How Common Mode Chokes Work

Coupled inductor characteristics directly affect conducted emissions and radiated emissions on power and signal lines.

EMI filters frequently employ coupled inductors to control common mode noise. But how do coupled inductors work? And what do they actually accomplish?

The most common type of coupled inductor is the common mode choke, a passive component that provides significant inductance for common mode signals while providing minimal inductance for differential mode signals. Common mode signals get blocked, while differential mode signals pass through unimpeded.

Less common are differential mode coupled inductors. For low-current applications these devices can provide the same inductance as two separate, uncoupled inductors, but in a smaller volume.

In this article we will focus primarily on common mode chokes, but the principles presented apply equally to differential mode coupled inductors.

Coupled versus Uncoupled

Coupled inductors differ from standard inductors in several ways. A standard inductor is a two-terminal device that has a single conductor wound into a coil, usually around a magnetically permeable core. A standard inductor is shown schematically on the left in the figure below. In contrast, a coupled inductor has two or more conductors wound on a single core. Schematically, common mode and differential mode inductors are as shown in the middle and right in the figure below.

A coupled inductor is most often a four-terminal device, but common mode chokes may have six terminals for 3-phase applications, or more for multi-conductor applications.

Coupled inductors provide high inductance in a small volume. Common mode chokes obtain high inductance by using a high permeability core. Their inductance is proportional to the number of turns squared, on each winding. Differential mode inductors achieve high inductance because their inductance is proportional to the square all winding turns on the core.

Electrical characteristics of coupled inductors vary significantly over the frequencies of interest for EMI filtering. The useful frequency range of most power chokes is from a few kilohertz to a few tens of megahertz. Signal line chokes work up to somewhat higher frequencies, topping out around 100 MHz.

Ideal Model versus Non-Ideal Model

A common mode choke may be adequately modeled as a sub-circuit comprised of several passive lumped circuit elements. The schematic below shows one model that accounts for the frequency-variable behavior of coupled inductors.

Not only does this model provide for the common mode inductance of the choke, but it also takes into account the effects three important parasitic elements:

a. Winding resistance

b. Interwinding capacitance

c. Leakage inductance

The paragraphs below discuss each in more detail.

Inductance

Inductance is a function of magnetic core permeability and the number of wire turns on the core. Inductance is proportional to the number of turns squared. Core permeability varies with material, temperature, dc bias, and frequency. Acurate inductor modeling requires consideration of each of these properties. However, for many applications, leakage inductance, interwinding capacitance and winding resistance are the dominant properties and are sufficient for predicting EMI filter performance.

Coupling Factor

The coupling factor of a common mode choke is a measure of how completely the magnetic flux from one winding couples to the other winding(s). For example, a coupling factor of 0.95 means 95% of the magnetic flux is coupled and 5% is not.

Magnetic flux that couples all windings on the core equates to mutual inductance between the windings. The flux that is not coupled, directly relates to leakage inductance. Leakage inductance is a parasitic element present in all physical inductors that can have a profound effect on filter performance.

To illustrate, consider a common mode choke that has a hypothetical coupling factor of 100%. Both windings are perfectly coupled. Current through either winding induces an equal current in the other winding. If the current through both windings is equal, the magnetic flux in the core is zero. All of the core flux generated by current in one lead cancels the core flux generated by current in the other lead.

However, if the coupling factor is less than 100%, say 0.95, then 5% of the magnetic flux is not coupled. Current in one winding does not induce equal current in the other winding, so the magnetic flux in the core is not zero, and some of the magnetic flux is outside the common mode choke.

The uncoupled flux has three effects:

1. If the magnetic flux in the core is strong enough, the core will begin to saturate. Partial saturation reduces the inductance of the device, making the choke less effective.

2. Flux that is not in the core is leakage flux. For a common mode choke, leakage flux introduces differential mode inductance. (Likewise, for a differential mode coupled inductor, leakage flux introduces common mode inductance.) This differential inductance interacts with other filter components, and while it does help provide additional differential mode filtering, it also introduces new resonances that have the potential to amplify circuit noise at the newly created resonant frequencies.

3. Stray magnetic fields may couple to nearby circuits and may be emitted from the device containing the common mode choke. For applications containing sensitive magnetic components, these stray magnetic fields may be problematic.

Interwinding Capacitance

Each turn of wire on a magnetic core has a small amount of capacitance to adjacent turns and to the wire and the core. This distributed capacitance for common mode chokes is typically between 10 and 50 pF, depending on how many turns of wire are used and construction and physical size of the inductor.

Although the inter-winding capacitance is distributed across the windings of the inductor, a lumped element capacitance from input terminal to output terminal is usually sufficiently accurate for most filter analyses.

The inter-winding capacitance, along with the inductance, sets the inductor self-resonant frequency (SRF), as shown in the red line in the graph below. At frequencies below the SRF the common mode choke behaves as an inductor, with impedance increasing in proportion to increasing frequency. Above the SRF the choke behaves as a series capacitor, and impedance decreases in proportion to increasing frequency.

A second SRF is present in real common mode chokes. It is caused by the parallel resonance of the leakage inductance and the interwinding capacitance. This secondary resonance is graphed above by the green line.

Winding Resistance

Resistance of the inductor windings has a subtle, but important effect on filter performance. Winding resistance is beneficial because it adds to the series impedance of the inductor and it provides some damping for the filter, especially at high frequencies. Since the inductor will resonant with filter capacitance and capacitive elements connected to the filter through cabling or other circuits, the damping effect of the winding resistance is beneficial. Winding resistance reduces the amplitude of signals amplified by filter component resonances.

The downside of winding resistance is thermal. Current flowing through the wires produces heat. Because the length of wire used to construct the inductor may several meters, and the power dissipation is concentrated in a relative small area, temperature rise may be significant.

Winding resistance also causes voltage drop between the input and output sides of the inductor, which may affect circuit operation if excessive.

Ac resistance of the windings is important. As frequency increases, resistance of the windings increases due to skin effect. For example, an inductor winding that has 25 mohm resistance at dc may have more than 100 mohm resistance at 1 MHz. The increase in resistance for a given increase in frequency is a function of wire gauge. Relative to dc resistance, ac resistance increases more for larger wire gauges.

Winding Method

The figure below, courtesy of Wurth Electronik shows two different ways a coupled inductor can be wound. Schematically the inductors are the same, but their high frequency characteristic differ somewhat.

Two-in-Hand

The inductor on the left is wound using a method called two-in-hand. Both wires are wound together through the core. This method provides very closely matched inductance for both wires. For a common mode choke this equates to better circuit balance and lower leakage inductance. Two-in-hand also provides lower inter-winding capacitance because the distance between successive turns is greater than for bank-wound inductors, discussed below.

Winding a coupled inductor two-in-hand, provides superior performance but at higher cost, since the inductor must be wound by hand. Due to its inherently higher cost, two-in-hand inductors are usually used in applications where price is less important than performance.

Three phase common mode chokes can also be wound this way by wrapping the three phase leads three-in-hand through the core as a triplet.

Bank Wound

The inductor on the right in the figure above is bank wound. The wires are wrapped an equal number of turns on opposite sides of the core. Bank wound inductors may be fabricated by machine, and are therefore less expensive than inductors wound two-in-hand.

Electrically, bank wound inductors have higher leakage inductance and higher inter-winding capacitance than inductors wound two-in-hand. However, for many applications bank wound inductors are adequate.

Because of their lower cost and mechanized production, bank wound coupled inductors are more readily available. Off-the-shelf common mode inductors are almost always bank wound.

Calculate conducted emissions produced by electronic circuits,