Using EMI Filters for Noise Suppression

Controlling electromagnetic interference (EMI) while ensuring an electronic system meets its functional requirements is a critical and complex process. Regulatory bodies worldwide are imposing strict limits on EMI from electronic systems. Due to the growing number of electronic devices in industrial, commercial, and consumer markets, electromagnetic compatibility (EMC) design has become an essential part of product development. This means that various electronic systems must be able to coexist and operate normally in the same environment without causing interference or being susceptible to it, placing greater demands on engineering skills.

Noise Emission and Immunity

Preventing a device from emitting noise is known as "emission suppression." Conversely, protecting a device from the effects of external noise is known as providing "noise immunity." Immunity refers to a device's ability to resist noise without malfunctioning, degrading in performance, or sustaining damage. While the term "electromagnetic susceptibility" (EMS) also describes a device's sensitivity to noise, "immunity" is often used as the antonym of "emission."

"Electromagnetic compatibility" (EMC) ensures that different electronic devices and systems can coexist and function correctly in the same electromagnetic environment without interfering with each other. It involves both ensuring a device has immunity to external electromagnetic disturbances and that it does not emit excessive electromagnetic radiation that could interfere with other equipment. "Electromagnetic interference" (EMI) refers to the adverse effects caused by the interaction of electromagnetic fields within an electronic system, which can include device malfunctions, data transmission errors, and signal quality degradation.

Noise Propagation Paths

After originating from a source, noise travels through many complex paths. It can be conducted through conductors or propagate in the form of radiation. When this noise reaches a piece of equipment, the equipment is affected by it.

Principles of Noise Suppression

To effectively suppress noise, it is essential to understand its source and propagation path. If the initial analysis is incorrect, it becomes impossible to determine whether a suppression technique has failed or if it was applied to the wrong source. The fundamental principle of noise suppression is to use EMI filters for conducted noise and shielding for radiated noise.

Methods for Suppressing Noise with EMI Filters

There are three primary methods for suppressing noise using EMI filters:

1. Differentiating between signals and noise based on frequency.
2. Utilizing the different conduction modes between signals and noise.
3. Using non-linear resistors (varistors) to suppress high-voltage surges.

Typical Filters

Filters used for EMI suppression are classified into four main types:

• Low-Pass Filter (LPF): This filter passes signals below a specified frequency and attenuates signals above it.
• High-Pass Filter (HPF): This filter passes signals above a specified frequency and attenuates signals below it.
• Band-Pass Filter (BPF): This filter passes signals only within a specific frequency range.
• Band-Stop Filter (BSF) or Band-Reject Filter (BRF): This filter rejects signals within a specific frequency range.

Most noise emitted from electronic devices has a higher frequency than the circuit's operational signals. Consequently, low-pass filters are commonly used as EMI filters because they pass low-frequency signals while attenuating high-frequency noise.

Insertion Loss

The noise suppression performance of an EMI filter is measured by its insertion loss, according to the method specified in MIL-STD-220A. Insertion loss is determined by measuring the voltage at the load with and without the filter inserted. It is expressed in decibels (dB). For example, an insertion loss of 20 dB means the noise voltage has been reduced to one-tenth of its original level.

This measurement is performed in a system with 50Ω input and output impedances. However, since the impedance of actual circuits is not always 50Ω, the filter's performance will differ from what is measured in a 50Ω system.

Low-Pass Filters

The most basic low-pass filters consist of the following two components:

1. A capacitor installed between the signal line and the ground (GND) line. As frequency increases, the capacitor's impedance decreases, shunting high-frequency noise to ground.
2. An inductor (coil) installed in series with the signal line. As frequency increases, the inductor's impedance increases, blocking noise from flowing down the signal line.

LPF: Component Values vs. Insertion Loss

In the frequency band where EMI problems occur, the insertion loss of a filter increases by 20 dB for every tenfold increase in frequency. Similarly, if the component value (capacitance or inductance) is increased tenfold, the insertion loss will increase by 20 dB. To achieve greater insertion loss, combination filters can be used.

Filter Input and Output Impedance

Insertion loss is measured under the condition of 50Ω input and output impedance. However, the impedance of real-world circuits is rarely 50Ω. A filter's actual effectiveness depends on the impedance of the circuit it is placed in. As a general rule, capacitors are more effective at suppressing noise in high-impedance circuits, while inductors are more effective in low-impedance circuits.

Capacitor Insertion Impedance Characteristics

This article describes the necessity and performance of capacitor-based EMI filters. An ideal capacitor's insertion loss increases indefinitely with frequency. However, a real capacitor's insertion loss begins to decrease after reaching a certain frequency, known as the self-resonant frequency (SRF).

The decrease in insertion loss above the SRF is due to the capacitor's equivalent series inductance (ESL) from its leads and internal construction, which forms a series resonant circuit with its capacitance. At frequencies above the SRF, the component behaves as an inductor, and its impedance starts to rise again, reducing its effectiveness as a bypass path to ground.

The Effect of ESL

For a given ESL, increasing or decreasing the capacitance value will not change the insertion loss at frequencies above the SRF. Therefore, to achieve better noise suppression at high frequencies beyond the SRF, it is necessary to select a capacitor with a higher SRF, which means choosing one with a lower ESL.

Insertion Loss of Typical Capacitors

For leaded capacitors, the insertion loss is measured with the leads trimmed to 1 mm.

Three-Terminal Capacitor Structure

Leaded two-terminal capacitors have a relatively high ESL because their leads act as inductors.

By adopting a three-terminal structure, the ESL is significantly reduced, resulting in better insertion loss characteristics compared to two-terminal capacitors.

Typical Three-Terminal Chip Capacitor

Electrodes are present on each dielectric sheet. The input and output terminals are at opposite ends and are connected through these electrodes, allowing the signal current to pass through the capacitor. Grounding the capacitor on both sides reduces the ESL on the ground path. This structure achieves very low ESL and provides a much higher self-resonant frequency.

Feed-through Capacitors

Another capacitor structure features a ground electrode encircling the dielectric, with the signal terminal passing through its center. A feed-through capacitor is used by creating an opening in a shielded enclosure and soldering the ground electrode directly to the enclosure (or board). Because this type of capacitor has virtually no ESL on either the ground or signal path, it can provide nearly ideal insertion loss characteristics.

The Effect of Equivalent Series Resistance (ESR)

The second factor causing non-ideal capacitor behavior is equivalent series resistance (ESR). ESR, arising from the electrodes and dielectric material, degrades insertion loss at resonance. Ceramic capacitors have very low ESR, whereas aluminum electrolytic capacitors have higher ESR.

The Effect of Non-Ideal Inductors

Just as capacitors are non-ideal due to ESL and ESR, inductors also exhibit non-ideal behavior. An inductor's insertion loss is not ideal. As frequency increases, the impedance of the inductor's parasitic capacitance decreases, causing the inductor's total impedance to drop when frequency exceeds its self-resonant frequency. As a result, high-frequency noise can bypass the inductor.

Ferrite Bead Inductors

Leaded ferrite bead inductors, commonly used as inductive EMI filters, have a simple structure where the lead passes through a ferrite core. This design minimizes stray capacitance. Indicating excellent performance with a self-resonant frequency of 1 GHz or higher due to low parasitic capacitance.

Understanding Ferrite Bead Inductors

In addition to low parasitic capacitance, ferrite beads have another important characteristic: at high frequencies, they cease to act as inductors and behave more like resistors, dissipating noise energy as heat.

SMD Ferrite Bead Inductor Structure

Electrodes are printed on ferrite sheets, forming via-hole electrodes. These sheets are then stacked to create a multilayer chip inductor. When higher impedance is required, the electrodes on each sheet are connected through vias to form a winding-style chip inductor.

Unlike general-purpose inductors, both of these chip types are designed to have minimal parasitic capacitance.

Impedance Characteristics of SMD Ferrite Beads

The impedance of a ferrite bead varies depending on its material and internal structure. When selecting a ferrite bead, it is necessary to consider both the impedance in the target noise band and the slope of the impedance curve.

Differential-Mode and Common-Mode Noise

Noise can be classified into two types based on its conduction mode:

1. Differential-mode noise travels in opposite directions on the signal (or VCC) line and the GND line. This noise can be suppressed by placing a filter on the signal or power line, as previously discussed.
2. Common-mode noise travels in the same direction on all lines. For example, on an AC power cord, the noise flows in the same direction on both wires. For a signal cable, noise travels in the same direction on all conductors within the cable.

Therefore, to suppress this type of noise, an EMI filter must be installed on all lines conducting the noise.

Noise Suppression with Common-Mode Chokes (1)

Common-mode chokes are used to suppress common-mode noise. This type of coil is made by winding signal or power lines together on a single ferrite core.

A common-mode choke acts as an inductor for common-mode currents because their magnetic fluxes add up within the ferrite core. This creates a high impedance to common-mode currents, making the choke more effective at suppressing common-mode noise than using multiple separate inductors.

Noise Suppression with Common-Mode Chokes (2)

For differential-mode currents, the magnetic fluxes cancel each other out within the ferrite core, resulting in very low impedance. This also minimizes issues with core saturation. Common-mode chokes are well-suited for suppressing common-mode noise on lines with large DC or low-frequency AC currents, such as AC/DC power lines. Because they do not significantly affect differential signals, they are also ideal for signal lines where waveform distortion is a concern, such as video signal lines.

The actual characteristics also include some differential-mode impedance, which must be considered if signal waveform integrity is critical when using a common-mode choke in a circuit.

Surge Suppression and Circuit Protection with Varistors

Varistors are used to protect circuits from high-voltage surges.

When a high-voltage surge hits a circuit, it can cause catastrophic damage. A capacitor can be installed on the signal line, but it cannot suppress a high-voltage surge effectively.

Therefore, when surge protection is required, a varistor is used as a voltage protection device. When an applied surge exceeds a specified voltage (the varistor voltage), the varistor clamps the voltage to protect the circuit.

Varistor Characteristics

When the applied voltage does not exceed the varistor's clamping voltage, the varistor acts like a capacitor. However, when the surge voltage exceeds this threshold, the varistor's impedance drops sharply. Since the circuit's input voltage is dependent on the varistor's internal resistance and the line impedance, this sharp decrease in impedance clamps the surge voltage.

A key selection criterion for a varistor is its peak pulse current rating. This is the maximum current a varistor can withstand for a specific pulse shape (e.g., an 8/20 µs waveform) without its voltage changing by more than 10%, typically measured with a 5-minute interval between pulses. If the peak pulse current rating is insufficient, the varistor may be damaged or destroyed.