The Two Candidates
For a switching converter inductor there are two mainstream material choices: a gapped ferrite core, which has high permeability and low loss but saturates sharply, and a distributed-gap powder core (iron, sendust, or a moulded composite), which has lower permeability and a soft saturation curve but higher core loss at high frequency. The choice is usually decided by what happens at the worst-case current, not by the loss at nominal.
Decision Table
| Criterion | Gapped ferrite | Powder / composite |
|---|---|---|
| Permeability | High effective (μe 30–200 with gap) | Low (10–90), stable with bias |
| Core loss at 100 kHz+ | Low with a good grade | Higher; rises quickly with frequency |
| Soft saturation | No: falls sharply past Bsat | Yes: inductance declines gradually, which is safer in overload |
| Inductance tolerance | Set by gap grinding, ±5–10 % | Set by material and dimensions, often ±10–20 % |
| Fringing loss | Significant near a concentrated gap | Low, gap is distributed |
| Cost at volume | Low for standard sizes | Higher per core, but no gap grinding step |
| Best fit | CCM boost, LLC, flyback transformers, filters | PFC chokes, high-ripple inductors, overload-tolerant designs, high-current low-inductance |
The practical decision is often made by the overload case. A PFC choke that must ride through a 120 % load surge without saturating is far easier with a powder core, whose inductance simply declines. A converter that operates in CCM at a fixed current is usually better served by ferrite, whose loss per unit of energy storage is lower.
The Loss Budget Comparison
A quick way to compare: at the same inductance and the same peak current, a powder core typically needs a larger core volume because its permeability is lower, and its core loss per unit volume is higher. The gapped ferrite wins on loss but its saturation is abrupt, so the design must leave headroom. At 100 kHz and above, the ferrite advantage widens; below 30 kHz and at high ripple, the powder core’s soft saturation often removes the need for that headroom and therefore for a bigger core.
What to Specify Either Way
- Inductance at the operating bias current, with the test method.
- Saturation definition: the current at which inductance falls to 90 % of nominal.
- Worst-case current, including transient and startup, not just rated.
- Core loss at the actual ripple and frequency, not at the catalogue reference point.
- Temperature rise at the real ambient and mounting.
Related reading
- Ferrite Core Gapping
- SMC vs Laminated Steel Cores
- Core Loss vs Copper Loss
- Inductance Tolerance and Performance
We build both constructions and will quote the two side by side with the loss and saturation numbers at your operating point. sales@sirui-tech.com.