Download as pdf or txt
Download as pdf or txt
You are on page 1of 96

Systems and Signals 244

Notes compiled by Arno Botha


Please contact me if you spot any mistakes
079 555 3232
22557318@sun.ac.za
Contents
1 Step Function ...................................................................................................... 6

1.1 Three Step-wise function .............................................................................. 6

1.2 Timeshift (Switch On) .................................................................................... 7

1.3 Switch Off ...................................................................................................... 7

1.4 Pulse ............................................................................................................. 8

1.5 Windowing..................................................................................................... 8

1.6 Impulse Function ........................................................................................... 9

1.6.1 Impulse Example .................................................................................. 10

1.7 Shifting Property of Impulse Functions ........................................................ 10

2 Some Revision .................................................................................................. 11

3 Modelling ........................................................................................................... 12

4 Laplace Transform ............................................................................................ 12

4.1 Laplace Transform Definition ...................................................................... 12

4.2 Laplace Transform of a derivative ............................................................... 13

4.3 Some Functional Transforms ...................................................................... 14

4.4 Applying Laplace Transform to KVL equation ............................................. 15

4.5 Inverse Laplace Transform.......................................................................... 17

4.6 Improper Fractions ...................................................................................... 18

4.7 Proper Rational Fractions ........................................................................... 18

4.8 Partial Fraction Expansion .......................................................................... 19

4.8.1 Method 1 - Identification ....................................................................... 19

4.9 Method 2 – Projection ................................................................................. 20

4.9.1 Non-repeating poles ............................................................................. 20

4.9.2 Repeating poles .................................................................................... 21

4.10 Initial Value Theorem ............................................................................... 22

4.11 Final Value Theorem................................................................................ 22


4.12 Equivalent S-Domain models ................................................................... 22

4.12.1 Resistor ............................................................................................. 23

4.12.2 Inductor ............................................................................................. 23

4.12.3 Capacitor ........................................................................................... 25

4.13 Kirchhoff’s Laws in S-Domain .................................................................. 26

4.14 Using Physical Insight .............................................................................. 26

4.15 Thevenin / Norton Equivalents ................................................................. 27

4.16 Mutual Inductance.................................................................................... 29

4.16.1 T-Equivalent ...................................................................................... 30

4.16.2 𝝅-Equivalent ...................................................................................... 31

4.17 Transfer Function ..................................................................................... 31

4.17.1 Time Invariant.................................................................................... 32

4.17.2 Impulse input ..................................................................................... 32

4.18 Convolution .............................................................................................. 33

4.18.1 Use of Convolution ............................................................................ 33

4.18.2 Derivation .......................................................................................... 33

4.18.3 Formal Definition ............................................................................... 33

4.18.4 Example ............................................................................................ 35

4.18.5 Steady State Sinusoidal Response ................................................... 37

5 Bode Plots ......................................................................................................... 37

5.1 Log Laws ..................................................................................................... 38

5.2 The decibel 𝒅𝑩 ............................................................................................ 38

5.3 Plotting Transfer Functions ......................................................................... 39

5.3.1 Functions with real poles and zeros ..................................................... 39

5.3.2 Functions with complex poles and zeros .............................................. 44

6 Filters ................................................................................................................ 47

6.1 Use of Filters ............................................................................................... 47


6.2 Low-Pass Filters.......................................................................................... 48

6.2.1 Capacitor Resistor Low-Pass Filter (RC) .............................................. 48

6.2.2 Inductor Resistor Low-Pass Filter (LR) ................................................. 48

6.2.3 Bode Plot of Low-Pass Filter ................................................................ 48

6.2.4 Time domain ......................................................................................... 49

6.2.5 High-Order Lowpass Filters .................................................................. 49

6.2.6 Realization of Higher-Order Lowpass Filters ........................................ 53

6.2.7 Impedance Scaling ............................................................................... 53

6.2.8 Frequency Scaling ................................................................................ 54

6.3 High-Pass Filters ......................................................................................... 55

6.3.1 Transformations.................................................................................... 55

6.3.2 Low-Pass to High-Pass Transform ....................................................... 56

6.4 Band-Pass Filters ........................................................................................ 57

6.4.1 Low-Pass to Band-Pass Transform ...................................................... 57

6.5 Band-Reject Filters...................................................................................... 60

6.5.1 Low-Pass to Band-Reject Transform .................................................... 60

6.6 Sallen-Key Active Filters ............................................................................. 62

6.6.1 Active Filters ......................................................................................... 62

6.6.2 Sallen-Key Sections ............................................................................. 63

7 The Fourier Series ............................................................................................ 70

7.1 Approximation Theory ................................................................................. 70

7.2 Fourier-Series and Signals .......................................................................... 71

7.3 Symmetry .................................................................................................... 73

7.4 Alternative Forms of Fourier-Series ............................................................ 75

7.4.1 Standard Form:..................................................................................... 75

7.4.2 Trigonometric Form: ............................................................................. 75

7.4.3 Exponential Form: ................................................................................ 76


7.5 Circuit Analysis............................................................................................ 76

7.6 Fourier Spectrum ........................................................................................ 77

7.7 Fourier Series of an impulse ....................................................................... 78

7.7.1 Example................................................................................................ 81

8 DC to DC converter ........................................................................................... 82

9 Fourier Transform ............................................................................................. 84

9.1 Convergence ............................................................................................... 84

9.2 Link between Laplace and Fourier .............................................................. 86

9.3 Important Properties .................................................................................... 88

9.4 Fourier Transform of derivative ................................................................... 91

9.5 Modulation................................................................................................... 92

9.6 Convolution ................................................................................................. 92

9.7 Example of transient response .................................................................... 93

9.8 Example of Steady-State response ............................................................. 95

10 Parseval’s Theorem........................................................................................ 96
1 Step Function

𝑓(𝑡) = 𝑢(𝑡)

0 ,𝑡 < 0
𝑢(𝑡) = {
1 ,𝑡 > 0

• Note this is a unit step function.

1.1 Three Step-wise function

0, 𝑡 < −Δ
1
𝑢(𝑡) = { ( ) 𝑡 + 0.5 , −Δ < 𝑡 < Δ

1, Δ<𝑡

• Note here that the smaller Δ becomes the steeper the slope of the function.

• By defining the step function in such a say makes it continuous and therefore
we can differentiate over it.
1.2 Timeshift (Switch On)
0, 𝑡 < 𝑎
𝑢(𝑡 − 𝑎) = {
1, 𝑡 > 𝑎

• Recall from high school that:


o Subtracting a value from 𝑡 will shift the graph right
o Adding a value to 𝑡 will shift the graph left

1.3 Switch Off


1, 𝑡 < 𝑎
𝑢(𝑎 − 𝑡) = {
0, 𝑡 > 𝑎

Or

𝑢(𝑎 − 𝑡) = 1 − 𝑢(𝑡 − 𝑎)
1.4 Pulse

• A pulse is generated by subtracting two time shifted functions


𝑢(𝑡 − 2) − 𝑢(𝑡 − 3)

1.5 Windowing
𝑓(𝑡) = (𝑎𝑡 + 𝑐) [ 𝑢(𝑡 − 𝑏1 ) − 𝑢(𝑡 − 𝑏2 ) ] = 𝑓1 (𝑡) 𝜔1 (𝑡)
• A pulse can therefore be used to mathematically “cut-out” a section of another
function.

1.6 Impulse Function

• If Δ → 0, we obtain a function of zero duration and infinite amplitude.


• The surface under the function is equal to 1, irrespective of the value of Δ.
• We will denote this as 𝛿(𝑡), and call it an impulse function.

𝑓(𝑡) = 𝛿(𝑡 − 𝑡0 )

• Infinite amplitude
• Zero duration
• Unity integral
• Not unique
1.6.1 Impulse Example

1 −|𝑡|/𝜖
𝑓(𝑡) = 𝑒
2𝜖

• As 𝜖 → 0, 𝑓(𝑡 = 0) → ∞
1
o Because 𝑒 0 = 1 and lim 2𝜖 = ∞
𝜖→0
|𝑡|
• 𝑒 − 𝜖 → 0 when 𝑡 = 0±
o Because as 𝜖 → 0 we get 𝑒 −∞ = 0


0 ∞ −𝑡/𝜖
𝑒 𝑡/𝜖 𝑒 1 1
∫ 𝑓(𝑡)𝑑𝑡 = ∫ 𝑑𝑡 + ∫ 𝑑𝑡 = + = 1
−∞ 2𝜖 0 2𝜖 2 2
−∞

• Various other functions also give similar results. These functions are
called impulse generating functions.

0, 𝑡 ≠ 0 ∞
𝑓(𝑡) = { and ∫−∞ 𝑓(𝑡)𝑑𝑡 = 1
∞, 𝑡 = 0

Also note
𝑑
𝛿(𝑡) = 𝑢(𝑡)
𝑑𝑡

1.7 Shifting Property of Impulse Functions


∞ 𝑎+𝜖
∫ 𝑓(𝑡)𝛿(𝑡 − 𝑎) 𝑑𝑡 = ∫ 𝑓(𝑡)𝛿(𝑡 − 𝑎) 𝑑𝑡
−∞ 𝑎−𝜖

• The limits can be changed because 𝛿(𝑡 − 𝑎) = 0 when 𝑡 ≠ 𝑎

𝑓(𝑡)𝛿(𝑡 − 𝑎) = 0, for all 𝑡 ≠ 𝑎

• 𝑡=𝑎
𝑓(𝑡) = 𝑓(𝑎)
𝛿(𝑡 − 𝑎) ≠ 0

• Adding all this information we get the following


Integral = 1
∞ ∞
∫ 𝑓(𝑡)𝛿(𝑡 − 𝑎) 𝑑𝑡 = ∫ 𝑓(𝑎)𝛿(𝑡 − 𝑎) 𝑑𝑡
−∞ −∞

Constant

= 𝑓(𝑎)

2 Some Revision

• Second order differential equation:


o Example
𝑑2𝑦
+ 𝜔𝑜 𝑦 = 0
𝑑𝑡 2

𝑦(𝑡) = 𝐴𝑐𝑜𝑠(𝜔𝑜 𝑡) + 𝐵𝑠𝑖𝑛(𝜔𝑜 𝑡)

• Now these differential equations are sometimes tedious to work with


and we do not always have a “quick” solution as we do with second
order differential equations.
• Therefore we use the Laplace Domain.
• The next chapter will describe how to transform switches into the
Laplace Domain.

3 Modelling
• We make use of the Switch On and Switch Off functions, as well as the
windowing function to remove switches in our diagrams. This is later useful
when working with Laplace transforms.

4 Laplace Transform

4.1 Laplace Transform Definition



𝐹(𝑠) = ℒ[ 𝑓(𝑡) ] = ∫ 𝑓(𝑡)𝑒 −𝑆𝑡 𝑑𝑡
0−

𝑆 = 𝜎 + 𝑗𝜔 (complex frequency)

• 𝐹(𝑆) is the Laplace Transform


• ℒ is the Laplace Operator
• Transforms between the t-domain (time domain) and s-domain (complex
frequency domain)
• The transform is one sided, for the integral to converge.
o It goes from 0− to ∞!
• 𝑅𝑒[𝑆] = 𝜎 must be positive, and large enough for the integral to converge

4.2 Laplace Transform of a derivative


• The derivation is in lecture 3.

For more information on this view


8.7 Laplace transform of a Derivative
in EngMath 214 summary.

• 𝑦(𝑆) can be obtained from an algebraic expression


• 𝑦(𝑡) can be obtained from the inverse Laplace transform
• This can be done for any order differential equation
4.3 Some Functional Transforms
• Translation in time-domain
−𝑎
o 𝑓(𝑡 − 𝑎)𝑢(𝑡 − 𝑎) ↔ 𝑒 𝑆 𝐹(𝑆)
• Translation in frequency-domain
o 𝑒 −𝑎𝑡 𝑓(𝑡) ↔ 𝐹(𝑆 + 𝑎)
• Scale change
1 𝑆
o 𝑓(𝑎𝑡) ↔ 𝑎 𝐹 (𝑎)

4.4 Applying Laplace Transform to KVL equation


Using Standard Loop equation principles, we get:
𝑑𝑖(𝑡) 1 𝑡
−𝑉𝑒 (𝑡) + [𝐿 ] + [ ∫ 𝑖(𝜏)𝑑𝜏 + 𝑉𝐶 (0)] + 𝑅𝑖(𝑡) = 0
𝑑𝑡 𝐶 0

Applying Laplace transform to all the terms using 𝑉𝐶 (0) = 0.

With no initial conditions, 𝑖(0− ) = 0

𝑉𝑒 (𝑆)
𝐼(𝑆) =
1
𝑆𝐿 + 𝑆𝐶 + 𝑅

𝑉𝑂
With 𝑉𝑒 (𝑡) = 𝑉𝑂 𝑢(𝑡) ⇒ 𝑉𝑒 (𝑆) = 𝑆

𝑉𝑂
𝐼(𝑆) = 𝑆
1
𝑆𝐿 + 𝑆𝐶 + 𝑅

𝑉𝑂
𝐼(𝑆) =
1
𝐿𝑆 2 + 𝑅𝑆 + 𝐶

This is the solution for 𝐼(𝑆) in the complex-domain, to find the time domain solution
𝑖(𝑡) we must inverse Laplace transform it.

As an example let
𝑅=2
𝐿=1
1
𝐶=
2

𝑉𝑂 𝑉𝑂
𝐼(𝑆) = =
𝑆2 + 2𝑆 + 2 (𝑆 + 1)2 + 1

Using the following identify from the table above


𝜔
⇔ 𝑒 −𝑎𝑡 sin (𝜔𝑡)
(𝑆 + 𝑎)2 + 𝜔 2

𝜔 = 1, 𝑎 = 1

𝑖(𝑡) = [𝑒 −𝑡 sin (𝑡)]𝑉𝑂 𝑢(𝑡)


-Called Damped Sine

Note you must always add 𝑢(𝑡) when doing the inverse Laplace transform because
the Laplace transform is only valid for the positive half of the domain.

4.5 Inverse Laplace Transform

• Given a function in the S-domain, we can transform this function to the t-


domain using the Inverse Laplace Transform, denoted as ℒ −1 [ 𝐹(𝑆) ]
• By definition ℒ −1 [ 𝐹(𝑆) ] = 𝑓(𝑡) if ℒ[ 𝑓(𝑡) ] = 𝐹(𝑆)
• Mathematically

𝑐+𝑗∞
1
ℒ −1 [ 𝐹(𝑆) ] = ∫ 𝐹(𝑆)𝑒 𝑆𝑡 𝑑𝑆
𝑗 × 2𝜋 𝑐−𝑗∞

Here, the integration path must be such that all singular points of 𝐹(𝑆) lie to
the left of the line 𝑅𝑒[𝑆] = 𝑐 in the complex plane.

• This transform is valid for any 𝑓(𝑡) ↔ 𝐹(𝑆) pair but can be quite difficult to
execute.
• When applied to differential equations in the time-domain we can show that
any S-domain function will be in form of a rational function such as:

𝑎𝑛 𝑆 𝑛 + 𝑎𝑛−1 𝑆 𝑛−1 + ⋯ + 𝑎0
𝐹(𝑆) =
𝑏𝑚 𝑆 𝑚 + 𝑏𝑚−1 𝑆 𝑚−1 + ⋯ + 𝑏0

• Instead of using the integral approach we break this into partial fractions and
then using the table provided above to solve for the time domain solution.

4.6 Improper Fractions


• Recall that if the degree of the numerator is larger than that of the
denominator then you must first do long division.

4.7 Proper Rational Fractions

• Any rational function can be expressed as

𝑁(𝑆) 𝐾(𝑆 + 𝑍1 )(𝑆 + 𝑍2 ) … (𝑆 + 𝑍𝑛 )


𝐹(𝑆) = =
𝐷(𝑆) (𝑆 + 𝑃1 )(𝑆 + 𝑃2 ) … (𝑆 + 𝑃𝑚 )

• Points where 𝑆 = −𝑍𝑖 where 𝑖 ∈ [1, 𝑛]


o Called Zeros
o 𝐹(𝑆 = −𝑍𝑖 ) = 0
o Roots of numerator
• Points where 𝑆 = −𝑃𝑖 where 𝑖 ∈ [1, 𝑚]
o Called Poles
o 𝐹(𝑆 = −𝑃𝑖 ) = ∞
o Roots of dominator

• Any rational function can be expressed as the sum of fractions, in terms of its
poles.
• This is called partial fraction expansion.

𝐾1 𝐾2
𝐹(𝑆) = + +⋯
𝑆 + 𝑃1 𝑆 + 𝑃2

Where 𝐾1 is called the residue, and 𝑃𝑛 and 𝐾𝑛 are complex.

4.8 Partial Fraction Expansion

• Determine Poles
o Roots of the denominator
• Calculate Residues
o Identification
o Projection

4.8.1 Method 1 - Identification

• For simple functions, residues can be calculated as in the following example

1 𝐾1 𝐾2 𝐾1 (𝑆 + 2) + 𝐾2 (𝑆 + 1)
𝐹(𝑆) = = + =
(𝑆 + 1)(𝑆 + 2) 𝑆 + 1 𝑆 + 2 (𝑆 + 1)(𝑆 + 2)
1 𝑆(𝐾1 + 𝐾2 ) + (2𝐾1 + 𝐾2 )
=
(𝑆 + 1)(𝑆 + 2) (𝑆 + 1)(𝑆 + 2)

• Now compare the coefficient of each term in the numerator

𝐾1 + 𝐾2 = 0
2𝐾1 + 𝐾2 = 1

• Now just solve it like you would in high school

4.9 Method 2 – Projection

4.9.1 Non-repeating poles


1 𝐾1 𝐾2 𝐾1 (𝑆 + 2) + 𝐾2 (𝑆 + 1)
𝐹(𝑆) = = + =
(𝑆 + 1)(𝑆 + 2) 𝑆 + 1 𝑆 + 2 (𝑆 + 1)(𝑆 + 2)

Multiply both sides with one of the poles

1 𝐾1 𝐾2
× (𝑆 + 1) = × (𝑆 + 1) + × (𝑆 + 1)
(𝑆 + 1)(𝑆 + 2) 𝑆+1 𝑆+2

Therefore we are left with


1 𝐾2 1
= 𝐾1 + × 0 = 𝐾1 =
−1 + 2 𝑆+2 1

• This only works if you have a pole of 1 degree, like no (𝑆 + 1)2


4.9.2 Repeating poles

• If 𝐹(𝑆) contains more than one pole at the same position, such a pole is called
‘repeating’, or ‘of order 𝑛’
o Where 𝑛 is the number of the poles at the position

• General equation for a pole of order 𝑛 expanded as

𝐾𝑛 𝐾𝑛−1 𝐾1
𝑛
+ 𝑛−1
+⋯+
(𝑆 + 𝑃) (𝑆 + 𝑃) (𝑆 + 𝑃)

𝐾𝑛 = [(𝑆 + 𝑃)𝑛 𝐹(𝑆)]𝑆=−𝑃


𝑑
𝐾𝑛−1 = [(𝑆 + 𝑃)𝑛 𝐹(𝑆)]𝑆=−𝑃
𝑑𝑆

𝑑 (𝑛)
𝐾𝑛 = [(𝑆 + 𝑃)𝑛 𝐹(𝑆)]𝑆=−𝑃
𝑑𝑆 (𝑛)

𝑑(𝑛)
o Recall that 𝑑𝑆(𝑛) means the 𝑛𝑡ℎ derivative, NOT POWER OF 𝑛

4.9.2.1 Example:
𝑆+2
𝐹(𝑆) =
(𝑆 + 1)2

• This example has a pole of order 2 at 𝑆 = −1

• The expansion of such a function is as follows:


𝑆+2 𝐾1 𝐾2
𝐹(𝑆) = 2
= 2
+
(𝑆 + 1) (𝑆 + 1) (𝑆 + 1)

• Using projection, we multiply both sides with (𝑆 + 1)2


(𝑆 + 2) = 𝐾1 + [(𝑆 + 1) 𝐾2 ]

• Evaluate both sides at 𝑆 = −1

(𝑆 + 2)𝑆=−1 = 𝐾1 + [(𝑆 + 1) 𝐾2 ]𝑆=−1

1 = 𝐾1
• To get 𝐾2 we take the derivative of the function
o We had
(𝑆 + 2) = 𝐾1 + [(𝑆 + 1) 𝐾2 ]
o Take the derivative
▪ 1 = (1)𝐾2
o Evaluate both sides at 𝑆 = −1
▪ 1 = 𝐾2

4.10 Initial Value Theorem

lim 𝑓(𝑡) = lim 𝑆 𝐹(𝑆)


𝑡→0+ 𝑆→∞

4.11 Final Value Theorem

lim 𝑓(𝑡) = lim 𝑆 𝐹(𝑆)


𝑡→∞ 𝑆→0

• These two theorems can be used to quickly determine circuit response.

4.12 Equivalent S-Domain models


4.12.1 Resistor

• In the S-Domain
𝑉(𝑆) = 𝑅 𝐼(𝑆)
𝑉(𝑆)
𝑅=
𝐼(𝑆)

4.12.2 Inductor

• In the S-Domain
𝑉(𝑆) = 𝐿[𝑆 𝐼(𝑆) − 𝑖(0− )]
• Let 𝐼0 = 𝑖(0− )
𝑉(𝑆) = 𝑆 𝐿 𝐼(𝑆) − 𝐿 𝐼0
• The initial condition on the current in the inductor, 𝑖(0− ) = 𝐼0, appears in the
S-plane equation in the same form as a series constant voltage source.

o Another form is
𝑉(𝑆) + 𝐿𝐼0 𝑉(𝑆) 𝐼0
𝐼(𝑆) = = +
𝑆𝐿 𝑆𝐿 𝑆

• Note that the term 𝑆𝐿 appears in all the equations in the same way as 𝑅 i.e.
multiplied by 𝐼(𝑆). If we denote 𝐼𝐿 (𝑆), 𝑉𝐿 (𝑆) as the voltage and current for the
𝑉𝐿 (𝑆)
S-domain inductor only, then = 𝑆𝐿
𝐼𝐿 (𝑆)
1
• We call 𝑆𝐿 the impedance of the inductor and 𝑆𝐿 the admittance.
𝑉𝐿 (𝑆)
𝑍𝐿 (𝑆) = = 𝑆𝐿
𝐼𝐿 (𝑆)

𝐼𝐿 (𝑆) 1
𝑦𝐿 (𝑆) = =
𝑉𝐿 (𝑆) 𝑆𝐿

4.12.3 Capacitor

• In the S-domain:
𝐼(𝑆) = 𝐶[𝑆 𝑉(𝑆) − 𝑉(0− )]
• Denotes 𝑉(0− ) = 𝑉0
𝐼(𝑆) = 𝑆 𝐶 𝑉(𝑆) − 𝐶𝑉0

• The initial condition on voltage appears in the S-plane equation in the same
way as a constant current source in parallel.
• We can model a capacitor in the S-domain as
• If 𝑉𝐶 (𝑆) and 𝐼𝐶 (𝑆) now refer to as the S-domain capacitor only voltage and
current
• 𝑍𝐶 (𝑆) – Impedance
• 𝑦𝐶 (𝑆) – Admittance

𝑉𝐶 (𝑆) 1
𝑍𝐶 (𝑆) = =
𝐼𝐶 (𝑆) 𝑆𝐶

𝐼𝐶 (𝑆)
𝑦𝐶 (𝑆) = = 𝑆𝐶
𝑉𝐶 (𝑆)

4.13 Kirchhoff’s Laws in S-Domain


• Yeh they work, big surprise

4.14 Using Physical Insight


• We may be asked to predict how the circuit will respond without doing any
calculations.
• For this we must use our understanding of the physical properties of the
components
o Voltage over a capacitor cannot change instantaneously
o Current through an inductor cannot change instantaneously
4.15 Thevenin / Norton Equivalents

• Replace Source with

• S-Plane Equivalent:
4.16 Mutual Inductance
4.16.1 T-Equivalent
4.16.2 𝝅-Equivalent

4.17 Transfer Function


• Working in the S-Domain enables us to relate any output variable (response)
to any input variable (excitation).
• This relationship is called a transfer function
• We discuss some characteristics of the transfer function:
o Time-invariant
o ℎ(𝑡) = ℒ −1 [𝐻(𝑆)] is the impulse response
• All our circuits can be written in the S-Plane as:
𝑌(𝑆) = 𝐻(𝑆)𝑋(𝑆)
o With 𝑋, 𝑌 being either voltage or current
• The transfer function 𝐻(𝑆) is
o Always a rational function of 𝑆
o Poles and Zeros always in complex pairs
o Stable systems have poles in the LHP
o Zeros may be anywhere

• Two important characteristics


𝑥(𝑡 − 𝑎)𝑢(𝑡 − 𝑎) ↔ 𝑒 −𝑎𝑠 𝑋(𝑆)
𝑦(𝑡 − 𝑎)𝑢(𝑡 − 𝑎) ↔ 𝑒 −𝑎𝑠 𝐻(𝑆)𝑋(𝑆) = 𝑒 −𝑎𝑠 𝑦(𝑆)

4.17.1 Time Invariant

• Delaying the input by 𝑎, delays the output by 𝑎


o ∴ Time Invariant

4.17.2 Impulse input

• Recall if
𝑥(𝑡) = 𝛿(𝑡)
𝑋(𝑆) = 1
𝑌(𝑆) = 𝑋(𝑆)𝐻(𝑆)
𝑌(𝑆) = 𝐻(𝑆)

𝑦(𝑡) = ℒ −1 [𝐻(𝑆)]

• For unit impulse as input, the output is the inverse Laplace transform of the
transfer function
• We call ℎ(𝑡) = ℒ −1 [𝐻(𝑆)] the impulse response .
• The impulse response contains all the information we need to calculate the
output for any input
• The impulse response is also the natural response of a circuit.

4.18 Convolution

• Given the impulse response ℎ(𝑡) of a circuit, and excitation, 𝑥(𝑡)

𝑦(𝑡) = ∫ 𝑥(𝜆)ℎ(𝑡 − 𝜆) 𝑑𝜆
−∞

= 𝑥(𝑡) ⊗ ℎ(𝑡)
• This is called the convolution integral

4.18.1 Use of Convolution

• Can we calculate the output of a circuit, given the input and the impulse
response, fully in the time domain?
o Yes, using the convolution integral!

4.18.2 Derivation

• The derivation can be found in Lecture 10

4.18.3 Formal Definition


𝑦(𝑡) = ∫ 𝑥(𝜆)ℎ(𝑡 − 𝜆) 𝑑𝜆 = 𝑥(𝑡) ⊗ ℎ(𝑡)


−∞

Or in another form, using


𝑡−𝜆 =𝑢
𝑑𝜆 = −𝑑𝑢
−∞ ∞

𝑦(𝑡) = ∫ −𝑥(𝑡 − 𝑢)ℎ(𝑢)𝑑𝑢 = ∫ 𝑥(𝑡 − 𝑢)ℎ(𝑢)𝑑𝑢 = ℎ(𝑡) ⊗ 𝑥(𝑡)


∞ −∞

• In the example above 𝑦(𝑡) = ℎ(𝑡) ⊗ 𝑥(𝑡)


• This means that ℎ(𝑡) is being convoluted with 𝑥(𝑡) (the changing one)
• Recall that when you have 𝑓(𝑥 − 𝑎) you are shifting the function right
• Similarly when you have 𝑓(𝑥 + 𝑎) you shift the function left
• Now in the above image 𝑥(𝜆) is shown, now 𝑥(−𝜆) is just the mirror about the
𝑦 − 𝑎𝑥𝑖𝑠
• So when they incorporate the shift by 𝑡
o 𝑥(𝑡 − 𝜆) = 𝑥(−[𝜆 − 𝑡])
o Which is the function 𝑥 shifted right by 𝑡 units and then mirrored about
the point at 𝑡
4.18.4 Example

𝑉𝑜𝑢𝑡 (𝑆) 𝑅 1
𝐻(𝑆) = = =
𝑉𝑔 (𝑆) 𝑅 + 𝑆𝐿 1 + 𝑆𝐿
𝑅

𝐿
ℎ(𝑡) = ℒ − [𝐻(𝑆)] = 𝑒 −(𝑅)𝑡 𝑢(𝑡) = 𝑒 −𝑡 𝑢(𝑡)

• For this example we will do 𝑦(𝑡) = 𝑥(𝑡) ⊗ ℎ(𝑡)

𝑦(𝑡) = ∫ 𝑥(𝜆)ℎ(𝑡 − 𝜆)𝑑𝜆


0

now ℎ(𝜆) = 1 at 𝜆 = 0
ℎ(𝑡 − 𝜆) = 1 at 𝑡 − 𝜆 = 0
or 𝑡 = 𝜆

• Now we can shift the graph


• For 𝑡 < 0:
𝑡

∫ 𝑥(𝜆)ℎ(𝑡 − 𝜆) = 0
0

• Since there is no overlap.

• For 0 < 𝑡 < 5


𝑡 𝑡

∫ 𝑥(𝜆)ℎ(𝑡 − 𝜆) = ∫(4𝜆)𝑒 −(𝑡−𝜆) 𝑑𝜆


0 0
𝑡
𝑡
= 4𝑒 −𝑡 ∫ 𝜆𝑒 𝜆 𝑑𝜆 = 4𝑒 −𝑡 [𝑒 𝜆 (𝜆 − 1)] = 4𝑡 − 4 + 4𝑒 −𝑡
0
0

• Now since our 𝑥(𝜆) has more than one “part” we must do two integrals in the
next step.

Final Answer is:


4.18.5 Steady State Sinusoidal Response

• Given a excitation 𝑥(𝑡) = 𝐴𝑐𝑜𝑠(𝜔0 𝑡 + 𝜙)𝑢(𝑡) and an S-plane transfer function


𝐻(𝑆)
• The output can be written directly in the time-domain as

𝑦(𝑡) = 𝐴| 𝐻(𝑗𝜔0 ) | cos( 𝜔0 𝑡 + 𝜙 + 𝜃(𝜔0 ) )

Where
𝐻(𝑆)|𝑠=𝑗𝜔0 = |𝐻(𝑗𝜔0 )| 𝑒 𝑗𝜃(𝜔0) = |𝐻(𝑗𝜔0 )| ∠𝐻(𝑗𝜔0 )

• Lecture 11 if you want to see derivation.

5 Bode Plots

• We work with two types of bode plots


o Magnitude |𝐻(𝑗𝜔)|𝑑𝐵
o Phase ∠𝐻(𝑗𝜔)
5.1 Log Laws

5.2 The decibel 𝒅𝑩


• Power ratios are very important in physics
• Power ratios in the real world span orders of magnitude, ie 10−10 → 10+5
• If we plot this on linear scale, differences in small values disappear.

𝑃
• The function 10 log10 (𝑃2 ), with 𝑃1 and 𝑃2 two power levels in 𝑤𝑎𝑡𝑡, has
1

become the standard.

• It is called the decibel or 𝑑𝐵

𝑃2 𝑃2
[ ] = 10 log10 ( )
𝑃1 𝑑𝐵 𝑃1
𝑉2
• As 𝑃 = , a voltage ratio can also be expressed in 𝑑𝐵, but then
𝑅

𝑃2 𝑉22 𝑉2
10 log [ ] = 10 log [ 2 ] = 20 [ ]
𝑃1 𝑉1 𝑉1

𝑉2 𝑉2
→ [ ] = 20 log [ ]
𝑉1 𝑑𝐵 𝑉1

5.3 Plotting Transfer Functions


𝑉𝑜𝑢𝑡 (𝑆)
• Given a transfer function 𝐻(𝑆) = 𝑉𝑖𝑛 (𝑆)

• Given a excitation signal 𝑉𝑖𝑛 (𝑡) = 𝐴 cos(𝜔0 𝑡) 𝑢(𝑡)


• The stead state output is 𝑉𝑜𝑢𝑡 (𝑡) = 𝐴 |𝐻(𝑗𝜔0 )| cos(𝜔0 𝑡 + 𝜃) 𝑢(𝑡)
o 𝜃 = ∠𝐻(𝑗𝜔0 )

𝑟𝑎𝑑
• As 𝜔 is the frequency ( 𝑖𝑛 ) of the sinusoidal signal cos (𝜔𝑡)
𝑠

o The two functions |𝐻(𝑗𝜔)| and 𝜃(𝑗𝜔) together are called the frequency
response of a system
o With |𝐻(𝑗𝜔)| being the magnitude/amplitude response
o With 𝜃(𝑗𝜔) being the phase response
• Bode plots make use of asymptote approximations

5.3.1 Functions with real poles and zeros


𝑉𝑜𝑢𝑡 𝑎1 𝑆 + 𝑎0
𝐻(𝑆) = =
𝑉𝑖𝑛 𝑏2 𝑆 2 + 𝑏1 𝑆

• We can write this function in a convenient form

𝑆
𝐾 (𝑍 + 1) 𝑃1 𝑆 + 𝑍1
1
𝐻(𝑆) = =𝐾
𝑆 𝑍1 𝑆(𝑆 + 𝑃1 )
𝑆( + 1)
𝑃1
• Where
o 𝑍1 is a zero
o 𝑃1 is a pole
o Both real

5.3.1.1 Magnitude Pot


𝑆 𝑗𝜔
𝐾 (𝑍 + 1) |( 𝑍 ) + 1|
1 1
|𝐻(𝑗𝜔)| = | | =𝐾
𝑆 𝑗𝜔
𝑆 (𝑃 + 1) |𝑗𝜔| |( ) + 1|
1 𝑠=𝑗𝜔
𝑃1

|𝐻(𝑗𝜔)|𝑑𝐵 = 20 log|𝐻(𝑗𝜔)|
𝑗𝜔 𝑗𝜔
= 20 log(𝐾) + 20 log |( ) + 1| − 20 log|𝑗𝜔| − 20 log |( ) + 1|
𝑍1 𝑃1

• 20 log(𝐾) is simply a straight horizontal line


𝑗𝜔
• 20 log | 𝑍 + 1|
1

𝜔 𝑗𝜔
o For ≪1 20 log |( ) + 1| ≃ 20 log(1) = 0
𝑍1 𝑍1
𝜔 𝑗𝜔 𝜔
o For 𝑍 ≫ 1 20 log |( 𝑍 ) + 1| ≃ 20 log (𝑍 )
1 1 1

= 20 log(𝜔) − 20 log(𝑍1 )

o For Real poles and zeros we simply join the two lines at ω=Z_1S
• −20 log|𝑗𝜔|
o = −20 log(𝜔)
𝑑𝐵
o Straight line though origin, with slope of −20 𝑑𝑒𝑐𝑎𝑑𝑒

𝑗𝜔
• −20 log |( 𝑃 ) + 1|
1

𝜔 𝑗𝜔
o For 𝑍 ≪ 1 −20 log |( 𝑃 ) + 1| ≃ −20 log(1) = 0
1 1

𝜔 𝑗𝜔 𝜔
o For ≫1 −20 log |( ) + 1| ≃ −20 log ( )
𝑍1 𝑃1 𝑃1

= −20 log(𝜔) + 20 log(𝑃1 )


• Now add them by EYE

5.3.1.2 Phase Plot


𝑗𝜔 𝑗𝜔
∠𝐻(𝑗𝜔) = ∠𝐾 + ∠ (1 + ) − ∠(𝑗𝜔) − ∠ (1 + )
𝑍1 𝑃1

𝑏
• Note ∠𝑎 + 𝑗𝑏 = tan−1 (𝑎)

• ∠𝐾

o As K is real
0
o ∠𝐾 = tan−1 (𝐾) = 0°
𝑗𝜔 𝜔
• ∠ (1 + 𝑍 ) = tan−1 (𝑍 )
1 1

𝜔 𝜔
o For 𝑍 ≪ 1 tan−1 (𝑍 ) ≃ tan−1(0) = 0°
1 1

𝜔 𝜔
o For 𝑍 ≫ 1 tan−1 (𝑍 ) ≃ tan−1(∞) = 90°
1 1

o Unlike in the magnitude case, these asymptotes do not meet


anywhere. We therefore use a third asymptote which passes through
the actual phase value at 𝜔 = 𝑍1 , with a slope which approximates the
actual slope at that point.
𝜔
o tan−1 (𝑍 ) |𝜔=𝑍1 = 45°
1

o Thus the line through 45° meets the 0° asymptote at 0.1 𝑍1 and the 90°
asymptote at 10 𝑍1 .

• −∠(𝑗𝜔)
o As 𝑗𝜔 is imaginary
o −∠(𝑗𝜔) = −90°

𝑗𝜔 𝜔
• −∠ (1 + ) = −tan−1 (𝑃 )
𝑃1 1

𝜔 𝜔
o For 𝑍 ≪ 1 − tan−1 (𝑃 ) ≃ − tan−1(0) = 0°
1 1

𝜔 𝜔
o For 𝑍 ≫ 1 − tan−1 (𝑃 ) ≃ − tan−1(∞) = −90°
1 1

o Unlike in the magnitude case, these asymptotes do not meet


anywhere. We therefore use a third asymptote which passes through
the actual phase value at 𝜔 = 𝑍1 , with a slope which approximates the
actual slope at that point.
𝜔
o − tan−1 (𝑍 ) |𝜔=𝑍1 = −45°
1

o Thus the line through −45° meets the 0° asymptote at 0.1 𝑍1 and the
90° asymptote at 10 𝑍1 .

Add by EYE

5.3.2 Functions with complex poles and zeros

𝑎0 𝑎0
𝐻(𝑆) = = 2
(𝑆 + 𝛼 − 𝑗𝛽)(𝑆 + 𝛼 + 𝑗𝛽) 𝑆 + 2𝜁𝜔𝑛 𝑆 + 𝜔𝑛2
• Where
o 𝜔𝑛2 = 𝛼 2 + 𝛽 2
o 𝜁𝜔𝑛 = 𝛼
o 𝑆 = −𝜁𝜔𝑛 ± √(𝜁𝜔𝑛 )2 − 𝜔𝑛2

• 𝜁 is called the damping factor


o 𝜁 < 1: complex roots (underdamped)
o 𝜁 = 1: real equal roots (critically damped)
o 𝜁 > 1: real distinct roots (overdamped)

5.3.2.1 Standard Bode Form


𝐾
𝐻(𝑆) = 2
𝑆 2𝜁𝑆
( ) +( )+1
𝜔𝑛 𝜔𝑛

5.3.2.2 Magnitude Plot

𝐾 𝐾
|𝐻(𝑆)|𝑠=𝑗𝜔 |𝑑𝐵 = | 2 | =| |
𝑗𝜔 2𝜁(𝑗𝜔) 𝜔2 2𝜁
(1 − 2 ) + 𝑗𝜔 (
(𝜔 ) + ( 𝜔 ) + 1
𝑛 𝑛 𝜔𝑛 𝜔𝑛 ) 𝑑𝐵
𝑑𝐵

𝜔2 2𝜁
= 20 log(𝐾) − 20 log |(1 − 2 ) + 𝑗𝜔 ( )|
𝜔𝑛 𝜔𝑛

• 20 log(𝐾)
o Straight line
𝜔2 2𝜁
• −20 log |(1 − 𝜔2 ) + 𝑗𝜔 (𝜔 )|
𝑛 𝑛

o For
▪ 𝜔 ≪ 𝜔𝑛 −20 log(1) = 0
𝜔2
▪ 𝜔 ≫ 𝜔𝑛 −20 log |− 𝜔2 | = −40 log(𝜔) + 40log (𝜔𝑛 )
𝑛

• Join the asymptote at 𝜔 = 𝜔𝑛


o |𝐻(𝑗𝜔𝑛 )|𝑑𝐵 = −20 log|2𝜁|

5.3.2.3 Phase Plot

2𝜁𝜔
(𝜔 )
𝑛
𝜃 = ∠𝐻(𝑗𝜔) = − tan−1 [ ]
𝜔 2
1 − (𝜔 )
𝑛

• We will approximate this with 3 asymptotes


o 𝜔 ≪ 𝜔𝑛 : 𝜃 ≃ 0° (low frequency)
o 𝜔 ≫ 𝜔𝑛 : 𝜃 ≃ −180° (high frequency)
o 𝜔 ≃ 𝜔𝑛 : Line running through the actual phase value at 𝜔 = 𝜔𝑛 ,
with a slope approximating the actual slope
▪ 𝜃(𝜔 = 𝜔𝑛 ) = − tan−1(∞) = −90°
▪ Slope at 𝜔 = 𝜔𝑛
2.3 𝑟𝑎𝑑⁄ 132 𝑑𝑒𝑔
• 𝑆𝑙𝑜𝑝𝑒 = − 𝑑𝑒𝑐 = − 𝜁

𝑑𝑒𝑐
𝜁
• Either end of those stars is
𝜔 = 𝜔𝑛 (4.81)±𝜁

6 Filters

6.1 Use of Filters


• To define receive bands, and to limit output to allowed bands
• Select frequency bands
• To remove unwanted signals from the output of a system
6.2 Low-Pass Filters

6.2.1 Capacitor Resistor Low-Pass Filter (RC)

6.2.2 Inductor Resistor Low-Pass Filter (LR)

6.2.3 Bode Plot of Low-Pass Filter

• Both these transfer functions above are of the form


1
𝐻(𝑆) =
𝑆
𝑃+1
1 𝑅
With 𝑃 = 𝑅𝐶 or 𝑃 = 𝐿
• For 𝜔 < 𝑃, the output equals the input, thus the signal is not attenuated at
all. This is called the passband.
• The width of the passband is called the Bandwidth (B)
• For 𝜔 > 𝑃, the signal is attenuated. This is called the stopband.
• For 𝜔 = 𝑃 is called the cut-off frequency. At this point, the attenuation of
1
the first order filter is 3𝑑𝐵 , as 20 log |𝑗+1| = −3

6.2.4 Time domain


𝑃1
ℎ(𝑡) = ℒ −1 [𝐻(𝑆)] = ℒ −1 [ ] = 𝑃1 𝑒 −𝑃𝑡 𝑢(𝑡)
𝑆+𝑃
• Small values of 𝑃 means a slowing decaying signal, and a small
bandwidth.
• Large values of 𝑃 means a rapidly decaying signal, and a wide bandwidth.
1
• Bandwidth ∝ 𝑑𝑒𝑐𝑎𝑦

6.2.5 High-Order Lowpass Filters

• Higher order filters are obtain using more poles.

1
𝐻(𝑆) =
𝑆 𝑆
(𝑃 + 1) (𝑃 + 1) …
1 2
• The position of these poles determine the frequency response.
• Two sets of polynomial functions are most often used in filter design,
namely Butterworth, and Chebyshev polynomials.

6.2.5.1 Butterworth

• Butterworth polynomials have a characteristic that


1
|𝐻(𝑗𝜔)|2 =
1 + 𝜔 2𝑛
With 𝑛 being the order of the polynomial

• The poles are available in a closed-form, and lie on the unit circle in the
LHP of the complex plane.
• At 𝜔 = 0, the first 𝑛 − 1 derivatives of |𝐻(𝑗𝜔)| are zero
o Therefore this function is called maximally flat.

6.2.5.2 Chebyshev

• Chebyshev polynomials have the characteristic that

1
|𝐻(𝑗𝜔)|2 =
1+ 𝜖 2 cos 2 [𝑛 cos −1 (𝜔)]
• For 𝜔 < 1, cos −1 (𝜔) is real
o 1 + 𝜖 2 cos 2 (𝜃) gives an equi-ripple function
• For 𝜔 > 1, cos −1 (𝜔) is imaginary
o 1 + 𝜖 2 cos 2 (𝜃) gives a monotonic function

• The poles of a Chebyshev function lie on an ellipse in the complex plane,


with 𝑆𝑘 = 𝛼𝑘 + 𝑗𝛽𝑘

1 1 𝜋 2𝑘 − 1
𝛼𝑘 = ± sinh ( sinh−1 ( )) sin ( )
𝑛 𝜖 2 𝑛
1 1 𝜋 2𝑘 − 1
𝛽𝑘 = ± cosh ( sinh−1 ( )) cos ( )
𝑛 𝜖 2 𝑛

𝑘 = 1 − 2𝑛
𝛼𝑘 < 0
6.2.6 Realization of Higher-Order Lowpass Filters

• The design of filter is a complex process, but is the only exact circuit
synthesis in any field of electronics.
• We will only look at active filter implementations of higher order filers.
• Normalized (i.e. 𝑅 = 1 and 𝜔𝑐 = 1 𝑟𝑎𝑑⁄𝑠) lowpass filters of Butterworth
and Chebyshev form are widely available in books.

6.2.7 Impedance Scaling

• We can multiply all the impedances in a circuit with the same constant,
without changing voltage or current transfer functions.
• Impedance scaling is very useful to obtain realistic component values.

6.2.8 Frequency Scaling


𝑆
• If we replace every 𝑆 in our S-Plane circuit with Ω
0

o The frequency response is scaled by Ω0


• We can scale the frequency response by Ω0 , by dividing the value of each
reactive element by Ω0
6.3 High-Pass Filters

6.3.1 Transformations

• In all the following, note that any magnitude function is always an even
function
|𝐻(𝑗𝜔)| = |𝐻(−𝑗𝜔)|
• Even though a negative frequency axis makes no physical sense, it is a
very useful mathematical tool.

6.3.2 Low-Pass to High-Pass Transform

Ω0
𝑆↔
𝑆
Ω
• If a function |𝐻𝐿𝑃 (𝑆)|𝑆=𝑗𝜔0 = 𝐻0 , the function |𝐻𝐿𝑃 ( 0)| will be 𝐻0 at
𝑆 𝑆=𝑗𝜔
Ω Ω
𝑗𝜔0 = 𝑗𝜔0 , or 𝜔 = − 𝜔0
0

• We choose 𝐻𝐿𝑃 (𝑆) as a Low-Pass transfer function, and define a new


function

Ω0
𝐻𝐻𝑃 (𝑆) = 𝐻𝐿𝑃 ( )
𝑆

• To achieve such a result, we can replace all occurrences of 𝑆 in the


Ω0
lowpass circuit, with , as the transfer function is simply a rational function
𝑆

of all the S-Plane circuit elements.


• We can replace every inductor of value 𝐿 in the Low-Pass S-Plane circuit,
1
with a capacitor of value 𝐶 ′ = Ω
0

• We can replace ever capacitor of value 𝐶 in the Low-Pass S-Plane circuit,


1
with an inductor of value 𝐿′ = Ω
0𝐶

6.4 Band-Pass Filters

6.4.1 Low-Pass to Band-Pass Transform


1 𝑆 Ω0
𝑆↔ [ + ]
∆ Ω0 𝑆

1 𝑆 Ω0
• If a function |𝐻𝐿𝑃 (𝑆)|𝑆=𝑗𝜔0 = 𝐻0 , the function |𝐻𝐿𝑃 (∆ [Ω + ])| will be
0 𝑆 𝑆=𝑗𝜔
1 𝑗𝜔 Ω 1 𝜔 Ω0
𝐻0 where ∆ [Ω + 𝑗𝜔0 ] = 𝑗𝜔0 or [ − ] = 𝜔0
0 ∆ Ω 0 𝜔

• We define a new function


1 𝑆 Ω0
𝐻𝐵𝑃 (𝑆) = 𝐻𝐿𝑃 ( [ + ])
∆ Ω0 𝑆
• Derivation in Lecture 14

• We obtain a Band-Pass response, with


o Ω0 the center frequency
𝜔4 −𝜔3
o ∆Ω0 the bandwidth or ∆= the relative or fractional bandwidth
Ω0

• Apply the same transform to all S-terms in all circuit elements in the S-
domain circuit.
6.5 Band-Reject Filters

6.5.1 Low-Pass to Band-Reject Transform


𝑆↔
𝑆 Ω
[Ω + 𝑆0 ]
0

• This is very similar to the 𝐿𝑃 ↔ 𝐵𝑃 transform with the following result

∆Ω0 ∆Ω0 2
𝜔 = −( √
)± ( ) + 4Ω20
2𝜔0 2𝜔0
6.6 Sallen-Key Active Filters

6.6.1 Active Filters

• For low frequencies (< 1 𝑀𝐻𝑧 )


o Capacitance and Inductance values of typical filters are very large
o There is a high ratio between the smallest and largest value
• Filters using operational amplifiers can be used very effectively in this
frequency range
o Such filters are called active filters
• Active filters are simple to design
o But stuffer from
▪ Sensitivity problems
▪ Instability (Oscillation)

• The most basic active section is

1
o If we now choose 𝑍1 = 𝑅1, 𝑍2 = 1
+𝑆𝐶2
𝑅2

𝑅
𝑉𝑜𝑢𝑡 ( 2)
𝑅1
o 𝐻(𝑆) = = − 𝑆[𝑅
𝑉𝑖𝑛 2 𝐶2 ]+1
o Which is a first order lowpass response
▪ With a cut-off frequency of
1
• 𝜔=𝑅
2 𝐶2

▪ And a lowpass gain of


𝑅
• − (𝑅2)
1

o While very simple, this standard active filter is very sensitive to the
characteristics of the operational amplifier.

6.6.2 Sallen-Key Sections

• Sallen-Key sections are second order sections


• For filter design, we introduce the 𝑄-Factor as
𝜔
o 𝑆 2 + [2𝜁𝜔𝑛 ]𝑆 + 𝜔𝑛2 = 𝑆 2 + [ 𝑄𝑛] 𝑆 + 𝜔𝑛2
1
o 𝑄 = 2𝜁

6.6.2.1 Low-Pass 2𝑛𝑑 order function

𝜔0 1
𝐻(𝑆) = 𝜔 = 2
𝑆 2 + [ 𝑄0 ] 𝑆 + 𝜔02 [ 𝑆 ] + [ 2𝜁 ] 𝑆 + 1
𝜔𝑛 𝜔𝑛

1
𝜔0 = 𝜔𝑛 and 𝑄 = 2𝜁
6.6.2.2 High-Pass 2𝑛𝑑 order function

6.6.2.3 Bandpass 2𝑛𝑑 order function

6.6.2.4 Higher-Order Filters


• A first order lowpass filter can be realized as

• If we need a 2𝑛𝑑 order filter with poles at 𝜔1 , 𝜔2 we cannot simply cascade a


1 1
section with a section with 𝜔1 = 𝑅 and a section with 𝜔2 = 𝑅 , as the
1 𝐶1 2 𝐶2

second section will load the first

• We can cascade sections if each section is


o Excited with a source with no output resistance
o Loaded by an infinite impedance
• This is called buffering
• An ideal operational amplifier is an ideal buffer as it has
o Zero output resistance
o Infinite input resistance
• Sallen-Key sections can be cascaded without any section changing the other
sections.

• To design higher order active filters, we therefore


o Calculate transfer function (even order)
o Express the transfer function as the product of second order functions
o Design a Sallen-Key section for each second order function
o Cascade the Sallen-Key sections
• Note if we had an odd-order function we set the above equation equal to 0!
7 The Fourier Series
• Any periotic function can be expanded into an infinite series of sine and
cosine terms

𝑓(𝑡) = 𝑎𝑣 + ∑[𝑎𝑘 cos(𝑘𝜔0 𝑡) + 𝑏𝑘 sin(𝑘𝜔0 𝑡)]


𝑘=1

• Where
2𝜋
o 𝜔0 = 𝑇

o 𝑇 is the period
• 𝑎𝑣 , 𝑎𝑘 , 𝑏𝑘 are calculated by projection

𝑡0 +𝑇
1
𝑎𝑣 = ∫ 𝑓(𝑡) 𝑑𝑡
𝑇
𝑡0

𝑡0 +𝑇
2
𝑎𝑘 = ∫ 𝑓(𝑡) cos(𝑘𝜔0 𝑡) 𝑑𝑡
𝑇
𝑡0

𝑡0 +𝑇
2
𝑏𝑘 = ∫ 𝑓(𝑡) sin(𝑘𝜔0 𝑡) 𝑑𝑡
𝑇
𝑡0

7.1 Approximation Theory


𝑓(𝑥) = ∑ 𝑎𝑛 𝑓𝑛 (𝑥)
𝑛=1

• If 𝑓𝑛 (𝑥) is orthogonal, 𝑎𝑛 can be calculated by projection

• Applications:
o Solving complex responses for a set of basic functions analytically
o Approximating functions with low-order functions (surrogate models)

7.2 Fourier-Series and Signals

• Basic form:

𝑓(𝑡) = 𝑎𝑣 + ∑[𝑎𝑘 cos(𝑘𝜔0 𝑡) + 𝑏𝑘 sin(𝑘𝜔0 𝑡)]


𝑘=1

o 𝑓(𝑡) single valued, with a finite number of discontinuities


2𝜋
o 𝑓(𝑡) is periotic with 𝑇, where 𝜔0 = 𝑇
𝑡 +𝑇
o ∫𝑡 0 |𝑓(𝑡)| 𝑑𝑡 exists
0

• 𝜔0 is called the fundamental frequency


• 𝑘𝜔0 is called the 𝑘 𝑡ℎ harmonic
• The Fourier-Series is the summation of different frequency components

o We can say 𝑓(𝑡) has two frequency components


▪ One at 𝜔0
▪ One at 3𝜔0
7.3 Symmetry
• If

𝑓(𝑡) = 𝑎𝑣 + ∑[𝑎𝑘 cos(𝑘𝜔0 𝑡) + 𝑏𝑘 sin(𝑘𝜔0 𝑡)]


𝑘=1

• If 𝑓(𝑡) has even symmetry


o 𝑓(𝑡) = 𝑓(−𝑡)

𝑇
2
2
𝑎𝑣 = ∫ 𝑓(𝑡) 𝑑𝑡
𝑇
0

𝑇
2
4
𝑎𝑘 = ∫ 𝑓(𝑡) cos(𝑘𝜔0 𝑡) 𝑑𝑡
𝑇
0

𝑏𝑘 = 0

• If 𝑓(𝑡) has odd symmetry


o 𝑓(𝑡) = −𝑓(−𝑡)
𝑎𝑣 = 𝑎𝑘 = 0
𝑇
2
4
𝑏𝑘 = ∫ 𝑓(𝑡) sin(𝑘𝜔0 𝑡) 𝑑𝑡
𝑇
0

• If 𝑓(𝑡) has half-wave symmetry


𝑇
o 𝑓(𝑡) = −𝑓 (𝑡 − 2)
𝑇
o 𝑓(𝑡) →shift right by 2 → invert

o 𝑎𝑣 = 0
o 𝑎𝑘 = 𝑏𝑘 = 0, for 𝑘 even

• If 𝑓(𝑡) has quarter-wave symmetry


𝑇
o Half-wave symmetry + symmetry around 4 point
o 𝑎𝑣 = 𝑎𝑘 = 0
o 𝑏𝑘 = 0 for 𝑘 even

7.4 Alternative Forms of Fourier-Series

7.4.1 Standard Form:


𝑓(𝑡) = 𝑎𝑣 + ∑[𝑎𝑘 cos(𝑘𝜔0 𝑡) + 𝑏𝑘 sin(𝑘𝜔0 𝑡)]


𝑘=1

7.4.2 Trigonometric Form:


𝑓(𝑡) = 𝑎𝑣 + ∑ 𝐴𝑛 cos(𝑛𝜔0 𝑡 − 𝜃𝑛 )
𝑛=1

𝑎𝑘 − 𝑗𝑏𝑘 = 𝐴𝑛 ∠ − 𝜃𝑛

∴ 𝐴𝑛 = |𝑎𝑘 − 𝑗𝑏𝑘 |
𝑏𝑘
𝜃𝑛 = tan−1 ( )
𝑎𝑘
7.4.3 Exponential Form:

𝑓(𝑡) = ∑ 𝐶𝑛 𝑒 𝑗𝑛𝜔0𝑡
𝑛=−∞

𝑡0 +𝑇
1
𝐶𝑛 = ∫ 𝑓(𝑡)𝑒 −𝑗𝑛𝜔𝑛𝑡 𝑑𝑡
𝑇
𝑡0

1 𝐴𝑛
𝐶𝑛 = (𝑎𝑘 − 𝑗𝑏𝑘 ) = ∠−𝜃
2 2

7.5 Circuit Analysis


• If

𝑣(𝑡) = 𝑉𝑑𝑐 + ∑ 𝑉𝑛 cos(𝑛𝜔0 𝑡 − 𝜃𝑣𝑛 )


𝑛=1

𝑖(𝑡) = 𝐼𝑑𝑐 + ∑ 𝐼𝑛 cos(𝑛𝜔0 𝑡 − 𝜃𝑖𝑛 )


𝑛=1


𝑉𝑛 𝐼𝑛
𝑃𝑎𝑣 = 𝑉𝑑𝑐 𝐼𝑑𝑐 + ∑ cos(𝜃𝑣𝑛 − 𝜃𝑖𝑛 )
2
𝑛=1


2
𝑉𝑛2
𝑉𝑟𝑚𝑠 = √𝑉𝑑𝑐 +∑( )
2
𝑛=1

• For a circuit with a transfer function


𝑉𝑜𝑢𝑡 (𝑆)
o 𝐻(𝑆) = 𝑉𝑖𝑛 (𝑆)

• With an input of 𝑉𝑖𝑛 (𝑡)


o Where

𝑉𝑖𝑛(𝑡) = 𝑎𝑣 + ∑[𝑎𝑘 cos(𝑘𝜔0 𝑡) + 𝑏𝑘 sin(𝑘𝜔0 𝑡)]


𝑘=1

• The output is simply

𝑉𝑜𝑢𝑡 (𝑡) = 𝑎𝑣 𝐻0 + 𝐴𝑘 ∑[𝑎𝑘 cos(𝑘𝜔0 𝑡 + 𝜃𝑘 ) + 𝑏𝑘 sin(𝑘𝜔0 𝑡 + 𝜃𝑘 )]


𝑘=1

o Where
▪ 𝐴𝑘 = |𝐻(𝑗𝑘𝜔)|
▪ 𝜃𝑘 = ∠𝐻(𝑗𝑘𝜔)

7.6 Fourier Spectrum


𝑓(𝑡) = ∑ 𝐶𝑛 𝑒 𝑗𝑛𝜔0𝑡
𝑛=−∞

• 𝐶𝑛 are the complex values of each frequency component of 𝑓(𝑡)


• 𝐶𝑛 drawn as a function of 𝜔 is called the spectrum of 𝑓(𝑡)

𝐴𝑛 = |𝑎𝑛 − 𝑗𝑏𝑛 |
𝜃𝑛 = tan−𝑡 (𝑏𝑛 /𝑎𝑛 )
7.7 Fourier Series of an impulse

𝑓(𝑡) = ∑ 𝛿(𝑡 − 𝑛𝑇)


𝑇
2
1 1
𝑎𝑣 = ∫ 𝛿(𝑡) =
𝑇 𝑇
𝑇

2

𝑇 𝑇
2 2
2 2 2
𝑎𝑘 = ∫ 𝛿(𝑡) cos(𝑘𝜔0 𝑡) 𝑑𝑡 = ∫ 𝛿(𝑡) cos(0) 𝑑𝑡 =
𝑇 𝑇 𝑇
𝑇 𝑇
− −
2 2

𝑇 𝑇
2 2
2 2
𝑏𝑘 = ∫ 𝛿(𝑡) sin(𝑘𝜔0 𝑡) 𝑑𝑡 = ∫ 𝛿(𝑡) sin(0) 𝑑𝑡 = 0
𝑇 𝑇
𝑇 𝑇
− −
2 2

• Therefore
∞ ∞
1 2
∑ 𝛿(𝑡 − 𝑛𝑇) = + ∑ cos(𝑘𝜔0 𝑡)
𝑇 𝑇
𝑛=−∞ 𝑘=1
7.7.1 Example
8 DC to DC converter
• This is an example of pulse width modulation

• Express 𝑉𝑖𝑛 (𝑡) as a Fourier-Series


9 Fourier Transform

𝐹(𝜔) = ℱ[𝑓(𝑡)] = ∫ 𝑓(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡


−∞


1
𝑓(𝑡) = ℱ −1 [𝐹(𝜔)] = ∫ 𝐹(𝜔)𝑒 𝑗𝜔𝑡 𝑑𝜔
2𝜋
−∞

There is a important tables document on the drive with a table of Fourier Transforms
and Inverse Fourier Transforms

9.1 Convergence
• The Fourier Integral converges when

o ∫−∞|𝑓(𝑡)| 𝑑𝑡 exists
o All discontinues are finite

• If not, we often use the transform of a function which approximates the actual
function in the limit.
9.2 Link between Laplace and Fourier

• If 𝑓(𝑡) = 𝑓1 (𝑡) + 𝑓2 (𝑡)


o With
▪ 𝑓1 (𝑡) = 0 for 𝑡 ≤ 0
▪ 𝑓2 (𝑡) = 0 for 𝑡 ≥ 0
• 𝐹(𝜔) = 𝐹1 (𝜔) + 𝐹2 (𝜔)
= ℒ[𝑓1 (𝑡)]𝑠=𝑗𝜔 + ℒ[𝑓2 (−𝑡)]𝑠=−𝑗𝜔
9.3 Important Properties

𝐹(𝜔) = ∫ 𝑓(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡


−∞

= ∫ 𝑓(𝑡)(cos(𝜔𝑡) − 𝑗𝑠𝑖𝑛(𝜔𝑡)) 𝑑𝑡
−∞
∞ ∞

= ∫ 𝑓(𝑡) cos(𝜔𝑡) 𝑑𝑡 − 𝑗 ∫ 𝑓(𝑡)sin (𝜔𝑡) 𝑑𝑡


−∞ −∞

= 𝐴(𝜔) + 𝑗𝐵(𝜔)
= |𝐹(𝜔)|𝑒 𝑗𝜃(𝜔)

• 𝐴(𝜔) → 𝑒𝑣𝑒𝑛
• 𝐵(𝜔) → 𝑜𝑑𝑑
• 𝐹(−𝜔) = 𝐹 ∗ (𝜔)
• |𝐹(−𝜔)| = |𝐹(𝜔)| (𝑒𝑣𝑒𝑛)
• 𝜃(−𝜔) = −𝜃(𝜔) (𝑜𝑑𝑑)
• ∴ If 𝑓(𝑡) is even, 𝐵(𝜔) = 0 or 𝐹(𝜔) is real
• ∴ If 𝑓(𝑡) is odd, 𝐴(𝜔) = 0 or 𝐹(𝜔) is imaginary
9.4 Fourier Transform of derivative
9.5 Modulation

9.6 Convolution
• If

𝑦(𝑡) = 𝑥(𝑡) ⊗ ℎ(𝑡) = ∫ 𝑥(𝜆)ℎ(𝑡 − 𝜆)𝑑𝜆


−∞

𝑦(𝜔) = ℱ[𝑦(𝑡)] = 𝑋(𝜔)𝐻(𝜔)

• If
𝑓(𝑡) = 𝑓1 (𝑡)𝑓2 (𝑡)

1
𝐹(𝜔) = ∫ 𝐹1 (𝑢)𝐹2 (𝜔 − 𝑢)𝑑𝑢
2𝜋
−∞

9.7 Example of transient response


9.8 Example of Steady-State response
10 Parseval’s Theorem

• If 𝑊 is the total energy in a signal 𝑓(𝑡)



1
𝑊= ∫ |𝐹(𝜔)|2 𝑑𝜔
2𝜋
−∞

Or


1
𝑊 = ∫ |𝐹(𝜔)|2 𝑑𝜔
𝜋
0

• |𝐹(𝜔)|2 is called an energy density


• Total energy in a signal is the integral of the spectrum magnitude squared.

You might also like