Title: Method and apparatus for determining properties of a transmission channel
Abstract: The invention concerns a time domain reflectometry method for estimating properties of a transmission channel, for instance a channel for transmitting electric or acoustic signals. This method includes the steps of generating, at one end of the channel, a plurality of pulses (40, 42, 44) covering differency frequency bands, and processing the echoes provided by these pulses at the same end of the channel. The frequency bands of the generated pulses are preferably overlapping. The invention is particularly applicable to the testing of ADSL services.
Patent Number: 6,865,256 Issued on 03/08/2005 to Descamps, et al.
| Inventors:
|
Descamps; Luc Fran.cedilla.ois (Gerpinnes, BE);
Doligez; Thierry Christian Marie (Boissy l'Aillerie, FR);
Duvaut; Patrick (Pointoise, FR)
|
| Assignee:
|
Alcatel (Paris, FR)
|
| Appl. No.:
|
740939 |
| Filed:
|
December 21, 2000 |
Foreign Application Priority Data
| Current U.S. Class: |
379/1.04; 379/1.01; 379/22.02; 379/24; 379/27.03; 379/21; 324/533 |
| Intern'l Class: |
H04M 001//24; H04M 003//08; H04M 003//22; 29.03; 30 |
| Field of Search: |
379/1.01,1.04,21,22,22.01,22.03,22.08,24,27.01,27.02,27.03,27.08,29.01,29.02
324/527,532,533,534
370/249
375/224
|
References Cited [Referenced By]
U.S. Patent Documents
| 5128619 | Jul., 1992 | Bjork et al. | 324/533.
|
| 5649304 | Jul., 1997 | Cabot | 455/67.
|
| 5751149 | May., 1998 | Oberg et al.
| |
| 6026145 | Feb., 2000 | Bauer et al. | 379/26.
|
| 6177801 | Jan., 2001 | Chong | 324/520.
|
| 6215855 | Apr., 2001 | Schneider | 379/22.
|
| 6266395 | Jul., 2001 | Liu et al. | 379/27.
|
| 6292468 | Sep., 2001 | Sanderson | 370/241.
|
| 6292539 | Sep., 2001 | Eichen et al. | 379/1.
|
| 6349130 | Feb., 2002 | Posthuma et al. | 379/1.
|
| 6366644 | Apr., 2002 | Sisk et al. | 379/1.
|
| 6385297 | May., 2002 | Faulkner et al. | 379/1.
|
| 6466649 | Oct., 2002 | Walance et al. | 379/22.
|
| Foreign Patent Documents |
| 0 926 841 | Jun., 1999 | EP.
| |
| 551818 | Mar., 1943 | GB.
| |
| WO 01/24482 | Apr., 2001 | WO | .
|
Primary Examiner: Tran; Quoc
Attorney, Agent or Firm: Sughrue Mion, PLLC
Claims
What is claimed is:
1. A time domain reflectometry method for determining properties of a
transmission channel, comprising:
generating, at one end of the channel, a plurality of pulses covering
different frequency bands through time domain reflectometry, and
processing, as received signals, the echoes provided by the plurality of
pulses at said one end of the channel;
wherein the generating is performed so that the frequency bands of adjacent
ones of the plurality of pulses overlap.
2. A method according to claim 1, wherein the overlapping of the frequency
bands is such that, after reflection and said processing, the frequency
spectrum of the plurality of pulses is practically flat.
3. A method according to claim 1, further comprising:
providing each of the plurality of pulses with a given amplification or
attenuation, and
providing pulses of the received signals with the corresponding attenuation
or amplification.
4. A method according to claim 1, further comprising subjecting the
received signals to a synchronous averaging.
5. A method according to claim 1, further comprising subjecting the
received signals to a matched filtering.
6. A method according to claim 1, further comprising suppressing noise, in
medium and high frequency pulses of the received signals, by
estimating the noise for the part of the received signal after the channel
end echo, and
determining a threshold above which the received signals are taken into
consideration.
7. A method according to claim 1, wherein the processing of the received
signals is performed so that the received signals are processed in their
own frequency bands, and then added.
8. A method according to claim 7, further comprising detecting the
variation with time of one or more of:
the modulus of the received signals, and
the frequency of the received signals.
9. A method according to claim 1, wherein the generating of the plurality
of pulses is performed so as to generate complex analytical pulses.
10. A method according to claim 1, further comprising selecting the
frequency bandwidth and the amplitude of low frequency pulses of the
received signals according to the channel attenuation and its compliancy
in terms of egress.
11. A method according to claim 1, wherein the plurality of pulses are
generated sequentially or simultaneously.
12. A method according to claim 1, wherein at least one of said properties
being determined comprises the locations of defects of the channel.
13. A method according to claim 1, wherein
said transmission channel comprises a telephone line between a central
office and a subscriber, and
the processing of the received signals is performed at the central office.
14. An apparatus for testing the properties of transmission channels
between a central office and a subscriber, comprising a time domain
reflectometry test circuit, wherein said time reflectometry test circuit
comprises:
a pulse generator generating a plurality of pulses, at one end of the line,
covering different frequency bands through time domain reflectometry, and
an echo processor processing the echoes provided by these pulses at the
same end of the channel;
wherein the different frequency bands of adjacent ones of the pulses are
overlapping.
15. An apparatus according to claim 14, wherein the different frequency
bands are overlapping.
16. An apparatus according to claim 15, wherein said echo processor
processes the reflected pulses such that the frequency spectrum is
practically flat after reflection and processing.
17. An apparatus according to claim 14, wherein the pulse generator
includes amplification or attenuation for each generated pulse, and said
apparatus includes complementary attenuation or amplification for each
received pulse.
18. An apparatus according to claim 14, further comprising a synchronous
averager for the received signals.
19. An apparatus according to claim 14, further comprising a matched filter
for the received signals.
20. An apparatus according to claim 14, further comprising amplification or
attenuation for each generated pulse and complementary attenuation or
amplification for each received pulse.
21. An apparatus according to claim 14, further comprising a processor
processing the received signals for each frequency band and an adder
adding the processed signals.
22. An apparatus according to claim 21, further comprising a detector
detecting the modulus of the received signals and/or the variation with
time of the frequency of the received signals.
23. An apparatus according to claim 14, further comprising a receiver
receiving complex analytical pulses.
24. An apparatus according to claim 14, further comprising a selector
selecting the frequency bandwidth and the amplitude of the low frequency
pulses according to the line attenuation and its compliancy in terms of
egress.
25. An apparatus according to claim 14 wherein said pulse generator
generates the pulses sequentially or simultaneously.
26. An apparatus according to claim 14, wherein said transmission channels
are telephone lines comprising copper pairs between a central office and a
subscriber.
Description
BACKGROUND OF THE INVENTION
The invention relates to a method and apparatus for determining properties
of a transmission line or channel, for instance a channel for transmitting
electric or acoustic signals. It concerns also a time domain reflectometry
method and an equipment for implementing this method.
In the field of telecommunications, the density of transmitted information
increases regularly. This increase is not always compatible with the
existing equipment, more particularly with the existing transmission
lines. In order to cope with this problem, several technologies are known,
for instance ADSL which means "Asymmetric Digital Subscriber Line" (or
other DSL services such as HDSL and VDSL). This technology provides the
possibility to transmit, with ordinary telephone lines, high data rate
from the network (a central office) to the subscriber and lower data rates
from the subscriber to the network.
ADSL is adapted for distances, between a central station or office and the
subscriber, which are comprised between 1.5 and 6 km. HDSL is adapted for
distances greater than 6 km, and VDSL is adapted for distances comprised
between 0.3 and 1.5 km.
In order to be able to comply with ADSL services, the telephone line must
satisfy standards of quality which are not necessarily fulfilled by all
telephone lines.
SUMMARY OF THE INVENTION
Before implementing ADSL services, it is therefore necessary to evaluate a
priori the line quality at the lowest cost as possible and with the
highest accuracy as possible. The present invention relates to a method
and an apparatus which determine, from a central office, the transmission
characteristics of a transmission line with a high signal to noise ratio
and a high accuracy.
The invention thus contributes to the line qualification.
Up to now an a priori line qualification requires to make tests at both
ends of the line. This means that one operator must be present at the
central office and another operator at the subscriber's location. Although
accurate, this a priori line testing is expensive.
The method according to the invention is characterized in that, for testing
a line, use is made of time domain reflectometry wherein, at one end of
the line, a plurality of pulses covering different bandwidths are
transmitted in the line and the corresponding echoes are processed, more
particularly filtered and reconstructed, at the same end, the echoes
providing information about the whole length of the line.
Up to now, time domain reflectometry has been used only for detecting the
first defect on a line. Time domain reflectometry is known to use only one
pulse of narrow or wide frequency band at one time and the transmitted
pulse provides only one echo. With the wide band "multifocal" (multi
echoes) time domain reflectometry according to the invention, several
echoes are provided which give information about the entire length of the
line. In fact, with the invention, which makes use of pulses at different
bandwidths, it is possible to optimize both the time accuracy and, with an
appropriate filtering, the signal to noise ratio along the whole line, at
any distance from the central office.
The plurality of pulses, called here "multifocality", allows to partially
compensate for limited dynamics of the A/D (analog to digital) converters.
In fact, with only one wide band pulse, it would be necessary to use a
pulse of very high amplitude in comparison to the A/D converter ability.
In a preferred embodiment, the pulses are analytical complex pulses and
contribute to a high signal to noise ratio.
Preferably the bandwidths of the transmitted pulses are overlapping and are
such that the reconstructed echoes present a flat spectrum, i.e. a perfect
line. In other words, without defects and without attenuation, the
response has practically a constant amplitude over the whole useful
frequency spectrum. In this way, the method provides directly the
reflectometric impulse response of a line, for instance of a copper pair,
generally a twisted pair.
As the pulses cover different bandwidths, in order to obtain this flat
spectrum, it is necessary that, after reflection and processing, the
modulus of the sum of the complex responses of the overlapping regions
equals the modulus (for instance one) in the non-overlapping region.
In a preferred embodiment, the spectrum amplitude on one side of a
band-width is decreasing according to a sine function and the overlapping
part of the neighboring bandwidth is increasing according to a cosine
function and the detected signal is the sum of squares of all components.
With this embodiment, after reflection, the first end of the first
bandwidth will vary as sin.sup.2 and the overlapping adjoining beginning
of the following bandwidth will vary as a cosin.sup.2. Therefore, in the
overlapping region: sin.sup.2 +cos.sup.2 =1.
With this embodiment, it is also possible to provide a given gain (or
attenuation) to each bandwidth in order that, for each of these
bandwidths, the pulse be compatible with the power limitations imposed by
standards. For instance, ADSL requests that the power be limited to -40
dbm/hertz (10.sup.-4 mwatt/hertz) and VDSL requests a limitation to -60
dbm/hertz (10.sup.-6 mwatt/hertz) for frequencies higher than 1 MHz.
Therefore, in order to obtain a flat spectrum, for each bandwidth, the
transmitted pulse has a given amplification (or attenuation) and the
received echo is provided with a complementary attenuation (or
amplification).
The echoes provide in the time domain, and/or in the frequency domain, an
information about the properties, generally the defects, of the line.
Moreover, the positions in time of the echoes represent the locations of
the defects.
In an embodiment, the frequency bandwidth and the amplitude of the low
frequency pulses are selected according to the line attenuation and its
compliancy in terms of egress. An egress compliant line is a line which
does not disturb neighbouring services. In other word means are provided
for selecting the frequency bandwidth and the amplitude of the low
frequency pulses.
The time domain reflectometry method according to the invention is not
limited to the estimation of the attenuation of telephone lines. More
generally, this method may be used for estimating transmission channels
which are dispersive, and subject to attenuation and noise.
These channels are not necessarily channels for transmitting electric
signals; the signals may be of a different nature, for instance acoustic
signals. The method provides the reflectometric impulse response with a
good accuracy and a good signal-to-noise ratio at any distance.
In brief, the invention concerns a time domain reflectometry method for
estimating properties of a transmission channel, for instance a channel
for transmitting electric or acoustic signals, which is characterized in
that it comprises the steps of generating, at one end of the channel, a
plurality of pulses covering different frequency bands, and of processing
the echoes provided by these pulses at the same end of the channel. The
invention therefore determines a priori the reflectometric impulse
response of this transmission line since it is to be noted that a spectrum
over a wide bandwidth is equivalent to a Dirac function.
The invention relates also to a method for testing the properties, such as
the attenuation, of telephone lines comprising copper pairs, for instance
twisted pairs, between a central office and a subscriber, which is
characterized in that it makes use of the time domain reflectometry.
The invention provides an apparatus for testing the properties, such as the
attenuation, of telephone lines comprising copper pairs, for instance
twisted pairs, between a central office and a subscriber, which is
characterized in that it comprises time domain reflectometry means.
BRIEF DESCRIPTION OF THE DRAWINGS
Other features and advantages of the invention will appear with the
following description of certain of its embodiments, this description
being made in connection with the following drawings, wherein:
FIG. 1 represents the application of a method according to the invention to
the estimation of defects of a telephone line for ADSL services,
FIG. 2 is a schematic diagram showing an equipment according to the
invention, this diagram representing also the operation of the equipment,
i.e. the method according to the invention,
FIGS. 3a and 3b show an example of pulses generated by the equipment
represented on FIG. 2, and
FIGS. 4a through 4f, FIGS. 5a through 5f, FIGS. 6a through 6f, and FIGS. 7
and 8 are diagrams showing signals at different locations on the receiving
side of the equipment represented on FIG. 2.
DETAILED DESCRIPTION OF THE INVENTION
The time domain reflectometry method according to the invention was
developed in order to test copper lines 10 between a central office 12 of
a telephone operator and subscribers 14. The goal of the test is to check
whether the copper pair 10 is able to comply with ADSL services, i.e.
whether its attenuation is inferior to a maximum attenuation imposed by
standards. Moreover, this method provides the possibility to localize
defects of the line 10.
In the schematic example shown on FIG. 1, a part 16 of the line 10 which is
relatively dose to the central office or station 12 is subject to
humidity; after this defect 16, in the direction from the office 12 to the
subscriber 14, the line presents a defect 18 corresponding to a bridge
tap, i.e. a derivation.
In the preferred embodiment of the method according to the invention, a set
of pulses are simultaneously or sequentially generated (sequentially in
the example) at the central station 12, each pulse covering a given
bandwidth and the bandwidths of all the pulses are overlapping. The whole
bandwidth formed by these pulses is wide, for instance, from 20 kHz to 6
MHz.
The set of pulses generated sequentially are transmitted from the office 12
on the line 10 and they are reflected by the defects 16, 18, as well as by
the line end 21, i.e. by subscriber 14. In fact, it is known that a
telephone set which is not in operation is a pure reflector.
The reflections or echoes are detected and processed at the central office
12 according to the equipment and method represented on FIGS. 2-8 and
these processed echoes provide information about the line 10 and its
defects.
In the simplified example represented on FIG. 2, the vertical line 22 on
the left represents the frequency and the diagram is, in the horizontal
direction, separated in five zones 24, 26, 28, 30, and 32 from the left to
the right. The zone 24 corresponds to the transmission of pulses at the
central office 12. The zone 26 represents the path of the pulses from the
central office 12 to the subscriber 14 and back from the subscriber 14 to
the central office 12. The zone 28 represents the processing of the
received echoes, each processing corresponding to a given band-width. The
zone 30 represents the sum ot the echoes which are processed in zone 28
and the zone 32 represents the spectrum of the transmitted, filtered and
reconstructed pulses.
In the simplified example, three pulses 40, 42, 44 are provided. The pulse
40 has a spectrum which covers the low frequencies, the pulse 42 has a
spectrum which covers the medium frequencies and the pulse 44 has a
spectrum which covers the high frequencies. For instance, the pulse 40
covers 12.5 kHz to 960 kHz, the pulse 42 covers 512 KHz to 2.6 MHz and the
pulse 44 covers 1.65 MHz to 6.3 MHz. The frequency bands of these pulses
are overlapping. More precisely, the higher part of the frequency band of
pulse 40 overlaps with the lower part of the frequency band of pulse 42
and the higher part of the frequency band of the pulse 42 overlaps with
the lower part of the frequency band of pulse 44.
Moreover, the overlapping portions of the spectra of the pulses are such
that, after reflection, reception and processing, the amplitude of the
added over-lapped parts equals the amplitude of the non-overlapping parts.
In other words, after reception, the spectrum of all the pulses is
practically flat, as shown on zone 32 of FIG. 2.
For instance, the higher part 40.sub.1 of the frequency band of pulse 40
varies as a sine function and the overlapping part of the lowest portion
42.sub.1 of the frequency band of pulse 42, varies as a cosine function.
As in zone, or step, 30, as explained herein after, the squares of the
amplitudes of the spectra are added, the overlapping regions 40.sub.1,
42.sub.1 provide after the final processing: sin.sup.2 +cos.sup.2 =1.
The method provides also, on the transmission side, a gain (amplification
or attenuation) for each pulse. For the sake of simplicity of the drawing,
only the amplification and attenuation for pulse 42 have been represented.
To the pulse 42 corresponds a gain represented by an amplifier 48, and on
the receiving side (zone 28), the inverse gain 50 is provided in order
that the resulting spectra (zone 32) be flat as explained herein above.
Each gain may be different from one bandwidth to another bandwidth, in
order to comply with the requirements of the standard which may impose
different constraints on the admissible maximum power for different
bandwidths.
On the receiving side, for each bandwidth, a processing is performed to
drastically reduce the noise: a synchronous averaging 52, followed by a
matched filtering 54 and a denoising 56. Each signal at the corresponding
bandwidth, after having been processed by the synchronous averaging 52,
the matched filtering 54 and the denoising 56, is submitted to a
reconstruction step (zone 30) 60 which sums the outputs of the processing
steps. Because of the matched filter properties, the summation of the
outputs of the processing steps is equivalent in the frequency domain, to
the following equation:
##EQU1##
where P.sub.i (f) is the transfer function for each pulse.
FIG. 3a shows the variation with time (in abscissa) of the pulse 40. The
curve 62 corresponds to the real part of the pulse, the curve 64
represents the variation of the imaginary part, and the curve 66
represents the envelope of the pulse.
The diagram of FIG. 3b shows the spectrum of the pulses 62 and 64, i.e. the
Fourier transform of said pulses. As shown, the spectrum, which extends
from 20 kHz to 200 kHz presents a flat part 68, a raising edge 70 and a
falling edge 72. As mentioned before, the falling edge has, for instance,
the shape of a sine function.
The diagrams of FIG. 4 represent the signals obtained after synchronous
averaging 52 and before matched filtering 54.
FIG. 4a represents the variation with time of the real part of the signal
for the pulses 44 at high frequencies, and FIG. 4b represents the
variation with time of the imaginary part for the signal at the output of
a synchronous averaging corresponding to the same high frequency pulses
44.
The diagrams of FIG. 4c and FIG. 4d correspond, respectively, to the real
and imaginary parts of the complex signal obtained after synchronous
averaging for the medium frequency pulses and FIG. 4e and FIG. 4f are
diagrams corresponding also to the real and imaginary parts of the signal
obtained at the output of the synchronous averaging 52 for the low
frequency pulses.
FIG. 4a and FIG. 4b show that, for high frequencies, the echoes present a
pulse 80. This pulse corresponds to a defect 16 dose to the central office
12, because the attenuation on the line increases sharply with the
frequency and the distance.
For medium frequencies (FIG. 4c and FIG. 4d), the diagram shows two echoes
82 and 84 corresponding to defects 16 and 18 and, for low frequencies
(FIG. 4e and FIG. 4f), the diagram shows several echoes corresponding to
defects 16, 18, and to the line end 21. It is recalled that, in this
example, the load coil defect 20 is not present.
The diagrams of FIG. 5 are similar to the diagrams of FIG. 4. but they
represent the echoes obtained after the matched filtering 54. The matched
filtering comprises a step of correlating the received pulse with the
transmitted pulse. This matched filtering provides a further sharp
decrease of the noise, as shown by comparison of the diagrams of FIG. 4a
to FIG. 4f with the corresponding diagrams of FIGS. 5a to 5f.
A further reduction of noise, more particularly for the high frequency
pulses is obtained with the denoising 56. This further noise suppression
comprises a step of determining a threshold below which the values of
signal and noise are set to zero, only the echoes which are above this
threshold being taken into account.
The threshold is, in an example, determined by an estimation of the noise
at the end (on the right of diagrams of FIG. 5) of the signals obtained
after matched filtering. In fact, the ending time corresponds to the end
of the line, at the subscriber's location, for which no signal can be
detected in medium and high frequencies; therefore, the signal end
corresponds, in practice, exclusively to noise for high and medium
frequencies. The noise is estimated by the variance of the signal at said
signal end and the threshold is determined by multiplying the square root
of this variance by a predetermined factor, for instance 2. More
precisely, the noise variance is estimated after the line end echo, i.e.
on a noise alone segment.
It is to be noted that the denoising is limited to the signals which appear
after the last echo. No denoising is performed on signals appearing before
the last echo.
The result of the denoising appears on the diagrams of FIG. 6a to FIG. 6f
which are similar to the diagrams of FIG. 4a to FIG. 4f.
FIG. 7 is a diagram showing the variation with time of the signal after the
summing of the outputs of the processing step for the three frequency
bands (output of adder 60).
This diagram shows that the time domain reflectometry method of the
invention provides, in this example, three echoes 96, 98 and 100. The
first echo corresponds mainly to defect 16, the second echo to defect 18
and the third to the line end 21.
On the diagram of FIG. 8, the abscissa is the time t and the ordinate is
the frequency f. It can be seen that, to the three lines 96, 98, 100,
correspond lines 102, 104 and 106. The line 102 extends on the whole
frequency band, the line 104 corresponds to the medium frequencies and the
last line 106 is limited to low frequencies.
Therefore, the diagrams obtained with FIG. 7 and FIG. 8 show that the
method according to the invention provides, with only one measurement,
information about the properties of the line and the defects, more
particularly about the location of such defects.
More generally, the method according to the invention provides directly an
estimation of the reflectometric impulse response of a line, with a good
time accuracy and a high signal-to-noise ratio at any distance.
*
