Why is a transformer Noisy? Constant Voltage Transformer (CVT) noise reduction
Noise is defined as unwanted sound.
Transformer 'humming has been known to soothe people (which makes it a sound) but generally it is thought to be a nuisance (which makes it a noise).
- Transformer noise is produced by the core
- The amount of noise is generally fixed by the design of the transformer
- Adjustments to a design to reduce the noise level can be made at cost but don't expect a huge reduction in the noise level
- Loading a transformer has little effect on the noise level
- Vibrations are produced as well as noise and these are just as important
- Noise Baffling can be done at a cost
Transformer noise is caused by a phenomenon called magnetostriction. In very simple terms this means that if a piece of magnetic sheet steel is magnetized it will extend itself. When the magnetization is taken away. I goes back to its original condition. A transformer is magnetically excited by an alternating voltage and current so that it becomes extended and contracted twice during a full cycle of magnetization.
A transformer core is made from many sheets of special steel. It is made this way to reduce losses and to reduce the consequent heating effect. If the extensions and contractions described above are taking place erratically all over a sheet, and each sheet is behaving erratically with respect to its neighbor, then you can get a picture of a moving, writhing construction when it is excited. Of course, these extensions are only small dimensionally, and therefore cannot usually be seen by the naked eye. They are, however, sufficient to cause a vibration, and as a result noise.
The act of magnetization by applying a voltage to a transformer produces a flux, or magnetic lines of force in the core. The degree of flux will determine the amount of magnetostriction (extensions and contractions) and hence, the noise level. This is the reason why Constant Voltage Transformers ( CVTs) are noisier than conventional transformers, their cores are SATURATED ie running at high flux
We have established that the transformer hum is caused by the extension and contraction of the core laminations when magnetized. Under alternating fluxes, we can expect this extension and contraction to take place twice during a normal voltage or current cycle. This means that the transformer is vibrating at twice the frequency of the supply, i.e. for 50 cycles per second supply frequency, the noise or vibration is moving at 100 cycles per second. This is called the fundamental noise frequency. Nothing is this world is ever perfect and so it is with transformer cores. Since the core is not symmetrical and the magnetic effects do not behave in a simple way, the resultant noise is not pure in tone. That is the noise or vibration produced is not only composed of a 100c/s frequency, we find from practical work that transformer noise is made up of frequencies of odd multiples of the fundamental known as 1st, 3rd, 5th and 7th harmonics.
This means we get noise frequencies of 100 (1st), 300 (3rd), 500 (5th), 700 (7th) cycles per second. They are not equally important for we find that the first and third harmonics predominate and produce most of the transformer sound.
It is important to know this because, with this knowledge, we can measure the amount of noise at these frequencies and determine whether amongst a number of other noises, we really are picking up a transformer noise.
It is usually asked - 'what proportion of the transformer noise is contributed by the windings and does the noise increase as the load increases?'. There are, of course, mechanical forces existing between individual conductors in a winding when the transformer is excited. These forces will produce a vibration and a noise, but only one which is pure in tone, i.e. at twice the exciting frequency 100 cps. This, however, is swamped by the fundamental and harmonics produced by the core. The difference between no load and full load, at constant flux density is usually no greater than 1 or 2dB. An exception to this is when special flux shields are placed inside a transformer tank to reduce stray flux effects.
It has been explained that the noise from a transformer is caused by mechanical movement of the individual lamination of the core under magnetization. The pulsation will cause not only air disturbances, thus producing noise, but also physical vibration of the core structure and everything attached to it. The vibration will have similar frequencies to those measured in the noise analysis. Baffling transformer noise and forgetting to isolate the vibrations will only lead to a disappointing result Remember noise is usually air borne. Vibration is ground borne.
We talk about dB's (decibels) but what do they really mean? In simple terms, we are trying to take what we hear and relate it to scientifically measurable terms. The decibel as used in acoustics are a measurement comparing the generated noise against a standard level. We refer to dBA. The 'A' part refers to a position on a sound level meter which more closely follows the human ear. It is important when taking measurements to specify if the noise level was taken on the 'A' weighted scale. It is always necessary to measure the background (ambient) noise level before you start and when you finish the tests. There has to be a difference between the ambient reading and the average noise level of 7dB ,or better, for it to be valid.
Attenuation and Baffling
The most obvious strategy is to place the transformer in a field miles away from habitation. The noise level drops away as the square of the distance from the noise.
- Put the object in room in which the walls, floor are massive enough to reduce the noise to a person listening on the other side. Noise is usually reduced (attenuated) as it tries to pass through a massive wall. Walls can be of brick, steel, concrete, lead, etc
- Put the object inside an enclosure which uses a limp wall technique. This is a method which uses two thin plates separated by viscous (rubbery) material. The noise hits the inner sheet - its energy (some) is used up inside the viscous material. The outer sheet should not vibrate
- Build a screen wall around the unit. This is cheaper than a full room. It will reduce the noise to those near the wall, but the noise will get over the screen and fall elsewhere (at a lower level). Screens have been made from wood, concrete, brick and with . dense bushes (although the latter becomes psychological)
- Do not make any reflecting surface coincident with half the wave length of the frequency. If a noise hits a reflecting surface at these dimensions it will produce what is called a standing wave. Standing waves will cause reverberations (echoes) and an increase in the sound level. If you hit these dimensions and get echoes you have to apply absorbent materials to the offending walls (fiberglass, wool, etc.)
- Isolate the core and coils of the transformer from the ground
- Use isolating materials guaranteed to eliminate transformer frequencies (at 100 cps upwards)
- Make sure all connections to a solid reflecting surface are flexible. This includes incoming cables, bus bars, stand off insulators, etc. Any solid connection from the vibrating transformer to a solid structure will transmit vibration
- Make sure shipping bolts are removed so that they do not short circuit anti-vibration pads
Expected Sound Level in Decibles (dBA)
|KVA||Constant Voltage Transformer||Conventional transformer|