Why Ungapped Cores Are Almost Never the Answer
Ferrite has a high permeability and a low saturation flux density. An ungapped ferrite core stores almost no energy before it saturates: the permeability that makes it good at coupling flux is exactly what makes it unsuitable for storing it. Gapping the core trades permeability for energy storage and stability, and how the gap is made and controlled decides the inductance tolerance, the saturation behaviour and a good part of the cost.
What the Gap Does
| Gap | Relative permeability | Inductance | Saturation current (approx) |
|---|---|---|---|
| Ungapped | ~2000 (effective) | Very high, tolerance poor | Low |
| 2 mm total | ~30 | Moderate | High |
| 4 mm total | ~15 | Lower | Higher |
The relationship is not linear and the numbers above are illustrative; the design point is that increasing the gap lowers inductance and raises the current at which the core saturates, while the energy storage capability rises because the core can tolerate a larger flux swing before saturation relative to the permeance it presents.
Three Ways to Make the Gap
- Ground centre-leg gap. The gap is machined into one leg, usually the centre leg, to a controlled depth. Inductance tolerance can be held to ±5–8 % with grinding and measurement, and the fringing field is contained inside the core. This is the standard for gapped inductors.
- Spacer gap. A non-magnetic shim between the core halves. Cheap and flexible, but tolerance depends on the shim thickness plus the flatness of the mating surfaces, and it is not repeatable across a production run.
- Distributed gap. The gap is distributed through the material, as in powder cores or in some SMC designs. Fringing losses are much lower and the saturation curve is soft, but effective permeability is low and the material cost per unit of energy storage is higher.
Fringing is the reason the method matters: at the gap, flux bulges outward and passes through the copper nearby, producing eddy losses in the winding adjacent to the gap. A concentrated 4 mm gap in a winding that sits right against it can add more loss than the core itself. The fix is physical — keep turns away from the gap, or split the gap into two smaller gaps — not electrical.
What to Write on the Drawing
- Inductance and the test conditions: frequency, bias current, and the fixture. A gapped inductor measured at 0 A and 1 kHz will read high and mean nothing.
- Tolerance, and whether it applies at the operating bias point or at zero bias.
- Saturation current definition: the current at which inductance falls to 90 % (or the agreed percentage), not “the saturation current” alone.
- Gap method: ground centre leg, spacer, or distributed, with the target gap and its tolerance.
- Whether the gap faces the winding: this decides fringing loss, and a supplier who can choose will choose the cheaper side.
Related reading
- How to Read a Magnetic Core Datasheet
- Inductance Tolerance and Converter Performance
- SMC vs Laminated Steel Cores
- Toroid vs EE vs Pot Core
We grind and measure gaps in-house and report inductance at the operating bias, with the fringing arrangement shown on the drawing. Magnetic Core engineering: sales@sirui-tech.com.