Bode Plots & S domain analysis

9/29/2001 Shahriar Rokhsaz 1

Frequency Response
• One can analyze a given network using S
domain and find out the frequency response
of the network, by equating s=j*w
– Capacitance 1/C.s
– Resistance R
– Inductance L.s

9/29/2001 Shahriar Rokhsaz 2

 A simple Network
I1 = I 2
V in ( s ) − V out ( s ) R
I1 =
R Vout
I 2 = V out ( s ) ⋅ Cs
V in ( s ) = V out ⋅ ( 1 + R ⋅ C ⋅ s )
Vin 1/Cs
V out(s) 1
=
V in(s) (1 + R ⋅ C ⋅ s ) I1
V out(s) 1 I2
=
V in(s) s
(1 + )
p
Freq. Term
Where
1
p = Pole
R ⋅C

9/29/2001 Shahriar Rokhsaz 3

Magnitude
• To find the magnitude
– s=j*w
– If j term is in the denominator multiply the
function by its conjugate
– Find magnitude

9/29/2001 Shahriar Rokhsaz 4

 Magnitude
Vout (s ) 1
=  s = jω
Conjugate Vin (s ) 1 +
s
p
jω jω
1 − 1 −
1 p p
∗ =
jω jω ω
1 + 1 − 1 + ( ) 2
p p P
ω ω
1 + (− ) 2
1 + ( ) 2
p p
=
ω  ω 
2
1 + ( ) 2
1 + (
2
p ) 
 p 
Vout (s ) 1
=
Vin (s )  ω 
1 + (
2
) 
Magnitude  p 

9/29/2001 Shahriar Rokhsaz 5

Phase
• Phase is the inverse img
θ (ω ) = tag −1 ( )
tangent of the real
imaginary term −ω
devided by the real
p
term θ (ω ) = tag −1 ( )
1
−ω
θ (ω ) = tag −1 ( )
Radiant p

9/29/2001 Shahriar Rokhsaz 6

 Bode Plot
• Bode plot is the frequency response of the
amplitude
V (ω ) 1
| H ( j ω ) |= out
=
V in  ω 2 
 1 + ( ) 
 p 

• So lets sweep the frequency from 0 to

infinite

9/29/2001 Shahriar Rokhsaz 7

Frequency Response
• Consider increasing the frequency by steps
of 10*p
1
H (0 ) = = 1
1 + 0
p 1 1
H (ω = ) = = ≈ 1
10 2 1
 p  1 +
  100
1 +  10 
 p 
 
 
1 1 1
H (ω = p ) = = =
 p 
2 1 + 1 2
1 +  
 p 
1 1 1
H (ω = 10 ⋅ p ) = = =
 10 ⋅ p 
2 1 + 100 101
1 +  
 p 

9/29/2001 Shahriar Rokhsaz 8

 Nepers & Decibel
• Consider the following representation of a
function
G ( jω ) = M (ω ) ⋅ exp jφ (ω )

Take the Natural Log of both sides, we get the log gain in
neper
ln(G ( jω )) = ln( M (ω ) ⋅ exp jφ (ω ) )
ln(G ( jω )) = ln( M (ω )) + jφ (ω ) Phase is in rad

9/29/2001 Shahriar Rokhsaz 9

Nepers & Decibel

• Consider the following representation of a
function
G ( jω ) = M (ω ) ⋅ exp jφ (ω )
The gain in dB = 20log(M(w)), Phase is in deg

Decibels = 8.68 * neper

Degree = 57.3 * radians
9/29/2001 Shahriar Rokhsaz 10
 Back To The Example
• Consider increasing (w/p) >> 1
20log(H(0)) = 20log(1) = 0
1
20log(H(ω = 10⋅ p)) = 20log ≈ −20dB
101
1
20log(H(ω = 100⋅ p)) = 20log ≈ −40dB
1+100 2

Hence, a 20dB per decade is the slope

9/29/2001 Shahriar Rokhsaz 11

Bode Plot
• This is a log log based plot
– The Y axis is in terms of 20*log(H(w))
• dB
– The X axis increases in a log fashion
• Increments of 10*the previous frequency
– log(w0/10*w0)=log(1)- log(10)=0-1
• One Decade
– For every Decade you get a 20 dB of attenuation

9/29/2001 Shahriar Rokhsaz 12

 Amplitude
(p=2.106)
− 12
0
− 1.086×10
5

10
-20dB
15
20 log( H( 1j⋅ω ) )
20

25

30

− 33.981 35
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz

1decade 13

Bode Plot
H (ω )
2 0 lo g ( 1 )

20dB/ De c

p ω
9/29/2001 Shahriar Rokhsaz 14
 Phase Plot
• Phase starts from 0
• Starts rolling off at (w/10)
• Gets to -45 deg. at the pole
• decays to -90 deg 10*w
− 1 − ω
θ ( ω ) = tag ( )
p
− p
p
θ ( ) = tag − 1 ( 100 ) = tag − 1 (. 01 ) ≈ 0 deg
100 p
− p
θ ( p ) = tag − 1 ( ) = tag − 1 ( − 1 ) = − 45 deg
p
− 100 ⋅ p
θ ( 100 ⋅ p ) = tag − 1 ( ) = tag − 1 ( − 100 ) ≈ − 90 deg
p

9/29/2001 Shahriar Rokhsaz 15

Phase
(p=2.106)
0o
−5
0
− 2.865× 10

-45o
20

40

φ (ω )
60 -90o

80
2.106
− 88.861 100
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 16

 Amplitude & Phase
(p=2.106)
− 12
0
− 1.086×10

20

20 log( H( 1j⋅ω ) ) 40

φ (ω )
60

80

− 88.861 100
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 17

Phase
0 Deg. Phase shift
-90 Deg. Phase shift

θ (ω )

-45 Deg. Phase shift

p ω
9/29/2001 Shahriar Rokhsaz 18
 Phase Magnitude
H (ω )
θ ( ω )

2 0 lo g ( 1 )

20dB/ De c

-45 Deg.

ω
p
9/29/2001 Shahriar Rokhsaz 19

Poles and Zeros

Im a g i n a ry
Im a g i n a ry

L Zeros R
e i
f g
t h
t
Re al
Re al

Poles
9/29/2001 Shahriar Rokhsaz 20
 Poles and Zeros Im a g in a ry
Im a g in a ry

Positive Phase Shift

Re al
Re al

Negative Phase Shift Negative Phase Shift Unstable System

9/29/2001 Shahriar Rokhsaz 21

Zero
• Magnitude
– A zero is created if the amplitude increases with
frequency
• Phase
– If zero is in the left hand plane the phase
increases
– If zero is in the right hand plane the phase acts
like a pole

9/29/2001 Shahriar Rokhsaz 22

 Magnitude
• Consider
s
H ( s ) = 1 +
Z
j ⋅ ω
H ( s = j ⋅ ω ) = 1 +
Z
 ω 
2

H ( jω ) = 1 +  
 Z 
   
2 2
Z Z 1
H (ω = ) = 1 +   = 1 +   ≈ 1
100  Z ⋅ 100   100 
 Z 
2

(ω = = 1 +   = 1 + (1 ) =
2
H Z ) 2
 Z 
 100 ⋅ Z 
2

H (ω = 100 ⋅ Z ) = 1 +   = 1 + (100 )2 ≈ 100

 Z 

9/29/2001 Shahriar Rokhsaz 23

Phase
• Consider
s
H ( s ) = 1 +
Z
j ⋅ ω
H ( s = j ⋅ ω ) = 1 +
Z
θ (ω ) = tan − 1
(ω / Z )
θ (ω ) = tan (ω / Z )
− 1

Z 1
θ (ω = ) = tan − 1 ( ) ≈ 0
100 100
θ (ω = Z ) = tan − 1
( 1 ) = 45
θ (ω = 100 ⋅ Z ) = tan − 1
( 1 ) = 90

9/29/2001 Shahriar Rokhsaz 24

 Amplitude
(z=2.106)
35
33.981

30

25

20
20 log( H( 1j⋅ω ) )
15 +20dB/Dec
10

− 12
1.086×10 0
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 25

Phase
(z=2.106)
+90o 88.861
100

80
+45o
60

φ (ω )
40

0o
20

−5
2.865×10 0
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 26

 Amplitude & Phase
(z=2.106)
100
88.861

80

20 log( H( 1j⋅ω ) ) 60

φ (ω )
40

20

− 12
1.086×10 0
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 27

Magnitude & Phase

H(ω)
2 0 d B / D e c

2 0 lo g ( 1 )

ω
Z
+45 Shift
θ (ω )

0 Shift +90 Shift

9/29/2001 Shahriar Rokhsaz

Z ω 28
 Magnitude
• Consider
s
H (s) = 1 −
Z
j ⋅ ω
H (s = j ⋅ ω ) = 1 −
Z
 ω 
2

H ( jω ) = 1 +  
 Z 
   1 
2 2
Z Z
H (ω = ) = 1 +   = 1 +   ≈ 1
100  Z ⋅ 100   100 
 Z 
2

H (ω = Z ) = 1 +   = 1 + (1 ) =
2
2
 Z 
 100 ⋅ Z 
2

H ( ω = 100 ⋅ Z ) = 1 +   = 1 + (100 ) 2 ≈ 10
 Z 

9/29/2001 Shahriar Rokhsaz 29

Phase
• Consider
s
H (s) = 1 −
Z
j ⋅ω − j ⋅ω
H (s = j ⋅ω ) = 1 − = 1 +
Z Z
θ ( ω ) = tan −1
(− ω / Z )
Z − 1
θ (ω = ) = tan − 1 ( ) ≈ 0
100 100
θ (ω = Z ) = tan − 1 ( − 1 ) = − 45
θ (ω = 100 ⋅ Z ) = tan −1
( − 1 ) = − 90

9/29/2001 Shahriar Rokhsaz 30

 Amplitude
(z=-2.106)
35
33.981

30

25

20
20 log( H( 1j⋅ω ) )
15 +20dB/Dec
10

− 12
1.086×10 0
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 31

Phase
(z=-2.106)
0o
−5
0
− 2.865× 10

-45o
20

40

φ (ω )
60 -90o

80
2.106
− 88.861 100
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 32

 Amplitude & Phase
(z=-2.106)
40
33.981

20

20 log( H( 1j⋅ω ) ) 20

φ (ω )
40

60

80

− 88.861 100
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 33

Magnitude & Phase

H(ω)
2 0 d B / D e c

θ (ω )
2 0 lo g ( 1 )
-90 Deg. Phase shift

0 Deg. Phase shift Z ω

θ (ω )

-45 Deg. Phase shift

ω
Z
9/29/2001 Shahriar Rokhsaz 34
 Canceling Effects
Zero s
1+
Vout ( s ) z
=
Vin ( s ) s
1+
p
Pole

• If z=p, then the pole and zeros will cancel each

other both is phase and amplitude

9/29/2001 Shahriar Rokhsaz 35

Amplitude & Phase

(z=p=2.106)
1
1

0.5

20 log( H( 1j⋅ω ) )
φ (ω )
0

0.5

−1 1
100 1 .10 1 .10 1 .10 1 .10 1 .10 1 .10
3 4 5 6 7 8
1 10
1 ω 10×10
7

9/29/2001 Shahriar Rokhsaz 36

 Two Pole System
• What if we have a two 1
pole system, and the H (s) =
s s
(1 + ) ⋅ (1 +
poles are not real?!?! p0 p1 )
1
=
s2 1 1
(1 + + s⋅( + )
p 0 ⋅ p1 p0 p1
ωn2
=
ωn
s2 + ⋅ s + ωn2
Q

9/29/2001 Shahriar Rokhsaz 37

Two Pole System

• Solve for the roots

ωn2
H (s) =
ωn
s2 + ⋅ s + ωn2
Q
ω n 
2

  − 4 ⋅ ωn2
ωn Q 
s 1, 2 = − ±
9/29/2001
2Q Shahriar Rokhsaz 2 38
 Two Pole System
• If below holds, the
roots are imaginary

ωn 
2

  − 4 ⋅ωn2 < 0
Q 
Q > 0 .5

9/29/2001 Shahriar Rokhsaz 39

Poles
ω n jω n  1 

− ± ⋅ − 4
2Q 2  Q2 
 
Re al 2 + Im 2
2
2  
 ωn   ω n  1
  + ⋅ − 4 
 2  2 
 2 Q    Q 

9/29/2001 Shahriar Rokhsaz 40

 Poles
Im a g in a ry

ω n

jω n  1 
⋅ − 4
2 Q2 
Re al

ωn
2Q

9/29/2001 Shahriar Rokhsaz 41

Stability vs. Pole Location

• A second order system with real poles will
be stable given p1>>p2
• A second order system with imaginary
poles will have the following solution
s1,2 = σ ± jωn and will have a step response of

v ( t ) = exp − tσ
(exp jω n t
+ exp − jω n t
)
v ( t ) = exp − tσ
cos( ω n t )
9/29/2001 Shahriar Rokhsaz 42
 Stability vs. Pole Location
v ( t ) = exp − tσ
(exp jω n t
+ exp − jω n t
)
v ( t ) = exp − tσ
cos( ω n t )
• if σ > 0

We have a ringing with an exponential decay

factor
• if σ < 0

We have a ringing with an exponential growing

oscillation
9/29/2001 Shahriar Rokhsaz 43

Stability vs. Pole Location

v ( t ) = exp − tσ
(exp jω n t
+ exp − jω n t
)
v ( t ) = exp − tσ
cos( ω n t )
• if σ = 0
We have oscillation

9/29/2001 Shahriar Rokhsaz 44

 Poles & Stability
Ima g ina ry

Re a l

9/29/2001 Shahriar Rokhsaz 45