If you are designing switch-mode power supplies, electromagnetic brakes, or pulse-forming networks, you eventually hit a wall where your inductor gets too hot, saturates, or simply fails to store the energy you calculated. The root cause almost always traces back to a misunderstanding of the magnetic energy density formula. It is not just an abstract physics equation; it is the governing rule for how much energy you can pack into a given volume of magnetic material before the physics fight back.
The fundamental magnetic energy density formula is:
u = B2 / (2μ)
Below, we will break down every symbol, track the units through two distinct bench problems, and walk through a real-world failure where ignoring the assumptions of this formula destroyed a 48V buck converter prototype.
The Magnetic Energy Density Formula: Symbols, Units, and Rearrangements
To use this formula on the bench, you need to know exactly what each variable represents and the strict SI units required. Mixing up relative and absolute permeability is the most common way to end up with an answer that is off by a factor of 1,000 or more.
| Symbol | Parameter | SI Unit | Typical Bench Context |
|---|---|---|---|
| u | Magnetic energy density | Joules per cubic meter (J/m³) | Used to calculate total stored energy when multiplied by core volume. |
| B | Magnetic flux density | Tesla (T) | Limited by Bsat (e.g., ~0.3T for ferrite, ~1.5T for silicon steel). |
| μ | Absolute permeability | Henries per meter (H/m) | The product of μ0 and μr. |
| μ0 | Vacuum permeability | 4π × 10-7 H/m (≈ 1.2566 × 10-6 H/m) | A physical constant. Never omit this. |
| μr | Relative permeability | Dimensionless | Material property (e.g., Air = 1, N87 Ferrite ≈ 2000). |
Rearranged Forms
Depending on what you are solving for on the whiteboard, you will need to rearrange the formula. Here are the most useful variants for magnetics design:
- Solving for Flux Density (B): B = √(2μu) — Useful for finding the peak B-field required to store a target energy density.
- Solving for Absolute Permeability (μ): μ = B2 / (2u) — Useful for selecting a core material or calculating the required effective permeability of a gapped core.
- Total Energy (U): U = u × Vcore — Where Vcore is the effective magnetic volume of the core in m³.
When This Formula Applies (And When It Breaks Down)
The magnetic energy density formula assumes a linear magnetic material. This means it assumes that μ is constant regardless of how much current (H-field) you push through the coil.
In the real world, ferromagnetic materials (ferrites, powdered iron, silicon steel) are highly non-linear. As B approaches the saturation limit (Bsat), μr plummets toward 1. If you use the initial μr from the datasheet to calculate energy density near saturation, your math will be dangerously wrong. The formula is perfectly accurate for air-core inductors, and it is accurate for gapped ferrite cores where the air gap dominates the reluctance and linearizes the effective B-H curve.
What a Realistic Answer Magnitude Looks Like
Having a mental benchmark for realistic energy density magnitudes prevents decimal-point errors:
- Air core (10 mT field): ~40 J/m³ (Very low, requires massive volume for useful energy).
- N87 Ferrite (0.3 T field, ungapped): ~18 J/m³ (Counter-intuitively lower than air for the same B, because μ is so high).
- Gapped Ferrite (0.3 T field, effective μr = 100): ~358 J/m³ (The sweet spot for SMPS inductors).
- Neodymium MRI Magnet (3.0 T field in air/vacuum): ~3.58 MJ/m³ (Megajoules per cubic meter—enough to turn a dropped wrench into a lethal projectile).
Bench Example 1: Sizing an Air-Core Inductor for a Pulse Circuit
Let us track the units explicitly. You are building a pulse-forming network for a xenon flashlamp trigger and need an air-core coil. The design requires a peak magnetic flux density of 50 mT (0.05 T).
Step 1: Identify knowns.
B = 0.05 T
μ = μ0 = 1.2566 × 10-6 H/m (since μr for air is 1)
Step 2: Plug into the formula.
u = (0.05)2 / (2 × 1.2566 × 10-6)
u = 0.0025 / (2.5132 × 10-6)
Step 3: Calculate and track units.
u = 994.7 (T2 / (H/m))
Dimensional Analysis Check: Tesla is (V·s)/m². Henry is (V·s)/A.
(T2) / (H/m) = [ (V²·s²)/m⁴ ] / [ (V·s)/(A·m) ] = (V·A·s) / m³.
Since Volts × Amps = Watts, and Watts × seconds = Joules, the units resolve perfectly to J/m³.
Outcome: The energy density is 994.7 J/m³. If your coil has an effective magnetic volume of 50 cm³ (50 × 10-6 m³), the total stored energy is U = 994.7 × 50 × 10-6 = 0.049 Joules. This is enough to trigger the flashlamp, but not enough to power it.
Bench Example 2: The Ferrite Core Saturation Trap (And How Gapping Fixes It)
This is where theory meets the workbench. In 2026, with wide-bandgap GaN and SiC converters switching at 1 MHz+, magnetics design is tighter than ever. Let us look at a scenario where the formula was misapplied.
The Setup
An engineer is designing a 100 µH inductor for a 48V-to-12V buck converter. Peak current is 5A. They select an EFD25 core made of TDK N87 ferrite. The datasheet states an initial relative permeability (μr) of 2,000. The core volume (Ve) is 5,550 mm³ (5.55 × 10-6 m³). The saturation flux density (Bsat) at 100°C is roughly 0.32 T.
The Flawed Numbers
The engineer wants to know the maximum energy the core can store right at the edge of saturation (B = 0.32 T). They use the magnetic energy density formula with the datasheet's initial μr:
- B = 0.32 T
- μ = 2000 × 1.2566 × 10-6 = 2.5132 × 10-3 H/m
- u = (0.32)2 / (2 × 2.5132 × 10-3) = 0.1024 / 0.0050264 = 20.37 J/m³
- Total Energy U = 20.37 × 5.55 × 10-6 = 0.00011 Joules (113 µJ)
The Outcome
The required energy for a 100 µH inductor at 5A is U = ½LI² = ½ × 100µ × 25 = 1,250 µJ. The engineer realizes the ungapped core stores 10 times less energy than needed. They wind more turns to hit 100 µH, but the core saturates violently at just 0.5A. The inductance collapses to 2 µH, the current spikes, and the high-side GaN FET explodes.
What Went Wrong (The Physics Fix)
The engineer fell for the high-permeability trap. As the formula shows, for a fixed B, a higher μ results in a lower energy density. High μ means the material requires very little current (H-field) to reach 0.32 T, so it stores very little energy in the process.
The Fix: You must introduce an air gap. By grinding a 1.0 mm gap into the EFD25 center leg, the effective permeability (μe) drops from 2,000 down to roughly 100. Let us rerun the formula at the same 0.32 T saturation limit:
- μeff = 100 × 1.2566 × 10-6 = 1.2566 × 10-4 H/m
- u = (0.32)2 / (2 × 1.2566 × 10-4) = 407.4 J/m³
- Total Energy U = 407.4 × 5.55 × 10-6 = 0.00226 Joules (2,260 µJ)
By lowering the permeability via an air gap, the energy density at saturation increased by a factor of 20. The core can now safely handle the 5A peak current without saturating, and the MOSFET survives. For deeper magnetics design workflows, referencing the Coilcraft Magnetics Design Tools or Georgia State University HyperPhysics modules on inductance will confirm how fringing flux around that gap slightly alters the effective volume, but the core math remains identical.
The Unit Mistakes That Will Fry Your Prototype
Warning: The following unit errors are responsible for the majority of magnetics calculation failures on the bench. Double-check these before ordering custom wound bobbins.
- Gauss vs. Tesla: Older datasheets and legacy US engineering docs often use Gauss. 1 Tesla = 10,000 Gauss. If you plug B = 3000 (thinking Gauss) into the formula without dividing by 10,000, your calculated energy density will be off by a factor of 100,000,000. Always convert to Tesla first.
- cm³ vs. m³ for Total Energy: The formula outputs J/m³. Core volumes in datasheets are almost always listed in mm³ or cm³. An EFD25 core is roughly 5.5 cm³. If you multiply your J/m³ result by 5.5 instead of 5.5 × 10-6, you will think your tiny ferrite bead can store enough energy to weld metal.
- Forgetting μ0: A surprisingly common mistake is plugging μr (e.g., 2000) directly into the μ slot in the denominator. μ must be the absolute permeability (μ0 × μr). Forgetting the 4π × 10-7 multiplier will make your calculated energy density billions of times smaller than reality.
- Ignoring Temperature Derating: Bsat is not a static number. For standard manganese-zinc ferrites (like N87 or PC40), Bsat drops from ~0.40 T at 25°C to ~0.32 T at 100°C. If you use the room-temperature Bsat in your formula, your inductor will saturate and fail the moment the converter reaches operating temperature.
Mastering the magnetic energy density formula is not about memorizing algebra; it is about understanding the physical trade-off between flux, permeability, and volume. When you respect the units and remember that air gaps are your friend for energy storage, your magnetics will run cooler, smaller, and without blowing your switching semiconductors.
