LECTURE 7 – ANALYSIS IN THE TIME DOMAIN

This section attempts to build up a consistent picture of how systems of various complexities
will respond to standard inputs driving them.

We recall

𝑌𝑌(𝑠𝑠) 𝑛𝑛(𝑠𝑠)
where 𝑋𝑋(𝑠𝑠)
= 𝐺𝐺(𝑠𝑠) ≡ 𝑑𝑑(𝑠𝑠)

𝑛𝑛(𝑠𝑠) is the numerator polynomial in s and 𝑑𝑑(𝑠𝑠) the denominator. If we want to know
what the transient response of the system might look like, it is 𝑑𝑑(𝑠𝑠) which determines all of
the exponential terms that appear back in time domain solultions. To find these
exponentials we equate 𝑑𝑑(𝑠𝑠) = 0. This is known as the CHARACTERISTIC EQUATION and is
fundamental to all studies in control. Note that

𝑛𝑛(𝑠𝑠) 𝐴𝐴 𝐵𝐵
𝑑𝑑(𝑠𝑠)
= (𝑠𝑠+𝑑𝑑 ) + (𝑠𝑠+𝑑𝑑 ) ⋯ etc
1 2

𝐴𝐴
and ℒ −1 {𝑠𝑠+𝑑𝑑 } = 𝐴𝐴𝑒𝑒 −𝑑𝑑1 𝑡𝑡 ... etc.
1

To discover the response of any system we must find ℒ −1 {𝐺𝐺(𝑠𝑠)𝑋𝑋(𝑠𝑠)}. We will now choose
𝐺𝐺(𝑠𝑠) and 𝑋𝑋(𝑠𝑠).

FOUR INPUTS 𝑋𝑋(𝑠𝑠)

We will work with the following inputs

IDENTITY 𝑥𝑥(𝑡𝑡) 𝑋𝑋(𝑠𝑠)

Unit impulse 𝛿𝛿(𝑡𝑡) 1
Unit step 𝑢𝑢(𝑡𝑡) 1�
𝑠𝑠
Unit ramp 𝑡𝑡 1�
𝑠𝑠 2
Unit cosine 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑠𝑠
(𝑠𝑠 2 + 𝜔𝜔 2 )

TWO SYSTEMS 𝐺𝐺(𝑠𝑠)

We use two systems that are commonly called:

𝐾𝐾 𝐾𝐾
FIRST ORDER 𝑠𝑠+𝑎𝑎
, SECOND ORDER 𝑠𝑠2 +2𝜉𝜉𝜔𝜔 2
,
𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

the order representing the multiplicity of differentiation. The form, with 𝐾𝐾, 𝑎𝑎, 𝜉𝜉, 𝜔𝜔𝑛𝑛 all being
specific constants, will become clear later on.

[The simplest system/input combinations will be fully analysed. The more complex ones will
be given by analogy.]
 𝐾𝐾
TAKE THE SYSTEM 𝑠𝑠+𝑎𝑎
With 𝛿𝛿(𝑡𝑡) A decaying exponential

𝐾𝐾 𝐾𝐾
𝑌𝑌(𝑠𝑠) = 𝑠𝑠+𝑎𝑎 ∙ 1 ∴ 𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠+𝑎𝑎}

∴ 𝑦𝑦(𝑡𝑡) = 𝐾𝐾𝑒𝑒 −𝑎𝑎𝑎𝑎

With u(𝑡𝑡) A rising exponential

𝐾𝐾 1 𝐾𝐾
𝑌𝑌(𝑠𝑠) = 𝑠𝑠+𝑎𝑎 ∙ 𝑠𝑠 ∴ 𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠(𝑠𝑠+𝑎𝑎)}

𝐾𝐾
∴ 𝑦𝑦(𝑡𝑡) = 𝑎𝑎 (1 − 𝑒𝑒 −𝑎𝑎𝑎𝑎 )

With 𝑡𝑡
A delayed ramp output
𝐾𝐾 1 𝐾𝐾
𝑌𝑌(𝑠𝑠) = (𝑠𝑠+𝑎𝑎) ∙ 𝑠𝑠2 ∴ 𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠2 (𝑠𝑠+𝑎𝑎)}

𝐾𝐾
∴ 𝑦𝑦(𝑡𝑡) = 𝑎𝑎2 (𝑒𝑒 −𝑎𝑎𝑎𝑎 + 𝑎𝑎𝑎𝑎 − 1)

𝑠𝑠
With 𝑠𝑠2 +𝜔𝜔2

𝐾𝐾 𝑠𝑠 𝐾𝐾 𝑠𝑠
𝑌𝑌(𝑠𝑠) = (𝑠𝑠+𝑎𝑎) ∙ (𝑠𝑠2 +𝜔𝜔2 ) ∴ 𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠+𝑎𝑎 ∙ 𝑠𝑠2 +𝜔𝜔2 }

𝐾𝐾 𝐾𝐾

∴ 𝑦𝑦(𝑡𝑡) = 𝑎𝑎
𝜔𝜔 2
𝑒𝑒 −𝑎𝑎𝑎𝑎 + 𝑎𝑎
1 𝑐𝑐𝑐𝑐𝑐𝑐(𝜔𝜔𝜔𝜔 − 𝜑𝜑)
�1+� � � 𝜔𝜔 2 2
𝑎𝑎 �1+� � �
𝑎𝑎

𝜔𝜔
𝜑𝜑 = 𝑡𝑡𝑡𝑡𝑡𝑡−1 𝑎𝑎 . Notice the first term is a transient and dies out after about 5/a seconds. We
will plot the second term that persists, against the input.

A delayed cosine of
different magnitude
 𝐾𝐾
TAKE THE SYSTEM 𝑠𝑠2 +2𝜉𝜉𝜔𝜔𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛 2

Before we can proceed, it has to be noted that the characteristic equation of this system
gives rise to roots that are in some ways not fully defined. We study the implications of this
point before proceeding to a full analysis. Look again at the possible roots of the
denominator and note the significance of the parameter 𝜉𝜉 known as the DAMPING RATIO
(or DAMPING COEFFICIENT).

𝐾𝐾
Case 1 (𝑠𝑠+𝛼𝛼)(𝑠𝑠+𝛽𝛽)
The roots are both real and different. Therefore, this system is
known to be OVERDAMPED and 𝜉𝜉 > 1.

𝐾𝐾
Case 2 (𝑠𝑠+𝛾𝛾)2
The roots are real and the same. Therefore, this system is known
to be CRITICALLY DAMPED and 𝜉𝜉 = 1.

𝐾𝐾
Case 3 𝑠𝑠2 +2𝜉𝜉𝜔𝜔𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛 2
The roots are a complex conjugate pair. Therefore the system is
known to be UNDERDAMPED and 0 < 𝜉𝜉 < 1.

𝐾𝐾
Case 4 𝑠𝑠2 +𝜔𝜔 2
Here 𝜉𝜉 = 0 and the roots are imaginary. Therefore the system is
totally UNDAMPED, that is, it will oscillate at all times.

A selection of system cases with inputs will now be analysed.

𝐾𝐾
WITH 𝜹𝜹(𝒕𝒕) 𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠2 +2𝜉𝜉𝜔𝜔 2
} OR one of the other forms above
𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

K
IF 𝜉𝜉 > 1 , y(t) = β−α �e−αt − e−βt �

IF 𝜉𝜉 = 1 , 𝑦𝑦(𝑡𝑡) = 𝐾𝐾𝐾𝐾𝑒𝑒 −𝛾𝛾𝛾𝛾

K 1
IF 𝜉𝜉 < 1 , y(t) = ω ∙ ∙ e−𝜉𝜉𝜉𝜉𝜉𝜉 ∙ sin[𝜔𝜔𝑛𝑛 �1 − 𝜉𝜉 2 t]
�1−𝜉𝜉 2

𝐾𝐾
IF 𝜉𝜉 = 0 , 𝑦𝑦(𝑡𝑡) = 𝜔𝜔 ∙ 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

Hence we see the bigger 𝜉𝜉 is the more damped the output is.
WITH 𝒖𝒖(𝒕𝒕)

Here is the analysis for an UNDERDAMPED system:

𝐾𝐾
Case 3 𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠(𝑠𝑠2 +2𝜉𝜉𝜔𝜔 }
2)
𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

𝐾𝐾 1
= 𝜔𝜔 2 �1 − ∙ 𝑒𝑒 −𝜉𝜉𝜔𝜔𝑛𝑛𝑡𝑡 ∙ 𝑠𝑠𝑠𝑠𝑠𝑠�𝜔𝜔𝑛𝑛 �1 − 𝜉𝜉 2 𝑡𝑡 + 𝜑𝜑��
𝑛𝑛 �1−𝜉𝜉 2

�1−𝜉𝜉 2
with 𝜑𝜑 = 𝑡𝑡𝑡𝑡𝑡𝑡−1 ( )
𝜉𝜉

To see what this looks like graphically, we reveal

𝑠𝑠𝑠𝑠𝑠𝑠(𝐴𝐴 + 𝐵𝐵) = 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 + 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐, with 𝐴𝐴 = 𝜔𝜔𝑛𝑛 �1 − 𝜉𝜉 2 𝑡𝑡 , 𝑏𝑏 = 𝜑𝜑

Hence we rewrite our output as

𝐾𝐾1 𝜉𝜉
𝑦𝑦(𝑡𝑡) = 2
�1 − ∙ 𝑒𝑒 −𝜉𝜉𝜔𝜔𝑛𝑛𝑡𝑡 ∙ 𝑠𝑠𝑠𝑠𝑠𝑠 �𝜔𝜔𝑛𝑛 �1 − 𝜉𝜉 2 𝑡𝑡� − 𝑒𝑒 −𝜉𝜉𝜔𝜔𝑛𝑛𝑡𝑡 ∙ 𝑐𝑐𝑐𝑐𝑐𝑐 �𝜔𝜔𝑛𝑛 �1 − 𝜉𝜉 2 𝑡𝑡��
𝜔𝜔𝑛𝑛 �1 − 𝜉𝜉 2

First term 𝐾𝐾
is just scaling
𝜔𝜔𝑛𝑛 2
- factor

Second term

Third term

Gives this A step function with

output decaying oscillations

WITH 𝒕𝒕

Here is the analysis for a CRITICALLY DAMPED system:

𝐾𝐾
Case 2 𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠2 (𝑠𝑠+𝛾𝛾)2 }
 𝐾𝐾 2 2
= 𝛾𝛾2 �𝑡𝑡 − 𝛾𝛾 + �𝑡𝑡 + 𝛾𝛾� 𝑒𝑒 −𝛾𝛾𝛾𝛾 �

Let us look at the output graphically.

WITH 𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄

Case 1 Complete the analysis for an OVERDAMPED system:

𝑠𝑠 𝐾𝐾
𝑦𝑦(𝑡𝑡) = ℒ −1 {𝑠𝑠2 +𝜔𝜔2 ∙ (𝑠𝑠+𝛼𝛼)(𝑠𝑠+𝛽𝛽)}

HINT: Use the partial fractions method.

Case 3 UNDERDAMPED

𝑠𝑠 𝐾𝐾
𝑦𝑦(𝑡𝑡) = ℒ −1 {(𝑠𝑠2 +𝜔𝜔2) ∙ 𝑠𝑠2 +2𝜉𝜉𝜔𝜔 2
}
𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

According to the partial fractions rules:

𝐴𝐴𝐴𝐴+𝐵𝐵 𝐶𝐶𝐶𝐶+𝐷𝐷
𝑌𝑌(𝑠𝑠) = 𝑠𝑠2 +𝜔𝜔2 + 𝑠𝑠2 +2𝜉𝜉𝜔𝜔 2
,
𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

which results in:

𝐾𝐾𝐾𝐾 1 1
𝑌𝑌(𝑠𝑠) = �𝜔𝜔 𝑠𝑠𝑠𝑠𝑠𝑠(𝜔𝜔𝜔𝜔 + 𝜑𝜑1 ) + ∙ 𝑒𝑒 −𝜉𝜉𝜔𝜔𝑛𝑛𝑡𝑡 ∙ 𝑠𝑠𝑠𝑠𝑠𝑠�𝜔𝜔𝑛𝑛 �1 − 𝜉𝜉 2 𝑡𝑡 + 𝜑𝜑2 �� ,
𝐾𝐾1 𝜔𝜔𝑛𝑛 �1−𝜉𝜉 2

with: 𝐾𝐾1 = �4𝜉𝜉𝜔𝜔𝑛𝑛 2 𝜔𝜔 2 + 𝜔𝜔𝑛𝑛 2 −𝜔𝜔 2 ,

 2𝜉𝜉𝜔𝜔𝑛𝑛 𝜔𝜔 2𝜉𝜉𝜔𝜔𝑛𝑛 3 �1−𝜉𝜉 2 �
𝜑𝜑1 = 𝑡𝑡𝑡𝑡𝑡𝑡−1 (− 𝜔𝜔 2 −𝜔𝜔2
) , 𝜑𝜑2 = 𝑡𝑡𝑡𝑡𝑡𝑡−1 [− 𝜔𝜔𝑛𝑛 2 −𝜔𝜔2
]
𝑛𝑛

1 𝐾𝐾
[Alternatively: 𝑦𝑦(𝑡𝑡) = ℒ −1 {(𝑠𝑠2 +𝜔𝜔2 ) ∙ (𝑠𝑠+𝛼𝛼)2 +𝑏𝑏2 }

𝐾𝐾𝐾𝐾 1 1
= 𝐾𝐾1 𝜔𝜔
� 𝑠𝑠𝑠𝑠𝑠𝑠(𝜔𝜔𝜔𝜔 + 𝜑𝜑1 ) + 𝑏𝑏 ∙ 𝑒𝑒 −𝑎𝑎𝑎𝑎 ∙ 𝑠𝑠𝑠𝑠𝑠𝑠(𝑏𝑏𝑏𝑏 + 𝜑𝜑2 )� ,

with 𝐾𝐾1 = [4𝑎𝑎2 𝜔𝜔2 + (𝑎𝑎2 + 𝑏𝑏 2 − 𝜔𝜔2 )]1/2 ,

2𝑎𝑎𝑎𝑎 2𝑎𝑎𝑎𝑎
𝜑𝜑1 = 𝑡𝑡𝑡𝑡𝑡𝑡−1 (− 𝑎𝑎2 +𝑏𝑏2−𝜔𝜔2) , 𝜑𝜑2 = 𝑡𝑡𝑡𝑡𝑡𝑡−1 (− 𝑎𝑎2 +𝑏𝑏2 −𝜔𝜔2 )

Again it is best to look at this graphically. There are just two terms in the solution.
𝐾𝐾𝐾𝐾
Ignore 𝐾𝐾 as a scaling constant.
1