Plan of the Lecture

I Review: prototype 2nd-order system

I Today’s topic: transient response specifications
Plan of the Lecture

I Review: prototype 2nd-order system

I Today’s topic: transient response specifications

Goal: develop formulas and intuition for various features of the

transient response: rise time, overshoot, settling time.
Plan of the Lecture

I Review: prototype 2nd-order system

I Today’s topic: transient response specifications

Goal: develop formulas and intuition for various features of the

transient response: rise time, overshoot, settling time.

Reading: FPE, Sections 3.3–3.4; lab manual

Prototype 2nd-Order System
!n2
H(s) =
s2 + 2⇣!n s + !n2
By the quadratic formula, the poles are:

p
s= ⇣!n ± !n ⇣ 2 1
⇣ p ⌘
= !n ⇣ ± ⇣ 2 1

The nature of the poles changes depending on ⇣:

I ⇣ >1 both poles are real and negative
I ⇣ =1 one negative pole
I ⇣ <1 two complex poles with negative real parts

s= ± j!d
p
where = ⇣!n , !d = !n 1 ⇣2
Prototype 2nd-Order System
!n2
H(s) = , ⇣<1
s2 + 2⇣!n s + !n2
The poles are
p
s= ⇣!n ± j!n 1 ⇣2 = ± j!d

Im

p
!d = !n 1 ⇣2 Note that
!n
2
' + !d2 = ⇣ 2 !n2 + !n2 ⇣ 2 !n2
= ⇣!n Re
0 = !n2
⇣!n
cos ' = =⇣
!n
2nd-Order Response
Let’s compute the system’s impulse and step response:
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 !n2
h(t) = L {H(s)} = L
(s + )2 + !d2
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 (!n2 /!d )!d
h(t) = L {H(s)} = L
(s + )2 + !d2
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 (!n2 /!d )!d
h(t) = L {H(s)} = L
(s + )2 + !d2
!n2 t
= e sin(!d t)
!d
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 (!n2 /!d )!d
h(t) = L {H(s)} = L
(s + )2 + !d2
!n2 t
= e sin(!d t) (table, # 20)
!d
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 (!n2 /!d )!d
h(t) = L {H(s)} = L
(s + )2 + !d2
!n2 t
= e sin(!d t) (table, # 20)
!d

I Step response:
⇢ ⇢
1 H(s) 1 !n2
L =L
s s[(s + )2 + !d2 ]
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 (!n2 /!d )!d
h(t) = L {H(s)} = L
(s + )2 + !d2
!n2 t
= e sin(!d t) (table, # 20)
!d

I Step response:
⇢ ⇢ 2
1 H(s) 1 + !d2
L =L
s s[(s + )2 + !d2 ]
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 (!n2 /!d )!d
h(t) = L {H(s)} = L
(s + )2 + !d2
!n2 t
= e sin(!d t) (table, # 20)
!d

I Step response:
⇢ ⇢ 2
H(s) + !d2
L 1 =L 1
s s[(s + )2 + !d2 ]
✓ ◆
t
=1 e cos(!d t) + sin(!d t)
!d
2nd-Order Response
Let’s compute the system’s impulse and step response:
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

I Impulse response:
⇢
1 1 (!n2 /!d )!d
h(t) = L {H(s)} = L
(s + )2 + !d2
!n2 t
= e sin(!d t) (table, # 20)
!d

I Step response:
⇢ ⇢ 2
H(s) + !d2
L 1 =L 1
s s[(s + )2 + !d2 ]
✓ ◆
t
=1 e cos(!d t) + sin(!d t) (table, #21)
!d
2nd-Order Step Response
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
✓ ◆
t
u(t) = 1(t) ! y(t) = 1 e cos(!d t) + sin(!d t)
!d
p
where = ⇣!n and !d = !n 1 ⇣ 2 (damped frequency)
2nd-Order Step Response
!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
✓ ◆
t
u(t) = 1(t) ! y(t) = 1 e cos(!d t) + sin(!d t)
!d
p
where = ⇣!n and !d = !n 1 ⇣ 2 (damped frequency)
yHtL
The parameter ⇣ is called
1.5 the damping ratio
I ⇣ > 1: system is
1.0 overdamped
I ⇣ < 1: system is
z=0.1
0.5 underdamped
z=0.9
I ⇣ = 0: no damping
z=1
t (!d = !n )
2 4 6 8 10 12 14
2nd-Order Step Response

!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
✓ ◆
t
u(t) = 1(t) ! y(t) = 1 e cos(!d t) + sin(!d t)
!d
p
where = ⇣!n and !d = !n 1 ⇣ 2 (damped frequency)
2nd-Order Step Response

!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
✓ ◆
t
u(t) = 1(t) ! y(t) = 1 e cos(!d t) + sin(!d t)
!d
p
where = ⇣!n and !d = !n 1 ⇣ 2 (damped frequency)

We will see that the parameters ⇣ and !n determine certain

important features of the transient part of the above step
response.
2nd-Order Step Response

!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
✓ ◆
t
u(t) = 1(t) ! y(t) = 1 e cos(!d t) + sin(!d t)
!d
p
where = ⇣!n and !d = !n 1 ⇣ 2 (damped frequency)

We will see that the parameters ⇣ and !n determine certain

important features of the transient part of the above step
response.

We will also learn how to pick ⇣ and !n in order to shape these

features according to given specifications.
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0
s+a
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0 (stable pole)
s+a
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0 (stable pole)
s+a
DC gain = 1 (by FVT)
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0 (stable pole)
s+a
DC gain = 1 (by FVT)

Step response:
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0 (stable pole)
s+a
DC gain = 1 (by FVT)
H(s) a 1 1
Step response: Y (s) = = =
s s(s + a) s s+a
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0 (stable pole)
s+a
DC gain = 1 (by FVT)
H(s) a 1 1
Step response: Y (s) = = =
s s(s + a) s s+a
y(t) = L {Y (s)} = 1(t) e at
1
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0 (stable pole)
s+a
DC gain = 1 (by FVT)
H(s) a 1 1
Step response: Y (s) = = =
s s(s + a) s s+a
y(t) = L {Y (s)} = 1(t) e at
1

yHtL
1.0

0.8

0.6

0.4

0.2

t
0.5 1.0 1.5 2.0
Transient Response Specifications: Rise Time
Let’s first take a look at 1st-order step response
a
H(s) = , a>0 (stable pole)
s+a
DC gain = 1 (by FVT)
H(s) a 1 1
Step response: Y (s) = = =
s s(s + a) s s+a
y(t) = L {Y (s)} = 1(t) e at
1

yHtL
1.0

0.8

0.6
Rise time tr : the time it
takes to get from 10% of
0.4 steady-state value to 90%
rise time tr
0.2

t
0.5 1.0 1.5 2.0
Rise Time
Step response: y(t) = 1(t) e at

yHtL
1.0

0.8

0.6
Rise time tr : the time it
takes to get from 10% of
0.4 steady-state value to 90%
rise time tr
0.2

t
0.5 1.0 1.5 2.0
Rise Time
Step response: y(t) = 1(t) e at

yHtL
1.0

0.8

0.6
Rise time tr : the time it
takes to get from 10% of
0.4 steady-state value to 90%
rise time tr
0.2

t
0.5 1.0 1.5 2.0

Examples of rise time:

I car — going from 0 to 60 mph in 7 sec
I oven — reach desired preheat temperature quickly
I thermostat, building climate control
I other examples?
Rise Time
Step response: y(t) = 1(t) e at

yHtL
1.0

0.8

0.6
Rise time tr : the time it
takes to get from 10% of
0.4 steady-state value to 90%
rise time tr
0.2

t
0.5 1.0 1.5 2.0

In this example, it is easy to compute tr analytically:

Rise Time
Step response: y(t) = 1(t) e at

yHtL
1.0

0.8

0.6
Rise time tr : the time it
takes to get from 10% of
0.4 steady-state value to 90%
rise time tr
0.2

t
0.5 1.0 1.5 2.0

In this example, it is easy to compute tr analytically:

at0.1 at0.1 ln 0.9
1 e = 0.1 e = 0.9 t0.1 =
a
Rise Time
Step response: y(t) = 1(t) e at

yHtL
1.0

0.8

0.6
Rise time tr : the time it
takes to get from 10% of
0.4 steady-state value to 90%
rise time tr
0.2

t
0.5 1.0 1.5 2.0

In this example, it is easy to compute tr analytically:

at0.1 at0.1 ln 0.9
1 e = 0.1 e = 0.9 t0.1 =
a
at0.9 at0.9 ln 0.1
1 e = 0.9 e = 0.1 t0.9 =
a
Rise Time
Step response: y(t) = 1(t) e at

yHtL
1.0

0.8

0.6
Rise time tr : the time it
takes to get from 10% of
0.4 steady-state value to 90%
rise time tr
0.2

t
0.5 1.0 1.5 2.0

In this example, it is easy to compute tr analytically:

at0.1 at0.1 ln 0.9
1 e = 0.1 e = 0.9 t0.1 =
a
at0.9 at0.9 ln 0.1
1 e = 0.9 e = 0.1 t0.9 =
a
ln 0.9 ln 0.1 ln 9 2.2
tr = t0.9 t0.1 = = ⇡
a a a
Transient Response Specs
Now let’s consider the more interesting case: 2nd-order response

!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
p
where = ⇣!n !d = !n 1 ⇣ 2 (⇣ < 1)
Transient Response Specs
Now let’s consider the more interesting case: 2nd-order response

!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
p
where = ⇣!n !d = !n 1 ⇣ 2 (⇣ < 1)
Im

p
!d = !n 1 ⇣2
!n

'
= ⇣!n Re
0
Transient Response Specs
Now let’s consider the more interesting case: 2nd-order response

!n2 !n2
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
p
where = ⇣!n !d = !n 1 ⇣ 2 (⇣ < 1)
Im

p
!d = !n 1 ⇣2
!n

'
= ⇣!n Re
0

✓ ◆
Step response: y(t) = 1 e t cos(!d t) + sin(!d t)
!d
Transient-Response Specs ✓ ◆
Step response: y(t) = 1 e t cos(!d t) + sin(!d t)
!d
yHtL
1.4
1.2
1.0
0.8
0.6
0.4
0.2
wn t
2 4 6 8 10 12 14
Transient-Response Specs ✓ ◆
Step response: y(t) = 1 e t cos(!d t) + sin(!d t)
!d
yHtL
1.4
1.2
1.0
0.8
0.6
0.4
0.2
wn t
2 4 6 8 10 12 14
I rise time tr — time to get from 0.1y(1) to 0.9y(1)
I overshoot Mp and peak time tp
I settling time ts — first time for transients to decay to
within a specified small percentage of y(1) and stay in
that range (we will usually worry about 5% settling time)
Transient-Response Specs ✓ ◆
Step response: y(t) = 1 e t cos(!d t) + sin(!d t)
!d
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
I rise time tr — time to get from 0.1y(1) to 0.9y(1)
I overshoot Mp and peak time tp
I settling time ts — first time for transients to decay to
within a specified small percentage of y(1) and stay in
that range (we will usually worry about 5% settling time)
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp small
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp small
I tp
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp small
I tp small
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp small
I tp small
I ts
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp small
I tp small
I ts small
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp small
I tp small
I ts small
Trade-o↵s among specs:
Transient-Response (or Time-Domain) Specs
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Do we want these quantities to be large or small?

I tr small
I Mp small
I tp small
I ts small
Trade-o↵s among specs: decrease tr ! increase Mp , etc.
 yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
Formulas for TD Specs: Rise Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
Formulas for TD Specs: Rise Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Rise time tr — hard to calculate analytically.

Formulas for TD Specs: Rise Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Rise time tr — hard to calculate analytically.

Empirically, on the normalized time scale (t ! !n t), rise times
are approximately the same
Formulas for TD Specs: Rise Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Rise time tr — hard to calculate analytically.

Empirically, on the normalized time scale (t ! !n t), rise times
are approximately the same

wn tr ⇡ 1.8 (exact for ⇣ = 0.5)

Formulas for TD Specs: Rise Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

Rise time tr — hard to calculate analytically.

Empirically, on the normalized time scale (t ! !n t), rise times
are approximately the same

wn tr ⇡ 1.8 (exact for ⇣ = 0.5)

1.8
So, we will work with tr ⇡ (good approx. when ⇣ ⇡ 0.5)
!n
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

tp is the first time t > 0 when y 0 (t) = 0

Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

tp is the first time t > 0 when y 0 (t) = 0

✓ ◆
t
y(t) = 1 e cos(!d t) + sin(!d t)
!d
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

tp is the first time t > 0 when y 0 (t) = 0

✓ ◆
t
y(t) = 1 e cos(!d t) + sin(!d t)
!d
✓ 2 ◆
0
y (t) = + !d e t sin(!d t)
!d
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

tp is the first time t > 0 when y 0 (t) = 0

✓ ◆
t
y(t) = 1 e cos(!d t) + sin(!d t)
!d
✓ 2 ◆
0
y (t) = + !d e t sin(!d t) = 0 when !d t = 0, ⇡, 2⇡, . . .
!d
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

tp is the first time t > 0 when y 0 (t) = 0

✓ ◆
t
y(t) = 1 e cos(!d t) + sin(!d t)
!d
✓ 2 ◆
0
y (t) = + !d e t sin(!d t) = 0 when !d t = 0, ⇡, 2⇡, . . .
!d
⇡
so tp =
!d
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
⇡
We have just computed tp =
!d
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
⇡
We have just computed tp =
!d
To find Mp , plug this value into y(t):
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
⇡
We have just computed tp =
!d
To find Mp , plug this value into y(t):

Mp = y(tp ) 1
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
⇡
We have just computed tp =
!d
To find Mp , plug this value into y(t):
✓ ✓ ◆ ✓ ◆◆
⇡ ⇡ ⇡
Mp = y(tp ) 1 = e !d cos !d + sin !d
!d !d !d
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
⇡
We have just computed tp =
!d
To find Mp , plug this value into y(t):
✓ ✓ ◆ ✓ ◆◆
⇡ ⇡ ⇡
Mp = y(tp ) 1 = e !d cos !d + sin !d
!d !d !d
✓ ◆
⇡
= exp
!d
Formulas for TD Specs: Overshoot & Peak Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
⇡
We have just computed tp =
!d
To find Mp , plug this value into y(t):
✓ ✓ ◆ ✓ ◆◆
⇡ ⇡ ⇡
Mp = y(tp ) 1 = e !d cos !d + sin !d
!d !d !d
✓ ◆ !
⇡ ⇡⇣
= exp = exp p — exact formula
!d 1 ⇣2
Formulas for TD Specs: Settling Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14
Formulas for TD Specs: Settling Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

⇢
|y(t0 ) y(1)|
ts = min t > 0 :  0.05 for all t0 t (here,
y(1)
y(1) = 1)
Formulas for TD Specs: Settling Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

⇢
|y(t0 ) y(1)|
ts = min t > 0 :  0.05 for all t0 t (here,
y(1)
y(1) = 1)

t
|y(t) 1| = e cos(!d t) + sin(!d t)
!d
Formulas for TD Specs: Settling Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

⇢
|y(t0 ) y(1)|
ts = min t > 0 :  0.05 for all t0 t (here,
y(1)
y(1) = 1)

t
|y(t) 1| = e cos(!d t) + sin(!d t)
!d
here, e t is what matters (sin and cos are bounded between
±1), so e ts  0.05
Formulas for TD Specs: Settling Time
yHtL
1.4
1.2 Mp
1.0
0.8
0.6
0.4
0.2
! n tr ! n tp ! n ts
wn t
2 4 6 8 10 12 14

⇢
|y(t0 ) y(1)|
ts = min t > 0 :  0.05 for all t0 t (here,
y(1)
y(1) = 1)

t
|y(t) 1| = e cos(!d t) + sin(!d t)
!d
here, e t
is what matters (sin and cos are bounded between
ln 0.05 3
±1), so e ts  0.05 this gives ts = ⇡
Formulas for TD Specs
2 + !2
!n2 d
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2

1.8
tr ⇡
!n
⇡
tp =
!d !
⇡⇣
Mp = exp p
1 ⇣2
3
ts ⇡
TD Specs in Frequency Domain
We want to visualize time-domain specs in terms of admissible
pole locations for the 2nd-order system
2 + !2
!n2 d
H(s) = =
s2 + 2⇣!n s + !n2 (s + )2 + !d2
where = ⇣!n
p
!d = !n 1 ⇣2
⇣ ⌘
Step response: y(t) = 1 e t cos(!d t) + sin(!d t)
!d

Im

p
!d = !n 1 ⇣2
!n

'
= ⇣!n 0
Re !n2 = 2
+ !d2
⇣ = cos '
Rise Time in Frequency Domain
Suppose we want tr  c (c is some desired given value)
Rise Time in Frequency Domain
Suppose we want tr  c (c is some desired given value)

1.8 1.8
tr ⇡ c =) !n
!n c
Rise Time in Frequency Domain
Suppose we want tr  c (c is some desired given value)

1.8 1.8
tr ⇡ c =) !n
!n c

Geometrically, we want poles to lie in the shaded region:

Im

1.8
!n =
c
Re
0

(recall that !n is the magnitude of the poles)

Overshoot in Frequency Domain
Suppose we want Mp  c
Overshoot in Frequency Domain
Suppose we want Mp  c
!
⇡⇣
Mp = exp p c
1 ⇣2
| {z }
decreasing function
Overshoot in Frequency Domain
Suppose we want Mp  c
!
⇡⇣
Mp = exp p c — need large damping ratio
1 ⇣2
| {z }
decreasing function
Overshoot in Frequency Domain
Suppose we want Mp  c
!
⇡⇣
Mp = exp p c — need large damping ratio
1 ⇣2
| {z }
decreasing function

Geometrically, we want poles to lie in the shaded region:

Im

'
Re
0
Overshoot in Frequency Domain
Suppose we want Mp  c
!
⇡⇣
Mp = exp p c — need large damping ratio
1 ⇣2
| {z }
decreasing function

Geometrically, we want poles to lie in the shaded region:

Im
⇣ ! ⇣
p = pn
1 ⇣2 !n 1 ⇣ 2
= = cot '
' !d
Re
0
Overshoot in Frequency Domain
Suppose we want Mp  c
!
⇡⇣
Mp = exp p c — need large damping ratio
1 ⇣2
| {z }
decreasing function

Geometrically, we want poles to lie in the shaded region:

Im
⇣ ! ⇣
p = pn
1 ⇣2 !n 1 ⇣ 2
= = cot '
' !d
Re
0 — need ' to be small
Overshoot in Frequency Domain
Suppose we want Mp  c
!
⇡⇣
Mp = exp p c — need large damping ratio
1 ⇣2
| {z }
decreasing function

Geometrically, we want poles to lie in the shaded region:

Im
⇣ ! ⇣
p = pn
1 ⇣2 !n 1 ⇣ 2
= = cot '
' !d
Re
0 — need ' to be small
Intuition: good damping !
good decay in 1/2 period
Settling Time in Frequency Domain
Suppose we want ts  c
Settling Time in Frequency Domain
Suppose we want ts  c

3 3
ts ⇡ c =)
c
Settling Time in Frequency Domain
Suppose we want ts  c

3 3
ts ⇡ c =)
c

Want poles to be sufficiently fast (large enough magnitude of

real part):
Im

3 Intuition: poles far to the

=
c left ! transients decay
Re faster ! smaller ts
0
Combination of Specs
If we have specs for any combination of tr , Mp , ts , we can easily
relate them to allowed pole locations:
Im

The shape and size of the

region for admissible pole
locations will change
Re depending on which
0 specs are more severely
constrained.
Combination of Specs
If we have specs for any combination of tr , Mp , ts , we can easily
relate them to allowed pole locations:
Im

The shape and size of the

region for admissible pole
locations will change
Re depending on which
0 specs are more severely
constrained.

This is very appealing to engineers: easy to visualize things, no

such crisp visualization in time domain.
Combination of Specs
If we have specs for any combination of tr , Mp , ts , we can easily
relate them to allowed pole locations:
Im

The shape and size of the

region for admissible pole
locations will change
Re depending on which
0 specs are more severely
constrained.

This is very appealing to engineers: easy to visualize things, no

such crisp visualization in time domain.
But: not very rigorous, and also only valid for our prototype
2nd-order system, which has only 2 poles and no zeros ...