Mains-Cable Ferrites: Common Mode, Differential Mode and Protective Earth
Mains-Cable Ferrites: Common Mode, Differential Mode and Protective Earth
A ferrite around an intact power cord can be a useful RF diagnostic and cure, but it is not the same circuit as a mains inlet filter or a manufacturer-qualified earth-line choke.
The safe question is not “Which ferrite mix always fixes the mains?” It is: which current is flowing, through which conductors, at what frequency, and where does it close its return path? An external ferrite around the complete, undamaged cord mainly impedes the cable’s net external current. A certified inlet filter can also treat differential noise. A choke placed only in protective earth is a safety-critical component choice, not a casual shack modification.
Mains-safety boundary: never disconnect, switch, bypass or deliberately weaken protective earth to solve RFI or hum. Do not insert a homemade component in PE or modify fixed wiring or mains equipment unless the work is designed, installed and verified by a competent person under the applicable equipment and installation rules. A qualified filter module with an earth-line choke is not evidence that any improvised PE choke is safe.
Three Arrangements That Must Not Be Confused
1. A ferrite around the complete flexible cord
A clamp-on ferrite or a large core around the intact cord surrounds line, neutral and PE without making electrical contact. The core responds to their algebraic current sum. Normal 50/60 Hz load current leaves through line and returns through neutral, so its magnetic field largely cancels. Any net current that returns through the environment, cable shields, other bonds or stray capacitance does not cancel and is the external or common-mode component the ferrite can impede.
This is also why the phrase “all three wires carry equal in-phase common-mode current” is too tidy for a real Class I appliance. The division among L, N and PE depends on filter capacitors, chassis coupling and other connected cables. What the enclosing core measures is the residual sum, not three assumed equal currents.
2. A mains inlet EMI filter
A complete filter may combine a common-mode choke in L and N, an X capacitor across L–N, Y capacitors from the supply conductors to PE, differential inductance, damping and a discharge resistor. Those parts have different jobs. The common-mode section works against noise referred to chassis or the environment; the X/series network treats differential noise between L and N; Y capacitors improve common-mode attenuation but create displacement current to PE.
That safety/performance package is why “add a capacitor” is not acceptable mains advice. IEC 60939-3:2024 covers passive EMI filter units for which safety tests are appropriate, while IEC 60384-14:2023 with Amendment 1:2025 specifies capacitors intended for connection to the supply mains. Rated voltage, current, temperature, leakage or touch current, discharge energy, insulation, creepage, clearance, fault behavior and the end-equipment standard all matter.
3. A qualified earth-line choke inside a filter module
The Schurter KFA and Schaffner FN9233E demonstrate that a manufacturer can add controlled RF impedance in the earth line while retaining the required protective function and approvals for specified variants. The FN9233E, for example, is a complete 250 VAC inlet filter with an integrated earth-line choke and defined current, high-potential, temperature and approval data.
That earth-line element is not galvanic isolation and it does not open PE. It targets high-frequency current. It should not be advertised as a universal cure for 50/60 Hz audio hum, because hum may come from magnetic coupling, leakage current, signal-reference wiring or voltage drop around a loop. The complete system and its fault path still have to meet the applicable safety requirements. IEC 60364-5-54 addresses earthing arrangements and protective conductors for low-voltage installations.
Common Mode and Differential Mode Need Different Tests
| Observed path | Useful first measurement | Likely treatment | What it does not prove |
|---|---|---|---|
| Net RF current on the whole cord | RF current probe around the complete cord | External cable ferrite; improve chassis/filter layout | That the noise originated inside the PSU |
| Noise voltage between L and N | Properly isolated LISN or differential probe in a qualified setup | Rated DM network, commonly X capacitance plus series impedance | That a whole-cord ferrite will provide enough DM loss |
| RF current on PE or interconnected chassis paths | Current measurements on intact PE and the other bonded cables | Correct bonding/layout; where justified, a qualified filter with earth-line choke | Permission to interrupt or improvise the safety conductor |
| Receiver noise that changes when a cord is moved | A/B test with a clamp-on ferrite and band-by-band current/noise readings | The measured treatment at the identified end of the cable | Compliance, causation or broadband performance by itself |
A whole-cord ferrite is mainly a common-mode component because ideal L–N differential fields cancel inside it. Real cancellation is not perfect, so the core may show some differential effect through leakage and asymmetry. Conversely, a conventional two-winding common-mode choke has leakage inductance that contributes some differential attenuation. “Ferrite means CM only” and “a mains filter stops everything” are both oversimplifications.
Ferrite Mix Is Only One Variable
Material labels do not specify the installed impedance. Fair-Rite describes material 31 as an EMI-suppression material spanning roughly 1 to 500 MHz, while its suppression guidance places material 43 broadly in the 20 to 300 MHz region. Those are material/application ranges, not guarantees that “31 is for these amateur bands” and “43 is for those bands.”
Core dimensions, the exact part, number of passes, winding spacing, cord diameter, source and load impedances, RF current and parasitic capacitance determine the result. More turns often increase low-frequency impedance roughly with the square of turns in the inductive region, but also increase inter-turn capacitance and move the high-frequency peak downward. Stacking unlike materials may help, do nothing useful, or create a shape that still misses the troublesome frequency. Measure the finished arrangement.
The practical figure is residual current reduction:
Reduction (dB) = 20 log10(|Ibefore| / |Iafter|)
Record frequency, cord position, equipment state, transmit power, probe calibration and noise-floor limit. Ferrite impedance alone does not determine attenuation because the source and return-path impedances are part of the circuit.
One Choke May Be Enough—or the Wrong Component
If one external mode dominates and a clamp-on adds enough impedance at the right point, one ferrite can solve the symptom. If the disturbance is differential, bypasses the ferrite through a data cable, couples through the enclosure, or comes from several frequencies with different path impedances, the same ferrite may do very little.
A mains filter design commonly combines a common-mode choke with X and Y capacitors because CM and DM paths are different. But filter insertion loss from a 50 Ω test graph is not automatically the attenuation in a high-impedance real mains network. Placement, chassis bonding, source/load orientation and separation of the dirty and clean sides can dominate high-frequency performance.
A Safe Diagnostic Sequence
- Define the symptom and frequency. Record which receiver band, transmitter mode, load state and connected cables reproduce it.
- Measure without exposing mains conductors. Clamp an RF current probe around the complete insulated cord. Compare with probes on coax, USB, Ethernet and other external cables.
- Try a reversible external ferrite. Keep the cord intact, respect bend radius and strain relief, prevent the core from abrading the jacket, and remeasure current and the actual symptom.
- Move the ferrite deliberately. The equipment end and supply end are different RF nodes. A successful location reveals something about the coupling path.
- If DM or internal filtering is implicated, stop at the enclosure boundary. Select a rated, approved filter for the equipment category and have competent personnel integrate it with correct PE bonding, fusing, insulation, leakage-current and thermal checks.
- Verify under worst intended operation. Check every relevant amateur band and equipment load, plus ferrite temperature, receiver noise, transmitted RFI and safety continuity. Never declare success from SWR or one noise reading alone.
Engineering takeaway: enclose the whole intact cord when you want to impede its external common-mode current. Use a properly rated filter when differential noise or mains-connected capacitors are involved. Treat any series impedance in PE as a qualified safety design, not a field modification. Then verify the result with current and symptom measurements.
Primary Sources Checked
- IEC 60939-3:2024 — passive EMI filter units for which safety tests are appropriate
- IEC 60384-14:2023 with Amendment 1:2025 — capacitors for EMI suppression and connection to the supply mains
- IEC 60364-5-54 — earthing arrangements and protective conductors
- TE Connectivity/Schaffner FN9233E inlet-filter product data
- Schurter KFA power-entry-module data sheet
- Würth Elektronik mains-filter design guide
- Fair-Rite material 31 data and suppression-material guidance
Mini-FAQ
- Should PE pass through an external ferrite with line and neutral? If the ferrite surrounds the complete intact Class I cord, yes: all conductors in that cable pass through together. The core then responds to the net current leaving the cable rather than interrupting PE.
- May I add my own choke only in protective earth? No casual modification is justified. Use only a qualified arrangement designed for the equipment and verified against the applicable safety requirements by competent personnel.
- Will a whole-cord ferrite stop differential PSU noise? Not reliably. Its main action is on net external common-mode current. Differential noise may require a rated L–N filter network and correct layout.
- Is mix 31 always low HF and mix 43 always upper HF? No. Those are broad material tendencies. The exact core, passes and parasitics set installed impedance, so measure the finished choke at the frequencies of interest.
- Can the ferrite become hot? Yes. RF common-mode energy is partly dissipated in a lossy ferrite. Significant temperature rise is evidence to reduce the current or redesign the path, not simply add more transmit power.
- Is one mains-cable choke always insufficient? No. One correctly placed choke may be enough for one dominant external mode. It will not cure a different mode or a bypass path merely because it is on the power cord.