Home / Sectors / Pulsed power and big science

Magnetics for pulsed power and big science

Klystron modulators, magnet power supplies, fusion coil supplies and accelerator RF stages. In pulsed work the average power tells you almost nothing. Peak current, volt-second product, rise time and droop are what set the design, and they pull in different directions.

Send us your spec

Where the magnetics sit

The stages, and what goes in each one.

Pulse transformer

Voltage step up into a klystron, magnetron or similar load.

The magnetic: The design is dominated by leakage inductance, which limits rise time, and by distributed capacitance, which does the same. The volt-second product sets the core, and droop sets the magnetising inductance.

Energy storage and smoothing

Charging supplies and magnet power supplies with heavy current.

The magnetic: High current chokes where conduction loss and the thermal path decide the winding, and where saturation at peak current is the limit that matters.

RF and accelerator stages

Drive and modulator stages around RF amplifiers.

The magnetic: Transformers and inductors designed for the repetition rate and the duty, with the material chosen for loss at the pulse spectrum rather than at a single frequency.

Isolation at high voltage

Filament, bias and auxiliary supplies floating at high potential.

The magnetic: Isolation transformers designed for the standoff and for low interwinding capacitance, so the floating supply does not couple the pulse back into the control system.

What decides it

What actually drives the design here.

Rise time

Leakage inductance and distributed capacitance both slow the edge. Reducing one usually raises the other, so the winding arrangement is the real design decision.

Volt-seconds, not volts

The core is sized by the product of voltage and pulse width. A short high voltage pulse and a long low voltage one can need the same core.

Droop

Magnetising inductance sets how much the flat top sags. If your load cares about flatness, that number is a specification, not an outcome.

Often one unit

Research hardware is frequently a single build. No minimum order quantity is not a commercial gesture here, it is the normal case.

The shape of it

Standards, range and typical parts.

Design inputs
Peak voltage and current, pulse width, rise time, repetition rate, droop budget
Measured and reported
Leakage inductance, magnetising inductance, distributed capacitance, hipot, partial discharge
Construction
Vacuum impregnated, potted in resin or in an aluminium housing for high voltage standoff
Typical parts
Pulse transformers, high current chokes, HV isolation transformers, saturable reactors

Questions

Pulsed power and big science, the usual questions.

What do you need to design a pulse transformer?
Peak voltage and current, pulse width, required rise time, repetition rate, the droop you can accept across the flat top, and the load. The volt-second product sets the core, the droop sets the magnetising inductance, and the rise time sets how hard we have to fight leakage and distributed capacitance.
Why does the rise time depend on the winding, not just the core?
Because the edge is limited by leakage inductance and by distributed capacitance, both of which come from how the winding is arranged rather than from the core. Interleaving reduces leakage but increases capacitance. Which one to favour depends on your source impedance and your load.
Can you build a single unit?
Yes. There is no minimum order quantity. Research and pulsed power work is very often one, and the same drawing scales later if the machine gets replicated.
Do you handle high voltage standoff and partial discharge?
Yes. Encapsulation, creepage and the insulation system are designed for the standoff, and partial discharge is tested and reported rather than assumed from a hipot pass.
Can you work from a schematic rather than a drawing?
Yes, and for pulsed work that is usually the better starting point. Send the circuit and the pulse you need at the load, and the first proposal comes back with estimated losses, the inductances and the expected hotspot temperature.

Start here

Send us your pulsed power spec.

A schematic and a waveform are enough to start. You get a first design proposal with estimated losses, the inductances, the dimensions and the expected hotspot temperature, before you have spent anything.

Send us your spec

All sectors we supply