Magnetic components are rarely the most expensive line on a bill of materials, but they are one of the easiest places to lose margin — and one of the hardest places to cut cost safely. The cost of a core-and-winding assembly is dominated by four things: raw core material, copper, labour, and the test and qualification burden. Knowing which one you are actually paying for is the difference between a 15% cost-down and a redesign.
1. Where the money actually goes
| Cost element | Typical share (wound toroid, mid volume) | Lever that moves it |
|---|---|---|
| Core material | 25–40% | Grade selection, geometry standardisation, AL tolerance |
| Copper winding | 15–30% | Current density, fill factor, Litz only where needed |
| Labour and winding time | 15–25% | Automation, fewer turns, single-layer designs |
| Insulation, bobbin, potting | 8–15% | Insulation class, varnish versus full encapsulation |
| Test and qualification | 5–12% | Test time per unit, AQL versus 100% test |
| Packaging and logistics | 5–10% | Tray density, Incoterms, consolidation |
A 30% share for core material means a 10% rise in ferrite powder price translates to roughly 3% on the part — which is why “the price went up because of raw material” is sometimes true and often an opening position.
2. The material is not interchangeable
Mn-Zn ferrite, Ni-Zn ferrite, iron powder, sendust, high flux, MPP, nanocrystalline and amorphous cores differ by a factor of five or more in loss density at a given frequency, and by 0.3–1.5 T in saturation flux density. Choosing the cheapest grade that meets the datasheet inductance frequently produces a part that runs hotter, needs a bigger case, or fails EMI. The Magnetic Core selection guide sets out the trade-off table.
3. Eight levers that cut cost without touching reliability
- Standardise the core geometry. A bespoke shape adds tooling amortisation and limits second sourcing. EE, ETD, toroidal and planar families usually already contain a size that fits.
- Relax AL tolerance only as far as the control loop allows. ±25% cores are cheaper than ±10%. If the converter tolerates it, take the saving; if it does not, the cost is real.
- Reduce turns by raising flux swing. One turn less on a 40-turn winding cuts copper and labour together, provided the core does not saturate at the worst-case transient.
- Use Litz wire only where the loss is actually significant. Below roughly 20 kHz with a small AC component, round wire is usually cheaper per watt of loss saved.
- Match insulation class to the real hot spot. Class H everywhere is a common over-specification; class B or F with a measured margin is cheaper and equally compliant.
- Replace full encapsulation with varnish where humidity is not extreme. Potting adds material, cure time and rework difficulty.
- Automate the winding, not the test. Winding time is the largest labour line; test time is a smaller lever but easier to reach.
- Consolidate part numbers. Two 90%-similar designs in one family cost more in inventory, tooling and quality records than one slightly over-designed part.
4. A worked example
A 150 µH, 12 A toroidal PFC inductor on a 60 µ sendust core: core $0.92, copper $0.38, labour $0.41, varnish $0.06, test $0.05, packaging $0.04 — $2.06 total. Moving to a standard size, reducing turns by 6%, and switching from Class H to Class F insulation removes roughly $0.21, about 10%, with no change to the thermal margin because the measured hot spot was 34 K below the Class F limit.
5. What not to cut
Do not remove the bias-current test, the hipot step on isolation parts, or the lot-level core certificate. Those three items are the cheapest insurance in the entire BOM: they catch the substitution, the wrong gap and the insulation fault before the part reaches a customer.
6. Buying the cost-down, not just quoting it
Ask for the cost breakdown by element, not just a new unit price. A supplier who can show the core material, copper weight, winding minutes and test time is a supplier who can engineer a cost-down with you. One who can only move the price is moving margin — and will move it back at the next renewal.
Shaanxi SiRui Technology produces the Magnetic Core series in-house and winds to customer drawings, so grade, gap and geometry can be optimised against the target cost instead of being bought in at whatever the distributor charges.
Send a current BOM line to sales@sirui-tech.com for a costed alternative.