How to find the root cause of medical device radiated emissions

A failing frequency on the chamber plot is rarely the root cause. The real fault may be a clock edge, a poorly bonded shield or a few microamps of common-mode current flowing along an external cable.

Effective troubleshooting of medical radiated emissions depends on separating three parts of the problem: the noise source, the coupling path and the structure acting as an antenna. Changing components without identifying all three can produce an apparent improvement that disappears when the enclosure is closed, the patient cable is fitted or the equipment under test, or EUT, enters another operating mode.

Start with a measurement you can reproduce

Before opening the enclosure, preserve the failing configuration. Record the EUT operating mode, software version, power source, cable types, cable positions, peripheral equipment, enclosure state, antenna polarisation, turntable angle and antenna height where applicable. Photographs are often more useful than written descriptions for cable geometry.

Medical electrical equipment may have mains leads, patient connections, data cables, transducer cables and functional earth conductors. Each can alter the common-mode current distribution. Moving one cable by 200 mm may change the measured field without changing the underlying noise generation. If the setup is not documented, the team cannot distinguish a genuine design improvement from a geometry-dependent result.

The applicable product requirements must be established before diagnosis begins. IEC 60601-1-2 addresses EMC for medical electrical equipment and systems, while its emissions requirements refer to other standards, including CISPR 11 in relevant cases. Product classification, intended environment, port configuration and operating mode affect the test plan. Engineers should check the latest active editions, amendments, test levels, frequency ranges, limits and documentation requirements rather than treating a previous product's plan as a template.

A compliant emissions receiver or suitable analyser-based development system can establish the spectral signature. Peak detection is useful for rapid investigative sweeps, but final detector selection, measurement bandwidth and measurement method must follow the applicable requirements. A peak result below a limit can sometimes reduce the need for slower detector measurements, depending on the standard and laboratory procedure, but a development scan should not be presented as formal compliance evidence without confirming that the complete method has been followed.

EMC Hire's radiated emissions measurement systems can support development investigations using equipment selected for the required frequency range and measurement task. Test equipment used for measured results is provided with calibration traceable through an appropriate ISO/IEC 17025 accredited calibration provider. This supports repeatability and comparison between development work and later formal testing. It does not, by itself, make an informal setup compliant with a standardised test method.

Read the spectrum as evidence, not just a pass or fail plot

The spacing and shape of emissions often reveal more than their absolute amplitude. A comb with regular spacing may point towards a digital clock, switch-mode converter or periodic data transfer. Broad spectral energy may arise from fast switching edges, arcing, motor commutation or multiple asynchronous sources. Narrow, stable lines usually justify checking oscillators and clocks first, while frequency movement with load or operating state can implicate a variable-frequency converter.

Calculate likely harmonics, but do not assume that the oscillator with the matching fundamental is automatically responsible. A 25 MHz clock harmonic and a 25 MHz switching artefact can occupy the same frequency. Triggering an oscilloscope from suspected sources, changing controlled operating parameters and comparing spectral movement provides stronger evidence than harmonic arithmetic alone.

Time correlation is particularly useful in medical devices with pumps, displays, wireless modules, heaters or acquisition cycles. If an emission rises only during display refresh or motor drive activity, capture that state deliberately. An average operating mode can hide the worst case and create false confidence before formal testing.

Locate board-level noise without mistaking the probe for an antenna test

Near-field probes help identify local electric and magnetic fields around processors, DC-DC converters, memory interfaces, connector regions and cable exits. They are diagnostic tools. Their readings do not directly predict a standardised far-field result because probe orientation, spacing and loading strongly influence amplitude.

Use a repeatable probe height and orientation, then compare relative changes at the failing frequencies. A magnetic-field probe is often productive around high di/dt current loops, inductors and switching nodes. An electric-field probe can expose high dv/dt conductors and poorly contained fields, although capacitive loading from the probe may disturb sensitive circuitry.

Board-level noise commonly escapes because the return path is discontinuous. A high-speed trace crossing a split reference plane forces return current around the discontinuity, enlarging the loop area and increasing coupling into nearby structures. Likewise, an unfiltered signal reaching an external connector can drive the attached cable in common mode even when the differential waveform appears clean.

Temporary modifications should answer one question at a time. Adding a correctly placed capacitor, slowing an edge under controlled conditions or applying local absorber material can test a hypothesis. Randomly covering the board with copper tape may lower one peak while introducing new resonances or unsafe clearances. In medical equipment, temporary conductive modifications must also be assessed against electric shock, insulation, leakage current, thermal and functional risks before power is applied.

Cabling often becomes the most efficient radiator

An external cable does not need a signal at the failing frequency to radiate. Parasitic capacitance between a noisy board and chassis can drive common-mode current onto a cable shield or conductor. Once cable dimensions become electromagnetically significant, very small currents can produce a measurable field.

A calibrated current probe, or a characterised probe used for relative debugging, can measure common-mode cable current at the suspect frequency. Move the probe along the cable and compare operating states. Strong current near the enclosure exit suggests coupling at the connector, shield termination or internal harness. Large changes caused by cable position indicate that the cable is part of the radiating structure, although position sensitivity alone does not identify the original source.

Clamp-on ferrites are useful diagnostic components when their impedance is appropriate at the frequency under investigation. If a ferrite around the complete cable reduces the chamber peak, common-mode cable current is implicated. Placing it around only selected conductors changes the experiment and may affect differential current or functional operation. Ferrite material, geometry, turns and bias all matter, so manufacturer impedance data should be checked rather than selecting a core by appearance.

Cable shields require controlled termination. A long shield pigtail adds inductance and degrades high-frequency bonding, allowing shield voltage to develop relative to the enclosure. A circumferential connection at the entry point generally provides lower transfer impedance at high frequencies, where the mechanical and safety design permits it. The final arrangement still has to satisfy the medical device's insulation, patient protection and functional requirements.

Shielding only works when current has somewhere to go

Shielding failures are frequently bonding failures. A conductive enclosure with painted seams, sparse fasteners or isolated panels may behave as several coupled antennas rather than one equipotential structure. Displays, ventilation apertures and connector cut-outs can also leak fields when their dimensions and local field orientation support coupling.

Use conductive foil, finger stock or absorber as a controlled diagnostic measure, not an immediate production solution. Bridging one seam at a time can identify a high-impedance joint. Temporarily bonding a connector shell to the enclosure can test whether shield current is being diverted through the PCB. If that change reduces emissions, the production fix may involve connector mounting, plating, gasket design or chassis bonding rather than extra filtering on every signal line.

Keep temporary bonds short and wide. A long wire may appear satisfactory at lower frequencies yet become inductive at higher harmonics, leading the team to dismiss a bonding hypothesis incorrectly. Document every modification and restore the baseline between experiments.

Separate radiated and conducted mechanisms

Noise observed in a radiated test can originate on the mains port. A switch-mode converter may place disturbance voltage onto the supply lead, which then contributes to radiation through common-mode current. Conversely, strong local board noise may couple onto the mains cable without producing the dominant conducted emissions result.

Investigating both paths can resolve ambiguity. A LISN is used for conducted emissions measurements on relevant power ports and provides a defined impedance and measurement connection. It is not a conducted immunity device. EMC Hire's conducted emissions test systems can help teams compare mains-port disturbance with the radiated spectrum where the product requirements call for that measurement.

Matching spectral features across current-probe, conducted-voltage and radiated measurements strengthens a diagnosis, but amplitude relationships are rarely one-to-one. Each setup measures a different physical quantity and transfer path.

Typical scenario

Consider an illustrative bedside medical device that exceeds a radiated emissions limit at several harmonics while connected to mains power, a patient cable and an Ethernet lead. The engineering team first repeats the measurement with the prescribed accessories and worst-case software mode. Antenna polarisation and turntable angle show that the dominant emission is associated with the cable side of the enclosure.

A near-field scan identifies matching board-level noise around a processor and its display interface. Yet disabling the display only partly reduces the chamber peak. A current probe then finds corresponding common-mode current on the patient cable. Temporarily improving the connector-shell bond produces a larger reduction than local shielding over the processor.

The likely mechanism is now clearer: the processor interface is the source, parasitic coupling into the chassis and connector region is the path, and the patient cable is the effective antenna. Treating only the source may not provide enough margin. Treating only the cable may leave susceptibility or consistency problems. A robust correction could require edge control, return-path improvement and a lower-impedance enclosure connection, subject to safety and functional review.

Early investigation allows these changes to be evaluated before tooling and formal testing. EMC Hire can support medical device EMC testing through equipment hire, accessible pre-compliance work, test-facility access, on-site measurements and practical setup support. Where appropriate, formal compliance testing can provide evidence for the technical file, EMC risk assessment, Declaration of Conformity and internal or stakeholder review. The manufacturer remains responsible for identifying the legislation, standards, conformity route and documentation applicable to the product.

When to Hire EMC Equipment

Hiring is well suited to root-cause work because the required instruments often change as the diagnosis develops. A receiver and antenna may establish the failing frequencies, near-field probes may locate board sources, and a current probe may then confirm the cable coupling path. Buying an incomplete system too early can lock the team into equipment with unsuitable bandwidth, dynamic range, detector capability or accessories.

Rental also addresses short development peaks without adding long-term storage, servicing and calibration overheads. It can provide the correct equipment for a defined debug window, support parallel projects or cover the period before a formal laboratory booking. This is particularly useful when emissions work is irregular and capital equipment may not suit the next programme.

Equipment access is only part of the decision. If the team lacks a controlled site, a facility booking may offer better repeatability. For large, installed or difficult-to-transport systems, on-site investigation may preserve the real cabling, grounding and peripheral configuration. EMC Hire can help select a proportionate route, from engineering measurements through to medical device EMC compliance support.

Common EMC Testing Mistakes to Avoid

Changing several variables together

Adding ferrites, foil and filter capacitors simultaneously may improve the plot, but it destroys diagnostic value. If the result later changes in production, nobody knows which mechanism mattered. Make one controlled change, measure it, record it and return to baseline.

Ignoring cable geometry

Coiling excess cable or routing it differently between tests changes antenna behaviour and mutual coupling. The resulting improvement may be a setup artefact rather than a design correction.

Using near-field amplitude as a compliance result

A strong local reading does not prove that a circuit dominates far-field emissions. Equally, a modest reading can feed an efficient cable antenna. Near-field probes locate candidates; the controlled radiated measurement verifies system-level effect.

Testing an easy operating mode

Idle software, inactive pumps or disconnected accessories can remove the emission mechanism. A passing result in an unrepresentative mode provides poor evidence and may fail to expose the configuration required by the product test plan.

Applying incorrect receiver settings

Using an arbitrary resolution bandwidth or detector changes the measured amplitude and can hide narrow or impulsive disturbances. Settings must correspond to the applicable frequency range and measurement method. Development shortcuts should be labelled clearly in the records.

Failing to preserve evidence

Plots without cable photographs, EUT modes, instrument settings, antenna details or modification records are difficult to reproduce. Weak records also make it harder to compare pre-compliance data with formal results or build a defensible technical file.

Frequently Asked Questions (FAQs)

Can a near-field probe identify the exact radiated emissions source?

It can identify candidate circuits and coupling regions, but it cannot independently prove which structure dominates the far-field result. Confirm the hypothesis by changing the suspected source or path and observing the controlled radiated measurement.

Should we investigate shielding or filtering first?

Follow the evidence. If common-mode current is present on an external cable, examine connector bonding, filtering and return paths. If energy escapes through a seam or aperture, shielding and bonding deserve attention. Starting with the easiest mechanical change often wastes time.

Why does the emission move when the cable is repositioned?

The cable is probably participating in the radiating structure or altering coupling to another structure. Repositioning changes its effective geometry and current distribution. It does not necessarily mean that the cable carries the original noise source.

Can pre-compliance testing prove conformity with IEC 60601-1-2?

No. Pre-compliance testing can find problems, provide calibrated engineering data and improve confidence before a formal programme, but it does not automatically demonstrate that every applicable configuration and requirement has been satisfied.

When should debugging move into a formal test facility?

Move when the design is stable enough to justify controlled, repeatable measurement, or sooner if ambient noise and site geometry prevent reliable diagnosis. A facility is also preferable when small margin changes must be resolved accurately.

Plan the next measurement

Root-cause work becomes faster when every measurement is designed to separate source, path and antenna. If your team needs suitable equipment, a controlled test environment or engineering support, contact the EMC Hire team on +44 (0)1462 817111 or email sales@emchire.co.uk.

EMC Hire can discuss an equipment hire quotation, on-site investigation, pre-compliance work, formal compliance testing where appropriate, or a booking at the EMC test facility. The aim is straightforward: obtain repeatable evidence early enough to make sound design decisions before late-stage redesign becomes the only option.