EMI, EMC, and Shielding Shielding and Enclosure Design Informational

How do I select the right EMI filter for an RF power supply to meet regulatory requirements?

An EMI filter for an RF power supply attenuates the conducted emissions generated by the power supply switching action, preventing them from reaching the AC mains (or the DC input bus) and causing interference. Selection criteria: (1) Identify the emission profile: measure the conducted emissions of the unfiltered power supply using a LISN and spectrum analyzer. Identify the frequencies and levels that exceed the regulatory limit (typically CISPR 32/FCC Part 15 limits). (2) Determine the required insertion loss (IL) at each critical frequency: IL_required = emission_level - limit + margin (typically 6-10 dB margin for production variation). Example: if the emission at 1 MHz is 75 dBuV and the limit is 56 dBuV: IL_required = 75 - 56 + 6 = 25 dB at 1 MHz. (3) Select the filter topology: common-mode (CM) filter: uses a CM choke (two windings on a common ferrite core). Attenuates CM emissions. The CM choke inductance determines the CM IL. Differential-mode (DM) filter: uses series inductors and shunt X-capacitors. The LC combination creates a low-pass filter that attenuates DM emissions. Pi-filter (C-L-C): provides 60 dB/decade rolloff above the cutoff frequency. T-filter (L-C-L): provides 60 dB/decade rolloff. Combined CM+DM filter: most commercial EMI filters include both CM and DM attenuation (CM choke + X-caps + Y-caps in a single module). (4) Verify the filter ratings: voltage rating: > max input voltage (120/240 VAC or DC bus voltage). Current rating: > max load current with derating for temperature. Leakage current: Y-capacitors are limited to prevent ground fault (typically < 0.5 mA for consumer products, < 3.5 mA for permanently connected equipment per IEC 60950). (5) Install and re-test: mount the filter at the power entry point (closest to the enclosure wall for best shielding effectiveness). Re-measure conducted emissions with the filter installed. Verify compliance with the limit.
Category: EMI, EMC, and Shielding
Updated: April 2026
Product Tie-In: Enclosures, Gaskets, Absorbers, Filters

EMI Filter Selection

EMI filter selection is a critical step in power supply design for any RF system. An improperly selected filter can fail to meet emissions requirements, introduce instability, or exceed safety limits.

Filter Topology Details

(1) First-order filter (single L or C): provides 20 dB/decade rolloff. Rarely sufficient for meeting modern emission limits. (2) Second-order (LC or CL): 40 dB/decade. Adequate for some low-noise designs. (3) Third-order (C-L-C pi-filter): 60 dB/decade. The most common topology for EMI filters. Provides high insertion loss with moderate component count. The two capacitors are shunt elements (line to neutral for X-caps, line to ground for Y-caps). The inductor is a series element (CM choke for CM rejection, DM inductor for DM rejection). (4) Multi-stage filter: cascade two or more filter sections for > 60 dB/decade rolloff. Used when emissions are very high or the limit is very stringent. Each additional LC section adds 40 dB/decade beyond the first order. However: each section adds cost, size, and potential resonance issues.

Component Selection

(1) X-capacitors (across line, L to N): provide DM attenuation. Safety-rated for across-the-line operation (will not fail short and create a fire hazard). X1: 4 kV surge rated. X2: 2.5 kV (most common). Values: 0.1-2.2 uF (film or ceramic). Self-healing (metallized film) for reliability. (2) Y-capacitors (line to ground): provide CM attenuation. Safety-rated for line-to-ground operation (limited capacitance to prevent excessive leakage current). Y1: 8 kV surge (for bridging basic+supplementary insulation). Y2: 5 kV (most common for consumer products). Values: 1-10 nF (limited by leakage current: I_leakage = V_line × 2×pi×f × C_Y. For C_Y = 4.7 nF at 60 Hz, 240 V: I = 0.43 mA). Maximum Y-cap per IEC 60950: 4.7 nF for Class I equipment (with ground), limited by leakage current specification. (3) Common-mode choke: a bifilar or multifilar winding on a high-permeability ferrite core (mu_r = 5000-15000, manganese-zinc or nickel-zinc ferrite). CM inductance: 1-50 mH. The CM choke has high impedance to CM signals and near-zero impedance to DM signals (the DM flux cancels in the core). CM choke frequency range: 10 kHz - 30 MHz (the core material determines the upper frequency). For higher frequencies: use a nickel-zinc ferrite (mu_r = 200-800, useful to 100 MHz). (4) DM inductor (series inductor): a single winding on a powdered iron or ferrite core. Inductance: 1-100 uH. Must not saturate at the full load current. For a 5 A power supply: use a core with sufficient energy storage (E = 0.5 × L × I^2).

Practical Considerations

(1) Filter placement: the EMI filter must be placed at the enclosure boundary (at the power entry point). If the filter is inside the enclosure with unfiltered wires running from the connector to the filter: those wires radiate the unfiltered emissions inside the enclosure, coupling to other internal wires and cables. The filter is then bypassed at high frequencies. Best practice: use a feedthrough filter or mount the filter module directly at the enclosure wall. (2) Resonance: the filter L and C components create a resonant circuit. At the resonant frequency: the filter may amplify emissions instead of attenuating them. Check: the resonant frequency should be well below the lowest emission frequency. If the resonance falls within the emission band: add damping (a resistor in series with the capacitor, or a lossy ferrite bead in series with the inductor). (3) Source and load impedance: the filter insertion loss depends on the source impedance (LISN = 50 ohms) and the load impedance (power supply input = varies). It is not just the filter alone that matters but the filter in the context of the system impedance. Measure the actual insertion loss with the LISN and the power supply connected (not on a 50-ohm bench network analyzer).

EMI Filter Equations
IL_required = emission - limit + margin (dB)
Pi-filter: 60 dB/decade rolloff
f_cutoff = 1/(2π√(LC))
Y-cap leakage: I = V×2πfC < 0.5 mA
CM choke: high Z to CM, low Z to DM
Common Questions

Frequently Asked Questions

Can I use just capacitors instead of a full LC filter?

Capacitors alone (without inductors) can provide some attenuation, but they are limited: (1) An X-capacitor across the line: provides DM attenuation at high frequencies. At 1 MHz with 0.47 uF: impedance = 0.34 ohms, forming a voltage divider with the LISN 50 ohms. Attenuation ≈ 20×log10(0.34/50) = -43 dB at 1 MHz. Sounds good, but: the capacitor ESL limits the high-frequency performance. A 0.47 uF film cap has ESL ≈ 10-20 nH, SRF ≈ 2-4 MHz. Above the SRF: the capacitor becomes inductive and provides no filtering. (2) Y-capacitors (line to ground): limited to 1-10 nF by safety requirements (leakage current limit). At 1 MHz with 4.7 nF: impedance = 34 ohms. Attenuation ≈ 20×log10(34/50) = -3.4 dB. Very little attenuation. For effective CM filtering: a CM choke (inductor) is essential. The choke provides high impedance in the CM path, working with the Y-caps to create a proper LC filter with 40+ dB/decade rolloff.

What about integrated EMI filter modules?

Commercial EMI filter modules (Schaffner, Schurter, TDK-Lambda, TE Connectivity) integrate all components in a single package: CM choke + X-caps + Y-caps. They are pre-designed for specific voltage/current ratings and regulatory standards. Advantages: guaranteed insertion loss (specified per CISPR 17), safety certified (UL, TUV, CSA), compact packaging (including IEC inlet connector on some models), and simplified design (no individual component selection needed). Disadvantages: limited customization (fixed topology and component values), higher cost than discrete components ($5-$30 per module vs $1-$5 for discrete), and the insertion loss specification may not match the actual system performance (depends on source/load impedance). For most commercial products: an integrated filter module is the fastest and safest path to compliance.

My power supply uses a DC-DC converter. Do I still need an EMI filter?

Yes. DC-DC converters (buck, boost, flyback) are significant sources of conducted emissions because: (1) They switch at 100 kHz - 2 MHz (generating harmonics up to 30+ MHz). (2) The switching current has fast edges (< 20 ns), creating high-frequency spectral content. (3) The input current waveform is a pulsed waveform (not smooth DC), creating ripple that conducts to the source. EMI filter considerations for DC-DC: (a) The LISN impedance for DC systems is 5-50 ohms (MIL-STD-461G uses 5 ohms, commercial standards use 50 ohms). (b) Y-capacitors are from the DC bus to chassis ground (no safety limitation on the capacitor value for DC systems, unlike AC mains). Larger Y-caps (100 nF - 10 uF) can be used for better CM filtering. (c) The DC-DC input filter must not create negative resistance instability with the converter feedback loop. Ensure the filter output impedance is much less than the converter input impedance at all frequencies (Middlebrook stability criterion).

Need expert RF components?

Request a Quote

RF Essentials supplies precision components for noise-critical, high-linearity, and impedance-matched systems.

Get in Touch