Power Electronics

Core Material (Magnetics)

Q: How do I choose between MnZn and NiZn ferrite for an RF inductor?

The deciding factor is operating frequency. Manganese-zinc (MnZn) ferrites offer very high initial permeability (mu_i of 750 to 15,000) but their bulk resistivity is low (1 to 100 ohm-cm), so eddy-current losses rise sharply above a few MHz. They dominate the 10 kHz to 2 MHz range for power transformers and EMI suppression. Nickel-zinc (NiZn) ferrites have lower permeability (mu_i of 10 to 1,500) but resistivity of 10^5 to 10^8 ohm-cm, keeping eddy losses low into the VHF and UHF bands; they are the standard choice from roughly 1 MHz to 500 MHz for RF chokes, baluns, and wideband transformers. As a rule, pick the lowest permeability that still gives the inductance you need, because lower-permeability NiZn grades have a higher useful frequency ceiling.

Q: Why does core permeability roll off with frequency, and where is the limit?

Permeability falls because the magnetic domain walls and electron spins cannot follow the field above a material-specific ferrimagnetic resonance frequency. The Snoek limit ties the two together: the product of the initial permeability and the resonance frequency is roughly constant for a given saturation magnetization, so a high-mu MnZn grade resonates at a low frequency while a low-mu NiZn grade resonates much higher. Practically this means a mu_i = 5,000 MnZn part is useful only to about 1 MHz, while a mu_i = 40 NiZn part stays usable to several hundred MHz. Above the resonance the complex permeability becomes strongly lossy (mu'' dominates), which is exactly what EMI suppression beads exploit.

Q: What core loss should I expect at 100 kHz and how is it modeled?

Core loss is usually estimated with the Steinmetz equation, P_v = k times f^alpha times B^beta, where P_v is volumetric loss in mW/cm^3, f is frequency, B is peak flux density, and k, alpha, and beta are fitted constants from the manufacturer datasheet. For a typical power ferrite such as a 3F3-class material at 100 kHz and 100 mT (0.1 T) peak, P_v is on the order of 60 to 100 mW/cm^3 at 100 degrees C. Loss splits into hysteresis (scales near linearly with frequency) and eddy-current plus residual losses (scale faster). Because the exponents alpha (1.3 to 1.7) and beta (2.3 to 2.7) are greater than one, halving the flux swing roughly cuts loss by a factor of five, which is why designers add turns or a larger core rather than push flux density.

/kor muh-TEER-ee-uhl (mag-NET-iks)/

Inside any wound magnetic component sits a magnetic substance that guides and concentrates the flux, and its choice sets the inductance, bandwidth, and loss of the part. Selected for its magnetic permeability, saturation flux density, and frequency-dependent core loss, the material determines whether a transformer is useful at line frequency, in the RF bands, or up to VHF. Manganese-zinc ferrite (μ_i 750 to 15,000) dominates below 2 MHz, nickel-zinc ferrite (μ_i 10 to 1,500) extends to several hundred MHz, and powdered iron handles high DC bias with low permeability. The trade-off is governed by the Snoek limit, which couples permeability against usable frequency.

Category: Power Electronics

Initial μ_i: 4 to 15,000

B_sat: 0.3 to 0.5 T (ferrite)

Selecting a Magnetic Core for the Frequency Band

A magnetic core is not a single material but a family, and the right pick depends almost entirely on where in the spectrum the part operates. The flux path stores and shapes magnetic energy; a higher permeability packs more inductance into fewer turns, but it also lowers the frequency at which the material stops behaving like an inductor and starts behaving like a lossy resistor. Power designers below 2 MHz reach for manganese-zinc (MnZn) ferrite because its initial permeability can exceed 10,000, while RF designers from 1 MHz into the VHF range use nickel-zinc (NiZn) ferrite, trading raw permeability for a bulk resistivity of 10⁵ to 10⁸ Ω-cm that suppresses eddy-current heating.

Three datasheet numbers anchor every selection: initial permeability μ_i, saturation flux density B_sat, and the volumetric loss density P_v at a stated frequency and flux swing. Ferrites saturate around 0.3 to 0.5 T, far below the 1.5 to 2.0 T of silicon steel, so RF cores must run at modest flux density to stay linear. Powdered iron and sendust-type alloys distribute a tiny air gap throughout the material, giving a soft saturation knee and excellent DC-bias tolerance at the cost of permeability rarely above 100.

The dominant physical limit is ferrimagnetic resonance. As frequency rises, domain walls and electron spins can no longer track the applied field, the real permeability μ′ collapses, and the loss term μ″ takes over. Snoek's relation states that the product of permeability and that resonance frequency is roughly fixed for a given saturation magnetization, which is precisely why no single ferrite serves both a 50 kHz flyback transformer and a 200 MHz wideband balun.

The Snoek Limit and Core Loss

Snoek Limit (permeability vs. frequency):
(μ_i − 1) × f_res ≈ (2/3) × γ × 4πM_s

Steinmetz Core Loss (volumetric):
P_v = k × f^α × B_pk^β (mW/cm³)

Inductance from core geometry:
L = A_L × N², A_L ≈ μ₀ × μ_e × A_e / l_e

Where μ_i = initial permeability, f_res = ferrimagnetic resonance, M_s = saturation magnetization, γ = gyromagnetic ratio, P_v = loss density, α ≈ 1.3 to 1.7, β ≈ 2.3 to 2.7, A_L = inductance factor (nH/turn²), N = turns, A_e = effective area, l_e = effective path length.

Core Material Comparison

Material	Initial μ_i	Useful Frequency	B_sat	Resistivity	Best Application
MnZn ferrite	750 to 15,000	10 kHz to 2 MHz	0.4 to 0.5 T	1 to 100 Ω-cm	Power transformers, EMI
NiZn ferrite	10 to 1,500	1 MHz to 500 MHz	0.3 to 0.4 T	10⁵ to 10⁸ Ω-cm	RF chokes, baluns
Powdered iron	4 to 100	50 kHz to 100 MHz	~1.0 to 1.4 T	Distributed gap	High DC-bias inductors
Sendust / Kool Mu	26 to 125	up to ~1 MHz	~1.0 T	Distributed gap	PFC chokes, filters
Nanocrystalline	15,000 to 100,000	10 kHz to 1 MHz	1.2 to 1.3 T	Tape-wound	Common-mode chokes

Common Questions

Frequently Asked Questions

How do I choose between MnZn and NiZn ferrite for an RF inductor?

Operating frequency decides it. MnZn ferrite gives very high μ_i (750 to 15,000) but low resistivity (1 to 100 Ω-cm), so eddy losses climb above a few MHz; it rules the 10 kHz to 2 MHz range. NiZn ferrite has lower μ_i (10 to 1,500) but resistivity of 10⁵ to 10⁸ Ω-cm, staying low-loss into VHF/UHF, so it is the choice from roughly 1 MHz to 500 MHz. Pick the lowest permeability that still hits your target inductance.

Why does core permeability roll off with frequency, and where is the limit?

Domain walls and electron spins cannot follow the field above a material-specific ferrimagnetic resonance. The Snoek limit fixes the product μ_i × f_res for a given saturation magnetization, so a μ_i = 5,000 MnZn grade is useful only to about 1 MHz while a μ_i = 40 NiZn grade reaches several hundred MHz. Above resonance the lossy term μ″ dominates, which is what EMI suppression beads exploit.

What core loss should I expect at 100 kHz and how is it modeled?

Use the Steinmetz equation, P_v = k × f^α × B^β in mW/cm³. A typical 3F3-class power ferrite at 100 kHz and 100 mT peak runs about 60 to 100 mW/cm³ at 100°C. Because α (1.3 to 1.7) and β (2.3 to 2.7) exceed one, halving the flux swing cuts loss roughly five-fold, so designers add turns or a larger core rather than push flux density.

Related Terms

← Core Loss Core Material (Radome) →