Sirui-Tech logo Sirui-Tech

Three Core Families, One Decision Framework

When a design needs high permeability with low loss at frequencies from 20 kHz to 200 kHz, the short list is almost always ferrite, amorphous metal, or nanocrystalline ribbon. They are not interchangeable upgrades of the same part; each has a different saturation level, a different loss mechanism, and a different cost structure. Choosing correctly is a matter of matching the dominant constraint in your design.

Property Comparison

Typical property ranges for the three families (values vary by grade and vendor)
Property MnZn ferrite Amorphous (Fe-based) Nanocrystalline
Saturation flux Bsat 0.4–0.55 T 1.2–1.6 T 1.1–1.3 T
Initial permeability µi 2,000–15,000 20,000–80,000 (with gap effects) 80,000–200,000
Core loss at 100 kHz, 0.2 T Moderate; rises sharply above 200 kHz Low to moderate Lowest of the three
Curie temperature 100–230 °C 350–400 °C ~570 °C
Mechanical form Pressed E, RM, PQ, toroid Tape-wound toroid or cut core Tape-wound toroid, cut core, ribbon
Relative cost per core Low Medium High

What Each Family Is Best At

Ferrite wins on cost and geometry freedom. Its low saturation flux is the limitation, so a ferrite design at high bias needs a larger core or an air gap, and the gap brings fringing loss and radiated field. Above roughly 300 kHz ferrite grades optimized for high frequency (MnZn power grades) remain practical up to the low MHz range with thin laminations and small flux swing.

Amorphous metal is the workhorse for 50/60 Hz to 20 kHz where high saturation and moderate loss matter: distribution transformers, CMC cores for high-current mains filters, and medium-frequency transformers in welding and induction heating. Its Curie temperature above 350 °C makes it tolerant of overload and poor cooling.

Nanocrystalline is the choice when the design must have both high permeability and low loss at 20–150 kHz with a small footprint: common mode chokes with high impedance per turn, precision current transformers, and resonant transformers in high-power converters. The trade is cost and brittleness — the ribbon is thin and the core is sensitive to mechanical stress, so clamping and potting must be designed carefully.

Decision Rules That Hold Up

  1. If Bsat is the binding constraint (high DC bias, inrush, or saturation during a fault), amorphous beats ferrite on size, and nanocrystalline is usually unnecessary.
  2. If loss at 50–150 kHz dominates and the window is tight, nanocrystalline usually wins on total volume even at a higher core price.
  3. If cost dominates and the design can accept a gap and more turns, ferrite remains the default through roughly 300 kHz.
  4. If the part must survive a short-circuit or an ambient above 120 °C with margin, ferrite’s Curie temperature becomes the risk, not its loss.
  5. If the choke must present high impedance from 10 kHz to 1 MHz with a small number of turns, the answer is nanocrystalline or a high-permeability ferrite with a comparably large core — compare both on volume, not on core price alone.

The Costs People Forget

Amorphous and nanocrystalline cores are tape-wound, so the winding window shape differs from a pressed E core: the bobbin is often custom, fill factor is lower, and the winding must handle a round or rectangular cut core. Nanocrystalline also needs stress relief — clamping force, potting compound, and mounting all affect permeability, and a poorly potted core can lose 20 % of its initial permeability. Ferrite needs no such treatment but pays for it in size at high bias.

Machining a gap in a tape-wound core costs more than molding a ferrite gap, and the gap tolerance affects inductance tolerance directly. If your design has a tight inductance window, check whether the chosen family can hold it before committing.

Related reading

Our Magnetic Core series covers ferrite, powder, amorphous, and nanocrystalline cores with the permeability, loss, and gap data needed for a fair comparison. Send your bias current, switching frequency, and volume target to sales@sirui-tech.com and we will return a side-by-side core proposal with the size and loss estimate for each family.