GDC21D003 Datasheet, PDF (51/77 Pages) Hynix Semiconductor – VSB Receiver

6.4.3 Error Estimation
The Equalizer uses a Least Mean Square(LMS)
algorithm and can do adaptation on the transmitted
binary training sequences as well as on the
transmitted data. The LMS algorithm computes
6
4
2
-4
(a) 2VSB slice
-2
-6
(b) 4VSB slice
GDC21D003
how to adjust the filter taps in order to reduce the
current error at the output of the Equalizer, by
generating an estimate of the current error in the
equalizer output signal using the slice level of
Figure 6.4.4.
7
5
3
1
-1
-3
-5
-7
(c) 8VSB slice
15
13
11
9
7
5
3
1
-1
-3
-5
-7
-9
-11
-13
-15
(d) 16VSB slice
: slice level
Figure 6.4.4 VSB Slice Level
The estimated error is used in the adaptive filter to
update its coefficients as following equations.
One is for feed-forward adaptive filter and the
other is for feedback adaptive filter. The error is
multiplied by data and adjusted by step-size, and
then added to the stored coefficient value to get
new coefficient value.
Ck
j
I ^{+ 1}
C^k
j
+
=
1_:
Ck
j
New
+
¥Äe^k
Vk
j
j ^th filter tap
coefficient
of
equalizer
C
:^k
j
Current
j
th
filter
tap
coefficient
of
equalizer
¥Ä : ^{step -size}
e^k
Current error value of equalizer
:
Vk
j
:
Current
filter
data
stored
in
th
j
filter
tap
of
Feedforward
C I k +1
j
=
Ck
j
+ ¥Äe^k
I ^k
j
Ck +1
j
:
New
th
j
filter
tap
coefficient
of
equalizer
C
:^k
j
Current
th
j
filter
tap
coefficient
of
equalizer
¥Ä : ^{step -size}
e^k
Current error value of equalizer
:
I ^k
j
:
Current
filter
decision
data
stored
in
th
j
filter
tap
of
Feedback
52