EE 370

Chapter II: Introduction to Signals

Lecture 4

Trigonometric Fourier Series

Let g(t) be one of the following: 1. a periodic signal with period T0 , 2. any signal in which we are interested in the time interval of t1 t t1 + T0 , The signal g(t) in the interval t1 t t1 + T0 can be represented by the

trigonometric Fourier series in terms of a sum of the following sinusoids:

{1, cos(0t), cos(20t), cos(30t), , sin(0t), sin(20t), sin(30t), }, or the compact Fourier series sinusoids: {1, cos(0t+1), cos(20t+2), cos(30t+3), }, where the frequency 0 and T0 are related by T0 = 2 / 0 . The basis for this representation is that the different sinusoids shown above are orthogonal. Two signals a(t) and b(t) are orthogonal over a period T of time if (1)

a(t ) b(t )dt = 0 .

T

This is true for all of the sinusoids given above in (1) over a period T0 = 2 / 0. The representation of g(t) is given by g (t ) = a0 + a1 cos(0t ) + a2 cos(20t ) + a3 cos(30t ) + K + b1 sin(0t ) + b2 sin(20t ) + b3 sin(30t ) + K t1 t t1 + T0

which can be written as

g (t ) = a0 + an cos(n0t ) + bn sin(n0t )
n =1

t1 t t1 + T0

or using the compact sinusoids as

g (t ) = C0 + Cn cos(n0t + n )
n =1

t1 t t1 + T0

The coefficients a0 , an , and bn shown above can be evaluated using

1 a0 = T0
t1 + T0

t1

g (t )dt ,

EE 370

Chapter II: Introduction to Signals

Lecture 4

The coefficient a0 represents the average (or the DC value) of the function. So, for functions that have zero DC, the coefficient a0 will be zero. 2 an = T0
t1 + T0

t1

g (t ) cos(n t )dt
0

n = 1, 2, 3,

2 bn = T0

t1 + T0

t1

g (t ) sin(n t )dt
0

n = 1, 2, 3,

A trigonometric identity in the form of an cos(n0t ) + bn sin(n0t ) = Cn cos(n0t + n ) , exits, where the relation between the coefficients an and bn and the coefficients Cn and n is given by
2 2 Cn = an + bn ,

and

n = tan 1

bn an

such that C0 = a0 .
Comments:

1.

2.

3. 4.

All cosine functions cos(n0t ) are even functions (they are symmetric about the yaxis). Since an odd function can never be represented in terms of even functions, an of the Fourier series for an odd function are always zero for all values of n. All sine functions sin(n0t ) are odd functions (they are antisymmetric about the yaxis or symmetric about the origin). Since an even function can never be represented in terms of odd functions, bn of the Fourier series for an even function are always zero for all values of n. For a periodic function that is not even and not odd, at least some of the coefficients an and some of the coefficients bn will be non zero. The frequency 0 is called the fundamental frequency of the periodic signal f(t) and the multiple of this frequency n0 is called the nth harmonic of this fundamental frequency. The fundamental frequency represents the lowest frequency component contained in f(t). Two signals: a sine wave with frequency 0 and a square wave with frequency 0 will sound similar when played using a speaker except that the square wave will also contain higher harmonics.

EE 370 5.

Chapter II: Introduction to Signals Lecture 4 The Fourier series of part of a nonperiodic signals is similar to the Fourier series of a periodic signal that is obtained by repeating that part of nonperiodic signal to the right and to the left.

Exponential Fourier Series

We also can represent the function g(t) in terms of complex exponentials as

g (t ) =

n =

D e
n

jn 0 t

= D0 +

n = ( n 0)

D e
n

jn 0 t

where Dn = 1 T0

T0

g (t )e jn 0 t dt .

Therefore, Dn is related to Cn and n as

1 | Dn |=| D n |= Cn , 2

Dn = Dn = n .

Note that the magnitudes and phases of the coefficients Dn are, respectively, known as the amplitude and phase spectrums of gTo(t), or | Dn | is amplitude spectrum of gTo(t) , and Dn is phase spectrum of gTo(t). Examples: 1. Find the Fourier series coefficients an and bn for a) the aperiodic signal g(t) = | t |, 0.5 t 1.5,

Solution: Although this signal is nonperiodic, we can still find its Fourier series expansion between the two points t = 0.5 to 1.5 as follows. We will consider T0 to be T0 = 1.5 (0.5) = 2 sec

0 = 2/T0 = rad/s

EE 370
a0 = = 1 t dt 2 0.5
1.5

Chapter II: Introduction to Signals

Lecture 4

1 1 t dt + 2 t dt 2 0.5 0

1.5

1 0 1 1.5 = t2 + t2 4 0.5 4 0 1 1 = ( 0 0.25 ) + ( 2.25 0 ) 4 4 = 0.625 1.5 1 an = t cos(n0t )dt 2 0.5

1 1 = t cos(n0t )dt + t cos(n0t )dt 2 0.5 20 Now we will have to integrate using the integration by parts method. That is

1.5

udv = uv
T

vdu
T

So, if we let we will get,

u=t du = dt

and and

dv = cos(n0t)dt, v = (1/n0) sin(n0t)

0 1.5 0 1.5 1 t 1 t 1 1 an = sin(n t ) sin(n t )dt + sin(n t ) sin(n t )dt 2 n 2 n n 0.5 0.5 n 0 0

Now completing the remaining integrations and evaluating the coefficients an becomes straight forward. Computing the coefficients an is performed in exactly the same manner.

b)

the periodic signal f(t) shown below

EE 370 c)

Chapter II: Introduction to Signals the periodic signal h(t) shown below

Lecture 4