EECS 16B Spring 2020 HW 3

This homework is due on Wednesday, February 19, 2020, at

11:59PM.

Self-grades are due on Monday, February 24, 2019, at 11:59PM.

1 Complex Numbers
A common way to visualize complex numbers is to use the complex plane.
Recall that a complex number 𝑧 is often represented in Cartesian form.

𝑧 = 𝑥 + 𝑗 𝑦 with Re{𝑧} = 𝑥 and Im{𝑧} = 𝑦

See Figure 1 for a visualization of 𝑧 in the complex plane.

Im{𝑧}

𝑦 𝑧
𝑟

𝜃 𝑥
Re{𝑧}

Figure 1: Complex Plane

In this question, we will derive the polar form of a complex number and
use this form to make some interesting conclusions.

a) Calculate the length of 𝑧 in terms of 𝑥 and 𝑦 as shown in Figure 1. This

is the magnitude of a complex number and is denoted by |𝑧| or 𝑟.
(Hint: Use the Pythagorean theorem.)

1
 HW 3 @ 2020-02-21 09:07:28-08:00

Solution

p
𝑟= 𝑥 2 + 𝑦 2 = |𝑧|

b) Represent 𝑥, the real part of 𝑧, and 𝑦, the imaginary part of 𝑧, in terms

of 𝑟 and 𝜃.

Solution

𝑥 = 𝑟 cos(𝜃) and 𝑦 = 𝑟 sin(𝜃)

c) Substitute for 𝑥 and 𝑦 in 𝑧. Use Euler’s identity1 𝑒 𝑗𝜃 = cos 𝜃 + 𝑗 sin 𝜃 to

conclude that,

𝑧 = 𝑟𝑒 𝑗𝜃 .

Solution

𝑧 = 𝑟 cos(𝜃) + 𝑗𝑟 sin(𝜃)
= 𝑟(cos(𝜃) + 𝑗 sin(𝜃))
= 𝑟𝑒 𝑗𝜃

d) In the complex plane, sketch the set of all the complex numbers such
that |𝑧| = 1. What are the 𝑧 values where the sketched figure intersects
the real axis and the imaginary axis?

1 also known as de Moivre’s Theorem.

2
EECS 16B Spring 2020 HW 3

Solution

Im{𝑧}

−1 1
Re{𝑧}

−𝑗

e) If 𝑧 = 𝑟𝑒 𝑗𝜃 , prove that 𝑧¯ = 𝑟𝑒 −𝑗𝜃 . Recall that the complex conjugate of a

complex number 𝑧 = 𝑥 + 𝑗 𝑦 is 𝑧¯ = 𝑥 − 𝑗 𝑦.

Solution

𝑧¯ = (𝑟(cos(𝜃) + 𝑗 sin(𝜃)))
= 𝑟(cos(𝜃) − 𝑗 sin(𝜃))
= 𝑟(cos(−𝜃) + 𝑗 sin(−𝜃))
= 𝑟𝑒 −𝑗𝜃

f) Show (by direct calculation) that,

𝑟 2 = 𝑧 𝑧¯ .

Solution
𝑧 𝑧¯ = 𝑟𝑒 𝑗𝜃 𝑟𝑒 −𝑗𝜃 = 𝑟 2 𝑒 𝑗𝜃−𝑗𝜃 = 𝑟 2 𝑒 0 = 𝑟 2

3
 HW 3 @ 2020-02-21 09:07:28-08:00

2 RLC Responses: Initial Part

Consider the following circuit like you saw in lecture:

𝐶
𝑡=0 𝑅 𝐿
𝐼𝐿
+ − + − + −
𝑉𝐶 𝑉𝑅 𝑉𝐿

+
𝑉𝑠 𝑡=0
−

Assume the circuit above has reached steady state for 𝑡 < 0. At time 𝑡 = 0,
the switch changes state and disconnects the voltage source, replacing it with a
short.
In this problem, the current through the inductor and the voltage across the
capacitor are the natural physical state variables since these are what correlate
to how energy is actually stored in the system. (A magnetic field through the
inductor and an electric field within the capacitor.)

a) Write the system of differential equations in terms of state variables

𝑥1 (𝑡) = 𝐼𝐿 (𝑡) and 𝑥2 (𝑡) = 𝑉𝐶 (𝑡) that describes this circuit for 𝑡 ≥ 0. Leave
the system symbolic in terms of 𝑉𝑠 , 𝐿, 𝑅, and 𝐶.

Solution
For this part, we need to find two differential equations, each including a
derivative of one of the state variables.
First, let’s consider the capacitor equation 𝐼𝐶 (𝑡) = 𝐶 𝑑𝑡𝑑 𝑉𝐶 (𝑡). In this circuit,
𝐼𝐶 (𝑡) = 𝐼𝐿 (𝑡), so we can write

𝑑
𝐼𝐶 (𝑡) = 𝐶 𝑉𝐶 (𝑡) = 𝐼𝐿 (𝑡) (1)
𝑑𝑡
𝑑 1
𝑉𝐶 (𝑡) = 𝐼𝐿 (𝑡). (2)
𝑑𝑡 𝐶
If we use the state variable names, we can write this as
𝑑 1
𝑥2 (𝑡) = 𝑥1 (𝑡), (3)
𝑑𝑡 𝐶
so now we have one differential equation.
For the other differential equation, we can apply KVL around the single
loop in this circuit. (Alternatively, we could just solve it directly and

4
EECS 16B Spring 2020 HW 3

substitute in for the desired voltage on the capacitor, which is a state

variable.) Going clockwise, we have

𝑉𝐶 (𝑡) + 𝑉𝑅 (𝑡) + 𝑉𝐿 (𝑡) = 0. (4)

Using Ohm’s Law and the inductor equation 𝑉𝐿 = 𝐿 𝑑𝑡𝑑 𝐼𝐿 (𝑡), we can write
this as
𝑑
𝑉𝐶 (𝑡) + 𝑅𝐼𝐿 (𝑡) + 𝐿 𝐼𝐿 (𝑡) = 0, (5)
𝑑𝑡
which we can rewrite as
𝑑 𝑅 1
𝐼𝐿 (𝑡) = − 𝐼𝐿 (𝑡) − 𝑉𝐶 (𝑡). (6)
𝑑𝑡 𝐿 𝐿
If we use the state variable names, this becomes
𝑑 𝑅 1
𝑥1 (𝑡) = − 𝑥 1 (𝑡) − 𝑥2 (𝑡), (7)
𝑑𝑡 𝐿 𝐿
and we have a second differential equation.
To summarize the final system is
𝑑 𝑅 1
𝑥 1 (𝑡) = − 𝑥1 (𝑡) − 𝑥2 (𝑡) (8)
𝑑𝑡 𝐿 𝐿
𝑑 1
𝑥 2 (𝑡) = 𝑥1 (𝑡). (9)
𝑑𝑡 𝐶

b) Write the system of equations in vector/matrix form with the vector state
𝑥 1 (𝑡)

variable 𝑥®(𝑡) = . This should be in the form 𝑑𝑡𝑑 𝑥®(𝑡) = 𝐴 𝑥®(𝑡) with a
𝑥 2 (𝑡)
2 × 2 matrix 𝐴.

Solution
By inspection from the previous part, we have
" #
𝑑
𝑑𝑡 𝑥 1 (𝑡) − 𝑅𝐿 𝑥1 (𝑡)
  
− 1𝐿
𝑑 = , (10)
𝑑𝑡 𝑥 2 (𝑡)
1
𝐶 0 𝑥2 (𝑡)
𝑑
which is in the form 𝑑𝑡 𝑥
®(𝑡) = 𝐴 𝑥®(𝑡), with

− 𝑅𝐿
 
− 1𝐿
𝐴= 1 . (11)
𝐶 0

c) Find the eigenvalues of the 𝐴 matrix symbolically.

(Hint: the quadratic formula will be involved.)

5
 HW 3 @ 2020-02-21 09:07:28-08:00

Solution
To find the eigenvalues, we’ll solve det(𝐴 − 𝜆𝐼) = 0. In other words, we
want to find 𝜆 such that
!
− 𝑅𝐿 − 𝜆

− 1𝐿
det(𝐴 − 𝜆𝐼) = det 1 (12)
𝐶 −𝜆
𝑅
 
1
= −𝜆 − − 𝜆 + (13)
𝐿 𝐿𝐶
𝑅 1
= 𝜆2 + 𝜆 + = 0. (14)
𝐿 𝐿𝐶
The Quadratic Formula gives
s 
2
1𝑅 1 𝑅 4
𝜆=− ± − . (15)
2𝐿 2 𝐿 𝐿𝐶

d) Under what condition on the circuit parameters 𝑅, 𝐿, 𝐶 are there going

to be a pair of distinct purely real eigenvalues of 𝐴?

Solution
For both eigenvalues to be real and distinct, we need the quantity inside
the square root to be positive. In other words, we need

𝑅2 4
− > 0, (16)
𝐿2 𝐿𝐶
or, equivalently,
r
𝐿
𝑅>2 . (17)
𝐶

e) Under what condition on the circuit parameters 𝑅, 𝐿, 𝐶 are there going

to be a pair of purely imaginary eigenvalues of 𝐴?

Solution
The only way for both eigenvalues to be purely imaginary is to have 𝑅 = 0.
In this case, the eigenvalues would be
r
1
𝜆 = ±𝑗 . (18)
𝐿𝐶

6
EECS 16B Spring 2020 HW 3

f) Assuming that the circuit parameters are such that there are a pair of
(potentially complex) eigenvalues 𝜆1 , 𝜆2 so that 𝜆1 ≠ 𝜆2 , find eigenvectors
𝑣®𝜆1 , 𝑣®𝜆2 corresponding to them.
(HINT: Rather than trying  to find the relevant nullspaces, etc., try to find
1
eigenvectors of the form where we just want to find the missing entry 𝑦. Can
𝑦
you see from the structure of the 𝐴 matrix why we might want to try that guess?)

Solution
The easy way is just to remember what an eigenvector is. We want
𝐴®
𝑣𝜆𝑖 = 𝜆1 𝑣®𝜆𝑖 . So, we can try to follow the hint:

− 𝑅𝐿 𝜆𝑖
      
− 𝐿1 1 1
= 𝜆𝑖 = (19)
1
𝐶 0 𝑦 𝑦 (𝜆 𝑖 )(𝑦)

We also know that:

𝑦
− 𝑅𝐿 − 𝑅𝐿 −
   
− 1𝐿 1 𝐿
= (20)
1
𝐶 0 𝑦 1
𝐶

Equating the two equations from above gives:

𝑦
𝜆𝑖 − 𝑅𝐿 −
  
= 𝐿 (21)
1
(𝜆 𝑖 )(𝑦) 𝐶

From the second row we see that 𝑦 = 1

𝜆𝑖 𝐶 . Now we find the eigenvectors
as:  
1
𝑣®𝜆1 = 1
𝜆1 𝐶

 
1
𝑣®𝜆2 = 1
𝜆2 𝐶

Alternatively, you can try to use the standard approach of finding the
nullspace of 𝐴 − 𝜆 𝑖 𝐼 to arrive at the same answer as above.

g) Assuming circuit parameters such that the two eigenvalues of 𝐴 are

distinct, let 𝑉 = [®𝑣𝜆1 , 𝑣®𝜆2 ] be a specific eigenbasis. Consider a coordinate
system for which we can write 𝑥®(𝑡) = 𝑉 e 𝑥® (𝑡). What is the 𝐴
e so that
𝑑 ® ®
𝑑𝑡 𝑥 (𝑡) = 𝐴e
e e 𝑥 (𝑡)? It is fine to have your answer expressed symbolically
using 𝜆1 , 𝜆2 .

7
 HW 3 @ 2020-02-21 09:07:28-08:00

Solution
𝑉 is given by:

 
1 1
𝑉= 1 1
𝜆1 𝐶 𝜆2 𝐶

We know that 𝑉 transforms from the 𝑥˜ coordinate frame to the 𝑥 coordinate

frame, 𝑉 −1 transforms back, and 𝐴 takes gives the relationship from 𝑥 to
𝑑
𝑑𝑡 𝑥.
𝑑
𝑥 to
Therefore to go from e 𝑥:
𝑑𝑡 e

 −1 
− 𝑅𝐿
  
e = 𝑉 −1 𝐴𝑉 = 1 1 − 1𝐿 1 1
𝐴 1 1 1 1 1
𝜆1 𝐶 𝜆2 𝐶 𝐶 0 𝜆1 𝐶 𝜆2 𝐶

" #
𝜆1 𝜆2 𝐶 𝜆21𝐶 − 𝑅𝐿
 
−1 −1 − 1𝐿 1 1
𝐴
e= 𝑉 𝐴𝑉 = 1 1 1 1
𝜆 1 − 𝜆 2 − 𝜆1 𝐶 1 𝐶 0 𝜆1 𝐶 𝜆2 𝐶

𝜆1
 
e = 𝑉 −1 𝐴𝑉 = 0
𝐴
0 𝜆2

You didn’t have to multiply things out explicitly. You could have just
noticed that the 𝐴 matrix times the 𝑉 matrix would give columns that
were 𝜆 𝑖 times 𝑣® 𝑖 each, and then multiplying that by 𝑉 −1 would just pick
out the 𝜆 𝑖 on the diagonals and zeros on the off-diagonals since 𝑉 −1𝑉 = 𝐼.

3 RLC Responses: Overdamped Case

Building on the previous problem, consider the following circuit with specified
component values:

10 nF
𝑡=0 1 kΩ 25 µH
𝐼𝐿
+ − + − + −
𝑉𝐶 𝑉𝑅 𝑉𝐿

+
𝑉𝑠 𝑡=0
−

8
EECS 16B Spring 2020 HW 3

Assume the circuit above has reached steady state for 𝑡 < 0. At time 𝑡 = 0,
the switch changes state and disconnects the voltage source, replacing it with a
short.
For this problem, we use the same notations as in Problem 2.

a) Suppose 𝑅 = 1 kΩ and the other component values are as specified in

𝑥® (0).
the circuit. Assume that 𝑉𝑠 = 1 Volt. Find the initial conditions for e
®
𝑥 is in the changed “nice” eigenbasis coordinates from the first
Recall that e
problem.

Solution
First of all, we must state the initial conditions for 𝑥®(0). If the circuit is in
steady state before 𝑡 = 0, then no current is flowing and the entire voltage
drop is across the capacitor. Therefore:

𝑥1 (0) = 𝐼𝐿 (0) = 0
𝑥2 (0) = 𝑉𝐶 (0) = 𝑉𝑠 = 1

Under these conditions, we can solve for

𝜆1 = −1.0 × 105 , 𝜆2 = −4.0 × 107

 
−1 −0.0025 −0.001
𝑉 =
1.0025 0.001
    
−0.0025 −0.001 0 −0.001
𝑥® (0) = 𝑉 −1 𝑥®(0) = =
1.0025 0.001 1 0.001
e

b) Continuing the previous part, find 𝑥1 (𝑡) = 𝐼𝐿 (𝑡) and 𝑥2 (𝑡) = 𝑉𝐶 (𝑡) for
𝑡 ≥ 0.

Solution
Plugging in for the component values gives:
 
−1.0 × 105 0
𝐴
e=
0 −4.0 × 107

These eigenvalues are the negative reciprocals of the relevant time con-
stants for these modes.
" #
𝑑
𝑥1 (𝑡) 𝑥1 (𝑡)
  
𝑑𝑡 e −1.0 × 105 0
, (22)
e
𝑑 =
𝑑𝑡 𝑥 2 (𝑡)
e 0 −4.0 × 107 𝑥2 (𝑡)
e

9
 HW 3 @ 2020-02-21 09:07:28-08:00

Therefore:

𝑥1 (𝑡) = 𝐾 1 𝑒 −1.0×10 𝑡
5
e
𝑥2 (𝑡) = 𝐾 2 𝑒 −4.0×10 𝑡
7
e

Solving for 𝐾 with the initial condition gives:

𝑥1 (𝑡) = −0.001𝑒 −1.0×10 𝑡

5
e
𝑥 2 (𝑡) = 0.001𝑒 −4.0×10 𝑡
7
e

® coordinates:
Converting back to the (𝑥)
 
1 1 ®
𝑥® (𝑡) =
𝑥®(𝑡) = 𝑉 e 𝑥 (𝑡)
−1000 −2.5
e

𝑥1 (𝑡) = −0.001𝑒 −1.0×10 𝑡 + 0.001𝑒 −4.0×10 𝑡

5 7

𝑥2 (𝑡) = 𝑒 −1.0×10 𝑡 − 0.0025𝑒 −4.0×10 𝑡

5 7

c) In the provided Jupyter notebook, move the sliders to approximately

𝑅 = 1𝑘Ω and 𝐶 = 10𝑛𝐹. Sketch 𝑉𝑐 (𝑡) and comment on its appearance.
Additionally, sketch the location of the eigenvalues on the complex plane.

Solution
𝑉𝑐 (𝑡) should look like a decaying exponential. The eigenvalues lie on the
real axis at coordinates (−1 × 105 , 0) and (−4 × 107 , 0).

4 RLC Responses: Undamped Case

Building on the previous problem, consider the following circuit with specified
component values:

10 nF
𝑡=0 0Ω 25 µH
𝐼𝐿
+ − + − + −
𝑉𝐶 𝑉𝑅 𝑉𝐿

+
𝑉𝑠 𝑡=0
−

10
EECS 16B Spring 2020 HW 3

Assume that the capacitor is charged to 𝑉𝑠 and there is no current in the

inductor for 𝑡 < 0. At time 𝑡 = 0, the switch changes state and disconnects the
voltage source, replacing it with a short.
For this problem, we use the same notations as in Problem 2.

a) Suppose 𝑅 = 0 kΩ and the other component values are as specified in

the circuit. Assume that 𝑉𝑠 = 1 Volt. Find the initial conditions for e𝑥® (0).
𝑥® is in the changed “nice” eigenbasis coordinates from the first
Recall that e
problem.

Solution
q q
Under these conditions, we can solve for 𝜆 = ±𝑗 1
𝐿𝐶 = ±𝑗 1
250×10−15
=
6
±𝑗2 × 10
𝜆1 = 𝑗2 × 106 , 𝜆2 = −𝑗2 × 106

Using the rule we derived earlier for finding 𝑉, we have

 
1 1
𝑉=
−50𝑗 50𝑗
 
0.5 0.01𝑗
𝑉 −1 =
0.5 −0.01𝑗

which lets us say

    
0.5 0.01𝑗 0 0.01𝑗
𝑥® (0) = 𝑉 −1 𝑥®(0) = =
0.5 −0.01𝑗 1 −0.01𝑗
e

b) Continuing the previous part, find 𝑥1 (𝑡) = 𝐼𝐿 (𝑡) and 𝑥2 (𝑡) = 𝑉𝐶 (𝑡) for
𝑡 ≥ 0.

Solution
Plugging in for the component values gives:

𝑗2 × 106
 
0
𝐴=
e
0 −𝑗2 × 106
" #
𝑑
𝑥1 (𝑡) 𝑗2 × 106 𝑥1 (𝑡)
  
𝑑𝑡 e 0
, (23)
e
𝑑 =
𝑑𝑡 𝑥 2 (𝑡)
e 0 −𝑗2 × 106 𝑥2 (𝑡)
e

Therefore:

11
 HW 3 @ 2020-02-21 09:07:28-08:00

6
𝑡
𝑥1 (𝑡) = 𝐾 1 𝑒 +𝑗 2×10
e
6
𝑡
𝑥2 (𝑡) = 𝐾 2 𝑒 −𝑗 2×10
e

Solving for 𝐾 with the initial condition gives:

6
𝑡
𝑥1 (𝑡) = 0.01𝑗𝑒 +𝑗 2×10
e
6
𝑡
𝑥2 (𝑡) = −0.01𝑗𝑒 −𝑗 2×10
e

® coordinates:
Converting back to the (𝑥)

" #
0.01𝑗𝑒 +𝑗 2×10 𝑡
 6
1 1
𝑥® (𝑡) =
𝑥®(𝑡) = 𝑉 e
−0.01𝑗𝑒 −𝑗 2×10 𝑡
6
−50𝑗 5 × 50𝑗

 
𝑥1 (𝑡) = 0.01𝑗𝑒 +𝑗 2×10 𝑡 − 0.01𝑗𝑒 −𝑗 2×10 𝑡
6 6
= −0.02 sin 2 × 106 𝑡

 
𝑥2 (𝑡) = 0.5𝑒 +𝑗 2×10 𝑡 + 0.5𝑒 −𝑗 2×10 𝑡
6 6
= cos 2 × 106 𝑡

c) In the provided Jupyter notebook, move the sliders to approximately

𝑅 = 0Ω and 𝐶 = 10𝑛𝐹. Sketch 𝑉𝑐 (𝑡) and comment on its appearance.
Additionally, sketch the location of the eigenvalues on the complex plane.
Are the waveforms for 𝑥1 (𝑡) and 𝑥2 (𝑡) “transient” — do they die out with
time?
Note: Because there is no resistance, this is called the “undamped” case.

Solution
No, these waveforms are sinusoids and do not die out over time. They
are not transient. The eigenvalues are located on the imaginary axis at
coordinates (0, −2 × 106 ) and (0, 2 × 106 ).

5 RLC Responses: Underdamped Case

Building on the previous problem, consider the following circuit with specified
component values:

12
EECS 16B Spring 2020 HW 3

10 nF
𝑡=0 1Ω 25 µH
𝐼𝐿
+ − + − + −
𝑉𝐶 𝑉𝑅 𝑉𝐿

+
𝑉𝑠 𝑡=0
−

Assume the circuit above has reached steady state for 𝑡 < 0. At time 𝑡 = 0,
the switch changes state and disconnects the voltage source, replacing it with a
short.
For this problem, we use the same notations as in Problem 2. You may
round numbers to make the algebra more simple.

a) Now suppose that 𝑅 = 1 Ω and the other component values are as specified
𝑥® (0).
in the circuit. Assume that 𝑉𝑠 = 1 Volt. Find the initial conditions for e
®
𝑥 is in the changed “nice” eigenbasis coordinates from the first
Recall that e
problem.

Solution
Under these conditions, we can solve for

𝜆1 = −0.02 × 106 + 𝑗2 × 106 , 𝜆2 = −0.02 × 106 − 𝑗2 × 106

" #
1 1
𝑉= 1 1
−0.0002+𝑗 0.02 −0.0002−𝑗 0.02

𝑗0.01
 
𝑥® (0) = 𝑉 −1 𝑥®(0) =
−𝑗0.01
e

b) Continuing the previous part, find 𝑥1 (𝑡) = 𝐼𝐿 (𝑡) and 𝑥2 (𝑡) = 𝑉𝐶 (𝑡) for
𝑡 ≥ 0.
(HINT: Remember that 𝑒 𝑎+𝑗𝑏 = 𝑒 𝑎 𝑒 𝑗𝑏 .)

Solution
−0.02 × 106 + 𝑗2 × 106
 
0
𝐴=
−0.02 × 106 − 𝑗2 × 106
e
0

13
 HW 3 @ 2020-02-21 09:07:28-08:00

" #
𝑑
𝑥1 (𝑡) −0.02 × 106 + 𝑗2 × 106 𝑥1 (𝑡)
  
𝑑𝑡 e 0
,
e
𝑑 =
𝑑𝑡 𝑥 2 (𝑡)
e 0 −0.02 × 106 − 𝑗2 × 106 𝑥2 (𝑡)
e
(24)
Therefore:

𝑥1 (𝑡) = 𝐾 1 𝑒 (−0.02×10
6
+𝑗 2×106 )𝑡
e
𝑥2 (𝑡) = 𝐾 2 𝑒 (−0.02×10
6
−𝑗 2×106 )𝑡
e
Solving for 𝐾 with the initial condition gives:

𝑥1 (𝑡) = 𝑗0.01𝑒 (−0.02×10

6
+𝑗 2×106 )𝑡
e
𝑥2 (𝑡) = −𝑗0.01𝑒 (−0.02×10
6
−𝑗 2×106 )𝑡
e
® coordinates:
Converting back to the (𝑥)
 
1 1
𝑥® (𝑡) =
𝑥®(𝑡) = 𝑉 e 𝑥® (𝑡)
−0.5 − 𝑗50 −0.5 + 𝑗50
e

𝑥 1 (𝑡) = 𝑗0.01𝑒 (−0.02×10 − 𝑗0.01𝑒 (−0.02×10

6
+𝑗 2×106 )𝑡 6
−𝑗 2×106 )𝑡

𝑥2 (𝑡) = (0.5 − 𝑗0.005)𝑒 (−0.02×10 + (0.5 + 𝑗0.005)𝑒 (−0.02×10

6
+𝑗 2×106 )𝑡 6
−𝑗 2×106 )𝑡

𝑥1 (𝑡) = 𝑗0.01𝑒 −0.02×10 𝑡 𝑒 𝑗 2×10 𝑡 − 𝑗0.01𝑒 −0.02×10 𝑡 𝑒 −𝑗 2×10 𝑡

6 6 6 6

𝑥 2 (𝑡) = (0.5 − 𝑗0.005)𝑒 −0.02×10 𝑡 𝑒 𝑗 2×10 𝑡 + (0.5 + 𝑗0.005)𝑒 −0.02×10 𝑡 𝑒 −𝑗 2×10 𝑡

6 6 6 6

   
𝑥1 (𝑡) = 𝑒 −0.02×10 𝑡
𝑗0.01𝑒 𝑗 2×10 𝑡 − 𝑗0.01𝑒 −𝑗 2×10 𝑡
= −0.020002𝑒 −0.02×10 𝑡 sin 2 × 106 𝑡
6 6 6 6

 
𝑥2 (𝑡) = 𝑒 −0.02×10 𝑡
(0.5 − 𝑗0.005)𝑒 𝑗 2×10 𝑡 + (0.5 + 𝑗0.005)𝑒 −𝑗 2×10 𝑡
6 6 6

   
= 𝑒 −0.02×10 𝑡 cos 2 × 106 𝑡 + 0.01 · 𝑒 −0.02×10 𝑡 sin 2 × 106 𝑡 .
6 6

c) In the provided Jupyter notebook, move the sliders to approximately

𝑅 = 1Ω and 𝐶 = 10𝑛𝐹. Sketch 𝑉𝑐 (𝑡) and comment on its appearance.
Additionally, sketch the location of the eigenvalues on the complex plane.
Are the waveforms for 𝑥1 (𝑡) and 𝑥2 (𝑡) “transient” — do they die out with
time?
Note: Because the resistance is so small, this is called the “underdamped”
case. It is good to reflect upon these waveforms to see why engineers
consider such behavior to be reflective of systems that don’t have enough
damping.

14
EECS 16B Spring 2020 HW 3

Solution

Yes, the waveforms are transient. They appear to be sinusoids that are
decaying exponentially. The eigenvalues should be located at coordinates
(−0.02 × 106 , 2 × 106 ) and (−0.02 × 106 , −2 × 106 ).

d) Notice that you got answers in terms of complex exponentials. Why did
the final voltage and current waveforms end up being purely real?

Solution

In this case, it’s because of the complex conjugacy of the quantities in the
problem. The eivenvalues and their associated eigenvectors were complex
conjugates, as were the transformed solutions e 𝑥1 (𝑡) and e
𝑥2 (𝑡). When we
applied the inverse transformation to e 𝑥 1 (𝑡) and e
𝑥2 (𝑡), we added together
many complex conjugate terms, and the imaginary parts cancelled out.

Now, was this just a fluke that just happened to line up perfectly? Is there
some 𝐴 matrix out there with real-valued entries that will result in a
complex solution? Or is something more profound going on?

It turns to be no fluke. If the entries in the 𝐴 matrix are real, and the initial
condition 𝑥®0 is real, then the solution to the differential equation 𝑑𝑡𝑑 𝑥® = 𝐴 𝑥®
with 𝑥®(0) = 𝑥®0 will also be real, regardless of whether the eigenvalues of 𝐴
are real, imaginary, or complex. If a matrix 𝐴 ∈ R𝑛×𝑛 has some complex
eigenvalues, then those eigenvalues will always arise in complex conjugate
pairs. Furthermore, the eigenvectors associated to those eigenvalues arise
on complex conjugate pairs. This will lead to the kind of cancellation that
you saw in here, every single time.

After all, the quantities that we observe in the world are always purely
real, so we would expect that the solutions to our models would also be
real-valued.

6 RLC Responses: Critically Damped Case

Building on the previous problem, consider the following circuit with specified
component values: (Notice 𝑅 is not specified yet. You’ll have to figure out what
that is.)

15
 HW 3 @ 2020-02-21 09:07:28-08:00

10 nF
𝑡=0 𝑅 25 µH
𝐼𝐿
+ − + − + −
𝑉𝐶 𝑉𝑅 𝑉𝐿

+
𝑉𝑠 𝑡=0
−

Assume the circuit above has reached steady state for 𝑡 < 0. At time 𝑡 = 0,
the switch changes state and disconnects the voltage source, replacing it with a
short.
For this problem, we use the same notations as in Problem 1.

a) For what value of 𝑅 is there going to be a single eigenvalue of 𝐴?

Solution
If the terms under the square root, i.e., the discriminant of the quadratic
formula, is 0, then we have a single value. More concretely,

𝑅2 4
− =0 (25)
𝐿 2 𝐿𝐶
𝑅2 4
⇒ 2 = (26)
𝐿 𝐿𝐶
r
𝐿
∴𝑅 = 2 (27)
𝐶

b) Find the eigenvalues and eigenspaces of 𝐴. What are the dimensions

of the corresponding eigenspaces? (i.e. how many linearly independent
eigenvectors can you find associated with this eigenvalue?)
For this part, assume the given values for the capacitor and the inductor,
as well as the critical value for the resistance 𝑅 that you found in the
previous part. It is easier to do the algebra with a non-symbolic matrix to
work with.

Solution
Our system’s matrix becomes,
 
−4 × 106 −4 × 104
𝐴= (28)
108 0

16
EECS 16B Spring 2020 HW 3

Our single eigenvector is,

𝑅
𝜆=− = −2 × 106 (29)
2𝐿
Hence, the eigenvector is a basis of the nullspace of 𝐴 − 𝜆𝐼,

𝑣1
    
−2 × 106 −4 × 104 0
= (30)
108 2 × 106 𝑣2 0

𝑣
   
1
Hence, the eigenvector, 𝑣® = 𝛼 1 = 𝛼 . We have only one eigen-
𝑣2 −50
vector, since we have a single dimensional nullspace.

c) In the provided Jupyter notebook, move the sliders to the resistance value
you found in the first part and 𝐶 = 10𝑛𝐹. Sketch 𝑉𝑐 (𝑡) and comment on
its appearance. Additionally, sketch the location of the eigenvalues on
the complex plane. What happens to the voltage and eigenvalues as you
slightly increase or decrease 𝑅?

Solution
At the 𝑅 = 100, 𝑉𝑐 (𝑡) appears to decay exponentially. A slight increase in
𝑅 causes the voltage to decay more slowly. A slight decrease in 𝑅 causes
a voltage undershoot and eventually oscillations. The eigenvalues have
converged into the same point at (−2 × 106 , 0). Increasing 𝑅 makes them
split into two points, and both points remain on the real axis. One point
goes towards the origin, while the other goes towards negative infinity.
Decreasing 𝑅 splits the eigenvalues back into their complex conjugates.

7 (OPTIONAL) Write Your Own Question And Provide a

Thorough Solution.
Writing your own problems is a very effective way to really learn material.
Having some practice at trying to create problems helps you study for exams
much better than simply solving existing practice problems. This is because
thinking about how to create an interesting problem forces you to really
consolidate your understanding of the course material.

8 Homework Process and Study Group

Citing sources and collaborators are an important part of life, including being
a student! We also want to understand what resources you find helpful and
how much time homework is taking, so we can change things in the future if
possible.

17
 HW 3 @ 2020-02-21 09:07:28-08:00

a) What sources (if any) did you use as you worked through the home-
work?
b) If you worked with someone on this homework, who did you work
with? List names and student ID’s. (In case of homework party, you can
also just describe the group.)
c) How did you work on this homework? (For example, I first worked by
myself for 2 hours, but got stuck on problem 3, so I went to office hours. Then I
went to homework party for a few hours, where I finished the homework.)
d) Roughly how many total hours did you work on this homework?

18