How to test rail electronics for conducted emissions

A rail electronics unit can appear quiet on the bench yet fail conducted emissions testing once its supply return, harness and enclosure are configured as they will be in the vehicle.

The measurement receiver is rarely the source of that discrepancy. More often, the test impedance, cable geometry or bonding arrangement has changed the noise current path.

Start with the applicable rail requirement

Rail conducted emissions work should begin with the product scope, installation environment and contractual test plan, not with an available LISN and a convenient receiver sweep. The EN 50121 series addresses electromagnetic compatibility within railway applications, but the relevant part depends on what is being assessed and where it will operate.

For rolling stock apparatus, EN 50121-3-2 is commonly relevant. Fixed installations, signalling and telecommunications apparatus may fall within other parts of the series, including EN 50121-4 where its scope applies. EN 50155 may also affect the wider environmental and electrical qualification of electronic equipment used on rolling stock, but it should not be treated as a substitute for identifying the applicable EMC emissions requirements.

Project specifications can add limits, configurations or operating conditions beyond a product standard. Operators and rolling-stock manufacturers may also define their own test plans. Confirm the latest active editions, frequency ranges, limits, detectors, bandwidths, port applicability and documentation requirements before selecting equipment.

The first engineering decision is whether the measurement concerns a DC or AC power port, a signal line, an auxiliary connection or another defined port. The coupling network, permissible loading and measurement quantity must match that port. A power LISN intended for a low-voltage laboratory product may be electrically unsuitable for a high-current rail converter, even if its nominal impedance looks familiar.

Define the noise path before connecting the receiver

Conducted emissions are normally a combination of differential-mode and common-mode currents. Differential-mode noise flows out on one supply conductor and returns on the other. Common-mode current leaves through multiple conductors together and returns through enclosure capacitance, chassis bonds, mounting structures or other parasitic paths.

This distinction matters. Adding a capacitor across the DC input may reduce differential-mode noise while doing little for common-mode current. Conversely, fitting capacitors from the lines to chassis can reduce common-mode voltage but increase leakage current, alter insulation behaviour or create problems elsewhere in the rail system.

Switch-mode converters are frequent sources. Their switching edges excite the input harness, transformer capacitance, heatsinks and enclosure. Long supply lines add inductance, while the capacitance between the equipment under test, or EUT, and the reference plane provides a return path that may not have existed during normal functional bench testing.

Use a current probe during investigation to compare current on individual lines with current around the complete harness. If the probe encloses both supply and return conductors, differential currents largely cancel and the remaining response can provide useful evidence of common-mode current. This is a diagnostic technique, not automatically the compliance measurement required by the standard.

LISN selection for rail power ports

A line impedance stabilisation network presents a defined impedance to the EUT over its characterised frequency range, isolates the measurement from supply-side RF noise and provides a receiver port for conducted disturbance voltage measurements. LISNs are used for conducted emissions. They are not conducted immunity generators or coupling devices.

Rail supplies complicate the selection. Check at least:

• Maximum continuous and transient supply voltage
• Continuous current and likely inrush current
• Number of lines and treatment of the return conductor
• Network impedance specified by the applicable method
• Characterised frequency range
• RF port attenuation or transducer factor
• Protective earth, chassis and reference-plane connections
• Safety provisions for stored energy and hazardous DC supplies

Do not select a LISN solely by current rating. Its impedance characteristic and insertion behaviour must suit the test method. Likewise, a nominally suitable high-current network can still produce misleading data if the RF port factor, cabling or grounding arrangement is not accounted for.

Follow the LISN manufacturer's documentation for connection sequence, earthing, discharge time, receiver protection and any external termination requirements. High-energy DC systems need a documented safe system of work. The RF input of a receiver can be damaged by transients or excessive residual mains-frequency energy, so an appropriate transient limiter, pulse limiter or attenuator may be needed where permitted by the measurement method.

EMC Hire's conducted emissions test equipment can be configured for development and pre-compliance work, subject to the voltage, current, frequency range and network requirements being established first.

Reference planes, lines, returns and harness layout

The reference plane is part of the RF circuit. Its dimensions, material, bonding and relationship to the EUT affect common-mode impedance. Placing the unit on an arbitrary wooden bench and clipping the LISN to a distant earth point creates a different circuit from a controlled test arrangement.

Mounting must represent the specified configuration. If the enclosure is bonded to a metal plane in service or during the prescribed test, use the defined bond. If it is isolated, do not add an informal earth lead to cure an unstable plot. That lead creates another RF return path and can suppress or amplify emissions depending on its length and resonance.

Harness routing needs the same discipline. Control:

• Exposed cable length between the EUT and LISN
• Height above the reference plane
• Separation between supply, signal and load cables
• Shield termination and connector backshell bonding
• Location and construction of representative loads
• Unused-port termination

Photograph the arrangement from several angles and record dimensions. A statement such as “cables laid on bench” is not enough to reproduce a marginal result six weeks later.

Receiver settings and measurement sequence

Rail conducted emissions measurements often cover frequencies associated with switching converters and their harmonics, but the exact measurement range comes from the applicable standard or contractual plan. Although 150 kHz to 30 MHz is common in many conducted emissions methods, it must not be assumed for every rail port or project.

Use the prescribed CISPR receiver bandwidth and detector for each part of the measurement. Peak detection is useful for a fast exploratory scan. Quasi-peak and average measurements may then be required against the relevant limits. A spectrum analyser can be effective for debugging, provided its input protection, resolution bandwidth, detector behaviour and overload performance are understood. It should not be assumed to reproduce a compliant measuring receiver under every condition.

Begin with a wide scan while monitoring overload indicators. Reduce attenuation only when the input level is known to be safe. Measure each applicable line or LISN port separately, maintaining the required termination on ports not being measured. Record receiver attenuation, preselection, detector, bandwidth, dwell or measurement time and any correction factors.

Operate the EUT in modes likely to maximise emissions. For a rail auxiliary converter, that may require representative input voltage, output loading, switching state, communication traffic and peripheral activity. An idle processor and unloaded converter can produce a clean plot that says little about the installed product.

Diagnosing a failed rail conducted emissions plot

Narrow harmonic families usually point towards a periodic switching source. Their spacing can identify a converter fundamental, clock or modulation process. A broad rise may indicate fast edges, unstable control behaviour, arcing, receiver overload or ambient noise entering through the supply.

Change one variable at a time. Compare line and return measurements, then use current probes, near-field probes and oscilloscope measurements with suitable high-voltage and bandwidth capability. Temporary filters can help identify the coupling mode, but a development fix must eventually be assessed for temperature, voltage stress, saturation, leakage, insulation coordination and surge performance.

Do not optimise solely against one static cable position. A modification that gains several decibels because a harness has been pressed against the enclosure may disappear when the harness is restored to the documented layout. Margin should come from controlling the source and coupling path, not from exploiting an accidental geometry.

Once conducted noise has been reduced, consider whether the same common-mode current can drive the harness as an antenna. A follow-up radiated emissions investigation is often technically sensible, even though it is a separate measurement requiring antennas and the appropriate test environment.

Typical scenario

Consider an illustrative 110 V DC rail control unit containing an isolated converter, processor board and several external communication lines. The prototype passes functional tests, but an exploratory current-probe scan shows strong switching harmonics on the input harness.

The team must establish the applicable EN 50121 part and customer test plan, then choose a LISN rated for the supply voltage, steady-state current and inrush. The EUT is installed over the specified reference plane with controlled supply-line and return routing. Representative loads and communication activity are maintained throughout the scan.

If the engineers use a low-current LISN or omit the return-side arrangement required by the method, the network may saturate, overheat or present the wrong RF impedance. Connecting chassis to the plane with an undocumented braid could also divert common-mode current and produce false confidence.

Early measurements allow the team to separate differential-mode converter noise from common-mode harness current before committing to PCB layout, filter packaging and production tooling. EMC Hire can support this work through equipment hire, accessible pre-compliance testing, test-facility access and practical setup advice. Formal compliance testing can then be planned with a more stable design and a clearer technical evidence trail.

When to Hire EMC Equipment

Rail projects often create irregular test demand. A receiver, high-current LISN and suitable accessories may be needed intensively during converter development, then remain unused for months. Hiring avoids committing capital to equipment whose voltage, current or frequency capability may not suit the next programme.

A defined hire window also reduces ownership overheads associated with storage, servicing, firmware management and calibration. It can expand an internal laboratory during project peaks or provide a specific network, probe or receiver for fault investigation.

Equipment selection still needs engineering review. Supply characteristics, applicable limits, EUT current, inrush, return topology and receiver protection should be agreed before dispatch. EMC Hire uses test equipment with calibration traceable through an appropriate ISO/IEC 17025 accredited calibration provider where calibration is relevant to the measurement. Suitable traceable calibration supports repeatability, comparison with later formal testing and confidence in recorded engineering data.

Where immunity investigation is also planned, use equipment intended for the separate method. CDNs are associated with conducted RF immunity methods such as IEC 61000-4-6 where called up by the product standard. They do not replace a LISN for emissions. See the conducted immunity equipment guidance when defining that work.

Common EMC Testing Mistakes to Avoid

Treating the DC return as an ideal zero-volt node

The return conductor carries both wanted current and RF disturbance current. Bonding it to the reference plane without test-plan authority changes the common-mode circuit and can invalidate the comparison with formal testing.

Using an electrically unsuitable LISN

An underrated network may overheat or be damaged. A network with the wrong impedance characteristic produces a measurement that cannot be compared defensibly with the specified limit.

Allowing the harness to move between scans

Cable movement changes capacitance to the plane and coupling between lines. The resulting amplitude shift can be mistaken for an effective filter change.

Measuring an unrepresentative operating mode

Testing only at nominal voltage and light load may miss converter modes that occur during start-up, low input voltage, regenerative conditions or peak processing activity.

Using peak results as the final compliance judgement

Peak scanning is efficient for finding disturbances, but the applicable limit may require quasi-peak or average detection. Detector choice must follow the current standard and test plan.

Failing to preserve setup evidence

Without photographs, cable dimensions, software versions, loading details and instrument settings, another laboratory may be unable to reproduce the result. That weakens the technical file and makes later diagnosis slower.

Frequently Asked Questions (FAQs)

Should both positive and return lines be measured?

Measure the ports and conductors required by the applicable method. In floating DC systems, both lines can carry significant common-mode disturbance. Do not assume the return is RF quiet merely because it is labelled 0 V.

Can a spectrum analyser replace a measuring receiver?

It can be useful for debugging and some pre-compliance measurements if its bandwidths, detectors, dynamic range, input protection and correction factors are suitable. Formal measurements should use equipment and settings that satisfy the applicable test requirements.

How much margin should a pre-compliance result have?

There is no universal figure. Required margin depends on setup uncertainty, unit variation, operating modes, cable tolerances and differences between development and formal facilities. Marginal data should trigger investigation rather than an assumption that production units will behave identically.

Does passing conducted emissions prove rail EMC compliance?

No. Other emissions and immunity requirements may apply, along with product, safety and contractual obligations. Testing is one part of the evidence supporting the manufacturer's conformity assessment and technical file.

Can pre-compliance data support CE or UKCA self-certification?

Calibrated engineering data can support risk assessment, mitigation records and technical documentation where self-certification is legally and technically appropriate. The manufacturer remains responsible for determining the applicable legislation, standards, conformity route and Declaration of Conformity requirements. EMC Hire provides commercial EMC and self-certification testing support, but pre-compliance results alone do not prove conformity.

Plan the measurement before the hardware arrives

A useful rail conducted emissions programme starts with the port definition, supply topology, operating modes and test-plan requirements. Those details determine the LISN, receiver protection, cable layout and measurement sequence.

EMC Hire can help review the setup, provide suitable hire equipment, arrange on-site testing, support pre-compliance debugging or discuss formal compliance testing where appropriate. Test-facility space can also be booked for controlled investigation before a formal programme.

To discuss supply ratings, lines, returns, power ports and harness configuration, contact the EMC Hire engineering team on +44 (0)1462 817111 or email sales@emchire.co.uk. Bringing the current standard, customer test plan and EUT electrical details to that discussion will make equipment selection considerably faster.