The eight IEC 60601-1 test failures we see most often in MedTech startup test labs are: leakage current exceeded, inadequate dielectric strength, creepage and clearance violations, EMC immunity failure, EMC emissions above the limit, thermal limits exceeded, mechanical strength failure, and alarm system non-conformity under the 60601-1-8 collateral. Every one of these is a design failure caught at the test bench, not a test-lab problem. Every one of them maps back to MDR Annex I Section 14 or Section 17 as the legal obligation, and to EN 60601-1:2006+A1+A12+A2+A13:2024 or EN 60601-1-2:2015+A1:2021 as the harmonised route. And every one of them is cheaper to prevent at schematic stage than to fix after first tape-out.

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

TL;DR

  • The eight most common IEC 60601-1 test failures all trace back to design decisions made before the first PCB tape-out. Not to test-lab bad luck.
  • Leakage current, dielectric strength, and creepage and clearance failures come from insulation strategy errors and PCB layout errors near the mains section.
  • EMC immunity and EMC emissions failures under EN 60601-1-2:2015+A1:2021 come from skipped shielding, filtering, and grounding work. And from missing essential performance definitions.
  • Thermal and mechanical failures come from enclosure and component choices made for cost or aesthetics without walking the 60601-1 clauses first.
  • Alarm system failures under the 60601-1-8 collateral come from teams applying a general-standard test plan to a device that generates clinical alarms and needs the collateral applied on top.
  • The fix for all eight failures is the same at the process level: open the ISO 14971 risk file first, walk the design against the relevant 60601-1 clauses before tape-out, run pre-compliance testing, and only then book the accredited lab.

Why these eight failures repeat across every startup batch

Test labs see the same mistakes every month. The names on the shipping labels change. The mistakes do not. And the reason is structural: founders treat electrical safety as a test-lab phase instead of as a design discipline, so the same design omissions keep arriving at the same test benches and failing in the same eight ways.

The point of this post is not to shame the pattern. The point is to name each failure so precisely that a founder reading this before first tape-out can stop it from reaching the test lab in the first place. Each failure below names the clause under EN 60601-1:2006+A1+A12+A2+A13:2024 or the collateral that catches it, the MDR Annex I obligation the failure violates, why it happens, and the design fix that prevents it. Read it with your current schematic on the screen. If any of the eight patterns matches something in your design, fix it now. Before the test-lab invoice arrives.

One frame before we start. These failures are not exotic. They are the boring, repeatable failures that any team can prevent with a half-day of review against the right clauses. The reason startups keep hitting them is not that the standard is obscure. It is that nobody on the team has read the relevant clauses yet, and the test lab is the first reader. That is the wrong first reader.

Failure 1. Leakage current exceeded

The test that catches it: Leakage current measurements under EN 60601-1:2006+A1+A12+A2+A13:2024 clauses on earth leakage current, touch current, and patient leakage current under normal condition and single fault condition.

The MDR obligation: Annex I Section 14. Construction of devices and interaction with their environment, specifically the requirement that devices can be used safely under normal conditions and in the event of a single fault condition.

Why it fails: The isolation barrier between the mains section and the patient-applied part is under-specified. Often the team chose a Type B applied part classification when the intended use required Type BF or CF, and the insulation scheme has fewer Means of Patient Protection than the classification demands. Sometimes the Y capacitors on the mains filter are sized without accounting for the permitted patient leakage budget, and their combined leakage exceeds the limit at 110 percent of rated mains voltage.

The fix: Classify the applied part against the intended clinical use before the power supply is selected. Budget patient leakage current at schematic stage. Y capacitors, transformer interwinding capacitance, and stray capacitance to ground all contribute, and they add, not offset. Select a medical-grade power supply whose datasheet already states compliance with the leakage limits relevant to your applied part classification. Do not expect a commercial power supply to pass without requalification.

Failure 2. Dielectric strength inadequate

The test that catches it: The dielectric strength (hipot) test under the insulation clauses of EN 60601-1:2006+A1+A12+A2+A13:2024, where specified AC or DC voltages are applied across each isolation barrier for one minute.

The MDR obligation: Annex I Section 14. Protection against electrical hazards under normal and single-fault conditions.

Why it fails: The isolation barrier does not physically withstand the test voltage because the insulation material, the transformer construction, or the optocoupler rating was chosen on cost rather than on Means of Protection. A 3 kV test is not a formality. It is the worst-case electrical stress the barrier has to survive in use. A barrier designed with a 1.5 kV margin will fail it.

The fix: Specify each isolation barrier explicitly against the required Means of Operator Protection or Means of Patient Protection before sourcing parts. Choose a transformer that carries a named medical-grade insulation rating. Choose optocouplers with a withstand voltage above the worst-case test. Do not fill the gap with a generic mains-rated capacitor across the barrier. It will track and fail.

Failure 3. Creepage and clearance violation

The test that catches it: The physical inspection and measurement of creepage distances (along the surface) and clearance distances (through air) on the PCB and inside the enclosure under the insulation clauses of EN 60601-1:2006+A1+A12+A2+A13:2024.

The MDR obligation: Annex I Section 14. Protection against electrical hazards, specifically the requirement that the construction of the device prevents dangerous contact with live parts under normal and fault conditions.

Why it fails: The PCB was laid out by someone who had not read the insulation clauses. The mains-side traces pass within a few millimetres of the secondary side in places where the required distance is much larger. Vias are placed inside the creepage gap. A mounting hole cuts through the required clearance zone. The layout looks compact and elegant and fails the first physical inspection.

The fix: Mark the required creepage and clearance zones on the PCB stack-up before the layout starts. Treat the mains-to-secondary gap as a no-trace, no-via, no-component zone on every layer. Use a slot in the PCB where necessary to extend the creepage distance along the surface. Have someone who has actually read the insulation clauses walk the layout before tape-out.

Failure 4. EMC immunity failure

The test that catches it: The radiated RF immunity test, the conducted RF immunity test, the ESD test, and the electrical fast transient test under EN 60601-1-2:2015+A1:2021.

The MDR obligation: Annex I Section 14, specifically Section 14.2(d), on resistance to electromagnetic disturbances, and Annex I Section 17.1 on reliability of electronic programmable systems in their intended environment.

Why it fails: The device resets, latches, loses display content, generates a false alarm, or degrades its essential performance when the test lab sweeps the radiated RF disturbance across the band. Or an ESD discharge to an exposed metal trim on the enclosure couples into a sensitive trace and locks a firmware state machine. Or no essential performance was defined in writing, so the test lab cannot declare a pass or fail at all and the campaign stops before it starts.

The fix: Define essential performance in writing before any test plan exists, tied to the ISO 14971 risk management file. Walk the schematic against the four EMC design pillars. Shielding, filtering, grounding, firmware robustness. Before tape-out. Route sensitive signals with continuous ground return. Filter every external interface. Design firmware to detect and recover from transient errors without crossing an essential performance limit. For the detailed walk-through, see post 511 on EMC requirements under EN 60601-1-2:2015+A1:2021.

Failure 5. EMC emissions exceeded

The test that catches it: Radiated emissions and conducted emissions measurements under EN 60601-1-2:2015+A1:2021, against the limits for the intended environment (professional healthcare facility or home healthcare).

The MDR obligation: Annex I Section 14. The requirement that the device does not introduce unacceptable risks into its environment, including disturbances to other equipment.

Why it fails: A clock harmonic radiates above the limit because the clock trace runs near an unshielded enclosure seam. Conducted emissions on the mains exceed the limit because the power entry has no input filter, or the filter is present but its ground return is interrupted by an insulating paint layer under a mounting screw. A switching converter with no damping or spread-spectrum option produces narrow peaks that cross the limit.

The fix: Select a power supply with a known-good EMI filter and a named certification level. Keep clock traces short, referenced to ground, and away from enclosure seams. Treat the enclosure seam as an EMC feature. Gasket it, close it, or move the seam. Bond the filter ground directly to the chassis without insulation in the path. Pre-scan for radiated and conducted emissions in a cheap pre-compliance facility before booking the accredited lab.

Failure 6. Thermal test exceeded

The test that catches it: The temperature measurement clauses of EN 60601-1:2006+A1+A12+A2+A13:2024, measuring touch temperature on the enclosure, on applied parts, and on accessible operator-touchable surfaces, under normal conditions and under single-fault conditions.

The MDR obligation: Annex I Section 14. The requirement that the device does not present thermal hazards to patients, users, or other persons.

Why it fails: An internal component. A voltage regulator, a power resistor, a motor driver. Runs hotter than the team expected and the enclosure surface above it exceeds the limit for operator contact. The limit depends on the surface material and whether the surface is patient-applied, operator-touchable, or not normally touched. The team checked the component datasheet, saw the junction temperature was fine, and did not check the external surface temperature against the standard's table.

The fix: Identify every heat source in the schematic. Model the thermal path from each source to every touchable surface. Choose an enclosure material with a higher permitted touch temperature where helpful. Reposition hot components away from touchable surfaces. Add internal heat spreaders or ventilation. And test thermal performance under the worst-case operating mode, not the idle mode.

Failure 7. Mechanical strength failure

The test that catches it: The mechanical strength clauses of EN 60601-1:2006+A1+A12+A2+A13:2024, including enclosure rigidity tests, impact tests, drop tests for portable equipment, and stability tests for equipment on stands or trolleys.

The MDR obligation: Annex I Section 14. Protection against mechanical hazards including impact, instability, and enclosure failure.

Why it fails: An injection-moulded enclosure chosen for cost cracks at an impact point and exposes live parts. A portable device fails its drop test because the internal mounting of the battery is not rated for the drop energy. A trolley-mounted device tips over during the stability test because the base is too narrow and the centre of mass is too high. A decorative bezel has sharp edges that become a hazard after the impact test cracks it.

The fix: Pick the enclosure material against the mechanical clauses, not only against tooling cost. For portable devices, run an internal drop test early. Before the test lab. To confirm the internal structure holds together. For trolley-mounted devices, check the tilt angle against the stability clause at the concept stage. For anything with a decorative element, treat the element as part of the enclosure and test it accordingly.

Failure 8. Alarm system (60601-1-8) failure

The test that catches it: The alarm system clauses of the collateral standard IEC 60601-1-8, applied on top of EN 60601-1:2006+A1+A12+A2+A13:2024 for devices that generate clinical alarms.

The MDR obligation: Annex I Section 17 on electronic programmable systems and the requirements for reliable performance and fault detection, combined with Annex I Section 14 for the physical alarm signalling.

Why it fails: The device generates alarms. It is a dosing system, a patient monitor, a ventilator, a therapy device. And the team built an alarm scheme based on engineering intuition instead of on the alarm system collateral. Alarm priorities are not distinguished by tone pattern, visual signal, or escalation logic in the way the collateral prescribes. Alarm-off states are not clearly indicated. Alarm reset behaviour does not meet the collateral's requirements. The test lab flags the entire alarm scheme and the software needs a redesign.

The fix: Identify at the concept stage whether your device generates alarms in the clinical sense. If yes, apply the alarm system collateral from the start. Not as a late addition. Design alarm priorities, signalling (auditory and visual), reset behaviour, and alarm-off indication against the collateral's clauses. And map the alarm logic back to the ISO 14971 risk management file, because the alarm system is a risk control measure for specific hazards and has to be traceable to those hazards.

The Subtract to Ship angle. Design for safety early

Every one of the eight failures above is an additive mistake dressed as a subtractive one. The team subtracted the up-front design work against the relevant 60601-1 clauses and added a month of test-lab iteration to pay for it. That trade is never worth it, because a week of schematic review against the clauses costs less than a single accredited-lab re-test, and the accredited lab is the wrong place to debug a design.

The Subtract to Ship move here is not to skip anything. It is to do the small amount of early work that removes the need for the large amount of late work. Open the ISO 14971 risk management file first. Identify the applicable clauses of EN 60601-1:2006+A1+A12+A2+A13:2024 and EN 60601-1-2:2015+A1:2021 for your device. Walk the schematic, the layout, the enclosure, and the firmware against those clauses before tape-out. Run pre-compliance at a cheap facility. Fix what it exposes. And only then book the accredited lab.

The MDR is the North Star. The standards are the efficient route. The eight failures above are what happens when the team forgets that both sentences have to be true at the same time. That the standard without the Regulation is a checklist, and the Regulation without the standard is an unstructured obligation. For the broader methodology see post 065 on the Subtract to Ship framework for MDR compliance.

Reality Check. Where do you stand?

  1. For each of the eight failure modes above, can you name the specific part of your current design that prevents it?
  2. Have you classified your applied parts and sized the isolation barrier before the power supply was sourced?
  3. Have the PCB creepage and clearance zones been marked on the stack-up before layout started, and walked by someone who has read the insulation clauses?
  4. Is essential performance defined in writing, tied to the risk management file, before the EMC test plan was drafted?
  5. Have you pre-scanned radiated and conducted emissions at a cheap facility before booking the accredited lab?
  6. Have you modelled the thermal path from every heat source to every touchable surface under worst-case operation?
  7. If your device generates clinical alarms, is the alarm system collateral in your standards list and applied at concept stage. Or was it added late?

Frequently Asked Questions

Which IEC 60601-1 test do startups fail most often? Across the batches we see, EMC immunity under EN 60601-1-2:2015+A1:2021 is the most frequent first failure, followed closely by creepage and clearance violations found by physical inspection under EN 60601-1:2006+A1+A12+A2+A13:2024. Both are design failures that survived to the test lab because nobody walked the schematic or the layout against the relevant clauses early enough.

Can I fix leakage current problems in firmware? No. Leakage current is a physical property of the insulation barrier, the Y capacitors, the transformer, and the stray capacitance to ground. Firmware cannot change it. Fixing a leakage failure means changing hardware. Typically the power supply, the isolation components, or the filter design.

Do these failure modes apply to battery-powered devices with no mains connection? Most of them do. ESD immunity, radiated RF immunity, radiated emissions, thermal limits, mechanical strength, creepage and clearance between internal voltage domains, and alarm system requirements all apply regardless of power source. Conducted mains emissions and mains-related dielectric tests do not apply if there is no mains connection, but the rest of the picture still does.

Is pre-compliance testing enough to catch these eight failures? Pre-compliance testing catches most of them if the pre-compliance scope covers the relevant clauses. It will not catch failures that come from missing documentation. For example, no essential performance defined in writing. Because those are not test failures, they are planning failures. Pre-compliance is cheap rehearsal, not a substitute for early design review against the clauses.

Do all devices need the alarm system collateral IEC 60601-1-8? No. Only devices that generate clinical alarms. Alarms that warn the user of a patient condition or a device condition that requires a response. Need the alarm system collateral. A device that generates only operator convenience beeps or user interface feedback sounds does not need it. The decision is made at concept stage and should be documented in the standards justification.

Does a failure at an accredited lab mean re-testing the whole standard? Usually no. Most Notified Body reviewers accept partial re-testing of the clauses affected by the corrective action, provided the manufacturer documents why the change does not affect other clauses. The change control process in the QMS, driven by the risk management file, is what determines the retest scope.

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), and Section 17 (electronic programmable systems). 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.
  3. 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.

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 and EN 60601-1-2:2015+A1:2021 are the harmonised tools that get you there. Useful, powerful, and only ever in service of the Regulation itself.