Sampling
Introduction
A-D convertor
Digital system
Fig. 3-1.
2/24
Sampling theory
Analog signal
Discrete signal
Sampler
Fig. 3-2.
3/24
Interval, T > 0
Fig. 3-3.
4/24
Fig. 3-4.
5/24
(3-1)
(t nT )
(3-2)
n =
x(nT ) (t nT )
(3-3)
n =
6/24
1
[ X ( j ) S ( j )]
2
(3-4)
By example 2-8
2
=
S ( j )
T
( k )
k =
(3-5)
T k =
(3-6)
7/24
Fig. 3-5.
8/24
Fig. 3-6.
9/24
(3-7)
Nyquist frequency
=
N
fN =
1
[ Hz ]
2T
=
[ rad / sec]
2 T
(3-8)
10/24
Reconstruction of continuous
signal from samples
x(nT )h(t nT )
(3-9)
n =
cT sin(c t )
c t
(3-10)
11/24
n =
x(nT )
cT sin(c (t nT ))
c (t nT )
(3-11)
Fig. 3-7.
12/24
c =
/T
Reconstructed signal with=
s /2
xr (t ) =
sin (t nT )
T
x(nT )
n =
T
Using convolution of sinc function
h(t ) =
(3-12)
(t nT )
sin( t / T )
t / T
(3-13)
Fig. 3-8.
13/24
Fig. 3-9.
(3-14)
14/24
Fig. 3-10.
15/24
Smoothing
filter
Fig. 3-11.
16/24
(3-15)
Fig. 3-12.
17/24
Fig. 3-13.
18/24
A/D converter
Converting the analog input signal into digital form
D/A converter
Converting processed signal back into analog form
Smoothing filter
Smoothing the reconstructed signal and removing unwanted high
frequency components
Anti-aliasing filter
A/D
Digital
D/A
Smoothing
(LPF)
converter
processors
converter
filter
Fig. 3-14.
19/24
Quantizer
Each signal sample is quantized into one of 2 B levels
Encoder
Encoding the discrete levels into distinct binary word each of length B bits
Lowpass
Quantizer
Encoder
filter
Fig. 3-15.
20/24
Quantization
Quantization error
e=
(n) x(n) x (n)
(3-16)
Quantization error
Fig. 3-16.
21/24
q/2
e2 = q 2 /12
Fig. 3-17.
(3-17)
22/24
Quantization interval
2Xm Xm
=
2B
2 B 1
=
q
(3-18)
(3-19)
Signal-to-noise ratio
x2
x2
=
=
SNR 10
log10 2 10 log10 2 B + 2 2
X
2
/12
m
e
(3-20)
3 22 B x2
Xm
= 10 log10
= 6.02 B + 4.77 20 log10
2
X
m
x
23/24
Example 3-1
1
,
p (e) = q
0,
q
q
e
2
2
otherwise
e2 = e 2 p(e) de
q
2
q
1 2
e de
q
q2
=
12
24/24