Generation and Demodulation of
Amplitude Modulated Signals
Gaurav S. Kasbekar
Dept. of Electrical Engineering
IIT Bombay
Introduction
• Consider a carrier wave:
1) 𝑐 𝑡 = 𝐴𝑐 cos(2𝜋𝑓𝑐 𝑡),
where 𝐴𝑐 > 0 is “carrier amplitude” and
𝑓𝑐 > 0 is “carrier frequency”
• Let 𝑚(𝑡) be message signal, which is a baseband
signal
• Amplitude modulated (AM) wave given by:
2) 𝑠 𝑡 = 𝐴𝑐 [1 + 𝑘𝑎 𝑚 𝑡 ]cos(2𝜋𝑓𝑐 𝑡)
where 𝑘𝑎 > 0 called “amplitude sensitivity”
• We will study circuits for generating AM signals and
demodulating them
Switching Modulator
• Switching modulator for generating AM signal shown in Fig. (a)
• 𝑐 𝑡 = 𝐴𝑐 cos 2𝜋𝑓𝑐 𝑡 : carrier signal and 𝑚(𝑡): message signal
• 𝑣1 𝑡 :
𝐴𝑐 cos 2𝜋𝑓𝑐 𝑡 + 𝑚(𝑡)
• Assume that diode acts as ideal switch:
presents zero (respectively, infinite) impedance when forward
(respectively, reverse) biased
• So plot of 𝑣2 𝑡 versus 𝑣1 𝑡 :
as shown in Fig. (b)
• Assuming that 𝑚 𝑡 ≪ 𝐴𝑐 , 𝑣2 𝑡 in terms of 𝑣1 𝑡 :
𝑣 𝑡 , 𝑐 𝑡 > 0,
𝑣2 𝑡 ≈ ቊ 1
0, 𝑐 𝑡 < 0.
Ref: “Communication Systems” by S. Haykin and M. Moher, 5th ed
• Recall: Switching Modulator (contd.)
𝑣1 𝑡 , 𝑐 𝑡 > 0,
1) 𝑣2 𝑡 ≈ ቊ
0, 𝑐 𝑡 < 0.
where 𝑣1 𝑡 = 𝐴𝑐 cos 2𝜋𝑓𝑐 𝑡 + 𝑚(𝑡)
• Relation in 1) can be written as:
2) 𝑣2 𝑡 ≈ 𝐴𝑐 cos 2𝜋𝑓𝑐 𝑡 + 𝑚 𝑡 𝑔𝑇0 (𝑡)
1
where 𝑇0 = 𝑓 and 𝑔𝑇0 (𝑡) is as in fig.
𝑐
• How can we obtain an AM signal from the signal 𝑣2 𝑡 in 2)?
• Fourier series of 𝑔𝑇0 (𝑡):
1 2 (−1)𝑛−1
𝑔𝑇0 𝑡 = 2 + 𝜋 σ∞
𝑛=1 2𝑛−1
cos 2𝜋𝑓𝑐 𝑡(2𝑛 − 1)
• AM signal obtained by:
passing 𝑣2 𝑡 through a band-pass filter of mid-band frequency 𝑓𝑐 and bandwidth 2𝑊
• Generated AM signal given by:
𝐴𝑐 4
1 + 𝜋𝐴 𝑚(𝑡) cos 2𝜋𝑓𝑐 𝑡
2 𝑐
• Amplitude sensitivity:
4
𝑘𝑎 = 𝜋𝐴
𝑐
Ref: “Communication Systems”
by S. Haykin and M. Moher, 5th
ed
Envelope Detector
• Recall: AM wave given by:
𝑠 𝑡 = 𝐴𝑐 [1 + 𝑘𝑎 𝑚 𝑡 ]cos(2𝜋𝑓𝑐 𝑡)
• Envelope of 𝑠 𝑡 has same shape as 𝑚 𝑡 if
following conditions satisfied:
1) 𝑘𝑎 𝑚 𝑡 < 1, for all 𝑡
2) 𝑓𝑐 ≫ 𝑊
• If 1) and 2) are satisfied, then “envelope
detection” can be used to recover 𝑚(𝑡) from
𝑠(𝑡)
• Next, we study an envelope detector circuit
widely used in commercial AM radio receivers
Envelope Detector (contd.)
• AM signal 𝑠(𝑡) shown in Fig. (b) is input to
envelope detector circuit shown in Fig. (a)
• Assume that:
internal resistance of voltage source is 𝑅𝑠
diode is ideal with resistance 𝑟𝑓 in forward bias
and infinite resistance in reverse bias
load resistance 𝑅𝑙 ≫ (𝑟𝑓 + 𝑅𝑠 )
• On a positive half-cycle of 𝑠(𝑡):
capacitor rapidly charges up to peak value of 𝑠(𝑡)
• When value of 𝑠(𝑡) falls below its peak value:
capacitor slowly discharges through load resistor
𝑅𝑙
discharging continues until 𝑠(𝑡) becomes greater
than capacitor voltage
• Then above process repeats
• Charging and discharging time constants:
𝑟𝑓 + 𝑅𝑠 𝐶 and 𝑅𝑙 𝐶 respectively
• What should be relation between 𝑟𝑓 + 𝑅𝑠 𝐶
and 𝑓𝑐 ?
1
𝑟𝑓 + 𝑅𝑠 𝐶 ≪
𝑓𝑐
• What should be relation between 𝑅𝑙 𝐶, 𝑓𝑐 and
𝑊?
1 1
≪ 𝑅𝑙 𝐶 ≪
𝑓𝑐 𝑊
Ref: “Communication
• Fig. (c) shows output of envelope detector Systems” by S. Haykin and M.
Moher, 5th ed
Envelope Detector (contd.)
• Fig. (c) shows output of
envelope detector
• Finally, message signal
𝑚(𝑡) can be obtained
form envelope detector
output by:
low-pass filtering
Ref: “Communication
Systems” by S. Haykin and M.
Moher, 5th ed
Generation of DSB-SC Signal
• Let 𝑚(𝑡) (respectively, 𝑐(𝑡)) be message signal (respectively,
carrier signal)
• Recall: in DSB-SC, modulated wave is product of 𝑚(𝑡) and 𝑐(𝑡):
𝑠 𝑡 = 𝑐 𝑡 𝑚 𝑡 = 𝐴𝑐 𝑚 𝑡 cos(2𝜋𝑓𝑐 𝑡)
• Multiplier Modulator:
Modulation achieved directly by multiplying 𝑚 𝑡 with 𝐴𝑐 cos(2𝜋𝑓𝑐 𝑡)
using an analog multiplier
• E.g. of analog multiplier:
variable gain amplifier in which gain parameter is controlled by 𝑚 𝑡
when 𝐴𝑐 cos(2𝜋𝑓𝑐 𝑡) is applied at input of this amplifier, output is
𝐴𝑐 𝑚 𝑡 cos(2𝜋𝑓𝑐 𝑡)
• In early days of DSB-SC, it was challenging to implement analog
multiplier circuit
as semiconductor technologies advanced, became easy to do so
• A classical modulator that avoids use of multiplication is switching
modulator
Switching Modulator
• Recall: we generated following AM signal using the
switching modulator shown in fig:
𝐴𝑐 4
1) 2
1+
𝜋𝐴𝑐
𝑚(𝑡) cos 2𝜋𝑓𝑐 𝑡
• How can we generate DSB-SC signal using switching
modulator?
𝐴𝑐
by subtracting cos 2𝜋𝑓𝑐 𝑡 from signal in 1)
2
Ref: “Communication
Systems” by S. Haykin and M.
Moher, 5th ed
Coherent Detection
• Recall: in DSB-SC, modulated wave is product of
𝑚(𝑡) and 𝑐(𝑡):
𝑠 𝑡 = 𝑐 𝑡 𝑚 𝑡 = 𝐴𝑐 𝑚 𝑡 cos(2𝜋𝑓𝑐 𝑡)
• 𝑚(𝑡) can be recovered from 𝑠 𝑡 by:
multiplying 𝑠 𝑡 with locally generated sinusoid that is
coherent (synchronized) in both frequency and phase
with carrier wave 𝑐 𝑡 used in modulator
low-pass filtering the product
• This method called “coherent detection” or
“synchronous demodulation”
Demodulation Using A Local Oscillator with Phase Error
• Difficult to locally generate sinusoid that is synchronized in both frequency and
phase with carrier wave 𝑐 𝑡 used in modulator
• Suppose locally generated sinusoid is:
𝐴′𝑐 cos(2𝜋𝑓𝑐 𝑡 + 𝜙)
• Signal obtained by multiplying 𝑠(𝑡) with 𝐴′𝑐 cos(2𝜋𝑓𝑐 𝑡 + 𝜙) and low-pass filtering
product:
𝐴𝑐 𝐴′𝑐
𝑚(𝑡)cos(𝜙)
2
• If 𝜙 is a constant, then:
±𝜋
If 𝜙 ≠ 2 , then demodulated signal is scaled, but undistorted version of 𝑚(𝑡), which is
acceptable
±𝜋
If 𝜙 = , then demodulated signal is 0
2
• However, in practice, phase error 𝜙 varies randomly with time due to:
random variations in communication channel (e.g., mobility of transmitter or receiver in
wireless communication)
𝐴𝑐 𝐴′𝑐
• Then demodulated signal is 𝑚(𝑡)cos(𝜙),
where cos(𝜙) varies with time, which is
2
distorted version of 𝑚(𝑡)
• Hence, techniques are required to maintain local oscillator at receiver in perfect
synchronism in both frequency and phase with carrier signal used to generate DSB-SC signal
in transmitter
resulting system complexity is the price that must be paid for suppressing carrier signal in AM to save
transmitted power
Costas Receiver
• As discussed, in DSB-SC demodulator, phase error 𝜙 often varies with time
results in distortion in demodulated signal
• Costas receiver designed for achieving phase synchronization between:
locally generated carrier at receiver and that used at transmitter
shown in fig.
• Phase discriminator multiplies its two inputs
• VCO configured such that if its input is:
positive, then phase of output decreases
negative, then phase of output increases
• Typically, phase error is small
𝜋 𝜋
• If initially, − < 𝜙 < , then:
2 2
eventually phase error 𝜙 driven to zero
Ref: “Communication
Systems” by S. Haykin and M.
Moher, 5th ed