Custom Ferrite Cores for High-Frequency Transformers

Most high-frequency transformers begin with a catalogue core, and for a large share of designs that is where they end. A standard shape and a standard material grade cover the electrical requirement, and the engineering effort goes into the winding. Custom core development becomes worthwhile when the standard options fail on something other than magnetics.

Understanding when that point is reached — and what to have ready when it is — saves a development cycle that would otherwise be spent discovering the constraint late.

When a catalogue core is not enough

Standard shapes exist in fixed proportions, and those proportions do not always match the mechanical reality of a design. The most common triggers for custom development are dimensional or thermal rather than electromagnetic.

• A height limit that no standard shape satisfies at the required power
• An isolation distance that does not fit the winding window of a standard core
• A footprint that must match an existing board outline
• A thermal path that requires a specific mounting or contact area
• A pin or terminal arrangement dictated by the assembly process
• A volume or cost point that a standard shape cannot reach

What to specify before development starts

Custom development proceeds from the operating point, not from a desired core drawing. The specification should establish the electrical, thermal and mechanical requirements together, because they interact — changing one usually moves the others.

• Input and output voltage, current and power level
• Switching frequency and duty cycle range
• Topology and whether the core stores or transfers energy
• Target primary inductance, leakage inductance and their tolerances
• Peak flux density with margin below saturation
• Isolation, creepage and clearance requirements with the applicable standard
• Maximum height, footprint and mounting method
• Ambient temperature, cooling method and allowed temperature rise
• Qualification requirements for the target market

Choosing the material grade

Material grade selection follows from frequency and loss budget. Manganese-zinc ferrite grades cover the power conversion range used by most high-frequency transformers, with different grades optimized for different loss and temperature behaviour. Where the frequency is higher and the power lower, nickel-zinc grades become relevant because their higher resistivity reduces eddy-current loss.

The grade also interacts with the operating temperature. A material with favourable loss characteristics at room temperature may behave differently at the hot-spot temperature the design actually reaches, so the loss curve should be evaluated at the real operating condition rather than at a nominal point.

Choosing the geometry

Geometry selection balances winding window against magnetic path length and thermal coupling. A larger window accepts more copper and provides room for the isolation barrier, but lengthens the magnetic path and increases core volume. A compact shape reduces volume but limits the turns that fit and can make isolation distances difficult to achieve.

For high-frequency transformers, shapes with a round centre leg and a circular winding window often fill the available space efficiently, while low-profile shapes suit height-constrained boards. The choice is usually made by evaluating two or three candidate geometries against the same operating point rather than by optimizing a single one.

Tooling, sampling and qualification

Custom cores require tooling, so the development sequence includes a tooling lead time before samples exist. That makes it important to settle the specification before committing: a late change to the geometry restarts the tooling cycle and can put the program schedule at risk.

The typical sequence runs from specification review, through material and geometry selection, to prototype samples, evaluation in the real circuit, refinement if needed, and then qualification and volume production. Each stage should have a defined acceptance criterion so that a problem surfaces at the earliest possible point rather than at qualification.

What to verify before production

• Inductance and leakage inductance within the specified tolerance
• Core loss measured at the real frequency and flux swing
• Temperature rise at worst-case ambient and load
• Isolation and creepage distances verified on the physical part
• Mechanical fit in the actual enclosure
• Consistency between samples from the first production lot

Supplier capabilities worth checking

Not every core supplier can support custom development, and the differences matter when a program is on a schedule. Useful capabilities to confirm include in-house material control, the ability to measure permeability, core loss and saturation rather than quoting typical values, tooling and prototyping capability, qualification support, and lot traceability for volume production.

It is also worth confirming that engineering, not sales, evaluates the application. A supplier who reviews the operating point before quoting tends to produce a usable recommendation rather than a part number.

Application examples

• On-board chargers and DC-DC converters with tight packaging
• Industrial power supplies with non-standard form factors
• Telecom rectifiers where height is constrained by the rack
• Medical and instrumentation supplies with stringent isolation
• Planar transformers where the core must match a PCB winding pattern

Development checklist

• Confirm a catalogue core genuinely cannot meet the requirement
• Document the electrical, thermal, mechanical and isolation requirements together
• Select material grade by loss at the real operating temperature
• Evaluate two or three geometries rather than optimizing one
• Freeze the specification before committing to tooling
• Define acceptance criteria for samples and first production

Getting started

The most useful first step is a specification review: send the operating point, the mechanical envelope and the isolation requirement, and ask for a recommendation on material grade and geometry before any tooling is discussed. That gives a clear view of whether a catalogue core can be adapted or whether development is genuinely needed.

See Ferrite Cores for High-Frequency Transformers for material and shape fundamentals, Ferrite Core Manufacturer in China: Sourcing Checklist for supplier evaluation, and the Ferrite Core Guide for available grades and shapes.

Frequently Asked Questions

When should I consider a custom ferrite core instead of a catalogue part?

When a standard shape cannot meet the requirement. The usual triggers are a height limit, an isolation distance that does not fit the winding window, a board outline that must be matched, a specific thermal or mounting requirement, or a cost and volume target that a standard shape cannot reach.

What information does a core manufacturer need to start custom development?

The operating point and the constraints together: input and output voltage, current and power, switching frequency and topology, target inductance and leakage with tolerances, peak flux density margin, isolation and creepage requirements, height and footprint limits, temperature conditions, and qualification requirements.

How long does custom ferrite core development take?

It depends on the geometry and the tooling required, and it runs through specification review, material and geometry selection, prototype samples, evaluation in the real circuit, refinement where needed, and then qualification. Because tooling has a lead time, settling the specification before committing avoids repeating the cycle.

Which ferrite material is used for high-frequency transformers?

Manganese-zinc grades cover most power conversion frequencies, with different grades optimized for different loss and temperature behaviour. Nickel-zinc grades become relevant at higher frequencies and lower power levels, where their higher resistivity reduces eddy-current loss.

What should be verified before releasing a custom core to production?

Inductance and leakage inductance within tolerance, core loss measured at the real frequency and flux swing, temperature rise at worst-case load and ambient, isolation and creepage distances on the physical part, mechanical fit in the enclosure, and consistency between samples and the first production lot.