Common-mode and differential-mode noise are not two names for the same interference. They differ in where the current returns, which means they couple differently, radiate differently, and respond to entirely different components. Fitting the wrong filter is the most common reason an interference problem survives three rounds of parts swapping.
The difference is the return path, not the amplitude
Differential-mode noise appears between two conductors of the same circuit, so its current leaves on one and comes straight back on the other. It is the same topology as the wanted signal, which is why it is hard to filter without also attacking what you want to keep. Common-mode noise appears between the whole cable and the surrounding reference, usually earth or a chassis. Its current leaves on every conductor together and finds a return through some path you never drew on the schematic: a mounting screw, a safety ground, a capacitance to a metal cabinet. Because the outgoing fields do not cancel, common-mode current radiates far more efficiently for the same amperage. A few microamps of common mode can fail an emissions test while milliamps of differential mode pass.
How to tell which one you have
The cleanest test is a current probe clamped around the full bundle. Clamp everything at once and the differential components cancel, so whatever the probe reads is common mode. Clamp a single conductor and you see the sum of both. Comparing the two readings separates them in a minute. Without a probe, a substitution test works: clip a common-mode choke over the whole bundle and see whether the symptom changes. If it does, you had common mode. If nothing happens, look at differential paths or at direct radiation. On audio systems, a noise that changes when you lift or reconnect an interconnect shield is usually common mode. A noise that scales with the signal level or persists with all cables removed except the pair is usually differential.
The components that treat each mode are physically different
A differential-mode filter is the familiar one: a series inductor and a capacitor across the line, sized so the cutoff sits below the noise and above the wanted band. In mains filters these are the X-class capacitors bridging line to neutral. A common-mode filter is a choke wound so that both conductors pass through the same core in the same sense, giving high impedance to current flowing together and almost none to current flowing in opposite directions. Paired with it are Y-class capacitors from each line to the chassis, which give common-mode current a short local return instead of a long one. Y-class values are deliberately limited because they also pass leakage current to ground, which is a shock-safety constraint, not an engineering preference.
Mode conversion means a clean source can still cause trouble
Real cables are never perfectly balanced. When the impedance from one conductor to ground differs slightly from the other, some differential current converts into common mode along the way, and some common mode converts back into differential at the far end. That conversion is why a well-filtered supply can still cause interference once it drives a long, asymmetric cable, and why a receiver can hear noise that a differential measurement at the source never showed. Balance is therefore a design goal in its own right: twisted pairs with a consistent twist rate, symmetric connector pinouts, and terminations that present matched impedances to ground on both sides. Improving balance often does more than adding filtering, because it stops the conversion rather than mopping up the result.
Why switching supplies generate mostly common mode
Inside a switching converter, a semiconductor switch drives a node through tens of volts in tens of nanoseconds. That node is usually bolted to a heatsink, and the heatsink has capacitance to the chassis. Every switching edge therefore injects a small charge pulse into the chassis, which has to return to the converter through whatever path exists, typically the mains cable. The result is broadband common-mode current with harmonics stretching well above the switching frequency. This is why input filters on switchers are dominated by common-mode chokes, why heatsink insulation and shield plates matter, and why a supply that measures quietly on the bench can misbehave once it is screwed into a different enclosure. The enclosure is part of the circuit even though no one drew it.
What to check before you act
- Measure the whole bundle with a clamp probe before choosing a filter.
- Verify the noise changes when you add a common-mode choke.
- Keep cable pairs balanced and twisted along the full run.
- Never increase line-to-chassis capacitance beyond the rated part class.
Common questions
Can one filter handle both modes?
Commercial mains filters usually combine both stages in one can, but the sections are separate inside. The specification sheet lists common-mode and differential-mode attenuation as two curves.
Does twisting a cable reduce common-mode noise?
Twisting mainly cancels differential-mode radiation and pickup. It improves balance, which indirectly limits mode conversion, but it does not present impedance to common-mode current.
The practical takeaway
Diagnose the mode before you buy a part. A clamp meter around the whole bundle answers the question in one reading, and that answer decides whether you need a common-mode choke, a differential filter, or better cable balance. Fitting the wrong one wastes money and leaves the interference exactly where it was.
References and further reading
- FCC: electromagnetic compatibility and radio frequency fundamentals
- NFPA: how the National Electrical Code treats grounding and bonding
- OSHA: wiring design and protection requirements including grounding
Numerical scenarios are illustrative unless identified otherwise. Follow the exact product instructions; component ratings and local installation requirements can differ.
