UNDERSTANDING REDUCING TRANSFORMER

TRANSFORMER NOISE NOISE AT THE SOURCE

Noise is defined as unwanted sound. The obvious question is why not re-
But, what is unwanted sound? A mel- duce the noise in the core by reducing
low sound to some, can be completely the amount of flux. Why? Because it is
unacceptable to others. Attending rock not that simple.
concerts with noise levels at eardrum
rattling levels is totally stimulating to Transformer voltages are fixed by
many people. Put those same people in system requirements, and the amount
a different environment, possibly next of magnetization, by the ratio of these
to a transformer, and there will be wild voltages to the number of turns in the
protestations. The difference then be- winding. The decision on what this
tween noise and sound is in the “ear of ratio of voltage to turns will be, is
the hearer”. Since it is necessary to place made for reasons, mainly economic.
electrical apparatus alongside a wide
WHAT MAKES A It means that the amount of flux at the
spectrum of people we have to accept the TRANSFORMER HUM? normal voltage is invariably fixed,
inevitable, that even under normal condi- Transformer noise is caused by a phe- thus setting the noise and vibration
tions, somebody will always complain. nomenon called magnetostriction. In level. Also, increasing (or decreasing)
very simple terms this means that if a magnetization does not increase or
piece of magnetic sheet steel is mag- decrease the magnetostriction by the
netized it will extend itself. When the same amount. In technical terms the
magnetization is taken away, it goes relationship is not linear. Therefore,
back to its original condition. A trans- when we are asked, as we invariably
former is magnetically excited by an are,– “can you reduce the noise level
alternating voltage and current so that at the source?” – the answer is that it
it becomes extended and contracted can be done, at a cost and for not much
twice during a full cycle of magnetiza- improvement in noise level.
tion.
EXTEND
TRANSFORMER NOISE
FREQUENCIES
Transformer “humming” has been FLUX
CONTRACT
We have established that the trans-
known to soothe people (which makes it former hum is caused by the extension
a sound) but generally it is reckoned to and contraction of the core laminations
be a nuisance (which makes it a noise). EXTEND when magnetized. Under alternating
The causes and reduction of transformer fluxes, we can expect this extension
noise has been the subject of many This extension and contraction is not and contraction to take place twice
learned papers for at least two genera- uniform, consequently the extension during a normal voltage or current
tions. It has come to prominence again, and contraction varies all over a sheet. cycle. This means that the transformer
mainly because transformers are placed is vibrating at twice the frequency
A transformer core is made from many of the supply, i.e. for 60 cycles per
closer to the populace—in high rise of- sheets of special steel. It is made this
fice buildings, apartments, shopping second supply frequency, the noise or
way to reduce losses and to reduce vibration is moving at 120 cycles per
malls and in their gardens. It is becoming the consequent heating effect. If the
even more necessary to locate these units second. This is called the fundamental
extensions and contractions described noise frequency.
carefully and some planning, preferably above are taking place erratically all
ahead of time, is needed. over a sheet, and each sheet is be- Nothing in this world is ever perfect
The remedies we use to counter pos- having erratically with respect to its and so it is with transformer cores.
sible objections to transformer noises neighbor, then you can get a picture of Since the core is not symmetrical and
are varied and in some cases, expensive, a moving, writhing construction when the magnetic effects do not behave in
because we cannot produce a blanket it is excited. Of course, these exten- a simple way, the resultant noise is not
remedy to cover all situations. It is ab- sions are only small dimensionally, pure in tone. That is the noise or vibra-
solutely necessary to consider each case and therefore cannot usually be seen tion produced is not only composed
on its merits, to apply the general rules by the naked eye. They are, however, of a 20c/s frequency, we find from
of acoustic technology and to be familiar sufficient to cause a vibration, and as a practical work that transformer noise
with the causes of transformer noise. result noise. is made up of frequencies of odd mul-
The techniques can be explained simply tiples of the fundamental known as 1st,
The act of magnetization by applying 3rd, 5th and 7th harmonics.
enough for anyone to understand and the a voltage to a transformer produces a
rules are, in the main easy to apply. flux, or magnetic lines of force in the
The BEST RULE however is to PLAN core. The degree of flux will determine
AHEAD. Finding out you have a noise the amount of magnetostriction (exten-
problem (or vibration problem) after the sions and contractions) and hence, the
placement of the unit is costly, time con- noise level.
suming and frustrating.

2
 NOISE WAVES However, each type requires special
consideration and treatments, and it is
important that these are understood.
Transformer size, requirements, and it
is important that these are understood.
ARE Transformer size, requirements and ap-
LIKE THIS plications will determine more exactly
where and how a transformer is placed,
NOT but there are certain treatments which
LIKE
THIS are common to all type. First, let us
consider how transformer noise is mea-
sured.
This means we get noise frequencies
of 120 (1st), 360 (3rd), 600 (5th), 840 Reducing (attenuating) these mechani- MEASURING
(7th) cycles per second. They are not cal pressure pulsations is vital to noise TRANSFORMER NOISE
equally important for we find that the and vibration control and consequently, We talk about dB’s (decibels) but what
first and third harmonics predominate isolating the core and coils of a trans- do they really mean? In simple terms,
and produce most of the transformer former, either in the tank or through a we are trying to take what we hear and
sound. tank, or just as the core and coils, is im- relate it to scientifically measurable
portant. Baffling transformer noise and terms. The decibel as used in acoustics
It is important to know this because, forgetting to isolate the vibrations will is a measurement comparing the pres-
with this knowledge, we can measure only lead to a disappointing result and is sure generated by a noise against some
the amount of noise at these frequen- something which should not be done. standard level. These decibels will vary
cies and determine whether amongst a according to the frequency of the noise,
number of other noises, we really are Remember noise is usually air borne.
but this is taken care of in the noise
picking up a transformer noise. Vibration is ground borne. They are
level meter.
very much connected.
WHAT ABOUT A We refer to dB. The “A” part refers to a
TRANSFORMER ON LOAD? LET US STOP AND position on a sound level meter which
It is usually asked – “what proportion SUMMARIZE more closely follows the human ear. It
of the transformer noise is contributed • Transformer noise is produced by the is important when taking measurements
core. to specify if the noise level was taken
by the windings and does the noise • The amount of noise is generally
increase as the load increases?”. There on the “A” weighted scale.
fixed by the design of the transformer.
are, of course, mechanical forces ex- • Adjustments to a design to reduce the Since the transformer is not necessarily
isting between individual conductors noise level can be made at cost but symmetrical, we cannot take one read-
in a winding when the transformer is don't expect a large reduction in the ing of noise level from a sound level
excited. These forces will produce a vi- noise level. meter and call the noise level of the
bration and a noise, but only one which • Loading a transformer has little effect unit. It is necessary to take many read-
is pure in tone, i.e. at twice the exciting on the noise level.
• Vibrations are produced as well as ings around the transformer and to aver-
frequency –120 cps. This, however, is age them. The resultant will become the
swamped by the fundamental and har- noise and these are just as important
as the noise. transformer noise level.
monics produced by the core.
The industry standard defining audible
The difference between no load and TRANSFORMER TYPES sound level testing for low voltage
full load, at constant flux density is We have established that the core
and coils of a transformer will, when dry-type transformers is NEMA ST 20
usually no greater than 1 or 2dB. An and IEEE C57.12.91-2011 for medium
magnetized, produce a hum (noise)
exception to this is when special flux and mechanical vibrations, but, the voltage dry-type transformers.
shields are placed inside a transformer transformer category will also have an
tank to reduce stray flux effects. effect on what happens once the noise The test procedure for low voltage
and vibration is produced. dry-type transformers is to take a
VIBRATION- minimum of four measurements one
There are three basic categories cur-
DON’T FORGET IT rently in use:
foot away from the contour of the
It has been explained that the noise transformer surface. The maximum
from a transformer is caused by me- • Those immersed in liquids - oils, interval between these measurements
chanical movement of the individual silicones, etc. is 3 feet. More measurements may be
lamination of the core under magne- • Those immersed in vapors and gases required depending on the height of the
tization. The pulsation will cause not – nitrogen, fluro-carbons, etc. transformer. The tested sound level is
only air disturbances, thus producing • Those mounted in air. the average of all the measurements.
A basic statement can be that irrespec- The computed average may be adjusted
noise, but also physical vibration of the tive of how transformer core and coils
core structure and everything attached depending on the ambient sound levels
are surrounded, noise and vibration in the test area. Similar test procedures
to it. The vibration will have similar will still be transmitted. Oil is incom-
frequencies to those measured in the pressible, and gas and air, we know, are defined in IEEE C57.12.90-2010
noise analysis. transmit sound very effectively. Until and C57.12.91-2011 for medium
we put units in absolute vacuums, we voltage liquid filled and dry-type
have to accept that they will transmit transformers respectively.
sound almost as if all were in air.

3
 • Do not make any reflecting surface
METHODS OF CUTTING AIR coincident with half the wave length of
BORNE NOISE the frequency. What does this mean?
• Put the object in room in which the Well, every frequency has a wave
walls, floor are massive enough to re- length. To find the wave length in air,
duce the noise to a person listening on for instance, you divide the speed of
the other side. Noise is usually reduced sound, in air (generally reckoned as
(attenuated) as it tries to pass through 1130 feet/second) by the frequency.
a massive wall. Walls can be of brick,
steel, concrete, lead, etc. e.g. 1130 = 9.42 feet
120
9.42 feet

To measure amounts of noise in each

frequency range you need a frequency
analyzer. This is a worthwhile acquisi- 1/2 wave length = 4.7 ft. et 120 cps
tion.
If a noise hits a reflecting surface at
It is always necessary to measure the these dimensions it will produce what
background (ambient) noise level is called a standing wave. Stand-
before you start and when you finish ing waves will cause reverberations
the tests. There has to be a difference • Put the object inside an enclosure (echoes) and an increase in the sound
between the ambient reading and the which uses a limp wall technique. level. If you hit these dimensions and
average noise level of 7dB or better, This is a method which uses two thin get echoes you have to apply absorbent
for it to be valid, otherwise you could plates separated by viscous (rubbery) materials to the offending walls (fiber-
be increasing the actual reading of the material. The noise hits the inner sheet glass, wool, etc.).
transformer. This sometimes makes – its energy (some) is used up inside
night owls of the testers! the viscous material. The outer sheet
should not vibrate.
METHODS OF CUTTING
STRUCTURE BORNE
VIBRATION
• Isolate the core and coils of the
transformer from the ground. In air
cooled dry types this means to isolate
the core and coil from its support on
the ground. For an oil filled unit it
means to isolate the core and coil from
its tank base, and isolate its tank base
from the supporting ground.
• Use isolating materials guaranteed to
SO NOW WE KNOW WHERE THE eliminate transformer frequencies (at
NOISE COMES FROM AND HOW 120 cps upwards). This is important.
Not many materials can do this. Seek
TO MEASURE IT. WHAT CAN WE advice on the best anti-vibration pads
DO ABOUT IT? • Build a screen wall around the unit. to use.
This is cheaper than a full room. It will
FLUID FILLED TRANSFORMER
reduce the noise to those near the wall,
First of all, accept that there is a noise but the noise will get over the screen FLEXIBLE ACOUSTIC
ENCLOSURE
and you are stuck with it. We have and fall elsewhere (at a lower level).
CONNECTIONS

to consider how to avoid making it a Screens have been made from wood,
nuisance to people. The most obvious concrete, brick and with dense bushes
strategy is to place the transformer in a (although the latter becomes psycho-
field miles away from habitation. The logical).
noise level drops away as the square of
the distance from the noise, but even
so, it would take a very large field to
hide it. However, we invariably have to
place transformers near people and we POSSIBLE ANTI-VIBRATION PAD POSITIONS
must face up to that fact.
We have both noise and vibration TYPICAL TRANSFORMER
to worry about and as we have said CONNECTIONS
NOISE is usually air borne, VIBRA- — DRY-TYPE TRANSFORMERS
TION is structure borne.

4
 TYPICAL TRANSFORMER CONNECTIONS He must insure that the core and coils A word of warning here. The noise
- DRY TYPE TRANSFORMER
HV FLEXIBLE
are correctly resiliently mounted, for level as measured and given by the
LV FLEXIBLE CONNECTION CONNECTION it they are not, the noise level will transformer manufacturer is usually
increase. The stiffness of the mounts for the core and coils inside a cubicle.
must be such that they do not weaken There is no way that the manufacturer
the installation by being too soft or can assess the effect on the transformer
spoil their attenuation properties by be- noise level by the location in which the
REMOVE REMOVE ing too hard. unit will be place. It is advisable that
SHIPPING SHIPPING
BOLTS BOLTS if a user wants to maintain a particular
The choice of the resilient must be noise level in a particular environment
LV
SWITCHGEAR TRANSFORMER
HV
SWITCHGEAR carefully considered. It has to absorb they should work backwards.
SECTION ANTI VIBRATION transformer frequencies which, by
PADS most commercial shock absorber sys- First of all assess what level is toler-
tems are very low. “Shore” hardness able (say 65dBA). Allow for the ef-
• Make sure all connections to a solid (resilience), ability to withstand the fect of the room (say 3dB). Allow for
reflecting surface are flexible. This in- environment and stiffness sufficient to the efficiency of all the connections
cludes incoming cables, busbars, stand carry the unit, are all important design (say 2dB) and as a result ask for a
off insulators, etc. Any solid connec- parameters. transformer to meet 60dBA! This will
tion from the vibrating transformer to a ensure that the required noise level is
solid structure will transmit vibration. Busbar or other connections to the core met. Advice on how to assess these
must not transmit vibrations. Flex- corrections is available within Federal
TYPICAL ARRANGEMENT FOR ibility is the key. Ventilators must be
H. V. ENTRY FROM BELOW Pacific.
carefully positioned. The core must be
TRANSFORMER TERMINALS
designed to avoid transformer frequen- The design of the room to house the
cies of half wave-length dimensions, or transformer is the next consideration.
CABLE FROM STRESS CONE
TO TERMINAL PROVIDES multiples of these dimensions. If this Avoiding half wave-lengths of trans-
SOME ANTI-VIBRATION
DAMPING BEST CONDITION
cannot be achieved, then consideration former noise, or multiples thereof, is
IS WHEN CABLES ARE LOOSE of damping material applied to the advisable. This includes dimensions in
CLAMP TO case, is required. This is an added cost all directions, including the ceiling. If
ENCLOSURE WALL and must be part of the arrangement these dimensions cannot be avoided,
with the user. then damping treatment is required
STRESS CONE remembering that transformer frequen-
Now comes the interface with the cies are involved. This is a caution
user. For shipment purposes, it is often against using acoustical treatments
necessary to “block out” the core and which are only effective for speech
3 CONDUCTOR CABLE
coils against the case, to avoid ship- frequencies. Choose damping materials
CONDUIT ping damage. This can include holding for the noise frequencies to be damped.
SAME PRINCIPLE APPLIES FOR
down bolts which if left in a fastened
END VIEW OF CUBICLE CABLE ENTRY FROM BELOW condition, can short circuit the anti-vi- Isolation of the transformer from the
bration effects of the resilient mounts. ground is vital.
The manufacturer can draw attention
• Make sure shipping bolts are re- to these bolts by painting them with Installation instructions must ensure
moved so that they do not short cir- a florescent paint, and advising his that nobody tightens down shipping
cuit anti-vibration pads. customer to remove all such marked bolts – but removes them. Connecting
bolts before use. All other blocking and cables must be as flexible as possible.
• Additional information is given in Ventilation ducts must be placed in po-
ANSI C57.94, Section 4.10. wedging, not part of the design, should
be carefully removed since these might sitions where these are effective ther-
interfere with the vibration isolation. mally without affecting the acoustic
WHAT CAN THE The user should be made aware of any performance.
MANUFACTURER DO? of these requirements. After this, it is After taking all the precautions, a noise
The manufacturer must first insure that up to the USER! survey after installation, with the trans-
he achieves the noise level as speci- former excited might be useful.
fied by the appropriate specification. If WHAT CAN THE USER DO?
something unusual is required by way The user thinking should start at the The most profitable thing a USER can
of a very low noise level then there conceptual stage. If he can, he must do is ‘THINK AHEAD!”
should be discussions and agreement consider if he has a noise problem
between the manufacturer and the user, before he specifies his transformer. If Save money, time and future headaches
as to what steps to take. Remember the he does, a noise survey including fre- by considering where to put a trans-
only course left to the manufacturer is quency analysis, would be advisable. former and if necessary consult the
usually to lower the flux density and If for instance, a building is only in the manufacturer for advice. Do that even
this means increased cost. There have conceptual stage, then a little thought if is only to warn the manufacturer of
to be trade-offs between cost and noise beforehand will make sure that trans- an impending problem. It will avoid
annoyance or treatments. formers are not placed in small rever- conflicts later on.
berant rooms next to a proposed board
If the manufacturer is only supplying room, sleeping areas, study areas or
core and coils, then what happens next other occupied areas where the normal
is in the hands of the user, assuming sound level is low. Closets under stairs
all noise level requirements have been seem, very popular for dry type trans-
met. If the core and coil is mounted in formers – but are usually acoustically
a containing cabinet then the manufac- bad. Some discussion with the manu-
turer has some precautions to take. facturer is useful at this point.

5
A SUMMARY - THE DO’S 12. External cooling to the enclosure other shipping hardware painted
requires flexible treatments to red should be removed. See
AND THE DON’TS the connecting pipework. installation manual for details.
It might be useful to review the salient
points and give some extra pointers: 13. Enclosure mounts should be 24. Remember transformers need
separate from the transformer cooling air in rooms. Be careful
1. Transformer noise is difficult base – or at least, isolated some- (acoustically) when you position
to change at the source. Flux how. air ducts, ventilators and grilles,
density reduction is the main etc.
thrust, but this means increased 14. All connections - cables, etc. to
cost. enclosed transformers should be 25. Pay attention to flexible connec-
flexible. tions inside rooms containing
2. Transformer core constructions transformers.
help to a degree. Reputable 15. Remember bushings vibrate and
manufacturers will use good losses (acoustic) are experienced 26. Select rooms which are not near
joints, flat steel, consistent through them. Flexible acoustic potential complaint areas.
thickness, good core supports, few protection between enclosure and
bolts, etc. bushings are needed. 27. Check the voltage on the system.
Increased flux density by having a
3 Transformer current loading 16. Bushings used in an enclosure higher than normal system voltage
has little or no effect on the noise might have to have a longer will raise the noise level. Most
level. ground sleeve to accommodate Federal Pacific transformers are
the enclosure roof distances. provided with tap connections to
4. Placing transformers in liquid accommodate higher voltages if
(oil) does not help since oils are 17. Pay close attention to access necessary. This will prevent the
incompressible. doors and removable covers flux density from being elevated.
on enclosures. Tight fits are See instruction manual for details
5. Vibration – isolating core and essential. on how to change these taps and
coils within a tank does assist make the connections properly
vibration isolation although 18. Watch the dimensions of rooms
isolation of the whole tank is still in which units are mounted. 28. When assessing the required
needed. Damp them if necessary, suitable noise level of a transformer work
for transformer frequencies. backwards from the required noise
6. Noise reduction by distance is level at a location. Consider the
the simplest form of attenuation. 19. Damping materials are
needed if standing waves or inefficiencies of the site.
If it can be achieved without cost–
excellent. Usually it cannot. reverberations are possible. 29. Consider very carefully where
20. Choose damping materials transformers will be mounted. Re-
7. Noise reduction by screens, silient structures such as wooden
bushes, etc., is the next simplest compatible with trans former
frequencies. mezzanines might be harmful.
but use should be made of the
topography of the site. Remember
the shadow effect means the noise 21. Carry out sound surveys
could be heard outside the shadow before and after installations. It has not been possible to give all the
of that screen. Remember to do a frequency points and suggestions that might as-
analysis so that transformer noise sist a user in producing a trouble free
8. Full enclosure is usually the can be differentiated. (noise) site. However, we are always
only option left to a troublesome available for assistance.
transformer. 22. Anticipate transformer noise
problems when accommodating The purpose of this leaflet is to make
9. Full enclosure can be made of them inside a building - espe- people aware of the important points.
any material with a high mass/ cially for dry types. If we have encouraged users to plan
weight ratio. Brick, concrete, ahead with their noise problems then
steel have been used. Expect 25- 23. Pay careful attention to
removing unnecessary bolting we have succeeded.
30dBA reductions.
or stiffening used originally
10. Full enclosures using masonry for shipping. Federal Pacific
products are not easily transformers are provided with
demountable. Prefabricated vibration isolators under the core
concrete block is the best for this and coil assembly. These isolators
application. are made ineffective during
shipment by shipping bars and/
11. Steel, mass or limp panel or tightened bolt-down hardware.
techniques make good Upon installation, this shipping
demountable enclosures. A 15 – hardware and fasteners should be
20dBA reduction is possible with removed or loosened according
properly designed enclosures. the installation manual. All

6
Notes

7
 601 Old Airport Road • Bristol, VA 24201
FP-TRANSNOISE-02/16 (276) 669-4084 • FAX (276) 669-1869
© 2016 Electro-Mechanical Corporation federalpacific.com