EMC case studies that show the value of pre-compliance

A product can pass every bench functional test and still enter formal EMC testing with a cable, enclosure seam or power converter acting as an efficient RF coupling path.

The examples below are illustrative engineering scenarios, not claimed customer cases. They show the process lessons that repeatedly matter during EMC pre-compliance testing: reproduce the problem, identify the coupling mechanism, change one variable at a time and preserve enough evidence to make the diagnosis defensible.

What an EMC case study should actually demonstrate

A useful EMC case study pre-compliance teams can learn from should explain more than the eventual fix. Adding a ferrite, changing a filter or bonding a panel may move a trace, but that does not prove the underlying mechanism was understood. The same weakness may return when the cable loom changes, a supplier substitutes a component or production tolerances alter the enclosure contact resistance.

Good investigation records connect four elements:

• The disturbance source, such as a switch-mode converter edge, processor clock or transient event.
• The coupling path, including common-mode cable current, capacitive coupling, poor bonding or conducted propagation through a power port.
• The affected receptor or measurement transducer.
• The evidence showing that a controlled change altered the result in the predicted direction.

Pre-compliance testing is particularly valuable because the setup can remain accessible. Probes can be moved, temporary bonds fitted and operating modes changed without consuming formal laboratory time. Calibrated engineering data can also support comparison with later formal measurements, although pre-compliance results do not automatically prove conformity.

Illustrative case study: a cable-driven radiated emissions failure

Consider a networked industrial controller with a plastic enclosure, an external DC supply and several long I/O cables. A development scan shows narrowband radiated emissions associated with the processor and its harmonics. The immediate temptation is to shield the PCB or slow every digital edge.

A better diagnosis starts with the cable currents. A current probe and spectrum analyser can identify whether common-mode current on an external cable follows the same spectral pattern seen by the measurement antenna. Near-field probes may then help localise likely PCB sources, but their readings must not be mistaken for a compliant radiated emissions measurement in dBµV/m.

If disconnecting one I/O cable removes the dominant radiation, the cable is probably part of the antenna system. That still leaves several possible causes: an interrupted return path at the connector, imbalance in a filter, excessive capacitance from a switching node to chassis, or a cable screen terminated through an inductive pigtail.

The team can compare controlled changes such as a short 360-degree screen termination, temporary chassis bonding or common-mode filtering at the connector. A long bonding lead is a poor diagnostic substitute. Its inductance rises with frequency and may prevent the intended RF return current from reaching the reference structure, leading the engineer to reject a sound mitigation concept.

The process lesson is straightforward. Diagnose common-mode current before redesigning the whole board. If the root cause is connector bonding, a processor layout revision may consume weeks without materially changing the radiated result.

Illustrative case study: conducted emissions traced to converter operation

A second product uses an off-line switch-mode power supply and passes functional testing across its load range. During an early conducted emissions investigation, disturbance voltage on the mains port approaches or exceeds the intended product-family limit over part of the spectrum.

For relevant power-port measurements, a suitable LISN provides a defined impedance and a measurement connection to the receiver. Conducted emissions are commonly investigated from 150 kHz to 30 MHz, but the applicable product or product-family standard must determine the actual range, configuration, limits, detector sequence and receiver bandwidth.

Peak detection is often useful for fast exploratory scans. Final comparison may require quasi-peak and average detectors at selected frequencies, using the CISPR receiver settings required by the applicable standard. Treating a peak trace as though it were a completed quasi-peak and average assessment can produce either unnecessary redesign or false confidence.

In this illustrative scenario, changing converter load shifts the spectral envelope. That observation links the disturbance to converter behaviour rather than to a fixed-frequency external source. Differential-mode and common-mode measurements, supported by appropriate probes and circuit knowledge, then guide the filter work.

Adding a larger capacitor without checking the noise mode can make matters worse. A differential capacitor cannot be expected to cure every common-mode path, while additional capacitance to chassis may alter leakage current, safety considerations and high-frequency return paths. Filter components also interact with source and load impedance. A change that looks effective on an open bench may perform differently when installed behind a LISN.

Early access to the product allows the team to compare component changes, cable placement and operating states before committing to a PCB redesign. The result is not merely a lower plot. It is evidence for why the chosen mitigation should remain effective in the representative configuration.

Illustrative case study: immunity failure caused by configuration

Pre-compliance has equal value for immunity. Imagine a control unit that resets during conducted RF immunity investigation on an external signal port. Where the applicable product standard calls up IEC 61000-4-6, a CDN or another coupling method specified by the test arrangement may be used to inject RF disturbance. A BCI probe must not be substituted simply because it is available. Bulk current injection belongs to applicable automotive, military, aerospace or product-specific current-injection procedures, with its own calibration and setup requirements.

The first task is to define failure correctly. A transient display artefact, a recoverable communications error and an unsafe latched output do not have the same significance. The applicable product standard may define performance criteria, while the engineering team must ensure that monitoring can detect the relevant degradation without itself changing the RF behaviour.

Suppose the reset occurs only with a particular sensor cable attached. Altering the cable route changes the threshold substantially. That is useful diagnostic evidence, but it also exposes a setup weakness. If the cable height, length, termination and routing are not controlled, repeated tests may appear contradictory.

Further investigation might find RF entering through signal-reference impedance and disturbing a reset line. Better filtering at the connector, improved reference bonding or changes to reset-line immunity may then be assessed individually. Simply increasing injected level until something fails provides little design insight. The value comes from mapping frequency-sensitive behaviour to a plausible circuit path.

Typical scenario

A product team is six weeks from a planned formal test programme. The enclosure is representative, but firmware, cable drawings and one power-supply option are still changing. Management must decide whether to book formal testing immediately or create a shorter pre-compliance window first.

A sensible plan begins with the intended markets, product function and likely product or product-family standards. CISPR 32 may be relevant to some multimedia equipment, CISPR 11 to certain industrial, scientific and medical equipment, and other products may fall under different standards entirely. IEC 61000-4-x documents provide basic immunity methods when called up by the applicable product requirements. The current published editions, scope, ports, test levels, frequency ranges, limits, configurations and documentation requirements must all be checked rather than inferred from a previous project.

The investigation should target the highest project risks: representative radiated and conducted emissions, likely immunity weaknesses, worst-case operating modes and configuration-sensitive cables. Firmware must exercise real data traffic and loads. An idle processor and unterminated interface can produce a reassuring result that disappears as soon as the product performs its intended function.

EMC Hire can support this work through equipment hire, accessible pre-compliance engineering, on-site testing and test facility access. Where appropriate, the team can also help plan formal compliance testing for self-certification applications without acting as a certification body. Equipment used for relevant measurements is calibrated with traceability through an appropriate ISO/IEC 17025 accredited calibration provider. Suitable traceability supports repeatability, comparison between development and formal testing, and stronger evidence for technical and regulatory review.

Finding a cable-current problem at this stage may allow a connector, bonding or filter change. Finding it after tooling release could require enclosure modification, PCB redesign and another formal booking. That is where early testing produces measurable risk reduction.

Turning measurements into defensible evidence

Pre-compliance records should be treated as engineering evidence rather than disposable debugging notes. Save receiver settings, detector type, bandwidth, transducer factors, photographs, cable dimensions, EUT orientation, software revision, loads, accessories and ambient conditions. Record failed experiments as well as successful ones.

Formal compliance testing supports more than a pass or fail decision. Robust evidence may contribute to the technical file, EMC risk assessment, RF environment gap analysis, mitigation records, Declaration of Conformity and customer documentation. For CE marking or other self-certification routes, the manufacturer or responsible economic operator remains responsible for determining the legislation, standards, conformity assessment route and documentation that apply. Testing alone does not complete every obligation.

Product teams preparing commercial equipment can review the practical issues around commercial CE marking and use the EMC test guides to structure early planning. Defence, automotive and aerospace programmes can also benefit from pre-compliance investigation, although contractual or sector requirements may require final work at an appropriately accredited laboratory.

When to Hire EMC Equipment

Hiring suits irregular development demand. A team may need a receiver, LISN, current probe, near-field probes or immunity equipment for two weeks, then have no comparable requirement for months. Purchasing around one project ties up capital and leaves the business responsible for storage, servicing, accessories and calibration.

Rental also reduces the risk of buying equipment that does not match a later programme. Receiver frequency coverage, detector capability, LISN ratings and coupling devices must align with the actual method. Equipment selected for conducted emissions cannot simply be repurposed for conducted immunity.

Short-term hire can expand capability during project peaks, support on-site diagnosis and keep a prototype accessible between tests. The useful question is not whether an instrument is generally described as EMC equipment. It is whether its ratings, calibration status, connectors, ancillary components and software are suitable for the defined test window.

Common EMC Testing Mistakes to Avoid

Treating cable position as an incidental detail

Cables are often dominant coupling structures. Moving one between scans can change common-mode current distribution and antenna efficiency, destroying repeatability. Photograph and dimension the routing.

Using the wrong network or coupling device

A LISN supports relevant conducted emissions measurements. A CDN supports specified conducted RF immunity methods. Confusing them produces a physically different test and an evidence trail that cannot support the intended assessment.

Running an unrepresentative operating mode

Reduced processor activity, absent peripherals or artificial loads may miss the worst emissions state or immunity response. Define operating modes from clock activity, power conversion, communications and safety-related functions rather than convenience.

Ignoring detector and bandwidth requirements

Peak, quasi-peak and average detectors are not interchangeable. Nor is one resolution bandwidth appropriate across every frequency range and standard. Incorrect settings can shift measured values and invalidate comparison with a limit.

Changing several variables at once

Replacing a filter, moving a cable and adding a chassis bond in one step may improve the trace, but the team learns nothing about which change worked. Production then inherits unnecessary cost and uncertain margin.

Failing to preserve the configuration

Missing firmware revisions, cable photographs, accessory details or calibration records make later reproduction difficult. A graph without setup evidence is weak support for diagnosis, formal planning or the technical file.

Frequently Asked Questions (FAQs)

How close must pre-compliance measurements be to formal results?

The required correlation depends on the decision being made. Early debugging may focus on trend and margin, while final readiness work needs a more representative site, configuration and measurement chain. Known differences should be quantified rather than hidden.

Can pre-compliance data be used in a technical file?

It may contribute useful calibrated engineering and mitigation evidence, particularly for a justified self-certification route. Its adequacy depends on applicable legislation, standards, uncertainty, setup control and product risk. The manufacturer must make that determination.

Should emissions or immunity be investigated first?

Start with the dominant technical and schedule risk. A noisy converter may justify early conducted emissions work, while exposed long cables and safety-related control functions may place immunity higher. Many programmes need both before design freeze.

Does a near-field probe result predict a radiated emissions pass?

No. Near-field probes are localisation tools. They help identify sources and compare changes, but they do not reproduce the calibrated antenna measurement geometry, site characteristics or limit comparison used for radiated emissions.

How much margin is enough?

No universal figure is defensible. Consider setup correlation, production spread, cable options, operating modes, component tolerances and measurement uncertainty. A small margin on an unstable configuration deserves more investigation than a repeatable result with a understood coupling path.

Discuss the next test window

If a prototype is approaching design freeze, EMC Hire can help define a proportionate route covering equipment hire, on-site diagnosis, pre-compliance testing, formal compliance testing where appropriate, or space at the EMC Hire test facility.

Speak with the engineering team on +44 (0)1462 817111 or email sales@emchire.co.uk to discuss the EUT, applicable test methods, available evidence and the equipment needed for a controlled test window.