Communications
Lecture 6
Sampling and Analog to digital Conversion:
Sampling is the process of converting a continuous time signal to discrete time signal, in
order to process and transmit the data digitally.
Analog signals can be digitized through sampling and quantization. This analog-to-digital
(A/D) conversion sets the foundation of modern digital communication systems.
In the A/D converter, the sampling rate must be large enough to permit the analog signal
to be reconstructed from the samples with su cient accuracy.
Quantization is the process of mapping continuous in nite values to a smaller set of discrete
nite values.
1
Let's consider the following examples:
2
SAMPLING THEOREM:
A band-limited signal g(t) of bandwidth B( Hz) can be reconstructed exactly from data
samples taken at a sampling rate fS if fS is greater than or equal to 2B( Hz).
For example, if the frequency of a sine wave is 100 Hz. We need to sample this at a rate
higher than 200 Hz (i.e. 200 samples per second) in order NOT to lose any data, i.e. to be
able to reconstruct the 100 Hz sine wave exactly.
Sampling g(t) at a rate of fs Hz means that we take uniform samples every Ts = 1
fs sec-
ond.
Uniform sampling can be accomplished by multiplying g(t) by an impulse train δTs (t) con-
sisting of unit impulses repeating periodically every Ts seconds.
This results in the sampled signal ḡ(t) shown in the gure.
The n th impulse, located at t = nTs has a strength g (nTs ) which is the value of g(t) at
t = nTs .
3
X
ḡ(t) = g(t)δTs (t) = g (nTs ) δ (t − nTs )
n
Because the impulse train δTs (t) is a periodic signal of period Ts , it can be expressed as an
exponential Fourier series:
∞
1 X jnωs t 2π
δTs (t) = e ωs = = 2πfs
Ts n=−∞ Ts
Therefore,
ḡ(t) = g(t)δTs (t)
∞
1 X
= g(t)ejn2πfs t
Ts n=−∞
ḡ(t) = g(t)δTs (t)
∞
1 X
= g(t)ejn2πfs t
Ts n=−∞
Based on the frequency-shifting property, the transform of the nth term is shifted bynfs
∞
1 X
Ḡ(f ) = G (f − nfs )
Ts n=−∞
4
This means that the spectrum Ḡ(f ) consists of G(f ), scaled by a constant 1/Ts , repeating
periodically with period fs = 1/Ts Hz as shown above.
Graphically from the gure above, perfect recovery of original signal is possible if there is
no overlap among the replicas in Ḡ(f ). This requires
fs > 2B
The minimum sampling rate fs = 2B required to recover g(t) from its samples ḡ(t) is called
the Nyquist rate, and the corresponding sampling interval Ts = 1/2B is called the Nyquist
interval.
If a signal is sampled below the Nyquist rate, them the issue of aliasing occurs. When a
continuous-time band-limited signal is sampled at a rate lower than Nyquist rate fs < 2B ,
then the successive cycles of the spectrum Ḡ(f ) of the sampled signal g(t) overlap with each
other as shown in the gure below:
In fact, aliasing is the phenomenon in which a high frequency component in the frequency-
spectrum of the signal takes identity of a lower-frequency component in the spectrum of the
sampled signal.
This leads to loss of information and imperfect reconstruction of the sampled signal. There-
fore, for successful sampling process
fs > 2B
Example : Given the signal x(t) = 3 cos 100πt, determine the minimum sampling rate.
ω = 2πf = 100π =⇒ f = 50 Hz
Since this is a sinusoidal signal, the bandwidth = 50 Hz. The minimum sampling frequency
i.e. Nyquist rate fs = 2B = 2 × 50 = 100 Hz.
Example: Given the signal x(t) = 5 cos 50πt + 4 sin 300πt + 3 cos 100πt, determine the
minimum sampling rate.
The frequency components are:
5 cos 50πt −→ f1 = 25 Hz
4 sin 300πt −→ f2 = 150 Hz
3 cos 100πt −→ f3 = 50 HZ
5
As the bandwidth is decided by the maximum frequency component, the bandwidth B = 150
Hz.
The minimum sampling frequency i.e. Nyquist rate fs = 2B = 2 × 150 = 300 Hz.
Example : Given the signal xa (t) = 4 cos 2000πt + 7 sin 6000πt − 6 cos 12000πt, , determine
the minimum sampling rate.
Signal Reconstruction:
If a signal is sampled at the Nyquist rate fs = 2B Hz, the spectrum Ḡ(f ) consists of
repetitions of G(f ) without any gap between successive cycles, as shown in part a of the
gure above. To recover g(t) from ḡ(t), we need to pass the sampled signal ḡ(t) through an
ideal low-pass lter (dotted area in part a). Such a lter is unrealizable in practice.
A practical solution to this problem is to sample the signal at a rate higher than the Nyquist
rate (fs > 2B or ωs > 4πB). This yields Ḡ(f ), consisting of repetitions of G(f ) with a nite
band gap between successive cycles, as shown in part b of the gure above. We can now
recover G(f ) from Ḡ(f ) by using a low-pass lter with a gradual cuto characteristic (dotted
area in part b ).
