CLASSROOM

Classroom

In this section of Resonance, we invite readers to pose questions likely to be raised in a

classroom situation. We may suggest strategies for dealing with them, or invite responses,
or both. “Classroom” is equally a forum for raising broader issues and sharing personal
experiences and viewpoints on matters related to teaching and learning science.

Harish Ravi1 and

Understanding DC Motors through Experiments Rajesh B Khaparde2

1Department of Electrical
We present here a set of simple experiments thro- Engineering, IIT Madras
ugh which various characteristics of a DC mo- Chennai 600036, India.
tor can be investigated. The experiments also Email:
[email protected]
provide insights on how to control a DC mo-
tor. We ¯rst study the back-emf versus angular 2
Homi Bhabha Centre for Sci-
speed and the torque versus current character- ence Education
istics. We then extract the moment of inertia TIFR VN Purav Marg, Mankhurd
of the rotor and the speed{torque curve for the Mumbai 400088, India.
Email:
load from the start-stop dynamics of the motor.
[email protected]
Finally, we study an alternate way of obtaining
the angular speed from the back-emf waveform.
1. Introduction
A three-pole DC motor is commonly used in many sim-
ple machines to convert electrical energy into mechanical
energy. Such a motor consists of a stator and a rotor.
The stator comprises of permanent magnets which pro-
vide a steady magnetic ¯eld and a pair of brushes as
Keywords
shown in Figure 1. The brushes are marked with the DC motor, motor constant, mo-
signs { and +, corresponding to the terminals of the tor speed control, speed–torque
power supply to which they are connected. The rotor curve, start-stop dynamics.
comprises of three coils (wound around three poles

RESONANCE  June 2010 561

 CLASSROOM

Figure 1. Schematic dia-

gram to illustrate the con-
struction of a three-pole DC
motor.

labeled 1, 2 and 3), a commutator ring having three

segments labeled A, B and C and an axle.
At the instant shown in Figure 1, current passes through
the coils 1 and 3 to produce a clockwise torque due to
repulsion of coil 1 and attraction of coil 3 by the right
(N side) magnet. Coil 2 is short-circuited by the left-
hand brush. Thus the rotor rotates clockwise and a
little while later, the current passes through all the three
coils and the rotor continues to rotate clockwise. The
interconnection of the coils to the commutator segments
is in such a way that the torque produced by the rotor
is always in the same direction [1].
2. Theoretical Analysis
2.1 Basic Equations for a DC Motor
When the rotor rotates in the presence of a steady mag-
netic ¯eld, a back-emf Eb is generated across the rotor
coils in accordance with Faraday's law and is expected to
be linearly proportional to the angular speed ! (hence-
forth called speed) of the rotor. Thus, we can write

Eb = k1 ! ; (1)
Figure 2. Equivalent circuit
where k1 is a constant.
representation of a DC
motor. A motor can be represented as a source of emf Eb con-
nected in series with its resistance R as shown in Figure
2.
Applying Kirchho®'s voltage law, we obtain
V ¡ Eb
I= ; (2)
R

562 RESONANCE  June 2010

 CLASSROOM

where I is the current passing through the motor and

V is the voltage applied across the motor. Lorentz force
law dictates that the motor torque ¿m is expected to be
linearly proportional to the current I passing through
the motor. That is,

¿m = k2I ; (3)

where k2 is a constant. The conservation of power gives

k1 = k2 = km ; where km is called the motor constant.
2.2 Speed{Torque Curve
A speed{torque curve (henceforth called s{t curve) is a
graph showing variation of torque with speed. The s{t
curve for a DC motor should be a straight line. We can
obtain the relation between motor torque ¿m and the
speed ! using (1), (2) and (3):

km V k2
¿m = ¡ m! : (4)
R R
A torque that opposes the motor torque is often called
the load torque ¿load . It is usually frictional and can be
considered as a static torque ¿s plus a viscous component
b!.
¿load = ¿s + b! : (5)
Thus, the motor torque decreases with increasing speed
while the load torque increases with increasing speed.
The motor operates at the point of intersection of the
s{t curves for the load and the motor as illustrated later
in Figure 3.
The motor torque
2.3 Start{Stop Dynamics
decreases with
Let us understand the dynamics of starting of a motor. increasing speed
If J indicates the moment of inertia of the rotor, then while the load
the law of rotation gives, torque increases
with increasing
¿m ¡ ¿load = J® ; (6) speed.

RESONANCE  June 2010 563

 CLASSROOM

Figure 3. s–t curves for the

motor and the load.

where ® is the angular acceleration (henceforth called

acceleration) of the rotor. Substituting for the torques
from (4) and (5), and assuming a constant voltage source,
we can write,

km (V ¡ km !) d!
¡ ¿s ¡ b! = J : (7)
R dt
After integrating (7), we obtain an exponentially in-
creasing function for speed !. Thus, the back-emf in-
creases with time and from (2), it follows that the cur-
rent should exponentially decrease. In a hypothetical
situation, when a motor has zero load and is driven by
a constant voltage power supply, the rotor would speed
up to the point where the back-emf equals the voltage
applied across the motor and hence the current would
In a hypothetical fall to zero.
situation, when a Now, let us understand the dynamics of stopping of a
motor has zero load motor. When the power supply is disconnected, only
and is driven by a the load torque ¿load is responsible for slowing down the
constant voltage rotor and hence we can write,
power supply, the
rotor would speed up d!
¿load = ¿s + b! = ¡J : (8)
to the point where the dt
back-emf equals the After integrating (8), we note that the speed ! should
voltage applied fall exponentially with time. One can stop the rotor
across the motor and quicker by shorting the motor terminals and thereby
hence the current allowing current to pass through the rotor coils which
would fall to zero. produces an opposing torque in accordance with Lenz's
law.

564 RESONANCE  June 2010

 CLASSROOM

Speed  (rpm)

Current I (A)
Time t  (10–3 s) Time t  (10–3 s)

To illustrate the above, we simulated the s{t curve and Figure 4. Graphs of speed
the start-stop dynamics obtained by solving the above versus time and current
equations. We used km = 9.65£10¡3 Vs, R = 2  , J = versus time for starting.
10¡7 kgm2 , b = 10¡5 Nms and ¿ s = 0.005 Nm.
Here, note that the steady-state speed in the loaded case
(Figure 4) is close to that obtained at the point of inter-
section of s{t curves for the load and the motor (Figure
3). Also note that the no load steady-state speed (Fig-
ure 4) matches with that obtained from the s{t curve
(Figure 3) for the motor at the point of intersection with
the X-axis. The time constants for the speed versus time
curve and the current versus time curve obtained from
the formulae were both equal to 2.1 ms.
2.4 Analysis of Back-Emf Waveform
Though the back-emf
Though the back-emf waveform should be an average waveform should be
DC, it has a small AC component due to the commu- an average DC, it has
tator brush arrangement and also due to the changing a small AC
angle of the rotor coils with respect to the magnetic
component due to the
¯eld. There would be sharp dips due to loss of contact
commutator brush
at the gaps of the commutator or due to shorting of
arrangement and also
one of the coils. If there are p poles, there would be p
due to the changing
gaps between the commutator segments as indicated
angle of the rotor coils
in Figure 1. Each gap would pass over each brush once
with respect to the
in a rotation. As there are two brushes, there will be
two dips in the voltage waveform due to each gap and magnetic field.

RESONANCE  June 2010 565

 CLASSROOM

The frequency of
hence, the frequency f of the dips should be,
dips would reveal
f = 2p! : (9)
the number of
poles in the motor Thus, the frequency of dips would reveal the number of
and could also be poles in the motor and could also be used to ¯nd the
used to find the speed of the rotor.
speed of the
rotor.
3. Experimental Study
In this investigation, we used two similar DC motors,
one was named as the test motor and the other was
called the drive motor. A variable DC power supply
was used to provide the necessary power to the motors
and a non-contact type digital tachometer was employed
to measure the speed of the rotor.
3.1 Back-Emf{Speed Relation
The test motor was clamped to a retort stand and run
at di®erent speeds by varying the voltage of the power
supply. For various voltages, the speed was measured
using the tachometer. The resistance R of the motor
was initially measured using a multimeter. The back-
emf was calculated from measured values of V and I
using (3). The back-emf versus speed was then plotted
and the value of motor constant km was determined.
3.2 Torque{Current Relation
A motor coupled to a pulley with a known radius was
clamped to a retort stand. A set of slotted masses
was attached to a string wound around the pulley and
was held suspended as shown in Figure 5. The current
passing through the motor (which caused an opposing
torque) was increased until the mass just moved down
upon releasing. Thus the balancing point was deter-
mined and the motor torque was counter-balanced by
the torque due to the mass. The current versus mass
curve was plotted to understand the current{torque re-
lationship.

566 RESONANCE  June 2010

 CLASSROOM

Figure 5. The arrangement

of mass hung from the pul-
ley.

3.3 Start-Stop Dynamics

Since the ¯rst experiment established the relationship
between back-emf and angular speed, we exploited it to
measure the transient rotational response of the motor.
The back-emf was calculated from the current which was
measured using a small series resistance (Figure 6). The
The motor was
motor was powered using a constant voltage source and
powered using a
the transient voltage curve across the series resistance
constant voltage
was monitored on a storage oscilloscope. This curve
gave information about the start dynamics as it revealed source and the
the speed at every time instant. The stop dynamics transient voltage
however, was obtained more easily as the back-emf was curve across the
directly available across the motor terminals while stop- series resistance
ping. The speed{time curves obtained can then be dif- was monitored on a
ferentiated to obtain instantaneous accelerations. A RC storage oscilloscope.
¯ltering circuit was used to ¯lter out the spikes in the
voltage waveform due to commutations. These commu-
tations were actually used in the next experiment to
measure speed. One must note that the moment of in-
ertia of the rotor can be obtained only by studying the
dynamics.

Figure 6. Circuit diagram

for studying the stopping
back-emf curve.

RESONANCE  June 2010 567

 CLASSROOM

The moment of inertia 3.3.1 Stop Dynamics: Suppose a running motor is

of the rotor can be disconnected from a power supply at t = 0. Just before
calculated from the disconnecting the power supply (i.e., t = 0¡ ), the rotor
initial acceleration and was in steady state, which means the motor torque ¿m
the steady-state was balancing the load torque ¿load . Thus, we can write,
current before
km It=0¡ = ¿load : (10)
disconnecting the
power supply. At an instant immediately after disconnecting the power
supply (i.e., t = 0+ ), the speed would not have changed
and thus, ¿load which was responsible for the initial de-
celeration also would not have changed. Thus, we can
write,
¿load = ¡J®t=0+ ; (11)
km It=0¡ = ¡J®t=0+ : (12)
Hence, the moment of inertia J of the rotor can be cal-
culated from the initial acceleration (actually decelera-
tion) and the steady-state current before disconnecting
the power supply using (12). The load torque at each
instant can then be calculated from the acceleration and
then, the s{t curve can be plotted for the load.
3.3.2 Start Dynamics: The start dynamics was ob-
tained by monitoring the voltage Vsr across a small known
series resistance. At any time instant while starting,

km I ¡ ¿load = J® (13)

Using the moment of inertia found in the previous ex-

Figure 7. Photograph of the periment, the load torque ¿load can be calculated at each
coupled motors to study instant from the acceleration and current using (13) and
back-emf waveform. then, the s{t curve can be plotted for the load.
3.4 Analysis of Back-Emf Waveform
A drive motor and a test motor were coupled using a
belt and two identical pulleys as shown in Figure 7. The
drive motor was connected to a variable DC power sup-
ply and was used to control the speed of the system. The

568 RESONANCE  June 2010

 CLASSROOM

back-emf waveform of the test motor was monitored on We studied the

an oscilloscope for di®erent speeds which were measured characteristics of four
using a non-contact type tachometer. The frequency of
DC motors and
the back-emf waveform was plotted against the mea-
obtained consistent
sured speed.
results which
4. Results and Discussion complemented the
theoretical predictions.
We studied the characteristics of four DC motors (a 7-
pole, a 5-pole and two 3-pole motors) and obtained con-
sistent results which complemented the theoretical pre-
dictions. We present below our observations and results
recorded for a 5-pole DC motor.
4.1 Back-Emf{Speed Relation
We obtained a straight line passing through the origin
for the back-emf versus speed graph as shown in Fig-
ure 8, whose slope gave km = 2:32 £ 10¡3 V=rpm =
2:22 £ 10¡2 Vs, and thus veri¯ed that the back-emf is
indeed linearly proportional to the speed as expected
from Faraday's law.
4.2 Torque{Current Relation
We obtained a straight line for the current{mass graph
which experimentally con¯rmed that the torque pro-
duced by the rotor is linearly proportional to the current
Figure 8.(left) A graph of
passing through the motor as expected from Lorentz law back-emf versus speed.
Figure 9. (right) A graph of
¿m = km I = ¡¿s + mgr : (14) current versus mass.
Back-emf Eb(V)

Current I (A)

Speed  (rpm) Mass m (g)

RESONANCE  June 2010 569

 CLASSROOM

The stopping back- We obtained km = 2.26£10¡2 Nm/A and the static fric-
emf waveform tional torque ¿ s = 3.42£10¡3 Nm. We wish to draw
looks linear near attention to the fact that the value of the motor con-
the end where the stant km obtained here is consistent with that obtained
static frictional from the back-emf{speed graph. Note that we would
torque dominates
have obtained a positive intercept in the current{mass
and causes a
graph if we had measured the current required for the
masses to just lift o® the surface.
uniform
deceleration. 4.3 Start{Stop Dynamics
This experiment was performed with a di®erent DC mo-
tor. We obtained the exponential graphs for stopping
and starting as shown in Figures 10 and 11, respectively.
Note that the stopping back-emf waveform looks linear
near the end where the static frictional torque dominates
and causes a uniform deceleration.
The moment of inertia J of the rotor was calculated to
be 1.3£10¡4 kgm2 from the initial deceleration and the
steady-state current before disconnecting the power sup-
ply. At each instant, the speed ! was calculated from the
back-emf and the acceleration ® was determined from
Figure 10.(left) Back-emf
the slope of the back-emf versus time graph. Also, the
versus time for stopping.
Figure 11. (right) Voltage
load torque ¿load was calculated from the acceleration
Vsr versus time for start- as described earlier and the s{t curve for the load was
ing. plotted for stopping and starting as shown in Figure 12.
Back-emf Eb(V)

Vsr (10–3 V)

Time t (s) Time t (s)

570 RESONANCE  June 2010

 CLASSROOM

(a) (b)

Torque  (10–3 Nm)

Torque  (10–3 Nm)

Speed  (rps) Speed  (rps)

Note that there is a considerable error in the value of the Figure 12. s–t curves for
torque calculated due to inaccuracies in the determina- the load obtained from (a)
tion of slope, motor constant and moment of inertia. stopping and (b) starting.
One may obtain more reliable results using a computer-
interfaced data-acquisition system.
4.4 Analysis of Back-Emf Waveform
We observed a back-emf waveform as shown in Figure
13. The amplitude of the waveform was much smaller The adjacent dips
than the average DC which was blocked. The waveform were due to two
had a periodic alternate repeating pattern. The adjacent different brushes
dips were due to two di®erent brushes while the alternate while the alternate
dips were due to the same brush. The adjacent dips look dips were due to
di®erent due to a slight asymmetry in the brushes. the same brush.
The adjacent dips
The graph of frequency of the back-emf waveform versus
speed of the rotor turned out to be a straight line as look different due
shown in Figure 14. The slope of the straight line gave to a slight
2p = 10.07, which was very close to the expected value asymmetry in the
of 10, as the motor had ¯ve poles. brushes.

Figure 13. Photograph of

back-emf waveform ob-
served on a storage oscil-
loscope.

RESONANCE  June 2010 571

 CLASSROOM

Figure 14. A graph of fre-

quency versus speed.

Frequency f (Hz)
Speed  (rps)

5. Conclusions
We have demonstrated a set of simple experiments thro-
ugh which the fundamental characteristics, like back-
emf{speed, and torque{current relationships of a DC
motor were experimentally con¯rmed. We then used
the back-emf{speed relation to examine the transient
rotational response of the rotor while starting and stop-
ping. From this, we obtained the moment of inertia of
the rotor and the load s{t curve. Finally, we studied an
alternate method of measuring the angular speed of the
rotor from the commutation dips. We suggest that this
set of simple experiments may be tried out in an under-
graduate teaching laboratory for physics and electrical
engineering students.
Suggested Reading
Acknowledgements
[1] http://www.stefanv.com/
rcstuff/qf200212.html
We are grateful to Profs. Arvind Kumar, Vijay Singh
[2] Exp Test 2-B, Indian Na- and H C Pradhan for providing the necessary support
tional Physics Olympiad, and facilities for this study under the NIUS programme.
OCSC, HBCSE, 2003. We are also extremely thankful to Profs. M L Ogala-
[3] Fitzgerald, Electric Ma-
purkar and D A Desai for the initial development, sup-
chinery, McGraw Hill,
2003.
port and for motivating us to take up the study. We ap-
[4] R L Weber, K V Manning preciate the help and support received from the project
et al, College Physics, 5th sta® of the physics laboratories at HBCSE.
edition, Tata McGraw
Hill, 1995.

572 RESONANCE  June 2010