Conducted emissions fixes you can try in one day

A failed conducted emissions plot rarely tells you which component to replace. It tells you that unwanted RF energy has found an efficient path onto a measured cable.

When the retest is days away, random component changes waste time and often obscure the original fault. The quickest conducted emissions fixes are controlled experiments that identify whether the dominant mechanism is differential-mode current, common-mode current, poor routing, ineffective filtering or enclosure coupling.

Start with the failed plot, not the circuit diagram

Obtain the original receiver trace, limit line, detector results, equipment under test configuration and photographs before touching the hardware. Record which power port was measured, the Line Impedance Stabilisation Network (LISN) arrangement, supply voltage, cable layout, operating mode and connected accessories.

Conducted disturbance voltage is commonly measured in dBµV using a LISN on relevant power ports. Many commercial standards cover a range from 150 kHz to 30 MHz, but this is not universal. The applicable product or product-family standard determines the frequency range, limits, detector requirements, receiver bandwidth, port configuration and operating conditions. Check its latest active edition rather than copying settings from an earlier project.

Peak scanning can locate disturbances quickly. Formal comparison with a limit may require quasi-peak and average detectors at specified frequencies. A peak trace below the applicable limit can be useful evidence during debugging, but using the wrong detector or resolution bandwidth can produce a misleading margin.

Look at the shape of the failure. Narrow, evenly spaced lines often correspond to switching frequencies, clock harmonics or modulation products. A broad rise may indicate switching-edge energy, diode recovery, brush noise or multiple unstable resonances. Frequencies that move when load or clock settings change provide particularly useful diagnostic evidence.

One-day conducted emissions fixes checklist

Prove whether the noise follows the operating mode

Run a short set of repeatable operating conditions: idle, maximum processing activity, representative load, communications active and each major power converter enabled where the design permits. Change one condition at a time.

If a spectral line moves when a converter frequency changes, work on that converter and its coupling path. If it remains fixed, investigate clocks, displays, communications interfaces and other periodic sources. Testing an unrepresentative idle mode may produce an attractive plot that has little value when the product is later exercised under its worst credible operating condition.

Re-route high di/dt loops before adding parts

Inspect the input capacitor, switching device, diode or synchronous rectifier and transformer loop. The physical loop area matters more than the neatness of the schematic. Long tracks between the input capacitor and switching stage add inductance, increasing voltage developed by fast current transitions and encouraging RF current to seek alternative returns through chassis, cable screens and measurement earth.

For a same-day experiment, add a suitable local capacitor directly across the switching-stage supply and return, using the shortest safe connection available. Keep temporary connections mechanically stable and observe voltage, ripple-current, temperature and safety ratings. A long flying lead can make a nominally suitable capacitor ineffective at RF because lead inductance dominates its impedance.

Move noisy internal wiring away from mains leads, DC input cables and I/O harnesses. Avoid running a switch-node conductor parallel to an external cable. Even without galvanic connection, capacitive and magnetic coupling can transfer switching energy onto the measured port.

Use ferrites as diagnostic components

Clamp-on ferrites are quick to fit and valuable for identifying common-mode cable current. Place the ferrite close to the point where the cable exits the equipment. That location restricts current before the cable becomes part of the coupling structure.

If practical, pass the complete cable through the core so both outgoing and return conductors are enclosed. This primarily adds common-mode impedance while allowing normal differential load current to cancel magnetically. Putting a core around only one conductor also affects differential current, risks saturation at higher load current and may disturb converter operation.

Core material and geometry determine impedance versus frequency. A ferrite described only by its outside diameter is not adequately specified. Check the manufacturer’s impedance curve, number of turns, bias behaviour, temperature range and current conditions. Extra turns can raise impedance approximately with the square of turns in a simplified low-level model, but inter-winding capacitance and core loss eventually limit the benefit.

A large improvement after fitting a common-mode ferrite is evidence about the coupling mechanism, not automatically a production solution. Mechanical retention, cable variability, thermal performance, safety spacing and repeatability still require assessment.

Change filter placement, not just filter values

An input filter must sit at the boundary between the noisy circuit and the external cable. If an unfiltered conductor travels across the PCB before reaching the filter, it can couple noise around the filter components. Likewise, routing filtered and unfiltered tracks alongside each other creates a parasitic bypass.

Try relocating a temporary capacitor or ferrite closer to the connector, while maintaining applicable safety requirements. For a differential-mode disturbance, an appropriately rated capacitor across the supply conductors may reduce RF voltage. Common-mode suppression may require a common-mode choke and capacitors to a suitable RF reference, where the product architecture and safety rules permit them.

Do not increase capacitance blindly. On mains-powered equipment, leakage current, discharge provisions, inrush behaviour and component safety classification constrain the available options. A filter can also interact with a converter’s negative input impedance, producing peaking or instability. Verify operation across expected supply and load conditions and consult the component manufacturer’s application data.

Improve grounding without creating a longer RF path

The word grounding hides several different functions: protective earthing, functional bonding, signal reference and RF return. They are not interchangeable.

For a metal enclosure, test whether a short, wide bond between the filter reference or connector panel and chassis reduces the disturbance. A narrow, unnecessarily long wire has appreciable inductance at several megahertz. It may look like a short circuit on a continuity meter while behaving as a poor RF connection.

Never modify protective-earth arrangements casually. Temporary debugging changes must preserve electrical safety, insulation, fault-current capability and the intended equipment class. If the emissions improve only after an unsafe bonding arrangement is introduced, the result identifies a mechanism but is not a permissible design fix.

Check enclosure bonding and cable screens

Paint, anodising, gaskets and loose fixings can prevent metal panels from forming a continuous RF enclosure. Measure and inspect joints, then trial a short, broad temporary bond across a suspect seam. A meaningful reduction points towards common-mode voltage between enclosure sections or an inadequate screen-current path.

Where a screened cable enters a conductive enclosure, a low-inductance circumferential termination generally controls high-frequency screen current better than a long pigtail. The pigtail’s inductance develops RF voltage, allowing current to couple into internal circuitry and other cables. Screen treatment must still match the interface design, safety requirements and applicable product standard.

Separate source suppression from port filtering

Reducing edge rate at the source can be more robust than trapping energy at every cable. Trial changes might include a gate resistor adjustment, an RC snubber based on measured ringing, improved transformer shielding or correction of a rectifier recovery problem.

These changes affect efficiency, switching loss, device stress and thermal performance. Monitor waveforms with correctly rated probes and appropriate measurement technique. A long oscilloscope probe ground lead can create apparent ringing that is not present at the device, leading to an incorrectly tuned snubber and extra dissipation.

Typical scenario

Consider an illustrative project involving a mains-powered controller that exceeds its conducted emissions limit at several harmonics of an internal switch-mode converter. Formal retesting is booked for the following week.

The team first recreates the laboratory operating mode using a suitable LISN, EMI receiver or appropriately configured analyser, transient protection and a documented cable layout. A peak scan provides rapid feedback, while final detector measurements are reserved for the frequencies that matter. The setup must follow the applicable standard closely enough that improvements are comparable with the earlier result.

A clamp-on ferrite around the complete mains cable reduces the harmonics. Moving the internal mains conductors away from the converter reduces them further. Those observations suggest common-mode coupling from the switching stage into the mains cable. The engineers can now investigate filter-to-chassis bonding, parasitic capacitance and enclosure current paths rather than changing every differential filter component.

Early investigation lowers the chance of arriving at the formal test with an unproven modification. EMC Hire can support this work through equipment selection, conducted emissions measurement equipment, pre-compliance engineering, test-facility access and on-site testing. Equipment used for relevant measurements is calibrated with traceability through an appropriate ISO/IEC 17025 accredited calibration provider, supporting repeatability and comparison between development and formal test data.

Pre-compliance results do not automatically demonstrate compliance. They can provide calibrated engineering data, improve confidence before formal testing and support a technical file or self-certification process where legally and technically appropriate. The manufacturer remains responsible for determining applicable legislation, standards, conformity assessment routes and documentation.

When to Hire EMC Equipment

Hiring is often more defensible than purchasing when conducted emissions work is tied to a short design sprint, an isolated failure or an irregular product programme. It gives the team access to a suitable LISN, receiver, probes and supporting equipment for a defined test window without committing capital to hardware that may not suit the next product’s voltage, current, frequency range or port type.

Ownership also brings storage, servicing, verification and calibration obligations. A receiver with inadequate detector functions, or a LISN with unsuitable current and voltage ratings, does not become useful simply because it is already on the asset register.

Rental can cover project peaks while preserving access to traceable measurement equipment. The engineering discussion should include supply type, maximum current, applicable standard, connector arrangements, frequency range, EUT environment and whether the work is diagnostic or intended to contribute to formal evidence.

Conducted emissions should not be confused with conducted immunity testing. A LISN supports conducted emissions measurement on relevant power ports. CDNs and suitable current-injection methods are used for conducted RF immunity where called up by the applicable method. EMC Hire also provides information on conducted immunity systems and radiated emissions systems when the wider test programme covers those separate phenomena.

Common EMC Testing Mistakes to Avoid

Changing several variables together

Adding a ferrite, moving a cable and replacing capacitors in one step may improve the plot, but it removes diagnostic value. If the result later changes in production, the team will not know which feature controlled the emission.

Ignoring LISN and earth layout

An unsuitable LISN configuration, inconsistent bonding or uncontrolled cable placement changes the RF impedance seen by the EUT. Results may then differ substantially from the formal setup, creating false confidence or a false failure.

Comparing incompatible traces

Plots taken with different detector types, attenuation, resolution bandwidths or operating modes cannot be treated as direct before-and-after evidence. Save instrument settings with every trace and note any overload protection or external attenuation.

Treating ambient signals as EUT emissions

Switch off the EUT or remove its power, where safe and methodologically appropriate, to establish the ambient baseline. A persistent signal may originate from nearby equipment or the supply environment rather than the product.

Failing to record the physical setup

Photograph cable positions, ferrite locations, enclosure fixings and temporary component connections. Without those records, a successful experiment may be impossible to reproduce at the next test or transfer into a controlled production design.

Frequently Asked Questions (FAQs)

Can a clamp-on ferrite be accepted as the final fix?

Potentially, if its impedance, material, thermal behaviour, mechanical retention, current conditions, safety impact and manufacturing controls are suitable. The production configuration must then be retested. A convenient laboratory ferrite is not automatically a controlled production component.

How can we distinguish common-mode from differential-mode noise?

A current probe around both supply conductors responds mainly to net common-mode current, while measurements around individual conductors contain both components. Response to a common-mode choke or whole-cable ferrite is another useful clue. Interpretation requires care because asymmetry and parasitic coupling can mix the modes.

Should we optimise to the peak or quasi-peak result?

Use peak detection for efficient scanning and debugging, then apply the detector requirements specified by the applicable standard at relevant frequencies. Do not assume peak, quasi-peak and average values are interchangeable.

How much margin should a pre-compliance result have?

No single margin suits every product. Consider measurement uncertainty, setup reproducibility, component tolerances, operating modes, supply variation and differences between development and formal facilities. A narrow pass deserves further investigation rather than optimistic rounding.

Can conducted emissions testing support CE or UKCA documentation?

Suitable test evidence may support the technical file, EMC risk assessment, mitigation records and Declaration of Conformity process where applicable. Testing alone does not complete every conformity obligation. Confirm the latest legislation, product-specific standards, test levels, limits, configurations and documentation requirements.

Plan the next measurement

If the project is close to retest, send EMC Hire the failed plot, applicable standard, EUT supply details, operating modes and photographs of the previous setup. The engineering team can help identify suitable hire equipment, arrange on-site testing, discuss formal compliance or pre-compliance work, or book space at the EMC Hire test facility.

For a practical discussion or equipment hire quotation, call +44 (0)1462 817111 or email sales@emchire.co.uk.