Fourier series

Jean Baptiste Joseph Fourier (1768-1830)

Born: 21 March 1768 in Auxerre, Bourgogne, France
Died: 16 May 1830 in Paris, France
• Joseph’s father was a tailor in Auxerre
• Joseph was the ninth of twelve children
• His mother died when he was nine and his father died the following
year Fourier demonstrated talent on math at the age of 14.
• In 1787 Fourier decided to train for the priesthood - a religious life
or a
• In 1793, Fourier joined the local Revolutionary Committee

Fourier studied the mathematical theory of heat conduction. He

established the partial differential equation governing heat diffusion and
solved it by using infinite series of trigonometric functions.

You have met in earlier Mathematics courses the concept of representing a function by an infinite series
of simpler functions such as polynomials. For example, the McLaurin series representing ex has the form

𝑥2 𝑥3
𝑒𝑥 = 1 + 𝑥 + + +⋯
2! 3!

or, in the more concise sigma notation,

∞
𝑥𝑛
∑ 𝑒𝑥 =
𝑛!
𝑛=0

(Remembering that 0! is defined as 1).

The basic idea is that for those values of x for which the series converges we may approximate the
function by using only the first few terms of the series.

Fourier series, which we discuss in this and the following Sections, are also usually infinite series but
involve sine and cosine functions (or their complex exponential equivalents) rather than polynomials.
They are widely used for approximating periodic functions. Such approximations are of considerable use
in science and engineering.

For example, elementary a.c. theory provides techniques for analyzing electrical circuits when the
currents and voltages present are assumed to be sinusoidal.
Fourier Series enable us to extend such techniques to the situation where the functions (or signals)
involved are periodic but not actually sinusoidal.

You may also see later that Fourier series sometimes have to be used when solving partial differential
equations.

Periodic Functions

A function f (t) is periodic if the function values repeat at regular intervals of the independent variable t.
The regular interval is referred to as the period.

If 𝑇 denotes the period we have

𝑓(𝑡 + 𝑇) = 𝑓(𝑡), for any value of 𝑡.

The most obvious examples of periodic functions are the trigonometric functions 𝑠𝑖𝑛 𝑡 and 𝑐𝑜𝑠 𝑡, both of
which have period 2𝜋 (using radian measure as we shall do throughout this chapter.) This follows since
𝑠𝑖𝑛 (𝑡 + 2𝜋) = 𝑠𝑖𝑛 𝑡 and cos(𝑡 + 2𝜋) = cos 𝑡

The amplitude of these sinusoidal functions is the maximum displacement from y = 0 and is clearly 1.
(Note that we use the term sinusoidal to include cosine as well as sine functions.) More generally we can
consider a sinusoid

𝑦 = 𝐴 sin 𝑛𝑡

which has maximum value, or amplitude, A and where n is usually a positive integer. For example
 𝑦 = sin 2𝑡
2𝜋
is a sinusoid of amplitude 1 and period 2
= 𝜋. The fact that the period is 𝜋 follows because

sin 2(𝑡 + 𝜋) = sin(2𝑡 + 2𝜋) = sin 2𝑡

for any value of t.

In general,
𝑦 = 𝐴 sin 𝑛𝑡
2𝜋
Has amplitude A, period and completes n oscillations when t changes by 2𝜋.
𝑛

Definition: we define the frequency of a sinusoid as the reciprocal of the period:

frequency =1/period

and the angular frequency (often denoted the Greek Letter 𝜔 (omega)) as

2𝜋
𝑎𝑛𝑔𝑢𝑙𝑎𝑟 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 = 2𝜋 × 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 =
𝑝𝑒𝑟𝑖𝑜𝑑

Thus

𝑦 = 𝐴 sin 𝑛𝑡
𝑛
Has frequency 2𝜋
and regular frequency n

.
Question: State the amplitude, period, frequency and angular frequency of
2𝑡
𝐼) 𝑦 = 5 cos 4𝑡 , 𝐼𝐼) 𝑦 = 6 sin
3

Harmonics

In representing a non-sinusoidal function of period 2𝜋 by a Fourier Series we shall see shortly that only
certain sinusoids will be required:

 𝐴1 𝑐𝑜𝑠 𝑡 (and 𝐵1 𝑠𝑖𝑛 𝑡) These also have period 2𝜋 and together are referred to as the first (or
fundamental) harmonic.

 𝐴2 𝑐𝑜𝑠 2𝑡 (and 𝐵2 𝑠𝑖𝑛 2𝑡) These have half the period, or double the frequency of the first
harmonic and are referred to as the second harmonic.
2𝜋
 𝐴3 𝑐𝑜𝑠 3𝑡 (and 𝐵3 𝑠𝑖𝑛 3𝑡) These have period 3
and constitute the third harmonic.

In general, the Fourier Series of a function of period 2𝜋 will require harmonics of the type 𝐴𝑛 𝑐𝑜𝑠 𝑛𝑡 (and
𝐵𝑛 𝑠𝑖𝑛 𝑛𝑡) where n = 1, 2, 3, . . .

Non-sinusoidal Periodic Functions

The following are examples of such functions (they are often called 'waves'):

Square Wave

Analytically we can describe this function as follows:

−1 −𝜋 < 𝑡 < 0
𝑓(𝑡) = {
+1 0 < 𝑡 < 𝜋

(which gives the definition over one period.)

𝑓(𝑡 + 2𝜋) = 𝑓(𝑡)

(which tells us that the function has period 2𝜋).

Saw-Tooth Wave

In this case we can describe the function as

follows:

𝑓(𝑡) = 2𝑡 , 0 < 𝑡 < 2

𝑓(𝑡 + 2) = 𝑓(𝑡)

2𝜋
Here the period is 2, the frequency is 1/2 and the angular frequency is 2
=𝜋

Triangular Wave

Here we can conveniently define the function using

−𝜋 < 𝑡 < 𝜋 as the basic period:

−𝑡 −𝜋 < 𝑡 < 0
𝑓(𝑡) = {
+𝑡 0<𝑡<𝜋
Or, more concisely, 𝑓(𝑡) = |𝑡|, −𝜋 < 𝑡 < 𝜋
together with the usual statement on periodicity

𝑓(𝑡 + 2𝜋) = 𝑓(𝑡)

Questions:
1- Write down an analytic definition for the following periodic function:

2- Sketch the graph of the following periodic functions showing all relevant values:

𝑓(𝑡) = 2𝑡 − 𝑡 2 , 0 < 𝑡 < 2, 𝑓(𝑡 + 2) = 𝑓(𝑡)

𝑡2
2
0<𝑡<4
𝑓(𝑡) = { 8 4 < 𝑡 < 6, 𝑓(𝑡 + 8) = 𝑓(𝑡)
0 6<𝑡<8
Introduction

We recall first a simple trigonometric identity:

1 1
cos 2𝑡 = 2 𝑐𝑜𝑠 2 𝑡 − 1 ⟺ 𝑐𝑜𝑠 2 𝑡 = + cos 2𝑡 (1)
2 2
Equation 1 can be interpreted as a simple finite Fourier Series representation of the periodic function
𝑓(𝑡) = 𝑐𝑜𝑠 2 𝑡 which has period 2𝜋.

We just note that the Fourier Series representation contains a constant term and a period 𝜋 term.

A more complicated trigonometric identity is

3 1 1
𝑠𝑖𝑛4 𝑡 = − cos 2𝑡 + cos 4𝑡
8 2 8

which again can be considered as a finite Fourier Series representation. (Do not worry if you are
unfamiliar with the result (2).)

Note that the function 𝑓(𝑡) = 𝑠𝑖𝑛4 𝑡 (which has period 𝜋) is being written in terms of a constant function,
𝜋 2
a function of period 𝜋 or frequency 1/𝜋 (the first harmonic) and a function of period or frequency (the
2 𝜋
second harmonic).

The reason for the constant term in both (1) and (2) is that each of the functions 𝑐𝑜𝑠 2 𝑡 and 𝑠𝑖𝑛4 𝑡 is non-
negative and hence each must have a positive average value.

Any sinusoid of the form 𝑐𝑜𝑠 𝑛𝑡 or 𝑠𝑖𝑛 𝑛𝑡 has, by symmetry, zero average value as, therefore, would a
Fourier Series containing only such terms.

A constant term can therefore be expected to arise in the Fourier Series of a function which has a non-
zero average value.

Functions of Period 𝟐𝝅

We now discuss how to represent periodic non-sinusoidal functions 𝑓 (𝑡) of period 2𝜋 in terms of
sinusoids, i.e. how to obtain Fourier Series representations.

As already discussed we expect such Fourier Series to contain harmonics of frequency

𝑛
, (𝑛 = 1,2,3, … )
2𝜋

and, if the periodic function has a non-zero average value, a constant term.

Definition: A Fourier Series representation of the general form

 ∞
𝑎0 𝑎0
𝑓(𝑡) = + 𝑎1 cos 𝑡 + 𝑎2 cos 2𝑡 + ⋯ . +𝑏1 sin 𝑡 + 𝑏2 sin 2𝑡 + ⋯ = + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)
2 2
𝑛=1
𝑎0
The reason for labelling the constant term as 2
will be discussed later.

The amplitudes 𝑎1 , 𝑎2 , … . , 𝑏1 , 𝑏2 , … of the sinusoids are called Fourier coefficients.

Obtaining the Fourier coefficients for a given periodic function 𝑓 (𝑡) is our main task and is referred to as
Fourier Analysis.

Before embarking on such an analysis it is instructive to establish, at least qualitatively, the plausibility of
approximating a function by a few terms of its Fourier Series.

Question: Consider the square wave of period 2𝜋 one period of which is shown in Figure

Write down:
a) the analytic description of this function,

b) whether you expect the Fourier Series of this function to contain a constant term, c) any other possible
features of the Fourier Series that you might expect from the graph of the square-wave function.

To be precise it is possible to show, and we will do so later, that the Fourier Series representation of this
square wave is
𝟖 𝟏 𝟏 𝟏
𝟐 + {𝒄𝒐𝒔 𝒕 − 𝒄𝒐𝒔 𝟑𝒕 + 𝒄𝒐𝒔 𝟓𝒕 − 𝒄𝒐𝒔 𝟕𝒕 + ⋯ }
𝝅 𝟑 𝟓 𝟕
i.e. the Fourier coefficients are
𝑎0 8 8 8
= 2, 𝑎1 = , 𝑎2 = 0, 𝑎3 = − , 𝑎4 = 0, 𝑎5 = ,…
2 𝜋 3𝜋 5𝜋

Note: as well as the presence of the constant term and of the cosine (but not sine) terms, that only odd
2𝜋 2𝜋 2𝜋
harmonics are present i.e. sinusoids of period 2𝜋, 3 , 5 , 7 , … or of frequency 1,3,5,7, … times the
1
fundamental Frequency 2𝜋.
We now show in the next figure graphs (for 0 < 𝑡 < 𝜋 only since the square wave and its Fourier
Series are even) of

a) the square wave

b) the first two terms of the Fourier Series

c) the first three terms of the Fourier Series

d) the first four terms of the Fourier Series

e) the first five terms of the Fourier Series

We can clearly see from the figure that as the number of terms is increased the graph of the Fourier Series
gradually approaches that of the original square wave - the ripples increase in number but decrease in
amplitude.
𝜋
The behaviour near the discontinuity, at 𝑡 = 2 , is slightly more complicated and it is possible to show
that however many terms are taken in the Fourier Series, some "overshoot" will always occur.

Orthogonality Properties of Sinusoids

As stated earlier, a periodic function 𝑓(𝑡) with period 2𝜋 has a Fourier Series representation
𝑎0
𝑓(𝑡) = + 𝑎1 cos 𝑡 + 𝑎2 cos 2𝑡 + ⋯ . +𝑏1 sin 𝑡 + 𝑏2 sin 2𝑡 + ⋯
2
∞
𝑎0
= + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡) (3)
2
𝑛=1
𝑎
To determine the Fourier coefficients 𝑎𝑛 , 𝑏𝑛 and the constant term 20 use has to be made of certain
integrals involving sinusoids, the integrals being over a range 𝛼, 𝛼 + 2𝜋, where _ is any number. (We will
normally choose 𝛼 = −𝜋).
𝜋 𝜋
Question: find ∫−𝜋 sin 𝑛𝑡 𝑑𝑡 and ∫−𝜋cos 𝑛𝑡 𝑑𝑡 where n is an integer.
𝜋 𝜋
find ∫−𝜋 sin 𝑛𝑡 𝑑𝑡 and ∫−𝜋cos 𝑛𝑡 𝑑𝑡 where n is an irrational number.

Notice: If we take integrate both side of (3) from −𝜋 to 𝜋, we can find 𝑎0 .

Notice: the constant term 𝑎0 in a trigonometric Fourier Series for a function of period 2𝜋 is

𝑎0 1 𝜋
2
= 2𝜋 ∫−𝜋 𝑓(𝑡) 𝑑𝑡 = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑓(𝑡) 𝑜𝑣𝑒𝑟 1 𝑝𝑒𝑟𝑖𝑜𝑑

Question:
𝜋
1- evaluate ∫−𝜋 sin 𝑛𝑡 cos 𝑚𝑡 𝑑𝑡 where n and m are any integers.
1
(sin 𝑛𝑡 cos 𝑚𝑡 = 2 [sin(𝑛 + 𝑚) 𝑡 + sin(𝑛 − 𝑚)𝑡])
𝜋
2- evaluate ∫−𝜋cos 𝑛𝑡 cos 𝑚𝑡 𝑑𝑡 where n and m are any integers
1
(cos 𝑛𝑡 cos 𝑚𝑡 = 2 [cos(𝑛 + 𝑚) 𝑡 + cos(𝑛 − 𝑚)𝑡])
𝜋
3- evaluate ∫−𝜋 sin 𝑛𝑡 sin 𝑚𝑡 𝑑𝑡 where n and m are any integers
1
(sin 𝑛𝑡 sin 𝑚𝑡 = [cos(𝑛 − 𝑚) 𝑡 − cos(𝑛 + 𝑚)𝑡])
2

Review: for integers n and m

𝜋
∫ sin 𝑛𝑡 cos 𝑚𝑡 𝑑𝑡 = 0
−𝜋

0 𝑛≠𝑚
𝜋
∫−𝜋cos 𝑛𝑡 cos 𝑚𝑡 𝑑𝑡 = { 𝜋 𝑛=𝑚≠0
2𝜋 𝑛=𝑚=0
𝜋
0 𝑛 ≠ 𝑚, 𝑛 = 𝑚 = 0
∫ sin 𝑛𝑡 sin 𝑚𝑡 𝑑𝑡 = {
−𝜋 𝜋 𝑛=𝑚
 Calculation of Fourier coefficients

Consider the Fourier Series for a function f (t) of period 2𝜋:

∞
𝑎0
𝑓(𝑡) = + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)
2
𝑛=1

To obtain the coefficients an (n =1,2,3,… ), we multiply both sides by cos mt where m is some positive
integer and integrate both sides from −𝜋 to 𝜋:

𝜋 ∞
𝑎0 𝜋 𝜋 𝜋
∫ 𝑓(𝑡) cos 𝑚𝑡 𝑑𝑡 = ∫ cos 𝑚𝑡 𝑑𝑡 + ∑(𝑎𝑛 ∫ cos 𝑚𝑡 cos 𝑛𝑡 𝑑𝑡 + 𝑏𝑛 ∫ cos 𝑚𝑡 sin 𝑛𝑡 )
−𝜋 2 −𝜋 −𝜋 −𝜋
𝑛=1

Using the orthogonality relations all the integrals in the summation give zero except for the case n = m
when, from the last review

𝜋 𝜋
1 𝜋
∫ cos2 𝑚𝑡 𝑑𝑡 = 𝜋 ⇒ ∫ 𝑓(𝑡) cos 𝑚𝑡 𝑑𝑡 = 𝑎𝑚 𝜋 ⇒ 𝑎𝑚 = ∫ 𝑓(𝑡) cos 𝑚𝑡 𝑑𝑡 𝑚 = 1,2,3 ….
−𝜋 −𝜋 𝜋 −𝜋

Similarly
1 𝜋
𝑏𝑚 = ∫ 𝑓(𝑡) sin 𝑚𝑡 𝑑𝑡 , 𝑚 = 1,2,3, …
𝜋 −𝜋

Summery: A function f (t) with period 2𝜋 has a Fourier Series

∞
𝑎0
𝑓(𝑡) = + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)
2
𝑛=1

The Fourier coefficients are

1 𝜋
𝑎𝑛 = ∫ 𝑓(𝑡) cos 𝑛𝑡 𝑑𝑡 𝑛 = 0,1,2,3 ….
𝜋 −𝜋
1 𝜋
𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡 , 𝑛 = 1,2,3, …
𝜋 −𝜋

In the integrals any convenient integration range 𝛼, 𝛼 + 2𝜋 may be used.

Example: We shall obtain the Fourier Series of the half-rectified square wave shown.
We have
1 0<𝑡<𝜋
𝑓(𝑡) = { , 𝑓(𝑡 + 2𝜋) = 𝑓(𝑡)
0 𝜋 < 𝑡 < 2𝜋

1 𝜋 1 𝜋
𝑎𝑛 = ∫ 𝑓(𝑡) cos 𝑛𝑡 𝑑𝑡 = ∫ 1 cos 𝑛𝑡 𝑑𝑡 = 0, 𝑛 = 1,2,3, …
𝜋 −𝜋 𝜋 0
1 𝜋
𝑎0 = ∫ 1 𝑑𝑡 = 1 →
𝜋 0

1 𝜋 1 𝜋 1 0 𝑛 = 2,4,6, …
𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡 = ∫ 1 sin 𝑛𝑡 𝑑𝑡 = (1 − cos 𝑛𝜋) = { 2
𝜋 −𝜋 𝜋 0 𝑛𝜋 𝑛 = 1,3,5, …
𝑛𝜋
∞
𝑎0 1 1
𝑓(𝑡) = + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡) = 1/2 + 2/𝜋 (sin 𝑡 + sin 3𝑡 + sin 5𝑡 + ⋯ . )
2 3 5
𝑛=1
∞ ∞
1 2 1 1 2 1
𝑓(𝑡) = + ∑ sin 𝑛𝑡 = + ∑ sin(2𝑘 − 1)𝑡
2 𝜋 𝑛 2 𝜋 2𝑘 − 1
𝑛=1,(𝑛 𝑜𝑑𝑑) 𝑘=1

Question: Obtain the Fourier Series of the square wave one period of which is shown:
Fourier Series for Functions of General Period
This is a straightforward extension of the period 2𝜋 case that we have already discussed.

if function 𝑓 (𝑡) with period 2𝜋 has a Fourier Series

∞
𝑎0
𝑓(𝑡) = + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)
2
𝑛=1

The Fourier coefficients are

1 𝜋
𝑎𝑛 = ∫ 𝑓(𝑡) cos 𝑛𝑡 𝑑𝑡 𝑛 = 0,1,2,3 ….
𝜋 −𝜋
1 𝜋
𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡 , 𝑛 = 1,2,3, …
𝜋 −𝜋

Suppose we now change the variable to t where

2𝜋
𝑥= 𝑡
𝑇
𝑇 𝑇
Thus 𝑥 = 𝜋 corresponds to 𝑡 = and 𝑥 = −𝜋 corresponds to 𝑡 = −
2 2

Hence regarded as a function of 𝑡, we have a function with period 𝑇.

2𝜋 2𝜋
Making the substitution 𝑥 = 𝑇
𝑡 , and hence 𝑑𝑥 = 𝑇
𝑑𝑡, in the expressions for 𝑎𝑛 and 𝑏𝑛 we obtain

𝑇
2 2 𝑛𝜋𝑡
𝑎𝑛 = ∫ 𝑓(𝑡) cos( ) 𝑑𝑡 𝑛 = 0,1,2,3 ….
𝑇 − 𝑇 𝑇
2
𝑇
2 2 𝑛𝜋𝑡
𝑏𝑛 = ∫ 𝑓(𝑡) sin( ) 𝑑𝑡 , 𝑛 = 1,2,3, …
𝑇 − 𝑇 𝑇
2

These integrals give the Fourier coefficients for a function of period 𝑇 whose Fourier Series is
∞
𝑎0 2𝑛𝜋𝑡 2𝑛𝜋𝑡
𝑓(𝑡) = + ∑(𝑎𝑛 cos( ) + 𝑏𝑛 sin( ))
2 𝑇 𝑇
𝑛=1

Various other notations are commonly used in this case e.g. it is sometimes convenient to write the period
𝑇 = 2𝐿 .
 2𝜋
Another alternative is to use the angualr frequency 𝜔 and put 𝑇 = 𝜔
.

Example: Find the Fourier Series of the function shown in the figure, viz a saw tooth wave with
alternative portions removed

Solution: the period 𝑇 = 2𝐿 = 4then 𝐿 = 2. The Fourier Series will have the form
∞
𝑎0 𝑛𝜋𝑡 𝑛𝜋𝑡
𝑓(𝑡) = + ∑(𝑎𝑛 cos( ) + 𝑏𝑛 sin( ))
2 2 2
𝑛=1

The Fourier coefficients are

1 2 𝑛𝜋𝑡
𝑎𝑛 = ∫ 𝑓(𝑡) cos( ) 𝑑𝑡 𝑛 = 0,1,2,3 ….
2 −2 2

1 2 𝑛𝜋𝑡
𝑏𝑛 = ∫ 𝑓(𝑡) sin( ) 𝑑𝑡 , 𝑛 = 1,2,3, …
2 −2 2

We have
0 −2 < 𝑡 < 0
𝑓(𝑡) = {
𝑡 0<𝑡<2
Then

1 2 𝑛𝜋𝑡 1 2 𝑛𝜋𝑡 2 0 𝑛 = 2,4,6, …

𝑎𝑛 = ∫ 𝑓(𝑡) cos( ) 𝑑𝑡 = ∫ 𝑡 cos( ) 𝑑𝑡 = 2 2 {cos 𝑛𝜋 − 1] = { 4
2 −2 2 2 0 2 𝑛 𝜋 − 2 2 𝑛 = 1,3,5, …
𝑛 𝜋
1 2 𝑛𝜋𝑡 1 2
𝑎0 = ∫ 𝑓(𝑡) cos( ) 𝑑𝑡 = ∫ 𝑡 𝑑𝑡 = 1
2 −2 2 2 0
2
1 2 𝑛𝜋𝑡 1 2 𝑛𝜋𝑡 2 − 𝑛 = 2,4,6, …
𝑏𝑛 = ∫ 𝑓(𝑡) sin( ) 𝑑𝑡 = ∫ 𝑡 sin( ) 𝑑𝑡 = − cos 𝑛𝜋 = { 𝑛𝜋
2 −2 2 2 0 2 𝑛𝜋 2
𝑛 = 1,3,5, …
𝑛𝜋

Hence, using all these results for the Fourier coefficients, the required Fourier Series is
 1 4 𝜋𝑡 1 3𝜋𝑡 1 5𝜋𝑡 2 𝜋𝑡 1 2𝜋𝑡
𝑓(𝑡) = − 2 {cos ( ) + cos ( ) + cos ( ) + ⋯ } + {sin ( ) − sin ( )
2 𝜋 2 9 2 25 2 𝜋 2 2 2
1 3𝜋𝑡
+ sin ( )…}
3 2

Question: Obtain the Fourier Series of the function

𝑓(𝑡) = 𝑡 2 , −1 < 𝑡 < 1, & 𝑓(𝑡 + 2) = 𝑓(𝑡)

Even and Odd Functions

In this Section we examine how to obtain Fourier series of periodic functions which are either even or
odd.

We show that the Fourier series for such functions is considerably easier to obtain as, if the signal is even
only cosines are involved whereas if the signal is odd then only sines are involved.

We also show that if a signal reverses after half a period then the Fourier series will only contain odd
harmonics.

Even and Odd Functions

We have shown in the previous Section how to calculate, by integration, the coefficients an (n = 0, 1, 2,
3,… ) and bn (n = 1, 2, 3,… ) in a Fourier Series.

Clearly this is a somewhat tedious process and it is advantageous if we can obtain as much information as
possible without recourse to integration.

Referring back to the examples in the previous Section we showed that the square wave had a Fourier
Series containing a constant term and cosine terms (i.e. all the Fourier coefficients bn were zero) while
the sawtooth function had a more complicated Fourier Series containing both cosine and sine terms as
well as a constant.
Question: Contrast the symmetry or otherwise of these two functions.

In general a function is called even if its graph is unchanged under reflection in the y-axis.
This is equivalent to
𝑓 (−𝑡) = 𝑓 (𝑡) 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑡

Obvious examples of even functions are

𝑡 2 , 𝑡 4 ; |𝑡|; 𝑐𝑜𝑠 𝑡; 𝑐𝑜𝑠 2 𝑡; 𝑠𝑖𝑛2 𝑡; 𝑐𝑜𝑠 𝑛𝑡

A function is said to be odd if its graph is symmetrical about the origin. This is to the condition

𝑓 (−𝑡) = − 𝑓 (𝑡) 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑡

The figure shows an example of an odd function.

Examples of odd functions are

𝑡; 𝑡 3 , 𝑡 5 ; |𝑡|; sin 𝑡 ; 𝑠𝑖𝑛 𝑛𝑡

A periodic function which is odd is the saw-tooth wave

Some functions are neither even nor odd. The periodic sawtooth is an example, as is the exponential
function 𝑒 𝑡 .

Question: State whether each of the following periodic functions is even or odd or neither.

A Fourier Series contains a sum of terms while the integral formulae for the Fourier coefficients 𝑎𝑛 and
𝑏𝑛 contain products of the type 𝑓 (𝑡) 𝑐𝑜𝑠 𝑛𝑡 and 𝑓 (𝑡) 𝑠𝑖𝑛 𝑛𝑡 (for the case where 𝑓 (𝑡) has period 2𝜋.)

We need therefore results for sums and products of functions.

Suppose, for example, g(t) is an odd function and h(t) is an even function.

Let 𝐹1 (𝑡) = 𝑔(𝑡)ℎ(𝑡) (product of odd and even functions)

So
𝐹1 (−𝑡) = 𝑔(−𝑡)ℎ(−𝑡) replacing 𝑡 by − 𝑡)
= (−𝑔(𝑡))ℎ(𝑡) since g is odd and h is even
= −𝑔(𝑡)ℎ(𝑡)
= −𝐹(𝑡)
so 𝐹1 (𝑡) is odd.
Now suppose 𝐹2 (𝑡) = 𝑔(𝑡) + ℎ(𝑡) (sum of odd and even functions)
𝐹2 (−𝑡) = 𝑔(−𝑡) + ℎ(−𝑡) = −𝑔(𝑡) + ℎ(𝑡)
So 𝐹2 (𝑡) is neither even nor odd.

Question: Use similar approaches to the above to investigate sums and products of

(i) two odd functions 𝑔1 (𝑡), 𝑔2 (𝑡)

(ii) two even functions ℎ1 (𝑡), ℎ2 (𝑡)

Notice:
products of Functions
 (even)×(even)=even
 (odd)×(even)=odd
 (odd)×(odd)=even

Sums of Functions
 (even)+(even)=even
 (odd)×(even)=neither
 (odd)×(odd)=odd

Useful properties of even and of odd functions in connection with integrals can be readily deduced if we
recall that a definite integral has the significance of giving us the value of an area:

𝑏
∫𝑎 𝑓(𝑡) 𝑑𝑡 gives us the net value of the shaded area, that above the t-axis being positive, that below being
negative.
 𝑎
Question: For the case of a symmetrical interval (−𝑎; 𝑎) deduce what you can about ∫−𝑎 𝑔(𝑡) 𝑑𝑡 and
𝑎
∫−𝑎 ℎ(𝑡) 𝑑𝑡 where g(t) is an odd function and h(t) is an even function.

Fourier Series Implications

Since a sum of even functions is itself an even function it is not unreasonable to suggest that a Fourier
Series containing only cosine terms (and perhaps a constant term which can also be considered as an even
function) can only represent an even periodic function.

Similarly a series of sine terms (and no constant) can only represent an odd function.

These results can readily be shown more formally using the expressions for the Fourier coefficients 𝑎𝑛
and 𝑏𝑛 .

Question: Recall that for a 2𝜋-periodic function

1 𝜋
𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡
𝜋 −𝜋

If 𝑓 (𝑡) is even, deduce whether the integrand is even or odd (or neither) and hence evaluate 𝑏𝑛 . Repeat
for the Fourier coefficients 𝑎𝑛 .

It should be obvious that, for an odd function f (t),

1 𝜋
𝑎𝑛 =∫ 𝑓(𝑡) cos 𝑛𝑡 𝑑𝑡 = 0
𝜋 −𝜋
1 𝜋 2 𝜋
𝑏𝑛 . = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡 = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡
𝜋 −𝜋 𝜋 0

Analagous results hold for functions of any period, not necessarily 2𝜋.

For a periodic function which is neither even nor odd we can expect at least some of both the 𝑎𝑛 and 𝑏𝑛
to be non-zero. For example consider the square wave function:
This function is neither even nor odd and we have already seen in the previous Section that its Fourier
Series contains a constant (1/2) and sine terms.
This result could be expected because we can write
1
𝑓(𝑡) = + 𝑔(𝑡)
2
where g(t) is as shown:

Clearly 𝑔(𝑡) is odd and will contain only sine terms. The Fourier series are in fact

1 2 1 1
𝑓(𝑡) = + (sin 𝑡 + sin 2𝑡 + sin 5𝑡 + ⋯ )
2 𝜋 3 5
And
2 1 1
𝑔(𝑡) = (sin 𝑡 + sin 2𝑡 + sin 5𝑡 + ⋯ )
𝜋 3 5
Exercise:
1- For each of the following functions deduce whether the corresponding Fourier Series contains
(i) sine or cosine terms or both (ii) a constant term

2. Confirm the result obtained for part 7 in the last exercise by finding the Fourier Series fully.
The function involved is
f(t) = |t| , − π < t < π, f(t + 2π) = f(t)
Fourier Series - Half-Range Series
In this Section we address the following problem:
"can we find a Fourier series expansion of a function defined over a finite interval?"

Of course we recognize that such a function could not be periodic (as periodicity demands an infinite
interval).

The answer to this question is yes but we must first convert the given non-periodic function into a
periodic function.

There are many ways of doing this. We shall concentrate on the most useful extension to produce a so-
called half-range Fourier series.

Half Range Fourier Series

So far we have shown how to represent given periodic functions by Fourier Series.

We now consider a slight variation on this theme which will be useful in future when we on solve Partial
Differential Equations.

Suppose that instead of specifying a periodic function we begin with a function 𝑓 (𝑡) defined only over a
limited range of values of t, say 0 < 𝑡 < 𝜋.

Suppose further that we wish to represent this function, over 0 < 𝑡 < 𝜋, by a Fourier Series. (This
situation may seem a little artificial at this point, but this is precisely the situation that will arise in solving
differential equations)
To be specific, suppose we define

𝑓(𝑡) = 𝑡 2 0 < 𝑡 < 𝜋

We shall consider the interval 0 < 𝑡 < 𝜋 to be half a period of a 2𝜋

periodic function. We must therefore define 𝑓 (𝑡) for −𝜋 < 𝑡 < 0
to complete the specification.

Question: Complete the definition of the above function by defining it over −𝜋 < 𝑡 < 0 such that the
resulting functions will have a Fourier Series
(i) containing only cosine terms (ii) containing only sine terms (iii) containing both cosine and sine terms.

The point is that all 3 periodic functions f1(t); f2(t); f3(t) will give rise to a different Fourier Series but
all will represent the function 𝑓 (𝑡) = 𝑡 2 over 0 < 𝑡 < 𝜋.
Fourier Series obtained by extending functions in this sort of way are often referred to as half-range
series.
Normally, in applications, we require either a Fourier Cosine Series (so we would complete a definition as
in (i) above to obtain an even, periodic function) or a Fourier Sine Series (for which, as in (ii) above, we
need an odd periodic function.)
The above considerations apply equally well for a function defined over an interval other than 0 < 𝑡 < 𝜋.

Example:
Obtain a half range Fourier Sine Series to represent the function

𝑓(𝑡) = 𝑡 2 , 0 < 𝑡 < 3.

Solution:
We first extend f (t) as an odd periodic function F(t) of period 6:
𝑓 (𝑡) = −𝑡 2 , −3 < 𝑡 < 0

We now evaluate the Fourier Series of F(t) by standard techniques but take advantage of the symmetry
and put 𝑎𝑛 = 0, n = 0, 1, 2,… .

Using the results for the Fourier Sine coefficients for period 𝑇 derived previously
𝑇
2 2 2𝑛𝜋𝑡
𝑏𝑛 = ∫ 𝐹(𝑡) sin( ) 𝑑𝑡
𝑇 −𝑇 𝑇
2

We put 𝑇 = 6 and, since the integrand is even (a product of 2 odd functions), we can write

2 3 2𝑛𝜋𝑡 2 3 2𝑛𝜋𝑡
𝑏𝑛 = ∫ 𝐹(𝑡) sin( ) 𝑑𝑡 = ∫ 𝑡 2 sin( ) 𝑑𝑡
3 0 6 3 0 6

(Note that we always carry out integration over the originally defined range of the function, in this case
0 < 𝑡 < 3.) We now have to integrate by parts,

2 27 54
𝑏𝑛 = {− cos 𝑛𝜋 + 3 3 (cos 𝑛𝜋 − 1)}
3 𝑛𝜋 𝑛 𝜋
 Then
−18
𝑛 = 2,4,6, …
𝑏𝑛 = 𝑓(𝑡) = { 𝑛𝜋
18 72
− 3 3 𝑛 = 1,3,5, …
𝑛𝜋 𝑛 𝜋
So the required Fourier Sine Series is
1 4 πt 18 2πt 1 4
F(t) = 18 ( − 3 ) sin ( ) − sin ( ) + 18 ( − ) sin (πt) − ⋯.
π π 3 2π 3 3π 27π3

Exercise:
Obtain a half-range Fourier Cosine Series to represent the function
𝑓(𝑡) = 4 − 𝑡, 0 < 𝑡 < 4

Note that the form of the Fourier Series (a constant of 2 together with odd harmonic cosine terms) could
be predicted if, in the figure, we imagine raising the t-axis by 2 units i.e. writing
𝑭(𝒕) = 𝟐 + 𝑮(𝒕)

Clearly 𝐺(𝑡) possesses half-period symmetry

𝐺(𝑡 + 4) = −𝐺(𝑡)
and hence its Fourier Series must contain only odd harmonics.
Fourier Series – Convergence
Introduction

In this Section we examine, briefly, the convergence characteristics of a Fourier series.

We have seen that a Fourier series can be found for functions which are not necessarily continuous (there
may be jumps in the curve) the only requirement that we have made is that the function be periodic.

We have seen that the more terms we take in the Fourier series the better is the approximation to the given
signal. But an obvious question to ask is what happens at the points of discontinuity?

What does the Fourier series converge to at these points? It must converge to something (finite) since a
Fourier series is a sum of very smooth continuous functions.

In this Section we give the answer to this question.

Convergence of a Fourier Series
We have now shown how to obtain a Fourier Series for periodic functions. We have suggested that we
would expect to be able to approximate such functions by using a 'few' terms of the Fourier Series.

The detailed question of the 'convergence' or otherwise of Fourier Series has not been discussed.
The reason for this is that the great majority of functions likely to be encountered in practice have Fourier
Series that do indeed converge and can therefore be safely used as approximations.

The precise conditions that have to be fulfilled for a Fourier Series to converge are known as Dirichlet
conditions after the French mathematician who investigated the matter.

The three conditions are listed in the following Key Point.

The Dirichlet conditions for the convergence of a Fourier Series of a periodic function f (t) are:

1. f (t) must have only a finite number of finite discontinuities, within 1 period

2. f (t) must have a finite number of maxima and minima over one period

3. the integral
𝑇
2
∫ |𝑓(𝑡)| 𝑑𝑡
𝑇
−
2
must be finite.
It follows, for example, that if f (t) is defined over (−𝜋, 𝜋) as one of the following functions

1
𝑡3, , 3𝑡 + 2
𝑡−4
and if 𝑓(𝑡 + 2𝜋) = 𝑓(𝑡) then f (t) can indeed be represented as a Fourier Series as each function satisfies
the Dirichlet conditions.

On the other hand, if, over (−𝜋, 𝜋), 𝑓(𝑡) is

1 1
, , tan 𝑡
𝑡 𝑡−2
then 𝑓 (𝑡) cannot be expanded in a Fourier Series because each of these functions has an infinite
discontinuity within (−𝜋, 𝜋).

If the Dirichlet conditions are satisfied at a point 𝑡 = 𝑡0 where 𝑓 (𝑡) is continuous then, as we would
expect, the Fourier Series at to viz.

∞
𝑎0 𝑛𝜋𝑡0 𝑛𝜋𝑡0
+ ∑[𝑎𝑛 cos ( ) + 𝑏𝑛 sin ( )]
2 𝑇 𝑇
𝑛=1

converges to the function value 𝑓(𝑡0 )

At a point, say 𝑡 = 𝑡1 , at which 𝑓 (𝑡) has a discontinuity then the series

∞
𝑎0 𝑛𝜋𝑡1 𝑛𝜋𝑡1
+ ∑[𝑎𝑛 cos ( ) + 𝑏𝑛 sin ( )]
2 𝑇 𝑇
𝑛=1

converges to
1
{𝑓(𝑡1− ) + 𝑓(𝑡1+ )}
2
where 𝑓(𝑡1− ) is the limit of 𝑓 (𝑡) as 𝑡 approaches 𝑡1 from the
left and 𝑓(𝑡1+ ) is the limit as t approaches 𝑡1 from the right.

Notice: If Dirichelet conditions are satisfied then at a point of continuity 𝑡 = 𝑡0

∞
𝑎0 𝑛𝜋𝑡0 𝑛𝜋𝑡0
𝑓(𝑡0 ) = + ∑[𝑎𝑛 cos ( ) + 𝑏𝑛 sin ( )]
2 𝑇 𝑇
𝑛=1

whereas at a point of discontinuity 𝑡 = 𝑡1 the Fourier Series converges to the average of the two limiting
values
 ∞
1 𝑎0 𝑛𝜋𝑡1 𝑛𝜋𝑡1
{𝑓(𝑡1− ) + 𝑓(𝑡1+ )} = + ∑[𝑎𝑛 cos ( ) + 𝑏𝑛 sin ( )]
2 2 𝑇 𝑇
𝑛=1

Example: Suppose we consider the square wave

Here 𝑓 (𝑡) has finite discontinuities at −𝜋, 0 and 𝜋. The Fourier Series of 𝑓 (𝑡) is

1 2 1 1
+ (sin 𝑡 + sin 3𝑡 + sin 5𝑡 + ⋯ )
2 𝜋 3 5

At 𝑡 = 𝜋/2, or example, where the function f (t) is continuous the square wave converges to 𝑓(𝜋/2) = 1.

On the other hand at 𝑡 = 𝜋 Series clearly has the value 1/2 (since all the sine terms are zero here).
1 1
This value 1/2 agrees with the average of the 2 limiting values of 𝑓 (𝑡) at 𝑡 = 𝜋/2 viz. (1 + 0) = .
2 2
If we actually put 𝑡 = 𝜋/2 in the Fourier Series we obtain

1 2 𝜋 1 3𝜋 1 5𝜋 1 2 1 1 1
+ (sin ( ) + sin( ) + sin( ) + ⋯ ) = + (1 − + − + ⋯ )
2 𝜋 2 3 2 5 2 2 𝜋 3 5 7

Since the series converges, as we have seen, to (𝜋/2) = 1 , we obtain the useful result

1 2 1 1 1
+ (1 − + − + ⋯ ) = 1
2 𝜋 3 5 7

1 1 1 𝜋
1− + − +⋯=
3 5 7 4
 Question: The function

0 −𝜋 < 𝑡 < 0
𝑓(𝑡) = { 2
𝑡 0<𝑡<𝜋

has Fourier Series

π2 cos 2t cos 3t 4 π π 4 π
− 2 (cos t − + − ⋯ ) + {(π − ) sin t − sin 2t + ( − ) sin 3t − sin 4t
6 4 9 π 2 3 9π 4
+ ⋯}
By using a suitable value of t show that
1 1 1 π2
1+ + + +⋯=
4 9 16 6
Note that if we do use 𝑡 = 0 in the Fourier Series (which converges to 𝑓 (0) = 0) we obtain another
infinite series but with alternating signs:

π2 1 1
− 2 (1 − + − ⋯ ) = 0
6 4 9
Or
1 1 π2
1− + −⋯=
4 9 12
 Fourier Series - The Complex Form
Introduction

In this Section we show how a Fourier series can be expressed more concisely if we introduce the
complex number
𝑖 = √−1

By utilizing the Euler formula:

𝑒 𝑖𝜃 = cos 𝜃 + 𝑖 sin 𝜃

we can replace the trigonometric functions by complex exponential functions.

By also combining the Fourier coefficients 𝑎𝑛 and 𝑏𝑛 into a complex coefficient 𝑐𝑛 through

1
𝑐𝑛 = (𝑎𝑛 − 𝑖𝑏𝑛 )
2

we find that, for a given periodic signal, both sets of constants can be found in one operation.

We also obtain Parseval's Theorem which has important applications in electrical engineering.

The complex formulation of a Fourier series is an important pre-cursor of the Fourier transform which
attempts to Fourier analyse non-periodic functions.

Complex Exponential Form of a Fourier Series

So far we have discussed the trigonometric form of a Fourier Series i.e. we have represented functions of
period 𝑇 in the terms of sinusoids, and possibly a constant term, using

∞ ∞
𝑎0 2𝑛𝜋𝑡 2𝑛𝜋𝑡
𝑓(𝑡) = + ∑ 𝑎𝑛 cos + ∑ 𝑏𝑛 sin
2 𝑇 𝑇
𝑛=1 𝑛=1

If we use the angular frequency

2𝜋
𝜔0 =
𝑇
We obtain the more concise form
∞ ∞
𝑎0
𝑓(𝑡) = + ∑ 𝑎𝑛 cos 𝑛𝜔0 𝑡 + ∑ 𝑏𝑛 sin 𝑛𝜔0 𝑡
2
𝑛=1 𝑛=1
We have seen that the Fourier coefficients are calculated using the following integrals:
𝑇
2 2
𝑎𝑛 = ∫ 𝑓(𝑡) cos 𝑛𝜔0 𝑡 𝑑𝑡 , 𝑛 = 0,1,2, … (1)
𝑇 −𝑇
2
𝑇
2 2
𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝜔0 𝑡 𝑑𝑡 , 𝑛 = 1,2,3, . .. (2)
𝑇 −𝑇
2

An alternative, more concise form, of a Fourier Series is available using complex quantities.

This form is quite widely used by Engineers, for example in Circuit Theory and Control Theory, and, in
this course, will lead naturally into a later Section on the Fourier Transform.

Revision of the Exponential Form of a Complex Number

Recall that a complex number in Cartesian form

z = a + ib
where a and b are real numbers and i2 = −1, can be written in polar form

z = r(cos θ + i sin θ)

Where r = |z| = √a2 + b 2 and θ, the argument or phase of z, is such that

a = r cos θ, b = r sin θ
A more concise version of the polar form of z can be obtained by defining a complex exponential quantity
𝑒 𝑖𝜃 by
𝑒 𝑖𝜃 = cos 𝜃 + 𝑖 sin 𝜃

(This is sometimes known as Euler's relation.) The polar angle 𝜃 is normally expressed in radians.
Replacing 𝑖 by −𝑖 we obtain

𝑒 −𝑖𝜃 = cos 𝜃 − 𝑖 sin 𝜃

Example: Write down in cosine, sine form

𝑖𝜋 𝑖𝜋
𝑒 6 , 𝑒− 6
Then write each in Cartesian form.
Solution:
𝑖𝜋 𝜋 𝜋 √3 1
𝑒 6 = cos + 𝑖 sin = +𝑖
6 6 2 2
−𝑖𝜋 𝜋 𝜋 √3 1
𝑒 6 = cos − 𝑖 sin = −𝑖
6 6 2 2
 1 1
Notice: cos 𝜃 = 2 (𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃 ) sin 𝜃 = 2 (𝑒 𝑖𝜃 − 𝑒 −𝑖𝜃 )

Using these results we can redraft an expression of the form

𝑎𝑛 cos 𝑛𝜃 + 𝑏𝑛 sin 𝑛𝜃
in terms of complex exponentials.
(This expression, with 𝜃 = 𝜔0 𝑡 , is of course the nth harmonic of a trigonometric Fourier Series.)

Clearly, we can now rewrite the trigonometric Fourier Series

∞
𝑎0
+ ∑( 𝑎𝑛 cos 𝑛𝜔0 𝑡 + 𝑏𝑛 sin 𝑛𝜔0 𝑡)
2
𝑛=1
As
∞
𝑎0
+ ∑( 𝑐𝑛 𝑒 𝑖𝑛𝜔0 𝑡 + 𝑐 ∗𝑛 𝑒 −𝑖𝑛𝜔0 𝑡 ) (3)
2
𝑛=1

A neater, and particularly concise, form of this expression can be obtained as follows:
𝑎0
Firstly write = 𝑐0 (which is consistent with the general definition of 𝑐𝑛 since 𝑏0 = 0)
2

The second term in the summation

∞

∑ 𝑐 ∗𝑛 𝑒 −𝑖𝑛𝜔0 𝑡 = 𝑐 ∗1 𝑒 −𝑖𝜔0 𝑡 + 𝑐 ∗2 𝑒 −𝑖2𝜔0 𝑡 + ⋯

𝑛=1

1
can be written, if we define 𝑐−𝑛 = 𝑐 ∗ 𝑛 = (𝑎𝑛 + 𝑖𝑏𝑛 ), as
2
−∞
−𝑖𝜔0 𝑡 −2𝑖𝜔0 𝑡 −3𝑖𝜔0 𝑡
𝑐−1 𝑒 + 𝑐−2 𝑒 + 𝑐−3 𝑒 + ⋯ = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝜔0 𝑡
𝑛=−1
Hence can be written
∞ −∞
𝑖𝑛𝜔0 𝑡
𝑐0 + ∑ 𝑐𝑛 𝑒 + ∑ 𝑐𝑛 𝑒 𝑖𝑛𝜔0 𝑡
𝑛=1 𝑛=−1

or, finally, in the extremely concise form

−∞

∑ 𝑐𝑛 𝑒 𝑖𝑛𝜔0 𝑡
𝑛=∞
The complex Fourier coefficients 𝑐𝑛 can be readily obtained as follows using (1) and (2) for 𝑎𝑛 , 𝑏𝑛 .
Firstly

𝑇
𝑎0 1 2
𝑐0 = = ∫ 𝑓(𝑡) 𝑑𝑡 (4)
2 𝑇 −𝑇
2

For 𝑛 = 1,2,3, .. we have

𝑇
1 1 2
𝑐𝑛 = (𝑎𝑛 − 𝑖𝑏𝑛 ) = ∫ 𝑓(𝑡) (cos 𝑛𝜔0 𝑡 − 𝑖 sin 𝑛𝜔0 𝑡)𝑑𝑡
2 𝑇 −𝑇
2

i.e.
𝑇
1 2
𝑐𝑛 = ∫ 𝑓(𝑡) 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡 (5)
𝑇 −𝑇
2

For 𝑛 = 1,2,3, .. we have

𝑇
1 1 2
𝑐−𝑛 = 𝑐 ∗𝑛 = (𝑎𝑛 + 𝑖𝑏𝑛 ) = ∫ 𝑓(𝑡) 𝑒 𝑖𝑛𝜔0 𝑡 𝑑𝑡
2 𝑇 −𝑇
2

This last expression is equivalent to stating that for 𝑛 = −1, −2, −3, ..

𝑇
1 2
𝑐𝑛 = ∫ 𝑓(𝑡) 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡 (6)
𝑇 −𝑇
2

The 3 equations (4), (5), (6) can thus all be contained in the one expression

𝑇
1
𝑐𝑛 = ∫2𝑇 𝑓(𝑡) 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡 , 𝑓𝑜𝑟 𝑛 = 0, ±1, ±2, ±3, ….n
𝑇 −
2

Then
Notice: A function 𝑓 (𝑡) of period 𝑇 has a complex Fourier series
−∞ 𝑇
1 2
𝑓(𝑡) = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝜔0 𝑡 , 𝑤ℎ𝑒𝑟𝑒 𝑐𝑛 = ∫ 𝑓(𝑡) 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡
𝑇 −𝑇
𝑛=∞ 2
For the special case 𝑇 = 2𝜋 so that 𝜔0 = 1 these formulae become particularly simple
−∞
1 𝜋
𝑓(𝑡) = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡 , 𝑤ℎ𝑒𝑟𝑒 𝑐𝑛 = ∫ 𝑓(𝑡) 𝑒 −𝑖𝑛𝑡 𝑑𝑡
2𝜋 −𝜋
𝑛=∞

Note the extremely concise form of these results.

Properties of the Complex Fourier Coefficients

Using properties of the trigonometric Fourier coefficients 𝑎𝑛 , 𝑏𝑛 we can readily deduce the following
results for the 𝑐𝑛 s:
𝑎0
 𝑐0 = 2
is always real.

 Suppose the periodic function f (t) is even so that all 𝑏𝑛 are zero.

Then since, in the complex form, the 𝑏𝑛 arise as the imaginary part of 𝑐𝑛 , it follows that for 𝑓 (𝑡) even the
coefficients 𝑐𝑛 (𝑛 = ±1, ±2, ±3, … ) are wholly real.

Question: If f (t) is odd, what can you deduce about the Fourier coefficients 𝑐𝑛 ?

Further Important Points About Complex Fourier Coefficients

1. Since
𝑇
1 2
𝑐𝑛 = ∫ 𝑓(𝑡) 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡
𝑇 −𝑇
2
then if 𝑓 (𝑡) is even, so 𝑐𝑛 will be real, we have two possible methods for evaluating 𝑐𝑛 :
𝑇 𝑇
(a) evaluate the integral above as it stands i.e. over the full range [− , ]
2 2

note carefully that the second term in the integrand is neither an even nor an odd function so the integrand
itself is

(even function)*(neither even nor odd function) = neither even nor odd function
Thus we cannot write
𝑇
2 2
𝑐𝑛 = ∫ 𝑓(𝑡) 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡
𝑇 0

(b) put 𝑒 −𝑖𝑛𝜔0 𝑡 = cos 𝑛𝜔0 𝑡 − 𝑖 sin 𝑛𝜔0 𝑡 so

𝑓(𝑡)𝑒 −𝑖𝑛𝜔0 𝑡 = 𝑓(𝑡) cos 𝑛𝜔0 𝑡 − 𝑖𝑓(𝑡) sin 𝑛𝜔0 𝑡 = (𝑒𝑣𝑒𝑛)(𝑒𝑣𝑒𝑛) − 𝑖(𝑒𝑣𝑒𝑛)(𝑜𝑑𝑑) = (𝑒𝑣𝑒𝑛) − 𝑖(𝑜𝑑𝑑)

Hence
 𝑇
2 2 𝑎𝑛
𝑐𝑛 = ∫ 𝑓(𝑡) cos 𝑛𝜔0 𝑡 𝑑𝑡 =
𝑇 0 2

𝑇
2. if 𝑓 (𝑡 + 2) = −𝑓(𝑡) then of course only odd harmonic coefficients 𝑐𝑛 (𝑛 = ±1, ±2, ±3, … . )
will arise in the complex Fourier Series just as with trigonometric series.

Example: Find the complex Fourier Series of the sawtooth wave shown:

We have

𝐴𝑡
𝑓(𝑡) = 0<𝑡<𝑇
𝑇
𝑓(𝑡 + 𝑇) = 𝑓(𝑡)

The period is 𝑇 in this case so

𝜔0 = 2𝜋/𝑇

Looking at the graph of f (t) we can say immediately

i. the Fourier Series will contain a constant term 𝑐0

ii. if we imagine shifting the horizontal axis up to 𝐴/2 the signal can be written

𝐴
𝑓(𝑡) = + 𝑔(𝑡)
2
where 𝑔(𝑡) is an odd function with complex Fourier coefficients that are purely imaginary.
Hence we expect the required complex Fourier series of f (t) to contain a constant term 𝐴/2 and complex
exponential terms with purely imaginary coefficients.

We have, from the general theory, and using 0 < 𝑡 < 𝑇 as the basic period for integrating.
 1 𝑇 𝐴𝑡 −𝑖𝑛𝜔 𝑡 𝐴 𝑇 −𝑖𝑛𝜔 𝑡
𝑐𝑛 = ∫ 𝑒 0 𝑑𝑡 = ∫ 𝑡𝑒 0 𝑑𝑡
𝑇 0 𝑇 𝑇2 0

We can evaluate the integral using parts:

𝑇
derivative Integrate 1 1
∫ 𝑡 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡 = {𝑡 𝑒 −𝑖𝑛𝜔0 𝑡 − 1 ( 𝑒 −𝑖𝑛𝜔0 𝑡 )}𝑇0
0 −𝑖𝑛𝜔0 (−𝑖𝑛𝜔0 )2
t 𝑒 −𝑛𝑖𝜔0 𝑡 𝑇 1
1 1 = 𝑒 −𝑖𝑛𝜔0 𝑇 − ( 𝑒 −𝑖𝑛𝜔0 𝑇 )
𝑒 −𝑖𝑛𝜔0 𝑡 −𝑖𝑛𝜔0 (−𝑖𝑛𝜔0 )2
−𝑖𝑛𝜔0
1
0 1 +
𝑒 −𝑖𝑛𝜔0 𝑡 (−𝑖𝑛𝜔0 )2
(−𝑖𝑛𝜔0 )2

But 𝜔0 = 2𝜋/𝑇 then

𝑒 −𝑛𝑖𝜔0 𝑡 = 𝑒 −𝑛𝑖2𝜋𝑡 = cos 2𝑛𝜋 − 𝑖𝑠𝑖𝑛 2𝑛𝜋 = 1

Therefore
𝑇
𝑇 1 𝑇
∫ 𝑡 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡 = − 2 (𝑒 −𝑛𝑖𝜔0 𝑇 − 1) =
0 −𝑖𝑛𝜔0 (−𝑖𝑛𝜔0 ) −𝑖𝑛𝜔0

Hence
𝐴 𝑇 𝑖𝐴
𝑐𝑛 = 2
( )= 𝑛 = ±1, ±2, ±3, ….
𝑇 −𝑖𝑛𝜔0 2𝜋𝑛
And
𝑖𝐴 −𝑖𝐴
𝑐−𝑛 = = = 𝑐 ∗𝑛
2𝜋(−𝑛) 2𝜋𝑛
Also

1 𝑇 𝐴𝑡 𝐴
𝑐0 = ∫ 𝑑𝑡 =
𝑇 0 𝑇 2

as expected.

Hence the required complex Fourier Series is

𝑛=∞
𝐴 𝑖𝐴 𝑒 𝑖𝑛𝜔0 𝑡 𝐴 𝑖𝐴 𝑒 𝑖(−2)𝜔0 𝑡 𝑒 𝑖(−1)𝜔0 𝑡 𝑒 𝑖𝜔0 𝑡 𝑒 𝑖2𝜔0 𝑡
𝑓(𝑡) = + ∑ = + {… + + + + + ⋯}
2 2𝜋 𝑛 2 2𝜋 −2 −1 1 2
𝑛=−∞,𝑛≠0
𝐴 𝑒 −2𝑖𝜔0 𝑡 𝑒 −𝑖𝜔0 𝑡 𝑒 𝑖𝜔0 𝑡 𝑒 𝑖2𝜔0 𝑡
= {… − 𝑖 −𝑖 +𝜋+𝑖 +𝑖 + ⋯}
2𝜋 2 1 1 2

We know that
 −𝑖𝑒 −𝑖𝜔0 𝑡 + 𝑖𝑒 𝑖𝜔0 𝑡 = −𝑖(cos 𝜔0 𝑡 − 𝑖 sin 𝜔0 𝑡) + 𝑖(cos 𝜔0 𝑡 + 𝑖 sin 𝜔0 𝑡) = −2 sin 𝜔0 𝑡

So
𝑛=∞
𝐴 𝐴 sin 𝑛𝜔0 𝑡
𝑓(𝑡) = − ∑
2 𝜋 𝑛
𝑛=1

Question: Find the complex Fourier Series of the 2𝜋 -periodic function

𝑓(𝑡) = 𝑒 𝑡 − 𝜋 < 𝑡 < 𝜋 , 𝑓(𝑡 + 2𝜋) = 𝑓(𝑡)

Parseval's Theorem
This is essentially a mathematical theorem but has, as we shall see, an important engineering
interpretation particularly in electrical engineering.

Parseval's Theorem states that if 𝑓 (𝑡) is a periodic function with period 𝑇 and if 𝑐𝑛 (𝑛 = ±1, ±2, ±3, … )
denote the complex Fourier coefficients of 𝑓 (𝑡) then
𝑇 𝑛=∞
1 2 2
∫ 𝑓 (𝑡) 𝑑𝑡 = ∑ |𝑐𝑛 |2
𝑇 −𝑇
2 𝑛=−∞

In words the theorem states that the mean square value of the signal 𝑓 (𝑡) over one period equals the sum
of the squared magnitudes of all the complex Fourier coefficients.

Proof of Parseval's Theorem

Assume 𝑓 (𝑡) has a complex Fourier Series of the usual form:
𝑇
2𝜋 1 2
𝑓(𝑡) = ∑𝑛=∞
𝑛=−∞ 𝑐𝑛 𝑒
𝑖𝑛𝜔0 𝑡
, 𝜔0 = 𝑇
, where 𝑐𝑛 = ∫ 𝑓(𝑡) 𝑒 −𝑖𝑛𝜔0 𝑡 𝑑𝑡
𝑇 −𝑇
2
Then
𝑛=∞ 𝑛=∞
2 (𝑡) 𝑖𝑛𝜔0 𝑡
𝑓 = 𝑓(𝑡)𝑓(𝑡) = 𝑓(𝑡) ∑ 𝑐𝑛 𝑒 = ∑ 𝑐𝑛 𝑓(𝑡)𝑒 𝑖𝑛𝜔0 𝑡
𝑛=−∞ 𝑛=−∞

𝑇 𝑇 𝑛=∞ 𝑛=∞ 𝑇 𝑛=∞

1 2 2 1 2 1 2
∫ 𝑓 (𝑡) 𝑑𝑡 = ∫ ∑ 𝑐𝑛 𝑓(𝑡)𝑒 𝑖𝑛𝜔0 𝑡 𝑑𝑡 = ∑ 𝑐𝑛 ∫ 𝑓(𝑡) 𝑒 𝑖𝑛𝜔0 𝑡 𝑑𝑡 = ∑ 𝑐𝑛 𝑐 ∗𝑛
𝑇 −𝑇 𝑇 −𝑇 𝑇 −
𝑇
2 2 𝑛=−∞ 𝑛=−∞ 2 𝑛=−∞
𝑛=∞

= ∑ |𝑐𝑛 |2
𝑛=−∞

which completes the proof.

Parseval's Theorem can also be written in terms of the Fourier coefficients 𝑎𝑛 , 𝑏𝑛 of the trigonometric
Fourier Series.

Recall that
𝑎0 𝑎𝑛 − 𝑖𝑏𝑛 𝑎𝑛 + 𝑖𝑏𝑛
𝑐0 = , 𝑐𝑛 = , 𝑛 = 1,2,3, … 𝑐𝑛 = , 𝑛 = −1, −2, −3, …
2 2 2

So
|𝑎𝑛 |2 + |𝑏𝑛 |2
|𝑐𝑛 |2 = , 𝑛 = ±1, ±2, ±3, …
4
𝑛=∞ 𝑛=∞ 𝑛=∞
|𝑎0 |2
2
|𝑎𝑛 |2 + |𝑏𝑛 |2 |𝑎0 |2 |𝑎𝑛 |2 + |𝑏𝑛 |2
∑ |𝑐𝑛 | = +2 ∑ = +∑
4 4 4 2
𝑛=−∞ 𝑛=1 𝑛=1

The engineering interpretation of this theorem is as follows:

Suppose 𝑓 (𝑡) denotes an electrical signal (current or voltage): then from elementary circuit theory f 2(t)
is the instantaneous power (in a 1 ohm resistor) so that
𝑇
1 2 2
∫ 𝑓 (𝑡) 𝑑𝑡
𝑇 −𝑇
2

is the energy dissipated in the resistor during one period.

Now a sinusoid wave of the form

𝐴 cos 𝜔𝑡 (𝑜𝑟 𝐴 sin 𝜔𝑡)

has a mean square value 𝐴2 /2 so a purely sinusoidal signal would dissipate a power 𝐴2 /2 in a 1 ohm
resistor.

Hence Parseval's Theorem in the form (*) states that the average power dissipated over 1 period equals
the sum of the powers of the constant (or d.c.) components and of all the sinusoidal (or alternating)
components.

Question: The triangular signal shown

has trigonometric Fourier Series (Even and Odd Functions)

𝑛=∞
𝜋 4 cos 𝑛𝑡
𝑓(𝑡) = − ∑
2 𝜋 𝑛2
𝑛=1,(𝑛 𝑜𝑑𝑑)

Use Parseval's Theorem to show that

𝑛=∞
1 𝜋4
∑ =
𝑛4 96
𝑛=1,(𝑛 𝑜𝑑𝑑)

Definition: Let f be a function with period 2𝜋 that is Riemann-integrable over a period. The general
formula for the Fourier series:
∞ ∞
𝑎0
𝑓(𝑥) = + ∑ 𝑎𝑛 cos(𝑛𝑥) + ∑ 𝑏𝑛 sin(𝑛𝑥)
2
𝑛=1 𝑛=1

o Where the coefficients are given by,

 1
• 𝑎0 = 𝜋 ∫𝑇 𝑓(𝑥)𝑑𝑥

1
• 𝑎𝑛 = 𝜋 ∫𝑇 𝑓(𝑥) cos(𝑛𝑥) 𝑑𝑥

1
• 𝑏𝑛 = 𝜋 ∫𝑇 𝑓(𝑥)sin(𝑛𝑥)𝑑𝑥

Series of complex numbers

The general formula for the Fourier series:
∞

𝑓(𝑥) = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑥
𝑛=1

o Where the coefficient is given by,

1
𝑐𝑛 = ∫ 𝑓(𝑥)𝑒 −𝑖𝑛𝑥 𝑑𝑥
2𝜋 𝑇

Example: let 𝑓 be a 2𝜋-periodic function determined by the 𝑓(𝑥) = 1 + 𝑥, 𝑥 ∈ (−𝜋, 𝜋) find the
Fourier series:
𝜋
1 1 𝜋 1 𝑥2
• 𝑎0 = ∫𝑇 𝑓(𝑥)𝑑𝑥 = ∫−𝜋(1 + 𝑥)𝑑𝑥 = (𝑥 + ) = 2.
𝜋 𝜋 𝜋 2 −𝜋

1 1 𝜋
• 𝑎𝑛 = 𝜋 ∫𝑇 (1 + 𝑥) cos(𝑛𝑥) 𝑑𝑥 = 𝜋 ∫−𝜋(+𝑥) cos(𝑛𝑥) 𝑑𝑥

1 1+𝑥 1 𝜋 1
= 𝜋( 𝑛
sin(𝑛𝑥) + 𝑛 2 cos(𝑛𝑥)) = 𝑛2 𝜋 (𝑐𝑜𝑠𝑛𝜋 − 𝑐𝑜𝑠𝑛𝜋) = 0
−𝜋

1 1 𝜋 1 −(1+𝑥) 𝜋 2
• 𝑏𝑛 = 𝜋 ∫𝑇 𝑓(𝑥) sin(𝑛𝑥) 𝑑𝑥 = 𝜋 ∫−𝜋(1 + 𝑥) sin(𝑛𝑥) 𝑑𝑥 = 𝜋 ( 𝑛
cos(𝑛𝑥)) = − 𝑛 𝑐𝑜𝑠𝑛𝜋
−𝜋

1, 𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛
cos(n𝜋) = {
−1, 𝑛 𝑖𝑠 𝑜𝑑𝑑
∞
2(−1)𝑛+1
𝑓(𝑥) = 1 + ∑ sin(𝑛𝑥)
𝑛
𝑛=1

Example: let 𝑓 be a 2𝜋-periodic function determined by the 𝑓(𝑥) = 𝑒 𝑥 , 𝑥 ∈ (−𝜋, 𝜋) find the Fourier
series:
1 1 𝜋 1 𝜋 1 𝜋
𝑐𝑛 = 2𝜋 ∫𝑇 𝑓(𝑥)𝑒 −𝑖𝑛𝑥 𝑑𝑥 = 2𝜋 ∫−𝜋 𝑒 𝑥 𝑒 −𝑖𝑛𝑥 𝑑𝑥 = 2𝜋 ∫−𝜋 𝑒 (1−𝑖𝑛)𝑥 𝑑𝑥 = [2𝜋(1−𝑖𝑛) 𝑒 (1−𝑖𝑛)𝑥 ] =
−𝜋
1 1
2𝜋(1−𝑖𝑛)
(𝑒 (1−𝑖𝑛)𝜋 − 𝑒 −(1−𝑖𝑛)𝜋 ) = 𝜋(1−𝑖𝑛) 𝑠𝑖𝑛𝑛𝜋.
 ∞
1
𝑓(𝑥) = ∑ 𝑠𝑖𝑛𝑛𝜋 𝑒 𝑖𝑛𝑥
𝜋(1 − 𝑖𝑛)
𝑛=1

Even and Odd function

• If 𝑓(𝑥) is even 𝑓(−𝑥) = 𝑓(𝑥) then 𝑏𝑛 = 0

∞
𝑎0
𝑓(𝑥) = + ∑ 𝑎𝑛 cos(𝑛𝑥)
2
𝑛=1

Example: cos(𝑥) is even function.

• If 𝑓(𝑥) is odd 𝑓(−𝑥) = −𝑓(𝑥) then 𝑎0 = 𝑎𝑛 = 0
∞

𝑓(𝑥) = ∑ 𝑏𝑛 sin(𝑛𝑥)
𝑛=1

Example: sin(𝑥) is odd function.

Example: as we know, 𝑓(𝑥) = 𝑥 4 , (−𝜋, 𝜋) is even function, find the Fourier series:
𝜋
1 2 𝜋 2 −4𝑥 3 24𝑥
• 𝑎𝑛 = 𝜋 ∫𝑇 𝑓(𝑥) cos(𝑛𝑥) 𝑑𝑥 = 𝜋 ∫0 𝑥 4 cos(𝑛𝑥) 𝑑𝑥 = 𝜋 ( 𝑛2
𝑐𝑜𝑠𝑛𝑥 + 𝑛 4 cos 𝑛𝑥) =
0
2 −4𝜋3 24𝜋 48 8𝜋2
𝜋
( 𝑛2 𝑐𝑜𝑠𝑛𝜋 + 𝑛4
cos 𝑛𝜋) = (−1)n (n4 − 𝑛2
)

𝜋4 48 8𝜋2
• 𝑓(𝑥) = 5
+ ∑∞ n
𝑛=1(−1) (n4 − 𝑛2
)cos(𝑛𝑥)

Example: we know that 𝑔(𝑥) = 𝑥, (−𝜋, 𝜋) is odd function, find the Fourier series:

1 2 𝜋 −2 𝑥 𝜋 −2 −2
𝑏𝑛 = ∫ 𝑔(𝑥)sin(𝑛𝑥) 𝑑𝑥 = ∫ xsin(𝑛𝑥)𝑑𝑥 = ( cos 𝑛𝑥) = cosn𝜋 = (−1)n+1
𝜋 𝑇 𝜋 0 𝜋 𝑛 0 𝜋 𝜋𝑛
−2
g(𝑥) = ∑∞
𝑛=1 𝜋𝑛 (−1)
n+1
𝑠𝑖𝑛 𝑛𝑥

LEMMA: Suppose that f is as in the definition of Fourier series then

1
1. The sequence of Fourier coefficients is bounded, and |𝑐𝑛 | ≤ ∫ |𝑓(𝑥)|𝑑𝑥 , ∀𝑛 ∈ 𝑁.
2𝜋 𝑇

2. The Fourier coefficients tend to zero as |𝑛| → 0.

Theorem: Suppose that 𝑓 is piecewise continuous and that all its Fourier coefficients are zero then
𝑓(𝑥) = 0 at all points where f is continuous.

Corollary: If two continuous functions 𝑓 & 𝑔 have the same Fourier coefficients the 𝑓 = 𝑔.
Theorem: If 𝑓 is continuous on T and its Fourier coefficients 𝑐𝑛 are such that ∑ |𝑐𝑛 | is convergent then
the Fourier series is convergent with sum 𝑓(𝑥) for all 𝑛 ∈ 𝑇 & 𝑓(𝑥) = ∑∞𝑛=1 𝑐𝑛 𝑒
𝑖𝑛𝑥

∞ ∞
𝑎0
𝑓(𝑥) = + ∑ 𝑎𝑛 cos(𝑛𝑥) + ∑ 𝑏𝑛 sin(𝑛𝑥).
2
𝑛=1 𝑛=1

Also the convergent is even uniform on T.

Theorem: let 𝑓 be a 2𝜋-periodic function if 𝑓 is piecewise smooth on [−𝜋, 𝜋] then for all points of
discontinuity 𝑥 ∈ [−𝜋, 𝜋]
we have
𝑓(𝑥 − )+𝑓(𝑥 + ) 𝑎0
2
= 2
+ ∑∞
𝑛−1(𝑎𝑛 cos(𝑛𝑥) + 𝑏𝑛 sin(𝑛𝑥))

Where as for points of continuity 𝑥 ∈ [−𝜋 , 𝜋] we have

∞
𝑎0
𝑓(𝑥) = + ∑(𝑎𝑛 cos(𝑛𝑥) + 𝑏𝑛 sin(𝑛𝑥))
2
𝑛=1

Review:
• cos(𝑎 ± 𝑏) = 𝑐𝑜𝑠𝑎𝑐𝑜𝑠𝑏 ∓ 𝑠𝑖𝑛𝑎𝑠𝑖𝑛𝑏
• sin(𝑎 ± 𝑏) = 𝑠𝑖𝑛𝑎𝑐𝑜𝑠𝑏 ∓ 𝑐𝑜𝑠𝑎𝑠𝑖𝑛𝑏
1
• 𝑐𝑜𝑠𝑎𝑐𝑜𝑠𝑏 = [cos(𝑎 + 𝑏) + cos(𝑎 − 𝑏)]
2
1
• 𝑠𝑖𝑛𝑎𝑠𝑖𝑛𝑏 = 2 [cos(𝑎 − 𝑏) − cos(𝑎 + 𝑏)]
1
• 𝑠𝑖𝑛𝑎𝑐𝑜𝑠𝑏 = 2 [sin(𝑎 + 𝑏) + sin(𝑎 − 𝑏)]

• 𝑐𝑜𝑠2𝑎 = cos2 𝑎 − sin2 𝑎 = 2 cos2 𝑎 − 1 = 1 − 2 sin2 𝑎

• 𝑠𝑖𝑛2𝑎 = 2𝑠𝑖𝑛𝑎𝑐𝑜𝑠𝑎
1
• cos2 𝑎 = (1 + 𝑐𝑜𝑠2𝑎)
2
1
• sin2 𝑎 = (1 − 𝑐𝑜𝑠2𝑎)
2
 3𝜋 𝜋
Example: 𝑓(𝑥) = 𝑠𝑖𝑛𝑥, [− , ]
2 2
𝜋
1 1 1
• 𝑎0 = 𝜋 ∫𝑇 𝑓(𝑡)𝑑𝑡 = 𝜋 ∫ 2𝜋 𝑠𝑖𝑛𝑥 𝑑𝑥 = 𝜋 [−𝑐𝑜𝑠𝑥] = 0
−
2

𝜋
1 1
• 𝑎𝑛 = 𝜋 ∫𝑇 𝑓(𝑡) cos(𝑛𝑡) 𝑑𝑡 = 𝜋 ∫ 2𝜋 sin(𝑥)𝑐𝑜𝑠(𝑛𝑥) 𝑑𝑥 = 0
−
2

𝜋 𝜋
1 1 1 1
• 𝑏𝑛 = ∫𝑇 𝑓(𝑡)sin(𝑛𝑡)𝑑𝑡 = ∫ 23𝜋 𝑠𝑖𝑛𝑥𝑠𝑖𝑛(𝑛𝑥)𝑑𝑥 = ∫ 23𝜋 [cos(1 − 𝑛) − cos(1 + 𝑛)]𝑑𝑥
𝜋 𝜋 − 𝜋 − 2
2 2

𝑓(𝑥) = ∑ 𝑏𝑛 sin(𝑛𝑥)
𝑛=1

DIRICHLET kernel
The function 𝐷𝑛
𝑁
1 1
𝐷𝑛 (𝑥) = ( + ∑ 𝑐𝑜𝑠𝑛𝑥 )
𝜋 2
𝑛=1

is called DIRICHLET kernel.

Also we have
𝑁
1
𝐷𝑁 (𝑥) = ∑ 𝑒 𝑖𝑛𝑥
2𝜋
𝑛=−𝑁

and
1
sin (𝑁 + ) 𝑥
𝐷𝑁 (𝑥) = 2
1
2𝜋sin(2)𝑥
1 1 1 1
Example: 𝐷5 (𝑥) = 𝜋 (2 + ∑5𝑛=1 𝑐𝑜𝑠𝑛𝑥 ) = 𝜋 (2 + 𝑐𝑜𝑠𝑥 + 𝑐𝑜𝑠2𝑥 + 𝑐𝑜𝑠3𝑥 + 𝑐𝑜𝑠4𝑥 + 𝑐𝑜𝑠5𝑥)

11𝑥
sin 2
𝐷5 (𝑥) = 𝑥
2𝜋 sin 2
FEJER kernel
The function
𝑁
1
𝐹𝑁 (𝑥) = ∑ 𝐷𝑁 (𝑥)
𝑁+1
𝑛=0

1
1 sin (𝑁 + 2) 𝑥
𝐹𝑁 (𝑥) = ( )
2𝜋(𝑁 + 1) 1
sin ( ) 𝑥
2
Is said to be FEJER kernel

Theorem: suppose that 𝑓 is 2𝜋-periodic, continuous and piecewise smooth. Let 𝑎𝑛 , 𝑏𝑛 and 𝑐𝑛 be the
Fourier coefficients of f defined before
𝑎0
o 𝑓(𝑥) = 2
+ ∑∞ ∞
𝑛=1 𝑎𝑛 cos(𝑛𝑥) + ∑𝑛=1 𝑏𝑛 sin(𝑛𝑥)

o 𝑓(𝑥) = ∑∞
𝑛=1 𝑐𝑛 𝑒
𝑖𝑛𝑥

1
• 𝑐𝑛 = 2𝜋 ∫𝑇 𝑓(𝑥)𝑒 −𝑖𝑛𝑥 𝑑𝑥

1
• 𝑎0 = ∫𝑇 𝑓(𝑥)𝑑𝑥
𝜋

1
• 𝑎𝑛 = 𝜋 ∫𝑇 𝑓(𝑥) cos(𝑛𝑥) 𝑑𝑥

1
• 𝑏𝑛 = 𝜋 ∫𝑇 𝑓(𝑥)sin(𝑛𝑥)𝑑𝑥

And let 𝑎𝑛′ , 𝑏𝑛′ , 𝑎0′ and 𝑐′𝑛 are the corresponding the Fourier coefficients of 𝑓 ′ , then

• 𝑎𝑛′ = 𝑛𝑏𝑛
• 𝑏𝑛′ = 𝑛𝑎𝑛
 • 𝑐𝑛′ = 𝑖𝑛𝑐𝑛

• 𝑎0′ = (𝑎𝑛 )′ = 0
o 𝑓 ′ (𝑥) = ∑∞ ∞
𝑛−1 𝑛𝑏𝑛 cos(𝑛𝑥) + ∑𝑛=1 𝑛𝑎𝑛 sin(𝑛𝑥)

o 𝑓 ′ (𝑥) = ∑∞
𝑛=−∞ 𝑖𝑛𝑐𝑛 𝑒
𝑖𝑛𝑥

Note: To find the Fourier series for 𝑓 ′ it is enough to take derivative tend by term.

Example: Find the Fourier series of 𝑓(𝑥) = 2𝑥, and ℎ(𝑥) = 𝑥, [−𝜋, 𝜋] when we know the Fourier
series of g(𝑥) = 𝑥 2 𝑖𝑛 [−𝜋, 𝜋]
Sol:
𝜋2 (−1)𝑛
f(𝑥) = 𝑥 2 = 3
+ 4 ∑∞
𝑛=1 𝑛2
𝑐𝑜𝑠𝑛𝑥

(𝑥 2 )′ = 2𝑥
−(−1)𝑛 𝑛(−1)𝑛+1
𝑓(𝑥) = 2𝑥 = (𝑥 2 )′ = 0 + 4 ∑∞
𝑛=1 𝑛2
𝑛𝑠𝑖𝑛𝑛𝑥 = 4 ∑∞
𝑛=1 𝑛2
𝑠𝑖𝑛𝑛𝑥

(−1)𝑛+1
= 4 ∑∞
𝑛=1 𝑛
𝑠𝑖𝑛𝑛𝑥

(−1)𝑛+1
𝑓(𝑥) = 2𝑥 = 4 ∑∞
𝑛=1 𝑛
𝑠𝑖𝑛𝑛𝑥

𝑓(𝑥) (−1)𝑛+1
ℎ(𝑥) = 𝑥 = 2
= 2 ∑∞
𝑛=1 𝑛
𝑠𝑖𝑛𝑛𝑥

Formula of the other periodic

𝑓(𝑥) = 𝑓(𝑥 + 𝑇); 𝑇 = 2𝐿

𝑓(𝑥) = 𝑓(𝑥 + 𝑇) = 𝑓(𝑥 + 2𝐿) [−𝐿, 𝐿]𝑜𝑟 [0, 2𝐿]𝑜𝑟 [𝑎, 𝑎 + 2𝐿]

If 𝑓(𝑥) is a 𝑇 = 2𝐿-periodic function, then

∞ ∞
𝑎0 𝑛𝜋 𝑛𝜋
𝑓(𝑥) = + ∑ 𝑎𝑛 cos + ∑ 𝑏𝑛 sin
2 𝐿 𝐿
𝑛=1 𝑛=1
1 𝐿
• 𝑎0 = 𝐿 ∫−𝐿 𝑓(𝑥)𝑑𝑥
1 𝐿 𝑛𝜋
• 𝑎𝑛 = ∫−𝐿 𝑓(𝑥) cos 𝑥 𝑑𝑥
𝐿 𝐿
1 𝐿 𝑛𝜋
• 𝑏𝑛 = 𝐿 ∫−𝐿 𝑓(𝑥)sin 𝐿
𝑥 𝑑𝑥
𝑛𝜋
𝑖 𝑥
• 𝑓(𝑥) = ∑∞
𝑛=−∞ 𝑐𝑛 𝑒 𝐿
 𝑛𝜋
1 𝜋
• 𝑐𝑛 = 2𝜋 ∫−𝜋 𝑓(𝑥)𝑒 −𝑖 𝐿 𝑥 𝑑𝑥
𝑛𝜋
1 𝜋
• 𝑐𝑛 = 2𝜋 ∫−𝜋 𝑓(𝑥)𝑒 −𝑖 𝐿 𝑥 𝑑𝑥

• If 𝑓(𝑥) is even

2 𝑙 𝑛𝜋
𝑎𝑛 = ∫0 𝑓(𝑥) 𝑐𝑜𝑛 = 𝑥 𝑑𝑥
{ 𝑙 𝑙
𝑏𝑛 = 0

• If 𝑓(𝑥) is odd

𝑎𝑛 = 𝑜
• {𝑏 = 2 𝑙 𝑛𝜋
𝑛 𝑙
∫0 𝑓(𝑥) sin ( 𝑙 ) 𝑥 𝑑𝑥

0, 𝑥<0
Example: 𝑔(𝑥) = { 𝜋𝑥
𝑠𝑖𝑛 2 , 0≤𝑥<2

1 𝑙 1 2 𝜋𝑥 1 −2 𝜋𝑥 2 −1 2
• 𝑎0 = ∫ 𝑔(𝑥) 𝑑𝑥 = ∫ 𝑠𝑖𝑛 2 𝑑𝑥 = [ 2 ( 𝜋 ) cos ( 2 )] = (−1 − 1) = 𝜋
𝑙 −𝑙 2 0 0 𝜋

1 2 𝑛𝜋 𝑛𝜋 1 2
• 𝑎𝑛 = ∫0 (𝑠𝑖𝑛 ) (𝑐𝑜𝑠 𝑥) 𝑑𝑥 = ∫0 sin(𝜋𝑥) 𝑑𝑥
2 2 2 4

1 2 (𝑛+1) 𝜋(1−𝑛)
= ∫ [𝑠𝑖𝑛𝜋 2 𝑥 − sin 𝑥] 𝑑𝑥
4 0 2

1 −2 𝜋(𝑛+1) 2 (1−𝑛)𝜋 2
= 4 (𝜋(𝑛+1) cos 2
𝑥 + 𝜋(1−𝑛)
𝑐𝑜𝑠 2
𝑥]
0
1 −2 2 2 2
= ( cos(𝑛 + 1) 𝜋 + cos(1 − 𝑛) 𝜋 + −
4 𝜋(𝑛+1) 𝜋(1−𝑛) 𝜋(𝑛+1) 𝜋(1−𝑛)

2 2 2 2
(𝜋(𝑛+1) − 𝜋(1−𝑛) + 𝜋(1−𝑛) − 𝜋(1−𝑛)) , 𝑛 = 2𝑘 1 2
(
𝜋 1−𝑛2
), 𝑛 = 2𝑘
• ={
0, 𝑛 = 2𝑘 − 1
{0, 𝑛 = 2𝑘 − 1

• 𝑏𝑛 = 0

1 2 𝜋𝑥 𝜋𝑥 1 2
• 𝑎1 = 2 ∫0 𝑠𝑖𝑛 2
cos 2 𝑑𝑥 = 4 ∫0 𝑠𝑖𝑛𝜋𝑥𝑑𝑥 = 0
 1 2 𝜋𝑥 𝜋𝑥 1
• 𝑏1 = 2 ∫0 𝑠𝑖𝑛 2
sin 2 𝑑𝑥 =2
1 1 𝜋 −2 2𝑘𝜋
• 𝑔(𝑥) = 𝜋 + 2 sin ( 2 ) + ∑∞
𝑛=1 𝜋(4𝑘 2 −1) (𝑐𝑜𝑠 𝑙
𝑥)

Half-ring Fourier series

𝑓(𝑥) = 𝑥, [−𝜋, 𝜋]

𝑥, [0,1]
• 𝑓𝑒𝑣𝑒𝑛 = {
−𝑥, [−1,0]

𝑥, [0,1]
• 𝑓𝑜𝑑𝑑 = {
−𝑥, [−1,0]

Odd and even extension

If 𝑓(𝑥) is a function on (0, 𝐿)𝑜𝑟 [0, 𝐿] then

𝑓(𝑥), [0, 𝐿]
• 𝑓(𝑥) = 𝑓𝑒𝑣𝑒𝑛 = {
𝑓(−𝑥), [−𝐿, 0]

𝑓(𝑥), [0, 𝐿]
• 𝐺(𝑥) = 𝑓𝑜𝑑𝑑 = {
−𝑓(−𝑥), [−𝐿, 0]

Definition: Suppose 𝑓 is integrable function on [0,1]

𝑎0
• The series 𝑓(𝑥) = 2
+ ∑∞
𝑛=1 𝑎𝑛 cos(𝑛𝑥)

1
𝑎0 = ∫ 𝑓(𝑥)𝑑𝑥
𝜋 𝑇

1
𝑎𝑛 = ∫ 𝑓(𝑥) cos(𝑛𝑥) 𝑑𝑥
𝜋 𝑇

is called the Fourier cosine series.

𝑎0
• The series 𝑓(𝑥) = 2
+ ∑∞ ∞
𝑛=1 𝑎𝑛 cos(𝑛𝑥) + ∑𝑛=1 𝑏𝑛 sin(𝑛𝑥)

1
𝑏𝑛 = ∫ 𝑓(𝑥)sin(𝑛𝑥)𝑑𝑥
𝜋 𝑇

is called the Fourier sine series.

Prestel's inequality and Bessel`s inequality

If 𝑓 is 2𝜋-periodic and Riemann inequality on [𝜋, − 𝜋] and the Fourier coefficient 𝑎𝑛 , 𝑎0 , 𝑏𝑛 , 𝑐𝑛 are
defined by before formula then
1 𝜋
∑∞ 2 2
𝑛=−∞|𝑐𝑛 | ≤ 2𝜋 ∫−𝜋|𝑓(𝑥)| 𝑑𝑥
• (𝑏𝑒𝑠𝑠𝑒𝑙`𝑠 𝑖𝑛𝑒𝑞𝑢𝑎𝑙𝑖𝑡𝑦) {
𝑎02 1 𝜋
2
+ ∑∞ 2 2 2
𝑛=1(𝑎𝑛 + 𝑏𝑛 ) ≤ 𝑛 ∫−𝜋|𝑓(𝑥)| 𝑑𝑥

𝑎02 1 𝜋
• (𝑝𝑒𝑟𝑒𝑠𝑟𝑒𝑙 𝑖𝑛𝑒𝑞𝑢𝑎𝑙𝑖𝑡𝑦) 2
+ ∑∞ 2 2 2
𝑛=1(𝑎𝑛 + 𝑏𝑛 ) = 𝑛 ∫−𝜋|𝑓(𝑥)| 𝑑𝑥

Rules and theorems

1) Linearly theorem;
If 𝑓(𝑥) and 𝑔(𝑥) are functions with T-periodic, then any arbitrary linear combination 𝑐1 𝑓 + 𝑐2 𝑔 are
periodic with T and:
∞ ∞
𝑎0 𝑛𝜋 𝑛𝜋
𝑓(𝑥) = + ∑ 𝑎𝑛 cos 𝑥 + ∑ 𝑏𝑛 sin 𝑥
2 𝐿 𝐿
𝑛=1 𝑛=1
∞ ∞
𝐴0 𝑛𝜋 𝑛𝜋
𝑔(𝑥) = + ∑ 𝐴𝑛 cos 𝑥 + ∑ 𝐵𝑛 sin 𝑥
2 𝐿 𝐿
𝑛=1 𝑛=1

𝑎 𝑛𝜋 𝑛𝜋 𝑛𝜋 𝑛𝜋
𝑐1 𝑓 + 𝑐2 𝑔 = 𝑐1 ( 20 + ∑∞
𝑛=1 𝑎𝑛 cos 𝐿
𝑥 + ∑∞
𝑛=1 𝑏𝑛 sin 𝐿
𝑥 ) + c2 (∑∞
𝑛=1 𝐴𝑛 cos 𝐿
𝑥 + ∑∞
𝑛=1 𝐵𝑛 sin 𝐿
𝑥)

2) The first shifting rule:

If 𝑓(𝑥) is periodic function with 𝑇 = 2𝐿 period.

If 𝑐𝑛 be the Fourier coefficients then 𝑓(𝑥 − 𝑎) is a periodic function with 𝑇 = 2𝐿 , and the Fourier
coefficients are equal.
𝑖𝑛𝜋𝑥
𝑐𝑛 , 𝑓(𝑥) → 𝑓(𝑥 − 𝑎), 𝑐𝑛 𝑒 − 𝐿 .

3) Inversion time (time inverse):

If 𝑓(𝑥) is periodic function and 𝑇 = 2𝐿 and 𝑐𝑛 are the Fourier coefficients then the Fourier coefficients of
𝑓(−𝑥) are equal 𝑐𝑛 .
4) The second shifting rule:
𝑖𝑛𝜋𝑥
𝑓(𝑥), 𝑇 = 2𝐿 → 𝑓(𝑥)𝑒 𝐿 , 𝑇 = 2𝐿
𝐶𝑛 → 𝐶𝑛−𝑎
5) Scaling rule:
If 𝑓(𝑥) is periodic function with 𝑇 = 2𝐿 period.
𝑇
If 𝐶𝑛 be the Fourier coefficients then 𝑓(𝑎𝑥) is periodic function with 𝑎
= 2𝐿

𝑓(𝑥) → 𝐶𝑛 , 𝑇 = 2𝐿
𝑇
𝑓(𝑎𝑥) → 𝐶𝑛 , = 2𝐿
𝑎

Application of Fourier series

Find the solution for following equation by Fourier series

𝑦′′ − 2𝑦 = 2𝑠𝑖𝑛𝛼𝑥 , 𝛼 ∉ 𝑍, [−𝜋, 𝜋]

2𝑠𝑖𝑛𝛼𝜋 ∞ (−1)𝑛+1 𝑛
𝑠𝑖𝑛𝛼𝑥 = ∑𝑛=1 2 2 𝑠𝑖𝑛𝑛𝑥
𝜋 𝑛 −𝛼

Suppose that
𝑎0
𝑦= 2
+ ∑∞
𝑛=1(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin 𝑛𝑥)

Then
𝑎0
𝑦′ = 2
+ ∑∞
𝑛=1(−𝑛𝑎𝑛 cos 𝑛𝑥 + 𝑛𝑏𝑛 sin 𝑛𝑥)

𝑎0
𝑦′′ = 2
+ ∑∞ 2 2
𝑛=1(−𝑛 𝑎𝑛 cos 𝑛𝑥 − 𝑛 𝑏𝑛 sin 𝑛𝑥)

∞ ∞
2 2
𝑎0
∑(−𝑛 𝑎𝑛 cos 𝑛𝑥 − 𝑛 𝑏𝑛 sin 𝑛𝑥) = −2 ( + ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin 𝑛𝑥))
2
𝑛=1 𝑛=1
∞
4𝑠𝑖𝑛𝛼𝜋 (−1)𝑛+1 𝑛
= ∑ 2 𝑑𝑖𝑛𝑛𝑥
𝜋 𝑛 − 𝛼2
𝑛=1
∞

= ∑[(−𝑛2 𝑎𝑛 − 2𝑎𝑛 ) cos 𝑛𝑥 + (−𝑛2 𝑏𝑛 − 2bn )sin 𝑛𝑥]−𝑎0

𝑛=1
∞
4𝑠𝑖𝑛𝛼𝜋 (−1)𝑛+1 𝑛
= ∑ 2 𝑠𝑖𝑛𝑛𝑥
𝜋 𝑛 − 𝛼2
𝑛=1

• 𝑎0 = 0

−(𝑛2 + 2)𝑎𝑛 = 0 → (𝑛2 + 2)𝑎𝑛 = 0 → 𝑎𝑛 = 0

 ∞
2
4𝑠𝑖𝑛𝛼𝜋 (−1)𝑛+1 𝑛
−(𝑛 + 2)𝑏𝑛 = ∑ 2
𝜋 𝑛 − 𝛼2
𝑛=1

4𝑠𝑖𝑛𝛼𝜋 ∞ (−1)𝑛 𝑛
• 𝑏𝑛 = ∑𝑛=1 2 2 2
𝜋 (𝑛 −𝛼 )(𝑛 +2)

4𝑠𝑖𝑛𝛼𝜋 ∞ (−1)𝑛 𝑛
• 𝑦= ∑𝑛=1 2 2 2 𝑠𝑖𝑛𝑛𝑥
𝜋 (𝑛 −𝛼 )(𝑛 +2)