Ferrite Chokes and Cores
Ferrite is the most important material in the interference-fighting toolkit. It is a ceramic compound that looks unremarkable — a dull gray or black ring, tube or tiny surface-mount bead — but its magnetic properties make it uniquely useful for suppressing interference at radio frequencies. Snap a ferrite clamp onto a USB or power cable and you can cut the noise it carries by 10 to 20 dB. Wind a few turns of a cable through a ferrite toroid and you have a common-mode choke that can reduce cable-borne interference by 20 to 40 dB. Put a ferrite bead in series with a supply rail on a circuit board and you keep high-frequency noise out of a sensitive circuit.
The challenge is that ferrite is not a single material. There are dozens of ferrite formulations, each optimized for a different frequency range, and using the wrong mix for your application can produce no useful suppression at all. A ferrite core optimized for 1 MHz will be nearly useless at 30 MHz. A core optimized for VHF will provide only weak suppression at 3 MHz. Getting the right mix for the right frequency is what separates effective ferrite use from expensive frustration.
What Is Ferrite and Why Does It Work?
Ferrite is a ceramic magnetic material made by combining iron oxide (Fe₂O₃) with one or more other metal oxides — typically zinc, manganese, nickel, or cobalt — and sintering the mixture at high temperature into a hard, dense ceramic. The result is a material with a very high relative magnetic permeability (µr), which is the measure of how much better it concentrates magnetic flux compared to empty air. Depending on the formulation, ferrite permeability ranges from about 40 (for high-frequency nickel-zinc mixes) to over 15,000 (for some low-frequency manganese-zinc mixes).
When a wire passes through a ferrite core, the core adds inductance to that wire. The inductance is proportional to µr multiplied by N² (the square of the number of turns) and divided by the path length of the magnetic circuit. More turns, or a higher permeability core, gives more inductance. At RF frequencies, this inductance appears as impedance — specifically, a complex impedance with both an inductive (reactive) component and a resistive (lossy) component. The resistive component is the key to understanding why ferrite chokes are effective.
The lossy behavior of ferrite at RF frequencies is called magnetic loss, and it is intentional. Ferrite is not a lossless magnetic core like a high-quality inductor core at its operating frequency. Instead, for common-mode choke applications, we want the ferrite to be lossy — to convert some of the common-mode RF energy into a small amount of heat rather than reflecting it back as a reactive impedance. A purely reactive choke would present high impedance but would also create a resonance with parasitic capacitances that could actually make things worse at certain frequencies. A lossy ferrite choke presents broad-band resistive impedance that simply absorbs common-mode energy over a wide frequency range without creating resonances.
The crucial advantage of a ferrite common-mode choke is its selectivity. When you wind a cable through a ferrite core, the differential-mode signal in the cable creates equal and opposite magnetic fluxes in the core — they cancel exactly, and the core has no effect on the wanted signal. Only common-mode current, flowing in the same direction on both conductors simultaneously, creates net flux in the core. The ferrite therefore impedes only the common-mode current, leaving the wanted signal completely unaffected. This is not a theoretical idealization — it is a real measurable property of the device, and it is why you can add a choke to a USB, audio or coaxial cable and see no change in the signal while the common-mode noise drops by 20 dB.
Common ferrite core types for interference suppression. The mix number determines the frequency range; the physical size determines how much cable can be wound through and how much power the core can handle.
View LargerFerrite Mix Numbers and Frequency Ranges
Ferrite materials are identified by "mix numbers" — standardized designations assigned by manufacturers Fair-Rite Products and Amidon Associates, two of the main suppliers to hobbyists and small manufacturers. The mix number tells you the ferrite formulation and, by implication, the frequency range where it is most effective as a common-mode choke. Using the wrong mix is a common mistake that leads to chokes that do nothing.
The underlying reason for different frequency performance is the magnetic loss mechanism. At low frequencies, ferrite cores can magnetize and demagnetize with each RF cycle without dissipating significant energy. As frequency increases, the magnetization cannot keep up and the material starts to lag behind — this lag shows up as magnetic loss (the resistive component of impedance). Each ferrite formulation has a specific frequency region where this loss is maximized and where it therefore provides the most effective suppression. Below that range the material is too efficient (too reactive, not lossy enough); above that range the permeability collapses and the material provides little impedance at all.
| Mix Number | Main Components | Relative Permeability (µr) | Best Frequency Range | Typical Use |
|---|---|---|---|---|
| Mix 31 | Manganese-zinc (MnZn) | ~1500 | 1–300 MHz (peak loss 25–250 MHz) | Cable common-mode chokes from 1 to 30 MHz; broadest low-frequency coverage; good all-round choice for power and signal cables |
| Mix 43 | Nickel-zinc (NiZn) | ~850 | 10 MHz–1 GHz (good 25–300 MHz) | General-purpose suppression from about 15 MHz to 300 MHz; very common in snap-on clamps for USB, video and data cables |
| Mix 61 | Nickel-zinc (NiZn) | ~125 | 200 MHz–2 GHz | VHF and UHF suppression; lower impedance than Mix 31 below 30 MHz but very effective above about 150 MHz |
| Mix 77 | Manganese-zinc (MnZn) | ~2000 | 0.1–10 MHz (peak loss 0.5–5 MHz) | Low-frequency suppression from about 0.5 to 5 MHz: switching-supply harmonics, AM broadcast interference, mains filters |
| Mix 52 | Nickel-zinc (NiZn) | ~250 | 10 MHz–500 MHz | Moderate suppression from 10 to 300 MHz; sometimes used where Mix 43 is unavailable |
The practical rule: use Mix 31 for the widest coverage from about 1 to 300 MHz, especially for switching-supply noise below 30 MHz; Mix 43 for general-purpose clamps on data and video cables where the problem is mostly above 20 MHz; Mix 77 for low-frequency noise around 0.5–5 MHz, such as switching-supply fundamentals and the AM broadcast band; and Mix 61 for VHF and UHF problems.
Mix 31 has become the most popular general choice for cable chokes because, in published measurements by Fair-Rite and independent engineers, it provides higher common-mode impedance from 1 to 30 MHz than any other readily available material while still working well into the VHF range. A well-wound FT-240-31 choke can provide 2,000 to 5,000 ohms of common-mode impedance from 2 to 30 MHz, compared with 500 to 2,000 ohms from an equivalent FT-240-43 choke.
Core Shapes and Sizes
Beyond the mix number, you need to choose the right physical form for your application. Ferrite is available in several shapes, and each has specific advantages.
Toroid ring cores are the most efficient shape for wound chokes. The toroid has no air gap — the magnetic path is completely through ferrite material — which gives it the highest inductance per turn and the most effective use of the ferrite material. The round shape also minimizes stray radiation from the core. For a choke on a cable carrying significant RF power, a toroid wound with the cable is the correct choice. Toroids also have the advantage that the winding direction does not matter for a choke application.
Split snap-on clamp cores are convenient because they can be added to existing cables without cutting or soldering. They consist of two ferrite halves in a plastic hinged housing that clips around the cable. The trade-off is that the split joint introduces an air gap that reduces effective permeability and therefore reduces impedance compared to an equivalent toroid. Snap-on clamps are useful for adding suppression to computer cables, USB cables, and audio cables where you do not want to disassemble connectors. Where heavy suppression is needed, a wound toroid is significantly more effective.
Binocular cores (also called two-hole cores or "balun cores") have two parallel holes side by side. Wire passes through the holes in a figure-8 pattern. They are often used for wideband RF transformers and some choke applications where the wire path through the core can be made shorter. For common-mode chokes, binocular cores are less common than toroids.
Rod and sleeve cores are cylindrical. They are used for suppression of interference on cables where the cable is threaded inside a ferrite tube or sleeve rather than wound around a toroid. The tube places the ferrite around the cable rather than through it, providing some common-mode suppression but less than an equivalent wound toroid.
Core size notation. The naming convention used by Amidon Associates is widely used by hobbyists and in application notes: FT-240-43 means "Ferrite Toroid, 2.40 inch outer diameter, Mix 43." The number between the FT prefix and the mix number is the outer diameter in hundredths of an inch. Common sizes are:
| Amidon Designation | Outer Diameter | Inner Diameter | Typical Use |
|---|---|---|---|
| FT-240 | 2.40 in (61 mm) | 1.40 in (36 mm) | Chokes on thick cables and coax; multi-turn mains and DC power chokes; higher power |
| FT-140 | 1.40 in (36 mm) | 0.90 in (23 mm) | Chokes on medium cables such as USB, Ethernet and RG-58 coax |
| FT-82 | 0.82 in (21 mm) | 0.52 in (13 mm) | Small chokes on thin cable and wire; low power |
| FT-50 | 0.50 in (13 mm) | 0.31 in (8 mm) | Single wires and small signal leads; low power only |
For chokes on cables carrying RF power of 100 watts or more, the FT-240 is the standard choice: it accommodates heavier cable and has the thermal mass to handle the heat from common-mode losses. The FT-140 suits medium cables and lower power. The FT-50 and FT-82 are for small signal wiring only — they are too small for high-power RF cables and will overheat.
How Many Turns? Impedance vs Frequency
The impedance of a ferrite choke increases with the square of the number of turns (N²) because each turn contributes inductance proportional to N, and inductance combines with the core's impedance characteristics to give total choke impedance. However, adding too many turns introduces parasitic capacitance between adjacent turns, which can short-circuit the choke at high frequencies, creating a self-resonance above which the choke loses effectiveness.
The impedance target for a common-mode choke depends on the impedance of the circuit it is fighting. The choke and the rest of the common-mode path form a voltage divider: the higher the choke's impedance compared with the rest of the path, the more it reduces the current. As a rule of thumb, aim for at least 500 ohms at the problem frequency, with 1,000 ohms or more being significantly better. With 1,000 ohms in a path whose other impedance is around 50 ohms, the common-mode current falls by a factor of about 20 (26 dB); with 5,000 ohms, by about 100 (40 dB). In high-impedance common-mode paths, such as a short cable at low frequencies, a choke helps much less.
An FT-240-43 toroid with 9 turns of RG-8X coaxial cable: according to published Fair-Rite and Amidon data, a single turn on this core gives approximately 14 µH inductance at low frequency. With 9 turns, inductance is 14 × 9² = 14 × 81 = 1,134 µH — but in practice the published impedance charts are more reliable than first-principles calculation, because they include the core losses that dominate at these frequencies.
Published data for this combination shows approximately:
- 1.8 MHz: ~200 ohms — marginal
- 3.5 MHz: ~600 ohms — adequate
- 7 MHz: ~1,200 ohms — good
- 14 MHz: ~2,000 ohms — excellent
- 28 MHz: ~1,500 ohms — good
This choke is excellent from about 7 to 30 MHz. For problems below 5 MHz, such as switching-supply noise, use Mix 31 or Mix 77 cores, or more turns.
Practical rules of thumb for common-mode choke winding:
- FT-240-31, 9–12 turns of RG-8X-sized cable: 1,000–5,000 ohms from 2–30 MHz — the best general choice
- FT-240-43, 8–10 turns: 600–2,000 ohms from 7–30 MHz
- FT-240-77, 10–14 turns: best below 5 MHz; less effective above 10 MHz
- Stacking two FT-240-31 cores: approximately doubles the impedance — use when maximum suppression is needed
Stacking cores — placing two or three toroids side by side and winding through them as a group — multiplies the impedance approximately in proportion to the number of cores. Two FT-240-31 cores wound together with 9 turns give approximately twice the impedance of one. This is useful when the common-mode drive is strong, such as on the cables of a high-power motor drive or inverter.
Winding Techniques for Cable Chokes
Winding a cable choke through a toroid is straightforward once you understand the conventions. The whole cable — jacket, shield and all its conductors — passes through the center hole of the toroid together. Each pass through the hole counts as one turn. The cable is not stripped or modified in any way; the ferrite acts only on the current that flows the same way in all the conductors.
Winding a 9-turn cable choke on an FT-240 toroid. Count each pass of the cable through the center hole as one turn. The winding direction does not matter; spacing the turns evenly and keeping the ends apart reduces parasitic capacitance.
View LargerStep-by-step procedure for winding a 9-turn choke on an FT-240 toroid:
- Allow about 2.5 meters (8 feet) of RG-8X-sized cable for 9 turns on an FT-240, with lead lengths left over. Thinner cables need less.
- Find the middle of the length and mark it lightly with tape. Start winding from the middle so the two free ends are about equal in length when you finish.
- Pass one end of the cable through the center hole of the toroid. This is turn 1.
- Bring the cable around the outside of the toroid and back through the center hole from the same side. This completes turn 2. Continue in the same direction.
- Count as you go: each pass through the center hole is one turn. Wind 9 turns, spaced evenly around the toroid — on an FT-240 they will occupy about two-thirds of the core.
- Secure the winding with cable ties so it cannot loosen, and fit connectors if the cable needs them.
A few important points about winding technique:
- Winding direction does not matter for a choke. Unlike a transformer, a choke works identically whether wound clockwise or counterclockwise.
- Spacing matters for parasitic capacitance. Tightly packed turns, and especially turns where the input and output ends of the winding come close together, add capacitance across the choke and lower its self-resonant frequency. Space the turns evenly and keep the two ends apart for the best high-frequency performance.
- For thick cables: Heavy coax or a thick mains cord may only fit 5–8 turns through an FT-240. Stack two cores, or use several snap-on clamps with the cable looped through each.
- For thin cables and wires: An FT-140 or FT-114 core is fine and allows more turns. An FT-140-31 with 12 turns of thin cable gives excellent suppression from 1 to 30 MHz.
- For mains cords: Use only intact, double-insulated cable, and never wind so tightly that the insulation is stressed. The choke works on all the conductors together, including the safety earth, so it does not affect safety.
Snap-On Clamps and Their Use
Snap-on ferrite clamps are the quick-fix tools of interference work. They need no soldering and no change to the cable, and can be added in minutes. They are not as effective as a properly wound toroid, but they are very useful on computer cables, USB cables, audio leads, sensor cables and power supply cords.
A single snap-on clamp on a cable gives the equivalent of one turn, and the air gap at the split joint reduces its effective permeability by around 30 to 50 percent compared with an equivalent toroid. For moderate suppression this is acceptable; for heavy suppression it is not enough.
You can increase effectiveness by looping the cable through the same clamp two or three times before closing it. Two passes give about four times the impedance of one (the N² relationship). Some clamps have a larger bore for exactly this purpose.
Place clamps as close as possible to the equipment that is the source or the victim of the noise: at the noisy device to stop noise getting onto the cable, and at the sensitive device to stop it getting in. On a cable between two problem devices, use one at each end.
When buying snap-on clamps, look for products with a stated material or impedance curve from established makers such as Fair-Rite, Würth Elektronik or TDK. Unbranded clamps may still help, but their performance is unpredictable.
Ferrite Beads, Saturation and Power
Ferrite beads on circuit boards. Surface-mount ferrite beads are tiny chokes for single conductors, used in series with supply rails and signal lines. Their datasheets give impedance at 100 MHz (for example “600 Ω at 100 MHz”) and an impedance curve. With capacitors on each side they form a low-pass filter that keeps high-frequency noise out of a sensitive circuit, such as the analog supply of an ADC or the supply to an RF module.
DC bias and saturation. A bead or choke carrying DC current in one direction is partly magnetized by it, and its impedance falls. Many small beads lose half or more of their rated impedance at their rated current. Always check the impedance-versus-DC-current curve, choose a bead rated well above the actual current, and remember that a common-mode choke avoids the problem: the equal and opposite supply and return currents cancel in the core, so even large DC currents do not saturate it.
Resonance. A ferrite bead combined with decoupling capacitors forms an LC circuit that can resonate at a low frequency and even amplify noise there. Beads with a high resistive component and a small series resistor or a lossy capacitor keep the resonance damped.
Power handling. On cables carrying RF power, the lossy ferrite converts some of the common-mode power to heat. With a small common-mode current the dissipation is tiny, but a large common-mode current can heat a small core quickly. A choke that becomes too hot to touch is a sign that the common-mode current is higher than expected, or that the core is too small. Ferrite also loses its magnetic properties above its Curie temperature, typically 130–250 °C, so overheating can make a choke stop working altogether.
Measuring Ferrite Impedance
The most reliable way to confirm that a choke provides enough impedance at the problem frequency is to measure it. A NanoVNA or other vector network analyzer sweeps the impedance across the frequency range in seconds.
Measurement procedure for a cable choke: connect the conductors of the cable together at each end, so the choke is measured as a single path through the ferrite. Connect one end to Port 1 of the VNA (center pin) and the other end to the VNA's ground (shell), with leads as short as possible. Set the VNA to display |Z| (impedance magnitude) and sweep the range of interest. For a single ferrite bead or a small choke, mount it in a short test fixture across the port in the same way.
Manufacturers publish impedance curves for their cores and beads, and Fair-Rite's catalog includes data for common turn counts on its toroids and clamps. Use these to choose a starting point, then measure to confirm.
⚖ Experiment: Winding and Testing a Cable Choke
Wind a cable choke on a ferrite toroid and measure its common-mode impedance with a NanoVNA. This experiment lets you check the core against published data and see how the number of turns changes the impedance curve.
- An FT-240-43 or FT-240-31 toroid core (Fair-Rite or Amidon)
- About 2.5 m (8 ft) of RG-58 coaxial cable or a two-core power cable
- A NanoVNA, with short test leads or an SMA-to-clip adapter
- Optional: a snap-on clamp and a surface-mount ferrite bead for comparison
- Wind the cable through the toroid: 9 turns, each pass through the center hole counting as one turn, spaced evenly around the core.
- At each end, connect all the cable's conductors together (center and shield of the coax, or both cores of the power cable).
- Calibrate the NanoVNA at Port 1 for 1–100 MHz using the open, short and load standards.
- Connect one end of the choke to the Port 1 center pin and the other end to the Port 1 shell, keeping the leads short. Display |Z| and record the impedance at 1, 3.5, 7, 14, 30 and 100 MHz.
- Unwind two turns to make a 7-turn choke and measure again. Then wind 11 turns and measure a third time, watching for a self-resonance peak followed by a fall in impedance.
- For comparison, measure a single pass of the cable through a snap-on clamp, and a ferrite bead on its own.
With 9 turns on FT-240-43, impedance rises from roughly 200 ohms at 1.8 MHz to a peak of 1,500–2,000 ohms somewhere around 10–20 MHz, then falls as parasitic capacitance takes over. The 7-turn choke has lower impedance at low frequencies but holds up to a higher frequency; the 11-turn choke has more impedance at low frequencies but its self-resonance moves lower. The single snap-on clamp gives only tens to a few hundred ohms — a clear demonstration of why multiple turns are worth the effort.
Frequently Asked Questions
Can I use any ferrite from a junk box?
Maybe, but the material matters greatly, and unmarked cores are hard to identify. Ferrite from old computer power supplies, monitors and surplus equipment may be anything from a low-frequency power material to a suppression material. A quick test is to wind a few turns and measure the impedance with a VNA: a suppression ferrite shows a broad, largely resistive impedance across the range you care about, while a power ferrite shows a sharp resonance and little loss. For predictable results, buy new cores with known mix numbers — they are inexpensive compared with the time spent troubleshooting a choke made from the wrong material.
How do I know if my choke has enough impedance?
Aim for at least 500 ohms of common-mode impedance at the problem frequency, with 1,000 ohms or more preferred. Measure it with a NanoVNA as described above: join the cable's conductors at each end, connect the two ends across Port 1, and sweep. If you see less than 500 ohms where you need it, add turns, stack a second core, or change to a material that works better at that frequency — Mix 31 or 77 for low frequencies, Mix 43 or 61 for higher ones. Then confirm the real result by checking the interference itself.
Will a ferrite choke affect the signal on the cable?
No — not when it is wound with the whole cable. The equal and opposite differential currents of the wanted signal cancel in the core, so the choke has no effect on the signal, the data rate, or the impedance the signal sees. It impedes only current flowing the same way on all the conductors. A ferrite bead on a single conductor is different: it acts on everything in that conductor, so it must be chosen so that its impedance is low at the signal's own frequencies.
Test Your Knowledge
Answer the questions below to check your understanding of this lesson. Every answer can be found in the lesson above.