Power inductors
Buck, boost and PFC chokes that have to hold their inductance at peak current. The gap, the core material and the winding are chosen together so the value survives your ripple and your temperature rise.
Home / Products / Custom Inductors
Custom Power Inductors
ConverNova designs and manufactures custom power, filter and resonant inductors for power electronics, from 50 Hz to 2 MHz. Based in Madrid, Spain, we return an engineering answer usually within 24 hours, deliver prototypes in 4 to 8 weeks, and manufacture from a single unit to production series with no minimum order quantity. Gapped, distributed gap and ungapped designs in ferrite, nanocrystalline, amorphous and powder, wound with litz from 0.03 mm strands, foil or edge wound flat wire.
Types
Three jobs, three different designs.
An inductor that stores energy and an inductor that blocks noise are not the same component with a different value. They are different designs.
Filter inductors
Differential mode filtering on the input or the output, sized from your real ripple spectrum and the attenuation your standard asks for, not from a rule of thumb.
Resonant inductors
Series and parallel resonant elements where the value is part of a tank. Tolerance is the specification here, so we hold it and we measure it.
The design decision
Where the gap goes.
In a gapped component the energy is stored in the gap, not in the ferrite. Where it sits decides the loss, the fringing and the noise.
Send us your spec| Approach | What it buys | What it costs |
|---|---|---|
| Discrete gap, ferrite | Low core loss, high energy density, tight value control | A fringing field at the gap, which heats the nearby copper unless the winding is kept back |
| Distributed gap, powder | Soft saturation, no fringing hotspot, graceful overload | Higher core loss, so it needs thermal room at high ripple |
| Graded or stepped gap | Fringing pushed away from the winding while keeping ferrite loss | More machining, so it earns its place in production volumes |
| Ungapped, nanocrystalline | Very high permeability for common mode and filtering | Saturates on any DC bias, so it is the wrong tool for energy storage |
Which one is right depends on your ripple current, your switching frequency and how much thermal headroom you have. Send the waveform and we will tell you which, and why.
Specification range
What we can build.
- Frequency
- 50 and 60 Hz line, through to 2 MHz for SiC and GaN switching stages
- Current
- From signal level to hundreds of amps, with the inductance held across the operating range
- Cores
- Ferrite (Ferroxcube, TDK), nanocrystalline (VAC, Hitachi), amorphous, powder (Magnetics, Micrometals). Custom machined geometries and production tooling available
- Gapping
- Discrete, stepped, graded or distributed, chosen for your loss and fringing budget
- Windings
- Litz from 0.03 mm strands, copper foil, edge wound flat wire, round magnet wire
- Construction
- Bobbin or bobbinless, vacuum impregnated, varnished, potted in resin or in an aluminum housing
- Temperature
- Minus 55 to plus 125 degrees C depending on materials and insulation class
- Test
- Inductance against DC bias, DC resistance, loss split, hipot, temperature rise under your real waveform
Straight answers
Custom inductors, the usual questions.
How do you stop it saturating at my peak current?
By sizing the core and the gap for your real peak current, not for a nominal
figure on a datasheet. The inductance against DC bias curve is part of the design proposal,
and the measured version ships with the part.
Gapped or distributed gap?
A discrete gap in ferrite stores energy efficiently and keeps core loss low,
but it pushes a fringing field into the nearby winding. A distributed gap spreads it out,
saturates softly and has no fringing hotspot, but the core loss is higher. Your ripple,
frequency and thermal headroom decide it, and we will tell you which and why.
Can you design for SiC and GaN converters?
Yes, up to 2 MHz switching. At those frequencies the loss balance moves:
winding loss from skin and proximity effect grows faster than core loss, so litz
construction and winding geometry start to matter more than the core choice.
What design information will I get?
Estimated core losses and winding losses, inductance, dimensions and the
expected hotspot temperature. We aim for inside 24 hours, before you have spent anything.
What is your minimum order?
One. Prototypes, samples and production come off the same drawing.
Construction
Inductor construction.
Stacked laminations, toroids and busbar terminations, chosen for the current and the bias rather than for the catalogue.
Start here
Send us your inductor spec.
Inductance, peak and RMS current, ripple, switching frequency, temperature and the space you have. Or just the waveform and the problem.