What Is Magnetostriction of Silicon Steel?
A transformer hums at a hundred hertz, and the sheet itself is the source. Magnetostriction of silicon steel explains that sound. The strain repeats at twice line frequency, so three parties share control of it. Jiangsu Zhongtong Weiye Special Steel Co., Ltd. traces the chain below, starting from silicon steel sheet as the raw input.
The Short Answer
Magnetostriction is a length change that follows the flux, and it runs at double the supply frequency. A 50 Hz winding therefore produces a 100 Hz tone, because the strain squares the rate. Three parties decide how loud that tone becomes: the mill, the core designer, and the assembler. Hence no single purchase order controls the noise, though buyers often expect one.
Why the Tone Doubles
Strain follows the square of the flux, and the flux swings through a full cycle each period. A squared sine carries a steady part plus a term at twice the rate. That doubling is arithmetic rather than accident, since a squared sine behaves this way. Therefore, a 50 Hz supply yields a 100 Hz hum, while a 60 Hz supply yields 120 Hz.
How Small the Movement Is
A 3 percent silicon sheet strains about one to three parts per million near 1.5 T. Over a 500 mm limb that reaches half a micrometre to one and a half, since the strain is tiny. A human hair measures near 70 micrometres, so the movement is roughly a hundredth of it. Hence the movement is tiny, yet the sound is not.

Why So Small a Movement Still Sings
At 100 Hz the acoustic wavelength in air reaches 3.43 metres, which dwarfs a half metre limb. The whole core therefore moves in phase, although a small source radiates inefficiently into air. The tank wall then does the amplifying, since a large panel flexes as one diaphragm. Therefore, the noise you hear belongs to the assembly rather than to the strip. That is why a strip swap rarely helps.
Party One — The Mill
Chemistry decides the baseline strain, because carbon, sulphur, and oxygen each disturb the lattice. A Goss texture lowers the strain along the rolling direction. An oriented grade then hums less along its easy axis. Grain size shifts the figure as well, and coating tension adds a small static bias. A non-oriented sheet carries a larger strain than an oriented one. Hence the mill sets a floor that no downstream step removes, so a noisy grade stays noisy.
Party Two — The Core Designer
Peak induction drives the strain hardest, since the curve steepens above 1.5 T, so a small cut pays. Crowding at a corner or a T-joint lifts the local flux beyond the nominal figure. A design running at 1.7 T in the limb may reach 1.9 T at a corner. Flux crowds where the path narrows. Therefore, a modest induction cut buys more quiet than a grade upgrade.
Party Three — The Assembler
Clamp pressure matters, since an over-tightened core adds stress and raises the strain. A finger-tight stack then hums lower than a cranked one. A loose stack rattles instead, so a broadband rattle lands on top of the tone. Joint gaps shift the local flux, while a step-lap spreads the effect over several teeth. Hence torque specs belong in the build sheet, rather than in a mechanic’s judgement.

Harmonics Beyond the Fundamental
A higher induction pushes the strain curve away from a clean square law, hence fourth and sixth harmonics appear. Those components sit at 200 and 300 Hz, because the ear is more sensitive there. A core running near saturation therefore sounds harsher rather than merely louder. Hence the peak induction figure matters twice, once for loss and once for tone.
What a Lower Loss Grade Does Not Fix
Loss and strain are separate properties, so a low-loss grade may hum the same or more. A buyer who swaps a grade to chase quiet often measures nothing new, since the strain curve barely moved. The flux density note explains why the two curves move apart. Therefore, quiet design needs a strain figure, not a watts figure.
How to Compare Two Cores
Measure at one metre with an A-weighted meter, and hold the induction and the mounting identical. A silicon electrical steel sheet from the same heat then removes a chemistry variable. A 3 dB gap means half the sound power, and that difference is audible without being dramatic. Repeat the test at the same hour, since background noise drifts. A cut core makes a steadier test object than a hand stack. Hence a simple protocol beats an expensive claim, since a repeatable test costs nothing.
What Goes Wrong
One buyer blames the strip for a hum that a loose clamp produced, though the sheet never moved. Another upgrades the grade while the design keeps 1.75 T at the corners, so nothing changes. A third measures two cores on different mounts and therefore reads a false gap. Those three errors arrive every quarter. A fourth mistake repeats a measurement on a different mount and calls it a trend.
What We Do Not Do
We do not melt or roll steel. We do not predict a sound level for an assembly we have not seen. Nor do we sell a grade as a noise fix.
Those limits keep the answer honest, and they keep a noise claim measurable.
Specify Your Order
Give the induction, the core mass, and the mounting style, since the last item decides the radiation. Add the supply frequency and the harmonic limit when a standard applies. A limit shapes both the grade choice and the test method. A CRGO coil then travels with its strain note alongside the loss curve. That reply usually takes a day.
Related Reading
The hysteresis note covers the other half of core loss. Our insulation note explains the coating that also sets tension. A hub page answers six basic questions.
Frequently Asked Questions
At twice line frequency.
About one part per million.
Rarely, since strain is not thickness driven.
Yes, within a defined torque range.
Not necessarily, since the properties differ.
- Email: vera.liang@ztwysteel.com
- Location: Xishan District, Wuxi, Jiangsu, China
- Supply range: CRGO and CRNGO 0.20–0.65 mm, four coating classes
- Slitting: 20–1,250 mm, ±0.05 mm, burr under 0.02 mm
- Documentation: MTC EN 10204 3.1 with grade, thickness and loss basis, 3.2 on request