Professional Documents
Culture Documents
Sasidhara Rao
Transformers
1 Introduction
Transformers can link two or more electric circuits. In its simple form two electric
circuits can be linked by a magnetic circuit, one of the electric coils is used for the creation
of a time varying magnetic filed. The second coil which is made to link this field has an
induced voltage in the same. The magnitude of the induced emf is decided by the number
of turns used in each coil. Thus the voltage level can be increased or decreased by changing
the number of turns. This excitation winding is called a primary and the output winding
is called a secondary. As a magnetic medium forms the link between the primary and the
secondary windings there is no conductive connection between the two electric circuits. The
transformer thus provides an electric isolation between the two circuits. The frequency on
the two sides will be the same. As there is no change in the nature of the power, the re-
sulting machine is called a ‘transformer’ and not a ‘converter’. The electric power at one
voltage/current level is only ‘transformed’ into electric power, at the same frequency, to an-
other voltage/current level.
Even though most of the large-power transformers can be found in the power systems,
the use of the transformers is not limited to the power systems. The use of the principle
of transformers is universal. Transformers can be found operating in the frequency range
starting from a few hertz going up to several mega hertz. Power ratings vary from a few
milliwatts to several hundreds of megawatts. The use of the transformers is so wide spread
that it is virtually impossible to think of a large power system without transformers. Demand
on electric power generation doubles every decade in a developing country. For every MVA
of generation the installed capacity of transformers grows by about 7MVA. These figures
show the indispensable nature of power transformers.
2 Basic Principles
coil of N turns. All these N turns link flux lines of φ Weber resulting in the Nφ flux linkages.
In such a case,
ψ = Nφ (2)
and
dφ
e=N volt (3)
dt
The change in the flux linkage can be brought about in a variety of ways
• coil may be static and unmoving but the flux linking the same may change with time.
• flux lines may be constant and not changing in time but the coil may move in space
linking different value of flux with time.
• both 1 and 2 above may take place. The flux lines may change in time with coil moving
in space.
These three cases are now elaborated in sequence below, with the help of a coil with a simple
geometry.
L
B
- +
B = Bm sin ωt (4)
where Bm is the peak amplitude of the flux density. ω is the angular rate of change with
time. Then, the instantaneous value of the flux linkage is given by,
em = Nφm .ω (7)
Fig. 2(b) shows the same example as above but with a small difference. The flux
density is held constant at B Tesla. The flux linked by the coil at the current position is
φ = B.L.X Weber. The conductor is moved with a velocity v = dx/dt normal to the flux,
cutting the flux lines and changing the flux linkages. The induced emf as per the application
of Faraday’s law of induction is e = N.B.L.dx/dt = B.L.v volt.(Here N=1)
Please note,the actual flux linked by the coil is immaterial. Only the change in the
flux linkages is needed to be known for the calculation of the voltage. The induced emf is
in step with the change in ψ and there is no phase shift. If the flux density B is distributed
sinusoidally over the region in the horizontal direction, the emf induced also becomes sinu-
soidal. This type of induced emf is termed as speed emf or rotational emf, as it arises out of
the motion of the conductor. The polarity of the induced emf is obtained by the application
of the Lenz’s law as before. Here the changes in flux linkages is produced by motion of the
conductor. The current in the conductor, when the coil ends are closed, makes the conductor
experience a force urging the same to the left. This is how the polarity of the emf shown in
fig.2b is arrived at. Also the mmf of the loop aids the field mmf to oppose change in flux
linkages. This principle is used in d.c machines and alternators.
The third case under the application of the Faraday’s law arises when the flux changes
and also the conductor moves. This is shown in Fig. 2(c).
The first term is due to the changing flux and hence is a transformer emf. The second
term is due to moving conductor or is a speed emf. When the terminals are closed such as
to permit a current the conductor experiences a force and also the mmf of the coil opposes
the change in flux linkages. This principle is used in a.c. machines where the field is time
varying and conductors are moving under the same.
The first case where there is a time varying field and a stationary coil resulting in
a transformer emf is the subject matter in the present section. The case two will be re-
visited under the study of the d.c machines and synchronous machines. Case three will be
extensively used under the study of a.c machines such as induction machines and also in a.c.
commutator machines.
Next in the study of the transformers comes the question of creating a time varying
filed. This is easily achieved by passing a time varying current through a coil. The winding
which establishes the field is called the primary. The other winding, which is kept in that
field and has a voltage induced in it, is called a secondary. It should not be forgotten that
the primary also sees the same time varying field set up by it linking its turns and has an
induced emf in the same. These aspects will be examined in the later sections. At first
the common constructional features of a transformer used in electric power supply system
operating at 50 Hz are examined.
3 Constructional features
Transformers used in practice are of extremely large variety depending upon the
end use. In addition to the transformers used in power systems, in power transmission and
distribution, a large number of special transformers are in use in applications like electronic
supplies, rectification, furnaces, traction etc. Here the focus is on power transformers only.
The principle of operation of these transformers also is the same but the user requirements
differ. Power transformers of smaller sizes could be air cooled while the larger ones are
oil cooled. These machines are highly material intensive equipments and are designed to
match the applications for best operating conditions. Hence they are ‘tailor made’ to a
job. This brings in a very large variety in their constructional features. Here more common
constructional aspects alone are discussed. These can be broadly divided into
1. Core construction
2. Winding arrangements
3. Cooling aspects
Transformer core for the power frequency application is made of highly permeable
material. The high value of permeability helps to give a low reluctance for the path of
the flux and the flux lines mostly confine themselves to the iron. Relative permeability µr
well over 1000 are achieved by the present day materials. Silicon steel in the form of thin
laminations is used for the core material. Over the years progressively better magnetic prop-
erties are obtained by going in for Hot rolled non-oriented to Hot rolled grain oriented steel.
Later better laminations in the form of cold Rolled Grain Oriented (CRGO), -High B (HiB)
grades became available. The thickness of the laminations progressively got reduced from
over 0.5mm to the present 0.25mm per lamination. These laminations are coated with a thin
layer of insulating varnish, oxide or phosphate. The magnetic material is required to have
a high permeability µ and a high saturation flux density, a very low remanence Br and a
small area under the B-H loop-to permit high flux density of operation with low magnetizing
current and low hysteresis loss. The resistivity of the iron sheet itself is required to be high
to reduce the eddy current losses. The eddy current itself is highly reduced by making the
laminations very thin. If the lamination is made too thin then the production cost of steel
laminations increases. The steel should not have residual mechanical stresses which reduce
their magnetic properties and hence must be annealed after cutting and stacking.
In the case of very small transformers (from a few volt-amperes to a few kilo volt-
amperes) hot rolled silicon steel laminations in the form of E & I, C & I or O as shown in
Fig. 3 are used and the core cross section would be a square or a rectangle. The percentage
of silicon in the steel is about 3.5. Above this value the steel becomes very brittle and also
very hard to cut. The saturation flux density of the present day steel lamination is about 2
Tesla.
Broadly classifying, the core construction can be separated into core type and
shell type. In a core type construction the winding surrounds the core. A few examples of
single phase and three phase core type constructions are shown in Fig. 4. In a shell type on
the other hand the iron surrounds the winding.
In the case of very small transformers the conductors are very thin and round.
These can be easily wound on a former with rectangular or square cross section. Thus no
special care is needed for the construction of the core. The cross section of the core also
would be square or rectangular. As the rating of the transformer increases the conductor size
(a ) (b)
(c)
10
1.phase
3.phase
HV LV core LV HV LV HV
also increases. Flat conductors are preferred to round ones. To wind such conductor on a
rectangular former is not only difficult but introduces stresses in the conductor, at the bends.
From the short circuit force with stand capability point of view also this is not desirable.
Also, for a given area enclosed the length of the conductor becomes more. Hence it results in
more load losses. In order to avoid all these problems the coils are made cylindrical and are
wound on formers on heavy duty lathes. Thus the core construction is required to be such as
to fill the circular space inside the coil with steel laminations. Stepped core construction thus
becomes mandatory for the core of large transformers. Fig. 5 shows a few typical stepped core
constructions. When the core size increases it becomes extremely difficult to cool the same
(Even though the core losses are relatively very small). Cooling ducts have to be provided
in the core. The steel laminations are grain oriented exploiting the simple geometry of the
transformer to reduce the excitation losses. The iron losses in the lamination, when the flux
11
0.14 0.14
0.16 0.1 0.1 0.07 0.12
0.16 0.12 0.07
0.09 0.09
0.53 0.42 0.3
0.71D
d d d
duct
duct
12
is oriented in the direction of grain orientation, is about 30% of that in the normal direction.
Another important aspect to be carefully checked and monitored is the air gaps in
Path of
flux
HV
LV
Windings
Core
(a) (b)
series in the path of the main flux. As the reluctance of air path is about 1000 times more
than that of the steel, an air path of 1mm will require a mmf needed by a 1 meter path in iron.
Hence butt joints between laminations must be avoided. Lap joints are used to pro-
vide alternate paths for flux lines thus reducing the reluctance of the flux paths. Some typical
constructional details are shown in Fig. 6. In some power transformers the core is built up
by threading a long strip of steel through the coil in the form of a toroid. This construction
is normally followed in instrument transformers to reduce the magnetizing current and hence
the errors.
13
Large cores made up of laminations must be rendered adequately stiff by the provi-
sion of stiffening plates usually called as flitch plates. Punched through holes and bolts are
progressively being avoided to reduce heating and melting of the through bolts. The whole
stack is wrapped up by strong epoxy tapes to give mechanical strength to the core which
can stand in upright position. Channels and angles are used for the frame and they hold the
bottom yoke rigidly.
3.2 Windings
14
HV LV
Core
LV
HV
(a)Concentric coil
LV
HV
Core
15
are common with shell type transformers. In the figure the letters L and H indicate the low
voltage and high voltage windings. In concentric arrangement, in view of the lower insulation
and clearance requirements, the LV winding is placed close to the core which is at ground
potential. The HV winding is placed around the LV winding. Also taps are provided on HV
winding when voltage change is required. This is also facilitated by having the HV winding
as the outer winding.
Three most common types of coils viz. helical, cross over and disc coils are shown in Fig. 8.
Helical coils
Disc coils cross over coils
Helical Windings One very common cylindrical coil arrangement is the helical winding.
This is made up of large cross section rectangular conductor wound on its flat side.
The coil progresses as a helix. This is commonly used for LV windings. The insulation
16
requirement also is not too high. Between layers no insulation (other than conductor
insulation) is needed as the voltage between layers is low. The complexity of this
type of winding rapidly increases as the current to be handled becomes more. The
conductor cross section becomes too large and difficult to handle. The eddy current
losses in the conductor rapidly increases. Hence two or more conductors have to be
wound and connected in parallel. The parallel circuits bring in problems of current
sharing between the circuits. Transpositions of the parallel paths have to be adopted
to reduce unequal current distribution. The modern practice is to use continuously
transposed and bunched conductors.
Cross over coils The second popular winding type is the cross over coil. These are made
of circular conductors not exceeding 5 to 6 sq mm in cross section. These are used for
HV windings of relatively small transformers. These turns are wound in several layers.
The length and thickness of each block is made in line with cooling requirements. A
number of such blocks can be connected in series, leaving cooling ducts in between the
blocks, as required by total voltage requirement.
Disc coils Disc coils consist of flat conductors wound in a spiral form at the same place
spiralling outwards. Alternate discs are made to spiral from outside towards the center.
Sectional discs or continuous discs may be used. These have excellent thermal prop-
erties and the behavior of the winding is highly predictable. Winding of a continuous
disc winding needs specialized skills.
Sandwich coils Sandwich windings are more common with shell type core construction.
They permit easy control over the short circuit impedance of the transformer. By
bringing HV and LV coils close on the same magnetic axis the leakage is reduced
and the mutual flux is increased. By increasing the number of sandwiched coils the
17
3.3 Insulation
Scaling advantages make the design of larger and larger unit sizes of transformers
economically attractive. This can be explained as below. Consider a transformer of certain
rating designed with certain flux density and current density. If now the linear dimensions
are made larger by a factor of K keeping the current and flux densities the same the core and
conductor areas increase by a factor of K 2 . The losses in the machine, which are proportional
to the volume of the materials used, increase by a factor of K 3 .The rating of the machine
increases by a factor of K 4 .
18
The surface area however increases by a factor of K 2 only. Thus the ratio of loss per
surface area goes on increasing by a factor of K. The substantial increase in the output is
the major attraction in going in for larger units. However cooling of the transformer becomes
more and more difficult. As the rating increases better cooling techniques are needed.
Simple air cooling of the transformers is adopted in dry type transformers. The limit
for this is reached by the time the rating is a few kVA. Hence air cooling is used in low
voltage machines. This method of cooling is termed as AN(Air Natural). Air Blast(AB)
method improves on the above by directing the blast of air at the core and windings. This
permits some improvement in the unit sizes.
OB(Oil Blast) method is an improvement over the ON-type and it directs a blast of
air on the cooling surface. In the above two cases the flow of oil is by natural convective
forces. The rate of circulation of oil can be increased with the help of a pump, with the
cooling at the surface remaining natural cooling to air. This is termed as OFN (Oil Forced
Natural). If now a forced blast of air is also employed, the cooling method become OFB(
Oil Forced Blast). A forced circulation of oil through a radiator is done with a blast of air
19
Main tank
Radiator
Tubes
(a)
Conservator
Bushing & Breather
water outlet
Radiator
oil pump
water inlet
(b)
Conservator&
Breather
Bushing
Radiator
(c)
20
Figure 9: Some Typical Cooling Arrangements
over the radiator surface. Substantial amount of heat can be removed by employing a water
cooling. Here the hot oil going into the radiator is cooled by a water circuit. Due to the
high specific heat of water, heat can be evacuated effectively. Next in hierarchy comes OFW
which is similar to OFB except that instead of blast of air a forced circulation of cool water
in the radiator is used in this. Some cooling arrangements are shown in Fig. 9.
In many large sized transformers the cooling method is matched with the amount
of heat that is required to be removed. As the load on the transformer changes the heat
generated within also changes. Suitable cooling method can be pressed into service at that
time. This gives rise to the concept of mixed cooling technique.
ON/OB Works as ON but with increased load additional air blast is adopted. This gives
the ratings to be in the ratio of 1:1.5
The temperature rise permitted in the British standard specification for power transformers
are tabulated below.
Type winding oil core
Class A Class B
◦
C ◦
C ◦
C
AN,AB 55 75 - As
ON,OB,OW 60 - 50 for
OFN,OFB 65 - 50 adjacent
OFW 70 - 50 winding
21
Even though the basic functions of the oil used in transformers are a) heat conduc-
tion and b) electrical insulation, there are many other properties which make a particular oil
eminently suitable. Organic oils of vegetative or animal origin are good insulators but tend
to decompose giving rise to acidic by-products which attack the paper or cloth insulation
around the conductors.
Mineral oils are suitable from the point of electrical properties but tend to form sludge.
The properties that are required to be looked into before selecting an oil for transformer
application are as follows:
Insulting property This is a very important property. However most of the oils naturally
fulfil this. Therefore deterioration in insulating property due to moisture or contami-
nation may be more relevant.
Viscosity It is important as it determines the rate of flow of the fluid. Highly viscous fluids
need much bigger clearances for adequate heat removal.
Purity The oil must not contain impurities which are corrosive. Sulphur or its compounds
as impurities cause formation of sludge and also attack metal parts.
Sludge formation Thickening of oil into a semisolid form is called a sludge. Sludge for-
mation properties have to be considered while choosing the oil as the oil slowly forms
semi-solid hydrocarbons. These impede flows and due to the acidic nature, corrode
metal parts. Heat in the presence of oxygen is seen to accelerate sludge formation. If
the hot oil is prevented from coming into contact with atmospheric air sludge formation
22
Acidity Oxidized oil normally produces CO2 and acids. The cellulose which is in the paper
insulation contains good amount of moisture. These form corrosive vapors. A good
breather can reduce the problems due to the formation of acids.
Flash point And Fire point Flash point of an oil is the temperature at which the oil
ignites spontaneously. This must be as high as possible (not less than 160◦ C from the
point of safety). Fire point is the temperature at which the oil flashes and continuously
burns. This must be very high for the chosen oil (not less than 200◦ C).
Inhibited oils and synthetic oils are therefore used in the transformers. Inhibited oils
contain additives which slow down the deterioration of properties under heat and moisture
and hence the degradation of oil. Synthetic transformer oil like chlorinated diphenyl has
excellent properties like chemical stability, non-oxidizing, good dielectric strength, moisture
repellant, reduced risk due fire and explosion.
It is therefore necessary to check the quality of the oil periodically and take corrective
steps to avoid major break downs in the transformer.
There are several other structural and insulating parts in a large transformer. These
are considered to be outside the scope here.
23
4 Ideal Transformer
Earlier it is seen that a voltage is induced in a coil when the flux linkage associated
with the same changed. If one can generate a time varying magnetic field any coil placed in
the field of influence linking the same experiences an induced emf. A time varying field can
be created by passing an alternating current through an electric coil. This is called mutual
induction. The medium can even be air. Such an arrangement is called air cored transformer.
Indeed such arrangements are used in very high frequency transformers. Even though the
principle of transformer action is not changed, the medium has considerable influence on the
working of such devices. These effects can be summarized as the followings.
1. The magnetizing current required to establish the field is very large, as the reluctance
of the medium is very high.
2. There is linear relationship between the mmf created and the flux produced.
5. It is very hard to direct the flux lines as we desire, as the whole medium is homogeneous.
If the secondary is not loaded the energy stored in the magnetic field finds its way
back to the source as the flux collapses. If the secondary winding is connected to a load then
part of the power from the source is delivered to the load through the magnetic field as a link.
The medium does not absorb and lose any energy. Power is required to create the field and
not to maintain the same. As the winding losses can be made very small by proper choice
of material, the ideal efficiency of a transformer approaches 100%. The large magnetizing
24
Primary
Leakage
x flux
Secondary
Mutual flux
(a)
Leakage flux
X
Primary
Mutual flux
Secondary
Iron core
(b)
25
1. Due to the large value for the permeance ( µr of the order of 1000 as compared to
air) the magnetizing current requirement decreases dramatically. This can also be
visualized as a dramatic increase in the flux produced for a given value of magnetizing
current.
2. The magnetic medium is linear for low values of induction and exhibits saturation type
of non-linearity at higher flux densities.
3. The iron also has hysteresis type of non-linearity due to which certain amount of power
is lost in the iron (in the form of hysteresis loss), as the B-H characteristic is traversed.
4. Most of the flux lines are confined to iron path and hence the mutual flux is increased
very much and leakage flux is greatly reduced.
5. The flux can be easily ‘directed’ as it takes the path through steel which gives great
freedom for the designer in physical arrangement of the excitation and output windings.
6. As the medium is made of a conducting material eddy currents are induced in the
same and produce losses. These are called ‘eddy current losses’. To minimize the
eddy current losses the steel core is required to be in the form of a stack of insulated
laminations.
From the above it is seen that the introduction of magnetic core to carry the flux
introduced two more losses. Fortunately the losses due to hysteresis and eddy current for
the available grades of steel is very small at power frequencies. Also the copper losses in the
26
1. Magnetic circuit is linear and has infinite permeability. The consequence is that a van-
ishingly small current is enough to establish the given flux. Hysteresis loss is negligible.
As all the flux generated confines itself to the iron, there is no leakage flux.
2. Windings do not have resistance. This means that there are no copper losses, nor there
is any ohmic drop in the electric circuit.
In fact the practical transformers are very close to this model and hence no major
departure is made in making these assumptions.
Fig. 11 shows a two winding ideal transformer. The primary winding has T1 turns and is
connected to a voltage source of V1 volts. The secondary has T2 turns. Secondary can be
connected to a load impedance for loading the transformer. The primary and secondary are
shown on the same limb and separately for clarity.
27
v1=V1mcosωt
µ
8
φ
io 0
+ +
~ e1 T1 i1
- + i2
+ +
e2 T2 v1=V1sinωt e1 e2
-
+ -
v1=V1cosωt µ
8
φ
i1
+ +
N e1 T1
-
-
i2
+
e2 T2
ZL
-
(c)Loaded machine
28
mmf I0 T1 I0 T1 Aµ
F lux φ = = l = . (9)
Reluctance µA
l
This current is the result of a sinusoidal voltage V applied to the primary. As the
current through the loop is zero (or vanishingly small), at every instant of time, the sum of
the voltages must be zero inside the same. Writing this in terms of instantaneous values we
have,
v1 − e1 = 0 (10)
where v1 is the instantaneous value of the applied voltage and e1 is the induced emf due to
Faradays principle. The negative sign is due to the application of the Lenz’s law and shows
that it is in the form of a voltage drop. Kirchoff’s law application to the loop will result in
the same thing.
29
Here ψ1peak is the peak value of the flux linkages of the primary. The same mutual
flux links the secondary winding. However the magnitude of the flux linkages will be ψ2peak =
T2 .φm . The induced emf in the secondary can be similarly obtained as ,
I1
I2
+
+
V1 E1 E2 V2
- -
The voltages E1 and E2 are obtained by the same mutual flux and hence they are
in phase. If the winding sense is opposite i.e., if the primary is wound in clockwise sense
and the secondary counter clockwise sense then if the top terminal of the first winding is
at maximum potential the bottom terminal of the second winding would be at the peak
potential. Similar problem arises even when the sense of winding is kept the same, but the
30
two windings are on opposite limbs (due to the change in the direction of flux). Hence in
the circuit representation of transformers a dot convention is adopted to indicate the ter-
minals of the windings that go high (or low) together. (Fig. 12). This can be established
experimentally by means of a polarity test on the transformers. At a particular instant of
time if the current enters the terminal marked with a dot it magnetizes the core. Similarly
a current leaving the terminal with a dot demagnetizes the core.
i1 T1 − i2 T2 = i0 T1 but i0 → 0 (16)
If the reference directions for the two currents are chosen as in the Fig. 12, then the above
equation can be written in phasor form as,
T2
I¯1 T1 = I¯2 T2
or I¯1 = .I¯2 (18)
T1
E1 T1 I2
Also = = E1 I1 = E2 I2 (19)
E2 T2 I1
31
Thus voltage and current transformation ratio are inverse of one another. If an impedance
of ZL is connected across the secondary,
Ē2 Ē2
I¯2 = ¯ or Z¯L = ¯ (20)
ZL I2
Ē1 T1 Ē2 T1
Z̄i = ¯ = ( )2 . ¯ = ( )2 .Z¯L (21)
I1 T2 I2 T2
V1 E1
I2
E2 V2
I1
θ1 θ2
φ φ
Finally, the phasor diagram for the operation of the ideal transformer is shown in
Fig. 13 in which θ1 and θ2 are power factor angles on the primary and secondary sides. As
32
the transformer itself does not absorb any active or reactive power it is easy to see that
θ1 = θ2 .
Thus, from the study of the ideal transformer it is seen that the transformer provides
electrical isolation between two coupled electric circuits while maintaining power invariance
at its two ends. However, grounding of loads and one terminal of the transformer on the
secondary/primary side are followed with the provision of leakage current detection devices
to safe guard the persons working with the devices. Even though the isolation aspect is
a desirable one its utility cannot be over emphasized. It can be used to step up or step
down the voltage/current at constant volt-ampere. Also, the transformer can be used for
impedance matching. In the case of an ideal transformer the efficiency is 100% as there are
no losses inside the device.
33
5 Practical Transformer
Finite permeability of the magnetic circuit necessitates a finite value of the current
to be drawn from the mains to produce the mmf required to establish the necessary flux.
The current and mmf required is proportional to the flux density B that is required to be
established in the core.
φ
B = µH; B= (22)
A
where A is the area of cross section of the iron core m2 . H is the magnetizing force which is
given by,
T1
H = i. (23)
l
where l is the length of the magnetic path, m. or
Aµ(iT1 )
φ = B.A = = permeance ∗ mmf (here that of primary) (24)
l
The magnetizing force and the current vary linearly with the applied voltage as long
as the magnetic circuit is not saturated. Once saturation sets in, the current has to vary in
34
a nonlinear manner to establish the flux of sinusoidal shape. This non-linear current can be
resolved into fundamental and harmonic currents. This is discussed to some extent under
harmonics. At present the effect of this non-linear behavior is neglected as a secondary
effect. Hence the current drawn from the mains is assumed to be purely sinusoidal and
directly proportional to the flux density of operation. This current can be represented by a
current drawn by an inductive reactance in the circuit as the net energy associated with the
same over a cycle is zero. The energy absorbed when the current increases is returned to
the electric circuit when the current collapses to zero. This current is called the magnetizing
current of the transformer. The magnetizing current Im is given by Im = E1 /Xm where Xm
is called the magnetizing reactance. The magnetic circuit being lossy absorbs and dissipates
the power depending upon the flux density of operation. These losses arise out of hysteresis,
eddy current inside the magnetic core. These are given by the following expressions:
Ph ∝ B 1.6 f (25)
Pe ∝ B 2 f 2 t2 (26)
For a constant voltage, constant frequency operation B is constant and so are these
losses. An active power consumption by the no-load current can be represented in the input
circuit as a resistance Rc connected in parallel to the magnetizing reactance Xm . Thus the
no-load current I0 may be made up of Ic (loss component) and Im (magnetizing component
35
as )
I¯0 = I¯c − j I¯m (27)
Ic2 Rc – gives the total core losses (i.e. hysteresis + eddy current loss)
2
Im Xm - Reactive volt amperes consumed for establishing the mutual flux.
Finite µ of the magnetic core makes a few lines of flux take to a path through the air.
Thus these flux lines do not link the secondary winding. It is called as leakage flux. As the
path of the leakage flux is mainly through the air the flux produced varies linearly with the
primary current I1 . Even a large value of the current produces a small value of flux. This
flux produces a voltage drop opposing its cause, which is the current I1 . Thus this effect of
the finite permeability of the magnetic core can be represented as a series inductive element
jxl1 . This is termed as the reactance due to the primary leakage flux. As this leakage flux
varies linearly with I1 , the flux linkages per ampere and the primary leakage inductance
are constant (This is normally represented by ll1 Henry). The primary leakage reactance
therefore becomes
xl1 = 2πf ll1 ohm (28)
A similar effect takes place on the secondary side when the transformer is loaded.
The secondary leakage reactance jxl2 arising out of the secondary leakage inductance ll2 is
given by
Finally, the primary and secondary windings are wound with copper (sometimes alu-
minium in small transformers) conductors; thus the windings have a finite resistance (though
36
r1 jxl1 I’2 φ
I1
+
Io +
V1
~ E1 T1
Rc jXm
-
r2 jxl2 I2
V2 ZL E2 T2
-
(a)Physical arrangement
r1 jXl1
I1 I’2 r2 jXl2
I2
Io
Ic Im
E1 E2 ZL V2
V1 Rc
jXm
(b)Equivalent circuit
37
small). This is represented as a series circuit element, as the power lost and the drop pro-
duced in the primary and secondary are proportional to the respective currents. These are
represented by r1 and r2 respectively on primary and secondary side. A practical transformer
sans these imperfections (taken out and represented explicitly in the electric circuits) is an
′
ideal transformer of turns ratio T1 : T2 (voltage ratio E1 : E2 ). This is seen in Fig. 14. I2
in the circuit represents the primary current component that is required to flow from the
mains in the primary T1 turns to neutralize the demagnetizing secondary current I2 due to
the load in the secondary turns. The total primary current
By solving this circuit for any load impedance ZL one can find out the performance of the
loaded transformer.
The circuit shown in Fig. 14(b). However, it is not very convenient for use due to
the presence of the ideal transformer of turns ratio T1 : T2 . If the turns ratio could be made
unity by some transformation the circuit becomes very simple to use. This is done here by
replacing the secondary by a ‘hypothetical’ secondary having T1 turns which is ‘equivalent ’
to the physical secondary. The equivalence implies that the ampere turns, active and reactive
power associated with both the circuits must be the same. Then there is no change as far
as their effect on the primary is considered. Thus
′ ′ I2 ′ ′ ′
V2 = aV2 , I2 = , r2 = a2 r2 , xl2 = a2 xl2 ZL = a2 ZL .
a
T1
where a -turns ratio T2
38
This equivalent circuit is as shown in Fig. ??(a). As the ideal transformer in this
case has a turns ratio of unity the potentials on either side are the same and hence they
may be conductively connected dispensing away with the ideal transformer. This particular
equivalent circuit is as seen from the primary side. It is also possible to refer all the pri-
mary parameters to secondary by making the hypothetical equivalent primary winding on
the input side having the number of turns to be T2 . Such an equivalent circuit having all
the parameters referred to the secondary side is shown in fig. 15.
The equivalent circuit can be derived, with equal ease, analytically using the Kir-
choff’s equations applied to the primary and secondary. Referring to fig. 14(a), we have (by
neglecting the shunt branch)
Multiply both sides of Eqn.34 by ‘a’ [This makes the turns ratio unity and retains the power
invariance].
aE2 = aV2 + aI2 (r2 + jxl2 ) but aE2 = E1 (36)
39
A similar procedure can be used to refer all parameters to secondary side. (Shown in fig. 15.)
40
6 Phasor diagrams
Io
Ic Im
Rc V’2 Z’L
V1 jXm
(a)
r1 jxl1 r’2 jx’l2 R jX
I1 I’2 I1 I’2
R=r1+r’2
Io
Ic Im
Z’L V’2 x=xl1+x’l2
V1 Rc V1 V’2
jxm
I1=I’2
(b) (c)
41
On similar lines to the ideal transformer the phasor diagram of operation can be
drawn for a practical transformer also. The positions of the current and induced emf phasor
are not known uniquely if we start from the phasor V1 . Hence it is assumed that the phasor
φ is known. The E1 and E2 phasor are then uniquely known. Now, the magnetizing and loss
components of the currents can be easily represented. Once I0 is known, the drop that takes
place in the primary resistance and series reactance can be obtained which when added to
E1 gives uniquely the position of V1 which satisfies all other parameters. This is represented
in Fig. 17(a) as phasor diagram on no-load.
42
V1
IoX l1
E1 Ior1 E2
Io
φ φ
Im Il
(a)No-load
V1
I1X l1
I1r1
E1 I2x2 I r
2 2
E2 I2
V2
I’2 Il
Io
φ φ
(b)On-load
43
7 Testing of Transformers
This is nothing but the resistance measurement of the windings by applying a small
d.c voltage to the winding and measuring the current through the same. The ratio gives
the winding resistance, more commonly feasible with high voltage windings. For low voltage
windings a resistance-bridge method can be used. From the d.c resistance one can get the
a.c. resistance by applying skin effect corrections.
44
V3
S
A2
Vs V1 a2 a2
A2
~ V2
+ +
V
-
A1 a1 a1
A1
(a)A.C.test (b)D.C.test
This is needed for identifying the primary and secondary phasor polarities. It is
a must for poly phase connections. Both a.c. and d.c methods can be used for detecting
the polarities of the induced emfs. The dot method discussed earlier is used to indicate the
polarities. The transformer is connected to a low voltage a.c. source with the connections
made as shown in the fig. 18(a). A supply voltage Vs is applied to the primary and the
readings of the voltmeters V1 , V2 and V3 are noted. V1 : V2 gives the turns ratio. If V3 reads
V1 −V2 then assumed dot locations are correct (for the connection shown). The beginning and
end of the primary and secondary may then be marked by A1 − A2 and a1 − a2 respectively.
If the voltage rises from A1 to A2 in the primary, at any instant it does so from a1 to a2 in
the secondary. If more secondary terminals are present due to taps taken from the windings
they can be labeled as a3 , a4 , a5 , a6 . It is the voltage rising from smaller number towards
larger ones in each winding. The same thing holds good if more secondaries are present.
45
Fig. 18(b) shows the d.c. method of testing the polarity. When the switch S is closed if the
secondary voltage shows a positive reading, with a moving coil meter, the assumed polarity
is correct. If the meter kicks back the assumed polarity is wrong.
A Io
Im Ic
V1 V V2
V1 jXm
Rc
(a)Physical Arrangement
(b)Equivalent Circuit
As the name suggests, the secondary is kept open circuited and nominal value
of the input voltage is applied to the primary winding and the input current and power are
measured. In Fig. 19(a) V, A, W are the voltmeter, ammeter and wattmeter respectively.
Let these meters read V1 , I0 and W0 respectively.Fig. 19(b) shows the equivalent circuit of
the transformer under this test. The no load current at rated voltage is less than 1 percent of
nominal current and hence the loss and drop that take place in primary impedance r1 + jxl1
due to the no load current I0 is negligible. The active component Ic of the no load current I0
46
represents the core losses and reactive current Im is the current needed for the magnetization.
Thus the watt meter reading
W0 = V1 Ic = Pcore (38)
W0
∴ Ic = (39)
V
q1
∴ Im = I02 − Ic2 or (40)
V1 V1
Rc = andXm = (41)
Ic Im
V1
Io
The parameters measured already are in terms of the primary. Sometimes the pri-
mary voltage required may be in kilo-Volts and it may not be feasible to apply nominal
voltage to primary from the point of safety to personnel and equipment. If the secondary
voltage is low, one can perform the test with LV side energized keeping the HV side open
circuited. In this case the parameters that are obtained are in terms of LV . These have to
be referred to HV side if we need the equivalent circuit referred to HV side.
47
Sometimes the nominal value of high voltage itself may not be known, or in doubt,
especially in a rewound transformer. In such cases an open circuit characteristics is first
obtained, which is a graph showing the applied voltage as a function of the no load current.
This is a non linear curve as shown in Fig. 20. This graph is obtained by noting the current
drawn by transformer at different applied voltage, keeping the secondary open circuited. The
usual operating point selected for operation lies at some standard voltage around the knee
point of the characteristic. After this value is chosen as the nominal value the parameters
are calculated as mentioned above.
The purpose of this test is to determine the series branch parameters of the equiv-
alent circuit of Fig. 21(b). As the name suggests, in this test primary applied voltage, the
current and power input are measured keeping the secondary terminals short circuited. Let
these values be Vsc , Isc and Wsc respectively. The supply voltage required to circulate rated
current through the transformer is usually very small and is of the order of a few percent
of the nominal voltage. The excitation current which is only 1 percent or less even at rated
voltage becomes negligibly small during this test and hence is neglected. The shunt branch
′
is thus assumed to be absent. Also I1 = I2 as I0 ≃ 0. Therefore Wsc is the sum of the
copper losses in primary and secondary put together. The reactive power consumed is that
absorbed by the leakage reactance of the two windings.
2 ′
Wsc = Isc (r1 + r2 ) (42)
Vsc
Zsc = (43)
I
qsc
′
(xl1 + xl2 ) = 2 − (r + r ′ )2
Zsc (44)
1 2
48
W
A
Vsc
V
(a)Physical Arrangement
r1 jxl1 r’2 jx’l2
Isc
Vsc
(b)Equivalent Circuit
49
Load Test helps to determine the total loss that takes place, when the transformer
is loaded. Unlike the tests described previously, in the present case nominal voltage is applied
across the primary and rated current is drown from the secondary. Load test is used mainly
1. to determine the rated load of the machine and the temperature rise
Rated load is determined by loading the transformer on a continuous basis and observ-
ing the steady state temperature rise. The losses that are generated inside the transformer
on load appear as heat. This heats the transformer and the temperature of the transformer
increases. The insulation of the transformer is the one to get affected by this rise in the
temperature. Both paper and oil which are used for insulation in the transformer start get-
ting degenerated and get decomposed. If the flash point of the oil is reached the transformer
goes up in flames. Hence to have a reasonable life expectancy the loading of the transformer
must be limited to that value which gives the maximum temperature rise tolerated by the
insulation. This aspect of temperature rise cannot be guessed from the electrical equivalent
circuit. Further, the losses like dielectric losses and stray load losses are not modeled in the
50
equivalent circuit and the actual loss under load condition will be in error to that extent.
Many external means of removal of heat from the transformer in the form of different cooling
methods give rise to different values for temperature rise of insulation. Hence these permit
different levels of loading for the same transformer. Hence the only sure way of ascertaining
the rating is by conducting a load test.
Further, the temperature rise of the transformer is due to the losses that take place
‘inside’ the transformer. The efficiency of the transformer is above 99% even in modest sizes
which means 1 percent of power handled by the transformer actually goes to heat up the
machine. The remaining 99% of the power has to be dissipated in a load impedance external
to the machine. This is very wasteful in terms of energy also. ( If the load is of unity power
factor) Thus the actual loading of the transformer is seldom resorted to. Equivalent loss
methods of loading and ‘Phantom’ loading are commonly used in the case of transformers.
The load is applied and held constant till the temperature rise of transformer reaches a
steady value. If the final steady temperature rise is lower than the maximum permissible
value, then load can be increased else it is decreased. That load current which gives the
maximum permissible temperature rise is declared as the nominal or rated load current and
the volt amperes are computed using the same.
51
In the equivalent loss method a short circuit test is done on the transformer. The
short circuit current is so chosen that the resulting loss taking place inside the transformer
is equivalent to the sum of the iron losses, full load copper losses and assumed stray load
losses. By this method even though one can pump in equivalent loss inside the transformer,
the actual distribution of this loss vastly differs from that taking place in reality. Therefore
this test comes close to a load test but does not replace one.
W 1
A
2Io Io Io
V1
V
I’2 I2 I’2
I2
W 2
A
Vs V
In Phantom loading method two identical transformers are needed. The windings
are connected back to back as shown in Fig. 22. Suitable voltage is injected into the loop
formed by the two secondaries such that full load current passes through them. An equiv-
52
alent current then passes through the primary also. The voltage source V1 supplies the
magnetizing current and core losses for the two transformers. The second source supplies
the load component of the current and losses due to the same. There is no power wasted
in a load ( as a matter of fact there is no real load at all) and hence the name Phantom
or virtual loading. The power absorbed by the second transformer which acts as a load is
pushed back in to the mains. The two sources put together meet the core and copper losses
of the two transformers. The transformers work with full flux drawing full load currents and
hence are closest to the actual loading condition with a physical load.
53
As stated earlier, transformers of various sizes, ratings, voltage ratios can be seen
being used in a power system. The parameters of the equivalent circuits of these machines
also vary over a large range. Also the comparison of these machines are made simple if all
the parameters are normalized. If simple scaling of the parameters is done then one has
to carry forward the scaling factors in the calculations. Expressing in percent basis is one
example of scaling. However if the scaling is done on a logical basis one can have a simple
representation of the parameters without the bother of the scaling factors. Also different
units of measurement are in use in the different countries (FPS, CGS, MKS, etc;). These
units also underwent several revisions over the years. If the transformer parameter can be
freed from the units then the system becomes very simple. The ‘per unit’ system is developed
keeping these aspects in mind. The parameters of the transformer are referred to some base
values and thus get scaled. In the case of power system a common base value is adopted
in view of different ratings of the equipments used. In the case of individual equipments,
its own nominal parameters are used as base values. Some base parameters can be chosen
as independent base values while some others become derived base parameters. Once the
base values are identified the per unit values are calculated for any parameter by dividing
the same by its base value. The units must be the same for both the parameters and their
bases. Thus the per unit value is a unit-less dimensionless number. Let us choose nominal
voltage and nominal current on the primary side of a transformer as the base values Vbase
and Ibase . Other base values like volt ampere Sbase , short circuit impedance Zbase can be
54
Normally Sbase and Vbase are known from name plate details. Other base values can be
derived from them.
V (volt)
Vp.u = ,
Vbase (volt)
I(Amps) I(amps)
Ip.u = = Sbase (48)
Ibase (amps) V base
Many times, when more transformers are involved in a circuit one is required to choose
a common base value for all of them. Parameters of all the machines are expressed on this
common base. This is a common problem encountered in the case of parallel operation of
two or more transformers. The conversion of the base values naturally lead to change in the
per unit values of their parameters. An impedance Zp.u.old on the old base of Sbaseold and
Vbaseold shall get modified on new base Sbasenew ,Vbasenew as
2
Vbase old Sbase new
Zp.u.new = (Zp.u.old. ) 2
(50)
Sbase old Vbase new
The term inside the bracket is nothing but the ohmic value of the impedance and this gets
converted into the new per unit value by the new Sbase and Vbase .
If all the equivalent circuit parameters are referred to the secondary side and per unit
values of the new equivalent circuit parameters are computed with secondary voltage and
55
current as the base values, there is no change in the per unit values. This can be easily seen by,
′
′ ′ S ′ 1
Zp.u. = Zohm . base
′2 but Zohm = .Zohm (51)
Vbase a2
Where a - is the turns ratio of primary to secondary
Z - impedance as seen by primary,
′
Z - impedance as seen by secondary.
′
Sbase = Sbase - as the transformer rating is unaltered.
′
Vbase = Vbase . a1
′
From the above relationships it can be seen that Zp.u. = Zp.u..
This becomes obvious if we realize that the mmf of the core for establishing a given
flux is the same whether it is supplied through primary or the secondary. Also the active
power and reactive power absorbed inside the transformer are not dependant on the winding
connected to supply. This is further illustrated by taking the equivalent circuit of a trans-
former derived earlier and expressing the same in per unit form.
Thus the per unit values help in dispensing away the scaling constants. The veracity
of the parameters can be readily checked. Comparison of the parameters of the machines
with those of similar ones throw in useful information about the machines. Comparing the
efficiencies of two transformers at any load one can say that the transformer with a higher
p.u.resistance has higher copper losses without actually computing the same.
Application of per unit values for the calculation of voltage regulation, efficiency and
load sharing of parallel connected transformers will be discussed later at appropriate places.
56
9 Voltage Regulation
Modern power systems operate at some standard voltages. The equipments work-
ing on these systems are therefore given input voltages at these standard values, within
certain agreed tolerance limits. In many applications this voltage itself may not be good
enough for obtaining the best operating condition for the loads. A transformer is interposed
in between the load and the supply terminals in such cases. There are additional drops
inside the transformer due to the load currents. While input voltage is the responsibility of
the supply provider, the voltage at the load is the one which the user has to worry about.
If undue voltage drop is permitted to occur inside the transformer the load voltage becomes
too low and affects its performance. It is therefore necessary to quantify the drop that takes
place inside a transformer when certain load current, at any power factor, is drawn from its
output leads. This drop is termed as the voltage regulation and is expressed as a ratio of
the terminal voltage (the absolute value per se is not too important).
The voltage regulation can be defined in two ways - Regulation Down and Regulation
up. These two definitions differ only in the reference voltage as can be seen below.
Regulation down: This is defined as ” the change in terminal voltage when a load current
at any power factor is applied, expressed as a fraction of the no-load terminal voltage”.
Expressed in symbolic form we have,
|Vnl | − |Vl |
Regulation = (52)
|Vnl |
Vnl and Vl are no-load and load terminal voltages. This is the definition normally used
in the case of the transformers, the no-load voltage being the one given by the power
57
supply provider on which the user has no say. Hence no-load voltage is taken as the
reference.
Regulation up: Here again the regulation is expressed as the ratio of the change in the
terminal voltage when a load at a given power factor is thrown off, and the on load
voltage. This definition if expressed in symbolic form results in
|Vnl | − |Vl |
Regulation = (53)
|Vl |
This definition is more commonly used in the case of alternators and power systems
as the user-end voltage is guaranteed by the power supply provider. He has to generate
proper no-load voltage at the generating station to provide the user the voltage he has asked
for. In the expressions for the regulation, only the numerical differences of the voltages are
taken and not vector differences.
In the case of transformers both definitions result in more or less the same value for
the regulation as the transformer impedance is very low and the power factor of operation is
quite high. The power factor of the load is defined with respect to the terminal voltage on
load. Hence a convenient starting point is the load voltage. Also the full load output voltage
is taken from the name plate. Hence regulation up has some advantage when it comes to its
application. Fig. 23 shows the phasor diagram of operation of the transformer under loaded
′
condition. The no-load current I0 is neglected in view of the large magnitude of I2 . Then
58
Re jXe
I’2
V1 V’2
A
O
θ B
C E
φ V’2 I2’Xe
I2’Re
I2’
(b)Phasor Diagram
59
′
I1 = I2 .
′ ′
V1 = I2 (Re + jXe ) + V2 (54)
p
OD = V1 = [OA + AB + BC]2 + [CD]2
q
′ ′ ′ ′ ′
= [V2 + I2 Re cos φ + I2 Xe sin φ]2 + [I2 Xe cos φ − I2 Re sin φ]2 (55)
Also,
′ ′
V1 = V2 + I2 .(Re + jXe ) (56)
′ ′
= V2 + I2 (cos φ − j sin φ)(Re + jXe )
′
|V1 | − |V2 |
q
∴ RegulationR = ′ = (1 + v1 )2 + v22 − 1 (57)
|V2 |
′
I2 Xe
ex = V2
′ =per unit reactance drop
as v1 and v2 are small.
v22 v2
∴ R ≃ 1 + v1 + − 1 ≃ v1 + 2 (61)
2(1 + e1 ) 2
(ex sin φ − er cos φ)2
∴ regulation R = er cos φ ± ex sin φ + (62)
2
60
Powers higher than 2 for v1 and v2 are negligible as v1 and v2 are already small. As
v2 is small its second power may be neglected as a further approximation and the expression
for the regulation of the transform boils down to
regulation R = er cos φ ± ex sin φ
The negative sign is applicable when the power factor is leading. It can be seen from
the above expression, the full load regulation becomes zero when the power factor is leading
and er cos φ = ex sin φ or tan φ = er /ex
or the power factor angle φ = tan−1 (er /ex ) = tan−1 (Re /Xe ) leading.
61
V1
D
V2 I2’Xe
θ
O φ
I2’Re C
I2’
A
B
Thus this expression may not be as convenient as the earlier one due to the square root
involved.
Fig. 25 shows the variation of full load regulation of a typical transformer as the
power factor is varied from zero power factor leading, through unity power factor, to zero
power factor lagging.
It is seen from Fig. 25 that the full load regulation at unity power factor is nothing but
the percentage resistance of the transformer. It is therefore very small and negligible. Only
with low power factor loads the drop in the series impedance of the transformer contributes
substantially to the regulation. In small transformers the designer tends to keep the Xe very
low (less than 5%) so that the regulation performance of the transformer is satisfactory.
62
3
%Regulation
2
power factor
1
-2
-3
-4
-5
63
A low value of the short circuit impedance /reactance results in a large short circuit
current in case of a short circuit. This in turn results in large mechanical forces on the
winding. So, in large transformers the short circuit impedance is made high to give better
short circuit protection to the transformer which results in poorer regulation performance.
In the case of transformers provided with taps on windings, so that the turns ratio can be
changed, the voltage regulation is not a serious issue. In other cases care has to be exercised
in the selection of the short circuit impedance as it affects the voltage regulation.
64
D.C Machines
1 Introduction
The steam age signalled the beginning of an industrial revolution. The advantages
of machines and gadgets in helping mass production and in improving the services spurred
the industrial research. Thus a search for new sources of energy and novel gadgets received
great attention. By the end of the 18th century the research on electric charges received
a great boost with the invention of storage batteries. This enabled the research work on
moving charges or currents. It was soon discovered ( in 1820 ) that, these electric currents
are also associated with magnetic field like a load stone. This led to the invention of an
electromagnet. Hardly a year later the force exerted on a current carrying conductor placed
in the magnetic field was invented. This can be termed as the birth of a motor. A better
understanding of the inter relationship between electric and magnetic circuits was obtained
with the enumeration of laws of induction by Faraday in 1831. Parallel research was contem-
porarily being done to invent a source of energy to recharge the batteries in the form of a
d.c. source of constant amplitude (or d.c. generator). For about three decades the research
on d.c. motors and d.c. generators proceeded on independent paths. During the second half
of the 19th century these two paths merged. The invention of a commutator paved the way
for the birth of d.c. generators and motors. These inventions generated great interest in the
generation and use of electrical energy. Other useful machines like alternators, transformers
and induction motors came into existence almost contemporarily. The evolution of these
machines was very quick. They rapidly attained the physical configurations that are being
used even today. The d.c. power system was poised for a predominant place as a preferred
system for use, with the availability of batteries for storage, d.c. generators for conversion of
mechanical energy into electrical form and d.c. motors for getting mechanical outputs from
electrical energy.
The limitations of the d.c. system however became more and more apparent
as the power demand increased. In the case of d.c. systems the generating stations and the
load centers have to be near to each other for efficient transmission of energy. The invention
of induction machines in the 1880s tilted the scale in favor of a.c. systems mainly due to
the advantage offered by transformers, which could step up or step down the a.c.voltage
levels at constant power at extremely high efficiency. Thus a.c. system took over as the
preferred system for the generation transmission and utilization of electrical energy. The
d.c. system, however could not be obliterated due to the able support of batteries. Further,
d.c. motors have excellent control characteristics. Even today the d.c. motor remains an
industry standard as far as the control aspects are concerned. In the lower power levels and
also in regenerative systems the d.c. machines still have a major say.
Electric machines can be broadly classified into electrostatic machines and electro-
magnetic machines. The electrostatic principles do not yield practical machines for commer-
cial electric power generation. The present day machines are based on the electro-magnetic
principles. Though one sees a variety of electrical machines in the market, the basic under-
lying principles of all these are the same. To understand, design and use these machines the
following laws must be studied.
Most of the present day machines have one or two electric circuits linking a common
magnetic circuit. In subsequent discussions the knowledge of electric and magnetic circuit
laws is assumed. The attention is focused on the Faraday’s law and Biot Savart’s law in the
present study of the electrical machines.
Faraday proposed this law of Induction in 1831. It states that if the magnetic
flux lines linking a closed electric coil changes, then an emf is induced in the coil. This
emf is proportional to the rate of change of these flux linkages. This can be expressed
mathematically,
dψ
e∝ (1)
dt
3
where ψ is the flux linkages given by the product of flux lines in weber that are linked
and N the number of turns of the coil. This can be expressed as,
dΦ
e∝N (2)
dt
Here N is the number of turns of the coil, and Φ is the flux lines in weber link-
ing all these turns. The direction of the induced emf can be determined by the application of
Lenz’s law. Lenz’s law states that the direction of the induced emf is such as to produce an
effect to oppose this change in flux linkages. It is analogous to the inertia in the mechanical
systems.
The changes in the flux linkages associated with a turn can be brought about by
(i) changing the magnitude of the flux linking a static coil
(ii) moving the turn outside the region of a steady field
(iii) moving the turn and changing the flux simultaneously
These may be termed as Case(i), Case(ii), and Case(iii) respectively.
This is now explained with the help of a simple geometry. Fig. 1 shows a rectan-
gular loop of one turn (or N=1). Conductor 1 is placed over a region with a uniform flux
density of B Tesla. The flux lines, the conductor and the motion are in mutually perpendic-
ular directions. The flux linkages of the loop is BLN weber turns. If the flux is unchanging
and conductor stationary, no emf will be seen at the terminals of the loop. If now the flux
alone changes with time such that B = Bm . cos ωt, as in Case(i), an emf given by
d
e= (Bm .L.N cos ωt) = −(Bm .L.Nω). sin ωt.
dt
= −jBm .L.Nω. cos ωt volt (3)
L
B
- +
dψ d(Bm .L.N) dX
e= = = Bm .L.N volts (4)
dt dt dt
but
dX
=v ∴ e = Bm .L.N.v volts (5)
dt
The emf induced in the loop is directly proportional to the uniform flux density under which
it is moving with a velocity v. This type of voltage is called speed emf (or rotational emf).
The Case(iii) refers to the situation where B is changing with time and so also is X. Then
the change in flux linkage and hence the value of e is given by
In this case both transformer emf and speed emf are present.
The Case(i) has no mechanical energy associated with it. This is the principle
used in transformers. One coil carrying time varying current produces the time varying field
and a second coil kept in the vicinity of the same has an emf induced in it. The induced emf
of this variety is often termed as the transformer emf.
The Case(ii) is the one which is employed in d.c. machines and alternators. A
static magnetic field is produced by a permanent magnet or by a coil carrying a d.c. current.
A coil is moved under this field to produce the change in the flux linkages and induce an emf
in the same. In order to produce the emf on a continuous manner a cylindrical geometry
is chosen for the machines. The direction of the field, the direction of the conductor of the
coil and the direction of movement are mutually perpendicular as mentioned above in the
example taken.
In the example shown above, only one conductor is taken and the flux ’cut’ by
the same in the normal direction is used for the computation of the emf. The second con-
ductor of the turn may be assumed to be far away or unmoving. This greatly simplifies the
computation of the induced voltage as the determination of flux linkages and finding its rate
of change are dispensed with. For a conductor moving at a constant velocity v the induced
emf becomes just proportional to the uniform flux density of the magnetic field where the
conductor is situated. If the conductor, field and motion are not normal to each other then
the mutually normal components are to be taken for the computation of the voltage. The
induced emf of this type is usually referred to as a rotational emf (due to the geometry).
cal energy conversion results in the generation of both transformer and rotational emf to be
present in the coil moving under a changing field. This principle is utilized in the induction
machines and a.c. commutator machines. The direction of the induced emf is
emf and
current
Force Motion F
B
(a) (b)
decided next. This can be obtained by the application of the Lenz’s law and the law of
interaction. This is illustrated in Fig. 3.
In Case(ii), the direction of the induced emf is as shown. Here again one could
derive the same from the application of the Lenz’s law. The changes in the flux linkages is
emf
current
F
Motion,Force
B
(a) (b)
brought about by the sweep or movement of the conductor. The induced emf, if permitted
to drive a current which produces an opposing force, is as shown in the figure. If one looks
closely at the field around the conductor under these conditions it is as shown in Fig. 2(a)and
(b). The flux lines are more on one side of the conductor than the other. These lines seem
to urge the conductor to the left with a force F . As F opposes v and the applied force,
mechanical energy gets absorbed in this case and the machine works as a generator. This
force is due to electro magnetic interaction and is proportional to the current and the flux
swept. Fig. 3(a)and (b) similarly explain the d.c.motor operation. The current carrying con-
ductor reacts with the field to develop a force which urges the conductor to the right. The
induced emf and the current are seen to act in opposite direction resulting in the absorption
of electric energy which gets converted into the mechanical form.
In Case (iii) also the direction of the induced emf can be determined in a
similar manner. However, it is going to be more complex due to the presence of transformer
emf and rotational emf which have phase difference between them.
Putting mathematically, in the present study of d.c.machines,
F = B.L.I Newton
When the generated voltage drives a current, it produces a reaction force on the
mechanical system which absorbs the mechanical energy. This absorbed mechanical energy
is the one which results in the electric current and the appearance of electrical energy in
the electrical circuit. The converse happens in the case of the motor. If we force a current
against an induced emf then the electrical power is absorbed by the same and it appears
as the mechanical torque on the shaft. Thus, it is seen that the motoring and generating
actions are easily changeable with the help of the terminal conditions.
D.C. machines are the electro mechanical energy converters which work from a d.c.
source and generate mechanical power or convert mechanical power into a d.c. power. These
machines can be broadly classified into two types, on the basis of their magnetic structure.
They are,
1. Homopolar machines
2. Heteropolar machines.
Homopolar generators
Even though the magnetic poles occur in pairs, in a homopolar generator the conductors
are arranged in such a manner that they always move under one polarity. Either north pole
or south pole could be used for this purpose. Since the conductor encounters the magnetic
flux of the same polarity every where it is called a homopolar generator. A cylindrically
symmetric geometry is chosen. The conductor can be situated on the surface of the rotor
with one slip-ring at each end of the conductor. A simple structure where there is only
one cylindrical conductor with ring brushes situated at the ends is shown in Fig. 4. The
excitation coil produces a field which enters the inner member from outside all along the
periphery. The conductor thus sees only one pole polarity or the flux directed in one sense.
A steady voltage now appears across the brushes at any given speed of rotation. The polarity
of the induced voltage can be reversed by reversing either the excitation or the direction of
10
+ N - Brush
A B
Flux
S
Field A B
+ N -
coil
11
rotation but not both. The voltage induced would be very low but the currents of very large
amplitudes can be supplied by such machines. Such sources are used in some applications
like pulse-current and MHD generators, liquid metal pumps or plasma rockets. The steady
field can also be produced using a permanent magnet of ring shape which is radially mag-
netized. If higher voltages are required one is forced to connect many conductors in series.
This series connection has to be done externally. Many conductors must be situated on the
rotating structure each connected to a pair of slip rings. However, this modification intro-
duces parasitic air-gaps and makes the mechanical structure very complex. The magnitude
of the induced emf in a conductor 10 cm long kept on a rotor of 10 cm radius rotating at
3000 rpm, with the field flux density being 1 Tesla every where in the air gap, is given by
e = BLv
3000
= 1 ∗ 0.1 ∗ 2π ∗ 0.1 ∗ = 3.14 volt
60
The voltage drops at the brushes become very significant at this level bringing down the
efficiency of power conversion. Even though homopolar machines are d.c. generators in a
strict sense that they ’generate’ steady voltages, they are not quite useful for day to day use.
A more practical converters can be found in the d.c. machine family called ”hetero-polar”
machines.
In the case of a hetero-polar generator the induced emf in a conductor goes through a
cyclic change in voltage as it passes under north and south pole polarity alternately. The
induced emf in the conductor therefore is not a constant but alternates in magnitude. For
12
c
B -
d
+
A
Load
Field coil
Pole
v
N
v
10
11 9
12 8
Commutator S1 7
A+ -
v
B
1 F1
S2
Armature core
v
F2
S3
6
F4
Yoke
F3
2
S4
v
3 5
4
13
a constant velocity of sweep the induced emf is directly proportional to the flux density
under which it is moving. If the flux density variation is sinusoidal in space, then a sine
wave voltage is generated. This principle is used in the a.c generators. In the case of d.c.
generators our aim is to get a steady d.c. voltage at the terminals of the winding and not
the shape of the emf in the conductors. This is achieved by employing an external element,
which is called a commutator, with the winding.
14
Fig. 6 shows a 2-pole, 12 coil, ring wound armature of a machine. The 12 coils
are placed at uniform spacing around the rotor. The junction of each coil with its neighbor
is connected to a commutator segment. Each commutator segment is insulated from its
neighbor by a mica separator. Two brushes A and B are placed on the commutator which
looks like a cylinder. If one traces the connection from brush A to brush B one finds that
there are two paths. In each path a set of voltages get added up. The sum of the emfs is
constant(nearly). The constancy of this magnitude is altered by a small value corresponding
to the coil short circuited by the brush. As we wish to have a maximum value for the output
voltage, the choice of position for the brushes would be at the neutral axis of the field. If
the armature is turned by a distance of one slot pitch the sum of emfs is seen to be constant
even though a different set of coils participate in the addition. The coil which gets short
circuited has nearly zero voltage induced in the same and hence the sum does not change
substantially. This variation in the output voltage is called the ’ripple’. More the number of
coils participating in the sum lesser would be the ’percentage’ ripple.
A rectangular flux density wave form has some advantages in the derivation
of the voltage between the brushes. Due to this form of the flux density curve, the induced
15
emf in each turn of the armature becomes constant and equal to each other. With this back
ground the emf induced between the brushes can be derived. The value of the induced in
one conductor is given by
Ec = Bav .L.v Volt (7)
where
Bav - Average flux density over a pole pitch, Tesla.
L- Length of the ’active’ conductor, m.
v- Velocity of sweep of conductor, m/sec.
If there are Z conductors on the armature and they form b pairs of parallel circuits between
the brushes by virtue of their connections, then number of conductors in a series path is
Z/2b.
The induced emf between the brushes is
Z
E = Ec . (8)
2b
Z
E = Bav .L.v. Volts (9)
2b
But v = (2p).Y.n where p is the pairs of poles Y is the pole pitch, in meters, and n is the
number of revolutions made by the armature per second.
Also Bav can be written in terms of pole pitch Y , core length L, and flux per pole φ as
φ
Bav = Tesla (10)
(L.Y )
φ Z φpZn
E= .L.(2p.Y.n). = volts (11)
(L.Y ) 2b b
The number of pairs of parallel paths is a function of the type of the winding chosen. This
16
When the armature is loaded, the armature conductors carry currents. These current
carrying conductors interact with the field and experience force acting on the same. This
force is in such a direction as to oppose their cause which in the present case is the relative
movement between the conductors and the field. Thus the force directly opposes the motion.
Hence it absorbs mechanical energy. This absorbed mechanical power manifests itself as the
converted electrical power. The electrical power generated by an armature delivering a
current of Ia to the load at an induced emf of E is EIa Watts. Equating the mechanical and
electrical power we have
2πnT = EIa (12)
where T is the torque in Nm. Substituting for E from Eqn. 11, we get
p.φ.Z.n
2πnT = .Ia (13)
b
This shows that the torque generated is not a function of the speed. Also,
it is proportional to ’total flux’ and ’Total ampere conductors’ on the armature, knowing that
Ia /2b is Ic the conductor current on the armature. The expression for the torque generated
can also be derived from the first principles by the application of the law of interaction. The
law of interaction states that the force experienced by a conductor of length L kept in a
17
The total work done by an armature with Z conductors in one revolution is given by,
φ
Wa = Bav .L.Ic .Z.(2p.Y ) Joules = .L.Ic .Z.2p.Y Joules (16)
L.Y
The work done per second or the power converted by the armature is,
Ia
AsIc = (18)
2b
Ia
= φ.p.Z.n. (19)
b
which is nothing but EIa .
The above principles can easily be extended to the case of motoring mode
of operation also. This will be discussed next in the section on motoring operation of d.c.
machines.
In the motoring operation the d.c. machine is made to work from a d.c. source and
absorb electrical power. This power is converted into the mechanical form. This is briefly
discussed here. If the armature of the d.c. machine which is at rest is connected to a d.c.
source then, a current flows into the armature conductors. If the field is already excited then
18
these current carrying conductors experience a force as per the law of interaction discussed
above and the armature experiences a torque. If the restraining torque could be neglected the
armature starts rotating in the direction of the force. The conductors now move under the
field and cut the magnetic flux and hence an induced emf appears in them. The polarity of
the induced emf is such as to oppose the cause of the current which in the present case is the
applied voltage. Thus a ’back emf’ appears and tries to reduce the current. As the induced
emf and the current act in opposing sense the machine acts like a sink to the electrical power
which the source supplies. This absorbed electrical power gets converted into mechanical
form. Thus the same electrical machine works as a generator of electrical power or the
absorber of electrical power depending upon the operating condition. The absorbed power
gets converted into electrical or mechanical power. This is briefly explained earlier with the
help of Figure 3(a) and 3(b). These aspects would be discussed in detail at a later stage.
19
As mentioned earlier the d.c. machines were invented during the second half of the 19th
century. The initial pace of development work was phenomenal. The best configurations
stood all the competition and the test of time and were adopted. Less effective options were
discarded. The present day d.c. generator contains most, if not all, of the features of the
machine developed over a century earlier. To appreciate the working and the characteristics
of these machines, it is necessary to know about the different parts of the machine - both
electrical and non-electrical. The description would also aid the understanding of the reason
for selecting one form of construction or the other. An exploded view of a small d.c.
1. Body
2. Poles
20
3. Armature
5. Commutating poles
6. Compensating winding
The constructional aspects relating to these parts are now discussed briefly in sequence.
Body The body constitutes the outer shell within which all the other parts are housed.
This will be closed at both the ends by two end covers which also support the bearings
required to facilitate the rotation of the rotor and the shaft. Even though for the
generation of an emf in a conductor a relative movement between the field and the
conductor would be enough, due to practical considerations of commutation, a rotating
conductor configuration is selected for d.c. machines. Hence the shell or frame supports
the poles and yoke of the magnetic system. In many cases the shell forms part of the
magnetic circuit itself. Cast steel is used as a material for the frame and yoke as the
flux does not vary in these parts. In large machines these are fabricated by suitably
welding the different parts. Those are called as fabricated frames. Fabrication as
against casting avoids expensive patterns. In small special machines these could be
made of stack of laminations suitably fastened together to form a solid structure.
Main poles Solid poles of fabricated steel with seperate/integral pole shoes are fastened
to the frame by means of bolts. Pole shoes are generally laminated. Sometimes pole
body and pole shoe are formed from the same laminations. Stiffeners are used on both
21
sides of the laminations. Riveted through bolts hold the assembly together. The pole
shoes are shaped so as to have a slightly increased air gap at the tips.
Inter-poles These are small additional poles located in between the main poles. These can
be solid, or laminated just as the main poles. These are also fastened to the yoke by
bolts. Sometimes the yoke may be slotted to receive these poles. The inter poles could
be of tapered section or of uniform cross section. These are also called as commutating
poles or compoles. The width of the tip of the compole can be about a rotor slot pitch.
Armature The armature is where the moving conductors are located. The armature is
constructed by stacking laminated sheets of silicon steel. Thickness of these lamination
is kept low to reduce eddy current losses. As the laminations carry alternating flux
the choice of suitable material, insulation coating on the laminations, stacking it etc
are to be done more carefully. The core is divided into packets to facilitate ventilation.
The winding cannot be placed on the surface of the rotor due to the mechanical forces
coming on the same. Open parallel sided equally spaced slots are normally punched in
the rotor laminations. These slots house the armature winding. Large sized machines
employ a spider on which the laminations are stacked in segments. End plates are
suitably shaped so as to serve as ’Winding supporters’. Armature construction process
must ensure provision of sufficient axial and radial ducts to facilitate easy removal of
heat from the armature winding.
Field windings In the case of wound field machines (as against permanent magnet excited
machines) the field winding takes the form of a concentric coil wound around the main
poles. These carry the excitation current and produce the main field in the machine.
Thus the poles are created electromagnetically. Two types of windings are generally
employed. In shunt winding large number of turns of small section copper conductor is
22
used. The resistance of such winding would be an order of magnitude larger than the
armature winding resistance. In the case of series winding a few turns of heavy cross
section conductor is used. The resistance of such windings is low and is comparable
to armature resistance. Some machines may have both the windings on the poles.
The total ampere turns required to establish the necessary flux under the poles is
calculated from the magnetic circuit calculations. The total mmf required is divided
equally between north and south poles as the poles are produced in pairs. The mmf
required to be shared between shunt and series windings are apportioned as per the
design requirements. As these work on the same magnetic system they are in the form
of concentric coils. Mmf ’per pole’ is normally used in these calculations.
Armature winding As mentioned earlier, if the armature coils are wound on the surface of
the armature, such construction becomes mechanically weak. The conductors may fly
away when the armature starts rotating. Hence the armature windings are in general
pre-formed, taped and lowered into the open slots on the armature. In the case of
small machines, they can be hand wound. The coils are prevented from flying out due
to the centrifugal forces by means of bands of steel wire on the surface of the rotor in
small groves cut into it. In the case of large machines slot wedges are additionally used
to restrain the coils from flying away. The end portion of the windings are taped at
the free end and bound to the winding carrier ring of the armature at the commutator
end. The armature must be dynamically balanced to reduce the centrifugal forces at
the operating speeds.
Compensating winding One may find a bar winding housed in the slots on the pole
shoes. This is mostly found in d.c. machines of very large rating. Such winding is
called compensating winding. In smaller machines, they may be absent. The function
23
1.Clamping cone
4 2.Insulating cups
3.Commutator bar
3 4.Riser
2
2 5.Insulating gasket
1
5
Commutator Commutator is the key element which made the d.c. machine of the present
day possible. It consists of copper segments tightly fastened together with mica/micanite
insulating separators on an insulated base. The whole commutator forms a rigid and
solid assembly of insulated copper strips and can rotate at high speeds. Each com-
mutator segment is provided with a ’riser’ where the ends of the armature coils get
connected. The surface of the commutator is machined and surface is made concentric
with the shaft and the current collecting brushes rest on the same. Under-cutting the
mica insulators that are between these commutator segments has to be done periodi-
cally to avoid fouling of the surface of the commutator by mica when the commutator
gets worn out. Some details of the construction of the commutator are seen in Fig. 8.
Brush and brush holders Brushes rest on the surface of the commutator. Normally
electro-graphite is used as brush material. The actual composition of the brush depends
on the peripheral speed of the commutator and the working voltage. The hardness of
the graphite brush is selected to be lower than that of the commutator. When the
24
brush wears out the graphite works as a solid lubricant reducing frictional coefficient.
More number of relatively smaller width brushes are preferred in place of large broad
brushes. The brush holders provide slots for the brushes to be placed. The connection
Pigtail
Pressure
spring
Brush
(a)
Radial Trailing
Reaction
Motion of commutator
(b)
Figure 9: Brush holder with a Brush and Positioning of the brush on the commutator
from the brush is taken out by means of flexible pigtail. The brushes are kept pressed
on the commutator with the help of springs. This is to ensure proper contact between
25
the brushes and the commutator even under high speeds of operation. Jumping of
brushes must be avoided to ensure arc free current collection and to keep the brush
contact drop low. Fig. 9 shows a brush holder arrangement. Radial positioning of the
brushes helps in providing similar current collection conditions for both direction of
rotation. For unidirectional drives trailing brush arrangement or reaction arrangement
may be used in Fig. 9-(b) Reaction arrangement is preferred as it results in zero side
thrust on brush box and the brush can slide down or up freely. Also staggering of the
brushes along the length of the commutator is adopted to avoid formation of ’tracks’
on the commutator. This is especially true if the machine is operating in a dusty
environment like the one found in cement plants.
Other mechanical parts End covers, fan and shaft bearings form other important me-
chanical parts. End covers are completely solid or have opening for ventilation. They
support the bearings which are on the shaft. Proper machining is to be ensured for
easy assembly. Fans can be external or internal. In most machines the fan is on the
non-commutator end sucking the air from the commutator end and throwing the same
out. Adequate quantity of hot air removal has to be ensured.
Bearings Small machines employ ball bearings at both ends. For larger machines roller
bearings are used especially at the driving end. The bearings are mounted press-fit
on the shaft. They are housed inside the end shield in such a manner that it is not
necessary to remove the bearings from the shaft for dismantling. The bearings must be
kept in closed housing with suitable lubricant keeping dust and other foreign materials
away. Thrust bearings, roller bearings, pedestal bearings etc are used under special
cases. Care must be taken to see that there are no bearing currents or axial forces on
the shaft both of which destroy the bearings.
26
4 Armature Windings
Main field
N
Commutator Compole field
& Brush X
X
x
x x
x
x
X
x x
Shaft Compensating
x x S
S winding
x x
v
x x
X
x
Armature
x
winding x
x
X X
Yoke
N X
Fig. 10 gives the cross sectional view of a modern d.c. machine show-
ing all the salient parts. Armature windings, along with the commutators, form the heart
of the d.c. machine. This is where the emf is induced and hence its effective deployment
enhances the output of the machine. Fig. 11(a) shows one coil of an armature of Gramme
ring arrangement and Fig. 11(b) shows one coil as per drum winding arrangement. Earlier,
a simple form of this winding in the form of Gramme ring winding was presented for easy
understanding. The Gramme ring winding is now obsolete as a better armature winding has
27
X Ν X Ν
X X
φ φ
Α
Α Α’
28
been invented in the form of a drum winding. The ring winding has only one conductor in
a turn working as an active conductor. The second conductor is used simply to complete
the electrical connections. Thus the effectiveness of the electric circuit is only 50 percent.
Looking at it differently, half of the magnetic flux per pole links with each coil. Also, the
return conductor has to be wound inside the bore of the rotor, and hence the rotor diameter
is larger and mounting of the rotor on the shaft is made difficult.
In a drum winding both forward and return conductors are housed in slots cut
on the armature (or drum). Both the conductors have emf induced in them. Looking at it
differently the total flux of a pole is linked with a turn inducing much larger voltage induced
in the same. The rotor is mechanically robust with more area being available for carrying
the flux. There is no necessity for a rotor bore. The rotor diameters are smaller. Mechanical
problems that existed in ring winding are no longer there with drum windings. The coils
could be made of single conductors (single turn coils) or more number of conductors in series
(multi turn coils). These coils are in turn connected to form a closed winding. The two sides
of the coil lie under two poles one north and the other south, so that the induced emf in
them are always additive by virtue of the end connection. Even though the total winding
is a closed one the sum of the emfs would be zero at all times. Thus there is no circulating
current when the armature is not loaded. The two sides of the coil, if left on the surface, will
fly away due to centrifugal forces. Hence slots are made on the surface and the conductors
are placed in these slots and fastened by steel wires to keep them in position. Each armature
slot is partitioned into two layers, a top layer and a bottom layer. The winding is called as
a double layer winding. This is a direct consequence of the symmetry consideration. The
distance, measured along the periphery of the armature from any point under a pole to a
similar point under the neighboring pole is termed as a pole pitch. The forward conductor
is housed in the top layer of a slot and the return conductor is housed in the bottom layer
29
B
S A
N
C
D
Lower
coil side
A’
N
Active
S A N A’ S
Inactive
Armature
C D
30
of a slot which is displaced by about one pole pitch. The junction of two coils is terminated
on a commutator segment. Thus there are as many commutator segments as the number of
coils. In a double layer winding in S slots there are 2S layers. Two layers are occupied by a
coil and hence totally there are S coils. The S junctions of these S coils are terminated on S
commutator segments. The brushes are placed in such a manner that a maximum voltage
appears across them. While the number of parallel circuits in the case of ring winding is
equal to the number of poles, in the case of drum winding a wide variety of windings are
possible. The number of brushes and parallel paths thus vary considerably. The physical
arrangement of a single coil is shown in Fig. 12 to illustrate its location and connection to
the commutators.
Fig. 13 shows the axial side view while Fig. 13-(b) shows the cut and spread view
of the machine. The number of turns in a coil can be one (single turn coils) or more (multi
turn coils ). As seen earlier the sum of the instantaneous emfs appears across the brushes.
This sum gets altered by the voltage of a coil that is being switched from one circuit to the
other or which is being commutated. As this coil in general lies in the magnetic neutral
axis it has a small value of voltage induced in it. This change in the sum expressed as the
fraction of the total induced voltage is called as the ripple. In order to reduce the ripple,
one can increase the number of coils coming in series between the brushes. As the number
of coils is the same as the number of slots in an armature with two coil sides per slot one is
forced to increase the number of slots. However increasing the slot number makes the tooth
width too narrow and makes them mechanically weak.
To solve this problem the slots are partitioned vertically to increase the number
of coil sides. This is shown in Fig. 14. In the figure, the conductors a, b and c belong to a
coil. Such 2/3 coils occupy the 2/3 top coil sides of the slot. In the present case the number
of coils in the armature is 2S/3S.
31
(a)End view
11 2’
10
3’
1’
12
1 2 3 4 5 6 7 8 9 10 11 12
N S N S
12
1’
2’
11
12 1 2 3 4 5 6 7 8 9 10 11
+ - + Motion
-
(b)Developed view
Press board
Copper
Mica Tape
Press board
33
As mentioned earlier, in a drum winding, the coils span a pole pitch where
ever possible. Such coils are called ’full pitched’ coils. The emf induced in the two active
conductors of such coils have identical emfs with opposite signs at all instants of time. If the
span is more than or less than the full pitch then the coil is said to be ’chorded’. In chorded
coils the induced emfs of the two conductor may be of the same sign and hence oppose each
other( for brief intervals of time). Slight short chording of the coil reduces overhang length
and saves copper and also improves commutation. Hence when the pole pitch becomes frac-
tional number, the smaller whole number may be selected discarding the fractional part.
Similar to the pitch of a coil one can define the winding pitch and commutator
pitch. In a d.c. winding the end of one coil is connected to the beginning of another coil
(not necessarily the next), this being symmetrically followed to include all the coils on the
armature. Winding pitch provides a means of indicating this. Similarly the commutator
pitch provides the information regarding the commutators to which the beginning and the
end of a coil are connected. Commutator pitch is the number of ’micas’ between the ends of
a coil. For all these information to be simple and useful the numbering scheme of the coils
and commutator segments becomes important. One simple method is to number only the
top coil side of the coils in sequence. The return conductor need not be numbered. As a
double layer is being used the bottom coil side is placed in a slot displaced by one coil span
′ ′
from the top coil side. Some times the coils are numbered as 1 − 1 , 2 − 2 etc. indicating
′ ′
the second sides by 1 , 2 etc. The numbering of commutators segments are done similarly.
The commutator segment connected to top coil side of coil 1 is numbered 1. This method
of numbering is simple and easy to follow. It should be noted that changing of the pitch
34
of a coil slightly changes the induced emf in the same. The pitch of the winding however
substantially alters the nature of the winding.
The armature windings are classified into two families based on this. They are
called lap winding and wave winding. They can be simply stated in terms of the commutator
pitch used for the winding.
35
Progressive Retrogressive
yc =+1 yc = -1
s1 s2 F1 F2 F2
s2 F3 s3
1 2 3 1 2 3 4
Coil span
s1 F1
1 2 _1
c+
p
36
1 2
S
N S N S
13 14 1 2 3 4 5 6 7 8 9 10 11 12 Motion
A1 + B1 - A2 + B2 -
37
equally.
The increase in the number of parallel paths in the armature winding brings
about a problem of circulating current. The induced emfs in the different paths tend to
differ slightly due to the non-uniformities in the magnetic circuit. This will be more with the
increase in the number of poles in the machine. If this is left uncorrected, circulating currents
appear in these closed parallel paths. This circulating current wastes power, produces heat
and over loads the brushes under loaded conditions. One method commonly adopted in d.c.
machines to reduce this problem is to provide equalizer connections. As the name suggests
these connections identify similar potential points of the different parallel paths and connect
them together to equalize the potentials. Any difference in the potential generates a local
circulating current and the voltages get equalized. Also, the circulating current does not
flow through the brushes loading them. The number of such equalizer connections, the
cross section for the conductor used for the equalizer etc are decided by the designer. An
example of equalizer connection is discussed now with the help of a 6-pole armature having
150 commutator segments. The coil numbers 1, 51 and 101 are identically placed under the
poles of same polarity as they are one pole-pair apart. There are 50 groups like that. In
order to limit the number of links to 5(say), the following connections are chosen. Then
1,11,21,31, and 41 are the coils under the first pair of poles. These are connected to their
counter parts displaced by 50 and 100 to yield 5 equalizer connections. There are 10 coils
connected in series between any two successive links. The wave windings shall be examined
next.
38
S N S N
20 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Motion
+ A1 - B1 + A2 - B2
(a)Winding layout
_ 5
Full pitch: 21/4=5.25 ~
Span : 1 to 6
Yc=
_1
C+
2
= 21
_1
+
2
=
11
10 }
Commutator pitch 1-11 for retrogressive winding
1-11-21-10-20-9-19-8-18-7-17-
v
A2
A1
6-16-5-15-4-14-3-13-2-12-1
v
B2
B1
(b)Parallel paths
39
In wave windings the coils carrying emf in the same direction at a time are all
grouped together and connected in series. Hence in a simple wave winding there are only
two paths between the brushes, the number of conductors in each path being 50 percent of
the total conductors. To implement a wave winding one should select the commutator pitch
as
C±1
yc = (21)
p
where C is the total segments on the commutator. yc should be an integer number; C and
p should satisfy this relation correctly. Here also the positive sign refers to the progressive
winding and the negative sign yields a retrogressive winding. yc = (C ± m)/p yields a multi-
plex wave winding of order m. A simple wave winding for 4 poles in 21 slots is illustrated in
Fig. 17. As could be seen from the figure, the connection to the next (or previous) adjacent
coil is reached after p coils are connected in series. The winding closes on itself after all the
coils are connected in series. The position for the brushes is indicated in the diagram.
It is seen from the formula for the commutator pitch, the choice of commutator
segments for wave winding is restricted. The number of commutator segments can only be
one more or one less than some multiple of pole pairs. As the number of parallel circuits is
2 for a simple wave winding irrespective of the pole numbers it is preferred in multi polar
machine of lower power levels.
As mentioned earlier the simple wave winding forms two parallel paths, duplex
wave winding has 2*2=4 etc. The coils under all the north poles are grouped together in
40
one circuit and the other circuit collects all the coils that are under all the south poles. Two
brush sets are therefore adequate. Occasionally people employ brush sets equal to the num-
ber of poles. This arrangement does not increase the number of parallel circuits but reduces
the current to be collected by each brush set. This can be illustrated by an example. A
4-pole wave connected winding with 21 commutator segments is taken. yc = (21 − 1)/2 = 10
. A retrogressive wave winding results. The total string of connection can be laid out as
shown below. If coil number 1 is assumed to be in the neutral axis then other neutral axis
coils are a pole pitch apart i.e. coils 6, 11, 16.
If the brushes are kept at commutator segment 1 and 6, nearly half the num-
ber of coils come under each circuit. The polarity of the brushes are positive and negative
alternately. Or, one could have two brushes at 11 and 16 or any two adjacent poles. By
having four brushes at 1, 6, 11 and 16 and connecting 1,11 and 6,16 still only two parallel
circuits are obtained. The brush currents however are halved. This method permits the use
of commutator of shorter length as lesser current is to be collected by each brush and thus
saving on the cost of the commutator. Fig. 17(b) illustrates this brush arrangement with
respect to a 21 slot 4 pole machine. Similarly proceeding, in a 6-pole winding 2,4 or 6 brush
sets may be used.
41
Due to the restrictions posed by lap and wave windings on the choice of number
of slots and commutator segments a practical difficulty arises. Each machine with a certain
pole number, voltage and power ratings may require a particular number of slots and com-
mutator segments for a proper design. Thus each machine may be tailor made for a given
specification. This will require stocking and handling many sizes of armature and commu-
tator.
42
5 Armature reaction
Earlier, an expression was derived for the induced emf at the terminals of the
armature winding under the influence of motion of the conductors under the field established
by field poles. But if the generator is to be of some use it should deliver electrical output to a
load. In such a case the armature conductors also carry currents and produce a field of their
own. The interaction between the fields must therefore must be properly understood in order
to understand the behavior of the loaded machine. As the magnetic structure is complex
and as we are interested in the flux cut by the conductors, we primarily focus our attention
on the surface of the armature. A sign convention is required for mmf as the armature and
field mmf are on two different members of the machine. The convention used here is that
the mmf acting across the air gap and the flux density in the air gap are shown as positive
when they act in a direction from the field system to the armature. A flux line is taken
and the value of the current enclosed is determined. As the magnetic circuit is non-linear,
the field mmf and armature mmf are separately computed and added at each point on the
surface of the armature. The actual flux produced is proportional to the total mmf and the
permeance. The flux produced by field and that produced by armature could be added to
get the total flux only in the case of a linear magnetic circuit. The mmf distribution due to
the poles and armature are discussed now in sequence.
Fig. 18 shows the distribution of mmf due to field coils over two pole pitches. It
is a step curve with the width being equal to the pole arc. The permeance variation at the
surface is given by Fig. 18 assuming the air gap under the pole to be uniform and neglecting
43
N S
mmf
Permeance
Practical
Flux density
Ideal flux density
44
the slotting of the armature. The no-load flux density curve can be obtained by multiplying
mmf and permeance. Allowing for the fringing of the flux, the actual flux density curve
would be as shown under Fig. 18.
N S
N-Pole
S-Pole
Generator Flux
mmf
Figure 19: Mmf and flux distribution under the action of armature alone carrying current
The armature has a distributed winding, as against the field coils which
are concentrated and concentric. The mmf of each coil is shifted in space by the number of
slots. For a full pitched coil, each coil produces a rectangular mmf distribution. The sum
of the mmf due to all coils would result in a stepped triangular wave form. If we neglect
slotting and have uniformly spaced coils on the surface, then the mmf distribution due to the
armature working alone would be a triangular distribution in space since all the conductors
carry equal currents. MMF distribution is the integral of the ampere conductor distribution.
45
This is depicted in Fig. 19. This armature mmf per pole is given by
1 Ic .Z
Fa = .
2 2p
where Ic is the conductor current and Z is total number of conductors on the armature. This
peak value of the mmf occurs at the inter polar area, shifted from the main pole axis by half
the pole pitch when the brushes are kept in the magnetic neutral axis of the main poles.
Brush axis
N S
A Generator B
c
D
C
Field a
flux
B
o’
A o
b
Armature flux
Total flux
The mmf of field coils and armature coils are added up and the re-
sultant mmf distribution is obtained as shown in Fig. 20.
46
This shows the decrease in the mmf at one tip of a pole and a substantial rise
at the other tip. If the machine has a pole arc to pole pitch ratio of 0.7 then 70% of the
armature reaction mmf gets added at this tip leading to considerable amount of saturation
under full load conditions. The flux distribution also is shown in Fig. 20. This is obtained
by multiplying mmf and permeance waves point by point in space. Actual flux distribution
differs from this slightly due to fringing. As seen from the figure, the flux in the inter polar
region is substantially lower due to the high reluctance of the medium. The air gaps under
the pole tips are also increased in practice to reduce excessive saturation of this part. The
advantage of the salient pole field construction is thus obvious. It greatly mitigates the
effect of the armature reaction. Also, the coils under going commutation have very little
emf induced in them and hence better commutation is achieved. Even though the armature
reaction produced a cross magnetizing effect, the net flux per pole gets slightly reduced,
on load, due to the saturation under one tip of the pole. This is more so in modern d.c.
machines where the normal excitation of the field makes the machine work under some level
of saturation.
In some small d.c. machines the brushes are shifted from the position of the mag-
netic neutral axis in order to improve the commutation. This is especially true of machines
with unidirectional operation and uni-modal (either as a generator or as a motor) operation.
Such a shift in the direction of rotation is termed ‘lead’ (or forward lead). Shift of brushes
in the opposite to the direction of rotation is called ‘backward lead’. This lead is expressed
in terms of the number of commutator segments or in terms of the electrical angle. A pole
pitch corresponds to an electrical angle of 180 degrees. Fig. 21 shows the effect of a forward
47
Rotation
a
Field c
flux
Armature flux
Total flux
Rotation
b’
a’ a
θ
b
S
(b)Calculation of demagnetizing mmf per pole
48
brush lead on the armature reaction. The magnetization action due to the armature is no
longer entirely cross magnetizing. Some component of the same goes to demagnetize the
main field and the net useful flux gets reduced. This may be seen as the price we pay for
improving the commutation. Knowing the pole arc to pole pitch ratio one can determine
the total mmf at the leading and trailing edges of a pole without shift in the brushes.
Fmax = Ff + α.Fa
where Ff is the field mmf, Fa is armature reaction mmf per pole, and α is the pole arc to
pole pitch ratio.
1 Z.Ic
Fa = . (23)
2 2p
The net flux per pole decreases due to saturation at the trailing edge and
hence additional ampere turns are needed on the pole to compensate this effect. This may
be to the tune of 20 percent in the modern d.c. machines.
The brush shift gives rise to a shift in the axis of the mmf of the armature
reaction. This can be resolved into two components, one in the quadrature axis and sec-
ond along the pole axis as shown in Fig. 21.(b) The demagnetizing and cross magnetizing
component of the armature ampere turn per pole can be written as
2θ
Fd = .Fa (24)
π
2θ
Fq = (1 − ).Fa (25)
π
where θ is the angle of lead . In terms of the number of commutator segments they are
Cl Ic Z Cl
Fd = .
C 4p
or .Ic .Z (26)
4p
C
49
As discussed earlier, for a given polarity of the field and sense of rotation, the
motoring and generating modes differ only in the direction of the armature current. Alter-
natively, for a given sense of armature current, the direction of rotation would be opposite
for the two modes. The leading and trailing edges of the poles change positions if direction
of rotation is made opposite. Similarly when the brush leads are considered, a forward lead
given to a generator gives rise to weakening of the generator field but strengthens the motor
field and vice-versa. Hence it is highly desirable, even in the case of non-reversing drives,
to keep the brush position at the geometrical neutral axis if the machine goes through both
motoring and generating modes.
The second effect of the armature reaction in the case of motors as well as
generators is that the induced emf in the coils under the pole tips get increased when a
pole tip has higher flux density. This increases the stress on the ‘mica’ (micanite) insulation
used for the commutator, thus resulting in increased chance of breakdown of these insulating
sheets. To avoid this effect the flux density distribution under the poles must be prevented
from getting distorted and peaky.
The third effect of the armature reaction mmf distorting the flux density is
that the armature teeth experience a heavy degree of saturation in this region. This increases
the iron losses occurring in the armature in that region. The saturation of the teeth may
be too great as to have some flux lines to link the thick end plates used for strengthening
50
the armature. The increase in iron loss could be as high as 50 percent more at full load
compared to its no-load value.
The above two effects can be reduced by providing a ’compensating’ mmf at
Commutating pole
s
Main pole
N S
N N
S N
Compensating
winding
s
the same spatial rate as the armature mmf. This is provided by having a compensating
winding housed on the pole shoe which carries currents that are directly proportional to the
armature current. The ampere conductors per unit length is maintained identical to that of
51
N S
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Rotation
mmf of
compensating
winding
Resultant
mmf
compole mmf
Armature
mmf
Main field
mmf
52
the armature. The sign of the ampere conductors is made opposite to the armature. This is
illustrated in Fig. 22 and Fig. 23 . Since the compensating winding is connected in series with
the armature, the relationship between armature mmf and the mmf due to compensating
winding remains proper for all modes of working of the machine. The mmf required to be
setup by the compensating winding can be found out to be
Ic .Z polearc
Fc = . (27)
4p polepitch
Under these circumstances the flux density curve remains unaltered under the poles between
no-load and full load.
The axis of the mmf due to armature and the compensating winding being
the same and the signs of mmf being opposite to each other the flux density in the region
of geometric neutral axis gets reduced thus improving the conditions for commutation. One
can design the compensating winding to completely neutralize the armature reaction mmf.
Such a design results in overcompensation under the poles. Improvement in commutation
condition may be achieved simply by providing a commutating pole which sets up a local
field of proper polarity. It is better not to depend on the compensating winding for improv-
ing commutation.
From the analysis of the phenomenon of armature reaction that takes place
in a d.c. machine it can be inferred that the equivalent circuit of the machine need not be
modified to include the armature reaction. The machine can simply be modelled as a voltage
53
source of internal resistance equal to the armature circuit resistance and a series voltage drop
equal to the brush contact drop, under steady state. With this circuit model one can arrive
at the external characteristics of the d.c. machine under different modes of operation.
5.1 Commutation
54
4 3 2 1 Ia Ia
2Ia
4 3 2 1
Motion
(a) 2Ia
tb
4 3 2 Ia 1 Ia
I2 i I1
4 3 2 1
x 2Ia
Thickness (b)
tb
dth 4 3 2 Ia Ia 1
Entering Edge Wi
Length
Leaving Edge
2Ia
4 3 2 1
(c) 2Ia
tb
55
5.1.1 Brushes
If the current density under the brush is assumed to be constant through out the
commutation interval, a simple model for commutation is obtained. For simplicity, the brush
thickness is made equal to thickness of one commutator segment. In Fig. 24(b), the brush
is initially solely resting on segment number 1. The total current of 2Ia is collected by
the brush as shown. As the commutator moves relative to the brush position, the brush
position starts to overlap with that of segment 2. As the current density is assumed to be
constant, the current from each side of the winding is proportional to the area shared on the
56
two segments. Segment 1 current uniformly comes down with segment 2 current increasing
uniformly keeping the total current in the brush constant. The currents I1 and I2 in brush
segments 1 and 2 are given by
x x
I1 = 2Ia (1 − ) and I2 = 2Ia (28)
tb tb
giving I1 + I2 to be 2 Ia .
Here ‘x’ is the width of the brush overlapping on segment 2. The process of commutation
would be over when the current through segment number 1 becomes zero. The current in
the coil undergoing commutation is
(I1 − I2 ) 2x
i = I1 − Ia = Ia − I2 = = Ia (1 − ) (29)
2 tb
tb
Tc = (30)
vc
where vc is the velocity of the commutator. This type of linear commutation is very close
to the ideal method of commutation. The time variation of current in the coil undergoing
commutation is shown in Fig. 25.(a). Fig. 25.(b) also shows the timing diagram for the
currents I1 and I2 and the current densities in entering edge αe , leaving edge αl and also the
mean current density αm in the brush. Machines having very low coil inductances, operating
at low load currents, and low speeds, come close to this method of linear commutation.
In general commutation will not be linear due to the presence of emf of self
induction and induced rotational emf in the coil. These result in retarded and accelerated
commutation and are discussed in sequence.
57
Ia 2Ia
I1 I2
i
αm = α’ = α"
0 Tc Time
Time of Time of Tc
communication -Ia 0 commutation
(a) (b)
Retarded commutation is mainly due to emf of self induction in the coil. Here
the current transfer from 1 to 2 gets retarded as the name suggests. This is best explained
with the help of time diagrams as shown in Fig. 26.(a). The variation of i is the change in
′
the current of the coil undergoing commutation, while i is that during linear commutation.
Fig. 26(b) shows the variation of I1 and current density in the brush at the leaving edge and
Fig. 26.(c) shows the same phenomenon with respect to I2 at entering edge. The value of
current in the coil is given by i undergoing commutation. αm is the mean current density in
the brush given by total current divided by brush area of cross section. αl and αe are the
current density under leaving and entering edges of the brush. As before,
I1 = Ia + i and I2 = Ia − i (31)
58
2Ia P α’=AB/AC
+Ia B
+Ia I1=Ia+i
i C αm
0 t Tc
i’
Q
-Ia 0 A t Tc
0 t Tc t
D
59
AB
αl = .αm (32)
AC
DF
αe = .αm (33)
DE
At the very end of commutation, the current density
′
di di
αe = αm . / (34)
dt dt
di 2Ia
= αm . /
dt Tc
If at this point di/dt = 0 the possibility of sudden breaking of the current and
hence the creation of an arc is removed .
60
α" α’ =AB/AC
α" =DB/DC
Ia D αm
C
α’
Ia B
0 i’ Tc
i
0 i’ Tc
i
B
-Ia A -Ia
(a1 ) (a2 )
α’ α"
Ia C S
A
αm B
Ia R α’ =PR/PQ
Q α" =SR/SQ
i Leaving edge
Entering
edge
0 Tc time
0 Tc time
i’
i
-Ia P -Ia
(b1 ) (b2 )
61
lead given to the brushes or by ii) having the field of suitable polarity at the position of the
brush with the help of a small pole called a commutating pole. In a non-inter pole machine
the brush shift must be changed from forward lead to backward lead depending upon gener-
ating or motoring operation. As the disadvantages of this brush shifts are to be avoided, it
is preferable to leave the brushes at geometric neutral axis and provide commutating poles
of suitable polarity (for a generator the polarity of the pole is the one towards which the
conductors are moving). The condition of commutation will be worse if commutating poles
are provided and not excited or they are excited but wrongly.
The action of the commutating pole is local to the coil undergoing commu-
tation. It does not disturb the main field distribution. The commutating pole winding
overpowers the armature mmf locally and establishes the flux of suitable polarity. The com-
mutating pole windings are connected in series with the armature of a d.c. machine to get
a load dependent compensation of armature reaction mmf.
The commutating pole are also known as compole or inter pole. The air gap
under compole is made large and the width of compole small. The mmf required to be
produced by compole is obtained by adding to the armature reaction mmf per pole Fa the
mmf to establish a flux density of required polarity in the air gap under the compole Fcp
.This would ensure straight line commutation. If sinusoidal commutation is required then
the second component Fcp is increased by 30 to 50 percent of the value required for straight
line commutation.
62
will be reduced by Fa * pole arc/pole pitch. This could have been predicted as the axis of
the compensating winding and armature winding is one and the same. Further, the mmf of
compensating winding opposes that of the armature reaction.
It is seen already that the equivalent circuit model of a d.c. machine becomes very
simple in view of the fact that the armature reaction is cross magnetizing. Also, the axis
of compensating mmf and mmf of commutating poles act in quadrature to the main field.
Thus flux under the pole shoe gets distorted but not diminished (in case the field is not
saturated). The relative connections of armature, compole and compensating winding are
unaltered whether the machine is working as a generator or as a motor; whether the load
is on the machine or not. Hence all these are connected permanently inside the machine.
The terminals reflect only the additional ohmic drops due to the compole and compensating
windings. Thus commutating pole winding, and compensating winding add to the resistance
of the armature circuit and can be considered a part of the same. The armature circuit
can be simply modelled by a voltage source of internal resistance equal to the armature
resistance + compole resistance + compensating winding resistance. The brushes behave
like non-linear resistance; and their effect may be shown separately as an additional constant
voltage drop equal to the brush drop.
The excitation for establishing the required field can be of two types a) Permanent
magnet excitation(PM) b) Electro magnetic excitation. Permanent magnet excitation is
63
Yoke
ly
lg lt lt lg lp
lp
Pole
Field
coil la
Armature
da
64
employed only in extremely small machines where providing a field coil becomes infeasible.
Also, permanent magnet excited fields cannot be varied for control purposes. Permanent
magnets for large machines are either not available or expensive. However, an advantage
of permanent magnet is that there are no losses associated with the establishment of the field.
The required ampere turns for establishing the desired flux per pole may be
computed by doing the magnetic circuit calculations. MMF required for the poles, air gap,
armature teeth, armature core and stator yoke are computed and added. Fig. 28 shows two
poles of a 4-pole machine with the flux paths marked on it. Considering one complete flux
loop, the permeance of the different segments can be computed as,
P = A.µ/l
Where P- permeance
A- Area of cross section of the part
mu- permeability of the medium
l- Length of the part
A flux loop traverses a stator yoke, armature yoke, and two numbers each
of poles, air gap, armature teeth in its path. For an assumed flux density Bg in the pole
region the flux crossing each of the above regions is calculated. The mmf requirement for
65
= B.A
From these expressions the mmf required for each and every part in the path of
the flux is computed and added. This value of mmf is required to establish two poles. It
is convenient to think of mmf per pole which is nothing but the ampere turns required to
be produced by a coil wound around one pole. In the case of small machines all this mmf
is produced by a coil wound around one pole. The second pole is obtained by induction.
This procedure saves cost as only one coil need be wound for getting a pair of poles. This
produces an unsymmetrical flux distribution in the machine and hence is not used in larger
machines. In large machines, half of total mmf is assigned to each pole as the mmf per pole.
The total mmf required can be produced by a coil having large number of turns but taking a
small current. Such winding has a high value of resistance and hence a large ohmic drop. It
can be connected across a voltage source and hence called a shunt winding. Such method of
excitation is termed as shunt excitation. On the other hand, one could have a few turns of
large cross section wire carrying heavy current to produce the required ampere turns. These
windings have extremely small resistance and can be connected in series with a large current
path such as an armature. Such a winding is called a series winding and the method of
excitation, series excitation. A d.c. machine can have either of these or both these types of
excitation.
These are shown in Fig. 29. When both shunt winding and series winding are
present, it is called compound excitation. The mmf of the two windings could be arranged to
aid each other or oppose each other. Accordingly they are called cumulative compounding
and differential compounding. If the shunt winding is excited by a separate voltage source
66
F1
A1
A1
F1
A1 F1 A1
67
then it is called separate excitation. If the excitation power comes from the same machine,
then it is called self excitation. Series generators can also be separately excited or self excited.
The characteristics of these generators are discussed now in sequence.
Ia=0
A2
Prime
mover E
n=const
Induced e.m.f
A1
Decreasing
Magnetisation
Increasing
magnetisation
+
F2 If
Vdc
e.m.f. due to Residual Magnetism
F1
- Exciting Current
(a) (b)
68
separate excitation is normally used for testing of d.c. generators to determine their open
circuit or magnetization characteristic. The excitation current is increased monotonically
to a maximum value and then decreased in the same manner, while noting the terminal
voltage of the armature. The load current is kept zero. The speed of the generator is held
at a constant value. The graph showing the nature of variation of the induced emf as a
function of the excitation current is called as open circuit characteristic (occ), or no-load
magnetization curve or no-load saturation characteristic. Fig. 30(b). shows an example. The
magnetization characteristic exhibits saturation at large values of excitation current. Due
to the hysteresis exhibited by the iron in the magnetic structure, the induced emf does not
become zero when the excitation current is reduced to zero. This is because of the remnant
field in the iron. This residual voltage is about 2 to 5 percent in modern machines. Separate
excitation is advantageous as the exciting current is independent of the terminal voltage
and load current and satisfactory operation is possible over the entire voltage range of the
machine starting from zero.
In a self excited machine, there is no external source for providing excitation current.
The shunt field is connected across the armature. For series machines there is no change in
connection. The series field continues to be in series with the armature.
Self excitation is now discussed with the help of Fig. 31.(a) The pro-
cess of self excitation in a shunt generator takes place in the following manner. When the
armature is rotated a feeble induced emf of 2 to 5 percent appears across the brushes de-
pending upon the speed of rotation and the residual magnetism that is present. This voltage
69
ms
210
oh
1500 rev/min
0
25
s
m
180
oh
0
17
1000 rev/min
s
ms
ohm
150 s
m
oh
oh
280
0
5
14
37
120
F2 A2 500 rev/min
90
Prime
mover 60
n
30
0 1.0
F1 A1
Exciting current,Amperes
Critical Resistance
160 180
s
hm
s
5o
hm
150
12
0o
25
120
120
ms
oh
60
80 90
40 60
30
200 400
0 1500
Total field circuit resistance, ohms speed in rev/min
70
D C Open circuit
characteristic
Q’
R’
Voltage
Q QL
R RL
P’
P PL Armature drop
P1 Q1 characteristic
O’ A’
0 P" Q" A
Excitation current If
Armature current Ia
gets applied across the shunt field winding and produces a small mmf. If this mmf is such
as to aid the residual field then it gets strengthened and produces larger voltage across the
brushes. It is like a positive feed back. The induced emf gradually increases till the voltage
induced in the armature is just enough to meet the ohmic drop inside the field circuit. Under
such situation there is no further increase in the field mmf and the build up of emf also stops.
If the voltage build up is ‘substantial’, then the machine is said to have ‘self excited’.
Fig. 31(b) shows the magnetization curve of a shunt generator. The field resistance
line is also shown by a straight line OC. The point of intersection of the open circuit charac-
71
teristic (OCC) with the field resistance line, in this case C, represents the voltage build up
on self excitation. If the field resistance is increased, at one point the resistance line becomes
a tangent to the OCC. This value of the resistance is called the critical resistance. At this
value of the field circuit resistance the self excitation suddenly collapses. See Fig. 31(c). In-
stead of increasing the field resistance if the speed of the machine is reduced then the same
resistance line becomes a critical resistance at a new speed and the self excitation collapses
at that speed. In this case, as the speed is taken as the variable, the speed is called the
critical speed. In the linear portion of the OCC the ordinates are proportional to the speed
of operation, hence the critical resistance increases as a function of speed Fig. 31.(b) and (d).
4. The speed of operation of the machine must be above the critical speed.
Remedial measures to be taken if the machine fails to self excite are briefly
discussed below.
1. The residual field will be absent in a brand new, unexcited, machine. The field may
be connected to a battery in such cases for a few seconds to create a residual field.
72
2. The polarity of connections have to be set right. The polarity may become wrong
either by reversed connections or reversed direction of rotation. If the generator had
been working with armature rotating in clockwise direction before stopping and if one
tries to self excite the same with counter clockwise direction then the induced emf
opposes residual field, changing the polarity of connections of the field with respect to
armature is normally sufficient for this problem.
3. Field circuit resistance implies all the resistances coming in series with the field winding
like regulating resistance, contact resistance, drop at the brushes, and the armature
resistance. Brush contact resistance is normally high at small currents. The dirt on
the commutator due to dust or worn out mica insulator can increase the total circuit
resistance enormously. The speed itself might be too low so that the normal field
resistance itself is very much more than the critical value. So ensuring good speed,
clean commutator and good connections should normally be sufficient to overcome this
problem.
4. Speed must be increased sufficiently to a high value to be above the critical speed.
5. The load switch must be opened or the load resistance is made very high.
The conditions for self excitation of a series generator remain similar to that of
a shunt machine. In this case the field circuit resistance is the same as the load circuit
resistance and hence it must be made very low to help self excitation. To control the field
mmf a small resistance called diverter is normally connected across the series field. To help
in the creation of maximum mmf during self excitation any field diverter if present must be
73
Terminal voltage
Open circuit PS=PQ-PR
characteristic
Q
Armature
S characteristic
A R B
External
characteristic
P
0
Load Current
open circuited.In a series generator load current being the field current of the machine the
self excitation characteristic or one and the same. This is shown in Fig. 33
Most of the compound machines are basically shunt machines with the series wind-
ing doing the act of strengthening/weakening the field on load, depending up on the con-
nections. In cumulatively compounded machines the mmf of the two fields aid each other
and in a differentially compounded machine they oppose each other. Due to the presence of
the shunt winding, the self excitation can proceed as in a shunt machine. A small difference
exists however depending up on the way the shunt winding is connected to the armature. It
can be a short shunt connection or a long shunt connection. In long shunt connection the
shunt field current passes through the series winding also. But it does not affect the process
of self excitation as the mmf contribution from the series field is negligible.
74
Both series field winding and shunt field winding are wound around the main
poles. If there is any need, for some control purposes, to have more excitation windings
of one type or the other they will also find their place on the main poles. The designed
field windings must cater to the full range of operation of the machine at nominal armature
current. As the armature current is cross magnetizing the demagnetization mmf due to pole
tip saturation alone need be compensated by producing additional mmf by the field.
The d.c. machines give rise to a variety of external characteristics with consid-
erable ease. The external characteristics are of great importance in meeting the requirements
of different types of loads and in parallel operation. The external characteristics, also known
as load characteristics, of these machines are discussed next.
Load characteristics are also known as the external characteristics. External char-
acteristics expresses the manner in which the output voltage of the generator varies as a
function of the load current, when the speed and excitation current are held constant. If
they are not held constant then there is further change in the terminal voltage. The terminal
voltage V can be expressed in terms of the induced voltage E and armature circuit drop as
V = E − Ia Ra − Vb (36)
75
Terminal voltage V B
Volts
0 A
Load current,Ia
As seen from the equation E being function of speed and flux per pole it will
also change when these are not held constant. Experimentally the external characteristics
can be determined by conducting a load test. If the external characteristic is obtained by
subtracting the armature drop from the no-load terminal voltage, it is found to depart from
the one obtained from the load test. This departure is due to the armature reaction which
causes a saturation at one tip of each pole. Modern machines are operated under certain
degree of saturation of the magnetic path. Hence the reduction in the flux per pole with
load is obvious. The armature drop is an electrical drop and can be found out even when
the machine is stationary and the field poles are unexcited. Thus there is some slight droop
in the external characteristics, which is good for parallel operation of the generators.
76
One could easily guess that the self excited machines have slightly higher droop
in the external characteristic as the induced emf E drops also due to the reduction in the
applied voltage to the field. If output voltage has to be held constant then the excitation
current or the speed can be increased. The former is preferred due to the ease with which it
can be implemented. As seen earlier, a brush lead gives rise to a load current dependent mmf
along the pole axis. The value of this mmf magnetizes/demagnetizes the field depending on
whether the lead is backward or forward.
For a given no-load voltage a self excited machine will have more voltage drop at
the terminals than a separately excited machine, as the load is increased. This is due to
the dependence of the excitation current also on the terminal voltage. After certain load
current the terminal voltage decreases rapidly along with the terminal current, even when
load impedance is reduced. The terminal voltage reaches an unstable condition. Also, in a
self excited generator the no-load terminal voltage itself is very sensitive to the point of inter-
section of the magnetizing characteristics and field resistance line. The determination of the
external characteristics of a shunt generator forms an interesting study. If one determines
the load magnetization curves at different load currents then the external characteristics
can be easily determined. Load magnetization curve is a plot showing the variation of the
terminal voltage as a function of the excitation current keeping the speed and armature cur-
rent constant. If such curves are determined for different load currents then by determining
the intersection points of these curves with field resistance line one can get the external
characteristics of a shunt generator. Load saturation curve can be generated from no-load
saturation curve /OCC by subtracting the armature drop at each excitation point. Thus
77
it is seen that these family of curves are nothing but OCC shifted downwards by armature
drop. Determining their intercepts with the field resistance line gives us the requisite result.
Instead of shifting the OCC downwards, the x axis and the field resistance line is shifted ‘up-
wards’ corresponding to the drops at the different currents, and their intercepts with OCC
are found. These ordinates are then plotted on the original plot. This is shown clearly in
Fig. 32. The same procedure can be repeated with different field circuit resistance to yield
external characteristics with different values of field resistance. The points of operation up to
the maximum current represent a stable region of operation. The second region is unstable.
The decrease in the load resistance decreases the terminal voltage in this region.
In the case of series generators also, the procedure for the determination of the
external characteristic is the same. From the occ obtained by running the machine as a sep-
arately excited one, the armature drops are deducted to yield external /load characteristics.
The armature drop characteristics can be obtained by a short circuit test as before.
Fig. 33 shows the load characteristics of a series generator. The first half of
the curve is unstable for constant resistance load. The second half is the region where series
generator connected to a constant resistance load could work stably. The load characteristics
in the first half however is useful for operating the series generator as a booster. In a booster
the current through the machine is decided by the external circuit and the voltage injected
into that circuit is decided by the series generator. This is shown in Fig. 35
78
- +
S1 S2
E2
A1 A2
F2 A2
Booster
Generator
E=E1 +- E2
E1
F1 A1
Main generator
79
IL
S2
If
F2 S1
A2
Load
Prime
Vf # mover
A1
F1
(a)-Connection
Over compounded
Level compounded
Under compounded
Shunt machine
Terminal voltage
Differential
compounding
Load current
(b)-Characteristics
80
D.C. generators are required to operate in parallel supplying a common load when
the load is larger than the capacity of any one machine. In situations where the load is small
but becomes high occasionally, it may be a good idea to press a second machine into operation
only as the demand increases. This approach reduces the spare capacity requirement and its
cost. In cases where one machine is taken out for repair or maintenance, the other machine
can operate with reduced load. In all these cases two or more machines are connected to
operate in parallel.
Parallel operation of two shunt generators is similar to the operation of two storage
batteries in parallel. In the case of generators we can alter the external characteristics easily
while it is not possible with batteries. Before connecting the two machines the voltages of
the two machines are made equal and opposing inside the loop formed by the two machines.
This avoids a circulating current between the machines. The circulating current produces
power loss even when the load is not connected. In the case of the loaded machine the
difference in the induced emf makes the load sharing unequal.
Fig. 37 shows two generators connected in parallel. The no load emfs are made
equal to E1 = E2 = E on no load; the current delivered by each machine is zero. As the load
is gradually applied a total load current of I ampere is drawn by the load. The load voltage
81
s1
v s2
A2 A2
Prime
mover Load
G1 G2
A1 A1
F2 F2
Vf1 Vf2
F1 F1
E1
V2 k V0 j
Total char
E2 a cteristic
V A B C
Terminal
V2
Voltage
I1 V1
I2
I=I1+I2
O D
Load current
82
under these conditions is V volt. Each machines will share this total current by delivering
currents of I1 and I2 ampere such that I1 + I2 = I.
Also terminal voltage of the two machines must also be V volt. This is dictated
by the internal drop in each machine given by equations
where Ra1 and Ra2 are the armature circuit resistances. If load resistance RL is known these
equations can be solved analytically to determine I1 and I2 and hence the manner in which
to total output power is shared. If RL is not known then an iterative procedure has to be
adopted. A graphical method can be used with advantage when only the total load current is
known and not the value of RL or V . This is based on the fact that the two machines have a
common terminal voltage when connected in parallel. In Fig. 38 the external characteristics
of the two machines are first drawn as I and II . For any common voltage the intercepts OA
and OB are measured and added and plotted as point at C. Here OC = OA + OB . Thus
a third characteristics where terminal voltage is function of the load current is obtained.
This can be called as the resultant or total external characteristics of the two machines put
together. With this, it is easy to determine the current shared by each machine at any total
load current I.
The above procedure can be used even when the two voltages of the machines
at no load are different. At no load the total current I is zero ie I1 + I2 = 0 or I1 = −I2 .
Machine I gives out electrical power and machine II receives the same. Looking at the voltage
equations, the no load terminal equation Vo becomes
83
As can be seen larger the values of Ra1 and Ra2 larger is the tolerance for the error
between the voltages E1 and E2 . The converse is also true. When Ra1 and Ra2 are nearly zero
implying an almost flat external characteristic, the parallel operation is extremely difficult.
Series generators are rarely used in industry for supplying loads. Some applications
like electric braking may employ them and operate two or more series generates in parallel.
Fig. 39 shows two series generators connected in parallel supplying load current of I1 and I2 .
If now due to some disturbance E1 becomes E1 + ∆E1 then the excitation of the machine I
increases, increasing the load current delivered. As the total current is I the current supplied
by machine II reduces, so also its excitation and induced emf. Thus machine I takes greater
and greater fraction of the load current with machine II shedding its load. Ultimately the
current of machine II becomes negative and it also loads the first machine. Virtually there is
a short circuit of the two sources, the whole process is thus highly unstable. One remedy is
for a problem as this is to make the two fields immune to the circulating current between the
machines. This is done by connecting an equalizer between the fields as shown in Fig. 39-a
. With the equalizer present, a momentary disturbance does not put the two machines out
of action. A better solution for such problems is to cross connect the two fields as shown in
Fig. 39-b. A tendency to supply a larger current by a machine strengthens the field of the
next machine and increases its induced emf . This brings in stable conditions for operation
rapidly.
84
I1 I1+I2
A2 + A2 +
I1 I2
A1 -
I1+I2 V
A1 -
S2 Equaliser S2
F1 F2
S1 S1 I1+I2
(a)Equalizer connection
I1+I2
I1 I2
A2 + A2 +
G1 G2
A1 - A1 - Load V
I1
S2 S2
I2
S1 I2 - S1
- I1+I2
85
F2 + A2 F2 + A2
Load V
A1 A1
- -
F1 F1
Equalizer
S2 S2
S1 S1
(a)Equalizer connection
F2 + A2 F2 + A2
Load V
A1 A1
- -
F1 F1
S2 S2
S1 S1
86
D.C. motors have a place of pride as far as electrical drives are considered. The
simplicity, and linearity of the control method makes them highly preferred machines in
precision drives. In spite of the great advancements in a.c. drives these machines are still
sought after by the industries. Apart from high precision application they are preferred in
stand alone systems working on batteries and high speed drives off constant voltage mains.
After the field is excited if we pass a current through the armature the rotor experiences
a torque and starts rotating. The direction of the torque can be readily obtained from the
law of interaction. These moving conductors cut the field and induce emf, usually called the
’back emf’ according to Lenz’s law and act as a sink of electrical power from the electrical
source. This absorbed power appears as mechanical power. The converted mechanical power
should overcome the frictional and iron losses before useful work could be done by the same.
The connections to the supply of a d.c. shunt motor are given in Fig. 41.
87
+ +
F2
A2 F2 A2
A1 F1 A1
F1
- -
(a)Separate excitation (b) Shunt excitation
s1
F2 +
s2
A2
DC
Supply
A1
F1
-
(c)Practical arrangement
88
Commonly used connection is where in both the field and the armature are
energized simultaneously Fig. 41(b). As the field has higher inductance and time constant
torque takes some time to reach the full value corresponding to a given armature current.
In Fig. 41.(c), the switch S1 is closed a few seconds prior to switch S2 . By then the field
current would have reached the steady value. So the torque per ampere is high in this case.
The only difference in the second connection Fig. 41.(a) is that the shunt field
winding is connected to a separate source. This connection is used when the armature and
field voltage are different as is common in high voltage d.c. machines. The field voltage is
kept low in such cases for the sake of control purposes. Here again the field circuit must
be energized prior to the armature. Suitable interlock should be provided to prevent the
armature switch being closed prior to / without closing of field circuit as the armature
currents reach very large values still not producing any torque or rotation. The relevant
equations for the motoring operation can be written as below
V − E − Ia Ra − Vb = 0 or E = V − Ia Ra − Vb (39)
p.φ.Z.n pZ
E= = Ke φ.n where Ke = (40)
b b
1 p.φ.ZIa 1 pZ
TM = . = Kt φIa where Kt = . (41)
2π b 2π b
dw
and TM − TL = J (42)
dt
where
TL - Load torque
TM - Motor torque
J - polar moment of inertia.
w - angular velocity = 2π.n
89
The first one is an electrical equation, the second and the third are electro
mechanical in nature and the last equation is the mechanical equation of motion. Ke and
Kt are normally termed as back emf constant and torque constant respectively. Under
steady speed of operation the fourth equation is not required. Using these equations one
can determine the torque speed characteristics of the machine for a given applied voltage.
These characteristics are similar to the external characteristics for a generator. Here the
torque on the machine is assumed to be varying and the corresponding speed of operation
is determined. This is termed as the torque speed characteristic of the motor.
Knowing the value of E and flux one can determine the value of the speed.
Also knowing the armature current and the flux, the value of the torque is found out. This
procedure is repeated for different values of the assumed armature currents and the values
are plotted as in Fig. 42-(a). From these graphs, a graph indicating speed as a function of
torque or the torque-speed characteristics is plotted Fig. 42-(b)(i).
As seen from the figure the fall in the flux due to load increases the speed due
to the fact that the induced emf depends on the product of speed and flux. Thus the speed
90
Line voltage
A No load speed B
Speed
E
F
Torque Flux
G
0 Armature current
(a)Load characteristics
(ii)
(i)
Speed
0 Torque
(b)Torque speed curve
91
of the machine remains more or less constant with load. With highly saturated machines
the on-load speed may even slightly increase at over load conditions. This effects gets more
pronounced if the machine is designed to have its normal field ampere turns much less
than the armature ampere turns. This type of external characteristics introduces instability
during operation Fig. 42(b)(ii) and hence must be avoided. This may be simply achieved by
providing a series stability winding which aids the shunt field mmf.
Following the procedure described earlier under shunt motor, the torque speed
characteristics of a series motor can also be determined. The armature current also happens
to be the excitation current of the series field and hence the flux variation resembles the
magnetization curve of the machine. At large value of the armature currents the useful flux
would be less than the no-load magnetization curve for the machine. Similarly for small
values of the load currents the torque varies as a square of the armature currents as the flux
is proportional to armature current in this region. As the magnetic circuit becomes more
and more saturated the torque becomes proportional to Ia as flux variation becomes small.
Fig. 43(a) shows the variation of E1 , flux , torque and speed following the above procedure
from which the torque-speed characteristics of the series motor for a given applied voltage
V can be plotted as shown in Fig. 43.(b) The initial portion of this torque-speed curve is
seen to be a rectangular hyperbola and the final portion is nearly a straight line. The speed
under light load conditions is many times more than the rated speed of the motor. Such
high speeds are unsafe, as the centrifugal forces acting on the armature and commutator
can destroy them giving rise to a catastrophic break down. Hence series motors are not
recommended for use where there is a possibility of the load becoming zero. In order to
92
Terminal voltage
Back emf
No load Useful
Magnetisation Flux
Torque, Flux and Speed curve
Developed Useful
Torque Torque
Speed
Load current
(a)Load characteristics
Speed
0 Torque
93
safeguard the motor and personnel, in the modern machines, a ‘weak’ shunt field is provided
on series motors to ensure a definite, though small, value of flux even when the armature
current is nearly zero. This way the no-load speed is limited to a safe maximum speed. It is
needless to say, this field should be connected so as to aid the series field.
Two situations arise in the case of compound motors. The mmf of the shunt field
and series field may oppose each other or they may aid each other. The first configuration
is called differential compounding and is rarely used. They lead to unstable operation of
the machine unless the armature mmf is small and there is no magnetic saturation. This
mode may sometimes result due to the motoring operation of a level-compounded generator,
say by the failure of the prime mover. Also, differential compounding may result in large
negative mmf under overload/starting condition and the machine may start in the reverse
direction. In motors intended for constant speed operation the level of compounding is very
low as not to cause any problem.
Cumulatively compounded motors are very widely used for industrial drives.
High degree of compounding will make the machine approach a series machine like charac-
teristics but with a safe no-load speed. The major benefit of the compounding is that the
field is strengthened on load. Thus the torque per ampere of the armature current is made
high. This feature makes a cumulatively compounded machine well suited for intermittent
peak loads. Due to the large speed variation between light load and peak load conditions, a
fly wheel can be used with such motors with advantage. Due to the reasons provided under
shunt and series motors for the provision of an additional series/shunt winding, it can be
94
seen that all modern machines are compound machines. The difference between them is only
in the level of compounding.
95
96
Speed
A B C D
Motors I and II
I II in parallel
0 Torque
97
In the case of series operation the motors shafts of the two machines are connected
to the same load and also the two armatures are series connected. This forces a common
armature current through both the machines and the torques developed by the machines
are proportional to the flux in each machine. Series operation of series motors is adopted
during starting to improve the energy efficiency. This method is ideally suited for shunt
and compound machines with nearly flat torque speed characteristics. Such machines can go
through high amount of dynamics without the fear of becoming unstable. This configuration
is used in steel mills. Having two smaller machines connected to the shaft is preferred over
there in place of one large machine as the moment of inertia of the motors is much reduced,
thus improving the dynamics.
98
Some elementary principles of application alone are dealt with here. The focus is
on the mechanical equation of dynamics which is reproduced here once again.
dw
TM − TL = J (43)
dt
Here TM and TL are the motor torque and the load torques respectively which are expressed
as functions of ω. Under steady state operation dω/dt will be zero. The application of
motors mainly looks at three aspects of operation.
1. Starting
2. Speed control
3. Braking
The speed of the machine has to be increased from zero and brought to the op-
erating speed. This is called starting of the motor. The operating speed itself should be
varied as per the requirements of the load. This is called speed control. Finally, the running
machine has to be brought to rest, by decelerating the same. This is called braking. The
torque speed characteristics of the machine is modified to achieve these as it is assumed
that the variation in the characteristics of the load is either not feasible or desirable. Hence
the methods that are available for modifying the torque speed characteristics and the actual
variations in the performance that these methods bring about are of great importance. When
more than one method is available for achieving the same objective then other criteria like,
initial cost, running cost, efficiency and ease operation are also applied for the evaluation of
the methods. Due to the absence of equipment like transformer, d.c. machine operation in
99
For the machine to start, the torque developed by the motor at zero speed must
exceed that demanded by the load. Then TM − TL will be positive so also is dω/dt, and the
machine accelerates. The induced emf at starting point is zero as the ω = 0 The armature
current with rated applied voltage is given by V /Ra where Ra is armature circuit resistance.
Normally the armature resistance of a d.c. machine is such as to cause 1 to 5 percent drop
at full load current. Hence the starting current tends to rise to several times the full load
current. The same can be told of the torque if full flux is already established. The machine
instantly picks up the speed. As the speed increases the induced emf appears across the
terminals opposing the applied voltage. The current drawn from the mains thus decreases,
so also the torque. This continues till the load torque and the motor torque are equal to
each other. Machine tends to run continuously at this speed as the acceleration is zero at
this point of operation.
The starting is now discussed with respect to specific machines.
100
If armature and field of d.c. shunt motor are energized together, large current is
drawn at start but the torque builds up gradually as the field flux increases gradually. To
improve the torque per ampere of line current drawn it is advisable to energize the field
first. The starting current is given by V /Ra and hence to reduce the starting current to a
safe value, the voltage V can be reduced or armature circuit resistance Ra can be increased.
Variable voltage V can be obtained from a motor generator set. This arrangement is called
Ward-Leonard arrangement. A schematic diagram of Ward-Leonard arrangement is shown
in Fig. 45. By controlling the field of the Ward-Leonard generator one can get a variable
voltage at its terminals which is used for starting the motor.
The second method of starting with increased armature circuit resistance can
be obtained by adding additional resistances in series with the armature, at start. The
current and the torque get reduced. The torque speed curve under these conditions is shown
in Fig. 46(a) . It can be readily seen from this graph that the unloaded machine reaches its
final speed but a loaded machine may crawl at a speed much below the normal speed. Also,
the starting resistance wastes large amount of power. Hence the starting resistance must
be reduced to zero at the end of the starting process. This has to be done progressively,
making sure that the current does not jump up to large values. Starting of series motor and
compound motors are similar to the shunt motor. Better starting torques are obtained for
compound motors as the torque per ampere is more. Characteristics for series motors are
given in fig. 47.
101
+ -
A2 A2 A2
Load variable
M G M
voltage
A1 A1 A1
+ + +
F2 F2 F2
constant
voltage
mains
F1
- F1
-
F1 -
(a)
+
A2
Variable Load
voltage
dc A1
Auto
Diode -
transformer
Constant bridge
voltage
ac mains
+
F2
Static Ward
Leonard system
F1
-
(b)
102
+ - Rext = 0
Rext
F2 A2 Rext increasing
E1
Speed
F1 A1
Constant voltage
source 0 Torque
(a)
v + -
If
F2 > A2
Rext If2 < If rated If2
Vf E1 If rated
Speed
F1 A1
Constant voltage
source
0
Torque
(b)
+ -
F2 A2 V1
Vf E1 V
V2
Speed
F1 A1
V3
V3 < V2 < V1
Variable voltage
source 103
0 Torque
(c)
v
S2 + -
Rext
S1
A2
Speed
E1
A1
Rext = 0
Rext > 0
Constant voltage
0 Torque
sources
(a)
+ -
S2
v
Rd
S1
Speed
A2
Rd =
8
E1
A1 Constant Rd reducing
voltage
0 Torque
sources
(b)
S2 + -
S1
A2
Variable
Speed
V Vrated
voltage
M
A1
V reducing
104
0 Torque
(c)
A2 F2
ra
Rn+1
A1
n+1
rn
R3 n F1
n-1 rn-1
R2
R1 r2
3
2
r1
1
(a)Physical connection
graphical method
Rn+1
Rn
Starting current with time
Rn-1 Imax
R3
Volts
R2 Imin
0 Ia R1
0
Imin Imax Time
105
If the starting resistor is reduced in uniform steps then the current peaks reached
as we cut down the resistances progressively increase. To ascertain that at no step does
the current jump to a large value non-uniform reduction of resistances must be assorted to.
This use of a non-uniform resistance step is called ‘grading’ of the resistors. The calculations
for a starter resistance of a shunt motor are shown below with the help of Fig. 48. In
the figure an n element or n+1 step starter is shown. The armature resistance when all
the external resistances are cut off is ra . The total armature circuit resistance at step 1 is
R1 = (r1 + r2 + ... + rn ) + ra . The field winding is connected across the supply. The starting
current reaches a maximum value Im ax when we move on to a step. One resistance element
is cut from the circuit when the current falls down to Im in . During the instant when the
element is cut the speed and hence the induced emf does not change but the current jumps
back to Im ax . Thus during the starting the current changes between two limits Im ax and
Im in. Writing the expression for the current before and after the resistance is changed on
step Ri and Ri+1 , we have
V −E V −E Im ax Ri
Im in = Im ax = or = (47)
Ri Ri+1 Im in Ri+1
Im ax R1 R2 Rn−1 Rn
= = = ... = = = k(say) (48)
Im in R2 R3 Rn Rn+1
r
n R1 R2 Rn R1 R1 R1
k = ∗ ∗ ... ∗ = = k= n (49)
R2 R3 Rn+1 Rn+1 ra ra
106
Sometimes the ratio k may be required to be fixed. Then the number of steps required can
be calculated as
R1 log Rra1
n log k = log ,n = (50)
ra log k
log R1 − log Rn
log k
Also, r r r
V V V
R= n
= n
= n+1
(51)
I1 ra RI2 ra I2 ra
From these expressions it is seen that to have the ratio k to be unity, the number
of steps should be infinity. Smaller the number of steps larger is the ratio of maximum to
minimum current. Also, it is not possible to choose n and k independently. Im ax is set
by the maximum possible starting current from the point of view of commutation. Im in
is found from the minimum torque against which the starting is required to be performed.
Similar method exists in the case of series motors and compound motors. In these cases the
ratio of currents and the ratio of fluxes are needed. The equation becomes non-linear and a
graphical method is normally adopted for the design of the resistances in those cases.
Resistance method of starting is cheaper and simple and hence is used univer-
sally. But it wastes energy in the starting resistor. Hence this method is not advised when
frequent starting of the motor is required. Ward-Leonard method gives a energy efficient
method of starting. With the help of a auto transformer and rectifier set one can get variable
voltage d.c. supply from a constant voltage a.c power source. This is some times called a
static Ward-Leonard arrangement. This method is becoming more popular over the rotating
machine counter part.
107
In the case of speed control, armature voltage control and flux control methods
are available. The voltage control can be from a variable voltage source like Ward-Leonard
arrangement or by the use of series armature resistance. Unlike the starting conditions the
series resistance has to be in the circuit throughout in the case of speed control. That means
considerable energy is lost in these resistors. Further these resistors must be adequately
cooled for continuous operation. The variable voltage source on the other hand gives the
motor the voltage just needed by it and the losses in the control gear is a minimum. This
method is commonly used when the speed ratio required is large, as also the power rating.
Field control or flux control is also used for speed control purposes. Normally
field weakening is used. This causes operation at higher speeds than the nominal speed.
Strengthening the field has little scope for speed control as the machines are already in a
state of saturation and large field mmf is needed for small increase in the flux. Even though
flux weakening gives higher speeds of operation it reduces the torque produced by the ma-
chine for a given armature current and hence the power delivered does not increase at any
armature current. The machine is said to be in constant power mode under field weakening
mode of control. Above the nominal speed of operation, constant flux mode with increased
applied voltage can be used; but this is never done as the stress on the commutator insulation
increases.
Thus operation below nominal speed is done by voltage control. Above the
nominal speed field weakening is adopted. For weakening the field, series resistances are used
for shunt as well as compound motors. In the case of series motors however field weakening
108
is done by the use of ’diverters’ . Diverters are resistances that are connected in parallel to
the series winding to reduce the field current without affecting the armature current.
When a motor is switched off it ‘coasts’ to rest under the action of frictional forces.
Braking is employed when rapid stopping is required. In many cases mechanical braking
is adopted. The electric braking may be done for various reasons such as those mentioned
below:
5. To improve the through put in many production process by reducing the stopping time.
In many cases electric braking makes more brake power available to the braking
process where mechanical brakes are applied. This reduces the wear and tear of the me-
chanical brakes and reduces the frequency of the replacement of these parts. By recovering
the mechanical energy stored in the rotating parts and pumping it into the supply lines
the overall energy efficiency is improved. This is called regeneration. Where the safety of
the personnel or the equipment is at stake the machine may be required to stop instantly.
Extremely large brake power is needed under those conditions. Electric braking can help
in these situations also. In processes where frequent starting and stopping is involved the
109
process time requirement can be reduced if braking time is reduced. The reduction of the
process time improves the throughput.
Basically the electric braking involved is fairly simple. The electric motor can
be made to work as a generator by suitable terminal conditions and absorb mechanical energy.
This converted mechanical power is dissipated/used on the electrical network suitably.
Braking can be broadly classified into:
1. Dynamic
2. Regenerative
These are now explained briefly with reference to shunt ,series and compound motors.
• Shunt machine
In dynamic braking the motor is disconnected from the supply and connected to a
dynamic braking resistance RDB . In and Fig. 49 this is done by changing the switch
from position 1 to 2 . The supply to the field should not be removed. Due to the
rotation of the armature during motoring mode and due to the inertia, the armature
continues to rotate. An emf is induced due to the presence of the field and the rotation.
This voltage drives a current through the braking resistance. The direction of this
current is opposite to the one which was flowing before change in the connection.
Therefore, torque developed also gets reversed. The machine acts like a brake. The
110
torque speed characteristics separate by excited shunt of the machine under dynamic
braking mode is as shown in Fig. 49(b) for a particular value of RDB . The positive
torque corresponds to the motoring operation. Fig. 50 shows the dynamic braking of
a shunt excited motor and the corresponding torque-speed curve. Here the machine
behaves as a self excited generator.
Below a certain speed the self-excitation collapses and the braking action becomes
Zero.
• Series machine
In the case of a series machine the excitation current becomes zero as soon as the
armature is disconnected from the mains and hence the induced emf also vanishes. In
order to achieve dynamic braking the series field must be isolated and connected to
a low voltage high current source to provide the field. Rather, the motor is made to
work like a separately excited machine. When several machines are available at any
spot, as in railway locomotives, dynamic braking is feasible. Series connection of all
the series fields with parallel connection of all the armatures connected across a single
dynamic braking resistor is used in that case.
• Compound generators
In the case of compound machine, the situation is like in a shunt machine. A separately
excited shunt field and the armature connected across the braking resistance are used.
A cumulatively connected motor becomes differentially compounded generator and the
braking torque generated comes down. It is therefore necessary to reverse the series
field if large braking torques are desired.
111
1 +
F2 + A2 RDB
Vf E
F1
- A1 2
1
-
(a)Connections
Torque 0
(b)Characteristics
112
1 +
F2 + A2 RDB
Vf E
F1
- A1 2
1
-
(a)Connections
Torque 0
(b)Characteristics
113
In regenerative braking as the name suggests the energy recovered from the rotating
masses is fed back into the d.c. power source. Thus this type of braking improves the energy
efficiency of the machine. The armature current can be made to reverse for a constant voltage
operation by increase in speed/excitation only. Increase in speed does not result in braking
and the increase in excitation is feasible only over a small range, which may be of the order of
10 to 15%. Hence the best method for obtaining the regenerative braking is to operate, the
machine on a variable voltage supply. As the voltage is continuously pulled below the value of
the induced emf the speed steadily comes down. The field current is held constant by means
of separate excitation. The variable d.c. supply voltage can be obtained by Ward-Leonard
arrangement, shown schematically in Fig. 51. Braking torque can be obtained right up to
zero speed. In modern times static Ward-Leonard scheme is used for getting the variable
d.c. voltage. This has many advantages over its rotating machine counter part. Static set
is compact, has higher efficiency, requires lesser space, and silent in operation; however it
suffers from drawbacks like large ripple at low voltage levels, unidirectional power flow and
low over load capacity. Bidirectional power flow capacity is a must if regenerative braking is
required. Series motors cannot be regeneratively braked as the characteristics do not extend
to the second quadrant.
8.3.3 Plugging
The third method for braking is by plugging.Fig. 52 shows the method of connection
for the plugging of a shunt motor. Initially the machine is connected to the supply with the
switch S in position number 1. If now the switch is moved to position 2, then a reverse
voltage is applied across the armature. The induced armature voltage E and supply voltage
114
+ -
If
F2 A2 Variable
votage
Vf E V source
F1 A1
(a)Physical connection
Speed
A
B V1
C
V2
V1 > V2
0 Torque
(b)Characteristics
115
2 RB
+ -
1
+
F2 A2
V
Vf
E
F1 2
- A1
1
(a)Physical connection
A
B
Speed
C 0 Torque
(b)Characteristics
116
V aid each other and a large reverse current flows through the armature. This produces a
large negative torque or braking torque. Hence plugging is also termed as reverse voltage
braking. The machine instantly comes to rest. If the motor is not switched off at this instant
the direction of rotation reverses and the motor starts rotating the reverse direction. This
type of braking therefore has two modes viz. 1) plug to reverse and 2) plug to stop. If we
need the plugging only for bringing the speed to zero, then we have to open the switch S
at zero speed. If nothing is done it is plug to reverse mode. Plugging is a convenient mode
for quick reversal of direction of rotation in reversible drives. Just as in starting, during
plugging also it is necessary to limit the current and thus the torque, to reduce the stress on
the mechanical system and the commutator. This is done by adding additional resistance in
series with the armature during plugging.
• Series motors
In the case of series motors plugging cannot be employed as the field current too gets
reversed when reverse voltage is applied across the machine. This keeps the direction
of the torque produced unchanged. This fact is used with advantage, in operating a
d.c. series motor on d.c. or a.c. supply. Series motors thus qualify to be called as
‘Universal motors’.
• Compound motors
Plugging of compound motors proceeds on similar lines as the shunt motors. However
some precautions have to be observed due to the presence of series field winding. A
cumulatively compounded motor becomes differentially compounded on plugging. The
mmf due to the series field can ’over power’ the shunt field forcing the flux to low values
or even reverse the net field. This decreases the braking torque, and increases the
duration of the large braking current. To avoid this it may be advisable to deactivate
117
the series field at the time of braking by short circuiting the same. In such cases the
braking proceeds just as in a shunt motor. If plugging is done to operate the motor
in the negative direction of rotation as well, then the series field has to be reversed
and connected for getting the proper mmf. Unlike dynamic braking and regenerative
braking where the motor is made to work as a generator during braking period, plugging
makes the motor work on reverse motoring mode.
It is seen from the earlier sections that the d.c.machine is capable of having variety of
torque-speed characteristics depending on the circuit conditions. The need for generating
these characteristics will be clear only when they are seen along with the characteristics of
the loads that they operate with. Even though a detailed treatment of motor load systems is
outside the scope here, it may be useful to look into the typical torque-speed characteristics
of some of the common loads.
Loads are broadly divided into,
Passive loads absorb the mechanical energy developed by the motors while
active loads are capable of working as both sinks and sources for mechanical energy. The
direction of rotation may be taken to be clockwise/counter clockwise rotation. Normally the
118
direction in which the load operates most of the time, is taken as the positive direction of
rotation. Any torque which accelerates the motor load system in the positive direction of
rotation is termed as a positive toque. With this rotation torques of motors, generators or
loads can be represented graphically on a four quadrantal diagram. The torque being taken
as an independent variable, is represented along the x-axis. Y-axis represents the speed.
Quadrants. I and III in Fig. 53(a) represent ‘forward motoring’ and ‘reverse motoring’ op-
eration respectively. Quadrants II and IV similarly represent generating/braking quadrants
as they absorb mechanical power and cause braking action.
Fig. 53(b) shows a few typical load characteristics on a four quadrantal dia-
gram.
The characteristics a, b,and c correspond to frictional torque, cutting torque and fan torque
respectively. While the frictional torque is not a function of speed, the cutting toque is pro-
portional to the speed and the fan torque varies as the square of the speed. These can only
absorb mechanical power and hence are represented in quadrantal II for positive direction
of rotation. Similar loads produce characteristics in quadrant IV for negative direction of
rotation.
Fig. 54 shows a typical behaviour of an active load. Here an elevator is taken as an example.
Here the counter weight is assumed to be heavier than the cage and similarly the loaded
cage in assumed to be heavier than the counter weight. As seen from the Fig. 54 the torque
is constant and depends on the difference in the weight of the case and the counter weight,
and the radius of the drum. The characteristics of the load exists in all the four quadrants
and is capable of delivering as well as absorbing mechanical power. Hence it is called as an
active load. The governing equation when the motor and a load are connected together is
dw
TM (w) − TL (w) = J (52)
dt
119
Speed
II I
Torque
III IV
(a)
b a Speed
Torque
a b
(b)
120
W
T W T
speed
Hoisting an
empty cage Hoisting a
loaded cage
Torque
o
T
W
T W
Lowering a
Lowering an loaded cage
empty cage
121
where TM (w) and TM (w) are motor and load torques respectively. J is the polar moment of
dw
inertia of the motor and load put together at the motor shaft. dt
is made positive when the
speed has to be increased in the positive direction and negative when reducing the speed.
Under steady operation TM (w) − TL (W ) = 0. Both motor and load torques are expressed as
functions of the speed. The speed at which motor and load torques are equal and opposite is
the steady state operating speed. By varying the characteristics of the motor (or the load),
this speed can be changed to suit our requirements. Normally the torque speed characteris-
tics of a load cannot be changed easily. Thus most speed control methods adopt, varying the
motor characteristics to achieve speed control. Some typical loads and the motors commonly
used to drive the same are tabulated in Table.
d.c. shunt motor lathes,fans,pumps disc and band saw drive requiring moderate torques.
d.c. series motor Electric traction, high speed tools
d.c. compound motor Rolling mills and other loads requiring large momentary toques.
122
A d.c. machine has to be tested for proper fabrication and trouble free operation.
From the tests one can determine the external characteristics needed for application of these
machines. Also, one can find the efficiency, rating and temperature rise of the machine.
Some of the tests are discussed in sequence now.
Even though any value of applied voltage can be used, the highest permissible
voltage/current is chosen during the test to minimize the errors. The armature circuit
consists of two resistances in series. They are armature winding resistance and resistance
due to the brushes and the brush drop. The brush contact drop behaves like a non-linear
resistance. To separate this from the armature circuit resistance and brush resistance a
number of v-i readings are taken. An equation of V = Vb + IRa form is fitted through
these test points shown graphically in Fig. 55. For large values of I the equivalent armature
resistance is taken to be V /I ohm. If the value of brush drop Vb can be neglected then the
armature resistance Ra = V /I ohm.
The OCC is of great value as it shows the mmf and hence the field current required
to generate a given voltage at any speed, on no load. It is a graph showing the variation
123
v
+
A
A2
DC Supply
A1 -
(a)Physical connection
I.V. Characteristic
dv
Ra =
di
dv
V di
Vb
0 I
(b)Characteristics
124
of the induced emf as a function of excitation current, when the speed is held constant,
with the load current being zero. It is also called the no-load saturation curve or no load
magnetization characteristic. This is experimentally determined by running the machine
as a separately excited generator on no-load at a constant speed and noting the terminal
voltage as a function of the excitation current. This curve can be used to find the OCC at
other speeds and also the self excited voltage when the machine works as a shunt generator.
In the case of short circuit test the armature is kept short circuited through an
ammeter. The machine is demagnetized and an extremely small field current is passed
through the field. The variation of the short circuit current as a function of excitation
current is plotted as the SCC. The speed is to be held constant during this test also. The
short circuit test gives an idea of the armature drop at any load current.
To assess the rating of a machine a load test has to be conducted. When the
machine is loaded, certain fraction of the input is lost inside the machine and appears as
heat, increasing the temperature of the machine. If the temperature rise is excessive then
it affects the insulations, ultimately leading to the breakdown of the insulation and the
machine. The load test gives the information about the efficiency of a given machine at any
load condition. Also, it gives the temperature rise of the machine. If the temperature rise
is below the permissible value for the insulation then the machine can be safely operated
at that load, else the load has to be reduced. The maximum continuous load that can be
125
delivered by the machine without exceeding the temperature rise for the insulation used, is
termed as the continuous rating of the machine. Thus the load test alone can give us the
proper information of the rating and also can help in the direct measurement of the efficiency.
The moment of inertia value is very important for the selection of a proper motor
for drives involving many starts and stops or requiring very good speed control characteris-
tics. The inertia can be determined by a retardation test.
The test works on the principle that when a motor is switched off from the
mains it decelerates and comes to rest. The angular retardation at any speed is proportional
to the retarding torque and is inversely proportional to the inertia. The torque lost at
any speed is calculated by running the motor at that speed steadily on no load and noting
the power input.From this power the losses that takes place in the armature and field are
deducted to get the power converted into mechanical form. All this power is spent in over
coming the mechanical losses at that speed. This can be repeated at any defined speed to
get the lost power (PL ) and torque lost (Tlost ) due to mechanical losses. In a retardation
test the motor speed is taken to some high value and the power to the motor is switched off.
The torque required by the losses is supplied by the energy stored in the motor inertia. The
lost torque at any speed can be written as
PL = Tlost .ω (53)
dw
Tlost = PL /w = J
dt
dw
Here the dt
is the slope of the retardation curve and the (Tlost ) is the torque required to be
126
met at the given speed. From these values the moment of inertia can be computed as
Tlost PL
J= dw
= kgm2 (54)
dt
w. dw
dt
A machine when loaded yields an output. The input to the machine is measured
at that operating point. The the efficiency in per unit is given as the ratio of output power
to input power.
output power
η = (55)
input power
Input power − power lost inside the machine
=
input power
output power
=
output power + power lost inside the machine
The first definition is used in the direct estimation of the efficiency . The other
two definitions are known as determination of efficiency using the loss segregation. For the
segregation of losses one must know the losses that take place inside a d.c. machine. The
losses that take place inside a d.c. machine can be listed as below.
127
7. Mechanical losses
8. Iron losses
Out of these items 1,2,7,8 and 9 will be present in all the d.c. machines. Out of
the remaining one or more may be present depending on which winding is present. These
losses change with temperature of operation. Mechanical losses vary with variation in speed.
Iron losses change with the degree of saturation and distortion of the shape of the field flux
distribution under the poles.
When a d.c. machine is loaded using a suitable load the output delivered by
the machine increases. The input requirement also increases along with the output. The
difference between the input and output powers is the power lost inside the machine as loss.
The efficiency of power conversion is given by the ratio of output power to input power.
Putting in mathematical form for a motor,
V I − losses
η= (56)
VI
for constant speed operation, the speed dependant losses remain constant. The load depen-
dant losses form the variable losses. While the loss that takes place in the brush drop in the
brushes is proportional to the load current, the loss that takes place in the resistance of the
armature is proportional to the square of the load current. Even though the loss that takes
place in a field winding is proportional to the square of the current through that winding, it
is classified under constant losses as the excitation current is held constant during loading.
128
Thus the total losses in a d.c. motor can be expressed in the form
PL = a + bI + cI 2 (57)
V I − PL A
η= = 1 − ( + B + CI) (58)
VI I
a b
When A = V
,B = V
and C = cV .
The term inside the brackets is sometimes referred to as the deficiency. For a
typical d.c.motor these are plotted in Fig. 56(a) as a function of the load current. The
curves a,b,c in the figure represent the efficiency curve taking one component of the loss
at a time. The curve d is the efficiency curve with all three components taken together.
The resultant curve exhibits a maximum. This can be easily seen from the graph that
this maximum occurs when constant losses equal the variable losses. AI = CI or A = CI 2 .
Fig. 56(b) depicts a typical output vs η curve of a d.c.machine.
129
b
c
a
Efficiency
d
current
Output
(b)Output Vs Efficiency
130