardware revision than the schematics in the documentation pack. Intake catches this on day one and the clock stops until new samples arrive.
  • Skipping the on-site intake day to save travel cost. The email loop that replaces the on-site engineer costs more than the travel.
  • Treating the draft report review as a negotiation on pass/fail. The draft review is for factual correction only. Technical disagreements are separate conversations with the test engineer, not markup on the draft.
  • Assuming the certificate is the deliverable. The Notified Body needs the full test report. A certificate without the report behind it is worth nothing at technical file review.
  • Forgetting that EN 60601-1-2 EMC testing is a separate booking at a separate facility. Founders who only plan for the general-standard lab discover the EMC timeline only after the general-standard work is done.

The Subtract to Ship angle on the test lab process

The subtractive move in the test lab process is not to test less than the Regulation requires. It is to walk in with a package so complete that the lab never has to stop the clock to ask a question. Every clarification email, every rebooking, every partial retest caused by a documentation gap is additive waste — work that would not have existed if the readiness phase had been taken seriously. Subtracting that waste requires the discipline to do the boring readiness work before the fun bench work.

The second subtractive move is to keep the device frozen from the moment the booking is confirmed. A device that is still changing when the samples ship is a device that will need a retest after the report lands, and the retest will cost the same as a fresh campaign on the affected categories. Freeze discipline is cheaper than retest discipline.

Underneath both of these: the obligation is MDR Annex I Section 14, the tool is EN 60601-1:2006+A1+A12+A2+A13:2024, and the test lab is where the tool is applied. Keeping that direction clear is what turns a first-pass clean report from a lucky outcome into a reproducible one. For the broader methodology that connects this thinking to the full certification path, see post 065 on the Subtract to Ship framework for MDR compliance.

Reality Check — Where do you stand?

  1. Have you identified a shortlist of accredited labs whose ISO/IEC 17025 scope explicitly names EN 60601-1:2006+A1+A12+A2+A13:2024 and the collaterals you need?
  2. Is your pre-test documentation package complete enough that a lab engineer could build a test plan without asking you any questions?
  3. Do you have a written freeze date for the hardware revision and firmware build you will ship to the lab, and is that date far enough in the future that you will actually hit it?
  4. Have you decided witness versus unattended for this first campaign, with a clear reason for the decision?
  5. Have you planned an engineer from your team to be on site for intake day and the first two bench days?
  6. Do you know how long your chosen lab typically takes from end-of-bench to signed report, and is that window in your go-to-market timeline?
  7. If the report comes back with a partial fail on one clause, do you already know which internal change-control process you will run to scope the retest?

Frequently Asked Questions

How long does the full IEC 60601-1 test lab process take from booking to signed report? For a well-prepared Class IIa or Class IIb active device, the end-to-end window is typically six to twelve weeks: four to eight weeks of bench plus documentation and intake work, plus two to four weeks of report drafting, review, and final signing. Unprepared campaigns with multiple clarification loops or partial failures can extend that window significantly.

What is the difference between the test report and the certificate? The test report is the clause-by-clause document that records every test executed, the measured values, the acceptance criteria, and the pass or fail determinations. The certificate, where issued, is a short attestation that the report exists and concludes pass. The technical file needs the full report. The certificate is a one-page summary for non-regulatory audiences.

Should I choose witness testing or unattended testing for my first campaign? Witness testing for a first campaign on a new device. The extra cost per day is repaid by faster ambiguity resolution and first-hand training for your engineer. Unattended testing is appropriate for experienced teams on second or third generation devices where the device under test is well understood and the lab has executed the same standards for the same team before.

What happens if a test fails mid-campaign? The lab records the failure in the draft report and continues with any clauses that do not depend on the failing one. The manufacturer then has a decision: attempt a design fix and rebook the affected clauses as a partial retest, or close out the campaign with the failure recorded and fix the design before a fresh campaign. The right choice depends on the root cause and the time cost of each option.

Can the lab write my test plan for me? No. The test plan depends on decisions the manufacturer owns — applied part classification, essential performance definition, standards scoping, single-fault analysis, intended use environment. The lab can review a draft test plan, point out gaps, and suggest sequencing improvements. It cannot write the plan from scratch because it does not own the inputs.

Do I need to retest everything if I change a single component? No. The scope of the retest is determined by the risk management file and the change control analysis. A component change on the mains side typically triggers a partial retest of the dielectric and leakage clauses. A component change on a non-safety, non-essential-performance function typically does not trigger a retest, provided the change control documentation shows the analysis. The default answer is neither "retest everything" nor "retest nothing" — it is "let the risk file decide."

Sources

  1. Regulation (EU) 2017/745 of the European Parliament and of the Council of 5 April 2017 on medical devices, Annex I Chapter II, Section 14 (construction of devices and interaction with their environment). Official Journal L 117, 5.5.2017.
  2. EN 60601-1:2006+A1+A12+A2+A13:2024 — Medical electrical equipment — Part 1: General requirements for basic safety and essential performance.

This post is part of the Electrical Safety & Systems Engineering Under MDR series in the Subtract to Ship: MDR blog. Authored by Felix Lenhard and Tibor Zechmeister. The MDR is the North Star. EN 60601-1:2006+A1+A12+A2+A13:2024 is the harmonised route. The test lab is where the route is verified — and the process above is how a prepared founder walks through it with a signed report at the end instead of a rebooking and an invoice.