Electromagnetic Compatibility (EMC)
Electromagnetic compatibility is the discipline of ensuring that electronic equipment functions satisfactorily in its electromagnetic environment without introducing intolerable electromagnetic disturbances to anything else in that environment. The definition cuts two ways at once: a compliant product must neither emit excessive interference nor be unduly disrupted by the interference around it. These two halves are known as emissions and immunity, and a system achieves EMC only when both are held within bounds across the frequency range in which it operates.
The spectrum that modern electronics must share is crowded. Switching power supplies, high-speed digital buses, motor drives, and radio transmitters all generate energy that can couple into nearby circuits, while the same circuits must tolerate electrostatic discharge, mains transients, and ambient radio-frequency fields. Because regulatory approval is mandatory in most markets before a product may be sold, EMC is not an optional refinement but a gating requirement that shapes architecture, layout, and enclosure design from the outset. This category examines how interference arises, how it propagates, how engineers measure it, and how they control it, with the emphasis falling where signal integrity places it: on the interconnect, meaning the traces, planes, connectors, and cables whose geometry decides how much energy escapes a design and how much finds its way in. The wider discipline, including its standards landscape, test facilities, sector-specific regimes, and specialized operating environments, is developed at length in the guide's Electromagnetic Compatibility and Interference category.
Articles in This Category
The Source-Path-Victim Model
Almost every EMC problem can be described with three elements: a source of electromagnetic energy, a coupling path that carries it, and a victim that is disturbed by it. The same device is usually both a source and a victim, which is why the discipline treats emissions and immunity as two views of one underlying physics. Interrupting any one of the three elements resolves the problem, so an engineer who cannot quiet a noisy source may instead break the path or harden the victim.
Coupling paths fall into two broad families. Conducted coupling carries interference along physical conductors such as power cords, signal cables, and ground connections, and dominates at lower frequencies, typically from about 150 kHz to 30 MHz in emissions testing. Radiated coupling carries energy through space as electromagnetic fields and dominates at higher frequencies, generally from 30 MHz upward. Within these families, coupling occurs through electric fields (capacitive coupling between conductors at different potentials), magnetic fields (inductive coupling between current loops), and direct radiation from any conductor that is electrically long relative to the wavelength. A cable attached to a circuit board is the most common offender, because it readily acts as both an antenna that radiates board noise and a conduit that injects external interference.
Current on conductors divides further into differential mode and common mode, and that distinction governs most practical work. Differential-mode current flows out on one conductor and returns on its partner, so the two contributions largely cancel in the far field and the residual radiation depends on the loop area enclosed between them. Common-mode current flows in the same direction on every conductor of a cable and returns through the chassis, the earth, or stray capacitance to the surroundings; nothing cancels it, so it radiates far more efficiently. Common-mode currents in the microampere range on a meter of attached cable are enough to breach a residential radiated-emission limit, which is why cables, connectors, and shield terminations attract attention out of all proportion to their apparent simplicity.
Emissions and Immunity
The two sides of EMC are evaluated against separate limits and with different test methods, but they share the same field physics and respond to the same design measures.
Emissions
Emissions testing verifies that a product does not pollute the spectrum beyond defined limits. Conducted emissions are measured at the power and signal ports through a line impedance stabilization network, a 50 ohm, 50 µH artificial mains network that presents a defined and repeatable impedance to the equipment and routes the disturbance voltage to the receiver. Radiated emissions are measured at a calibrated distance on an open-area test site or in a semi-anechoic chamber, with the antenna scanned in height, the turntable rotated, and both polarizations recorded so that the worst case is captured. Limits are expressed as voltages or field strengths in decibel form, referenced to one microvolt or one microvolt per meter.
Under CISPR 32, the emission standard for multimedia equipment, the Class B radiated limit for the residential environment is 40 dB(µV/m) from 30 to 230 MHz and 47 dB(µV/m) from 230 MHz to 1 GHz when measured at three meters, equivalent to 30 and 37 dB(µV/m) at ten meters. The Class B quasi-peak limit at the mains port falls from 66 to 56 dB(µV) across 150 to 500 kHz, holds at 56 dB(µV) up to 5 MHz, and rises to 60 dB(µV) from 5 to 30 MHz, with average-detector limits ten decibels lower throughout. Class A limits, which apply to equipment intended for commercial and industrial environments, are correspondingly relaxed. Equipment containing high-frequency internal clocks is measured above 1 GHz as well, where peak and average detectors replace the quasi-peak. In the United States the FCC enforces an analogous scheme for unintentional radiators under Title 47, Part 15, again separating the stricter Class B for residential use from Class A for commercial and industrial products.
Immunity
Immunity, or susceptibility, testing confirms that a product continues to operate when subjected to representative disturbances. The basic test methods are standardized in the IEC 61000-4 series and applied at graded severity levels chosen to match the intended environment.
- Electrostatic discharge (IEC 61000-4-2). A charged simulator delivers discharges to accessible surfaces and to nearby coupling planes. Severity levels span ±2 kV to ±8 kV for contact discharge and ±2 kV to ±15 kV for air discharge. The current rises in roughly a nanosecond, so the event is a broadband problem rather than merely a high-voltage one.
- Radiated radio-frequency field immunity (IEC 61000-4-3). The equipment is illuminated in an anechoic chamber by an amplitude-modulated field swept from 80 MHz upward, commonly at 3 V/m for ordinary environments and 10 V/m for industrial ones. The range extends into the gigahertz region to represent the mobile and wireless transmitters that operate near modern equipment.
- Electrical fast transient, or burst (IEC 61000-4-4). Bursts of pulses with a rise time of about five nanoseconds are coupled onto power and signal lines, reproducing the interference generated when inductive loads and relay contacts are switched.
- Surge (IEC 61000-4-5). A combination wave, specified as a 1.2/50 µs open-circuit voltage and an 8/20 µs short-circuit current, simulates the energy delivered by nearby lightning and by power-system switching. Levels are applied line to line and line to earth.
- Conducted radio-frequency immunity (IEC 61000-4-6). Disturbances are injected onto cables from 150 kHz to 80 MHz, the range in which a cable is a poor antenna but still an effective conduit.
Further parts of the series cover power-frequency magnetic fields (IEC 61000-4-8) and voltage dips, short interruptions, and voltage variations (IEC 61000-4-11). Results are judged against defined performance criteria rather than a simple pass or fail: one criterion requires normal operation throughout the disturbance, another allows temporary degradation that recovers without intervention, and a third allows a loss of function that the operator or the system may reset. Because those criteria are generic, the product specification must state what counts as acceptable behavior for each function, and writing that specification honestly is part of the engineering work.
Measurement Practice
Limits mean nothing without a specified way to measure them, so EMC rests on basic standards that define the instruments themselves. The CISPR 16 series specifies measuring receivers, antennas, artificial mains networks, test sites, and measurement methods, and that specification is what allows a laboratory in one country to reproduce another laboratory's result.
Three detectors dominate the numbers. The peak detector captures the highest instantaneous amplitude and settles quickly, which makes it the natural choice for scanning; it never reads below the others, so a spectrum that passes on peak passes outright. The quasi-peak detector applies standardized charge and discharge time constants that weight a disturbance according to how often it repeats, and most limits below 1 GHz are written against it. The average detector reports mean amplitude and therefore discriminates against impulsive noise, exposing narrowband carriers that a broadband spectrum might otherwise mask. Resolution bandwidth is part of the specification rather than an operator's choice: 9 kHz from 150 kHz to 30 MHz, 120 kHz from 30 MHz to 1 GHz, and 1 MHz above 1 GHz. Comparing a measurement made in the wrong bandwidth with a published limit is a common and expensive mistake.
Full compliance testing is slow and costly, so most development relies on pre-compliance work. Near-field probes locate the radiating structure on a board, a clamp-on current probe measures common-mode current on a cable, and a bench receiver or spectrum analyzer with a line impedance stabilization network yields a conducted-emission profile that tracks the accredited result closely enough to guide design. Absolute accuracy on the bench is limited, so engineers use these measurements comparatively. The useful question is not the exact decibel figure but whether a change improved the emission, by how much, and at which frequencies.
Control Techniques
The practical toolkit for achieving EMC reduces to a small set of complementary techniques, applied together rather than in isolation.
- Source suppression. Reducing the disturbance where it originates is the cheapest control of all: slowing driver edge rates to the slowest that timing allows, damping clock lines with series resistors, snubbing the switching node of a converter, and choosing switching frequencies whose harmonics avoid sensitive bands.
- Grounding and return paths. A continuous, low-impedance reference and a well-defined return path keep signal currents on tight loops, which limits both the radiation a circuit produces and the noise it picks up. A large share of EMC failures trace back to a compromised return path, most often a trace crossing a plane split or changing reference layers without a nearby return via.
- Shielding. A conductive enclosure reflects and absorbs fields, attenuating both emissions and incoming interference. Its effectiveness is set by the integrity of the enclosure rather than by the thickness of the metal: seams, ventilation slots, display windows, and cable penetrations act as apertures whose leakage grows as their longest dimension approaches a half wavelength, and these usually govern the real-world result.
- Filtering. Filters on power and signal lines divert high-frequency energy back toward its source before it reaches a cable or a sensitive input. Common-mode chokes, feed-through capacitors, and ferrite beads target the common-mode currents that drive most cable radiation, and they work only when placed at the boundary they defend, with a short, low-inductance connection to the chassis at the point where the cable enters.
- Layout and partitioning. Controlling trace geometry, minimizing loop area, separating noisy and quiet domains, keeping high-speed clocks and switching nodes away from connectors and enclosure openings, and reserving a quiet perimeter for input and output circuitry prevent interference at the board level, where it is least expensive to address.
Each measure is most economical when adopted early. Interference designed out during schematic capture and stackup planning costs little, whereas the same problem discovered during compliance testing may force enclosure changes, added filtering, a board respin, or all three, with the schedule penalty of a second visit to the test house.
EMC and Signal Integrity
EMC and signal integrity examine the same currents from opposite sides. Signal integrity asks whether the receiver recovers the transmitted waveform; EMC asks what that same waveform does to everything around it. Both therefore converge on two physical quantities above all others: the loop area of the signal-return path and the rise time of the switching edge.
Rise time, not clock frequency, sets the bandwidth of an emission. A digital edge carries harmonic energy far beyond its fundamental, and a faster edge extends that content upward, so an unnecessarily fast driver adds measurable emission while buying no useful timing margin. Return-path discontinuities show the same dual character: a trace that crosses a plane split forces its return current onto a detour, which distorts the signal and enlarges the radiating loop at the same time. Measures that help one discipline usually help the other, which is why EMC belongs with the stackup and routing decisions rather than after them.
The exceptions are worth knowing, because they reveal what the limits actually measure. Spread-spectrum clocking modulates a clock so that its energy spreads across a wider band; the amplitude recorded within the receiver's resolution bandwidth falls, yet the total energy emitted does not change, and the added jitter must still fit within the link's timing budget. Similarly, a shield connected to the board reference at one end only may satisfy a low-frequency grounding argument while leaving the cable free to radiate at high frequency. EMC rewards measures that reduce the offending current, and merely rearranges the problem when they do not.
Standards and Compliance
EMC requirements are codified by regional regulators and international bodies. The International Special Committee on Radio Interference, universally known by its French initials CISPR, operates as a special committee of the IEC and develops the emission and immunity standards adopted across much of the world. Its publications are transposed into the harmonized EN standards that underpin the European Union's EMC Directive, 2014/30/EU. A manufacturer whose apparatus meets a harmonized standard cited in the Official Journal of the European Union gains a presumption of conformity, draws up a declaration of conformity, and affixes the CE marking. For multimedia equipment the pairing is EN 55032 for emissions and EN 55035 for immunity, which together replaced the earlier separate standards for information technology equipment and broadcast receivers.
In the United States, the FCC governs radio-frequency devices under Title 47, Part 15, separating unintentional radiators such as digital devices, covered by Subpart B, from intentional radiators such as transmitters, covered by Subpart C. Since 2017 the supplier's declaration of conformity has consolidated the earlier verification and declaration-of-conformity procedures for unintentional radiators, while intentional radiators continue to require certification through an accredited body.
Sector-specific regimes layer further requirements on top of the general ones, because both the electromagnetic environment and the cost of failure differ sharply by industry.
- Automotive. CISPR 25 governs component and vehicle emissions measured in an absorber-lined shielded enclosure, and the ISO 11452 series defines the matching component immunity methods. Test field strengths run far above commercial levels, because a module may sit within a meter of a transmitting antenna.
- Aerospace. RTCA DO-160 defines environmental and EMC qualification for airborne equipment, with Section 20 covering radio-frequency susceptibility and Section 21 covering radio-frequency emissions, alongside sections addressing lightning-induced transients and electrostatic discharge.
- Defense. MIL-STD-461 specifies emission and susceptibility requirements by test designation, such as CE102 for conducted emissions on power leads and RS103 for radiated susceptibility, with the applicable set and limits tailored to the platform.
- Medical. IEC 60601-1-2, the EMC collateral standard within the IEC 60601 family, frames its requirements around basic safety and essential performance, and accounts for the wireless devices now carried routinely into clinical spaces.
Demonstrating conformity typically combines design analysis, pre-compliance measurement during development, and formal testing at an accredited laboratory before a product reaches the market. Documentation carries as much weight as measurement: the technical file, the test reports, and the declaration of conformity are what a market-surveillance authority examines, and they must remain valid for the product as shipped, including its cables, accessories, and configuration options.
Conclusion
Treated as an early design constraint rather than a final hurdle, EMC becomes a matter of disciplined engineering rather than guesswork. The physics is compact: a source, a path, and a victim, with common-mode current on cables and compromised return paths accounting for the majority of real failures. The remedies are equally few, namely suppression at the source, sound grounding and return paths, shielding, filtering, and careful layout, and each costs little when applied at the schematic and stackup stage. The topics in this category develop the threads in depth: the conducted and radiated mechanisms by which energy escapes, the immunity a product must demonstrate, and the design strategies that hold both within their limits.
Related Topics
The pages below approach the same physics from the interconnect side within signal integrity, or from the compliance side within the guide's dedicated electromagnetic compatibility category.