Common Mode Choke: Design Rules and Process Limits

A common mode choke has two windings on one core, arranged so that the current flowing out on one conductor and back on the other produces no net flux. It presents a high impedance to the current that flows on both conductors together, and almost none to the current the circuit actually uses.

How the Selectivity Works

The two windings are wound so that the differential current magnetises the core in opposite directions and cancels. The choke therefore carries the full load current without saturating, provided the windings are well matched.

The common mode current flows in the same direction through both windings, so the fluxes add and the choke presents the full inductance of the winding. That asymmetry is the whole function of the component.

The cancellation is not perfect. The two windings differ slightly in their position on the core and in their leakage, so a small differential inductance remains. That residual is a leakance inductance and can be a useful part of a filter.

impedance and Frequency

The impedance of the choke is specified at a frequency, usually a hundred kilohertz or a megahertz, and it is the magnitude of the complex impedance rather than the inductance alone. Two chokes with the same inductance can differ by a factor of two in impedance at the frequency that matters.

The impedance peaks at the self resonant frequency of the winding and falls above it, where the winding capacitance takes over and the choke begins to look capacitive. Above that frequency the choke can actually worsen the common mode noise.

A good selection starts from the frequency of the emission that has to be suppressed and looks for a choke with a high impedance there. Choosing by inductance alone is a common and avoidable mistake.

saturation current and Rated Current

The rated current is the current the windings can carry without overheating, and the saturation current is the differential current at which the core begins to saturate. The two are different figures and both have to be respected.

As the core approaches saturation, the impedance falls and the choke stops working exactly when the load current is largest. A choke that is correct in every other respect and undersized in saturation current will fail an emission test at full load only.

Where the differential current is large, a choke with an air gap or a distributed gap material is used. The gap reduces the permeability and therefore the impedance, so the design trades impedance for current capability.

leakage inductance and Its Effect

leakage inductance is the differential inductance that remains because the coupling between the windings is imperfect. It appears in series with the load current and contributes to the differential mode attenuation of the filter.

A large leakage inductance is helpful in a filter, because it removes the need for a separate differential inductor. It also causes a voltage drop at the load current and can resonate with the filter capacitors, which is why the damping has to be checked.

Where the leakage inductance is quoted, it is usually given as a small fraction of the common mode inductance. Where it is not quoted, it can be measured with an impedance analyser between the two windings with the other ends shorted.

winding capacitance and Self Resonance

winding capacitance is the capacitance between the two windings and between each winding and the core. It provides a path that bypasses the inductance, and it sets the self resonant frequency of the component.

A choke with a high inductance and a large number of turns has a large winding capacitance and a low self resonance, which limits how high in frequency it can work. A two stage filter with a smaller choke in each stage usually outperforms one large choke.

The construction influences the capacitance. A toroidal winding with separated sections has a lower capacitance than a single layer wound over itself, and a bobbin with physically separated windings is lower still.

Choosing the Core Material

Ferrite is the common material for high frequency filters, with manganese zinc types for the lower frequencies and nickel zinc types for the higher ones. The choice follows the frequency band where the attenuation is needed.

Nanocrystalline and amorphous cores give a high permeability in a small volume and are used where space is limited or where a high impedance is needed at a low frequency. The cost is higher and the saturation behaviour differs.

Powdered iron cores have a lower permeability and a much higher saturation flux, so they are used where a large differential current has to be carried. They provide less impedance, which limits their use to the lower part of the spectrum.

Layout and Installation

Place the choke close to the connector so that the noise it removes cannot couple to anything else. A choke placed in the middle of the board filters the noise after it has already been radiated by the traces ahead of it.

Keep the input and output sides of the choke physically separated. Capacitive coupling between the two sides provides a path that bypasses the choke, and the benefit is lost even though the schematic is correct.

Route the two conductors of the pair together through the choke and keep them together afterwards. Separating them creates a loop that radiates, and the common mode noise that was suppressed upstream is regenerated. The wider practice for filter layout is described in our guide to EMI suppression design principles.

Two Stage Filtering

A single choke rarely meets a demanding limit on its own. Two stages, each with a smaller inductance and its own capacitor, give more attenuation than one large choke, because the total winding capacitance is shared and the self resonance of each stage is higher.

The capacitor between the stages sets the impedance that the second choke works into, and its value is chosen from the frequency of the emission. Damping that capacitor with a series resistor stops the two stages interacting and producing a peak that leaves the filter worse than a single stage.

Verification and Faults

Measure the common mode impedance with an impedance analyser to confirm the datasheet value in the actual assembly, because the surrounding copper and the mounting can change it. The differential impedance measured at the same time gives the leakage inductance.

Check the saturation behaviour at the maximum load current and at the maximum temperature. The impedance at a low signal level can be perfect while the impedance at full load is a fraction of it.

A filter that passes on the bench and fails in the enclosure often has a coupling path around the choke rather than a choke problem. The release checks that keep such a filter effective are collected in our PCB design release checklist, and the assembly points are listed in judging PCB quality.

FAQ

Should I choose a choke by inductance or by impedance? By impedance at the frequency of the noise. Two chokes with the same inductance can differ greatly in the frequency band that matters.

What is leakage inductance? The differential inductance left because the windings are not perfectly coupled. It helps the differential mode attenuation and adds a small series impedance to the load current.

Why does my choke stop working at full load? It is saturating. Check the saturation current rather than the rated current, and consider a gapped core.