Sleeved and Clip-On Ferrites Are Not for QRO

Let’s cut through the marketing: at QRO, ferrite shape is not a power rating. A tiny, unknown clip-on is usually a poor choice for a high-power HF feed line—but a documented, large clamp-on with enough turns can be a legitimate choke. A toroid can also fail if the material, winding, impedance, or thermal design is wrong.

The engineering question is not “clip-on or toroid?” It is: what is the measured common-mode impedance across the operating bands, how much common-mode current flows, and does the complete assembly remain thermally stable at the intended duty cycle?

Scope: this article concerns 1:1 coaxial common-mode chokes—sometimes called current baluns or feed-line isolators—used on HF transmitting systems. Transformer baluns and differential-mode filters require a different analysis.

What a Coaxial Choke Actually Sees

In normal differential-mode operation, RF current flows out on the coax centre conductor and returns on the inner surface of the shield. Those equal and opposite currents produce magnetic fields that largely cancel outside the coax. A ferrite core placed around the entire cable therefore does not normally process the transmitter’s full differential power.

The ferrite responds mainly to the unwanted current flowing on the outside of the shield. That is common-mode current. It can make the feed line radiate, alter the antenna pattern, carry noise into the antenna, and bring RF back into the shack.

For a matched 50 Ω system, 400 W corresponds to about 141 V RMS and 2.83 A RMS. At 1.5 kW, the line carries about 274 V RMS and 5.48 A RMS. Those numbers matter for the coax, connectors, voltage spacing, and fault margin—but they are not automatically the current and voltage exciting the ferrite. Antenna balance, feed-line routing, choke location, load impedance, and common-mode path determine that excitation.

SWR raises local stress. With 1.5 kW of forward power and a 2:1 SWR, |Γ| = 0.333 and the idealised voltage maximum is approximately 365 V RMS, or 516 V peak, before line loss is considered. That is why a “legal-limit” claim without a stated SWR and choke location says very little.

The Two Numbers That Matter: R and X

A choke’s common-mode impedance is complex and varies with frequency:

The reactance X stores and returns energy. The resistance R represents loss that damps common-mode resonances. A high, predominantly resistive impedance over the operating band is often the most predictable broadband solution because it suppresses current without simply moving a feed-line resonance elsewhere.

The heating relationship needs one important qualification. If common-mode current were fixed, more resistance would dissipate more heat. In the real circuit, however, the choke’s impedance reduces that current. If the choke dominates the common-mode path and is mainly resistive, then approximately:

That is why an under-designed 500 Ω choke can run much hotter than a 5 kΩ choke in the same system: the stronger choke reduces the current far more effectively.

Why Small One-Pass Clip-Ons Usually Disappoint

The main limitation is often not mysterious saturation. It is simply insufficient impedance. One pass through a core is one turn, and a small HF suppression core may provide only tens to a few hundred ohms. That can help on a USB, control, DC, or audio cable while being far short of what is needed to control a transmitting antenna’s common-mode circuit.

For perspective, Fair-Rite’s unusually large 289 g mix-31 split core, part 2631181381, is specified at a typical 67 Ω at 1 MHz, 193 Ω at 10 MHz, and 288 Ω at 25 MHz for one turn. Even this large device would require roughly 26 identical one-pass cores to reach 5 kΩ at 10 MHz if their impedances added ideally. A small no-name clip-on will generally offer much less—and without a known material or impedance curve, there is no defensible calculation at all.

Adding turns can be far more effective. Below the region where winding capacitance and resonance dominate, impedance rises approximately with the square of the turns:

That relation is only a starting estimate. More turns also add inter-turn capacitance, change the winding geometry, and move the self-resonance. At upper HF, adding turns can eventually make performance worse. The actual assembly—not just the bare core—must be measured.

What the Split-Core Gap Really Does

A split core contains a small magnetic gap at its mating surfaces. Because air has far lower permeability than ferrite, the gap reduces effective permeability and therefore reduces low-frequency inductance and impedance. Fair-Rite’s analysis shows that the effect is especially strong in high-permeability materials and recommends clean mating surfaces and the smallest practical gap.

The gap does not inherently create an RF “hotspot.” Heating comes from magnetic loss, conductor and dielectric loss, poor heat removal, or excessive electrical stress. The plastic shell may restrict cooling, and a small core has little thermal mass, but those are thermal-design issues—not proof that every split core is unusable.

Heating, Saturation, and Curie Temperature Are Different Things

Ferrite has a complex permeability: the storage component, μ′, contributes mainly to reactance; the loss component, μ″, contributes mainly to resistance. That loss is useful because it turns unwanted common-mode RF into heat—but the choke must have enough total impedance and enough thermal capacity to do so safely.

Heat alone does not prove saturation. Saturation is a nonlinear magnetic condition reached when flux density becomes too high. In a coaxial choke, the differential-mode fields largely cancel, so saturation is driven mainly by the remaining common-mode ampere-turns. Core loss and cable loss can produce substantial heating before saturation is reached.

Temperature still matters. Fair-Rite lists an initial-permeability temperature coefficient of about 1.3%/°C from 20–70°C and a Curie temperature above 130°C for mix 31. Mix 43 has similarly temperature-dependent permeability and a Curie temperature above 130°C. These are material properties—not recommended operating temperatures for a finished choke. The coax dielectric, jacket, connector, adhesive, enclosure, and mechanical stresses may set much lower limits. A design should never rely on approaching the Curie point.

For example, Fair-Rite specifies 0–85°C operation for its suppression parts in plastic cases and warns that the case can soften and deform above 85°C. Uncased ferrite may tolerate more, but the manufacturer also warns that rapid temperature change can crack a core through thermal shock. The limit of the lowest-rated part governs the assembly.

As a choke warms, its impedance can shift. Severe heating can soften a jacket, distort a dielectric, damage a connector, crack a core, or push the ferrite toward a region where its magnetic properties deteriorate. The scientifically correct conclusion is therefore not “hot ferrite always saturates,” but: a temperature-dependent choke must be tested as a complete electromechanical assembly.

Inductive Reactance Is Not Automatically a Failure

Every practical choke contains resistance and reactance. Below its resonance it may look mainly inductive; near a broad resonance it may look mainly resistive; above resonance, winding capacitance may dominate. Inductive reactance does not inherently “distort” a properly operating 50 Ω differential signal, because the choke is acting on the separate common-mode circuit.

The problem with a highly reactive choke is predictability. Its reactance can resonate with the feed line’s common-mode capacitance and inductance, shifting current peaks rather than damping them. For a multiband HF installation, a high resistive component over the intended bands is usually a more robust design target than a large magnitude of reactance at one spot frequency.

There Is No Honest “2 kW” Rating Without Test Conditions

A useful choke rating must state, or allow you to determine:

• frequency range and the measured R, X, and |Z| across it;
• antenna/load condition and expected common-mode voltage or current;
• continuous or average duty cycle;
• maximum SWR and whether the choke is before or after a tuner;
• coax, connector, winding radius, core count, and core material;
• ambient temperature, enclosure, airflow, and allowable temperature rise.

“1–2 kW ICAS” without those conditions is a label, not an engineering specification. High SWR or a tuner that transforms to a high impedance can raise RF voltage well above the matched-line value. A choke that is comfortable on SSB into a balanced 50 Ω load may overheat during long RTTY transmissions on a badly unbalanced antenna.

Choosing Material Without Folklore

A manufacturer’s material frequency range describes the ferrite, not the completed multi-turn choke. Core dimensions, number of cores, turns, turn spacing, coax, and parasitic capacitance determine the final impedance curve.

A Defensible QRO Design and Test Procedure

1. Define the electrical case. List every operating band, maximum power, SWR, tuner position, modulation mode, transmit duty cycle, ambient temperature, and enclosure condition.
2. Set a common-mode impedance goal. A widely used engineering starting point is about 5 kΩ of predominantly resistive impedance on each operating band. It is not a formal standard or a guarantee. QRO, high common-mode voltage, or a badly unbalanced antenna can require more impedance and more core mass.
3. Use identified materials. Record the manufacturer, part number, mix, core dimensions, number of cores, coax type, turns, and winding layout. “Black ferrite” is not a specification.
4. Do not treat PTFE as magic. PTFE coax adds temperature and voltage margin, but quality PE-dielectric coax can also handle high power when used within its data-sheet limits. Respect connector ratings, bend radius, conductor loss, and local compression where the cable passes through the core.
5. Measure the finished choke. Measure common-mode R and X across the full frequency range, preferably with a calibrated two-port method suitable for high impedance. In a suitable series-through fixture, complex impedance can be derived from Z = 2Z0(1 − S21) ÷ S21. A one-port analyser near several kΩ can be dominated by fixture error. Also verify low differential-mode insertion loss and acceptable 50 Ω return loss.
6. Measure common-mode current if possible. An RF current probe on the outside of the coax, used at repeatable locations, reveals whether the choke is reducing the current that matters.
7. Run a thermal soak. Start at low power, then increase in controlled steps using the real antenna/load and the highest-average-duty operating mode. A dummy load can check the differential path, but it does not reproduce the antenna installation’s common-mode circuit. Monitor the core, cable, and connectors until temperature stabilises at each step. Repeat on the bands where impedance is lowest.
8. Stop on drift. End the test if temperature keeps climbing, the jacket softens or smells, a connector discolours, the core cracks, or SWR changes with temperature. Find the cause before applying full power again.
9. Re-test cold. After the assembly cools, repeat the impedance sweep and inspect it for permanent electrical or mechanical change.

What a Changing VSWR Is Telling You

A correctly built coaxial common-mode choke should add little loss or mismatch to the 50 Ω differential path. But installing one can still change the measured feed-point impedance if the feed line was previously acting as part of the antenna. That immediate change is not necessarily choke failure; it may be evidence that common-mode current was significant.

A VSWR that drifts as the assembly heats is different. It points toward a temperature-dependent ferrite response, cable deformation, connector trouble, arcing, or another differential-path fault. Shut down and inspect. “The meter went crazy” is a symptom, not a diagnosis.

Bottom Line

Small, unknown, one-pass clip-ons are usually the wrong tool for a QRO HF transmission line. Their common-mode impedance is typically too low, their material is often undocumented, and their thermal margin is difficult to defend.

But the scientifically accurate rule is stronger than “clip-ons bad, toroids good”:

Engineering References

1. Fair-Rite: 31 Material Data Sheet — material properties, temperature coefficient, Curie temperature, and intended suppression use.
2. Fair-Rite: Mix-31 Split Round Cable Core 2631181381 — dimensions, mass, and measured one-turn impedance.
3. Fair-Rite: Low-Frequency EMI Suppression and Split-Core Air Gaps — turns, effective permeability, gap effects, bias, and temperature dependence.
4. Fair-Rite: 43 Material Data Sheet and 52 Material Data Sheet — manufacturer descriptions and magnetic properties.
5. Jim Brown, K9YC: A Ham’s Guide to RFI, Ferrites, Baluns, and Audio Interfacing — common-mode choke theory, the 5 kΩ design goal, power dissipation, and measured constructions.
6. ARRL Contest Update: Ferrite Choking Materials — discussion of measured HF choke behaviour and mix-31 construction.
7. Fair-Rite: General Considerations for Suppression and Storage and Operating Conditions — material ranges, bias and temperature derating, and plastic-case limits.
8. Keysight: Impedance Measurement Methods — suitable impedance ranges, calibration, and the two-port series-through equation.

Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.

Questions or measurements to share? Feel free to contact RF.Guru.