Error detection/correction and fault detection/recovery – Pulse or data error handling – Digital data error correction
Reexamination Certificate
2000-03-09
2003-01-07
Baker, Stephen M. (Department: 2133)
Error detection/correction and fault detection/recovery
Pulse or data error handling
Digital data error correction
C341S059000, C714S769000
Reexamination Certificate
active
06505320
ABSTRACT:
FIELD OF INVENTION
The present invention relates to communication systems, such as the recording and reproduction of binary data in disk storage systems for digital computers. In particular, the present invention relates to a commuted read/write channel employing a multiple-rate channel encoder/decoder (ENDEC) which increases storage capacity by reducing or eliminating pad bits required in prior art configurations.
BACKGROUND OF THE INVENTION
Disk storage system (e.g., magnetic and optical disk drives) are essentially communication systems wherein the storage medium, head and read/write electronics constitute the communication channel. User data are transmitted through a communication medium by recording the data to the disk during a write operation and sensing the recorded data during a read operation.
FIG. 1A
illustrates the typical format of a magnetic disk storage medium comprising a plurality of concentric, radially spaced data tracks
2
partitioned into a number of data sectors
4
. A typical format for a data sector
4
is shown in
FIG. 1B
as comprising a preamble
3
for acquiring the frequency and phase of the recorded data, a sync mark
5
for symbol synchronizing to the recorded data, channel encoded user data
7
, and appended ECC redundancy symbols
9
for detecting and correcting errors in the user data. The disk may also include embedded servo sectors
6
which facilitate positioning a head with respect to the disk during write and read operations. In order to achieve a more constant linear bit density, the tracks are typically banded together to form zones and the data rate is increased from the inner to outer diameter zones. This is illustrated in
FIG. 1A
wherein the disk is partitioned into an inner zone
10
comprising seven data sectors per track and an outer zone
8
comprising fourteen data sectors per track. In practice, the disk is actually partitioned into several zones with the data rate increasing incrementally from the inner to outer diameter zones.
The user data are typically encoded in order to increase the effective signal-to-noise ratio (SNR) which facilitates higher linear recording densities leading to an increase in storage capacity. Two types of codes are typically employed: a channel code and an error correction code (ECC). A channel code typically increases the effective SNR by attenuating a noise source within the recording channel. For example, a run-length limited (RLL) channel code constrains the minimum spacing between consecutive medium transitions representing the recorded data which reduces intersymbol interference. An RLL channel code may also constrain the maximum spacing between consecutive medium transitions in order to reduce errors in bit synchronizing to the data. These two run-length constraints on the minimum and maximum spacing between medium transitions are typically referred to as (d,k) respectively.
A channel code is typically augmented by an ECC code, such as the well known Reed-Solomon ECC code, which increases the effective SNR by encoding the data according to a minimum Hamming distance which defines the correction power of the ECC code. When noise corrupts a transmitted (recorded) codeword, the received codeword can still be successfully decoded as long as the erroneous bits do not violate the minimum Hamming distance.
A typical prior art configuration for the channel and ECC encoders/decoders is illustrated in FIG.
2
. The user data
12
received from the host are first encoded by an ECC encoder
14
which generates a number of ECC redundancy symbols
16
by dividing the user data represented as a data polynomial by a generator polynomial. The user data
12
are passed through a multiplexer
18
and encoded by a channel encoder
20
, such as a rate 16/17 RLL encoder
20
, and the encoded user data are written to the disk
22
by a write modulator
24
. The ECC redundancy symbols
16
are then passed through the multiplexer
18
, encoded by the channel encoder
20
, and written to the disk
22
by the write modulator
24
. During read back, a data detector
26
detects an estimated data sequence
28
from a read signal
30
emanating from a head (not shown) positioned over a selected track of the disk
22
. The estimated data sequence
28
is decoded by a channel decoder
32
, such as a 16/17 RLL decoder, which implements the inverse operation of the channel encoder
20
. The channel decoded data
34
is then decoded by an ECC decoder
36
into user data
38
transmitted to the host.
The code rate (input-bits/output-bits) of the channel ENDEC shown in the prior art read/write channel of
FIG. 2
is typically low in order to minimize the error propagation for the ECC code. For example, when using a rate 16/17 RLL ENDEC with a byte oriented ECC code, an error in RLL decoding
32
a first byte may propagate into a neighboring byte or bytes which must also be corrected by the ECC code. This error propagation problem increases relative to the number of user data bits encoded by the channel encoder, which directly affects the code rate of the channel ENDEC. Thus, the number of user data bits encoded is typically selected to be low in order to reduce error propagation which places an upper bound on the code rate. This is undesirable since a higher code rate allows more user data to be written to the disk.
The code rate limitation of the prior art read/write channel of
FIG. 2
is overcome by “commuting” the channel architecture such that the user data is first encoded by the channel encoder, and then encoded by the ECC encoder. This allows for a higher rate channel ENDEC since the error propagation problem is avoided by first ECC decoding the detected data during a read operation, and then passing the ECC corrected data through the channel decoder.
An example prior art commuted read/write channel is disclosed in the above referenced U.S. patent application entitled “DISK STORAGE SYSTEM EMPLOYING ERROR DETECTION AND CORRECTION OF CHANNEL CODED DATA, INTERPOLATED TIMING RECOVERY, AND RETROACTIVE/SPLIT-SEGMENT SYMBOL SYNCHRONIZATION.” The general configuration and operation of the commuted read/write channel disclosed in the aforementioned patent application is illustrated in FIG.
3
. The user data
12
received from the host is first encoded by a high rate channel encoder
40
, such as a high rate RLL encoder
40
. The channel encoded data
42
passes through a multiplexer
44
and is written to the disk
22
by the write modulator
24
. The channel encoded data
42
is simultaneously encoded by an ECC encoder
14
which generates the ECC redundancy symbols
16
over the encoded data. The ECC redundancy symbols
16
are then encoded by a low rate channel encoder
46
, such as a low rate RLL encoder
46
, so that the ECC redundancy symbols
16
satisfy the desired channel constraints when written to the disk
22
. The channel encoded ECC redundancy symbols
48
are then passed through multiplexer
44
and written to the disk
22
by the write modulator
24
.
During a read operation, the data detector
26
detects an estimated data sequence
28
from the read signal
30
. The estimated data sequence
28
representing the user data is passed through a multiplexer
50
and input into the ECC decoder
36
. The estimated data sequence
28
representing the encoded ECC redundancy symbols is then decoded by a low rate channel decoder
52
, such as a low rate RLL decoder
52
, which implements the inverse operation of the low rate channel encoder
46
. The decoded ECC redundancy symbols
54
are then passed through multiplexer
50
and input into the ECC decoder
36
which detects and corrects errors in the estimated data sequence
28
representing the user data. The corrected user data sequence
56
is then decoded by a high rate channel decoder
58
, such as a high rate RLL decoder
58
, which implements the inverse operation of the high rate channel encoder
40
. The decoded user data
38
is then transmitted to the host. The error propagation problem inherent in the prior art read/write channel of
FIG. 2
is avoided s
Turk Stephen A.
Vis Marvin L.
Zook Christopher P.
Baker Stephen M.
Cirrus Logic, Incorporated
Shifrin Dan
LandOfFree
Multiple-rate channel ENDEC in a commuted read/write channel... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multiple-rate channel ENDEC in a commuted read/write channel..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multiple-rate channel ENDEC in a commuted read/write channel... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3023541