The normal hum of a transformer comes from magnetostriction: the core steel lamination stretches and contracts with the magnetic flux twice per cycle, so it sounds at twice the grid frequency (120 Hz on 60 Hz grids, 100 Hz on 50 Hz grids). Winding noise under load and noise from fans and pumps add to it. A change is what should concern you: a hum louder than before, clicking, crackling or knocking, which may indicate overvoltage, loose parts or internal discharges.
Where the hum comes from
A transformer core is made of silicon steel laminations. When magnetic flux runs through them, the laminations change dimensions slightly: they lengthen and shorten. This phenomenon is called magnetostriction and it does not depend on the direction of the flux, only on its magnitude. That is why the core vibrates twice per cycle: on a 60 Hz grid (North America, Colombia, Mexico) the fundamental tone is 120 Hz; on a 50 Hz grid (Europe, most of Asia and Africa) it is 100 Hz. Harmonics of that tone also appear.
The vibration passes from the core to the structure, the oil and the tank, which radiates it as sound. This core noise exists even when the transformer carries no load and depends mainly on the flux density at which it works: the higher the applied voltage, the more flux density and the more noise. For transformer overfluxing and load rejection, see transformer overfluxing and load rejection.
Other sources add to it:
- Load noise. The current in the windings produces electromagnetic forces that also vibrate at twice the grid frequency; they grow with load.
- Cooling. Fans and pumps on ONAF and OFAF transformers, often noisier than the core itself.
- Energization. When the transformer is switched on, the inrush current saturates the core for a few cycles and the sound is louder and different; this is normal and disappears within seconds.
The core and its losses are explained in transformer core and windings.
What the technical literature says about the tone and its harmonics
A Chinese study on the relationship between core magnetostriction and the noise characteristics of the tank of power transformers confirms one fact. Core vibration has twice the frequency of the excitation current, that is 100 Hz on a 50 Hz grid. The noise measured outside the tank also contains whole multiples of that frequency: 200, 300 and 400 Hz. On a 60 Hz grid those tones are 120, 240, 360 and 480 Hz. The hum of a healthy transformer is therefore a set of spectral lines and not a pure tone. More on how transformer noise is measured with microphones on a contour in transformer noise measurement.
Other research from the Chinese power sector, including a 2025 article in the journal 电源学报 (Dianyuan Xuebao), identifies one of the main reasons the vibration and noise of a large transformer increase. It is the DC imbalance in the core, called DC magnetic bias. That article, which simulates and compares against field measurements, also notes that the bias alters the magnetostriction of the steel and increases losses.
As for how to measure the sound level, the international standard is IEC 60076-10 (2016 edition, which replaced the 2001 one). It defines sound pressure and sound intensity methods to determine the sound power level of transformers, reactors and their cooling equipment. It is intended mainly for factory measurements; on site it is applied as far as possible, and conditions can be very different because of the proximity of other objects, including other transformers. A technical summary of the standard explains that core noise depends on the flux density in the laminations and the properties of the steel. It is therefore independent of the load current, while the forces in the windings depend on the current. In North America, sound level measurement is covered by IEEE C57.12.90, and China uses a national equivalent, GB/T 1094.10. On the sound power level of a transformer: transformer sound power and far field.
Normal noises and alarm noises
| Sound | Probable cause | What to do |
|---|---|---|
| Low, constant hum | Core magnetostriction | Normal; record as a reference |
| Hum that rises at certain hours | Higher operating voltage at those hours, or more load | Measure voltage and load; check the tap changer position |
| Hum louder than months ago | Sustained overvoltage, DC in the load, harmonics or loose clamping | Measure voltage and harmonics; inspect fixings and tank |
| Vibrating panels, covers or pipes | Mechanical resonance of loose parts with the core tone | Tighten and dampen; not an electrical fault |
| Crackling, sizzling | Partial discharges or corona at bushings or connections | Check soon: thermography, bushing inspection, gas analysis |
| Internal clicking or knocking | Internal discharges, intermittent arc or a loose part inside the tank | Gas analysis and electrical tests before continuing to operate |
| Bearing noise | Worn fans or pumps | Cooling system maintenance |
The causes most often seen in plants
Voltage above nominal. If the grid delivers a high voltage and the de-energized tap changer is in a position that raises it further, the core works at higher flux density. It sounds louder, heats more and draws more magnetizing current. In extreme cases this is called overexcitation, which large transformers monitor with the 24 V/Hz function.
DC in the core. Some loads, such as half-wave rectifiers or certain faulty equipment, introduce a DC component that pushes the core toward saturation during half a cycle. The sound becomes harsher and even harmonics appear.
Load harmonics. Drives and rectifiers add frequencies that excite other resonances of the structure and change the timbre of the hum. See variable frequency drive harmonics.
Loose clamping. Over the years and with short-circuit stresses, the core clamping bolts, the winding spacers or the tank parts loosen. Noise increases and, in the case of the windings, so does the risk at the next short circuit.
Discharges and arcs. A contaminated bushing, a loose connection or an insulation defect produces clicking or crackling. This is the noise that matters most, because it can precede a failure.
- Is there clicking, crackling or knocking?
- Yes: Possible discharges or arcing; gas analysis, thermography and electrical tests soon
- No: Is the operating voltage above nominal?
- Yes: Check the tap changer position and the grid voltage
- No: Are covers, panels or pipes vibrating?
- Yes: Mechanical resonance; tighten fixings
- No: Measure harmonics and DC component; check internal clamping
How it is diagnosed
- Voltage and harmonics measurement at the terminals over several days.
- Thermography of bushings and connections, and a close visual and aural inspection with the equipment in service, keeping safe distances.
- Dissolved gas analysis on oil-filled transformers: it distinguishes a thermal fault, partial discharges and arcing. See dielectric oil analysis.
- Out-of-service tests when there are signs of damage: excitation current (changes in the core), winding resistance, insulation power factor and, if winding movement is suspected, frequency response analysis.
- Sound level measurement, when the noise is an environmental problem or must be compared with the guaranteed value. The measurement standard is IEC 60076-10, intended mainly for the factory; on site, background noise and reflections require careful interpretation.
When the problem is the neighborhood
Sometimes the transformer is healthy and the problem is that it annoys people: a transformer room next to offices or a substation near homes. The solutions are mechanical: anti-vibration mounts, decoupling of the structure and pipes, acoustic barriers or enclosures and, on a new unit, specifying a maximum noise level at purchase.
How we help
If the sound of your transformer has changed, our team performs the on-site diagnosis in Colombia, or elsewhere on request: thermography, oil analysis, electrical tests and voltage and harmonics measurement, with a report that says whether intervention is needed and how urgently. Reviewing results and reports remotely is available worldwide. See our power transformer testing and our testing and commissioning service. Write to us with the equipment nameplate and a description of the noise.
Related guides
- Plant transformer failures
- Transformer core and windings
- Transformer excitation current
- 24 V/Hz overexcitation
- Why a transformer overheats
- Why a transformer burns out
- Dielectric oil analysis
- Transformer maintenance
- Transformer Overfluxing and Load Rejection
- Transformer Sound Power: No-Load, Load and Far Field
- Transformer Noise Measurement: IEC 60076-10 Method
References
- IEC 60076-10, Power transformers – Part 10: Determination of sound levels.
- IEEE C57.12.90, Standard Test Code for Liquid-Immersed Distribution, Power, and Regulating Transformers (sound measurement).
- IEEE C57.104 and IEC 60599, interpretation of dissolved gases.
Frequently asked questions
Is it normal for a transformer to make noise?
Yes. Every energized transformer hums because of core magnetostriction, even with no load. The level depends on the design, the size and the flux density the core works at. What is not normal is a change in that sound or the appearance of clicking and crackling.
Why is my transformer humming louder than before?
The usual causes are a higher operating voltage than before (more flux density in the core), a DC component or harmonics in the load, core clamping or tank parts that have loosened, or fans with worn bearings. If there is also clicking, there may be internal discharges and it should be checked soon.
Does a transformer hum mean it is going to burn out?
The normal hum does not. A new crackling, sizzling or knocking noise can precede a failure, because it may come from partial discharges, an arc at a connection or a loose part. It is confirmed with dissolved gas analysis, thermography and electrical tests.