A NOTE ON FOURIER SERIES OF HALF WAVE RECTIFIER,

FULL WAVE RECTIFIER AND UNRECTIFIED SINE WAVE

Jambunatha Sethuraman*
Vinayaka Mission’s Kirupananda Variyar Engineering College
Salem Tamil Nadu India

ABSTRACT: There is always an inherent phase difference between a sinusoidal

input and output (response) for a linear passive causal system. This is explained in
detail and even in the Fourier series of a periodic ‘causal’ function, this principle can
be elegantly used with profit and better understanding. Several illustrations are give
in support of this novel idea. There need not be a special section for Fourier
cosine/sine transforms as this approach covers them also.

[email protected]
 INTRODUCTION: f ( x ) is a periodic function with period 2 L. It can be expended
as a Fourier series in the interval – L< x < L:

a0 ∞ nπx nπx
f ( x )=
2 n=1 n (
+ ∑ a cos
L
+b n sin
L
. ) (1)

¿ the function space spanned by the basic set of orthogonal functions

{sin nπxL ,cos nπxL }, n=0,1, 2 , 3 ,… the function receives the co-ordinates {a ,a , b }. 0 n n

nπ n
Each sinusoidal wave has angular frequency ω n= ∧frequency f n= . The
L 2L
1
fundamental frequency f 1= . The set of Fourier coefficients has
2L
( 2 n+1 ) entities with n→ ∞[1].

x Time t Space x
fn Temporal frequency Spatial frequency

L
1 nπx
a n= ∫ f ( x ) cos dx ,
2 L −L L

L
1 nπx
b n= ∫ f ( x ) sin dx ,
2 L −L L

L
1
a 0= ∫ f ( x ) dx , (2)
L −L

Half-wave rectifier output wave form and its Fourier series

f ( t )= Asin ω0 t 0 ≤ t ≤ T .
{0−T ≤t ≤0

For negative duration of the period , the function is has zero value . The coefficients are
evaluated as
 A 2A 1 1 1 A
Asin ω 0 t=
π
− (
π 1.3
cos 2 ω 0 t +
3.5
cos 4 ω0 t+
5.7 2 )
cos 6 ω 0 t +… + sin ω 0 t ,

where n is even . It is convenient to write the coefficients as

a0 A
= , a =¿ ¿ Note that there is only one coefficient which survives in sine series,
2 π n
namely b 1 . all the other coefficients vanish. It is interesting. Why it is so, will be
clear later. See Fig. 1.

Fig .1 :Fourier coefficients of half −wave rectifier . All coeffidients

a 0 , a n ( n> 0 )∧b n ( n> 0 ) are present . A scale factor of 2 is due to doubling the function in
the period.

Full−wave rectifier: The wave form can be represented as

f ( t )= Asin ω 0|t|−T ≤ t ≤T .

The period is 2T . For the full-wave rectifier the Fourier coefficients are given by
 a0 2 A
= , an=¿¿
2 π
See Fig .2 below .

Fig .2 :Fourier coefficients of full−wave rectifier .

Note all bn are zero . Only a0∧an coefficients survive as expected . A scale factor 2
is due to doubling the existence of the function.

Unrectified sine wave: This is a pure sine wave given by

Asin ω 0 t−T ≤ t ≤ T .

The Fourier coefficients are a0 =0 , all an are zeros ,∧only one bn coefficient , namely
b1
b 1 survives∧is equal ¿ A . b 1=2 A , b 2=b3 =… .=0. = A.
2

This is expected since it is a pure sine wave and has only one Fourier component,
viz., sin ω 0 t and hence only b 1 contributes. Again a factor 2 is present due to
‘doubling’ the function in the period. See Fig. 3.
 Fig . 3 : A pure sine waveis an odd extension of half wave rectifier .
It has only b1 coefficient ∧all other bn ( n>1 ) vanish .
Again a factor is present due ¿ thedoubling of the wave∈the period .

The assertion is the the half-wave rectifier contains the coefficients of full-wave
rectifier and (unrectified) pure sine wave. This is interesting. A closer analysis
shows that full-wave rectifier and pure sine wave are respectively even and odd
extensions of half-wave rectifier! If half-wave rectifier is extended as an even
function (full-wave rectifier) only the cosine coefficients survive and sine
coefficients (odd) vanish. A factor 2 arises due to the period is doubled. If the half-
wave rectifier is extended as an odd function, i.e., pure sine wave only the odd sine
coefficients survive and all even coefficients vanish. This is an important concept
and can be applied to all so called Fourier sine/cosine series. In both extensions, a
factor 2 arises due to the function is doubled in the period.

REDUNDANT EXERCISES: It is not necessary to teach Fourier sine /cosine series and
they are redundant in the sense that they are special cases of Fourier series of a
causal periodic function.

Our approach is further strengthened by the following exercise: See. Fig. 4.

 Fig. 4.

( 12 ) coefficient
A rectangle periodic function with ‘duty cycle’ 50% has only a 0 ¿

¿ all a n> 0 vanish . b n coefficients exist .

When extended as an odd function the . b n are simply doubled. When extended as
an even function, it becomes a continuous straight line with constant value 1. Hence
a0
=1∧a0=2. There is consistency in this approach. It is well-known that a constant
2
function has only ‘dc’ term as there is no undulation or change in the function. In
Fourier analysis, it is sometimes regarded as ‘useless’ term having no information;
but it is not so. Its role is important and serves as a ‘canvas for painting’. The next
illustration with a causal triangle function is also self evident and proves beyond
doubt our assertion. See. Fig. 5.
 Fig. 5.

A function is said to be causal if it is zero for negative range. For example the

1 x> 0
1
2{
Heaviside function is causal : H ( x ) = x=0 .
0 x< 0
(3)

Exponential function used to describe radioactivity is also causal.

− λt
f ( x )= N 0 e .
{ 0
(4)

Half wave rectifier is causal because for negative duration of the period, the wave is
zero.
CAUSALITY AND QUADRATURE RESPONSE: One might have noticed
that when a cosine periodic force is acting on a damped harmonic
oscillator , in the response (displacement) there is a component
proportional to cosine periodic force and also a component of
displacement proportional to sine of the periodic force! This is surprising
as no sine periodic force was applied. Yet the system responds as if a sine
periodic force were also applied. This displacement is said to be ‘in
quadrature’ response and that proportional to cosine force is said to be
‘in phase’ response. The differential equation of a damped harmonic
oscillator (DSHO) is [5, 6]

d 2 x (t) dx(t) 2
+ ω0 x ( t )= Cδ ( t ) t >0 ω 0> α ,
dt 2
+2 α
dt {
0t <0
(5)

ω
where 0 is the natural frequency and 2α is the damping constant
per unit mass and C is the strength of the impulse. x ( t ) is the
displacement and ẋ ( t )=v (t) is the velocity of the particle. It can be shown
that C=v 0 [2]. The velocity is given by [2]

sin ω 0 t

{ −αt
(
v ( t )= v 0 e cos ω 0 t−α
0 t <0
ω0 )
t >0
. (6)

where ω 0 is the natural frequency of the DSHO . If a periodic force

f 0 cosωt is applied ¿time t ≥ 0 , where f 0 is the force per unit mass.

The velocity is now given by

ω20
f0
[ ]
V (t)=
{
[( ) ]
ω−
ω20
ω
2

+α 2
αcosωt − ω−

0 t <0
(
ω )
sinωt t> 0
, (7)

Thus there are two components for the velocity: first ‘in-phase’
α
cosωt
2 2
component
[( ) ]
ω−
ω
ω
0
+α 2
and ‘in-quadrature’ component
 ω 20
( ω−
ω )
sinωt . It is surprising that there is a response proportional
2
ω 20
(ω−
ω )
+ α2

to sinωt and that oscillating force was not applied. Nevertheless the
response has that component also. A causal linear passive system always
produces an impulse response with both ‘in-phase’ and ‘in-quadrature’
response. We can write eq. (.) as

V ( t ) =A ( ω ) cos ⁡(ωt +ϕ ( ω ))

f0
ω 20
where
A ( ω )=
ω20
2
and tan ( ω− )
(8)
√( ω−
ω )
+α 2
ϕ ( ω )=
α
ω
.

The in-phase and in quadrature responses are not independent of each

other as the the principle of causality ascertains that that for a physically
realizable system it is not possible to give an arbitrary characteristic for
the in-phase response without setting up a definite in-quadrature
response and hence a definite phase characteristic. For more information
the reader is referred to ref. 2 and 3 in Academia.edu.

References:

1. E. Butkov, Mathematical physics, Addison-Wesley Pub Co (Reading), 1968, Ch. 7

2. Jambunatha Sethuraman, “A convolution approach to damped harmonic
oscillator’,
https://www.academia.edu/5385350/CONVOLUTION_APPROACH_TO_DAMPED
_HARMONIC_OSCILLATOR
3. Jambunatha Sethuraman, ‘Hilbert transform, Causality, analyticity, and
Fraunhofer diffraction’,
https://www.academia.edu/5453386/Causality_Analyticity_Hilbert_transform_
and_Fraunhofer_diffraction