Principles of Communications

Lecture 7: Analog Modulation

Techniques (5)
Chih-Wei Liu
National Chiao Tung University
cwliu@twins.ee.nctu.edu.tw
Outlines

Linear Modulation
Angle Modulation
Interference
Feedback Demodulators
Analog Pulse Modulation
Delta Modulation and PCM
Multiplexing

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 2
Analog Pulse Modulation

Message waveform is represented by regularly spaced

sample values (sample signals) discrete in time.
Historically, these methods are the early attempts to
achieve modern communications. They are in the twilight
zone between analog and digital modulations.
Today, their basic forms can still be found in some
electronic components such as ADC.

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 3
Analog/Digital Pulse-coded
Analog pulse modulation: A pulse train is used as the
carrier wave. Some characteristic feature of each pulse (e.g.,
amplitude, duration, or position) is used to represent message
samples.
PAM pulse amplitude
PDM pulse duration
PPM pulse position
Digital Pulse Modulation: Messages are discrete-amplitude
(finite levels) samples.
DM delta modulation
PCM pulse-code modulation
Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 4
Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 5
Pulse-Amplitude Modulation (PAM)
The amplitude of each pulse corresponds to the value of the
message signal m(t) (at the leading edge of the pulse).

t ( nTS + 0.5 )
mc ( t ) = m ( nTS )
n =
t 1, t < W 2
where =
W 0, otherwise
The pulse generator can be considered as a filter.
t 0.5
c
m ( t ) = m ( t ) h ( t ), h ( t ) =

M ( f ) = M ( f ) H ( f ), H ( f ) = sincf e j f
c

m ( t ) = n = m ( nTs ) ( t nTs ) sampled message

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 6
PAM Signals

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 7
Demodulation

Recover M(f) m(t) samples Mc( f )

M ( f ) =
Recover M(f) m(t) message H( f )

m c(t) 1/H(f) m (t) m(t)

LPF
equalizer
Equalizer: Recover distorted signals particularly when
the distortion method is known or estimated.

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 8
Pulse-Width Modulation (PWM)
Pulse width the values of message
Spectrum: complicated (Fourier-Bessel spectra)

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 9
PWM/PPM Generation
Mod: <Carlson,
Fig.6-3-2>

Demod: area of
pulse.
Low-pass filtering
(integration)

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 10
Pulse-Position Modulation (PPM)

Pulse position the values of message

x (t ) = g (t t )
n =
n

Spectrum: complicated
Demodulation: (1) LPF & integration
(2) convert PPM to PWM LPF
Conversion of PPM or PWM to PAM: a ramp generator
(re)starts at kTs and stops at tk. (next page)

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 11
Conversion Between PWM & PPM

<Carlson, Fig.6-3.3> (integration)

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 12
Delta Modulation
m(t) samples (analog amplitude) difference binary
or m(t) difference binary samples
Operations:
(1) d (t ) = m(t ) m s (t )
0 , d(t) 0
(2) (t ) = threshold (d (t )) =
0 , d (t ) < 0

(3) xc (t ) = samples of (t ) = (t ) (t - nTs ) = (nT ) (t nT )
s s
n =- n =

(nT ) ( nT )d
t
(4) ms (t ) = S S
n =-

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 13
DM Signal
Generation

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 14
Slope Overload
The message signal m(t) has a slope greater than can be
followed by the stair-step approximation ms(t).
Assume step-size = 0 slope (max) = 0/Ts

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 15
Solution to Overload

Adaptive delta modulation -- adjust the step-size 0

based on xc(t).
Idea: If m(t) constant, xc(t) alternates in sign
make 0 .
If m(t) ( or ) rapidly, xc(t) has the same polarity
make 0.
Method: Detect the trend of signal

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 16
Adaptive DM

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 17
ADM Receiver
Transmit step-size or regenerate the step-size at the
receiver according to pre-decided rules.

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 18
Pulse-Code Modulation (PCM)
m(t) samples (analog amplitude) quantized
samples binary representation binary modulated
waveform (ASK (AM), PSK (PM), FSK (FM) to be
discussed in Commun.II )

Main advantages of digital communication

more reliable communication
Main disadvantages of digital communication
wide BW (reduced by compression)
complicated circuits ( cost reduced by VLSI)

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 19
PCM Signal Generation

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 20
BW of PCM
Assume the number of quantization levels=q=2n
Message BW = W
Sampling rate = 2W
2nW binary pulses/second
1
Assume maximum width of pulse, =
2nW
transmission BW knW, k=constant
Hence, B k2Wlog2q
Recovered message error is due to mainly quantization
error. Thus, q error B

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 21
Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 22
Multiplexing

A number of data sources share the same

communication channel.
z Frequency-Division Multiplexing (FDM)
z Quadrature Multiplexing (QM)
z Time-Division Multiplexing (TDM)

hmhang, EE/NCTU Nov 2009

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 2323
FDM
Several
message
signals are
translated,
using
modulation,
to different
spectral
locations
and added
to form a
baseband
signal.

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 24
Example:
Stereophonic
FM

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 25
Quadrature Multiplexing (QM)
Quadrature-carrier multiplexing: transmit two signals on
the same carrier frequency. (not exactly FDM)
Note that cos and sin are orthogonal.
QM Quadrature Amplitude Modulation (QAM)

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 26
<Modulation>
xc (t ) = AC [m1 (t ) cos ct + m2 (t )sin ct ].

<Coherent Demodulation>
If the receiver has a carrier phase error, i.e.,
LO(t ) = 2cos(ct + ).
xr (t ) 2cos(ct + )
= AC [m1 (t )cos m2 (t )sin ]
+AC [m1 (t )cos(2ct + ) + m2 (t )sin(2ct + )].
yDD (t ) = AC [m1 (t )cos m2 (t )sin ]. (ideal : 0 )

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 27
Time-Division Multiplexing (TDM)
Each message signal occupies a small time slot in every
Ts second.

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 28
BW of TDM

A rough estimate of BW
Baseband message BW = Wi . There are N channels.
Samples per T second = 2WiT .
N
Total samples per T second: ns = 2WiT . Or,
i =1
N
Total samples per second = 2W .
i =1
i

N
Total baseband BW to accommodate all sources = B W .
i =1
i

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 29
Example:
Digital
telephony
system

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 30
Comparison of MUX

FDM: simple to implement,

inter-modulation distortion (crosstalk) due to nonlinear
channels
TDM: less crosstalk (in memoryless channels),
difficult to keep synchronization (frame structure, header),
digital (sampled) signals
QM: efficient use of channel, crosstalk between I & Q
channels (needs coherent demodulation)

Commun.-Lec7 cwliu@twins.ee.nctu.edu.tw 31