EMC Magnetic Cores: Mechanism and Selection

A magnetic core fitted around a cable is one of the few electromagnetic compatibility measures that can be added to an existing product without changing the board. It produces no signal, amplifies nothing, and does only one thing: it removes high frequency energy from a path where that energy should not exist. Because it is simple, it is often specified loosely, and a core chosen by permeability rather than by impedance curve frequently does very little.

Understanding why requires looking at how the material behaves as frequency rises. The behaviour is not a simple filtering action, and the region of the impedance curve in which the component operates determines whether it absorbs the disturbance or merely reflects it. A core that is effective at one frequency can increase the problem at another.

Two Mechanisms in One Component

An EMC magnetic core is a ferrite, a ceramic crystal of iron, manganese, zinc and nickel oxides sintered at high temperature. At low frequency the material presents a high permeability and behaves as an inductor: the impedance is dominated by the inductive component, and the disturbance encountering it is reflected rather than dissipated. As frequency rises, domain rotation and eddy currents inside the material convert electromagnetic energy into heat, and the resistive component of the impedance takes over.

The component is therefore effective in two different ways in two different bands, and the transition between them is the region where it works best. In the resistive region the noise energy is converted to heat efficiently, which removes it from the circuit permanently. In the inductive region the reflection reduces the current that reaches the load but the energy remains in the system, and a high quality factor at that frequency can even produce resonance and amplification. This is why a core must be selected so that its resistive region covers the frequency of the disturbance.

Material Families

Ferrite materials divide broadly into two families, and the choice between them follows from the frequency of the problem. Manganese zinc ferrites have higher permeability and higher saturation flux density, which makes them effective at lower frequencies and in applications where a direct current bias is present. Nickel zinc ferrites have lower permeability but higher resistivity, which pushes their useful range to higher frequencies where eddy losses would otherwise dominate a manganese zinc part.

Within each family the supplier offers grades with different permeability values, and higher permeability is not automatically better. A high permeability material reaches its resistive region at a lower frequency and may be entirely inductive in the band of interest, while a lower permeability grade of the same size can present a higher impedance at the frequency that matters. Selection should compare impedance against frequency for the candidate parts at the intended operating conditions, not the permeability figure alone.

Reading an Impedance Curve

The specification that answers the design question is the impedance against frequency, usually presented as a family of curves for different grades and sizes, with the resistive and reactive components distinguished. The useful figure is the magnitude of the impedance at the frequency of the disturbance, and a secondary consideration is how much of that impedance is resistive, because the resistive part is what converts the energy to heat.

Two features of the curve deserve attention. The first is the frequency at which the peak impedance occurs, which is where the material transitions into its lossy region. The second is the shape of the curve above that frequency, because a core that is effective over a decade of frequency is more useful than one with a narrow peak. Where the disturbance covers a band rather than a single frequency, a grade with a broader, lower peak often performs better than one with a sharp maximum.

Where and How to Fit a Core

The effect of a core depends on which conductors pass through it. Fitting it around all the conductors of a cable, including the return, presents the impedance to the common mode current that circulates on the group, which is the usual application for reducing cable emissions. Fitting it around a single conductor presents the impedance to the differential current in that conductor, which is used when a specific line carries unwanted high frequency content.

Position matters as well as configuration. Placing the core close to the source of the disturbance reduces the length of cable that can radiate or receive before the impedance is applied, while placing it at the connector of the affected equipment addresses the current flowing on the outside of the cable shield. For a retrofit, a clamp-on core allows the effect to be assessed before a permanent solution is chosen.

Practical Limits

Three limitations determine whether a core will work in a given situation. The first is saturation: a high direct current through the conductor drives the material toward saturation, which reduces its permeability and therefore its impedance. Where the cable carries supply current, the grade and the size must be chosen with that bias in mind rather than from a small signal curve.

The second is the space available. A core that fits the cable and the enclosure has a limited cross section, and the impedance obtainable from a given material is bounded by its dimensions. Where the required attenuation cannot be achieved with a single core, several cores distributed along the cable perform better than one large core at a single point, because each applies its impedance to the current before it has been re-established by coupling.

The third is temperature. The loss mechanism that makes the core useful also makes it warm, and the material properties change with temperature. Where the component is fitted inside an enclosure with other heat sources, the impedance at the operating temperature is the relevant figure. The placement and thermal considerations that apply are the same as those in thermal management PCB design, and the cable level practice that often accompanies a core is described in PCB cable assembly and in EMI suppression principles.

Combining a Core With Other Measures

A core is one element of a suppression strategy and rarely solves a problem on its own. Where the disturbance is generated by a switching converter, the first step is to reduce the loop area that radiates it, using the layout guidance in DC-DC converter routing, and a core on the cable then addresses what remains. Where the problem is a ground potential difference between two pieces of equipment, the core can reduce the current that flows but the return path still has to be defined.

The order of measures matters for cost. Layout changes are free after a redesign but expensive after production, filtering components add cost per unit, and a core fitted to a cable is usually the least invasive retrofit. Starting with the measure that is cheapest to implement permanently, and using the core to confirm the diagnosis rather than to compensate for a design that radiates, produces a solution that holds when the cable is replaced or the enclosure is changed.

FAQ

Does a higher permeability core always attenuate more? No. Permeability describes low frequency behaviour. What matters is the impedance at the disturbance frequency, and a lower permeability grade can present a higher impedance in the band of interest.

Should the core go around all conductors or one? Around all conductors of a group to suppress common mode current, and around a single conductor to suppress differential mode content on that line. The configuration determines which current the impedance opposes.

Why did adding a core make the problem worse at another frequency? Because in its inductive region the component reflects energy rather than absorbing it, and a high quality factor can produce resonance that amplifies coupling. The grade must be chosen for the band being addressed.