Chap2 - AM-part I
Chap2 - AM-part I
1
Chap. 2: Amplitude (Linear) Modulation
Now that we are familiar with the basic signal analysis techniques, we want to move
ahead to discuss operational communication systems.
• Amplitude Modulation (AM) and its variants;
• Angle Modulation with special case - FM;
Message signal: m(t) baseband signal with bandlimited to B Hz. For efficient
transmission, m(t) has to be transmitted over a bandpass communication channel.
• This requires a shift of the range of frequencies in m(t) into other frequency
ranges suitable for transmission over a communication channel (usually shifting
up in frequency);
• At the receiver, this process has to be reversed (down shifting in frequency);
• Why do we need to move to higher frequency band – antenna size, available
bandwidth, multiple access, etc. (for example, cellular phone, antenna size - λ/4.
For f1 = 3k, λ1 = c/f1 = 105 m, λ1 /4 = 2.5 × 104 m; while f2 = 900M ,
λ2 = c/f2 = 1/3m, λ2 /4 = 8cm.)
2
Definition: The process by which some characteristics of a carrier signal is varied in
accordance with a modulation signal is called MODULATION.
• A common form of the carrier we will use, is a sinusoidal wave. Let
c(t)
carrier signal
narrowband
3
• time-varying amplitude a(t)
• generalized phase θ(t)
• instantaneous frequency fi (t):
1 d
fi (t) = θ(t)
2π dt
If a(t) changes as m(t), then we have amplitude modulation (AM). If θ(t) changes as
m(t), then we have angle modulation (PM, FM).
In the case of AM,
a(t) ∼ m(t)
4
Section 2.1: DSB-SC
Let us consider the simple and intuitive AM case. Let the message signal to be m(t)
with a bandwidth of B Hz. The carrier signal is
c(t) = Ac cos(2πfc t)
with fc >> B.
m(t) M (f )
K
t f
0
c(t) C (f )
Ac /2
Ac
t f
−fc 0 fc
t f
−fc 0 fc
5
Then the AM wave is given by the expression
6
– Demodulation: DSB-SC → m(t) shift in frequency again
Ac
φ(t) cos(2πfc t) = Ac m(t) cos2 (2πfc t) = [1 + cos(4πfc t)]m(t)
2
Ac Ac
= m(t) + m(t) cos(4πfc t)
2 2
Ac
after low-pass filter, the output is 2 m(t).
7
Section 2.1.1: Generation of DSB-SC Signals
A non-linear device spreads the spectrum of its input. By controlling what is used as
an input to the non-linear device(s) and by eliminating what is not needed we may
generate the desired DSB-SC amplitude modulated signal. Consider a non-linear
device
y(t) = ax(t) + bx2 (t)
which is used in the following system
z(t) DSB-SC
Bandpass
Filter
8
x1 (t) = m(t) + cos(ωc t) x2 (t) = m(t) − cos(ωc t)
y1 (t) = ax1 (t) + bx21 (t) y2 (t) = ax2 (t) + bx22 (t)
z(t) = y1 (t) − y2 (t) (2)
= a[m(t) + cos(ωc t)] + b[m(t) + cos(ωc t)]2
−a[cos(ωc t) − m(t)] − b[cos(ωc t) − m(t)]2
= am(t) + a cos(ωc t) + bm2 (t) + 2bm(t) cos(ωc t) + b cos2 (ωc t)
am(t) − a cos(ωc t) − bm2 (t) + 2bm(t) cos(ωc t) − b cos2 (ωc t)
= 2am(t) + 4bm(t) cos(ωc t) (3)
9
m(t) x1 (t) Non-Linear
y1 (t)
Device
z(t) DSB-SC
Bandpass
Filter
Then the first term in (3) can be filtered out and the second term can be kept by
applying a bandpass filter (BPF). The BPF is tuned to fc with a bandwidth 2B Hz.
10
(2) Switching Modulator
and
∑
m(t) · w(t) = Cn m(t) cos(nωc t + θn )
n
such that
1∑ 1∑
F[m(t)·w(t)] = Cn M (f )∗[δ(f −nfc )+δ(f +nfc )] = Cn [M (f +nfc )+M (f −nfc )]
2 n 2 n
11
Thus, the spectrum of the product m(t)w(t) is the spectrum of the modulating
waveform, M (f ) shifted to ±fc , ±2fc , ±3fc , · · · and scaled by Cn . If we pass this
signal through a BPF of bandwidth 2B Hz and tuned to fc , then we have the desired
12
modulated signal m(t) cos(2πfc t).
The above operation is also known as mixing, frequency conversion, or heterodyning.
w(t) fundamental
frequency: fc
t
m(t)
A M (f )
t f
0
y(t)
Y (f )
t f
−fc 0 fc
ϕDSB-SC (t)
ΦDSB-SC (f )
t f
−fc 0 fc
13
Analysis of Switching Modulator
Let’s expand Fourier Series for a gate function w(t), with duty cycle η = 0.5. It is the
periodic extension of its basic function x(t) = rect( T t/2 ) with a period T :
∞
∑ ∞
∑ ( )
t − nT
w(t) = x(t − nT ) = rect
n=−∞ n=−∞
T /2
( ) (π )
1 T 1 T 1
Dn = X(nf0 ) = sinc π · n · = sinc n . (4)
T 2T T 2 2 2
Obviously,
D0 = 1/2; Dn = 0 when n equals even numbers
14
Let’s express w(t) as trigonometrical form as
∞
∑
w(t) = C0 + Cn cos(nωc t + θn )
n=1
For n = 1,
1 (π) 1 2 1
D1 = sinc = · =
2 2 2 π π
1 ( π ) 1 2 1
D−1 = sinc − = · =
2 2 2 π π
Similarly, D3 = −1/(3π), D5 = 1/(5π), · · · . Therefore, the coefficients of {Cn } are:
D =1 n=0
0 2
Cn = (5)
2Dn = (−1)(n−1)/2 · 2 n = 1, 3, 5, 7, · · · ,
nπ
15
The gate function can be written as
[ ]
w(t) = 1
2 + 2
π cos(ωc t) − 1
3 cos(3ωc t) + 1
5 cos(5ωc t) + · · · .
(6)
so
[ ]
1 2 1 1
m(t)w(t) = m(t) + m(t) cos(ωc t) − m(t) cos(3ωc t) + m(t) cos(5ωc t) + · · · .
2 π 3 5
After bandpass filter centered at ωc , we have DSB-SC signal
2
φ(t) = m(t) cos(ωc t) (7)
π
16
Section 2.1.2: DSB-SC Demodulation
• There are other AM generation techniques. Refer to the course reference text for
further information and discussion.
• We will use the terms mixing / frequency conversion / heterodyning
interchangeably. These terms refer to the process:
multiplication + bandpass filtering
• Now that we know how to generate DSB-SC amplitude modulated signals, let us
investigate how we can undo what we have done
Demodulation
17
DSB-SC: Demodulation
18
DSB-SC: Demodulation – cont’d
M(f)
m(t)
K
t f
0 B
cos(2πfc t)
KAc /2 ΦDSB-SC (f)
t
f
0 fc
cos2 (2πfc t)
t f
0 B 2fc
19
Section 2.1.3: Synchronization for DSB-SC Demodulation
Modulator Demodulator
ϕ(t)
m(t) Channel LPF K ′ m(t)
~ ~
Local Local
Oscillator-1 Oscillator-2
20
Synchronization for Coherent Demodulation
21
Unsynchronized Demodulation
l2 (t) = cos(2π(fc + ∆f )t + ϕ0 )
22
We assume that the error/mismatch between LO1 and LO2 is small, i.e., ∆f << fc .
Therefore, the 2nd component in (8) will be centerred about f = 2fc and will be
filtered out by the demodulator LPF. Thus, the LPF output equals
Ac
y(t) = m(t) cos(2π∆f t + ϕ0 )
2
Thus, y(t) is oscillating at a slow rate. For example, let m(t) = e(−2t) . The
demodulator output is an oscillation function as shown in the figure below.
23
1
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5
−0.5
−1
0 0.5 1 1.5 2 2.5
24
Assume that ∆f = 0, then
Ac
y(t) = m(t) cos(ϕ0 )
2
It can be written as
Ac
2 m(t) if ϕ0 = 0 (best)
Ac
y(t) = 2 m(t) cos(π/2) =0 if ϕ0 = π/2 (worse)
varying degrees of attenuation Other values of ϕ0
25
Section 2.2: Amplitude Modulation (AM) – Large Carrier
• We have seen that the use of suppressed carrier waveforms requires fairly
complicated circuitry at the receiver in order to acquire and maintain phase
synchronization (expensive receivers).
• Especially in applications where we have one (or few) TX and a much larger
number of RXs, it makes economic sense that the RXs are as simple as possible.
• To address this issue we entertain the idea of transmitting a carrier in the same
frequency band as the DSB-SC, which is DSB-LC
• This approach will certainly ruin the DC response of m(t), but many signals do
not have any DC content anyway. Hence we have
and
1 Ac
ΦAM (f ) = [M (f + fc ) + M (f − fc )] + [δ(f + fc ) + δ(f − fc )]
2 2
26
AM-LC Spectrum
ΦAM (f ) carrier
DSB-SC
f
−fc
0 fc
27
AM-LC Compared with DSB-SC
m(t) M (f )
t f
0
ΦDSB-SC (f )
ϕDSB-SC (t)
t f
0 fc
ϕAM (t)
ΦAM (f )
t f
0 fc
28
Section 2.2.1: Envelope Detection
We will refer to the envelope of the signal, where we use definition of the envelope as
follows: given the signal E(t) cos(ωc t), E(t) is the envelope if E(t) ≥ 0 for all t. In a
more general sense we will use |E(t)| as the envelope of E(t) cos(ωc t).
For an AM signal φ(t) = [Ac + m(t)] cos(ωc t), amplitude and envelope functions are:
amplitude = Ac + m(t)
envelope = |Ac + m(t)|
29
Concept of Envelope
t t
30
Condition for Envelope Detection
31
Envelope Detection Analysis
32
Modulation Index
• If m(t) ≥ 0 for all t, there is no sense to add a carrier signal Ac cos(ωc t). We can
apply envelope detection to the signal m(t) cos(ωc t). Therefore, we are interested
to investigate the case that m(t) has negative amplitude for some t.
• Let us define mp = maxt |m(t)|. Then,
µ = mp /Ac
0≤µ≤1
33
Other Definition of Modulation Index
Let
Amax = max A(t) = max[Ac + m(t)] Amin = min A(t) = min[Ac + m(t)]
t t t t
Then
Amax − Amin
Modulation Index : µ= (11)
2Ac
Amax − Ac
Positive Modulation Index : µ+ =
Ac
Ac − Amin
Negative Modulation Index : µ− =
Ac
34
mp = max |m(t)| = Am (12)
t
Amax = max[Ac + m(t)] = Ac + Am
t
Amin = min[Ac + m(t)] = Ac − Am
t
35
Hence Am = µAc and
Therefore,
36
Case 1: µ = Am /Ac < 1
E(t) Emax = Ac + Am
37
Case 2: µ = Am /Ac = 1
38
Case 3: µ = Am /Ac > 1
Emax = Ac + Am
A(t): amplitude function
E(t): envelope function Emin = Ac − Am < 0
E(t) = |Ac + Am cos ωm t| ̸= A(t) such that E(t) is no longer related to m(t)
39
Another example: un-symmetric m(t)
40
• Using the concept of mp ,
mp 0.8
µ= =
Ac Ac
• Using the concept of Amax and Amin
Amax − Amin 0.55
µ = = (14)
2Ac Ac
Amax − Ac 0.8
µ+ = =
Ac Ac
Ac − Amin 0.3
µ− = =
Ac Ac
• We observe all different modulation index results since m(t) is not symmetric
with respect to horizontal axis.
41
2.2.2: Sideband Power and Carrier Power
The advantage of envelope detection in AM has its price. In AM, the carrier term
does not carry any information, and hence, the carrier power is wasted. Let’s look at
an AM signal
φAM (t) = Ac cos ωc t + m(t) cos ωc t
The carrier power Pc and sideband power Ps are given as
A2c 1 2
Pc = Ps = m (t)
2 2
Observations
• Any power used to transmit the carrier is for Envelope detection purpose.
• Sidebands are a function of m(t). Should maximize sideband power Ps .
• Define power efficiency
useful power Ps m2 (t)
η= = =
total power P c + Ps A2c + m2 (t)
42
Single Tone AM Spectrum Analysis 1
43
Single Tone AM Spectrum Analysis 2
44
Single Tone AM Spectrum Analysis 3
45
Therefore, for single-tone modulation,
• Modulating signal: m(t) = Am cos ωm t = µAc cos ωm t
• Carrier: c(t) = Ac cos ωc t with fc >> fm
• Modulation index: µ = Am /Ac
• Sideband and Carrier Power
( )2
m2 (t) 1 Am A2m µ2 A2c A2c
Ps = =2· = = , Pc =
2 2 2 4 4 2
Hence
µ2 A2c
Ps µ2
η= = A2c
4
µ2 A2c
= 0≤µ≤1
Pc + Ps + 2 + µ2
2 4
1
It can be seen that η increases monotonically with µ, and ηmax = 3 occurs at µ = 1.
46
Section 2.2.3: Generation of AM Signals
We can use the same DSB-SC generation methods except that an additional carrier
component Ac cos(ωc t) needs to be added.
47
vbb′ (t) = [Ac cos ωc t + m(t)] · w(t) (15)
[ ( )]
1 2 1 1
= [Ac · cos ωc t + m(t)] + cos ωc t − cos 3ωc t + cos 5ωc t − · · ·
2 π 3 5
Ac 2
= cos ωc t + m(t) cos ωc t + other terms
2 π
48
Section 2.2.4-Demodulation of AM Signals
49
These operations can also be described analytically
50
(b) Envelope detector
The biggest selling point for the AM signals was the simple/inexpensive detection,
i.e., envelope detector (ED). ED is a simple but effective device well suited to the
demodulation of AM signals with µ ≤ 1. Ideally the output of the ED follows the
envelope of φ(t).
The output of the ED is then lowpass filtered to eliminate the ripple, followed by
blocking out the DC component.
51
52
53
Section 2.2.5: AM Practical Applications
54
55
AM: Examples
56