EE443

ENERGY CONVERSION II
Lecture 2: DC Machines
 DC Machines
• A DC machine has two windings.

1. Field winding is on the stator and produces a field flux.

2. Armature winding is on the rotor. The armature current interacts with the stator field flux to
produce torque on the rotor shaft.
DC Machine Model
1. Electrical Model

𝑑𝑑
𝑣𝑣𝑎𝑎 = 𝑅𝑅𝑎𝑎 𝑖𝑖𝑎𝑎 + 𝐿𝐿𝑎𝑎 𝑖𝑖 + 𝑒𝑒𝑎𝑎
𝑑𝑑𝑑𝑑 𝑎𝑎

𝑑𝑑
𝑣𝑣𝑓𝑓 = 𝑅𝑅𝑓𝑓 𝑖𝑖𝑓𝑓 + 𝐿𝐿𝑓𝑓 𝑖𝑖
𝑑𝑑𝑑𝑑 𝑓𝑓
2. EMF equation
𝑒𝑒𝑎𝑎 = 𝐾𝐾𝑒𝑒 𝜔𝜔𝑚𝑚
3. Torque Equation
𝑇𝑇𝑒𝑒 = 𝐾𝐾𝑇𝑇 𝑖𝑖𝑎𝑎
4. Mechanical Model
𝑑𝑑𝜔𝜔𝑚𝑚
𝐽𝐽 + 𝐵𝐵 𝜔𝜔𝑚𝑚 = 𝑇𝑇𝑒𝑒 − 𝑇𝑇𝐿𝐿
𝑑𝑑𝑑𝑑
Dynamic control of DC machine
• We covered control at steady state only:
1. Armature voltage control

2. Armature resistance control

3. Field Flux control

• To control the transient response, we need a complete model of the machine.

1. Electric Model
In the time domain, 𝑻𝑻𝒆𝒆

𝑑𝑑
𝑣𝑣𝑎𝑎 = 𝑅𝑅𝑎𝑎 𝑖𝑖𝑎𝑎 + 𝐿𝐿𝑎𝑎 𝑖𝑖𝑎𝑎 + 𝑒𝑒𝑎𝑎
𝑑𝑑𝑑𝑑

In the s-domain, using Laplace Transformer,

𝑉𝑉𝑎𝑎 = 𝑅𝑅𝑎𝑎 𝐼𝐼𝑎𝑎 + 𝑠𝑠 𝐿𝐿𝑎𝑎 𝐼𝐼𝑎𝑎 + 𝐸𝐸𝑎𝑎

Or

1
𝐼𝐼𝑎𝑎 = 𝑉𝑉𝑎𝑎 − 𝐸𝐸𝑎𝑎
𝑅𝑅𝑎𝑎 + 𝑠𝑠 𝐿𝐿𝑎𝑎
2) EMF
𝐸𝐸𝑎𝑎 = 𝐾𝐾𝐸𝐸 𝜔𝜔𝑚𝑚 = 𝑘𝑘𝐸𝐸 Φf 𝜔𝜔𝑚𝑚

𝜔𝜔𝑚𝑚 𝐸𝐸𝑎𝑎
𝐾𝐾𝐸𝐸
3) Torque Equation
𝑇𝑇𝑒𝑒 = 𝐾𝐾𝑇𝑇 𝐼𝐼𝑎𝑎 = 𝑘𝑘 𝑇𝑇 Φf 𝐼𝐼𝑎𝑎

𝐾𝐾𝑇𝑇
4) Mechanical Model
In the time domain,
𝑑𝑑𝜔𝜔𝑚𝑚
𝐽𝐽 + 𝐵𝐵𝜔𝜔𝑚𝑚 = 𝑇𝑇𝑒𝑒 − 𝑇𝑇𝐿𝐿
𝑑𝑑𝑑𝑑
In the s-domain, using Laplace Transformer,
𝐽𝐽𝑠𝑠𝜔𝜔𝑚𝑚 + 𝐵𝐵𝜔𝜔𝑚𝑚 = 𝑇𝑇𝑒𝑒 − 𝑇𝑇𝐿𝐿

Or
1
𝜔𝜔𝑚𝑚 = 𝑇𝑇 − 𝑇𝑇𝐿𝐿
𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝑒𝑒
The Complete Model

𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑
𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖

𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
𝐾𝐾𝑇𝑇

𝐾𝐾𝐸𝐸
Transfer function
1
𝜔𝜔𝑚𝑚 = 𝑇𝑇 − 𝑇𝑇𝐿𝐿
𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝑒𝑒

𝐾𝐾𝑇𝑇 1
= 𝐼𝐼𝑎𝑎 − 𝑇𝑇
𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿

𝐾𝐾𝑇𝑇 1 1
= 𝑉𝑉𝑎𝑎 − 𝐸𝐸𝑎𝑎 − 𝑇𝑇
𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿

𝐾𝐾𝑇𝑇 1
= 𝑉𝑉𝑎𝑎 − 𝐾𝐾𝐸𝐸 𝜔𝜔𝑚𝑚 − 𝑇𝑇
𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿

𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝐾𝐾𝑇𝑇 1

1+ 𝜔𝜔𝑚𝑚 = 𝑉𝑉𝑎𝑎 − 𝑇𝑇
𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿

𝐽𝐽𝐽𝐽 + 𝐵𝐵 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 + 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝜔𝜔𝑚𝑚 = 𝐾𝐾𝑇𝑇 𝑉𝑉𝑎𝑎 − 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚𝑚

𝐾𝐾𝑇𝑇 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎

𝜔𝜔𝑚𝑚 = 𝑉𝑉
𝑎𝑎 − 𝑇𝑇
𝐽𝐽𝐿𝐿𝑎𝑎 𝑠𝑠 2 + 𝐽𝐽𝑅𝑅𝑎𝑎 + 𝐵𝐵𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐵𝐵 + 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝐽𝐽𝐿𝐿𝑎𝑎 𝑠𝑠 2 + 𝐽𝐽𝑅𝑅𝑎𝑎 + 𝐵𝐵𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐵𝐵 + 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝐿𝐿

•Finally,

𝐾𝐾𝑇𝑇 𝑅𝑅
1 𝑠𝑠 + 𝐿𝐿 𝑎𝑎
𝐽𝐽𝐿𝐿𝑎𝑎 𝑎𝑎
𝜔𝜔𝑚𝑚 = 𝑉𝑉𝑎𝑎 − 𝑇𝑇
𝑅𝑅 𝐵𝐵 𝑅𝑅𝑎𝑎 𝐵𝐵 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝑅𝑅
𝐽𝐽 𝑠𝑠 2 + 𝑎𝑎 + 𝐵𝐵 𝑅𝑅𝑎𝑎 𝐵𝐵 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝐿𝐿
𝑠𝑠 2 + 𝐿𝐿 𝑎𝑎 + 𝑠𝑠 + + 𝐽𝐽𝐿𝐿 𝑠𝑠 + +
𝑎𝑎 𝐽𝐽 𝐽𝐽𝐿𝐿𝑎𝑎 𝑎𝑎 𝐿𝐿𝑎𝑎 𝐽𝐽 𝐽𝐽𝐿𝐿𝑎𝑎 𝐽𝐽𝐿𝐿𝑎𝑎
Characteristics equation
2
𝑅𝑅𝑎𝑎 𝐵𝐵 𝑅𝑅𝑎𝑎 𝐵𝐵 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸
𝑠𝑠 + + 𝑠𝑠 + + =0
𝐿𝐿𝑎𝑎 𝐽𝐽 𝐽𝐽𝐿𝐿𝑎𝑎 𝐽𝐽𝐿𝐿𝑎𝑎

• Based on the poles of the transfer function, the response may be:
1) Over damped (poles are real distinct).

2) Critically damped (poles are real repeated).

3) underdamped (Complex poles)

4) Undamped (Imaginary poles)

Location of the poles
Special case B = 0
• The characteristics equation becomes:
2
𝑅𝑅𝑎𝑎 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸
𝑠𝑠 + 𝑠𝑠 + =0
𝐿𝐿𝑎𝑎 𝐽𝐽𝐿𝐿𝑎𝑎
• Define the electrical time constant,

𝐿𝐿𝑎𝑎
𝜏𝜏𝑎𝑎 =
𝑅𝑅𝑎𝑎
• And the mechanical time constant,

𝐽𝐽 𝑅𝑅𝑎𝑎
𝜏𝜏𝑚𝑚 =
𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸

• Then the characteristics equation becomes

1 1 1
𝑠𝑠 2 + 𝑠𝑠 + =0
𝜏𝜏𝑎𝑎 𝜏𝜏𝑎𝑎 𝜏𝜏𝑚𝑚
• And the poles are
1 1 1 𝜏𝜏𝑎𝑎
𝑠𝑠1,2 = − ± −
2 𝜏𝜏𝑎𝑎 𝜏𝜏𝑎𝑎 4 𝜏𝜏𝑚𝑚
Special case B = 0
If,

• 𝜏𝜏𝑚𝑚 > 4 𝜏𝜏𝑎𝑎 → 𝜁𝜁 > 1 → 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 → 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜

• 𝜏𝜏𝑚𝑚 = 4 𝜏𝜏𝑎𝑎 → 𝜁𝜁 = 1 → 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 → 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑

• 𝜏𝜏𝑚𝑚 < 4 𝜏𝜏𝑎𝑎 → 𝜁𝜁 < 1 → 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 → 𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢

Another representation
• Alternatively, we can define the natural frequency,

𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸
𝜔𝜔𝑛𝑛2 =
𝐽𝐽𝐿𝐿𝑎𝑎

• And the damping factor

𝑅𝑅𝑎𝑎 𝐽𝐽
𝜁𝜁 =
2 𝐿𝐿𝑎𝑎 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸

• Then, the characteristics equations is,

𝑠𝑠 2 + 2 𝜁𝜁 𝜔𝜔𝑛𝑛 𝑠𝑠 + 𝜔𝜔𝑛𝑛2 = 0
Example
Consider a separately excited DC motor with the following parameters,
• The armature resistance is 0.26 Ω,
• the armature inductance is 1.7 mH,
• the inertia of moment J of this motor is 0.00252 Kg m2,
• and the emf and torque constants are 0.4078 V/(rad/s).
a) Calculate the locations of the two poles and predict the transient response of the
speed to an armature voltage at no-load.
b) Predict the transient response when this motor is driving the load with the inertia
of moment of 6J.
Solution
a) The time constants, b) Jnew = J + 6 J = 0.0176 Kg m2

𝐿𝐿𝑎𝑎 1.7 × 10−3

𝜏𝜏𝑎𝑎 = = = 6.5 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝐽𝐽𝑛𝑛𝑛𝑛𝑛𝑛 𝑅𝑅𝑎𝑎 0.0176 × 0.26
𝑅𝑅𝑎𝑎 0.26 𝜏𝜏𝑚𝑚 = = = 27.6 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚
𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 0.4078 2
𝐽𝐽 𝑅𝑅𝑎𝑎 0.00252 × 0.26 • The poles,
𝜏𝜏𝑚𝑚 = = = 3.93 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚
𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 0.4078 2
• The poles,
1 1 1 𝜏𝜏𝑎𝑎
𝑠𝑠1,2 = − ± − = −59, −93.8
2 𝜏𝜏𝑎𝑎 𝜏𝜏𝑎𝑎 4 𝜏𝜏𝑚𝑚
1 1 1 𝜏𝜏𝑎𝑎
𝑠𝑠1,2 = − ± − = −76.47 ± 𝑗𝑗 181.57 • The damping ratio,
2 𝜏𝜏𝑎𝑎 𝜏𝜏𝑎𝑎 4 𝜏𝜏𝑚𝑚

• The damping ratio,

𝑅𝑅𝑎𝑎 𝐽𝐽𝑛𝑛𝑛𝑛𝑛𝑛
𝜁𝜁 = = 1.0269
2 𝐿𝐿𝑎𝑎 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸
𝑅𝑅𝑎𝑎 𝐽𝐽
𝜁𝜁 = = 0.3881 • Because ζ > 1, we can expect the system to have an over-damped
2 𝐿𝐿𝑎𝑎 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸
response without overshoot.
• Because ζ < 1, we can expect the system to have an under-
damped response and an overshoot.
Sketching the step response
• We can use MATLAB to sketch the step response. For the previous example,
Ra = 0.26;
La = 1.7e-3;
J = 0.00252;
K = 0.4078;

H = tf(K/(J*La),[1 Ra/La K*K/(J*La)]);

step(H)
Overshoot (%𝑂𝑂𝑆𝑆) and Settling time (Ts)
• Overdamped system • Underdamped system

𝜁𝜁𝜁𝜁
%𝑂𝑂𝑂𝑂 = 0 −
%𝑂𝑂𝑂𝑂 = 𝑒𝑒 1−𝜁𝜁 2 × 100%
4
𝑇𝑇𝑠𝑠 ≈ 4
𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑇𝑇𝑠𝑠 ≈
𝜁𝜁𝜔𝜔𝑛𝑛
The dominant pole is the pole closest to zero.
The Transfer Function
𝑇𝑇𝐿𝐿

𝐾𝐾𝑇𝑇 𝑅𝑅𝑎𝑎
1 𝑠𝑠 +
𝐽𝐽𝐿𝐿𝑎𝑎 𝐿𝐿𝑎𝑎
𝜔𝜔𝑚𝑚 = 𝑉𝑉𝑎𝑎 − 𝑇𝑇
𝑠𝑠 2 +
𝑅𝑅𝑎𝑎 𝐵𝐵
+
𝑅𝑅 𝐵𝐵 𝐾𝐾 𝐾𝐾
𝑠𝑠 + 𝑎𝑎 + 𝑇𝑇 𝐸𝐸 𝐽𝐽 𝑠𝑠 2 + 𝑅𝑅𝑎𝑎 + 𝐵𝐵 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐵𝐵 + 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝐿𝐿
𝐿𝐿𝑎𝑎 𝐽𝐽 𝐽𝐽𝐿𝐿𝑎𝑎 𝐽𝐽𝐿𝐿𝑎𝑎 𝐿𝐿𝑎𝑎 𝐽𝐽 𝐽𝐽𝐿𝐿𝑎𝑎 𝐽𝐽𝐿𝐿𝑎𝑎
Closed Loop Control
Current (Torque) Control
• In this stage we control the current only.

𝐾𝐾𝑝𝑝𝑝𝑝

𝐾𝐾𝑖𝑖𝑐𝑐
PI Controller
• The PI controller consists of a proportional term 𝐾𝐾𝑝𝑝𝑝𝑝 , which depends on the present error, and an integral term
𝐾𝐾𝑖𝑖𝑖𝑖 , which depends on the accumulation of past and present errors as follows,

𝐾𝐾𝑖𝑖𝑖𝑖 𝐾𝐾𝑝𝑝𝑝𝑝 𝑠𝑠 + 𝐾𝐾𝑖𝑖𝑖𝑖

𝐺𝐺𝑐𝑐 𝑠𝑠 = 𝐾𝐾𝑝𝑝𝑝𝑝 + =
𝑠𝑠 𝑠𝑠

• The PI Controller could also be represented as,

1
𝐺𝐺𝑐𝑐 𝑠𝑠 = 𝐾𝐾𝑝𝑝𝑝𝑝 (1 + )
𝑇𝑇𝑝𝑝𝑝𝑝 𝑠𝑠
• Where,
𝐾𝐾𝑝𝑝𝑝𝑝
𝑇𝑇𝑝𝑝𝑝𝑝 =
𝐾𝐾𝑖𝑖𝑖𝑖
Closed Loop Transfer function
• We can derive the closed loop transfer function as,

𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾𝑖𝑖𝑖𝑖
𝑠𝑠 +
𝐿𝐿𝑎𝑎 𝐿𝐿𝑎𝑎 𝑠𝑠/𝐿𝐿𝑎𝑎
𝐼𝐼𝑎𝑎 = 𝐼𝐼𝑎𝑎∗ − 𝐸𝐸𝑎𝑎
𝑅𝑅𝑎𝑎 + 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾𝑖𝑖𝑖𝑖 𝑅𝑅𝑎𝑎 + 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾𝑖𝑖𝑖𝑖
𝑠𝑠 2 + 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝐿𝐿𝑎𝑎 𝑠𝑠 2 + 𝐿𝐿𝑎𝑎 𝑠𝑠 + 𝐿𝐿𝑎𝑎
• Where the transfer function between the reference and actual current is,

𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾
𝑠𝑠 + 𝑖𝑖𝑖𝑖
𝐼𝐼𝑎𝑎 𝐿𝐿𝑎𝑎 𝐿𝐿𝑎𝑎
=
𝐼𝐼𝑎𝑎∗ 𝑅𝑅𝑎𝑎 + 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾𝑖𝑖𝑖𝑖
𝑠𝑠 2 + 𝑠𝑠 +
𝐿𝐿𝑎𝑎 𝐿𝐿𝑎𝑎

• By selecting the proportional gain 𝐾𝐾𝑝𝑝𝑝𝑝 and an integral gain 𝐾𝐾𝑖𝑖𝑖𝑖 , we can change current control characteristic.
Design of PI Controller
• Let the desired closed-loop transfer function be,
𝐼𝐼𝑎𝑎 𝑝𝑝1 𝑝𝑝2 𝑠𝑠 + 𝑧𝑧 𝑝𝑝1 𝑝𝑝2 𝑠𝑠 + 𝑧𝑧
∗ = = 2
𝐼𝐼𝑎𝑎 𝑧𝑧 𝑠𝑠 + 𝑝𝑝1 𝑠𝑠 + 𝑝𝑝2 𝑧𝑧 𝑠𝑠 + 𝑝𝑝1 + 𝑝𝑝2 𝑠𝑠 + 𝑝𝑝1 𝑝𝑝2
Then

𝑅𝑅𝑎𝑎 + 𝐾𝐾𝑝𝑝𝑝𝑝
𝑝𝑝1 + 𝑝𝑝2 =
𝐿𝐿𝑎𝑎

𝐾𝐾𝑖𝑖𝑖𝑖
𝑝𝑝1 𝑝𝑝2 =
𝐿𝐿𝑎𝑎
𝐾𝐾𝑖𝑖𝑖𝑖
𝑧𝑧 =
𝐾𝐾𝑝𝑝𝑝𝑝
• One way to design the system is pole-zero cancellation and choosing 𝑝𝑝1 as a constant,
𝑝𝑝1 = 𝜔𝜔𝑐𝑐𝑐𝑐 𝑎𝑎𝑎𝑎𝑎𝑎 𝑧𝑧 = 𝑝𝑝2
• Resulting in,

𝐼𝐼𝑎𝑎 𝜔𝜔𝑐𝑐𝑐𝑐
∗ = (𝜔𝜔𝑐𝑐𝑐𝑐 𝑖𝑖𝑖𝑖 𝑡𝑡𝑡𝑡𝑡 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏)
𝐼𝐼𝑎𝑎 𝑠𝑠 + 𝜔𝜔𝑐𝑐𝑐𝑐
• Then we can find,
𝐾𝐾𝑝𝑝𝑝𝑝 = 𝜔𝜔𝑐𝑐𝑐𝑐 𝐿𝐿𝑎𝑎 and 𝐾𝐾𝑖𝑖𝑖𝑖 = 𝜔𝜔𝑐𝑐𝑐𝑐 𝑅𝑅𝑎𝑎
Bandwidth
• The settling time is
4
𝑇𝑇𝑠𝑠 =
𝜔𝜔𝑐𝑐𝑐𝑐

𝜔𝜔𝑐𝑐𝑐𝑐 = 2𝜋𝜋𝜋𝜋𝜋 𝑟𝑟𝑟𝑟𝑟𝑟/𝑠𝑠 → 𝑇𝑇𝑠𝑠 = 1.3 𝑚𝑚𝑚𝑚 𝜔𝜔𝑐𝑐𝑐𝑐 = 2𝜋𝜋𝜋𝜋𝜋𝜋 𝑟𝑟𝑟𝑟𝑟𝑟/𝑠𝑠 → 𝑇𝑇𝑠𝑠 = 0.6 𝑚𝑚𝑚𝑚
Capability Curve
• Any machine has a rated voltage and rated current.

• The capability curve shows the range of torque and speed that can be attained without exceeding the rated voltage or
rated current.
𝑇𝑇𝑒𝑒 = 𝐾𝐾𝑇𝑇 𝑖𝑖𝑎𝑎 = 𝐾𝐾𝑇𝑇′ 𝑖𝑖𝑓𝑓 𝑖𝑖𝑎𝑎

• If we keep the field and armature current at the rated value, then we can achieve the rated torque at any speed up
to a limit called the rated speed.
𝑒𝑒𝑎𝑎 = 𝐾𝐾𝐸𝐸 𝜔𝜔𝑚𝑚 = 𝐾𝐾𝐸𝐸′ 𝑖𝑖𝑓𝑓 𝜔𝜔𝑚𝑚

• In order to increase the speed for a constant 𝐸𝐸𝑎𝑎 , we must reduce the field current 𝑖𝑖𝑓𝑓 .

𝑒𝑒𝑎𝑎 𝑒𝑒𝑎𝑎 𝑖𝑖𝑎𝑎 𝑃𝑃𝑚𝑚

𝑇𝑇𝑒𝑒 = 𝐾𝐾𝑇𝑇′ 𝑖𝑖𝑓𝑓 𝑖𝑖𝑎𝑎 = 𝐾𝐾𝑇𝑇′ 𝑖𝑖𝑎𝑎 = =
𝐾𝐾𝐸𝐸′ 𝜔𝜔𝑚𝑚 𝜔𝜔𝑚𝑚 𝜔𝜔𝑚𝑚
Capability Curve
Feedforward Control
Anti-Windup controller
• The output of the PI controller must be limited to the rated voltage. The limits are +Vrated and -Vrated.

• The gain of an anti-windup controller is normally selected as,

1
𝐾𝐾𝑎𝑎 =
𝐾𝐾𝑝𝑝
 Speed Control
• The speed controller is be placed in the outer loop of the current controller (Cascaded Control)

• If the bandwidth of the speed loop is much lower than the current loop bandwidth 𝜔𝜔𝑐𝑐𝑐𝑐 ,
𝑖𝑖𝑎𝑎 𝜔𝜔𝑐𝑐𝑐𝑐
∗ = ≈1
𝑖𝑖𝑎𝑎 𝑠𝑠 + 𝜔𝜔𝑐𝑐𝑐𝑐
• And the transfer function could be derived as,

𝐾𝐾𝑇𝑇 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾 𝐾𝐾
𝑠𝑠 + 𝑇𝑇 𝑖𝑖𝑖𝑖
𝐽𝐽 𝐽𝐽 𝑠𝑠
∗
𝜔𝜔𝑚𝑚 = 𝜔𝜔𝑚𝑚 − 𝑇𝑇𝐿𝐿
𝐾𝐾𝑇𝑇 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾 𝐾𝐾 𝐾𝐾𝑇𝑇 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾 𝐾𝐾
𝑠𝑠 2 + 𝑠𝑠 + 𝑇𝑇 𝑖𝑖𝑖𝑖 𝑠𝑠 2 + 𝑠𝑠 + 𝑇𝑇 𝑖𝑖𝑖𝑖
𝐽𝐽 𝐽𝐽 𝐽𝐽 𝐽𝐽
Speed Control
𝐾𝐾
• Clearly, we have a zero at 𝑧𝑧 = 𝐾𝐾 𝑖𝑖𝑖𝑖
𝑝𝑝𝑝𝑝

• Let,

𝐾𝐾𝑇𝑇 𝐾𝐾𝑝𝑝𝑝𝑝
= 𝜔𝜔𝑐𝑐𝑠𝑠
𝐽𝐽
• Then

𝑠𝑠
𝜔𝜔𝑐𝑐𝑐𝑐 (𝑠𝑠 + 𝑧𝑧) ∗ 𝐽𝐽
𝜔𝜔𝑚𝑚 = 2 𝜔𝜔𝑚𝑚 − 2 𝑇𝑇
𝑠𝑠 + 𝜔𝜔𝑐𝑐𝑐𝑐 𝑠𝑠 + 𝜔𝜔𝑐𝑐𝑐𝑐 𝑧𝑧 𝑠𝑠 + 𝜔𝜔𝑐𝑐𝑐𝑐 𝑠𝑠 + 𝜔𝜔𝑐𝑐𝑐𝑐 𝑧𝑧 𝐿𝐿
• 𝛼𝛼𝑠𝑠 which determines the dynamic performance of the speed controller and is limited by the bandwidth of the current controller. And is selected as five to ten times less
than that of current loop bandwidth,
𝜔𝜔𝑐𝑐𝑐𝑐
𝜔𝜔𝑐𝑐𝑐𝑐 =
10
• The damping ratio could be derived as,

1 𝜔𝜔𝑐𝑐𝑐𝑐
𝜁𝜁 =
2 𝑧𝑧
• If z is selected as one-fifth of 𝜔𝜔𝑐𝑐𝑐𝑐 , then 𝜁𝜁 = 1.118. This design will result in,

𝐽𝐽𝜔𝜔𝑐𝑐𝑐𝑐
𝐾𝐾𝑝𝑝𝑝𝑝 =
𝐾𝐾𝑇𝑇

2
𝐽𝐽𝜔𝜔𝑐𝑐𝑐𝑐
𝐾𝐾𝑖𝑖𝑖𝑖 =
5 𝐾𝐾𝑇𝑇
2DOF Speed PI controller
Position control
• In many applications we need to control the position of the motor such as in robotics and tracking applications.
Position of the motor could be controlled using a third control loop with a bandwidth selected as five to ten times
less than that of speed loop bandwidth,
𝜔𝜔𝑐𝑐𝑠𝑠
𝜔𝜔𝑐𝑐𝑝𝑝 =
10
• since the bandwidth of the position loop is much lower than the current and speed loops,
𝜔𝜔𝑚𝑚
∗ ≈1
𝜔𝜔𝑚𝑚

𝐾𝐾𝑝𝑝𝑝𝑝
Summary
Model of DC Machines
𝑇𝑇𝐿𝐿 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑
𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜

𝐾𝐾𝑇𝑇 𝑅𝑅𝑎𝑎
1 𝑠𝑠 +
𝐽𝐽𝐿𝐿𝑎𝑎 𝐿𝐿𝑎𝑎
𝜔𝜔𝑚𝑚 = 𝑉𝑉𝑎𝑎 − 𝑇𝑇
𝑠𝑠 2 +
𝑅𝑅𝑎𝑎 𝐵𝐵
+
𝑅𝑅 𝐵𝐵 𝐾𝐾 𝐾𝐾
𝑠𝑠 + 𝑎𝑎 + 𝑇𝑇 𝐸𝐸 𝐽𝐽 𝑠𝑠 2 + 𝑅𝑅𝑎𝑎 + 𝐵𝐵 𝑠𝑠 + 𝑅𝑅𝑎𝑎 𝐵𝐵 + 𝐾𝐾𝑇𝑇 𝐾𝐾𝐸𝐸 𝐿𝐿
𝐿𝐿𝑎𝑎 𝐽𝐽 𝐽𝐽𝐿𝐿𝑎𝑎 𝐽𝐽𝐿𝐿𝑎𝑎 𝐿𝐿𝑎𝑎 𝐽𝐽 𝐽𝐽𝐿𝐿𝑎𝑎 𝐽𝐽𝐿𝐿𝑎𝑎
Current (Torque) Control
• In this stage we control the current only.

𝐾𝐾𝑝𝑝𝑝𝑝
𝐾𝐾𝑝𝑝𝑝𝑝 = 𝜔𝜔𝑐𝑐𝑐𝑐 𝐿𝐿𝑎𝑎 and 𝐾𝐾𝑖𝑖𝑖𝑖 = 𝜔𝜔𝑐𝑐𝑐𝑐 𝑅𝑅𝑎𝑎

𝐾𝐾𝑖𝑖𝑐𝑐
 Speed Control
• The speed controller is be placed in the outer loop of the current controller (Cascaded Control)

• If the bandwidth of the speed loop is much lower than the current loop bandwidth 𝜔𝜔𝑐𝑐𝑐𝑐 ,
𝑖𝑖𝑎𝑎 𝜔𝜔𝑐𝑐𝑐𝑐
∗ = ≈1
𝑖𝑖𝑎𝑎 𝑠𝑠 + 𝜔𝜔𝑐𝑐𝑐𝑐
• And the transfer function could be derived as,

𝐾𝐾𝑇𝑇 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾 𝐾𝐾
𝑠𝑠 + 𝑇𝑇 𝑖𝑖𝑖𝑖
𝐽𝐽 𝐽𝐽 𝑠𝑠
∗
𝜔𝜔𝑚𝑚 = 𝜔𝜔𝑚𝑚 − 𝑇𝑇𝐿𝐿
𝐾𝐾𝑇𝑇 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾 𝐾𝐾 𝐾𝐾𝑇𝑇 𝐾𝐾𝑝𝑝𝑝𝑝 𝐾𝐾 𝐾𝐾 𝐽𝐽𝜔𝜔𝑐𝑐𝑐𝑐 2
𝐽𝐽𝜔𝜔𝑐𝑐𝑐𝑐
𝑠𝑠 2 + 𝑠𝑠 + 𝑇𝑇 𝑖𝑖𝑖𝑖 𝑠𝑠 2 + 𝑠𝑠 + 𝑇𝑇 𝑖𝑖𝑖𝑖
𝐽𝐽 𝐽𝐽 𝐽𝐽 𝐽𝐽 𝐾𝐾𝑝𝑝𝑝𝑝 = and 𝐾𝐾𝑖𝑖𝑖𝑖 =
𝐾𝐾𝑇𝑇 5 𝐾𝐾𝑇𝑇
Position control
• In many applications we need to control the position of the motor such as in robotics and tracking applications.
Position of the motor could be controlled using a third control loop with a bandwidth selected as five to ten times
less than that of speed loop bandwidth,
𝜔𝜔𝑐𝑐𝑠𝑠
𝜔𝜔𝑐𝑐𝑝𝑝 =
10
• since the bandwidth of the position loop is much lower than the current and speed loops,
𝜔𝜔𝑚𝑚
∗ ≈1
𝜔𝜔𝑚𝑚

𝐾𝐾𝑝𝑝𝑝𝑝

𝐾𝐾𝑝𝑝𝑝𝑝 = 𝜔𝜔𝑐𝑐𝑝𝑝
Summary
Converter for DC Machines
Power Electronics Converters
Switching Converter
• Mode 1
S1 is closed S2 is open

𝑣𝑣𝑠𝑠 = 𝑉𝑉𝑔𝑔

• Mode 2
S1 is open S2 is closed

𝑣𝑣𝑠𝑠 = 0
 𝑣𝑣𝑠𝑠
𝑉𝑉𝑔𝑔

1
• For a switching period 𝑇𝑇𝑠𝑠 , we will be switching between mode 1 and mode 2. The switching frequency 𝑓𝑓𝑠𝑠 is =
𝑇𝑇𝑠𝑠
• We have two modes of operation:
-Mode 1: S1 is ON, and S2 is OFF for some time 𝑡𝑡1
-Mode 2: S1 is OFF, and S2 is ON for some time 𝑡𝑡2
𝑡𝑡𝑜𝑜𝑜𝑜 + 𝑡𝑡𝑜𝑜𝑜𝑜𝑜𝑜 = 𝑇𝑇𝑠𝑠
• We define the duty cycle 𝐷𝐷 as:

𝑡𝑡𝑜𝑜𝑜𝑜
𝐷𝐷 =
𝑇𝑇𝑠𝑠
• Then
𝑡𝑡𝑜𝑜𝑜𝑜 = 𝐷𝐷𝑇𝑇𝑠𝑠 𝑎𝑎𝑎𝑎𝑎𝑎 𝑡𝑡𝑜𝑜𝑜𝑜𝑜𝑜 = 1 − 𝐷𝐷 𝑇𝑇𝑠𝑠
Fourier Analysis
• In this course we’ll deal with function which are periodic. For example,

• Periodic function: is function which repeats its values every period T.

𝑓𝑓 𝑡𝑡 = 𝑓𝑓 𝑡𝑡 + 𝑚𝑚𝑚𝑚 𝑚𝑚 = 1,2,3, … .
• We define the fundamental frequency as,

2𝜋𝜋
𝜔𝜔1 =
𝑇𝑇
 Fourier stated that:
• Any periodic function can be represented by an infinite sum of sine or cosine functions that are harmonics (integer multiples) of the fundamental frequency.

𝑣𝑣 𝑡𝑡 = 𝑎𝑎0 + � 𝑎𝑎𝑛𝑛 cos 𝑛𝑛𝜔𝜔1 𝑡𝑡 + 𝑏𝑏𝑛𝑛 sin 𝑛𝑛𝜔𝜔1 𝑡𝑡

𝑛𝑛=1

• Where n is an integer n = 1, 2, 3, …. and 𝑎𝑎0 , 𝑎𝑎𝑛𝑛 𝑎𝑎𝑎𝑎𝑎𝑎 𝑏𝑏𝑛𝑛 are known as the Fourier coefficients.

• The harmonic frequency 𝜔𝜔𝑛𝑛 is defined as an integer multiple of the fundamental frequency

𝜔𝜔𝑛𝑛 = 𝑛𝑛 𝜔𝜔1

• For any periodic function, we can find 𝑎𝑎0 , 𝑎𝑎𝑛𝑛 and 𝑏𝑏𝑛𝑛 to represent that function in the Fourier Series from. To find the Fourier Coefficients

1 𝑇𝑇 1
𝑎𝑎0 = � 𝑣𝑣(𝑡𝑡) 𝑑𝑑𝑑𝑑 = × 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 (𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 𝑡𝑡𝑡𝑡𝑡 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑜𝑜𝑜𝑜 𝑜𝑜𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝)
𝑇𝑇 𝑜𝑜 𝑇𝑇

2 𝑇𝑇
𝑎𝑎𝑛𝑛 = � 𝑣𝑣(𝑡𝑡) cos 𝑛𝑛𝜔𝜔1 𝑡𝑡 𝑑𝑑𝑑𝑑
𝑇𝑇 𝑜𝑜

2 𝑇𝑇
𝑏𝑏𝑛𝑛 = � 𝑣𝑣(𝑡𝑡) sin 𝑛𝑛𝜔𝜔1 𝑡𝑡 𝑑𝑑𝑑𝑑
𝑇𝑇 𝑜𝑜

• Note that the coefficient 𝑎𝑎0 is the average of the periodic function. It is also known as the dc component.
 Filter
• An ideal filter can eliminate the fundamental frequency and all the harmonics.
Switching Converter

• Using Fourier analysis, 𝑣𝑣𝑠𝑠 𝑡𝑡 has a dc component:

1 𝑇𝑇𝑠𝑠 1 𝐷𝐷𝑇𝑇𝑠𝑠 𝑇𝑇𝑠𝑠

𝑉𝑉𝑠𝑠 = < 𝑣𝑣𝑠𝑠 > = 𝑎𝑎0 = � 𝑣𝑣𝑠𝑠 (𝑡𝑡) 𝑑𝑑𝑑𝑑 = � 𝑉𝑉𝑔𝑔 𝑑𝑑𝑑𝑑 + � 0 𝑑𝑑𝑑𝑑 = 𝐷𝐷𝑉𝑉𝑔𝑔
𝑇𝑇𝑠𝑠 𝑜𝑜 𝑇𝑇𝑠𝑠 𝑜𝑜 𝐷𝐷𝑇𝑇𝑠𝑠

• To eliminate the harmonics and keep the dc component only, we will use a low-pass filter LPF.
Example
• For a switching converter, if Vg = 6 V, the switching frequency is 10 kHz, and the duty cycle is
25%. Sketch the output voltage vs(t) and find the average value.

𝑉𝑉𝑠𝑠 = 𝐷𝐷 𝑉𝑉𝑔𝑔 = 0.25 × 6 = 1.5 𝑉𝑉

Buck Converter
Buck Converter

• Since the harmonics are eliminated,

𝑣𝑣 𝑡𝑡 ≈ 𝑉𝑉𝑠𝑠 = 𝐷𝐷 𝑉𝑉𝑔𝑔

• In this example, if 𝐷𝐷 = 0.5,

𝑣𝑣 𝑡𝑡 = 𝑉𝑉 = 𝐷𝐷 𝑉𝑉𝑔𝑔 = 0.5 × 100 = 50 𝑉𝑉
Buck converter
• Since the duty cycle D is limited between 0 and 1,

0 ≤ 𝐷𝐷 ≤ 1

• The output voltage,

0 ≤ 𝑉𝑉 ≤ 𝑉𝑉𝑔𝑔

• The output voltage cannot be negative, and it is less than the input voltage.
Small-Ripple Approximation
• the output voltage v(t) is well approximated by its dc component V, with the small ripple
term vripple(t) neglected.

• The small-ripple approximation is applied to the inductor currents and capacitor voltages of
the converter, which are continuous variables.
Buck Converter
• Eventually, the point is reached where the increase in inductor current during subinterval 1 is equal to the
decrease in inductor current during subinterval 2.

• The converter waveforms are periodic:

𝑖𝑖𝐿𝐿 (𝑛𝑛𝑇𝑇𝑠𝑠 ) = 𝑖𝑖𝐿𝐿 ( 𝑛𝑛 + 1 𝑇𝑇𝑠𝑠 )
Periodic Steady State
• When we first energize a circuit, there will be transients for a short time. After a while, the circuit will reach
steady state.

DC Steady state AC Steady state Periodic Steady state

Inductor volt-second balance

𝑉𝑉𝐿𝐿𝐿 𝐷𝐷 + 𝑉𝑉𝐿𝐿𝐿 1 − 𝐷𝐷 = 0
Capacitor amp-second balance

𝑖𝑖𝐶𝐶𝐶 𝐷𝐷 + 𝑖𝑖𝐶𝐶𝐶 1 − 𝐷𝐷 = 0
Buck Converter Modes

𝑑𝑑𝑣𝑣𝑐𝑐 𝑖𝑖𝑐𝑐 𝑖𝑖𝐿𝐿 𝑉𝑉 𝑑𝑑𝑣𝑣𝑐𝑐 𝑖𝑖𝑐𝑐 𝑖𝑖𝐿𝐿 𝑉𝑉

= ≈ − = ≈ −
𝑑𝑑𝑑𝑑 𝐶𝐶 𝐶𝐶 𝑅𝑅𝑅𝑅 𝑑𝑑𝑑𝑑 𝐶𝐶 𝐶𝐶 𝑅𝑅𝑅𝑅
Then
• Using the inductor volt-sec balance
𝑉𝑉 = 𝐷𝐷 𝑉𝑉𝑔𝑔

• Using the capacitor ampere-sec balance

𝑉𝑉
𝐼𝐼𝐿𝐿 =
𝑅𝑅
Buck Converter Ripple
• Inductor current ripple

• Capacitor Voltage Ripple

Example
• In the buck dc-dc converter has the following parameters:
𝑉𝑉𝑔𝑔 = 50𝑉𝑉
𝐷𝐷 = 0.4
𝐿𝐿 = 400 𝜇𝜇𝜇𝜇
𝐶𝐶 = 100𝑢𝑢𝑢𝑢
𝑓𝑓𝑠𝑠 = 20 𝑘𝑘𝑘𝑘𝑘𝑘
𝑅𝑅 = 20Ω
• Calculate,
a) The output voltage 𝑉𝑉.
b) The maximum and minimum inductor current.
c) The output voltage ripple.
Solution
• 𝑉𝑉 = 𝐷𝐷𝑉𝑉𝑔𝑔 = 0.4 50 = 20𝑉𝑉
𝑉𝑉𝑜𝑜 20
• 𝐼𝐼𝑅𝑅 = 𝐼𝐼𝐿𝐿 = = = 1𝐴𝐴
𝑅𝑅 20

𝑉𝑉 1−𝐷𝐷 20 1−0.4
• ∆𝑖𝑖𝐿𝐿 = = = 0.75 𝐴𝐴
𝑓𝑓𝑠𝑠 2𝐿𝐿 20×10 2×400×10−6
3

• 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 = 𝐼𝐼𝐿𝐿 + ∆𝑖𝑖𝐿𝐿 = 1.75𝐴𝐴 𝑎𝑎𝑎𝑎𝑎𝑎 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 = 𝐼𝐼𝐿𝐿 − ∆𝑖𝑖𝐿𝐿 = 0.25𝐴𝐴

∆𝑉𝑉 1−𝐷𝐷
• = = 0.23%
𝑉𝑉 8 2𝐿𝐿𝐿𝐿 𝑓𝑓𝑠𝑠2
Switching Schemes
The boost converter
The boost converter
1
𝑉𝑉 = 𝑉𝑉
1 − 𝐷𝐷 𝑔𝑔
Buck-Boost converter
The buck-boost converter
𝐷𝐷
𝑉𝑉 = − 𝑉𝑉
1 − 𝐷𝐷 𝑔𝑔
Full Bridge (H-bridge)