EMC testing for medical devices is the test campaign that demonstrates a medical electrical device behaves correctly in its intended electromagnetic environment and does not disturb other equipment in that environment. The test plan is defined by EN 60601-1-2:2015+A1:2021, the harmonised collateral standard for electromagnetic compatibility, applied on top of the general standard EN 60601-1:2006+A1+A12+A2+A13:2024. The evidence feeds MDR Annex I Section 14.5 (and the connected Section 14.2(d) and Section 17.1). EMC has two halves. Emissions and immunity. And three intended-environment classes (professional healthcare facility, home healthcare, and special environments). A startup that treats EMC testing as a single line item without understanding the halves and the environments almost always pays for the test slot twice.

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

TL;DR

  • EMC testing for medical devices is governed by EN 60601-1-2:2015+A1:2021, the harmonised collateral standard applied together with EN 60601-1:2006+A1+A12+A2+A13:2024. The evidence maps to MDR Annex I Section 14.5 and the related Sections 14.2(d) and 17.1.
  • EMC has two halves. Emissions limits what the device radiates or conducts into the environment. Immunity requires the device to keep working safely when the environment pushes disturbances into it.
  • EN 60601-1-2:2015+A1:2021 defines three intended-environment categories. Professional healthcare facility, home healthcare, and special environments. And the test levels scale with the category.
  • Immunity acceptance criteria are tied to the device's basic safety and essential performance, which must be defined in writing before the first test is run. No essential performance on paper means no valid EMC test report.
  • Pre-compliance EMC testing at a cheaper facility is the single biggest cost-control lever in the EMC process. The accredited lab should be the performance, not the debug bench.

Why EMC testing is a two-halves problem

Every experienced EMC test engineer has watched the same mistake a dozen times. A startup books an EMC slot, drops the device off, and expects a single pass-or-fail verdict. When the results land, half the device passed and half did not. And the founder is surprised that "EMC" turned out to be several different measurements with different setups, different limits, and different root causes.

EMC is not one test. It is two halves. The halves exist because the regulation cares about two different things.

Emissions asks a question about the rest of the world. Is your device putting so much electromagnetic energy into its surroundings. As radiated waves through space, or as conducted noise back through its power cable. That it disturbs other equipment? A device that passes all its own functional tests but drowns out the ECG monitor on the next trolley is unsafe and non-compliant.

Immunity asks a question about the device itself. When the environment pushes energy into the device. A nurse's static discharge, a mobile phone transmitter, a nearby motor switching on, a voltage dip during a thunderstorm. Does the device keep doing what it is supposed to do, safely, with no loss of essential performance? A device that passes all its emissions tests but resets itself whenever a cleaner's radio is keyed is also unsafe and also non-compliant.

Both halves must be addressed. A clean EMC file demonstrates both. The test plan, the lab setup, the instrumentation, and the acceptance criteria are all different between the two halves. Which is why the calendar time and the cost per campaign are both larger than founders expect the first time through.

The three environments. Professional, home, special

EN 60601-1-2:2015+A1:2021 does not apply the same test levels to every device. The standard defines three intended-environment categories, and the test levels (for both halves) scale with the category.

Professional healthcare facility environment. Hospitals, clinics, doctor's offices, surgical theatres, intensive care units. The environment is controlled. Equipment is installed by professionals, mains supply is relatively clean, and the other equipment nearby is also subject to EMC limits. Emissions limits and immunity levels are calibrated to this baseline.

Home healthcare environment. The patient's home, a care home, a vehicle, or any location where the device is used outside a clinical facility, often by a lay user. The environment is far less controlled. Mains supply is noisy, radios and phones are close, static build-up is higher, and the surrounding electronics are consumer-grade and more sensitive to emissions. EN 60601-1-2:2015+A1:2021 applies tighter emissions limits and higher immunity levels to devices intended for this environment, because the real-world electromagnetic weather is worse.

Special environments. Specific settings that neither professional nor home covers. For example, proximity to active HF surgical equipment, magnetic resonance environments, or military-grade environments. Special-environment requirements are additional, typically stricter, and not every device needs them.

The category is a manufacturer decision, written into the intended use and the instructions for use, and locked in before the test plan is drafted. A device labelled for professional healthcare must not be tested only to the home category (insufficient) and does not need to be tested to home category (wasteful). A device labelled for home healthcare must meet the home category levels. No shortcut to the professional baseline exists. Getting the category wrong on paper is expensive to fix after the lab visit.

The EMC test sequence. What the week looks like

A realistic EMC campaign under EN 60601-1-2:2015+A1:2021 has a recognisable shape once the readiness work is done. The specifics depend on the device and the environment category, but the sequence is broadly stable.

Intake and test plan confirmation. The lab engineer reviews the documentation package. Essential performance definition, environment category, applicable standard set, worst-case configuration. And confirms the test plan. Ambiguity here stops the clock.

Emissions measurements. Radiated emissions are measured in an anechoic chamber or on an open area test site, across the specified frequency range, with the device in its worst-case operating mode. Conducted emissions are measured on the mains cable using a line impedance stabilisation network, across the specified frequency range. Both produce a plot with limit lines; any peak above the line is a fail.

Immunity exposures. The device is exposed to each specified disturbance in turn, in its worst-case operating mode, with the acceptance criterion observed by the lab engineer and, where possible, by instrumentation. The disturbance types in EN 60601-1-2:2015+A1:2021 include radiated radio-frequency electromagnetic fields, conducted disturbances induced by radio-frequency fields, electrostatic discharge, electrical fast transients and bursts, surges, voltage dips and interruptions, and power-frequency magnetic fields. Proximity fields from RF wireless communications equipment are a distinct test under the standard, reflecting the real-world presence of mobile phones close to the device.

Worst-case cable and configuration sweeps. Cables are rotated, extended to the maximum length stated in the instructions for use, and oriented to expose the weakest coupling path. Accessories are connected. The device is operated in the mode the risk file identifies as most vulnerable.

Report generation. The lab records the measured values, the limit lines, the acceptance criteria, and the pass or fail determination for each test, in a clause-by-clause report that references EN 60601-1-2:2015+A1:2021 and its amendment set explicitly.

Expect the bench portion of a clean first-pass EMC campaign to run one to two weeks for a simple device and longer for a device with many cables, many modes, or a particular standard layered on top. Report drafting adds another two to four weeks.

What counts as a pass

"Pass" on an emissions test is a measured value below the limit line, across the specified frequency range, in every tested configuration. The limit lines are published in the standard and depend on the environment category. There is no interpretive room. Either the peak is below the line or it is above it.

"Pass" on an immunity test is harder, because the standard does not define "still working" for your specific device. It defines a set of performance categories and requires the manufacturer to map each immunity test to the category that matches the device's basic safety and essential performance. For one test, "no degradation" may be the only acceptable outcome. For another, "temporary degradation with automatic recovery within a specified time" may be acceptable. For a third, "loss of function followed by user-detectable alarm" may still be acceptable. What is never acceptable is a loss of basic safety or an undetected loss of essential performance.

This is why the essential performance definition must be in writing before the test plan is drafted. The lab engineer cannot tell you what counts as pass. Only you can, because only you know what the device must do to be safe and effective for its intended use. The risk management file under EN ISO 14971:2019+A11:2021 is where the essential performance definition lives and where the acceptance criteria are justified.

A test plan without a documented essential performance definition fails the same day it arrives. A test plan with a well-documented essential performance definition sets up a clean pass-or-fail answer for every immunity test.

Common EMC failure modes

The failure patterns we see repeatedly in EMC test reports are not random. They cluster around a small set of design and documentation omissions that would have been cheap to prevent.

  • Radiated emissions above the limit around clock harmonics. The clock trace routes near an unshielded enclosure seam, or the oscillator has no damping, or the PCB stack-up has no continuous ground reference under the high-speed traces.
  • Conducted emissions above the limit on the mains. There is no input filter on the power entry, or the filter is present but its ground return is broken by paint under a mounting screw.
  • Radiated immunity failure at a cellular band. The device firmware has no recovery path from a transient signal glitch, and an input latches in an unsafe state. The fix is sometimes hardware (a trace-level filter), sometimes firmware (watchdog and recovery logic), and often both.
  • ESD failure on a user interface surface. A metallic trim couples the discharge directly to a sensitive trace because the enclosure design has no dedicated static discharge path to chassis.
  • Electrical fast transient failure on a long signal cable. An unshielded cable terminates on an unprotected input, and the transient energy walks straight into the controller.
  • Surge failure on a mains-powered device. The input protection does not include a surge protection device sized for the environment category, and the standard's surge pulse destroys the front end.
  • No essential performance defined. The test campaign cannot start because no acceptance criterion exists for immunity. A documentation failure that costs a full week or more.

Every item on this list is cheaper to prevent at the schematic and enclosure stage than to debug at the accredited lab.

Design-for-EMC tips that actually save money

EMC is a design problem, not a test problem. The money-saving moves happen before the first board is laid out.

  • Define essential performance early. Before schematic freeze, not the week before the lab visit. The essential performance definition drives the acceptance criteria, which drive the test plan, which drives the hardware and firmware design decisions that determine whether the device will pass.
  • Design the four pillars together. Shielding, filtering, grounding, firmware robustness. These are not four separate workstreams. A filter with a broken ground is not a filter. A shield with a firmware glitch path is not a shield. Walk the schematic against all four as a single exercise.
  • Classify the environment once, honestly. Professional, home, or special. Under-classifying is non-compliant. Over-classifying is wasteful. The category should reflect the actual intended use, not a marketing aspiration.
  • Layout matters more than component choice. A good filter laid out badly loses to a cheap filter laid out well. Invest in PCB layout review before tape-out, not after.
  • Run pre-compliance before booking the accredited slot. A cheap pre-compliance sweep at a pre-compliance facility or in-house finds the radiated emissions peaks and the ESD weak spots while they are cheap to fix. The accredited lab is for the final measurement, not the first discovery.
  • Freeze the device before the samples ship. Hardware revision, firmware build, cable set, accessories. A device that is still changing when it arrives at the lab is a device that will need a retest after the report lands.

Common mistakes startups make

  • Booking the accredited EMC lab before essential performance is defined in writing. The immunity acceptance criteria have nowhere to come from and the campaign stalls on day one.
  • Treating emissions and immunity as a single test. Separate budgets, separate setups, separate failure modes, separate design responses. They need separate planning.
  • Classifying the device for the professional healthcare environment because it is "cheaper" when the intended use is actually home healthcare. Non-compliance is not a cost saving.
  • Skipping pre-compliance to save the pre-compliance bill and then paying for two or three full accredited campaigns instead.
  • Bringing a device with cables shorter than the maximum length stated in the instructions for use. The standard requires the worst-case configuration, which includes the longest specified cables, and a short-cable test invalidates the report.
  • Forgetting that EN 60601-1-2:2015+A1:2021 only tests EMC. The general standard EN 60601-1:2006+A1+A12+A2+A13:2024 still has to be satisfied separately. Usually at a separate lab booking.

The Subtract to Ship angle. Match EMC scope to the real environment

The additive instinct on EMC is to over-scope "just to be safe." Test to the home healthcare environment on a device that will only ever be used in hospitals. Apply immunity levels from special environments a particular standard does not require. Add shielding that the risk file does not justify. Every one of these additions costs real money, real weeks, and real engineering attention.

The Subtract to Ship move is not to skimp. It is to match the EMC scope to the actual intended environment and the actual applicable collateral, and then to put the saved engineering budget into getting the few things that do apply right at the schematic stage. Test to the environment the device is intended for. Apply the immunity levels that the standard specifies for that environment. Do not invent additional scope. And do spend the readiness week walking the design against the four pillars. Because that week pays for itself ten times over against the cost of a single rework iteration after a failed lab visit.

The obligation is MDR Annex I Section 14.5, together with Sections 14.2(d) and 17.1. EN 60601-1-2:2015+A1:2021 is the harmonised route to proving you meet it. The accredited lab is where the route is verified. Keeping that direction clear is what turns a first-pass clean report into a reproducible outcome rather than a lucky 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. Is the intended environment (professional healthcare, home healthcare, or special) written down in the instructions for use and in the standards scoping document?
  2. Have you defined essential performance in writing, with specific enough limits that a lab engineer can map each immunity test to a pass-or-fail criterion without asking you questions?
  3. Have you walked the schematic against the four EMC design pillars. Shielding, filtering, grounding, firmware robustness. Before tape-out?
  4. Have you run a pre-compliance EMC sweep before booking the accredited slot?
  5. Do the cables and accessories you plan to ship to the lab match the maximum lengths and the worst-case configuration stated in the instructions for use?
  6. Have you planned for two separate lab bookings. One for the general standard under EN 60601-1:2006+A1+A12+A2+A13:2024 and one for EMC under EN 60601-1-2:2015+A1:2021. Or only for one?
  7. For each expected immunity test, do you know in advance what "still working" means for your device, based on the essential performance definition?

Frequently Asked Questions

What is the difference between emissions and immunity in EMC testing? Emissions measures what the device puts out into the environment. Radiated or conducted. And compares it to limit lines defined by the intended environment category. Immunity measures how the device behaves when the environment pushes disturbances into it and requires that basic safety and essential performance are maintained. Both halves are required for any medical electrical device under EN 60601-1-2:2015+A1:2021.

Which standard governs EMC testing for medical devices under MDR? EN 60601-1-2:2015+A1:2021 is the harmonised collateral standard for electromagnetic compatibility of medical electrical equipment. It is applied together with the general standard EN 60601-1:2006+A1+A12+A2+A13:2024, not instead of it. Following the collateral standard gives presumption of conformity with MDR Annex I Section 14.5 and the connected Section 14.2(d) and Section 17.1.

Do I need separate lab bookings for general-standard testing and EMC testing? Usually yes. General-standard testing under EN 60601-1:2006+A1+A12+A2+A13:2024 and EMC testing under EN 60601-1-2:2015+A1:2021 use different instrumentation and often different facilities. EMC labs need anechoic chambers and specific radiated-field generation equipment. Some facilities host both capabilities, but the test bookings are still separate slots with separate reports.

What intended environments does EN 60601-1-2 distinguish? The standard distinguishes between professional healthcare facility environments, home healthcare environments, and special environments. Home healthcare applies tighter emissions limits and higher immunity levels than professional healthcare, because the real-world electromagnetic conditions are less controlled. Special environments add further requirements for specific settings such as proximity to active HF surgical equipment.

Can pre-compliance EMC testing replace the accredited lab campaign? No. Pre-compliance testing finds large failures cheaply so the accredited lab visit does not become a debug bench. The accredited lab report, from an ISO/IEC 17025 accredited facility, is the evidence that enters the technical file. Pre-compliance is the rehearsal; the accredited campaign is the performance that counts.

What happens if the device has no mains connection. Does EMC still apply? Yes. Battery-powered devices still radiate, still couple through any signal cables, still have to survive electrostatic discharge, radiated RF fields, and power-frequency magnetic fields in the intended environment. Some specific tests. Mains conducted emissions, voltage dips and interruptions. Do not apply when there is no mains connection, but the rest of the EMC plan does.

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 14.5 (electromagnetic compatibility), together with Section 14.2(d) (resistance to electromagnetic disturbances) and Section 17.1 (reliability of electronic programmable systems in their intended environment). Official Journal L 117, 5.5.2017.
  2. EN 60601-1-2:2015+A1:2021. Medical electrical equipment. Part 1-2: General requirements for basic safety and essential performance. Collateral Standard: Electromagnetic disturbances. Requirements and tests.
  3. 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 60601-1-2:2015+A1:2021 is the harmonised route to the EMC obligations in Annex I Section 14.5. The accredited lab is where the route is verified. And the two halves of EMC, emissions and immunity, each need their own honest planning before the device ever arrives at the chamber door.