How to Select Magnetic Core Material for High Frequency Applications?
For 1–100MHz power applications, evaluate core loss density, permeability stability, saturation flux density and cost. FeNi metal alloy cores offer Bs ≥1.5T with stable permeability for MHz metal inductors; ferrite fits low-power >500kHz designs; Sendust and MPP suit mid-frequency EMI and filter roles. Define frequency, current, temperature and loss budget before choosing.
The switching frequency of power electronics keeps climbing — AI GPU VRMs, server power, and compact chargers now operate in the 1–100MHz range. At these frequencies the magnetic core, not the copper winding, usually sets the limit of efficiency and size. Here is a practical selection method.
Start with three numbers
- Switching frequency (MHz): drives core loss and material choice
- Peak current (A): decides whether saturation (Bs) is the constraint
- Temperature budget (°C): ferrite loses permeability near Curie temperature
Compare the main core materials
| Property | FeNi (Iron Nickel) | Ferrite | Sendust | MPP |
|---|---|---|---|---|
| Saturation Bs | ≥1.5T | 0.3–0.5T | ~1.05T | ~0.75T |
| Permeability range | 14–160 | 300–15000 | 26–125 | 14–550 |
| Core loss @100kHz | ≤300 kW/m³ | Very low (<100kHz–1MHz) | Moderate | Low |
| Temp stability | Excellent (−40 to +180°C) | Fair (Curie limit) | Good | Good |
| Typical use | MHz metal inductors, AI GPU VRM, EV | Low-power HF, transformers | EMI filters, PFC | Filters, precision inductors |
Why FeNi for MHz metal inductors
AI GPU and server power rails need high current at high frequency without saturation. FeNi powder cores combine Bs ≥1.5T with a distributed air gap, so inductance remains stable under heavy DC bias while core loss stays bounded at MHz switching. That is why YUTE develops molded metal-composite inductors for the high switching frequencies used in AI accelerator VRM designs; frequency capability is confirmed per part number from measured data.
Selection checklist
- Current is high and DC bias is large → prioritize Bs and bias stability (FeNi)
- Frequency is above 500kHz and power is low → ferrite is cost-effective
- EMI filtering / PFC → Sendust or MPP for controlled permeability
- Temperature swings or automotive grade → require AEC-Q200 qualification at component level
- Prototype first: request samples and measure loss and temperature rise at your real operating point
YUTE Magnetics offers free technical evaluation of core material selection. Send your frequency, current and size targets and our engineers will respond on the next business day.
Dive deeper into Fe-Ni material properties and the FeNi powder core series in our High Flux Cores Guide, or compare ferrite options in the Ferrite Core Guide.
Frequently Asked Questions
How do I choose a core material for high frequency applications?
Start from three numbers: switching frequency, peak current and temperature budget. Those determine whether saturation or core loss is the binding constraint, which in turn points to FeNi for high current, ferrite for low-power high frequency, or Sendust and MPP for mid-frequency filtering roles.
Which core material is used for MHz switching inductors?
Metal alloy powder or composite cores are typically used at MHz switching, because they combine high saturation with a permeability that stays reasonably flat well beyond 10 MHz. Manganese-zinc ferrite becomes lossy at those frequencies, though it remains the better choice at lower frequency.
Why is core loss the limiting factor at high frequency?
Core loss rises steeply with both frequency and flux swing, so at MHz switching the magnetic material can dissipate more power than the copper winding. Once core loss dominates, improving the winding no longer improves efficiency.
What is the difference between FeNi, Sendust and MPP?
FeNi offers the highest saturation (Bs ≥1.5 T) for high-current work. Sendust balances cost against loss for PFC and EMI filtering in the mid-frequency range. MPP has the lowest core loss among metal powder cores but lower saturation, suiting precision filter and resonant applications.
Should I prototype before committing to a core material?
Yes. Datasheet loss curves are measured under specific conditions, and the real ripple waveform usually differs. Measuring loss and temperature rise on a prototype at the actual operating point is the only reliable confirmation.