Lecture 6:

Frequency Response Bode

Plots and Nyquist Diagrams
Today’s Lecture
Frequency Response
Amplitude Ration and Phase
System Analysis using
Bode Plot
Nyquist Diagram
System Identification
Frequency Response
Up to now – Time response
How the system behaves to a specific input with respect to
time
Now – Frequency Response
System response to a sinusoidal input
Range of frequencies used
Used for system identification
Used for stability analysis (next lecture)
Frequency Response
System behavior determined from the steady state
response to a sinusoidal input in the form:
𝑅 = 𝐴 sin 𝜔𝑡
Sine wave used:
Easy to analyze
Easy to generate
Easy to measure experimentally
Frequency Response
Sinusoidal applied to LINEAR system:
Output will be sinusoidal
Output amplitude proportional to input amplitude
Harmonic input produces harmonic output at same frequency
Variation in Amplitude and Phase
Function of Frequency
Frequency Response
Input G(s) Output
Asint Bsin(t+)

A
B

𝐵
 Amplitude Ratio (AR) 𝐴𝑅 = (dimensionless)
𝐴
Frequency Response
Input G(s) Output
Asint Bsin(t+)

One Period = 360

Frequency Response
Input G(s) Output
Asint Bsin(t+)

 Phase Shift (angle)  (degrees)

 If output follows input → LAG → negative
Frequency Response
Data gained experimentally
Can be gained from OL transfer function
Input: 𝑅 = 𝐴 sin 𝜔𝑡
𝑅 can be represented as: 𝑅 = 𝐴𝑒 𝑗𝜔𝑡 (𝑗 = −1)
Take first derivative:
𝑑𝑅
𝑅 = 𝐴 sin 𝜔𝑡 → = 𝑅𝑠 = 𝑗𝜔𝐴 𝑗𝜔𝑡
𝑑𝑡
Replace s terms in TF with 𝑗𝜔
Moving from s-domain to frequency domain
Frequency Response
First order system:
1 1
𝐺 𝑠 = → 𝐺 𝑗𝜔 =
1 + 𝜏𝑠 1 + 𝑗𝜔𝜏

Second order system:

𝜔𝑛2 𝜔𝑛2
𝐺 𝑠 = 2 2 → 𝐺 𝑗𝜔 =
𝑠 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛 −𝜔 2 + 2𝜁𝜔𝑛 𝑗𝜔 + 𝜔𝑛2

Note:  is deriving frequency, n is natural frequency

 Frequency Response
Consider second order transfer function:
5
𝐺 𝑠 = 2
2𝑠 + 3𝑠 + 1

Substitute for s:
5
𝐺 𝑗𝜔 =
2 𝑗𝜔 2 + 3 𝑗𝜔 + 1 5
𝐺 𝑗𝜔 =
1 − 2𝜔 2 + 𝑗 3𝜔
5
𝐺 𝑗𝜔 =
−2𝜔 2 + 3𝑗𝜔 + 1
Frequency Response
Complex quantity that depends on frequency:
5
𝐺 𝑗𝜔 =
1 − 2𝜔 2 + 𝑗 3𝜔
Multiply by complex conjugate to separate real and
imaginary parts:
5 1 − 2𝜔2 − 𝑗 3𝜔
𝐺 𝑗𝜔 =
1 − 2𝜔 2 + 𝑗 3𝜔 1 − 2𝜔 2 − 𝑗 3𝜔
5 1 − 2𝜔2 − 𝑗 3𝜔
𝐺 𝑗𝜔 =
1 − 2𝜔 2 2 + 𝑗 3𝜔 1 − 2𝜔 2 − 𝑗 3𝜔 1 − 2𝜔 2 + +𝑗 3𝜔 −𝑗 3𝜔
 Frequency Response
5 1 − 2𝜔2 − 𝑗 3𝜔
𝐺 𝑗𝜔 =
1 − 2𝜔 2 2 + 𝑗 3𝜔 1 − 2𝜔 2 − 𝑗 3𝜔 1 − 2𝜔 2 + +𝑗 3𝜔 −𝑗 3𝜔

5 1 − 2𝜔2 − 𝑗 3𝜔
𝐺 𝑗𝜔 =
1 − 4𝜔 2 + 4𝜔 4 + 9𝜔 2

5 − 10𝜔2 15𝜔
𝐺 𝑗𝜔 = 4 2
−𝑗 4
4𝜔 + 5𝜔 + 1 4𝜔 + 5𝜔 2 + 1
Frequency Response
Complex number in Re + jIm form:
5 − 10𝜔2 15𝜔
𝐺 𝑗𝜔 = 4 2
−𝑗
4𝜔 + 5𝜔 + 1 4𝜔 4 + 5𝜔 2 + 1

Real part: 5 − 10𝜔2

𝑅𝑒 =
4𝜔 4 + 5𝜔 2 + 1

Imaginary part: 15𝜔

𝐼𝑚 = − 4
4𝜔 + 5𝜔 2 + 1
Frequency Response
From complex number theory:
Amplitude ratio:
𝐴𝑅 = 𝑅𝑒 2 + 𝐼𝑚2

Phase angle:
−1
𝐼𝑚
𝜙 = 𝑡𝑎𝑛
𝑅𝑒

Used to generate a table of results

Frequency Response
Frequency Frequency
Re Im AR Phase Re Im AR Phase
(red/s) (red/s)
0.01 4.996502 -0.14993 4.99875 -1.7187 0.5 1 -3 3.162278 -71.5649
0.02 4.986025 -0.2994 4.995006 -3.43636 1 -0.5 -1.5 1.581139 -108.435
0.03 4.968625 -0.44798 4.98878 -5.15198 2 -0.41176 -0.35294 0.542326 -139.398
0.05 4.913459 -0.74072 4.968979 -8.57298 3 -0.22973 -0.12162 0.259938 -152.102
0.1 4.66489 -1.42803 4.878571 -17.0205 5 -0.0933 -0.02856 0.097571 -162.979
0.2 3.812997 -2.48674 4.552232 -33.1113 10 -0.02457 -0.0037 0.024845 -171.427
0.3 2.765785 -3.03562 4.106646 -47.6629 20 -0.00622 -0.00047 0.00624 -175.705
Representing Data
Some methods
Nyquist Diagram
Bode Plot
Nyquist Diagram
Polar plot on Argand diagram of open loop frequency
locus (i.e. 𝐺 𝑗𝜔 𝐻 𝑗𝜔 )
Amplitude ratio is radial coordinate.
Phase angle is the angular coordinate
 >> nyquist(sys)
5 − 10𝜔2 15𝜔
𝐺 𝑗𝜔 = 4 2
−𝑗 4
4𝜔 + 5𝜔 2 + 1
Nyquist Diagram 4𝜔 + 5𝜔 + 1

Example
Frequency
AR Phase
(red/s)
0.01 4.99875 -1.7187
0.02 4.995006 -3.43636
0.03 4.98878 -5.15198
0.05 4.968979 -8.57298
0.1 4.878571 -17.0205
0.2 4.552232 -33.1113
0.3 4.106646 -47.6629
0.5 3.162278 -71.5649
1 1.581139 -108.435
2 0.542326 -139.398
3 0.259938 -152.102
5 0.097571 -162.979
10 0.024845 -171.427
20 0.00624 -175.705
 >> bode(sys)
5 − 10𝜔2 15𝜔
𝐺 𝑗𝜔 = 4 2
−𝑗
4𝜔 + 5𝜔 + 1 4𝜔 4 + 5𝜔 2 + 1
Bode Plots
Two logarithmic plots
Amplitude ratio vs frequency
Phase angle vs frequency
 Bode Plots
Amplitude measured in dB, converted by: 𝐴𝑅 𝑑𝐵 = 20 log 𝐴𝑅
Frequency
AR AR(db)
(red/s)
0.01 4.99875 13.97723
0.02 4.995006 13.97072
0.03 4.98878 13.95989
0.05 4.968979 13.92534
0.1 4.878571 13.76585
0.2 4.552232 13.16449
0.3 4.106646 12.26975
0.5 3.162278 10
1 1.581139 3.9794
2 0.542326 -5.31479
3 0.259938 -11.7026
5 0.097571 -20.2135
10 0.024845 -32.0953
20 0.00624 -44.096
 Bode Plots
Phase measured in degrees:
Frequency
Phase
(red/s)
0.01 -1.7187
0.02 -3.43636
0.03 -5.15198
0.05 -8.57298
0.1 -17.0205
0.2 -33.1113
0.3 -47.6629
0.5 -71.5649
1 -108.435
2 -139.398
3 -152.102
5 -162.979
10 -171.427
20 -175.705
Bode Plots
Constructing Bode plots
Sections of TF can be represented as straight lines = asymptotic
approximations
Example:
Bode Plots
Each part of TF has certain type of frequency response
Building blocks:
Gain 𝐾
Differentiator 𝑠
1
Integrator
𝑠 1
First order lead 1 + 𝜏𝑠 /lag
Second order lead/lag 1 + 𝜏𝑠
1
𝑠 2 2𝜁 𝑠 2 2𝜁
2+𝜔 𝑠+1 2+𝜔 𝑠+1
𝜔𝑛 𝑛 𝜔𝑛 𝑛
Bode Plots
Gain
𝐺 𝑠 =𝐾
𝐴𝑅 𝑑𝐵 = 20 log 𝐾
𝑃ℎ𝑎𝑠𝑒 = 0°

𝐺 𝑠 =5
Note 20log(5) = 13.9794
Bode Plots
Differentiator
𝐺 𝑠 =𝑠
+20 dB/decade rise
Phase = +90
Bode Plots
Integrator
𝐺 𝑠 = 1 𝑠
-20 dB/decade fall
Phase = -90
Bode Plots
1st order lead
𝐺 𝑠 = 1 + 𝜏𝑠
Conner frequency: 1 𝜏 +20 dB/decade rise
True curve 3dB above break
point

𝐺 𝑠 = 1 + 10𝑠
Bode Plots
1st order lag
1
𝐺 𝑠 =
1 + 𝜏𝑠 -20 dB/decade fall
Conner frequency: 1 𝜏
True curve 3dB below break
point

1
𝐺 𝑠 =
1 + 10𝑠
Bode Plots
2nd order lead
𝑠 2 2𝜁
𝐺 𝑠 = 2+ 𝑠+1 +40 dB/decade rise
𝜔𝑛 𝜔𝑛
Conner frequency: 𝜔𝑛
True curve at break point depend on 𝜁

𝐺 𝑠 = 0.006944𝑠 2 + 0.1𝑠 + 1
𝜁 = 0.6
𝜔𝑛 = 12
Bode Plots
2nd order lead
𝑠 2 2𝜁
𝐺 𝑠 = 2+ 𝑠+1 +40 dB/decade rise
𝜔𝑛 𝜔𝑛
Conner frequency: 𝜔𝑛
True curve at break point depend on 𝜁

𝐺 𝑠 = 0.006944𝑠 2 + 0.2𝑠 + 1
𝜁 = 1.2
𝜔𝑛 = 12
Bode Plots
2nd order lag
1
𝐺 𝑠 = 2
𝑠 2𝜁
2+𝜔 𝑠+1
𝜔𝑛 𝑛 -40 dB/decade fall
Conner frequency: 𝜔𝑛
True curve at break point depend on 𝜁

1
𝐺 𝑠 =
0.006944𝑠 2 + 0.2𝑠 + 1
𝜁 = 0.6
𝜔𝑛 = 12
Bode Plots
2nd order lag
1
𝐺 𝑠 = 2
𝑠 2𝜁
2+𝜔 𝑠+1
𝜔𝑛 𝑛 -40 dB/decade fall
Conner frequency: 𝜔𝑛
True curve at break point depend on 𝜁

1
𝐺 𝑠 =
0.006944𝑠 2 + 0.2𝑠 + 1
𝜁 = 1.2
𝜔𝑛 = 12
 𝐵 𝑠 5 5
= =
𝐸 𝑠 0.1𝑠 + 0.7𝑠 + 𝑠 𝑠 0.1𝑠 2 + 0.7𝑠 + 1
3 2
Bode Plots
Example:
Draw the Bode diagram for the following transfer function:
𝐵 𝑠 5 𝐵 𝑗𝜔 5
= =
𝐸 𝑠 𝑠 1 + 0.2𝑠 1 + 0.5𝑠 𝐸 𝑗𝜔 𝑗𝜔 1 + 0.2𝑗𝜔 1 + 0.05𝑗𝜔
𝐴𝑅 𝑑𝐵
= 20 log 5 − 20 log 𝜔 − 20 log 1 + 0.2𝜔 2 − 20 log 1 + 0.05𝜔 2

∅ = −90° − 𝑡𝑎𝑛−1 0.2𝜔 − 𝑡𝑎𝑛−1 0.05𝜔

Bode Plots
Tank each term separately:
(a) 𝑅1 = 20 log 5 = 14 𝑑𝐵
Bode Plots
Tank each term separately:
(b) 𝑅2 = −20 log 𝜔
Bode Plots
(c) 𝑅3 = −20 log 1 + 0.2𝜔 2

1 1
Break point at 𝜔 = = = 5 𝑟𝑎𝑑/𝑠
𝜏 0.2
for 𝜔 < 5 𝑟𝑎𝑑/𝑠 slope = 0 dB/dec
for 𝜔 > 5 𝑟𝑎𝑑/𝑠 slope = -20 dB/dec
Bode Plots
(d) 𝑅4 = −20 log 1 + 0.5𝜔 2

1 1
Break point at 𝜔 = = = 2 𝑟𝑎𝑑/𝑠
𝜏 0.5
for 𝜔 < 2 𝑟𝑎𝑑/𝑠 slope = 0 dB/dec
for 𝜔 > 2 𝑟𝑎𝑑/𝑠 slope = -20 dB/dec
Bode Plots
(d) Combining plots:

14 𝑑𝐵
−20 log 𝜔
𝜔 = 5 𝑟𝑎𝑑/𝑠
𝜔 = 2 𝑟𝑎𝑑/𝑠
Bode Plots
(d) Combining plots:

14 𝑑𝐵
−20 log 𝜔
𝜔 = 5 𝑟𝑎𝑑/𝑠
𝜔 = 2 𝑟𝑎𝑑/𝑠
Bode Plots
Phase lag given by:
∅ = −90° − 𝑡𝑎𝑛−1 0.2𝜔 − 𝑡𝑎𝑛−1 0.05𝜔
∅ = ∅1 + ∅2 + ∅3
 - rad/s ∅1 - deg ∅2 - deg ∅3 - deg ∅ - deg
1 -90 -11 -3 -104
5 -90 -45 -7 -149
10 -90 -63 -27 -180
20 -90 -76 -45 -211
50 -90 -84 -68 -242
System Identification
Can use Bode Plot to estimate transfer function
Use same rules as before
Example
System Identification
Step 1: Plot amplitude and phase data on Bode plot
Step 2: Fit asymptotic approximations to amplitude data
Slopes must be 0 dB/dec, ±20 dB/dec, ±40 dB/dec etc.
Approximations must pass through the data points at the
lowest and highest frequencies
Step 3: Use relationships to determine values for transfer
function
System Identification
Following experimental data obtained from a frequency
response test undertaken on control system
Freq – (rad/s) R - dB Phase - deg
2 8 -112
6 -2.5 -212
12 10 -152
30 26 -202
90 52 -245
200 72 -260
System Identification
System Identification
System Identification
System Identification
Amplitude: Asymptotic slopes
Low frequency = -20 dB/decade
High frequency = -60 dB/decade
No intermediate asymptotes
Low frequency amplitude and phase
 -20 dB/decade and phase → -90
1
Indicates a integrator term
𝑠
System Identification
High frequency amplitude and phase
 -60 dB/decade and phase → -270
Indicates a 3rd order system
Break point
Only one internal break point at intersection of low and high
frequency asymptotes
oSlope change = -40 dB/decade
oBreak point frequency = 20 rad/s
oDifference between point of intersection and true curve = -6 dB
Indicates two coincident 1st order lags
System Identification
System transfer function
Integrator and two coincident first order lags:
𝐾
𝑇𝐹 = 2
𝑠 1 + 𝜏𝑠
Transfer function terms
Break point frequency:
1 1 1
𝜔𝑏𝑝 = → 𝜏= = = 0.05 𝑠𝑒𝑐
𝜏 𝜔𝑏𝑝 20
System Identification
Transfer function terms
Integral term supposed to cross zero at 1 rad/s
In this case, integral term crosses at 5 rad/s → integrator is 14
dB higher
Therefore, gain: 20 log 𝐾 = 14 𝑑𝐵
log 𝐾 = 0.7
𝐾 = 100.7 = 5.01
5.01
𝑇𝐹 = 2
𝑠 1 + 0.05𝑠
Conclusion
Introduction to Frequency response
Determining Amplitude Ratio and Phase
Representing diagrammatically
Nyquist
Bode
Plotting frequency response
System Identification