Note: Descriptions are shown in the official language in which they were submitted.
CA 02456526 2004-O1-30
FAULT CHARACTERIZATION USING INFORMATION INDICATIVE OF ECHO
RELATED APPLICATIONS
[0001] The prE~se=~t application is a continuation-in-part (CIP)
of the following co-pending United States patent application:
Serial No. 10/317,H46, (Attorney Docket No. 100.491US01), filed
December 12, 2002, titled ~~FAtJLT CHARACTERIZATION USING
INFORMATION INDICATIVE OF ECHO."
TECHNICAL FIELD
[0002] The following description relates to telecommunications
in general and to digital subscriber line (DSL) devices in
particular.
BACKGROUND
[0003] Telecommunication service providers use a variety of
techniques to troubleshoot faults occurring in systems that use a
copper twisted-pair telephone line (also referred to here as a
"local loop"). Typically, a fault such as an open circuit or a
short circuit in a local loop is found using either a handheld
testset or by a mechanized metallic loop test (MLT) system. These
devices are normally able to measure the distance from the test
device to the loop fault. In some situations, this approach can
reduce the circuit restoral time as well as associated labor cost.
[0004] Although nearly all plain old telephone service (POTS)
circuits are attached to such an MLT system, a large number of
fielded high speed digital subscriber line (HDSL) circuits are
not. These HDSL circuvts include HDSL, HDSL2, and HDSL4 circuits,
Attorney Docket Nc_ 1:.:J.56'~JS(:~i 1
CA 02456526 2004-O1-30
which are also collectively referred to here as "HDSLx" circuits.
These HDSLx circuits typically carry critical DSl data and have
mean-time-to-restoral (MTTR) times that are typically subject to
service level agreements. Returning these HDSLx circuits to
service quickly after a fault is typically a high priority for
service providers.
SUMMARY
[0005] In one a?nbodiment, a method of analyzing an analyzed
fault associated with an analyzed communication medium includes
receiving a plurality of sets of first echo canceller coefficients
from an echo cancell.er. Each of the plurality of sets of first
echo canceller coefficients has an associated degree of
convergence. The method further includes selecting one of the
plurality of sets of first echo canceller coefficients based on
the degree of convergence associated with each of the plurality of
sets of first echo canceller coefficients. The method further
includes receiving an echo canceller profile. In such an
embodiment, the echo canceller profile includes at least one set
of profile echo canceller coefficients. The method further
includes correlating the selected set of first echo canceller
coefficients with at least one of the sets of profile echo
canceller coefficients included in the echo canceller profile. The
method further includes characterizing the analyzed fault based on
the correlation between the selected set of first echo canceller
coefficients and the at least one of the sets of profile echo
canceller coefficients included in the echo canceller profile.
[0006] In one embodiment, a method of analyzing an analyzed
fault associated with an analyzed communication medium includes
Attorney Docket Nc. lUU.561JS01 2
CA 02456526 2004-O1-30
receiving a plurality of second sets of information indicative of
echo and selectina one of the second sets of information
indicative of echo. The method further includes correlating a
first set of information indicative of echo associated with the
analyzed communication medium with the selected second set. The
method further incluaes characterizing the analyzed fault based on
the correlation between the first set and the selected second set.
[0007] In one embodiment, a method of analyzing an analyzed
fault associated with an analyzed communication medium includes
receiving an echo canc:eller profile. In such an embodiment, the
echo canceller profile includes a plurality of sets of profile
echo canceller coefficients. The method includes correlating a set
of first echo canceller coefficients received from an echo
canceller with at least a subset of the plurality of sets of
profile echo canceller coefficients included in the echo canceller
profile to create a correlation coefficient for each set of the
profile echo canceller coefficients included in the subset. The
method further includes characterizing the analyzed fault based on
the correlations between the set of first echo canceller
coefficients and the subset of the plurality of the sets of
profile echo canceller coefficients included in the echo canceller
profile. In such an embodiment, characterizing the analyzed fault
includes selecting the set of profile echo canceller coefficients
having a highest correlation coefficient. In such an embodiment,
characterizing the analyzed fault further includes when the
highest correlation coefficient is greater than a threshold
correlation value, providing a first attribute associated with the
selected set of profile echo canceller coefficients.
[0008] In one embodiment, a line interface unit that analyzes
an analyzed fault associated with an analyzed communication medium
Attorney ~o~~xe~ Nu. ~;J.~clJS~=. 3
CA 02456526 2004-O1-30
includes a first interface adapted to couple the line interface
unit to a first communication link. The line interface unit
further includes an echo canceller coupled to the first
communication link and a controller coupled to the echo canceller.
The controller is adapted to receive a plurality of sets of first
echo canceller coefficients from the echo canceller. In such an
embodiment, each of the sets of first echo canceller coefficients
has an associated degree of convergence. The controller is further
adapted to select one of the plurality of sets of first echo
canceller coefficiE:nts based on the degree of convergence
associated with each of the plurality of sets of first echo
canceller coefficients. The controller is further adapted to
receive an echo canceller profile. In such an embodiment, the echo
canceller profile includes at least one set of profile echo
canceller coefficients. The controller further adapted to
correlate the selected set of first echo canceller coefficients
with at least one of the sets of profile echo canceller
coefficients included in the echo canceller profile. The
controller is further adapted to characterize the fault based on
the correlation between the selected set of first echo canceller
coefficients and the at least one of the sets of profile echo
canceller coefficients included in the echo canceller profile.
[0009] In one. embodiment, a line interface unit that analyzes
an analyzed fault associated with an analyzed communication medium
includes a first interface adapted to couple the line interface
unit to a first communication link. The line interface unit
further includes an echo canceller coupled to the first
communication link and a controller coupled to the echo canceller.
In such an embodiment, the controller is adapted t: receive a
plurality of second sets of information indicative of echo and
select one of the second sets of information indicative of echo.
Attorney Docket Nc, lu~.~611JS01 4
CA 02456526 2004-O1-30
In such an embodirC:ent, the controller is further adapted to
correlate a first set of information indicative of echo associated
with the analyzed communication medium with the selected second
set. In such an embodiment, the controller is further adapted to
characterize the analyzed fault based on the correlation between
the first set and the selected second set.
[0010] In one embodiment, a line interface unit that analyzes
an analyzed fault associated with an analyzed communication medium
includes a first interface adapted to couple the line interface
unit to a first communication link. The line interface unit
further includes an echo canceller coupled to the first
communication link and a controller coupled to the echo canceller.
In such an embodiment, the controller is adapted to receive an
echo canceller profile. In such an embodiment, the echo canceller
profile includes a plurality of sets of profile echo canceller
coefficients. In such an embodiment, the controller is further
adapted to corr;~late a set of first echo canceller coefficients
received from an echo canceller with at least a subset of the
plurality of ser.s of profile echo canceller coefficients included
in the echo canceller profile to create a correlation coefficient
for each set of the profile echo canceller coefficients included
in the subset. In such an embodiment, the controller is further
adapted to characterize the analyzed fault based on the
correlations between the set of first echo canceller coefficients
and the subset of the plurality of the sets of profile echo
canceller coefficients included in the echo canceller profile. In
such an embodiment, the controller is further adapted, in order to
characterize the analyzed fault, to: select the set of profile
echo canceller -_:oerficients having the highest correlation
coefficient, an~~, when the highest correlation coefficient is
greater than a t:hrF~~shold correlation value, provide a first
Attorney Docket No. ~U0.561US01 5
CA 02456526 2004-O1-30
attribute associated with the selected set of profile echo
canceller coefficients.
[0011] In one embodiment, a telecommunication device that
analyzes an analyzed fault associated with an analyzed
communication medium includes an interface adapted to couple the
telecommunication device to a communication medium and an echo
canceller coupled to the interface. In such an embodiment, the
telecommunication device is adapted to receive a plurality of sets
of first. echo canceller coefficients from the echo canceller. In
such an embodiment, each of the sets of first echo canceller
coefficients has an associated degree of convergence. In such an
embodiment, the telecommunications device is further adapted to
select one of the plurality of sets of first echo canceller
coefficients based on the degree of convergence associated with
each of the plurality of sets of first echo canceller
coefficients. In such an embodiment, the telecommunications device
is further adapted to receive an echo canceller profile. In such
an embodiment, the echo canceller profile includes at least one
set of profile echo canceller coefficients. In such an embodiment,
the telecommunications device is further adapted to correlate the
selected set of first echo canceller coefficients with at least
one of the sets of profile echo canceller coefficients included in
the echo canceller profile. In such an embodiment, the
telecommunications device is further adapted to characterize the
fault based on the correlation between the selected set of first
echo canceller coefficients and the at least one of the sets of
profile echo canceller coefficients included in the echo canceller
profile,
[0012] In one embodiment, a telecommunication device that
analyzes an analyzed fault associated with an analyzed
Attorney Jockey No. 10~~.56~iJS(ll 6
CA 02456526 2004-O1-30
communication medium includes an interface adapted to couple the
telecommunication device to a communication medium and an echo
canceller coupled to the interface. In such an embodiment, the
telecomrnunicatiou device is adapted to receive a plurality of
second sets of information indicative of echo and select one of
the second sets of information indicative of echo. In such an
embodiment, the telecommunications device is further adapted to
correlate a first set of information indicative of echo associated
with the analyzed communication medium with the selected second
set, and characterize the analyzed fault based on the correlation
between the first set and the selected second set.
[0013] In one embodiment, a telecommunication device that
analyzes an analyzed fault associated with an analyzed
communication mediu;n includes an interface adapted to couple the
telecommunication .device to a communication medium and an echo
canceller coupled tc the interface. In such an embodiment, the
telecommunication device is adapted to receive an echo canceller
profile. In such a:z embodiment, the echo canceller profile
includes a plurality of sets of profile echo canceller
coefficients. In such an embodiment, the telecommunications device
is further adapted to correlate a set of first echo canceller
coefficients received from an echo canceller with at least a
subset of the plurality of sets of profile echo canceller
coefficients included in the echo canceller profile to create a
correlation coefficient for each set of the profile echo canceller
coefficients inclt;ded in the subser_. In such an embodiment, the
telecommunications device is further adapted to characterize the
analyzed fault based o:i the correlations between the set of first
echo canceller coefficients and the subset of the plurality of the
sets of profile echo canceller coefficients included in the echo
canceller profile. In such an embodiment, the telecommunication
Attorney Docket No. 100.561USOi 7
CA 02456526 2004-O1-30
device is further adapted, in order to characterize the analyzed
fault, to: select the set of profile echo canceller coefficients
having the highest correlation coefficient, and, when the highest
correlation coefficient is greater than a threshold correlation
value, provide a first attribute associated with the selected set
of profile echo canceller coefficients. The details of one or
more embodiments of the claimed invention are set forth in the
accompanying drawings and the description below. Other features
and advantages will become apparent from the description, the
drawings, and the claims.
DRAWINGS
[0014] FIG. 1 is a block diagram illustrating echo in a
telecommunication device coupled to a communication medium.
[0015] FIG. 2 is a flow diagram of one embodiment of a method
of analyzing a fault associated with a communication medium.
[0016] FIG. 3 is a block diagram of one embodiment of an HDSL2
line interface unit.
[0017] FIG. 4 is a chart showing exemplary echo responses of a
hybrid circuit rrom one implementation of an HDSL2 line interface
unit under various open-circuit faults.
[0018] FIG. 5 is a chart showing exemplary echo responses of a
hybrid circuit from one implementation of an HDSL2 line interface
unit under various short-circuit faults.
[0019] FIG. 6 is a chart showing exemplary correlation
coefficients between e~~ho responses generated by an echo canceller
from one embodiment of an HDSL2 line card with various echo
Attor:my GocKet No. _-._.'~6~U.:; 8
CA 02456526 2004-O1-30
canceller coefficients associated with faults located at various
distances from a hybrid circuit.
[0020] FIG. 7A is a flow diagram of an embodiment of a method
of receiving a second set of information indicative of echo.
[0021) FIG. 7B is a flow diagram of an embodiment of a method
of receiving a second set of information indicative of echo.
[0022] FIG. 8A is a flow diagram of an embodiment of a method
of receiving a first set of information indicative of echo.
[0023] FIG. 8B is a flow diagram of an embodiment of a method
of receivi:ig a first set of information indicative of echo.
[0024] FIG. 9 is a f'~ow diagram of one embodiment of a method
of characterizing a fault based on a correlation between first and
second sets oz information indicative of echo.
[0025] Like reference numbers and designations in the various
drawings indicate like elements.
DETAILED DESCRIPTION
[0026] FIG. 1 is a block diagram illustrating echo 100 in a
telecommunication device 102 coupled to a communication medium
104. Device 102 is coupled to a far-end device 106 via the
communication medium 104. Device 102 includes an interface 108
that couples the rest of the components of the device 102 to the
communication medium 104. Typically, there is echo 100 that
occurs in the device 102. Echo 100 is caused by impedance
mismatches between the device 102 and the communication medium
104. A fault, such as an open circuit or a short circuit,
Attorney Docket Nc. 100.561US01 9
CA 02456526 2004-O1-30
occurring in the communication medium 104 between the device 102
and the far-end device 106 can cause impedance mismatches and, as
a result, echo.
(0027] For example, in one embodiment illustrated with dashed
lines in FIG. 1, device 102 is a line interface unit. In such an
embodiment, the far-end device 106 includes a modem 132, and the
communication medium 104 includes a twisted-pair telephone line
134. The lisle interface unit is coupled to the modem 132 via the
twisted-pair telephone line 134.
[0028] The interface 108 includes a hybrid circuit 136. Hybrid
circuit 136 converts the two-line twisted-pair telephone line 134
into a 4-line connection that is coupled to a transceiver 150.
This 4-line connection includes a separate 2-line transmit path
138 and a receive path 140.
[0029] Hybrid circuit 136 is designed to isolate signals on the
transmit path 138 from signals on the receive path 140.
Theoretically, hybrid circuit 136 can achieve perfect isolation
under matched impedance r_onditions. In practice, however, perfect
impedance matching is unlikely and a portion of the signal
transmitted on the transmit path 138 reflects back along the
receive path 140. This reflection is included in echo 100 and is
referred to here as hybrid echo 160. In such an embodiment, other
reflections (referred to here as line echo 162? included in echo
100 may also result from normal loop conditions such as bridge
taps, wire gauge changes, and the like. The line echo 162 may
also result from faults occurring in the twisted-pair telephone
line 139.
[0030] Echo nay be acceptable and even desirable in POTS
service since echo allows a speaker to hear his or her own
Attorney 'r~ockeL Nc. 1C~(:.Se~US;;; i0
CA 02456526 2004-O1-30
attenuated voice through the ear piece. However, in some
situations, echo limits data transmission in HDSLx service, since
echo can create a high "noise floor," which can limit the signal-
to-noise ratio (SNR). In HDSLx applications, an echo canceller
142 is typically used to cancel at least a portion of the echo
100. In one embodiment, echo canceller 142 is implemented using
an adaptive filter that estimates the amount of echo 100 by
constantly updating a set of echo coefficients. The estimate of
the echo 100 is subtracted from the received signal from the
receive path 190. In such an implementation, the echo canceller
142 is typically implemented in an application-specific integrated
circuit (ASIC) .
[0031] FIG. 2 is a flow diagram of one embodiment of a method
200 of analyzing a fault associated with a communication medium.
The fault to be analyzed is referred to here as the "analyzed
fault." The communication medium that is analyzed is referred to
here as the "analyzed communication medium" or "analyzed medium."
The analyzed fault can be, for example, a short circuit or an open
circuit. Moreo~,rer, it is to be understood that analysis may
determine that there is no fault in the analyzed communication
medium. For example, this may be because there is no fault at all
or that any fault is located somewhere other than in the analyzed
communication medium. In one embodiment, the analyzed
communication medium is a twisted-pair telephone line coupled to
an ilDSLx line interface unit. In such an embodiment, a far-end
device aucr. as ..xn :jDSLx modem or a remote terminal need not be
coupled to the ether end of the telephone line in order to analyze
the analyzed fault. E_nbcdiments of method 200 are suitable, for
example, for use with a telecommunication device that employs an
echo canceller and communicates over one or more twisted-pair
telephone lines.
Attorney Docket Ne. 100.561US01 11
CA 02456526 2004-O1-30
[0032] Method 200, shown in FIG. 2, includes receiving a first
set of information indicative of echo associated with the analyzed
communication medium (block 202). In one embodiment, the first
set of information is a set of echo canceller coefficients from an
echo canceller coupled to the analyzed communication medium. For
example, in such an embodiment, the set of echo canceller
coefficients includes echo canceller coefficients currently used
by the echo canceller. The set o.f echo canceller coefficients are
received, in such an embodiment, from the echo canceller by a
controller. In other embodiments, the first set of information
(for example, a set of echo canceller coefficients) is retrieved
from a memory such as a RAM or ROM, or is otherwise generated or
calculated.
[0033] Method 200 also includes receiving a second set of
information indicative of echo (block 209). In one embodiment,
the second set of information is a set of echo canceller
coefficients corresponding to a known fault in the analyzed
communication mF~diurn or a communication medium similar to the
analyzed communication medium. For example, in one implementation
of such an embodiment, prior to normal operation, an echo
canceller coupled to a communication medium similar to the
analyzed communication medium is operated with the known fault in
the communication medium. Then, a set of echo canceller
coefficients resulting from operating the echo canceller with the
known fault is saved and stored in a memory such as ROM or RAM for
later retrieval. In such an embodiment, a controller or other
device retrieves the set of saved echo canceller coefficients from
the memory. In one implementation of such an embodiment, multiple
sets of echo canceller coefficients are generated under various
conditions (including, for example, various fault conditions and
various communicat?~on media). In other embodiments, the second
Atto~~ey .')ccket Nv. ,. .. ~E;1J_'~1 '2
CA 02456526 2004-O1-30
set of information (for example, a set of echo canceller
coefficients) is otherwise generated or calculated.
[0034] Method 200 also includes correlating the first set of
information with the second set of information (block 206). For
example, in one embodiment, a first set of echo canceller
coefficients are correlated with a second set of echo canceller
coefficients. In one implementation of such an embodiment, a
normalized covariance correlation is performed in order to obtain
a correlation coefficient between the first set of echo canceller
coefficients and the second set of echo canceller coefficients.
Other ways of correlating the first and second sets of echo
canceller coefficients can be used, for example, using higher
order statistics and/or neural networks.
[0035] Then, the analyzed fault is characterized based on the
correlation between the first set of information and the second
set of information (block 208). For example, in one embodiment
(shown in FIG. 2 with dashed lines), a first and second set of
echo canceller coefficients are correlated and a resulting
correlation coefficient is compared to a threshold coefficient
value (block 252). If the correlation coefficient is greater than
(or greater than or equal to) the threshold coefficient value, a
determination is made that the analyzed fault has an attribute
associated with the second set of echo canceller coefficients
(block 254). I:v implementations where a known fault is associated
with the second ser_ of echo canceller coefficients, a
determination i~ made that the analyzed fault has an attribute of
the known fault associated with the second set of echo canceller
coefficients. For example, in one case, a determination is made
that the analyzed fault. is of the same type (for example, an open
Attorney Docket No. 10~.561USC-~ 13
CA 02456526 2004-O1-30
circuit or short circuit) as the known fault associated with the
second set of echo canceller coefficients.
[0036] In implementations of such an embodiment, if the
correlation coei~ic~_E~;~t is less than (or less than or equal to)
the threshold coefficients value, a determination is made that
there is not a fault in the communication having an attribute
associated with the second set of echo canceller coefficients. In
other implementations, no determination is made if the correlation
coefficient is less than the threshold coefficients value.
[0037] In another embodiment, a known fault having a known
location within (or otherwise in relation to) to the analyzed
medium (or a rnedimn similar to the analyzed medium) is associated
with the second set of echo canceller coefficients. In such an
embodiment, if a correlation. coefficient between a first set of
echo canceller coefficients and the second set of echo canceller
coefficients is greater than (or greater than or equal to) a
threshold coefficient value, a determination is made that the
analyzed fault is of the same type as the known fault and is at
the same location within the analyzed medium as the known fault.
The second set of echo canceller coefficients, in one
implementation of such an embodiment, includes several groups of
echo canceller coefficients associated with multiple known faults
located at various locations within the analyzed medium (or a
medium similar to the analyzed medium). Each group of echo
canceller coefficients is associated with one of the multiple
known faults, each of which is located at one of the multiple
locations within the analyzed medium (or a medium similar to the
analyzed medium). When an analyzed fault is characterized, the
first set of echo canceller coefficients is correlated with one or
more of the groups of echo canceller coefficients included in the
Atrorney ~ocxst No. 1,;~~.~~1'JSJ1 14
CA 02456526 2004-O1-30
second set of echo canceller coefficients. If a correlation
coefficient resulting from correlating the first set of echo
canceller coefficients with a particular group is greater than a
threshold correlation value, then a determination is made that the
analyzed fault is at the location associated with the particular
group.
[003$] In one implementation of such an embodiment, this
location data is used (for example, by a service provider) to
identify whether the analyzed fault is within a central office or
outside of the central office in the outside plant. In such an
implementation, locations that are within the central office and
locations that are outside of the central office in the outside
plant are identified. If, as result of performing an embodiment
of method 200, an analyzed fault is identified as having a
location that is associated with the central office, a
determination is made that the analyzed fault is within the
central office. If an analyzed fault is identified as having a
location that is associated with the outside plant, a
determination is rr~ade that the analyzed fault is in the outside
plant.
[0039] Whether the analyzed fault is within the central office
or outside of the central office in the outside plant can provide
a useful data point for service providers. A service provider can
use this data point to assist in the dispatch of repair
technicians. Io so;ne situations, this has value because central
office technicians and outside plant technicians have different
skill levels, work rules, and the like. For example, if a loop
fault is identified as being in the central office, then an
expensive outside plant service dispatch can be avoided.
Alternatively, if a loop fault is identified as being in the
Attorney :ecket No. 1;:'~.'~6i'JSO:~ 15
CA 02456526 2004-O1-30
outside plant, then an outside plant technician can be immediately
dispatched to minimize MTTR.
[0040] In other embodiments, the analyzed fault is
characterized in other ways. Although method 200 is depicted in
FIG. 2 with the elements of method 200 occurring in a particular
order, it is to be understood that the elements of method 200 can
occur in a different order or certain elements can occur in
parallel.
[0041] Method 200 allows a fault associated with a
communication medium to be analyzed using information indicative
of echo associated with the communication medium. In those
embodiments where such information is used for other purposes
(such as echo cancellation), resources and costs associated with
providing such a fault analysis capability can be reduced by using
existing functionality to obtain such information,
[0042] Moreover, embodiments of method 200 that make use of an
echo canceller can be operated in a single-ended manner that does
not require a far-end device to be coupled to the communication
medium in order for the fault to be characterized. Such
embodiments can allow a service provider (such as a provider of
HDSLx service) to characterize a fault without having to access
equipment located at a customer's premise. This can reduce costs
and delays associated with sending a technician to a customer's
premise in order to access such equipment.
[0043] FIG. 3 is a block diagram of one embodiment of an HDSLx
line interface unit 300 (also referred to here as a "line card"
300). In one embodiment, the line card 300 is used to implement
the methods and apparatus described here. Line card 300 is used
to send and receive DSl traffic over an HDSLx communication link
Attorney ::oc:c~t Nc. l:_:_;.SE,i:J~:~~~ 16
CA 02456526 2004-O1-30
using at least one twisted-pair telephone line 340 (also referred
to here as a "local loop" or "loop"). For example, in one
embodiment, the line card 300 is an HDSL2 line interface unit that
is used to send and receive DS1 traffic over an HDSL2 link using a
single twisted-pair telephone line.
[0044] The line card 300 includes an upstream interface 302 and
a downstream interface 304. Upstream interface 302 and downstream
interface 309 couple the line card 300 to an upstream link and a
downstream link, respectively. In the embodiment shown in FIG. 3,
the upstream link is a DSX-1 link that is cross-connected to a
time division-multiplexing network. The upstream interface 302
couples the line card 300 to the DSX-1 link and includes, for
example, a T1 framer 308 and a DSX-1 pre-equalizer 310. In the
embodiment shown in FIG. 3, the downstream link is an HDSLx link
such as an HDSL, HDSL2, or HDSL9 link. The downstream interface
304 couples the line card 300 to the HDSLx link. The HDSLx link
is implemented using the twisted-pair telephone line 340. The
downstream interface 304 includes, for example, an HDSLx framer
312, an HDSLx transceiver 314, an echo canceller 316, and a hybrid
circuit 318.
[0045] The line card 300 includes a power supply 320 for
providing power to the various components of the line card 300.
Also, in the embodiment shown in FIG. 3, the power supply 320
includes a current sensor 321. Current sensor 321, in one
implementation, is used to perform a current test by applying a
predetermined voltage to the telephone line 340 and measuring the
resulting current. The line card 300 also includes a controller
322. For example, in the embodiment shown in FIG. 3, the
controller 322 includes a programmable processor 324 (such as a
microprocessor) and a memory 326. Memory 326 includes both read-
Attorney Docket ~Io. i0C.56'~USO1 17
CA 02456526 2004-O1-30
only memory ("ROM") 328 and random access memory ("RAM") 330.
Although memory 326 is shown in FIG. 3 as having a separate ROM
328 and RAM 330, other memory configurations can be used, for
example, using scratchpad memory included in the programmable
processor 324.
[0046] Line card 300 also includes a craft interface 332.
Craft interface 332 includes, for example, a universal
asynchronous receiver-transmitter ("UART") that couples an RS-232
serial port to the controller 322. A user can connect a portable
computer or other data terminal to the serial port and communicate
with an embedded control program executing on the programmable
processor 324. Alternatively, the user can communicate with the
embedded control program over an embedded operations channel
carried among the DS1 traffic handled by the line card 300.
[0047] The hybrid circuit 318 converts a 2-wire, full-duplex
twisted-pair telephone line 340 into a separate 2-wire transmit
path 394 and a separate 2-wire receive path 346. Echo canceller
316 is used to cancel at least a portion of any echo. In one
embodiment, echo canceller 316 is implemented using an adaptive
filter that estimates the amount of echo by constantly updating a
set of echo coefficients. The estimate of the echo is subtracted
from the received signal from the receive path 346. In such an
embodiment, the echo canceller 316 is implemented in an ASIC.
[0048] In operation, the line card 300 receives DSl traffic
from the downstream link on the downstream interface 304. The
incoming DS1 traffic is formatted as HDSL frames. The downstream
interface 304 processes the incoming frames and communicates the
DS1 traffic to the upstream interface 302. The upstream interface
302 formats the DS1 traffic into T1 frames and transmits the
frames out on the' upstream link. A similar process occurs in
Attorney Docket No. lOC.561USGl 18
CA 02456526 2004-O1-30
reverse for DSl traffic received on the upstream interface 302
from the upstream link. The incoming DS1 traffic is formatted as
T1 frames. The upstream interface 302 processes the incoming
frames and communicates the DS1 traffic to the downstream
interface 304. The downstream interface 304 formats the DS1
traffic into ;-1DSL Frames and transmits the frames out on the
downstream link. Although FIG. 3 depicts an HDSLx line interface
unit, other telE-ec:ommunications devices can be used to implement
the techniques described here. For example, G.SHDSL or
asynchronous digital subscriber line (ADSL) devices can be used.
[0049] Fmbodi.ments of method 200 can be implemented using line
card 300. In one such embodiment, the analyzed communication
medium is the telephone line 340 (referred to here as the
"analyzed line"), which is coupled to the line card 300. In this
embodiment, an HDSL2 modem or other remote device need not be
coupled to the other end of the analyzed line in order to analyze
an analyzed fault.
[0050] In this embodiment, the second set of information
indicative of echo includes an echo canceller profile. The echo
canceller profile includes at least one set of echo canceller
coefficients (referred to here as the "profile echo canceller
coefficients"). In one implementation of this embodiment, the
profile echo canceller coefficients are generated by operating the
same line card 300 used to analyze the analyzed fault while the
line card 300 is coupled to the analyzed line. In other
implementations, the profile echo canceller coefficients are
generated using other techniques. For example, in one such other
implementation, the profile echo canceller coefficients are
generated by operating the same line card 300 used to analyze the
analyzed fault while the line card 300 is coupled to a telephone
Attorney pocket No. l;J~.561,JSQ_ 19
CA 02456526 2004-O1-30
line other than the analyzed line. In other implementations, the
profile echo canceller coefficients are generated by operating a
line card other than the line card 300 used to analyze the
analyzed fault. For example, in one such implementation, a line
card of the same type as the line card 300 used to analyze the
analyze fault is used.
[0051] In one implementation of this embodiment, the set of
profile echo canceller coefficients is generated as a part of the
manufacturing process. The echo canceller of a line card (for
example, the line card 300 used to analyze the analyzed fault or a
line card similar thereto) is operated with one or more known
fault conditions in the telephone line to which the line interface
is coupled (for example, the analyzed line or another telephone
line). The echo canceller coefficients for the known fault
conditions are saved and included in the echo canceller profile.
During operation, the saved echo canceller profile is retrieved
and used. In other implementations, a similar procedure is used
to generate the echo canceller profile during installation of the
line card 300.
[0052] For example, in one implementation of this embodiment,
the set of profile echo canceller coefficients are generated by
operating the echo canceller while the echo canceller is coupled
to a twisted-pair telephone line that is configured to model the
expected operating environment of the line card. In one example,
an echo cancelier is coupled to a telephone line having sections
with different attributes. For example, in one such
implementation, an canceller is coupled to a telephone line having
a 1000 foot section that uses 24 AWG wire and an 8000 foot section
that uses 26 AWG wire. Examples of other line configurations
include using a single-gauge telephone line (far example, 26 AWG
AtLOrney ~'oc,tet No. 1'~,::.~67;1~: 20
CA 02456526 2004-O1-30
wire). Then, multiple sets of profile echo canceller coefficients
are generated under various fault conditions. In other examples,
echo canceller coefficients generated using multiple line
configurations are included in the echo canceller profile.
Moreover, in other implementations, the echo canceller
coefficients included in the echo canceller profile are
calculated, for example, by running simulations (or using other
mathematical te.-:h~iiau~~s) that model operation of the echo
canceller wader Various fault conditions and various line
configuration: .
[0053] In other .implementations of this embodiment, the echo
canceller profile is created in other ways. For example, the echo
canceller profile is created by calculating a set of profile echo
canceller coefficients in one such other implementation. In
implementations of this embodiment, the echo canceller profile is
periodically updated (for example, by recalculating a set of
profile echo canceller coefficients) and saved during operation of
the line card 30C. The echo canceller profile is updated, for
example, to reflect changes in analyzed line or other operating
conditions.
[0054] In this embodiment, the first set of information
indicative of e~~ho associated with the analyzed line includes a
set of echo canc~el'er ~voefiicients generated by the echo canceller
316 of the line card 300. The set of echo canceller coefficients
generated by tht e;:'no canceller 316 in this embodiment is referred
to here as the "set of first echo canceller coefficients." In one
implementation, the set of first echo canceller coefficients
includes echo canceller coefficients generated by the echo
canceller 316 during operation of the line card 300 in a fault
test mode. The set ef first echo canceller coefficients is
Attorney ~ockrt ?~~~. _",~.~S:JSC~ 22
CA 02456526 2004-O1-30
received, in such au implementation, by the programmable processor
324 from the echo canceller 316. In other implementations, the
set of first echo canceller coefficients includes echo canceller
coefficients generated by the echo canceller 316 during the normal
operating mode of the line card 300.
[0055] FIGS. 4 and 5 illustrate the creation of one example of
an echo canceller profile. FIG. 4 is a chart 400 showing
exemplary echo responses of a hybrid circuit 318 from one
implementation of an HDSL2 line interface unit 300 coupled to a
telephone line 340 with open-circuit faults located at various
distances from the hybrid circuit 318. Lines 402, 404, and 406 are
echo responses of the hybrid circuit 318 when the telephone line
340 to which the HDSL2 line interface unit 300 is connected has an
open-circuit fault located 0 feet, 500 feet, and 1000 feet,
respectively, from the hybrid circuit 318. FIG. 5 is a chart 500
showing exemplary echo responses of a hybrid circuit 318 from one
implementation of an HDSL2 line interface unit 300 coupled to a
telephone line 340 with short-circuit faults located at various
distances from the hybrid circuit 318. Lines 502, 504, and 506
are echo responses of the hybrid circuit 318 when the telephone
line 340 to which the HDSL2 line interface unit 300 is connected
has a short-circuit fault located 0 feet, 500 feet, and 1000 feet,
respectively, from the hybrid circuit 318. Echo coefficients
corresponding to each of the echo responses shown in FIGS. 4 and 5
are determined by the echo canceller 316 included in the HDSL2
line interface unit 300. The resulting echo coefficients are
included in the echo canceller profile.
(0056] In this embodiment, the first set of information
indicative of echo ass«ciated with the analyzed line and the
second set of information indicative of echo associated with the
Attorney Docket Cdo. 100.561US0~ 22
CA 02456526 2004-O1-30
analyzed line are correlated by correlating the set of first echo
canceller coefficients with at least one of the sets of profile
echo canceller coefficients included in the echo canceller
profile. For example, in one embodiment, a normalized covariance
correlation is perforrned. For each normalized covariance
correlation that is performed, a correlation coefficient is
generated. Other ways of correlating the set of first echo
canceller coefficients with at least one of the sets of profile
echo canceller coefficients included in the echo canceller profile
can be used. In orie implementation, the programmable processor
324 of the controller 322 is programmed to perform the
correlation. In other implementations, a computer or other device
is coupled to the line card 300 (for example, via the craft
interface 332 or via an embedded operations channel). The
computer or other device then performs the correlation.
[0057] Then, the analyzed fault is characterized based on the
correlation between the set of first echo canceller coefficients
and at least one of the sets of profile echo canceller
coefficients inclu de a in the echo canceller profile. For example,
in one implementat.icon, it is determined if there is a high
correlation between r_h~ set of first echo canceller coefficients
and at least one of the sets of profile echo canceller
coefficients included in the echo canceller profile. If there is,
a determination is made that the analyzed fault has an attribute
associated with the set of profile echo canceller coefficients
with which the set of first echo canceller coefficients is highly
correlated. Moreover, if there is not a high correlation between
the set of first echo canceller coefficients and at least one of
the sets of echo canceller coefficients included in the echo
canceller profile, a determination is made that the analyzed fault
does not have an attribute associated with the sets of profile
At ~cr:~ey uocket vc . =~'. '~ r.l'JS:: ~ 23
CA 02456526 2004-O1-30
echo canceller coe'ficients included in the echo canceller
profile. It is to be understood, however, that in other
implementations, if there is not a high correlation between the
set of first echo canceller coefficients and at least one of the
sets of profile echo canceller coefficients included in the echo
canceller profile, no determination is made and additional
techniques are used to analyze the analyzed fault, if desired.
[0058] The determination as to whether there is a high
correlation between the set of first echo canceller coefficients
and at least one of the sets of profile echo canceller
coefficients included in the echo canceller profile, in one
implementation, is made by comparing a correlation coefficient to
a threshold correlation value. FIG. 6 is a chart showing
exemplary correlation coefficients between echo responses
generated by an echo canceller 316 from one embodiment of an HDSL2
line card 300 with various echo canceller coefficients associated
with faults located at various distances from a hybrid circuit
318. Line 602 shows the correlation coefficients resulting from
correlating a 0 feet, open-circuit echo response (line 402 in FIG.
4) with the 0 feet (line 402 in FIG. 4), 100 feet (not shown in
FIG. 4), 200 feet (not shown in FIG, 4), 300 feet (not shown in
FIG. 9), 400 feet (not shown in FIG. 4), 500 feet (line 404 in
FIG. 4), 600 feet (not shown in FIG. 4), 700 feet (not shown in
FIG. 9), 800 feet (not shown in FIG. 9), 900 feet (not shown in
FIG. 4), and 1000 feet (line 406 in FIG. 4) open-circuit echo
responses. Line 604 snows the correlation coefficients resulting
from correlating the 0 feet, short-circuit echo response (line 502
in FIG. 5) and the 0 feet (line 502 in FIG. 5>, 100 feet (not
shown in FIG. 5), 200 feet (not shown in FIG. 5), 300 feet (not
shown in FIG. 5), 400 'eet (not shown in FIG. 5), 500 feet (line
504 in FIG. 5), 000 feet (not shown in FIG. 5), 700 feet (not
Attorney Docket No. 10U.561US01 24
CA 02456526 2004-O1-30
shown in FIG. 5), 800 feet (not shown in FIG. 5), 900 feet (not
shown in FIG. 5), and 1000 feet (line 506 in FIG. 5) short-circuit
echo responses. Line 606 shows the cross-correlation coefficients
resulting from correlating the 0 feet, open-circuit echo response
(line 402 in FIG. 4) and the 0 feet (line 502 in FIG. 5), 100 feet
(not shown in FIG. '_~), 200 feet (not shown in FIG. 5), 300 feet
(not show:i in FI,~. 5), 400 feet (not shown in FIG. 5), 500 feet
(line 504 in fIG. 5), 600 feet (not shown in FIG. 5), 700 feet
(not shown in FIG. 5), 800 feet (not shown in FIG. 5), 900 feet
(not shown in FIG. 5), and 1000 feet (line 506 in FIG. 5) short-
circuit echo responses. Line 608 shows the cross-correlation
coefficients resulting from correlating the 0 feet, short-circuit
echo response (line 502 in FIG. 5) and the 0 feet (line 402 in
FIG. 4), 100 feet (not shown in FIG. 4), 200 feet (not shown in
FIG. 9), 300 feet (not shown in FIG. 4), 400 feet (not shown in
FIG. 4), 500 feet (line 404 in FIG. 9), 600 feet (not shown in
FIG. 4), 700 feet (not shown in FIG. 4), 800 feet (not shown in
FIG. 4) , 900 ~eEet (riot shown in FIG. 4) , and 1000 feet (line 406
in FIG. 4) open-circuir_ echo responses.
[0059] Based on FIG. 6, a threshold correlation value of around
0.9 can be used to determine if a first set of echo canceller
coefficients is highly correlated with echo canceller coefficients
associated with line 902 from FIG. 4 (that is, echo canceller
coefficients associated with an open-circuit fault located 0 feet
from a hybrid circuit 318). Also, a threshold correlation value
of around 0.9 can be used to determine if a first set of echo
canceller coefficients is highly correlated with echo canceller
coefficients associated with line 502 from FIG. 5 (that is, echo
canceller coefficients associated with an short-circuit fault
located 0 feet from a hybrid circuit 318).
Attorney :~cck:eL ~r;.. .G;;. ;~'1'.7~:;:. 25
CA 02456526 2004-O1-30
[0060] In one implementation of this embodiment, the
programmable processor 324 is programmed in a suitable manner to
carry out the processing of method 200. The programmable
processor 324 is programmed by storing appropriate program
instructions in memory 326. The program instructions are retrieved
from memory 326 and executed on programmable processor 324. The
program instruc~.vons are operable to cause the programmable
processor 324 to carry out the processing of method 200. In other
implementations, a computer or other device is coupled to the line
card 300 (for example, via the craft interface 332 or via an
embedded operations channel). The computer or other device then
performs all or a portion of the processing of method 200.
[0061] FIG. ~'F. is a flow diagram of one embodiment of a method
700 of receiving a first set of information indicative of echo.
This embodiment is implemented using an echo canceller coupled to
the analyzed corrimunication medium. Method 700 includes receiving
multiple sets of echo canceller coefficients from an echo
canceller coupled to the analyzed communication medium (block
702). In this embodiment of method 700, the convergence process
of the echo canceller is performed multiple times with the
analyzed fault existing in the analyzed communication medium.
Each time the convergence process is performed a different set of
echo canceller coefficients is generated. Each set of echo
canceller coefficients includes information indicative of the
degree of convergence achieved by the echo canceller in generating
that set. of echo canceller coefficients. In one implementation of
this embodiment, each set of echo canceller coefficients received
from the echo cari~_eller includes a received level that indicates
the degree of convergence that was reached for that set of echo
canceller coefficients. In one such implementation, the received
level has a value ranging, for example, from a value of 127 to a
Attorney ;locket No. iG~.5~ilUSC's 26
CA 02456526 2004-O1-30
value of 1000 or higher, where a lower value represents a higher
degree of convergence. In one such implementation, the process is
repeated three times to generate three sets of echo canceller
coefficients, ear_h set. having an associated received level value.
[0062] Method 70U also includes selecting at least one of the
multiple sets of echo canceller coefficients based on the degree
of convergence achieved by each of the multiple sets of echo
canceller coefficients (block 704). The selected set of echo
canceller coefficients is then used as the first set of
information indicative of echo in subsequent processing, for
example, as described above in connection with method 200. In one
implementation shown in F'IG. 7 using dashed lines, selecting at
least one of the. multiple sets of echo canceller coefficients
based on the degree of convergence achieved by each of the
multiple sets of echo canceller coefficients includes determining
for which set of e~.ho canceller coefficients the echo canceller
achieved the highest degree of convergence (block 706) and the set
of echo cancel.lc>r coefficients having the highest degree of
convergence is selc,cted (block 708) and is used for subsequent
processing as the firsr_ set of information indicative of echo, for
example, as described above in connection with embodiments of
method 200.
[0063] FIG. 7B is a flow diagram of one embodiment of a method
750 of receiving a first set of information indicative of echo.
In the embodiment of method 750 shown in FIG. 7B, selecting at
least one of the multiple sets of echo canceller coefficients
based ora the degrees of convergence includes additional processing
to determine if the set of echo canceller coefficients that has
the highest degree of ~~onvergence is suitable for use in
characterizing the analyzed fault. For example, it may be the
Attorney Docket :VC. ~ ~... 'i6i'.1S0: '%'
CA 02456526 2004-O1-30
case that the echo canceller, in some situations, is unable to
achieve a degree of convergence that is desirable for reliable
fault characterization. Such additional processing is intended to
identify such situations. Moreover, it may be the case that even
though a high degree ef convergence is not achieved, the resulting
echo canceller ,,~.oefficients may nevertheless be suitable for use
in the fault characterization process. For example, in one
implementation :~'_ such an embodiment, if the set of echo canceller
coefficients having the highest degree of convergence and the set
of echo cancellc~r coefficients having the second highest degree of
convergence are highly correlated, the set of echo canceller
coefficients having the highest degree of convergence is likely to
be suitable for fault characterization.
[0064] Method 750 includes receiving multiple sets of echo
canceller coefficients from an echo canceller coupled to the
analyzed medium (block 752) and determining for which set of echo
canceller coefficients the echo canceller achieved the highest
degree of convergence (block 754) as described above in connection
with method 700. If the highest degree of convergence achieved by
the echo canceller is greater than (or greater than or equal to) a
convergence threshold value (checked in block 756), then the set
of echo ca.°~celler ::oefficients having the highest degree of
convergence i:~ selected (block 758) and is used as the first set
of information lr:dicative of echo for subsequent processing as
described above.
[0065] If the highest degree of convergence achieved by the
echo canceller is less than (or less than or equal to) the
convergence threshold value, then it is determined for which set
of echo canceller coefficients the echo canceller achieved the
second highest degree of convergence (block 760). Then, the set
Attorney Docket tJr~. l;it,~.~olLJS'il 28
CA 02456526 2004-O1-30
of echo canceller coefficients having the highest degree of
convergence and the set of echo canceller coefficients having the
second highest degree of convergence are correlated (block 762).
For example, in one implementation, this correlation is performed
in the same way in which the first and second sets of information
indicative of echo are correlated (for example, by executing the
same or similar software or firmware routines). If the
correlation between the set of echo canceller coefficients having
the highest degree of convergence and the set of echo canceller
coefficients having the second highest degree of convergence is
greater than (or greater than or equal to) a threshold correlation
value (checked in block 764), then the set of echo canceller
coefficients having the highest degree of convergence is selected
(block '764) and is used as the first set of information indicative
of echo for subseauent processing as described above. For
example, in one implementation, the threshold correlation value is
the same as the threshold correlation value used in embodiments of
method 200. If the correlation between the set of echo canceller
coefficients having the highest degree of convergence and the set
of echo canceller coefficients having the second highest degree of
convergence is less than (or less than or equal to) the threshold
correlation value, then additional echo canceller coefficients are
obtained and method 700 is repeated (looping back to block 702).
It is to be understood, however, that in other embodiments other
processing occurs in the event that the correlation is not greater
than the threshold correlation value, for example, generating an
alarm or other error indication.
[0066] By rec~,iving multiple sets of echo canceller
coefficients and selecting the set having the highest degree of
convergence for usw as the first set of information indicative of
echo, the accuracy of the fault characterization process can be
Attorney Docket No, i0~.561USCi 29
CA 02456526 2004-O1-30
improved. Also, the accuracy of the fault characterization
process, in some embodiments, can be improved by performing
additional processing to determine if the set of echo canceller
coefficients having the highest degree of convergence is suitable
for use in characterizing the fault. Such embodiments may be used
where the echo c:anceller frequently does not achieve a degree of
convergence that: i.s desirable for reliable fault characterization.
[0067] fIG. 8 is a slow diagram of one embodiment of a method
800 of receiving a second set of information indicative of echo.
Embodiments of method 800 are suitable For use with embodiments of
methods 200, 700, 7~0, and 900. Method 800 includes receiving
multiple second sets of information indicative of echo (block 802)
and selecting a second set of information indicative of echo from
the multiple second sets (block 804). The selected second set of
information indicative of echo is used for subsequent processing
(for example, as described above in connection with embodiments of
method 200).
[0068] In one embodiment of method 800, each of the multiple
sets of informatie:Z is associated with one type of fault. The
second set of i_Wormatio:n indicative of echo is selected, in such
an embodiment, based on the fault type. For example, method 850
is one exemplary embodiment of method 800 and is shown in FIG. 8B.
Implementations of method 850 are suitable for use with
embodiments of tree line interface unit 300 of FIG. 3. At least
two echo canceller profiles are used in method 850. One echo
canceller profile includes multiple sets of profile echo canceller
coefficients, where each set of profile echo canceller
coefficients is associated with a first type of fault (for
example, a short circuit) located at one of multiple locations in
a telephone line (for example, every 100 feet). The other echo
Attorney Docket Vo. LOC.5E1L;SC1 3C
CA 02456526 2004-O1-30
canceller profile i:~cluaes multiple sets of profile echo canceller
coefficients, where each set of echo canceller coefficients is
associated with a seconc, fault type (for example, an open circuit)
located at one oWnult:iple locations in a telephone line (for
example, every 100 feet). The echo canceller profiles in such an
embodiment are generated, for example, in one of the ways
described above.
[0069] Method 850 includes receiving multiple sets of echo
canceller profiles (block 852). Method 850 further includes
determining if the analyzed fault is a short circuit or an open
circuit fault (block 854). In one such implementation, this
determination i5 made by performing a current test using
functionality irucluded in the power supply of a line interface
unit. For example, in the embodiment of line interface unit 300
shown in FIB. 3, power supply 321 includes a current sensor 321.
A currerut test is performed by applying a predetermined voltage
(for example, 180 volts) to the analyzed line and measuring the
current in the ar~alyzecx line using the current sensor 321. If the
measured current is less than a first threshold current value, an
undercurrent condition exists and the analyzed fault is considered
to be an open circuit fault. If the measured current is greater
than a second threshold current value, an overcurrent condition
exists and the analyzed fault is considered to be a short circuit
fault.
[0070] Method 85~' also includes selecting an echo canceller
profile based o~~ wher_her the analyzed fault is a short circuit
fault or a:~ ope~~ -_ircuit fault (block 856) . For example, in one
implementation, if it is determined that the analyzed fault is a
short-circuit fault, then the echo canceller profile associated
with the short-circ~uir. fault type is selected. If it is
Attorney Docket No. 100.561US01 ~1
CA 02456526 2004-O1-30
determined that the analyzed fault is an open-circuit fault, then
the echo canceller profile associated with the open-circuit fault
type is selected. The selected echo canceller profile is then
used as the second set of information indicative of echo for
subsequent processing as described above.
[0071] In embodiments of methods 800 and 850, the number of
correlations between the first set of information indicative of
echo and various items included in the second set of information
indicative echo tr_~at are performed can be reduced. For example,
in embodiments where two echo canceller profiles are used, one of
the echo canceller profiles is used for subsequent correlation
processing (that is, one or more of the sets of profile echo
canceller coefficients included in the selected echo canceller
profile is correlated with a first set of echo canceller
coefficients), and no correlation processing need be performed on
the non-selected echo c:anceller profile. This can reduce the
amount of resources needed to perform the fault characterization
processing and/or. reduce the amount of time needed to perform such
processing. Moreover, this can improve the accuracy with which
the analyzed is characterized. For example, in those situations
where it is more difficult to distinguish between open circuit and
short. circuit conditions at certain fault locations (for example,
where the analyzed fault is relatively far from the hybrid circuit
or other telecomrn~~aications devir_e), the accuracy with which an
analyzed fault is characterized can be improved with such
embodiments of methods 800 and 850.
[0072] FIG. 9 is a flow diagram of one embodiment of a method
900 of characterizing a fault based on a correlation between first
and second sets of information indicative of echo. Method 900 is
used in embodiments of method 200 where the first set of
Attor~:ey Docket No. .:.:.'.~61;JS:': 32
CA 02456526 2004-O1-30
information indlcat.ive of echo includes a first set of echo
canceller coefficients and the second set of information
indicative of echo includes multiple second sets of echo canceller
coefficients. In such embodiments, the first set of echo
canceller coefficients is correlated with each of the multiple
second sets of echo canceller coefficients to generate a
correlation coefficient for that second set of echo canceller
coefficients.
[0073] Method 900 includes selecting the second set of echo
canceller coefficients having the highest correlation with the
first set of echo canceller coefficients (block 902). For
example, in one impleme:~station, this is done by selecting the
second set of echo canceller coefficients having the highest
correlatio:~ coerficienn. with the first set of echo canceller
coefficients. N.et~iod 900 also includes determining if the
correlation coefzicient associated with selected second set is
greater than (or greater than or equal to) a threshold correlation
value (block 904). If the correlation coefficient associated with
the selected second set is greater than (or greater than or equal
to) the threshold correlation value, method 900 provides one or
more attributes associated with the selected second set of echo
canceller coefficients along with the correlation coefficient
(block 906). For example, in one implementation implemented using
an embodiment of the ?-iDSL. lire interface unit 300 of FIG. 3, the
one or more attributes associated with the selected second set and
the correlation c=oefficient for the selected second set are
communicated to a terminal attached to the craft port of the line
interface unit c:r t.o a management system executing on a management
card coupled to r.he line interface unit, for example, via a system
backplane. In such an implementation, the one or more attributes
that are communi:,ated include the type of fault (for example, a
Attorruey Docket No. i00.561US01 33
CA 02456526 2004-O1-30
short circuit or an open circuit) and the location of the fault
(for example, 50C feet from a hybrid circuit coupled to the
analyzed line) associated with the selected second set, along with
the correlation coefficient for the selected second set.
[0074] If the correlation coefficient associated with the
selected second set of echo canceller coefficients is less than
(or less than or eaual to) the threshold correlation value, method
900 selects the N second sets of echo canceller coefficients
having the N highest correlations with the first set of echo
canceller coefficients, where N is greater than one (block 908).
One or more attributes associated with the N selected second sets,
along with the correlation coefficients associated with the
selected second sets, are provided (block 910). In one
implementation of such an embodiment, N is equal to four (4) and
the four second sets of echo canceller coefficients having the
four highest correlation coefficients are selected. In one
implementation of such an embodiment using an embodiment of the
HDSL line interface unit 300 of FIG. 3, the type and location of
the known fault associated with each of the four selected second
are provided along with the correlation coefficients for each of
the selE.~cted second sets. Such an approach allows the system to
which this information is provided to perform additional fault
analysis. For example, in one implementation, the additional
fault analysis includes selecting the attributes associated with
one of the N selected second sets based on additional data not
available to the device used to implement method 900.
[0075] The methods and techniques described here may be
implemented in digital electronic circuitry, or with a
programmable pro;,essor (for example, a special-purpose processor
or a general-purpose processor such as a computer) firmware,
Attorney Dor_het Nc. _0:;.~FM JSfj;: 34
CA 02456526 2004-O1-30
software, or in combinations of them. Apparatus embodying these
techniques may include appropriate input and output devices, a
programmable pry>cessor, and a storage medium tangibly embodying
program instructions for execution by the programmable processor.
A process embodying these techniques may be performed by a
programmable processor executing a program of instructions to
perform desired functions by operating on input data and
generating appropriate output. The techniques may advantageously
be implemented in one or more programs that are executable on a
programmable system including at least one programmable processor
coupled to receive data and instructions from, and to transmit
data and instructio::s to, a data storage system, at least one
input, device, a:ud at lE:ast one output device. Generally, a
processor will rc~~r;~ive instructions and data from a read-only
memory and/or a random access memory. Storage devices suitable
for tangibly embodying computer program instructions and data
include all forms of non-volatile memory, including by way of
example semiconductor memory devices, such as EPROM, EEPROM, and
flash memory devices; magnetic disks such as internal hard disks
and removable disks; magneto-optical disks; and DVD disks. Any of
the foregoing may be supplemented by, or incorporated in,
specially-designed application-specific integrated circuits
(ASICs) .
[0076] A number of embodiments of the invention defined by the
following claims have been described. Nevertheless, it will be
understood that various :modifications to the described embodiments
may be made withc7ur_ departing from the spirit and scope of the
claimed inventic:u. Accordingly, other embodiments are within the
scope of the fol:owing claims.
Attorney Docket No. 1~~0,561L1S01 35
