A 680µH common-mode choke suppresses conducted EMI by blocking noise currents that flow in the same direction on both conductors while allowing normal differential currents to pass largely unimpeded. It's most effective when its impedance peak aligns with your actual noise frequency range—something datasheets don't always make clear. Below I'll walk through how it works, where it fits in real filter circuits, and what to check beyond the 680µH number.
Key Takeaways
- The 680µH value is measured under test conditions; real filtering performance depends on the impedance curve across your noise spectrum, not the inductance alone.
- Common-mode chokes handle noise flowing in the same direction on both lines; differential-mode noise requires separate filtering (X caps, differential inductors, or tighter switching loops).
- Saturation, thermal rise, and leakage inductance can all degrade performance in ways the nominal inductance spec won't reveal—verify these under your max load and inrush conditions.
- In high-speed interfaces (USB 3.x, HDMI, DisplayPort), even small leakage inductance or parasitic capacitance can wreck eye diagrams; most data-line chokes use far lower inductance and specialized core materials.
- EMI filter placement, grounding, and loop layout often matter as much as component choice; I've seen identical chokes yield 10+ dB difference in conducted emissions simply from trace routing changes.
680µH Common-Mode Choke Overview
A 680µH common-mode choke is a passive two-winding inductor wound on a shared magnetic core, placed in series with power or signal lines to attenuate common-mode interference. You'll find it in AC input filters, switching power supplies, motor drives, communication interfaces, and automotive systems where cable-conducted noise can cause EMC failures or upset sensitive circuits.
The 680µH rating describes the common-mode inductance measured at a specific test frequency (often 10 kHz or 100 kHz) and low signal level. In my experience, this number is a starting point, not a guarantee. What actually suppresses noise is the choke's impedance magnitude across the frequencies where your EMI problem lives—typically a few hundred kHz to several MHz in switchers, sometimes extending to tens of MHz in high-speed digital systems. Always request or measure the impedance-versus-frequency curve; a choke with 680µH at 10 kHz may present only modest impedance at 1 MHz if its self-resonant frequency (SRF) is too low or core loss is high.
Beyond inductance, check rated current (both DC and any AC ripple component), DC resistance (which dictates copper loss and voltage drop), insulation rating for mains circuits, core material (NiZn ferrite for higher frequencies, MnZn for lower), temperature range, and saturation behavior under inrush or fault conditions.
How a 680µH Common-Mode Choke Works
The choke uses two windings on a single magnetic core. During normal operation, differential current—your intended power or signal—flows in opposite directions through the two windings. The resulting magnetic fluxes largely cancel inside the core, so the choke presents only its small leakage inductance and DC resistance to the wanted current.
Common-mode noise behaves differently. When unwanted noise currents flow in the same direction on both conductors (relative to ground or chassis), the magnetic fields add constructively inside the core. This produces high common-mode impedance, which blocks the noise path and prevents interference from propagating along cables or polluting the power bus.
In switching power supplies, common-mode noise originates from fast dV/dt at MOSFET drain nodes, parasitic capacitance between primary-side switching elements and grounded heat sinks or chassis, high-frequency displacement currents through transformer inter-winding capacitance, and return-current imbalances. These mechanisms can inject noise from low hundreds of kHz well into the MHz range. A 680µH choke will only help if its impedance is substantial across the actual noise spectrum you measure during pre-compliance or formal EMC testing.
One common mistake is assuming the choke behaves as an ideal 680µH inductor at all frequencies. In reality, parasitic winding capacitance creates a self-resonant peak, and above SRF the component looks capacitive, losing its filtering function. Core losses also increase with frequency, which can be beneficial (damping resonances) or detrimental (reducing impedance magnitude), depending on the circuit.
Common-Mode Noise vs Differential-Mode Noise
Electronic systems generate both common-mode and differential-mode noise, and each requires a different filtering approach.
Common-mode noise appears when unwanted currents flow in the same direction on multiple conductors relative to a reference plane (typically ground or chassis). A common-mode choke is specifically designed to present high impedance to this noise without significantly impeding the wanted differential signal.
Differential-mode noise flows in opposite directions between conductors—for example, ripple current in a switching converter's input or output path. Because the magnetic fields cancel in a common-mode choke, differential-mode noise sees only the small leakage inductance, which is usually insufficient for strong attenuation. To suppress differential-mode interference, designers use X capacitors across the line pair, discrete differential inductors, multi-stage LC filters, and tight switching-loop layout to minimize di/dt loop area.
In practice, most EMI filters combine both techniques: a common-mode choke for common-mode suppression, X caps or differential inductors for differential-mode suppression, Y caps to shunt residual common-mode current to ground, and careful component placement to avoid coupling noise back into the circuit. A 680µH choke alone won't solve a differential-mode problem, and stacking higher common-mode inductance won't help if your dominant noise is differential.
Common Applications of 680µH Common-Mode Choke
Switching Power Supplies
Common-mode chokes appear in flyback, buck, boost, and LED driver designs where fast MOSFET transitions couple noise onto input or output cables. Placing the choke near the AC or DC input reduces conducted emissions and helps the supply meet CISPR 32, EN 55032, or FCC Part 15 limits. In my projects, I've seen a well-chosen common-mode choke provide 10–20 dB attenuation in the 150 kHz–30 MHz conducted-emissions band, but only when the choke's impedance peak overlaps the dominant noise frequencies.
AC/DC Power Filters
In mains-input filters, the common-mode choke sits between the AC entry point (after fuse and MOV) and the X capacitor or rectifier stage. This position blocks common-mode noise from traveling back onto the power line and also reduces external transients entering the equipment. Safety standards (IEC 60950-1, IEC 62368-1) impose minimum creepage and clearance requirements, so the choke package and winding construction must meet these limits.
Automotive Electronics
Automotive systems—CAN bus, LIN, battery management, DC/DC converters—operate in electrically harsh environments with transient voltages, load dumps, and radiated fields. Common-mode chokes help maintain communication integrity and reduce radiated emissions from cable harnesses. Automotive designs typically require AEC-Q200-qualified components with proven thermal cycling, vibration resistance, and humidity performance.
Industrial and Communication Systems
Industrial PLCs, motor drives, and fieldbus interfaces use common-mode chokes to isolate subsystems and prevent ground loops from coupling noise between modules. Communication equipment—Ethernet switches, base stations, line cards—also uses chokes to reduce cable-radiated emissions and improve immunity to external interference.
High-Speed Interfaces
For USB 2.0, a 680µH common-mode choke can reduce cable emissions without significantly degrading the 480 Mbps signal, provided leakage inductance and parasitic capacitance remain low. However, USB 3.x (5 Gbps and above), HDMI 2.0/2.1, and DisplayPort require much lower inductance (often sub-10µH) and specialized low-leakage, low-capacitance designs to avoid eye-diagram closure, jitter accumulation, and impedance mismatches. In those applications, a 680µH part would almost certainly fail signal-integrity testing.
Practical EMI Filter Example
A common filter topology for a switching power supply AC input stage is:
AC Input → Fuse → MOV → 680µH Common-Mode Choke → X Capacitor → Rectifier Bridge
| Component | Main Function | Practical Note |
|---|---|---|
| Fuse | Overcurrent protection | Opens circuit during fault; must coordinate with inrush limiter if present |
| MOV (Metal Oxide Varistor) | Surge suppression | Clamps transient overvoltage before it reaches active circuits |
| 680µH Common-Mode Choke | Common-mode EMI attenuation | Blocks noise flowing in the same direction on L and N relative to ground |
| X Capacitor | Differential-mode EMI attenuation | Placed across L and N to shunt line-to-line switching noise |
| Rectifier Bridge | AC-to-DC conversion | Feeds the bulk capacitor and downstream DC rails |
In this arrangement, the common-mode choke handles noise that couples to chassis or ground through parasitic capacitance, while the X cap addresses differential switching ripple. Y capacitors (line-to-ground and neutral-to-ground) are often added in parallel with the choke windings to provide an additional common-mode shunt path, but Y-cap current must stay within safety limits (typically a few milliamps leakage).
Layout considerations:
- Place the choke as close as possible to the AC input connector to prevent noise from radiating off the PCB traces before filtering.
- Keep high-frequency switching nodes (MOSFET drain, rectifier diode cathode) away from the choke and input traces to avoid capacitive coupling that bypasses the filter.
- Use short, low-inductance connections for X and Y capacitors; long traces add parasitic inductance that degrades high-frequency attenuation.
- Verify creepage and clearance distances for mains voltage per IEC 60950-1 or IEC 62368-1.
I've debugged designs where moving the common-mode choke 2 cm farther from the AC input increased conducted emissions by 8 dB at 3 MHz because the unfiltered trace segment acted as an antenna. Component placement matters as much as component choice.
Specifications and Selection Guide
| Specification | Selection Guidance |
|---|---|
| Rated Current | Must handle continuous RMS current plus any AC ripple component without exceeding the temperature rise limit or driving the core into saturation. Check both the thermal current limit (set by copper loss and ambient temperature) and the saturation current limit (where inductance drops significantly). Inrush current may briefly exceed steady-state rating, so verify the choke can tolerate short-duration overload. |
| DC Resistance (DCR) | Lower DCR reduces copper loss, heat, and voltage drop. In a 5 A application, a choke with 0.1 Ω DCR dissipates 2.5 W and drops 0.5 V; a 0.02 Ω part dissipates only 0.5 W and drops 0.1 V. |
| Impedance-Frequency Curve | The actual impedance magnitude at your noise frequencies determines filtering performance. Request or measure the curve from 10 kHz to 100 MHz. A choke with 680µH at 10 kHz but SRF at 2 MHz may provide little attenuation at 5 MHz, where your EMI problem peaks. |
| Leakage Inductance | Leakage inductance is the small series inductance seen by differential-mode signals. In power filters, moderate leakage can provide useful differential-mode attenuation. In high-speed data lines, excessive leakage degrades eye diagrams and increases jitter. |
| Self-Resonant Frequency (SRF) | Operate well below SRF to maintain inductive behavior. Near or above SRF, parasitic capacitance dominates and the choke may actually increase noise coupling. |
| Core Material | NiZn ferrite offers higher resistivity and performs better at MHz frequencies; MnZn ferrite has higher permeability and lower loss at kHz but becomes lossy above a few MHz. Choose NiZn for switching noise in the 1–30 MHz range, MnZn for lower-frequency applications or where higher inductance in a smaller package is needed. |
| Package and Reliability | Consider PCB footprint, through-hole versus SMD, creepage/clearance for safety, thermal rating, AEC-Q200 qualification for automotive, and ingress protection for harsh environments. |
When choosing between two chokes with similar inductance, I prioritize the impedance curve over the nominal inductance value. A choke with lower inductance but higher impedance at the problem frequency will outperform a higher-inductance part whose impedance has already rolled off.
Verification and Testing
A 680µH common-mode choke must be validated in the actual circuit because EMI performance depends on switching frequency, load current, cable type and length, grounding topology, PCB layout, and interaction with other filter components. A part that looks ideal on paper may deliver disappointing results if its impedance peak misses your noise band or if parasitic coupling bypasses the filter.
EMI Testing
Conducted emissions testing per CISPR 32 or FCC Part 15 is the primary validation method. Use a line-impedance stabilization network (LISN) and spectrum analyzer to measure noise on the AC mains from 150 kHz to 30 MHz. Compare results with and without the common-mode choke to quantify its attenuation. Near-field probes or current probes can help identify noise sources and verify that the choke is blocking the intended path.
In one project, a 680µH choke provided excellent attenuation from 500 kHz to 5 MHz but made emissions worse at 8 MHz because its self-resonance coincided with a switching harmonic, creating a parallel resonance with the input capacitance. We fixed it by adding a small series resistor to dampen the resonance—an approach that sacrifices some peak attenuation but broadens the effective filtering bandwidth.
Thermal Testing
Measure the choke's surface temperature at maximum continuous current and worst-case ambient temperature. Excessive heating may indicate copper loss from high DCR, core loss from AC flux, or partial saturation. Any of these will reduce long-term reliability and may degrade EMI suppression if core permeability drops with temperature.
Impedance Verification
If you have an impedance analyzer or vector network analyzer, measure the choke's impedance magnitude and phase from 10 kHz to 30 MHz. Compare the measured curve against the datasheet. Significant deviation may indicate a counterfeit part, wrong core material, or a batch variation.
Signal Integrity (for Data Lines)
For high-speed interfaces, measure eye diagrams, jitter, and return loss (S₁₁, S₂₂) to confirm the choke does not degrade signal quality. However, in AC power filters, signal-integrity testing is irrelevant; focus on EMI measurement, thermal performance, and impedance characteristics.
EMI Problems and Troubleshooting
| Problem | Possible Cause | Recommended Fix |
|---|---|---|
| EMI failure at high frequency (>10 MHz) | Insufficient impedance above choke's SRF, or parasitic capacitance bypassing the choke | Select a choke with higher SRF or add a second smaller choke in series; check for capacitive coupling between input traces and switching nodes |
| Eye-diagram degradation in data lines | Excessive leakage inductance or common-mode impedance | Use an ultra-low-leakage, low-inductance choke designed for high-speed signals |
| Choke overheating | High DCR, excessive ripple current, or partial core saturation | Choose a lower-DCR part, increase the core size, or reduce AC flux by lowering inductance and adding a second stage |
| Limited EMI improvement despite high inductance | Poor PCB layout, long unfiltered traces, ground loops, or differential-mode noise (which the choke doesn't address) | Move choke closer to input, shorten traces, add X caps for differential noise, improve grounding, and check cable shielding |
| Resonance peak causing worse emissions at specific frequency | Choke SRF interacting with input capacitance | Add series damping resistor (a few ohms) or choose choke with different SRF |
One failure mode I've encountered is assuming a larger inductance value always improves filtering. In reality, higher inductance often means larger core size, more turns, higher inter-winding capacitance, and lower SRF—any of which can degrade high-frequency performance. Sometimes two smaller chokes in series, each with lower inductance and higher SRF, outperform a single high-inductance part.
Frequently Asked Questions
Why does a 680µH common-mode choke suppress noise without blocking normal current?
The choke uses two coupled windings on a shared core. Normal differential current flows in opposite directions through the windings, so the magnetic fluxes cancel and the choke presents only minimal leakage inductance and DC resistance. Common-mode noise currents flow in the same direction, causing the magnetic fields to add and create high impedance that blocks the noise path.
What tradeoffs should I consider when selecting a 680µH common-mode choke?
Higher inductance can improve low-frequency attenuation but often increases size, weight, DC resistance, cost, and parasitic capacitance, which lowers SRF and may harm high-frequency performance or signal integrity. In power filters, balance inductance against thermal limits and PCB space. In data lines, prioritize low leakage and low parasitic capacitance over high inductance.
Why do two systems with the same choke show different EMC results?
EMC performance depends on the entire system: PCB layout, grounding, cable routing, shielding, switching-loop area, and the choke's position relative to noise sources. I've measured 10+ dB variation in conducted emissions by changing only the choke's PCB placement and trace routing, with no change to the choke itself.
What signs suggest I need a 680µH common-mode choke?
Excessive conducted emissions in the 150 kHz–30 MHz range, failed EMC testing, communication errors or resets correlated with switching activity, and measurable common-mode current on cables (detectable with a current probe) all indicate a common-mode noise problem. Differential-mode noise—ripple on power rails, switching artifacts in the time domain—requires different filtering.
Why doesn't increasing inductance always improve EMI performance?
Because EMI suppression depends on impedance at the noise frequency, not inductance alone. Higher inductance may lower SRF, increasing parasitic capacitance and reducing high-frequency impedance. It also increases DCR, heat, and voltage drop. Moreover, if noise is bypassing the choke through poor layout, ground loops, or capacitive coupling, adding inductance won't help. Effective EMI control requires matching the choke's impedance curve to the measured noise spectrum, optimizing layout to prevent bypass paths, and combining common-mode and differential-mode filtering as needed.
