Effective Design of EMI Filter for Enhanced Performance in Electronics
Key Takeaways
- EMI filter design targets both conducted emissions (150 kHz–30 MHz) and radiated emissions (30 MHz–1 GHz), with compliance to CISPR and FCC standards built into the design from day one.
- The core of most power-line EMI filters combines differential-mode LC stages with common-mode chokes, sized from measured spectra rather than guesswork.
- DOREXS follows an impedance-based, measurement-driven design flow to avoid oversized, overcostly filters while guaranteeing EMC compliance margins.
- Layout, parasitics, and safety clearances often dominate real-world filter performance more than ideal schematics—these must be treated as primary design constraints.
- This article walks through concrete design steps, example topologies, typical component ranges, and how DOREXS delivers custom filters and transformers tailored to specific converters.
Introduction to EMI and the Need for Filters
Modern switch-mode power supplies have revolutionized electronics with their compact size and high efficiency, but they come with a hidden cost: electromagnetic interference. Whether you’re designing a 65 W USB-PD adapter for consumer devices or a 1 kW servo drive for industrial automation, the fast switching edges and high di/dt transients in these converters inevitably generate noise that can wreak havoc on nearby equipment. Every time a MOSFET or GaN transistor switches at hundreds of kilohertz—or even megahertz—it creates voltage spikes and current harmonics that propagate through power lines and radiate into the surrounding environment.
Regulatory limits force designers to address this noise head-on. Standards like CISPR 11, CISPR 22/32, and FCC Part 15 define strict conducted emissions limits from 150 kHz to 30 MHz and radiated emissions limits from 30 MHz to 1 GHz, requiring that power supplies include properly designed input filter stages.
The risks of unfiltered EMI extend far beyond failed certification tests:
-
Medical equipment interference: High frequency noise can corrupt monitoring signals in patient-connected devices, creating safety hazards in clinical settings
-
Industrial control corruption: Data bus errors and communication failures in PLCs and automation networks can halt production lines
-
Audio/visual artifacts: Consumer products may exhibit buzzing, screen flickering, or wireless connectivity problems
-
Certification failure: A single dB over the limit at an accredited test lab means redesign costs and delayed market entry
For many offline power supplies, the EMI filter can easily consume 20–30% of the total volume and represent a significant portion of the bill of materials. This makes emi filter design a critical optimization task rather than an afterthought bolted onto a finished design.
DOREXS specializes in EMI filters and transformers as a B2B manufacturer, supporting sectors like medical devices, industrial automation, and home appliances with both standard catalog filters and fully custom solutions. The company’s design philosophy centers on measurement-driven optimization—delivering compliant filters without the excess cost and size of overengineered alternatives.
Types of EMI: Radiated vs. Conducted, Common-Mode vs. Differential-Mode
Understanding the “type” of EMI is the first step toward choosing the correct filter topology and components. Without this classification, engineers often waste time and money applying the wrong suppression techniques to the wrong noise modes.
Conducted Emissions
Conducted emissions refer to noise that appears on power lines, typically measured in the frequency range between 150 kHz and 30 MHz. Test procedures defined by CISPR or FCC standards use a Line Impedance Stabilization Network (LISN) connected between the equipment under test and a spectrum analyzer. The LISN provides a standardized 50 Ω impedance and isolates the measurement from external mains noise, allowing repeatable quantification of the interference your converter injects back onto the AC or DC supply lines.
Radiated Emissions
Radiated emissions occupy the 30 MHz to 1 GHz range (and sometimes higher). These emissions often originate from fast switching nodes and cable structures acting as unintended antennas. While line filters cannot directly suppress radiated fields, they play a critical role by reducing the high-frequency currents that feed these antenna structures. A well-designed filter cuts off the noise source before it can couple into cables and chassis elements.
Differential-Mode Noise
Differential-mode noise flows between line and neutral (or between positive and negative DC rails). This noise relates directly to the pulsating input current drawn by converters—every buck converter, PFC stage, or inverter front-end draws current in pulses synchronized to its switching frequency. The resulting harmonics appear as DM noise on the power lines. DM currents flow in a loop through the converter and back through the supply, making them relatively straightforward to attenuate with series inductance and shunt capacitance.
Common-Mode Noise
Common-mode noise flows from both lines simultaneously to chassis or earth ground through parasitic capacitance. The primary culprits include MOSFET drain-to-heatsink capacitance, transformer interwinding capacitance, and PCB coupling to metal enclosures. CM currents can be particularly troublesome because they find unintended return paths through the safety ground conductor, potentially causing EMI issues that seem disconnected from the obvious power flow.
In most practical systems, DM dominates near the switching frequency and its low harmonics (tens to hundreds of kHz), while CM tends to dominate at higher frequencies (several MHz and above). This frequency-dependent behavior directly influences component selection: DM filters need effective attenuation in the lower portion of the conducted band, while CM filters must maintain high impedance well into the MHz range.
Regulatory Limits and Design Targets
Designers must translate abstract EMI standards into concrete dBµV limits on a measurement plot. Without clear numerical targets, filter design becomes guesswork—either resulting in failed tests or wastefully overbuilt filters.
Relevant Standards
The standards relevant to DOREXS customers span multiple industries that rely on power quality solutions and EMI filter applications:
|
Standard |
Application |
Typical Frequency Range |
|---|---|---|
|
CISPR 32 |
IT/multimedia equipment |
150 kHz – 30 MHz (conducted), 30 MHz – 6 GHz (radiated) |
|
CISPR 11 |
Industrial, scientific, medical equipment |
150 kHz – 30 MHz (conducted), 30 MHz – 1 GHz (radiated) |
|
IEC 60601-1-2 |
Medical device EMC |
Per CISPR 11 with additional immunity requirements |
|
CISPR 25 |
Automotive electronics |
150 kHz – 108 MHz (conducted), 150 kHz – 2.5 GHz (radiated) |
|
FCC Part 15 |
Consumer electronics (US) |
150 kHz – 30 MHz (conducted), 30 MHz – 40 GHz (radiated) |
Measurement Methodology
Conducted limits are specified as quasi-peak and average values in dBµV at the LISN port. Limit curves vary across the 150 kHz–30 MHz band, with different equipment classes (Class A for industrial environments, Class B for residential) having different thresholds. Class B limits are typically 10 dB more stringent than Class A, reflecting the expectation of closer proximity to sensitive consumer equipment.
Setting Design Targets
DOREXS typically designs for at least 6 dB of headroom below regulatory limits to account for production spread, component aging, and test-lab variation. A filter that passes certification with only 1–2 dB of margin may fail in production when component tolerances stack unfavorably.
Setting clear attenuation targets guides component sizing:
-
“Need 40 dB DM attenuation at 300 kHz” → determines minimum DM inductor and X capacitor values
-
“At least 25 dB CM attenuation above 5 MHz” → sets CM choke inductance and Y capacitor requirements
These quantified targets transform filter design from trial-and-error into systematic engineering.
Systematic EMI Filter Design Flow
Too many engineers approach EMI filtering through trial-and-error: grab a filter from the catalog, test the system, tweak components when it fails, repeat until something works. This approach wastes development cycles and often results in filters that are either inadequate or grossly oversized. DOREXS employs a structured, measurement-based design flow that reduces iterations and material cost while ensuring compliance.
Three-Step Impedance-Based Process
The systematic approach follows three core steps:
-
Characterize the noise source: Measure the converter’s conducted emissions spectrum and estimate its source impedance across the frequency range of interest
-
Identify line and load impedances: Understand the LISN behavior and real-world mains characteristics that the filter will see
-
Synthesize the filter: Design a network that shapes total system impedance to keep emissions under limits with adequate margin
Measurement Setup
To measure noise accurately, connect your equipment under test to a LISN and capture data with a spectrum analyzer configured for the appropriate detector (quasi-peak and average per CISPR requirements). Operate the converter at worst-case conditions: maximum load, minimum input voltage, and highest expected ambient temperature. These conditions typically produce the highest switching stress and therefore the worst EMI.
Document the failing frequencies and magnitudes. For example, a 24 V, 5 A DC/DC buck converter operating at 250 kHz might show harmonics exceeding CISPR limits at 250 kHz, 500 kHz, and 750 kHz by 15–20 dB, with CM emissions peaking around 5–10 MHz.
Calculating Required Attenuation
With measured data in hand, calculate the attenuation needed at each problematic frequency:
Required Attenuation = Measured Level – Limit + Margin
If your 250 kHz harmonic measures at 85 dBµV and the limit is 66 dBµV, you need at least 19 dB of attenuation—plus 6 dB margin, totaling 25 dB minimum at that frequency.
This measurement-driven approach is exactly how DOREXS, as an EMI filter supplier and manufacturer, sizes both standard catalog filters and custom solutions. When customers share preliminary EMI data, engineers can rapidly converge on appropriate topologies without iterative guessing.
Design of Differential-Mode EMI Filters
DM filtering is implemented with series inductors and shunt capacitors forming low pass networks, typically placed just after the input connector or bridge rectifier. These lc filter stages prevent the pulsating current drawn by the converter from propagating back to the mains.
Basic LC Stage
The simplest DM filter consists of a series filter inductor (LDM) and a shunt filter capacitor (CX, an X-rated capacitor between line and neutral). The cutoff frequency follows the standard relationship:
fc ≈ 1 / (2π√(LDM × CX))
This cutoff should sit at least one decade below the lowest problematic noise harmonic to ensure adequate attenuation at the frequencies that matter.
Numeric Example
For a 200 kHz switching converter with strong noise at 200–400 kHz:
-
Target cutoff frequency: 20–30 kHz
-
Typical component values: LDM ≈ 1–3 mH, CX ≈ 47–220 nF (X2 safety-rated for 230 VAC systems)
A 2 mH inductor paired with a 100 nF capacitor yields fc ≈ 11 kHz, providing roughly 40 dB of attenuation at 200 kHz—often sufficient for a single-stage design.
Real-World Limitations
Ideal calculations assume perfect components, but real-world parasitics flatten attenuation at higher frequencies:
-
Capacitor ESR and parasitic inductance (ESL) limit high-frequency bypass effectiveness
-
Inductor winding capacitance creates parallel resonances that can actually boost noise at certain frequencies
-
PCB stray inductance adds uncontrolled impedance in the filter path
When a single LC stage cannot achieve required attenuation across the full frequency range, options include:
-
Adding a second LC stage for steeper rolloff
-
Including damping networks (RC snubbers or resistor-damped capacitors) to flatten resonance peaks
-
Using multiple capacitor types in parallel to extend effective bandwidth
DOREXS optimizes DM chokes not only for inductance value but also for saturation behavior at peak inrush and fault currents, thermal rise under continuous load, and safety creepage distances appropriate for different mains insulation categories.
Design of Common-Mode EMI Filters
CM noise is strongly influenced by parasitic capacitance between switching nodes and chassis or earth. Because these stray coupling paths exist in every power converter, CM filtering is often the dominant factor in passing tests above a few MHz.
The Common-Mode Choke
The core element of CM filtering is the common-mode choke: two (or three for three-phase systems) windings on a high-permeability ferromagnetic core. The windings are arranged so that differential currents produce opposing magnetic fields that cancel in the core, while common-mode currents produce additive fields that magnetize the core strongly.
This elegant design presents high impedance to CM currents without significantly increasing the differential-mode voltage drop or copper loss. The result: high attenuation of CM noise with minimal impact on normal power flow.
Component Value Ranges
Typical CM inductance values depend on the application:
|
Application |
Typical CM Inductance |
|---|---|
|
Single-phase AC (230 VAC) |
1–30 mH per winding |
|
DC systems (48 V telecom) |
100 µH – several mH |
|
Three-phase industrial |
1–10 mH per winding |
Higher inductance provides more CM attenuation but requires larger cores and more copper, increasing both size and cost.
Y-Capacitors
Y-capacitors connect from line and neutral to earth (PE), creating a defined CM return path. These capacitors improve high-frequency attenuation but are limited by safety considerations:
-
Leakage current limits: IEC 60950 allows up to 3.5 mA for IT equipment; IEC 60601 for medical devices may require less than 100 µA for patient-connected equipment
-
Safety ratings: CY1 capacitors provide reinforced insulation; CY2 provides basic insulation
For medical applications under IEC 60601, extremely low leakage requirements often constrain Y capacitor values to single-digit nanofarads, forcing designers to rely more heavily on optimized CM chokes.
DOREXS offers both standard CM chokes and integrated CM+DM filter modules. For applications requiring ultra-low leakage—such as BF or CF-rated medical equipment—the company can redesign core materials and winding structures to maximize CM attenuation while minimizing capacitive coupling to earth.
Component Parasitics, Layout, and Real-World Performance
Even a perfectly calculated filter will underperform if PCB layout, wiring, and component parasitics are not carefully controlled. Engineers often discover that their filter provides 20 dB less attenuation than expected because noise bypasses the filter through unintended coupling paths.
Capacitor Parasitics
Every capacitor has ESR (equivalent series resistance) and ESL (equivalent series inductance) that limit its effectiveness at higher frequencies:
|
Capacitor Type |
Typical ESL |
Best Frequency Range |
|---|---|---|
|
Aluminum electrolytic |
10–50 nH |
Below 100 kHz |
|
Film (X-rated) |
5–15 nH |
Up to 1–5 MHz |
|
Ceramic (MLCC) |
0.5–2 nH |
Up to 100 MHz+ |
Using multiple capacitor types in parallel widens effective attenuation bandwidth. A film capacitor handles low frequencies while a smaller ceramic shunts high-frequency noise.
Inductor Parasitics
Inductor winding capacitance creates a self-resonant frequency (SRF) above which the inductor behaves like a capacitor. For filter inductors wound with many turns of fine wire, SRF may occur at just a few MHz—precisely where CM noise often peaks. Above SRF, attenuation actually decreases.
Layout Guidelines
Layout determines whether your filter functions as designed or serves as an expensive ornament:
-
Minimize loop area: High-frequency paths should be as short and direct as possible
-
Isolate noisy nodes: Keep switch node traces physically separated from filter input traces
-
Avoid parallel routing: High-dv/dt traces running parallel to quiet traces couple noise capacitively
-
Use short, wide traces: Reduce parasitic inductance in filter connections
-
Ground planes: Provide low-impedance return paths for filter capacitors
A poorly laid out filter with ideal component values can easily perform 10–20 dB worse than expected, negating careful electrical design.
DOREXS’ design philosophy includes early 3D and layout reviews for custom filters. Where needed, the company provides enclosure and shielding recommendations to ensure the implemented filter achieves modeled attenuation in the customer’s mechanical context.
DOREXS EMI Filter Design Philosophy and Product Integration
DOREXS manufactures single-phase, three-phase, IEC inlet, PCB-mount, feed-through, and DC EMI filters, along with transformers used in EMI-critical power stages. The company serves industrial automation, medical device, consumer electronics, and home appliance manufacturers who need reliable EMC solutions.
Measurement-Driven Design
DOREXS emphasizes a “design-by-measurement” philosophy: sizing filters based on measured spectra and target margins rather than generic catalog assumptions. This approach avoids both under-filtering (which causes test failures) and excessive over-design (which wastes cost and board space).
When engineering teams share their conducted emissions data—even preliminary bench measurements—DOREXS can rapidly assess whether a standard filter will suffice or whether modifications are needed.
Collaborative Engineering
DOREXS collaborates with customer engineering teams through multiple touchpoints:
-
Reviewing schematics and layouts for EMI-critical issues
-
Proposing filter topologies (two-stage LC, π-filters, integrated CM+DM chokes)
-
Providing SPICE or S-parameter models for co-simulation with customer power stages
-
Recommending component substitutions when supply chain constraints arise
Integrated Solutions
For sectors like medical devices and industrial automation, DOREXS considers power-quality aspects beyond pure EMI suppression:
-
Inrush behavior: Integrated NTC thermistors or inrush limiters prevent capacitor charging surges from tripping breakers
-
Leakage current: Component selection ensures compliance with stringent medical leakage limits
-
Overload endurance: Thermal margins accommodate fault conditions and worst-case operating points
-
Surge protection: MOVs and TVS devices can be integrated into the filter assembly
DOREXS supports customers from early concept through certification, offering technical consulting, on-site or remote troubleshooting of EMI test failures, and tailored redesigns for second-generation products.
Design Examples and Application-Specific Considerations
EMI filter design differs significantly between applications. A consumer adapter, industrial drive, and medical power supply each face unique constraints that shape filter topology and component selection.
Consumer SMPS: 150 W Class II Adapter
A typical 90–264 VAC, 150 W Class II adapter uses AC EMI filter stages:
-
One-stage DM LC filter: Series inductor (1–2 mH) plus X2 capacitor (100–220 nF)
-
CM choke: 5–15 mH with modest Y capacitor (2.2–4.7 nF) to stay within leakage limits
-
Space-saving approach: PCB-mount filter modules that integrate all elements in a compact footprint
Class II (double-insulated) equipment has no earth connection, so Y capacitors connect to a floating reference rather than true earth—simplifying some aspects while requiring careful attention to CM return paths.
Industrial Drive: 3-Phase 400 VAC, 5 kW
Higher power industrial drives using single-phase 220 V AC EMI filters face different challenges:
-
Higher DM currents: Require larger inductors with appropriate saturation margins
-
3-phase CM chokes: Three-winding chokes on toroidal cores provide balanced CM suppression
-
PFC stage emissions: Strong conducted emissions around the PFC switching frequency (typically 50–150 kHz) often require dedicated filtering
-
Mechanical considerations: Rail-mount or panel-mount filters with robust terminals for high-current connections
Medical Device: 1 kW Imaging Subsystem (IEC 60601-1-2)
Medical equipment compliance under IEC 60601-1-2:2014 demands:
-
Ultra-low leakage: Patient and touch leakage currents measured in microamps constrain Y capacitor values severely
-
Enhanced CM chokes: When Y capacitors cannot provide adequate CM shunting, larger CM chokes with optimized high-frequency performance become essential
-
Safety approvals: Components must carry appropriate medical safety certifications (UL 60601, IEC 60601)
DOREXS can adapt standard EMI/EMC filter designs—modifying Y-cap values, changing core materials, adding additional DM stages—into application-specific variants without requiring full custom tooling for every project.
Images and Visual Elements to Include
This article benefits from visual elements that reinforce technical concepts:
-
CM vs. DM current path illustration: Shows L/N/PE conductors with current loops clearly distinguished
-
EMI test setup: LISN, spectrum analyzer, EUT, and grounding layout
-
Basic LC DM filter schematic: Fuse, surge limiter, bridge rectifier, DM inductor, X-capacitors
-
CM choke and Y-capacitor topology: Annotated with current paths and safety markings
-
Good vs. bad PCB layout comparison: Identical schematics with different trace routing showing coupling effects
-
DOREXS product lineup: Single-phase and three-phase filters with visible branding
-
Application block diagrams: Filter location relative to mains connector, PFC, DC/DC stage
Each diagram should use consistent symbol conventions (L for inductors, C for capacitors, GND/PE for earth) with color coding to distinguish CM from DM current paths where helpful.
Summary and Best Practices
-
Classify EMI correctly: Distinguish between conducted/radiated and CM/DM modes before selecting filter topology
-
Measure first, design second: Base filter specifications on actual emissions data, not assumptions
-
Apply DM and CM stages appropriately: Use LC stages for DM suppression and CM chokes with Y capacitors for CM—each mode requires its own solution
-
Respect parasitics and layout: Component placement and PCB routing often matter more than ideal schematic values
-
Build in margin: Design for 6–10 dB below regulatory limits to accommodate production variation
Successful EMI filter design balances electrical performance, safety requirements, mechanical constraints, cost targets, and certification margins. The most effective approach treats EMI filtering as a first-class design concern from the earliest schematic revision—not a problem to solve after the first failed pre-compliance test.
Always validate filters with representative worst-case operating conditions. A filter that passes at room temperature and 50% load may fail at high ambient temperature and maximum output current.
Engineering and procurement teams seeking reliable EMI solutions should involve DOREXS early in their design cycle. Whether you need standard catalog filters or custom designs tailored to unique mechanical and electrical constraints, early collaboration reduces time-to-market and certification risk.
FAQ
How do I choose between a catalog EMI filter and a custom design?
Catalog filters work well for common input voltages (115/230 VAC, 24/48 VDC) and power ranges with typical noise profiles. They offer faster availability and lower unit cost due to volume production. Custom filters become necessary when you face tightly constrained mechanical layouts, unusual noise spectra, or stringent leakage limits—situations common in medical applications.
Start with a standard DOREXS filter matching your current and voltage requirements. If pre-compliance tests show insufficient attenuation, excessive margin (suggesting you’re paying for unneeded performance), or mechanical fit issues, move to a custom variant. Sharing your EMI test data and mechanical constraints with DOREXS enables rapid feasibility assessment for modified or custom solutions.
Can EMI filters fix poor PCB layout or transformer design?
EMI filters significantly reduce conducted noise, but they cannot fully compensate for fundamental design problems. An extremely poor PCB layout with large loop areas, unshielded transformers radiating stray fields, or very high parasitic capacitances from switch nodes to chassis will continue causing EMI issues regardless of filter size.
Address EMI at the noise source first: add snubbers to reduce ringing, slow switching edges where thermal margins allow, optimize transformer winding arrangements, and implement proper grounding. Then use a filter to achieve final regulatory margins. DOREXS often reviews customer layouts and transformer designs to ensure the filter and power stage form a coherent EMI strategy rather than fighting each other.
What information does DOREXS need to design a custom EMI filter?
Key data for custom filter development includes:
-
Input voltage range (e.g., 90–264 VAC or 18–72 VDC)
-
Maximum continuous and surge current requirements
-
Switching frequency and power topology
-
Target standards (CISPR class, medical IEC 60601, automotive CISPR 25)
-
Mechanical constraints (dimensions, mounting style, connector preferences)
-
Leakage current limits and creepage/clearance requirements
Providing preliminary LISN measurements showing conducted emissions at worst-case operating points enables more precise attenuation targeting. Safety documentation requirements (UL, ENEC, CSA certifications) should also be communicated so DOREXS can select appropriate insulation systems and component approvals.
How does an EMI filter affect power quality and efficiency?
A well-designed filter introduces small additional losses—mainly in choke copper and core dissipation plus capacitor losses—typically representing a fraction of a percent of total power supply efficiency. For a 90% efficient 100 W supply, filter losses might add 0.2–0.5 W.
Large inductances and capacitances can influence other system behaviors:
-
Inrush current: Input capacitors require charging current at turn-on
-
Power factor: Added reactive elements may affect PF in certain topologies
-
Dynamic response: LC resonances can interact with control loop stability
DOREXS accounts for these effects during design, often integrating bleed resistors for capacitor discharge or coordinated NTCs for inrush management. Verify efficiency and thermal performance with the filter installed to confirm acceptable impact for your application.
When should I involve EMI filter design in my project timeline?
EMI considerations belong in the first revision of your power stage schematic and PCB layout—not the week before certification testing. Treating EMI as an afterthought leads to expensive redesigns when filters don’t fit in the available space or when fundamental layout issues prevent any filter from achieving compliance.
Reserve physical space on the PCB or in the enclosure for at least a two-stage filter option. This flexibility allows adjustments after early bench tests without major mechanical redesign. DOREXS supports early design reviews, helping teams predict EMI challenges and pre-select suitable EMI/EMC filter platforms before hardware is frozen. This proactive approach reduces both development time and certification risk.
EMI in Elevator Control Panels: Causes, Risks and Filter-Based Solutions
Common Mode Chokes vs EMI Filters: Which EMI Solution is Right for Your Application?