USB-C Charger Noise in RF Stations: Measure CM and DM Separately
USB-C Charger Noise in RF Stations: Measure CM and DM Separately
A USB supply can meet its DC-voltage specification and still disturb a receiver. The useful diagnosis separates differential ripple, common-mode cable current, direct radiation and receiver coupling—at declared ports, loads and frequencies.
“The charger is noisy” is a symptom, not a mechanism. Identify the affected receiver, frequency, port and cable; reproduce the interference; then measure the mode that carries it. A ferrite, an LC filter or an isolated supply is useful only when it interrupts the measured coupling path without violating USB operation, current, temperature, EMC or electrical-safety requirements.
RF.Guru measurement boundary: the USB power products discussed below are evaluated with controlled TEM-cell emissions and coupling measurements, together with separate conducted-noise measurements. Results belong to the declared load, supply, cable, port, fixture, frequency span, detector and test plane; the installed station should still be checked with its actual receiver and wiring.
Four Different Observations Get Called “Noise”
| Observation | Possible mechanism | Useful first measurement |
|---|---|---|
| Buzz or whine from the adapter | Mechanical vibration of magnetics or capacitors, often load/control-mode dependent | Acoustic spectrum and load sweep; do not infer RF emissions from sound alone |
| Ripple or spikes between VBUS and return | Differential-mode switching current, load transients, cable impedance or probing artefact | Bandwidth-defined differential voltage at the device port with a low-inductance probe connection |
| Current on the complete USB cable relative to the room | Common-mode displacement current through transformer, PCB, shield, chassis and environmental capacitances | Calibrated or repeatable RF current probe around the entire cable |
| Receiver noise that changes when equipment moves | Direct radiation, cable radiation, near-field coupling, shield/reference current or receiver overload | Near-field scan, cable-current map and controlled distance/routing tests |
A charger can exhibit one, several or none of these. Audible silence does not prove RF cleanliness. A clean VBUS-to-return trace does not prove low common-mode current. A raised receiver noise floor does not prove the product fails a legal emission limit.
Define DM and CM at the Cable Boundary
For the two power conductors, differential current flows out on VBUS and returns on the power-return conductor. In the ideal differential mode the signed currents cancel when a current probe encloses both conductors:
iDM ≈ (iVBUS − ireturn)/2
iCM ≈ iVBUS + ireturn for a two-conductor screening definition
A real USB cable also contains shield, data and configuration conductors. A whole-cable probe responds to the net longitudinal current on everything it encloses. Current can return through the USB shield, another connected cable, receiver ground, mains protective earth, transformer capacitance, a user's body and distributed capacitance to the room. The reference structure must therefore be drawn, not implied by the word “ground.”
Differential fields tend to cancel when the outbound and return paths are close, but cancellation is not perfect. Common-mode current can radiate efficiently when the cable and its return structure form an electrically significant antenna, yet cable length alone does not set radiation. Current magnitude and distribution, shield termination, routing, nearby conductors and the environment decide.
How an Isolated Charger Still Passes RF Current
Galvanic isolation removes the intended conductive path between mains and SELV output. It does not remove capacitance. Texas Instruments' current transformer-EMI guidance shows the central mechanism: a switching waveform on the primary side drives common-mode current through primary-to-secondary parasitic capacitance. The first-order displacement-current relation is:
iCM(t) ≈ CPS × dV(t)/dt
The complete current also depends on every return impedance. Transformer construction, switch-node waveform, Y-capacitor network, PCB, enclosure, cable, load and station grounding all matter. A Class II two-wire adapter can have a floating low-voltage output and still couple RF through its barrier and surroundings.
Faster edges contain energy at higher harmonics and more readily excite parasitic resonances. This is a design challenge, not an indictment of GaN. A well-designed GaN adapter can be quieter than a poor silicon design; topology, controlled slew rate, transformer geometry, snubbing, shielding, layout and filtering determine the measured result.
Higher power and smaller size do not mathematically imply more interference. They can make thermal density, switching-edge and filter-volume trade-offs harder, but the answer is the product's measured emissions under the relevant load and cable conditions.
What USB-C Power Delivery Changes—and What It Does Not
USB-IF's current document library lists USB Power Delivery Specification Revision 3.2 Version 1.2.2, dated May 2026. USB-IF also documents Extended Power Range operation up to 240 W with suitable sources, sinks and full-featured cables. That maximum is an ecosystem capability, not a rating for every USB-C connector, cable or charger.
USB Type-C attachment and USB-PD communication occur on the configuration channel. Negotiated voltage and load changes can move a converter among operating modes and therefore change its spectrum. The PD protocol is not automatically the dominant interferer, and a changing spur is not proof that CC signalling caused it. Compare controlled power contracts and loads while measuring the same port.
A 5 V-only inline power filter may not pass CC or USB data and may not support PD negotiation. If placed after a source that provides default 5 V, it must still meet the load's voltage-drop, inrush, current and connector requirements. Never assume a USB-C-shaped interface supports every Type-C or PD feature.
Compliance Is a Boundary, Not a “Radio Quiet” Guarantee
In the EU, the current consolidated EMC Directive 2014/30/EU requires equipment to meet essential electromagnetic-compatibility requirements. Mains equipment within the voltage scope also engages the Low Voltage Directive 2014/35/EU, subject to the exact product and applicable legislation. IEC 62368-1:2023 provides a hazard-based safety framework for external power supplies within its scope. CISPR 32:2015 with Amendment 1:2019 defines reproducible emission requirements for multimedia equipment within its scope.
Those are not interchangeable statements:
- meeting a standardized emission limit does not mean zero energy at every amateur frequency or zero disturbance beside a sensitive antenna;
- a receiver complaint does not by itself prove legal non-compliance, because test configuration, detector, bandwidth, distance, limits and uncertainty matter;
- a CE marking is not an independent performance award and does not reveal the product's noise spectrum;
- low measured RF noise does not establish electrical safety, insulation class, creepage, clearance, touch current, fire safeguards or conformity documentation.
Mains-safety boundary: do not open a charger or probe its primary with an ordinary grounded oscilloscope. Do not modify barrier capacitors, transformer screens, insulation or mains filtering as an RFI experiment. Use intact products with applicable documentation and competent safety assessment.
A Receiver Test Is Useful Screening, Not Proof
A portable receiver or SDR can quickly reveal correlation:
- Choose a frequency without a strong wanted signal and record receiver settings, antenna and baseline.
- Run the same load from the suspect supply, then from a suitable battery or reference supply. Keep cables and device state fixed.
- Move the receiver probe—not the entire station—near the adapter body and along the cable.
- Add one known ferrite configuration or filter and repeat at the same load and geometry.
- Change cable route, length or shield connection only one variable at a time.
A reduction after adding a whole-cable ferrite supports the hypothesis that common-mode current on that cable contributes. It does not prove all remaining noise is common mode, identify the original source inside the charger, or quantify compliance. The ferrite also changes the external network and can move a resonance.
Measurements That Separate the Paths
Differential output voltage
Measure VBUS-to-return at the device end under declared steady, burst and transient loads. State oscilloscope bandwidth, probe type, termination and whether a ground spring or true differential probe was used. A long probe ground lead can manufacture spikes. Record the DC drop as well as the AC spectrum because series filtering can make a 5 V load fail before it improves RF reception.
Whole-cable common-mode current
Place an RF current probe around the entire USB cable and measure several positions. Use the same fixture, orientation, receiver bandwidth and load. A single low point may be a current minimum. If absolute current is required, calibrate probe transfer impedance and include cable placement and instrument uncertainty.
Conducted and radiated emissions
Use the applicable artificial network, detector, bandwidth, load and cable arrangement for a standards result. A bench spectrum analyser trace is pre-compliance evidence only. For direct radiation and system coupling, use a repeatable near-field scan or appropriate chamber/site method, then confirm the receiver improvement in the installed station.
Noise coupling into the receiver
Measure at the receiver input when possible. An antenna-port spectrum, sensitivity/desense check or substitution source separates a genuine RF input from audio, USB-data or power-rail coupling after the RF front end.
Choose the Remedy From the Measured Mode
| Measured problem | Candidate control | Required qualification |
|---|---|---|
| Differential ripple or spikes at the load | LC/π filtering, feed-through capacitance, local decoupling, lower-impedance cable | Load stability, inrush, voltage drop, ripple current, resonance and transient response |
| Whole-cable common-mode current | Common-mode choke, ferrite sleeve/clamp, shield/reference redesign, lower-barrier-capacitance supply | Complex common-mode impedance over the affected band, DC current, imbalance, temperature and parasitic bypass |
| Direct adapter radiation | Distance, orientation, shielding/enclosure redesign or a different supply | Repeatable near-field and receiver measurements; do not compromise cooling or insulation |
| Receiver overload or internal coupling | Receiver filtering, port isolation, cable routing, grounding/bonding correction | Signal-level, linearity, safety and complete-system tests |
Filter placement is not universally “nearest the charger” or “nearest the device.” Put the barrier where it prevents the noisy cable section or protects the sensitive port, then verify both sides. Minimizing the downstream unfiltered length often helps, but source impedance, load impedance and parasitic return paths can reverse the intuitive result.
A USB filter has electrical limits too
- input/output voltage, continuous and inrush current, connector temperature and contact resistance;
- DC resistance and load-end voltage at worst current;
- common-mode impedance and differential insertion loss versus frequency and termination;
- capacitor ripple current, resonance, damping and converter stability;
- shield/data/CC continuity and permitted USB or PD modes;
- transient, ESD, insulation and enclosure requirements appropriate to the port.
RF.Guru USB Power Options
Inline 5 V USB power filter: the RF.Guru 5 V USB-C to USB-A filter is intended to suppress common-mode and differential-mode RF energy on a 5 V power path. It is an EMI filter rather than a safety-rated galvanic-isolation barrier. The source must provide default 5 V, or a suitable USB-PD trigger must establish 5 V upstream; voltage drop, load current, inrush and connector temperature remain part of the application check.
Mains-isolated 5 V supply: the RF.Guru isolated 230 VAC to 5 V/3 A supply combines mains-to-output isolation with filtering and protection for installations that need both conversion and RF-noise control. Its emissions and coupling are evaluated in controlled TEM-cell setups, while conducted noise is measured separately at defined ports and loads. Use the current product documentation for electrical, safety and conformity limits.
Choose by coupling path: use the inline filter when the existing 5 V source is suitable but RF current on the USB power connection needs control. Use the isolated supply when the power source itself and the mains-to-output path must be addressed. Confirm the result in the installed station because cable routing, load state, receiver susceptibility and external return paths can change the outcome.
A Repeatable Station Workflow
- Record the symptom. Frequency range, receiver mode/bandwidth, antenna, signal level, affected port and device state.
- Freeze the geometry. Mark charger, cable, receiver and bonding positions.
- Substitute power. Compare a suitable reference supply or battery without changing the load or signal path.
- Map current and fields. Probe the complete cable at several points and scan the adapter/device separately.
- Measure VBUS differentially. Use declared bandwidth and a suitable load profile.
- Change one control. Ferrite, filter, cable, route, distance or supply—one at a time.
- Check operation. Voltage margin, inrush, USB/PD function, temperature and faults must still pass.
- Verify the receiver outcome. Repeat the original sensitivity/noise-floor test and retain raw data.
- Escalate correctly. Use a competent EMC laboratory for compliance claims and qualified safety evidence for mains equipment.
Conclusion
USB-C did not repeal EMC physics. Switching edges, transformer capacitance, cable current and environmental return paths can create interference; topology and careful design can also control it. Separate DM voltage, CM current, direct radiation and receiver coupling, then qualify the remedy at the exact voltage, current, load, cable, frequency and safety boundary.
Primary Technical Sources
- USB-IF: current USB Power Delivery specification library and USB Power Delivery overview
- Texas Instruments: How to characterize a power transformer for EMI performance
- Texas Instruments: EMI considerations for isolated AC-DC power supplies
- Würth Elektronik ANP044e: common-mode and differential-mode interference in switched converters
- IEC CISPR 32:2015 with AMD1:2019: multimedia-equipment emission requirements
- IEC 62368-1:2023: safety requirements for AV/ICT equipment and external supplies within scope
- EU EMC Directive 2014/30/EU and EU Low Voltage Directive 2014/35/EU
Mini-FAQ
- Is USB Power Delivery itself usually the RF-noise source? Not necessarily. PD changes contracts and operating points, but converter switching, parasitic coupling, cable current and the load must be measured before assigning the source.
- Does a clean VBUS ripple trace prove that the USB cable is RF quiet? No. VBUS-to-return is a differential measurement. Common-mode current can flow on the complete cable relative to its environment while differential ripple remains small.
- Does a clamp ferrite prove the problem is common mode? A repeatable reduction supports cable common-mode current as a contributor, but the ferrite also changes the external network. It does not locate every source or quantify compliance.
- Should a USB filter go near the charger or near the device? Place it where measurement shows it interrupts the noisy cable segment or protects the sensitive port, then verify both sides, voltage drop, current, temperature and USB operation.
- Does CE marking mean a charger cannot disturb an HF receiver? No. Compliance uses defined tests and limits; it does not mean zero emissions at every amateur frequency or guarantee compatibility in every close-proximity installation.
- Can a mains-isolated supply still carry common-mode RF to its output? Yes. Parasitic capacitance and intentional EMI components can pass displacement current across the isolation barrier even though there is no intended conductive path.
- Can every USB-C inline filter pass a negotiated PD voltage? No. A 5 V power filter may omit CC and data paths. Verify its connector wiring, voltage, current, inrush and PD compatibility before use.