Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
I,t. ~3~9~
METHOD AND APPARATUS FOR DECODING MANCHESTER ENCODED DATA
This invention relates to a method and apparatus for decoding
Manchester (also known as split phase) encoded data.
It is well known to use Manchester encoding of binary data for
example for transmission of the data. In Manchester encoding, a data
'l' is represented by the two-bit word 10, and a data 'O' is
represented by the opposite two-bit word 01. The transmitted bit
rate is thus twice the data rate. Advantages of Manchester encoding
include a high signal transition dens1ty (changes between 'O' and '1'
bits) which facilitates clock recovery, a null d.c. component in the
transmitted signal spectrum, and the ability to detect data errors as
sequence violations (e.g. the two-bit words 00 and 11 represent errors
rather than valid data).
In decoding Manchester encoded data, it is necessary for the
decoder to be synchronized to the two-bit word boundaries. For
example, a series of data 'l's is encoded as a bit sequence
...10101010...; if the decoder is out of phase with the ward
boundaries, this will be incorrectly interpreted as a series of data
'O's, i.e. the bit sequence ...0101010.... In the event that the
decoder is in phase with the word boundaries, transmission errors may
occur in the Manchester encoded bit sequence, such errors appearing to
the decoder as sequence violations which can cause an erraneous phase
slip. In this event the decoder operates out of phase with the word
bnundaries until further sequence violations produce a subsequent
phase slip.
Such erroneous phase slips can lead to significant problems,
especially in continuous (as distinct from packetized) digital
transmission systems in which they may cause loss of frame
synchronization~
Accordingly3 an object of this invention is to provide an
improved method and apparatus for decoding Manchester encoded data, in
which erroneous phase slips due to bit errors are reduced or
substantially aYoided.
According to one aspect this invention provides a method of
decoding data represented by a bit stream in which lO and 01 bit
sequences represent respective data bits, comprising the steps of:
1 3 ~
producing a clock signal with a period corresponding to the duration
of each data bit; producing decoded data from the bit stream in
dependence upon the clock signal; detecting in the bit stream
consecutive bits having the same binary value; in response to such
detection, advancing a state machine towards a first state or towards
a second state in dependence upon the phase of the clock signal, the
first and second states representing respectively synchronized and
out-of-synchronism phases of the clock signal with respect to the data
bit sequences, the state machine having at least one intermediate
state between the first and second states; and in response to the
state machine having the second state, changing the phase of the clock
signal and advancing the state machine towards the first state.
In general, the state machine has m~1 states, where m is an
integer equal to 2 or more. In response to each pair of consecutive
bits in the bit sequence having the same binary value, which bits in
the absence o~ errors occur on opposite sides of a word (data bit)
boundary (i.e. they are the second bit representing a data bit of a
first value and the first bit representing a following data bit of the
opposite value), the state machine is advanced towards the first
state if the clock signal phase is such that these consecutive bits
cross a word boundary (a phase assertion as described below~, and is
advanced towards the second state if the clock signal phase ;s such
that these consecutive bits are part of a single word (a phase error
as described below).
Preferably the state machine comprises a counter and the step
of advancing the state machine towards the first and second states
comprises increasing and decreas~ng a count of the counter.
Conveniently the step of advancing the state machine towards the first
state in response to the state machine reaching the second state
comprises changing the state machine to the first state after changing
the phase of the clock signal.
According to another aspect this invention provides a decoder
for decoding data represented by a bit stream in which 10 and 01 bit
sequences represent respective data bits, comprising: first means for
producing a clock signal with a period corresponding to the duration
of each data bit; second means for producing decoded data from the bit
stream in dependence upon the clock signal; a state machine having a
13~3~
first state representing a synchronized phase of the clock signal with
respect to the data bit sequences7 a second state, and at least one
intermediate state between the first and second states; third means
responsive to the occurrence in the bit stream of two consecutive bits
having the same binary value for changing the state of the state
machine in a direction towards or away from the first state in
dependence upon the phase of the clock signal; and means responsive to
the state machine having the second state for changing the phase of
the clock signal and for changing the state of the state machine in a
direction towards the first state.
Preferably the state machine comprises an up/down csunter and
the first, intermediate, and second states of the state machine
comprise first, intermediate, and second counts respectively of the
counter. Conveniently the means for changing the state of the state
machine in response to the state machine having the second state
comprises means for resetting the counter from the second count to the
first count.
In an embodiment of the invention described in detail below,
the first means comprises a flip-flop responsive to the state machine
not having the second state for producing the clock signal with a
frequency which is half the bit rate of the bit stream, the second
means comprises a flip-flop responsive to the clock signal and the bit
sequence for producing the decoded data by sampling of the bit
sequence once for each period of the clock signal, and the third means
comprises means for delaying the bit sequence by one bit duration and
Exclusive-OR gating means responsive to the bit sequence and the
delayed bit sequence for enabling a change in the state of the state
machine.
The invention will be further understood from the following
description with reference to the accompanying drawings, in which:
Fig. 1 is a generalized state diagram represent;ng the
operation of a decoder in accordance with an embodiment of this
invention;
Fig. 2 schematically illustrates a circuit diagram of such a
decoder; and
Fig. 3 is a timing diagram illustrating signals which can
occur ;n operat;on of the decoder of Fig. 2.
- ` ~ 3~
Referring to Fig. 1, a Manchester encoded data decoder in
accordance with an embodiment of this invention can have any one of
m+1 states, three of which are represented by circles labelled
STATE 0, STATE m-1, and STATE m, m being an integer equal to 2 or
more. Thus in its simplest form the decoder can have three stat~s 0,
19 and 2, and more generally the decoder can have any desired number
of states 0, 1, 2, ...m-1, m.
Transitions between states of the decoder occur in response to
phase errors and phase assertions, as discussed further below. More
part kularly, in any of the states 0 to m-1 the occurrence of a phase
error produces a transition to the ne~t higher state 1 to m
respectiYely. In any of the states 1 to m-1 ~and optionally also m,
as shown by a broken line 10 in Fig. 1 from STATE m to STATE m-1) the
occurrence of a phase assertion produces a transition to the next
lower state 0 to m-2 (and optionally m-1) respectively. In the state
0, the occurrence of a phase assertion maintains the deccder ;n this
state, as shown by a line 12. From STATE m (in the optional case in
the absence of a phase assertion) the decoder automatically effects a
phase slip, shifting by one bit its recognition of the two-bit word
boundaries in the encoded data, and reverts to a lower state as shown
by a line 14 in Fig. 1. ~n Fig. 1 and in the embodiment of the
decoder described below this lower state is STATE 0, but it could
instead be an intermediate one of the states 1 to m-1 td~sirably in
the lower part of this range).
In a Manchester encoded bit stream, in the absence of bit and
phase errors two consecutive occurrences of the same bit (1 or 0)
only occur across a two-bit word boundary. ~n decoders in accordance
with embodiments of this inventian~ a phase error, causing a
transit~on to the next higher state as discussed above, is deemed to
occur if both bits of a two-bit word have the same binary value. A
phase assertion as discussed above occurs if there is a two-bit word
boundary between two consecutive occurrences of the same bit value in
the encoded data. This will become more clear from the description
below with reference to Figs. 2 and 3 of the drawings.
In the Manchester encoded data decoder of Fig. 2, an up/down
counter 20, which can have a count of 0, 1, 2 ... m-1, m, constitutes
a state machine with states corresponding to those of Fig. 1 as
~3 ~
described above. From STATE m there is no return to STATE m-1, i.e.
there is no line 10 as in Fig. 1, but there is only a phase slip (line
14) with a return to STATE 0. It should be appreciated that the
counter operation may be readily modified to modify these parameters
as desired. In addition, in this decoder the counter 20 counts up or
down by one in response to a phase error or assertion respectively,
consistent with the state diagram of Fig. 1, but it should be
appreciated that this also may be modified as desired. For example,
the counter 20 could be modified to increase its count by 2 or more in
response to each phase error, and to decrease its count by one in
response to each phase assertion. The manner in wh;ch such
modifications may be implemented is well within the knowledge of those
of ordinary skill in the art, given the description herein.
In addition to the up/down counter 20, the decoder of Fig. 2
comprises four D-type ~lip-flops 22, 24, 26, and 28, a gate 30, and an
inverter 32. Each of the flip-flops has a data input D, a clock input
C, and a non-inverting output Q; the flip-flop 28 also has an
;nvert;ng output -Q. In addit;on, each of the flip-flops 26 and 2~
has a hold input H; a log;c 1 at the ;nput H ma;nta;ns the flip-flop
in its current state, regardless of the si~nals at its inputs C and D.
Each of these fl;p-flops can, for examplle, compr;se a conventional D-
type flip-flop w;th a two-input data sellector, controlled by the
input H, for select;ng as a data ;nput e;ther the data ;nput as shown
in Fig. 2 (H=O) or the Q output of the flip-flop (H=1).
The gate 30 is a two-input Exclusive-OR gate with an ;nverting
output, connected to a count enable input E of the counter 20 to
enable the counter to count up or down when both inputs o~ the gate 30
have the same binary value. The counter 20 also has a clock input C,
an up/down control ;nput U/-D wh;ch controls the count direction (up
with a logic 1, down with a logic O, at this input), and an output M
at which the counter produces a logic 1 when it has its maximum count
m, and otherwise produces a logic 0.
A Manchester encoded data stream is supplied via a receive
data line 34 to the input D of the flip-flop 22, and a rece;ve clock
signal, recovered in known manner from the receive data, is supplied
via a line 36 to the clock inputs C of all of the flip-flops 22, 24,
26, and 28, and via the inverter 32 to the clock input C of the
~ 3 ~
counter 20. The output Q of the flip-flop 22 is connected to one
input of the gate 30 and to the input D of the flip-flop 24, whose
output Q is connected to the other input of the gate 30. Thus
recei~e data bits on the line 34 are clocked successively throu0h the
flip-flops 22 and 24, and the gate 30 enables the counter 20 whenever
two consecutive bits of the receive data have the same binary value,
i.e. whenever there is a phase error or a phase assertion.
The flip-flop 28 has its input D connected to its output -Q,
and its input H connected to the counter output M, so that unless the
counter 20 has its max;mum count m this flip-flop divides the
frequency of the receive clock by two to produce at its output Q7 and
hence on a line 38 connected thereto, an output clock signal at the
decoded data rate. ~his output Q of the flip-flop 28 is also
connected to the input U/-D of the counter 20, to control the count
direction and distinguish between phase errors and phase assertions.
When the counter 20 has its maximum count m, its output M is a logic 1
to ;nhibit toggling of the flip-flop 28 for one period of the receive
clock on the line 36, thereby to effect a phase slip. From this
maximum count m, the counter 20 is internally synchronously cleared,
t~ a count of 0, by the next falling edge of the receive clock.
The flip-flop 26 has its input D connected to the output Q of
the flip-flop 2~, its input H connected to the output -Q of the flip-
flop 28, and its output Q connected to an output line 40 at which it
produces the decoded data.
The operat~on of the decoder will be further understood from
the t;m;ng d;agram ;n Fig. 3 wh;ch illustrates, from top to bottom,
the receive clock on the line 36, a Manchester encoded receive data
stream on the line 34, the data represented thereby, the Q output
signals of the flip-flops 22 and ~4, the output of the gate 30
const;tuting the counter enable input sigr,al, arrows representing the
timing and direction of state transitions (arrows up indicating phase
errors, arrows down ;ndicating phase assertions), the count or value
of the counter 20, the output M of the counter 20, the output clock on
the line 38, the decoded data on the line 40, the data represented
thereby, and times which are referred to below.
It is assumed in Fig. 3 that the decoder is initially out-of-
synchronism with the word boundaries of the Manchester encoded receive
7 ~ 3 ~
data, and that the counter 20 has an initial value or count of m-2.
It is further assumed that the receive data signal, shown as a solid
line, contains errors in three bits marked by asterisks, where the
receive data signal should properly follow the broken line indicated
at these bit positions. The last of these bit errors, because it
occurs in the first bit of a two-bit Manchester encoded word, produces
a corresponding error in the decoded data, also shown by an asterisk.
As already indicated above, the counter is enabled in
response to each pair of consecutive bits of the receive data which
have the same value. Thus for example the receive data bits
immediately before and after a time tl are both 0, resulting hal~ a
receive clock period later in the counter enable signal becoming 1 ~or
one receive clock period. In the middle of this clock per;od, at a
time t2, with the falling edge of the receive clock the counter 20 is
clocked to counk in a direction determined by the state of the output
clock at this time. At this time the output clock is 1, representing
(erroneously because the detector is out of synchronism with the
receive data word boundaries) a phase error, so that the counter 20
counts up, as indicated by the upwardly directed state transition
arrow at the time t2, from the counter value m-2 to m-1.
Similarly, tha receive data bits immediately before and after
a time t3 are both 1, resulting one receive clock period later at a
time t4, when the output clock is 1 again representing a phase error,
in the counter 20 being incremented to its maximum count m.
Consequently, at the time t4 the counter ou~put M becomes 1. One
receive clock period later, at a time t5, with a falling edge of the
receive clock the counter is synchronously cleared to a count o~ 0,
a~d the counter output M again becomes 0. Between the times t4 and t5
the counter output M inhibits toggling of the flip-flop 28, so that
the output clock phase is reversed, corresponding to a phase sl;p as
represented by the line 14 in Fig. 1. As-a result, the decoder is
synchronized to the encoded data word boundaries, so that (except for
bit errors) the data is subsequently correctly decoded.
At times t6, t7, tlO, and tl3, in response to correctly
interpreted word boundaries on both sides of which the encoded data
bits have the same binary value, the counter is enabled while the
output clock is O representing a phase assertion. At these times the
~ 3 ~
counter value is 0 so that it can not be further decremented and
remains 0, corresponding to the path represented by the line 12 in
Fig. 1.
At times t8, tll, and tl4, in each case in consequence of a
b;t error ;n the receive data as discussed above and shown by an
asterisk, the counter is enabled when the output clock is 1,
representing an error, and the counter is therefore incremented to a
count value of 1. In each case shortly afterwards, at times t9, tl2,
and tl5 respectively, again in conse~uence of the bit error in the
receive data the counter is again enabled when the output clock is 0,
representing a phase assertion and returning the counter to the count
value of 0. Thus single bit errors do not result in phase slips of
the decoder, in contrast to the initial phase error and consequent
phase slip at about the time t5.
In fact, an error density in the receive data stream of nearly
one in four bits is necessary to produce an erroneous phase slip of
the decoder, or to prevent the decoder from becoming synchronized to
the Manchester encoded word boundaries.
As can be seen from Fig. 3, the decoded data is produced by
the flip-flop 26 effectively sampling the delayed receive data bit
sequence at the output Q of the flip-flop 24 once during each period
or cycle of the output clock on the line 38, with the rising edge of
the receive clock when the output clock is a logic 1. 3ecause this
sampling is effectively a sampling of the first bit in each two-bit
data word, of the three bit errors shown by the asterisks in the
receive data only the third bit error, occurring in a first bit
position, produces a corresponding error in the decoded data. Apart
from this error, the data is correctly decoded after the time t5 when
phase synchronism is established. As will be appreciated by those of
ordinary skill in the art, the counter value itself may be used to
provide an indication of the reliability of the decoded data, and if
desired of the bit error rate.
Although a particular embod;ment of the invention has been
described in detail, it should be appreciated that numerous
modifications, variations, and adaptations may be made thereto within
the scope of the invention as defined in the claims. In particular,
in addition to the specific alternatives and options which have
9 ~
already been mentioned it should be appreciated that the state machine
constituted by the up/down counter 20 could instead be constituted by
a shift register, accumulator, or other integrating or summing means,
with appropriate modif;cation of the associated decoder circuitry.
