Sketching Bode Plots

by Hand

MECH 3140
Lecture #

1
Recall what a bode plot is

• It is the particular solution to a LTI

differential equation for a sinusoidal input
• It shows how the output of a system will
responds to different input frequencies
(including constant inputs, i.e. ω=0)
• It is plotted as gain and phase where gain
𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
is the and phase is the phase
𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
difference between the input sinusoid and
output sinusoid
2
Why know how to sketch?
• Matlab can plot a bode plot and calculate the gain and phase from any
transfer function
• So why do we need to know how to sketch (approximately draw) one?
– Its not to be a pain
– Its not because we don’t like matlab
• By understanding how to sketch a bode plot we can do two things as
systems engineer where we might not have the transfer function to begin
with
– Design
• If I am designing a system I don’t have a transfer function to put into matlab – I am
building the transfer function
• By understanding how to sketch a bode plot, I understand how to make a system
with a given response
– Analysis
• I might just be given data to look at and need to understand what the system might
be (i.e. be able to back out the transfer function from frequency response)
• I might be analyzing a specific phenomena and by understanding how to sketch a
bode plot I can understand what is happening or how to fix/change the phenomena.
3
Bode Sketching Rules
• You will notice that these sketching rules essentially
come from the gain and phase calculations
– The rules for 1st and 2nd order systems are “extended” by
understanding how gains multiply and phases add.
– From these relationships you will notice that the rule for a
pole is inverted (or flipped) for the rule from a zero
• Gain and phase of the numerator vs. denominator

𝑛𝑛𝑛𝑛𝑛𝑛(𝑗𝑗𝜔𝜔) 𝑎𝑎(𝑗𝑗𝑗𝑗) 𝑐𝑐 𝑗𝑗𝑗𝑗

𝐺𝐺 𝜔𝜔 = 𝐻𝐻(𝑗𝑗𝜔𝜔) = = �
𝑑𝑑𝑑𝑑𝑑𝑑(𝑗𝑗𝜔𝜔) 𝑏𝑏(𝑗𝑗𝑗𝑗) 𝑑𝑑 𝑗𝑗𝑗𝑗

∅ 𝜔𝜔 = ∠𝐻𝐻 𝑗𝑗𝑗𝑗 = ∠𝑛𝑛𝑛𝑛𝑛𝑛 𝑗𝑗𝑗𝑗 − ∠𝑑𝑑𝑑𝑑𝑑𝑑 𝑗𝑗𝑗𝑗 = ∠𝑎𝑎 𝑗𝑗𝑗𝑗 − ∠𝑏𝑏 𝑗𝑗𝑗𝑗 + ∠𝑐𝑐 𝑗𝑗𝑗𝑗 − ∠𝑑𝑑 𝑗𝑗𝑗𝑗

4
Bode Sketching Rules
• Draw your vertical and horizontal axis
– Label horizontal axis ω (rad/s)
– Label the top vertical axis Gain
– Label the bottom vertical axis Phase (or φ)
• Mark the magnitude of all zeros and poles on
the frequency axis
– It is sometimes helpful to draw the poles and zeros
on the s-plane (or at least calculate their
magnitudes and damping ratios if they are under-
damped)

5
Bode Sketching Rules (Gain)
• Start by sketching the asymptotes:
– Start from the DC Gain
• If the DC gain is 0 or ∞ then start with a line of correct
slope (depending on the number of zeros or poles at the
origin)
– Move in frequency until you hit the frequency (in
magnitude) of a pole or zero
• This is the absolute value of the pole or zero if it is real
• This is the ωn of the pole or zero if it is complex
– The slope then changes +1 for every zero and -1
for every pole
• Note this means the slope would change +2 for a
complex zero or -2 for a complex pole. 6
Bode Sketching Rules (Gain)
– Label the slope of the asymptotes
• Label the gain at DC and at the corners
– Use your slopes and start from the DC gain
• This requires you understand logarithmic plots (and
their “slopes”)
𝐺𝐺2 𝜔𝜔 𝜔𝜔
– �𝐺𝐺1 = (𝜔𝜔2)𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 ∴ 𝐺𝐺2 = 𝐺𝐺1 ( 2)𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝜔𝜔1
1

• If the DC Gain is 0 or ∞, then you must select another

portion of the graph to start your gain from
– Generally this means calculating the gain from the transfer
function for a particular frequency to use as your starting point

7
Bode Sketching Rules (Gain)

– Calculate (and label) the gain at the corners

• Real Pole: 0.707Gc
• Real Zero: 1.414Gc
𝐺𝐺𝑐𝑐
• Complex Poles:
2𝜁𝜁
• Complex Zeros: 2𝜁𝜁𝐺𝐺𝑐𝑐
• Draw the Gain line

8
Bode Sketching Rules (Phase)
• Start with the phase at ω=0
– This will be zero unless there is a pole or zero at
the origin
• In which case the phase will be
90*z (where z is the number of zeros at the origin)
-90*p (where p is the number of poles at the origin)
– The phase changes +90/zero and -90/pole at the
magnitude of the zero or pole
• Half of the phase change (+45/zero, -45/pole) occurs at
the magnitude of the eigenvalue

9
Comments

• Note that these sketching rules produce

very rough ideas for the bode plot (or the
frequency response)
– But the provide insight into design and
analysis.

10
Maxon DC16 Spec Sheet

https://www.maxongroup.com/medias/sys_master/root/8833376059422/19-EN-94.pdf 11
Example #1 (12 V Maxon DC 16 Motor)
𝜔𝜔(𝑠𝑠) 𝐾𝐾𝐼𝐼 0.015
= =
𝑉𝑉𝑠𝑠 (𝑠𝑠) 𝐿𝐿𝐿𝐿𝑠𝑠 2 + 𝐽𝐽𝐽𝐽 + 𝐿𝐿𝐿𝐿 𝑠𝑠 + (𝐾𝐾𝐼𝐼 𝐾𝐾𝑏𝑏 + 𝑅𝑅𝑅𝑅) 6.405𝑒𝑒 −11 𝑠𝑠 2 + 2.006𝑒𝑒 −6 𝑠𝑠 + 2.249𝑒𝑒 −4

• Two eigenvalues: S=-3.12e4 , -112

• τ1=0.032 ms, τ2=8.9 ms
• Note the dominant time constant is the “mechanical”
time constant on the data sheet.
• GDC=66.7 rad/sec/Volt (this matches the no load
speed on the data sheet).
Im
Re

-3.12e4 -112
12
Hand Sketch

13
>> bode(Ki,[J*L J*R+L*b Ki*Kb+R*b])

14
Example #2 (LRC Circuit)
𝐼𝐼(𝑠𝑠) 𝐶𝐶𝐶𝐶 0.01𝑠𝑠 L=1 H
= = R=100 Ω
𝑉𝑉(𝑠𝑠) 𝐿𝐿𝐿𝐿𝑠𝑠 + 𝑅𝑅𝑅𝑅𝑅𝑅 + 1 0.01𝑠𝑠 2 + 𝑠𝑠 + 1
2 C=10 mF

0.1𝑗𝑗 Im
𝐺𝐺 10 = = 0.01
10𝑗𝑗 + (1 − 1)

Re
-99 -1

15
Hand Sketch

16
>> bode([C 0],[L*C R*C 1])

17
Example #3 (PID Controller)

𝐹𝐹 𝑡𝑡 = 𝐾𝐾𝑝𝑝 𝑒𝑒 + 𝐾𝐾𝑑𝑑 𝑒𝑒̇ + 𝐾𝐾𝐼𝐼 � 𝑒𝑒𝑒𝑒𝑒𝑒

𝐹𝐹(𝑠𝑠) 𝐾𝐾𝐼𝐼 𝐾𝐾𝑑𝑑 𝑠𝑠 2 + 𝐾𝐾𝑝𝑝 𝑠𝑠 + 𝐾𝐾𝐼𝐼 𝑠𝑠 2 + 15𝑠𝑠 + 50

= 𝐾𝐾𝑝𝑝 + 𝐾𝐾𝑑𝑑 𝑠𝑠 + 𝑒𝑒 = =
𝑒𝑒(𝑠𝑠) 𝑆𝑆 𝑠𝑠 𝑠𝑠

Im

Re
-10 -5

18
Hand Sketch

-1 +1

• Notice positive phase shift (output “leads” input)

– This is a property of derivative control
• It “anticipates” the output by using the derivative
• However, notice the increasing gain
– Amplifies noise!
19
>> bode([1 15 50],[1 0])

20
Example #4 (Random System)
50 50
𝐻𝐻 𝑠𝑠 = 2 = 3
(𝑠𝑠 + 2𝑠𝑠 + 100)(𝑠𝑠 + 1) 𝑠𝑠 + 3𝑠𝑠 2 + 102 + 100

Im

99

10

Re
-1

− 99

21
Hand Sketch

22
>>bode(50,[1 3 102 100])

23
Example #5 (1/4 Car Model)

• Used to model vehicle ride (vertical motion

only)
• For this example, we will ignore the
damping
m1
x1
k1

m2 F
x2
k2

24
¼ Car Transfer Functions
𝑥𝑥1 (𝑠𝑠) 𝑘𝑘1
=
𝐹𝐹(𝑠𝑠) 𝑚𝑚1 𝑚𝑚2 𝑠𝑠 4 + (𝑚𝑚1 𝑘𝑘1 + 𝑚𝑚1 𝑘𝑘2 + 𝑚𝑚2 𝑘𝑘2 )𝑠𝑠 2 + 𝑘𝑘1 𝑘𝑘2
𝑥𝑥2 (𝑠𝑠) 𝑚𝑚1 𝑠𝑠 2 + 𝑘𝑘1
=
𝐹𝐹(𝑠𝑠) 𝑚𝑚1 𝑚𝑚2 𝑠𝑠 4 + (𝑚𝑚1 𝑘𝑘1 + 𝑚𝑚1 𝑘𝑘2 + 𝑚𝑚2 𝑘𝑘2 )𝑠𝑠 2 + 𝑘𝑘1 𝑘𝑘2

Im Im
60 60

6 6
Re Re

6 6

60 60
25
X1 Hand Bode Sketch (Car)

26
X2 Hand Bode Sketch (Wheel)

27
X1 Bode Plot (Car)

>>bode(k1,[m1*m2 0 m1*(k1+k2)+m2*k2 0 k1*k2])

28
X2 Bode Plot (Wheel)
>>bode([m1 0 k1],[m1*m2 0 m1*(k1+k2)+m2*k2 0 k1*k2])

29
Effect of Wheel Imbalance

• What is the force acting on the wheel due

to a wheel imbalance
– Same as the reaction forces on the pendulum
you did in a prior Matlab homework
Ry
� 𝐹𝐹𝑦𝑦 = 𝑚𝑚𝑦𝑦̈ = 𝑅𝑅𝑦𝑦 − 𝑚𝑚𝑚𝑚
ω ω
R θ 𝑦𝑦̈ = 𝑅𝑅𝜃𝜃̈ sin 𝜃𝜃 − 𝑅𝑅 𝜃𝜃̇ 2 cos(𝜃𝜃)
m
If ω=const
mg 𝑦𝑦̈ = −𝑅𝑅 𝜃𝜃̇ 2 cos(𝜃𝜃)
𝜃𝜃 = 𝜔𝜔𝜔𝜔 + 𝜃𝜃0

𝑅𝑅𝑦𝑦 = 𝑚𝑚𝑚𝑚 − 𝑚𝑚𝑚𝑚𝜔𝜔2 sin(ω𝑡𝑡 + 𝜃𝜃0 )

30
Effect of Wheel Imbalance

• It creates a sinusoidal force input!

– Frequency of the sinusoidal input is the speed
of the wheel
– Notice the magnitude of the input force
increases by frequency (i.e. speed) squared!
• 2nd order low pass filter would not remove

𝑅𝑅𝑦𝑦 = 𝑚𝑚𝑚𝑚 − 𝑚𝑚𝑚𝑚𝜔𝜔2 sin(ω𝑡𝑡 + 𝜃𝜃0 )

31
𝑥𝑥1 (Car position)
𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
• Remember the Bode Plot is
𝑖𝑖𝑖𝑖𝑜𝑜𝑜𝑜𝑜𝑜

32
𝑥𝑥2 (Wheel position)
𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
• Remember the Bode Plot is
𝑖𝑖𝑖𝑖𝑜𝑜𝑜𝑜𝑜𝑜

33
Calculating Gain & Phase Graphically

• Also, since the gain and phase are calculated by

evaluating the magnitude and phase of the transfer
function at jω, you can also do it graphically.
• Recall that jω is essentially the imaginary axis on
the s-plane
– Can find the gain and phase by calculating lengths from
the imaginary axis to the poles and zeros and the angles
from the poles and zeros to the frequency on the
imaginary axis:
𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝜔𝜔 𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑡𝑡𝑡𝑡 𝑎𝑎𝑎𝑎𝑎𝑎 𝑡𝑡𝑡𝑡𝑡 𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧
𝐺𝐺 𝜔𝜔 =
𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝜔𝜔 𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑡𝑡𝑡𝑡 𝑎𝑎𝑎𝑎𝑎𝑎 𝑡𝑡𝑡𝑡𝑡 𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧

∅ 𝜔𝜔 = � ∅𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧 − � ∅𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
34
Graphical Example #1

Im
𝑠𝑠 + 4 4
𝐻𝐻 𝑠𝑠 = 4 2
𝑠𝑠 + 3 5
ϕ2 ϕ1 Re
-4 -3

Φ1=53

Φ2=45
4 2
𝐺𝐺 4 =
5

𝜙𝜙 4 = −7 𝑑𝑑𝑑𝑑𝑑𝑑

35
Graphical Example #2

Im
ϕ1
𝑠𝑠 + 3 4
𝐻𝐻 𝑠𝑠 = 2
𝑠𝑠 + 12𝑠𝑠 + 52 5
ϕ3 Re
-6 -3
10
ϕ2 Φ1=0
-4
Φ2=53
5
𝐺𝐺 4 = = 0.0833
6 ∗ 10 Φ3=53

𝜙𝜙 4 = 0 𝑑𝑑𝑑𝑑𝑑𝑑

36