Note: Descriptions are shown in the official language in which they were submitted.
CA 02354483 2001-07-30
A METHOD FOR LOOP AND RATE DEPENDENT POWER CUTBACK
The present invention relates generally to transfer of data using Digital
Subscriber Loop
(DSL) technology, and specifically to a method for reducing transmitter power
used for
the transfer.
BACKGROUND OF THE INVENTION
Remote access and retrieval of data is becoming increasingly popular in data
communication. The proliferation of the Internet has provided a vast network
of
information that is available to the general public. As the Internet grows and
technology
advances, this information is becoming increasingly voluminous and the details
are
become increasingly intricate. What used to comprise mainly text information
has grown
to include still and moving images as well as sound. The increase in the
volume of
information to be transferred has presented a need for a high-speed Internet
connection,
since traditional telephone modems communicate at speeds to slow for efficient
communication.
One proposal for high-speed communication is the introduction of Digital
Subscriber
Line (DSL) technology. The various DSL technologies include ADSL, HDSL, SDSL,
SHDSL and ISDN BRI DSL systems. One of the most attractive features of DSL is
that
it is implemented using an infrastructure that already exists. DSL shares
copper twisted
pair lines typically used for telephone communication. However, only a small
portion of
the available bandwidth of the twisted pair line (0 to 4kHz) is used for Plain
Old
Telephone Service (POTS). DSL takes advantage of the available frequency
spectrum
from 4kHz to approximately 1.1 MHz for transmitting data.
Asymmetric DSL (ADSL) is currently the most practical form of DSL technology,
and
therefore the most widely implemented. ADSL is asymmetric in that its
downstream (to
a subscriber) capacity is larger than its upstream (from the subscriber)
capacity. An
ADSL transceiver unit at a central office a central office or remote loop
carrier (ATU-C)
is used for sending downstream information and receiving upstream information.
An
ADSL transceiver unit at a remote location or user end (ATU-R) is used for
receiving
1
CA 02354483 2001-07-30
downstream information and sending upstream information. Typically, a Discrete
Multi-
tone (DMT) scheme is used. The spectrum from 4kHz to 1.lMHz is divided into
256
sub-channels, or tones, each having a bandwidth of 4.3125kHz. Each sub-channel
uses
Quadrature Amplitude Modulation (QAM) to carry 2 to 15 bits/QAM symbol.
A predefined power level is used by the ATU-C for transmitting the downstream
signal.
At the ATU-R, the achievable downstream rate is a function of the received
signal level
and of the receiver input-referenced noise levels. Where the loop attenuation
is modest
and the target downstream rate is lower than the achievable rate, the
predefined ATU-C
transmission level can be much higher than necessary. Therefore, an
unnecessary amount
of power is consumed by the ATU-C and additional crosstalk noise is induced in
adjacent
DSL lines. Reducing the ATU-C transmit level would both save power at the ATU-
C
and reduce crosstalk noise, improving the quality of signal on adjacent lines.
From a
crosstalk perspective, it is beneficial if the ATU-C transmit power can be
reduced during
both initialization and steady state (also referred to as "Showtime")
operation. If the
transmit power and resulting crosstalk is only reduced on entry to Showtime,
the higher
crosstalk levels during initialization may cause excessive errors and even
force re-
initialization on adjacent DSL lines.
The benefit of such a power cutback can be significant in Digital Loop Carrier
applications, for example, where subscriber loops are typically shorter than
seen at
central office/resident line interfaces and where power and thermal budgets
are tight.
Assuming a DMT transmit signal with peak-to-average-ratio (PAR) of l4.SdB,
downstream power cutback can yield significant savings on line driver power
alone.
ANSI and ITU compliant ATU-Cs reduce their transmitter power on very short
loops to
avoid overloading the ATU-R. This reduction in downstream transmission power
is often
referred to as a "politeness cutback". 'The politeness cutback is determined
in accordance
with a loop loss estimate from an upstream signal path and applies on loops
shorter than
approximately 2-3kft 26AWG-equivalent. Other downstream transmitter power
cutbacks
are also specified in the ITU splitterless ADSL standard (G.992.2, which is
also known as
2
CA 02354483 2001-07-30
G.lite) for reducing the downstream signal level to the ATU-R. This cutback is
designed
to address ADSL signal levels that could cause distortion in the presence of
an off hook
telephone set.
However, even with the implementation of the politeness cutback and off hook
cutback,
the ATU-C often transmits at a greater power than necessary and there is no
agreed upon
mechanism to implement a general downstream power cutback. It is an object of
the
present invention to obviate or mitigate at least some of the above mentioned
disadvantages.
SUMMARY OF THE INVENTION
In accordance with an aspect of the present invention, there is a provided a
method for
reducing power required for transmitting a signal from a first transceiver to
a second
transceiver. The method comprises the steps of estimating an excess amount of
power
used by the first transceiver for transmitting the signal, reducing the first
transceiver's
power use by the excess amount of power to a reduced power level, and
transmitting the
signal from the first transceiver using the reduced power level. The reduced
power level
achieves a transmission rate of the signal within a predefined tolerance of
its preferable
rate.
BRIEF DESCRIPTION OF THE DRAWINGS
Embodiments of the invention will now be described by way of example only with
reference to the following drawings in which:
Figure la is a graph illustrating a relationship between transmission rate and
loop
length for signals transmitted at a plurality of power levels, in the presence
of
forty-nine G.Lite disturbers;
Figure lb is a graph illustrating a relationship between transmission rate and
loop
length for signals transmitted at a plurality of power levels, in the presence
of
ten HDSL disturbers;
3
CA 02354483 2001-07-30
Figure lc is a graph illustrating a relationship between transmission rate and
loop
length for signals transmitted at a plurality of power levels, in the presence
of
five adjacent binder T1 disturbers;
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
ATU-C power requirements can be reduced when there is an excess signal-to-
noise ratio
(SNR) margin at the ATU-R receiver. It is possible to determine an amount of
cutback
possible given a target or maximum downstream data rate and a target or
maximum
downstream SNR margin. The target data rate is the rate at which the ATU-C is
to
transmit the downstream signal and the SNR margin is a margin against SNR
degradation
for a specified bit-error-rate.
Referring to figures 1 a through 1 c, graphs illustrate attainable downstream
rates over loop
lengths of 0-l2kft (26AWG). Each graph has a plot illustrating the estimated
transmission rates with respect to the loop length with a DS power cutback of
OdB, 6dB,
l2dB, and lSdB in the presence of crosstalk from other DSLs on adjacent pairs.
Full-rate
ADSL (G.dmt and T1.413) are also illustrated for comparison. The attainable
rates have
been estimated assuming a 4dB SNR margin, a 3dB coding gain, use of downstream
carriers 36-127, and a downstream receiver noise floor of -136.8dBm/Hz. The
politeness
cutback (of 0 to l2dB) is not shown in the graphs. However, the politeness
cutback is
only applied for loop lengths up to approximately 2kft (26AWG).
Referring to figure 1 a, the crosstalk scenario is caused by forty-nine (49)
other G.lite
disturbers. This case is typical of a residential neighbourhood, where T1 or
HDSL
services in the same or adjacent binders is less common. It should be noted
that if all
local subscribers are being served off power-cutback capable digital loop
carriers
(DLCs), the G.lite crosstalk levels would actually be lower and the rates
would be
improved over those shown.
4
CA 02354483 2001-07-30
Referring to figure lb, the crosstalk scenario is caused by ten (10) HDSL
disturbers. For
this case, full capacity (I.SMbps) is possible with lSdB of cutback for loop
lengths of up
to 9kft (26AWG), within Carrier Serving Area (CSA) loop engineering rules
Referring to figure lc, the crosstalk scenario is caused by five (5) adjacent
binder T1
disturbers. With T1 disturbers in adjacent binders, full capacity is possible
on a loop of
length of up to 9kft with approximately 6dB of cutback.
In order to implement an appropriate power cutback, it is necessary to either
obtain an
estimate of the SNR at the receiver or develop a method for determining the
amount of
signal cutback possible while achieving the target downstream rate and SNR
margin. A
signal-to-additive-noise (as opposed to signal-dependent-noise such as mis-
equalization
error) estimate permits a more accurate estimate of the SNR degradation
resulting from a
signal level reduction. Once the cutback is determined, it is communicated
between the
ATU-C and the ATU-R and implemented by the ATU-C.
Estimating the SNR and determining an optimal amount of downstream transmit
signal
power cutback is most easily accomplished by the ATU-R receiver. Thus far,
some
proposals have been made in the ITU standards body to permit this and/or
require the
ATU-R to minimize downstream transmit power via the per-DMT-carrier gains when
the
maximum downstream SNR margin is exceeded. These methods are not required or
supported in the current ADSL standards nor do they address ATU-C transmit
(and
generated crosstalk) levels during initialization. As a result, it is
preferable to have
methods for estimating downstream SNR and/or potential cutback in the ATU-C.
The techniques described herein describe a downstream power cutback mechanism
that
reduces downstream transmit power based on a metric of excess SNR or capacity
at the
downstream receiver (ATU-R). The following techniques can be implemented with
an
ATU-R compliant with the current ADSL standards (ITU-T 6.992.1, 6.992.2 and
ANSI
T1.413 issue 2).
5
CA 02354483 2001-07-30
The amount of downstream transmit power reduction that can be tolerated for a
given
capacity target is a function of the loop and the crosstalk environment, which
together
determine the SNR per receiver sub-Garner. Unfortunately, this information is
not
available to the ATU-C. This information may be inferred from the bits per
carrier i (B;s)
and gains per carrier i (G;s) sent during initialization. Specifically, bits
and gains
information from the remote-end is transmitted during R-B&G, toward the end of
the
initialization. However, at this point it is too late to implement a
downstream power
cutback without impacting the ATU-R, which will at least need to adapt its
receiver
automatic gain control (AGC).
Therefore, in the absence of explicit information from the ATU-R, several
approaches to
initialization are disclosed that the potential for power savings without
significant
capacity losses. For all cases, the amount of cutback versus the measured
upstream
power can be an operator-specified parameter, as will be appreciated by a
person skilled
in the art.
A) Reduce downstream power based on measured upstream power alone
Downstream power cutback based on received upstream power at the ATU-C is
already
part of the current standards for the politeness cutback. The politeness
cutback is used on
very short loops, approximately 0 to 2kft 26AWG equivalent, so there is no
impact on
downstream capacity regardless of the crosstalk noise environment at the ATU-
R. The
politeness cutback is performed to minimize the peak signal handling
requirements of the
ATU-R receiver on short loops. Since it is invoked on only very short loops,
it has no
impact on power consumption for a typical loop of length 6-8kft
Referring to figures 1-3, it is possible to cutback the downstream power for
loops longer
than 2kft 26AWG-equivalent and maintain an acceptable transmission rate.
However, the
degree to which the power is cutback depends on one's willingness to risk a
failure
during full initialization. For example, a cutback of l2dB is employed where
the loop
length is estimated to be less than 9kft 26AWG-equivalent. Assuming that the
loop
length and equivalent downstream loop losses are adequately estimated from the
6
CA 02354483 2001-07-30
upstream power measured at the ATU-C and there are no adjacent binder T1
disturbers,
the l2dB cutback will not significantly reduce the DS capacity below
approximately
1.SMbps.
Unexpected capacity losses may still occur if the ATU-R is not capable of
meeting
standard test cases because of a noisy front end or AM ingress noise. In those
cases, the
cutback should be reduced or eliminated and a second full-initialization
triggered.
Referring to table 1 below, a downstream power cutback is determined as a
function of an
estimated average of upstream loop attenuation. The estimated average loop
attenuation
provides an estimate of the loop length. The upstream loop attenuation is
defined as the
difference between the upstream reference power measured in dBm and the total
upstream power measured by the ATU-C on subcarriers 7-18 during stage R-
REVERB1
of the initialization. The upstream reference power is defined as the total
power used by
1 S the ATU-R for transmitting subcarners 7-18 using the R-REVERB 1 transmit
power
spectral density (PSD) level.
Average >15>14 >13>12>11 >10>9 >8 >7 >6 >5 >4 >3 >2 >1 >0
U/S
loop attenuation
(dB)
D/S Power 0 1 2 3 4 5 6 6 7 8 8 9 10 11 12 12
cut
back (dB)
Table 1
Table 1 has been constructed to minimize achievable downstream rate loss with
a 6dB
SNR margin in the presence of 24 HDSL or 24 DSL near-end crosstalk (NEXT)
sources.
A person skilled in the art will appreciate that it is also possible to
provide different
cutback tables based on operator-provisioned parameters. These parameters
include a
maximum downstream rate and SNR margin. Generally, the cutback is more
aggressive
when the downstream rate and/or SNR margin are reduced, and less aggressive
when
they are increased.
CA 02354483 2001-07-30
The operator can also be given the capability of overriding this mechanism via
an
additional OAM-parameter. It is preferable that the maximum amount of cutback
is kept
to 12 dB as some implementations may have problems in applying higher cutbacks
in the
analog domain. A cutback applied in the digital domain results in higher
requirements for
the dynamic range of a digital-to-analogue converter (DAC) in the transmitter.
In order to implement the above in accordance with the current ADSL standards,
the
ATU-C adjusts its transmit power once the downstream power cutback has been
determined. This is achieved early enough (before transmission of C-REVERB)
such
that the ATU-R can adjust its AGC without triggering a new initialization.
Note that this
mechanism is similar to the existing politeness cutback but is more aggressive
in reducing
downstream transmit levels. This cutback is motivated by power savings, unlike
the
politeness cutback, which attempts to avoid overloading an ATU-R receiver on
very short
loops.
B) Two passes through full-initialization during initial install or when on-
hook loop
conditions change significantly
In an alternate embodiment, the ATU-C forces a second full-initialization once
it has
received the B;s, G;s, performance (SNR) margin, and attainable rate from the
ATU-R
near the end of the first initialization procedure. A per-carrier SNR (linear
scale) at the
ATU-R is estimated as 103*B' + G' + ",~'°~i20, where i is the carrier
number, B; is the number
of bits on carrier i, G; is the gain of carrier i in dB, and margin is the SNR
margin
reported by the ATU-R (in dB). The ATU-C uses the downstream SNR to perform a
downstream rate (capacity) estimate and determines how much of a cutback can
be
tolerated, for the target downstream rate and margin.
The line rate capacity estimate, less forward error correction (FEC) overhead,
in bits per
symbol period with no cutback is given by
SNR.
C = ~ b; ; b; = round (logz (1 + r ' )) and 2 S b; <_ b",a" Equation 1
a
8
CA 02354483 2001-07-30
where bm~ is the maximum # of bits/carner supported in the ATU-R receiver, and
r =1o~9~g+~0-3>~~0
is the SNR gap for a bit-error-rate of 10-~ with IOdB margin and 3dB coding
gain, for
6.992.2.
'The maximum cutback is determined such that the capacity with that cutback,
C', is
greater than or equal to 0.98 x C, where
SNR. l cutback
C'= ~ b;'; b'; = round (loge (1 + ' r )) and 2 <_ b; <_ b",a,~ Equation 2
r
The cutback is determined in linear form, corresponding to a multiple of 2dB
The initialization procedure is then repeated with the ATU-C transmitting at a
power
level including the cutback throughout the initialization procedure.
It should be noted that the ATU-R B;s, G;s, and performance margin are based
on
measurements in C-MEDLEY that include both additive noise (e.g. crosstalk) and
signal-
level-dependent noise, particularly mis-equalization error. The ATU-C has no
direct
information on the relative contribution of these two noise or error types. If
mis-
equalization error is the dominant limitation of SNR, then downstream transmit
levels
could be reduced more aggressively than the case where additive (signal-level-
independent) noise is dominant.
This process also requires approximately twice as long (up to about 20
seconds) to
initialize the link, but only needs to be invoked whenever a change has
occurred in the
noise environment at the ATU-R, for example, when a HDSL service is newly
deployed.
Such a change in environment may be detected as a degradation in performance
during
Showtime, including a reduction in the noise margin, an increase in detected
CRC errors,
9
CA 02354483 2001-07-30
and the like. Alternately an enviroment change may be detected on a failed
power
cutback initialization following a period when the communication link is
powered down.
C) Reduce DS power by excess margin reported by ATU R
Negotiation of transmission rates for the communication link occurs over a
series of
handshakes between the ATU-C and the ATU-R. The ATU-C begins by providing the
ATU-R with a list of four (4) rate options to be met. In yet an alternate
embodiment, the
ATU-C further transmits the required SNR margin to be met. The ATU-R responds
to
those options indicating the highest rate, if any, that can be supported. It
also provides
the ATU-C with the average downstream loop attenuation and the SNR margin at
that
rate. The values should be approximately the same for all carriers if the G;s
have been
calculated to equalize the SNR margin across all carriers. If the ATU-R is
able to support
the highest transmission rate, with a SNR margin in excess of the desired
margin, then a
power reduction in the amount of the difference between the desired SNR margin
and
actual SNR margin can be realized. The difficulty with this approach is that
the ATU-R
is not necessarily designed to adapt to a signal level change at this point in
the
initialjzation, forcing a new full initialization.
D) Over-specify margin in first round of rate negotiations
In yet an alternate embodiment, during the first round of rate negotiations,
the ATU-C
specifies SNR margin comprising a minimum SNR margin and an additional 'N' dB
of
SNR margin. If the ATU-R responds (in R-MSG-RA) that it can support a high
link rate
at the inflated SNR margin, the ATU-C drops its transmitter power by 'N' dB
and sets the
required margin lower by an equivalent amount for the second round of rate
negotiations.
If, however, the ATU-R cannot support the target transmission rate with the
inflated
margin, the ATU-C does not cutback its transmit power, but still reduces the
minimum
required SNR margin (by 'N') for the second round of negotiations. As with the
previous
embodiment, the difficulty with this approach is that the ATU-R is not
necessarily
designed to adapt to a signal level change at this point in the
initialization, forcing a new
full initialization
CA 02354483 2001-07-30
More optimal techniques than those described above require changes to the
existing
standards. The changes allow for a faster initialization and maximize the
achievable
cutback.
E) ATU R signals attainable cutback implicitly in downstream G;s based on SNR
measured in C MEDLEY
In yet an alternate embodiment the ATU-R can indicate a power cutback
implicitly by
reducing the G;s to be used on each of the downstream carriers where there is
excess SNR
margin, based on SNR measurements made during C-MEDLEY . If so, this sets the
downstream power cutback desired. With the current ADSL standards, one cannot
assume that a vendor's ATU-R will specify the G;s in this manner. That is, the
ATU-R
may use the G;s only to equalize the margin on each carrier, while keeping an
overall
unnecessarily large SNR margin. The G;'s are communicated to the ATU-C late in
the
initialization process. As a result, any change in transmit gain must be exact
as there is
no time for the ATU-R to adapt its receiver to an imprecise gain change before
the start
of SHOWTIME. Practically, this causes the signal to be reduced in the digital
domain,
before the DAC, and can place excess demands on the DAC dynamic range. This
method
has the disadvantage of not applying the cutback until the Gi's are
implemented on entry
to Showtime. As a result, the transmit level, and resulting crosstalk into
adjacent lines,
remains high through initialization.
As described in method B above, the SNR measured by the ATU-R in C-MEDLEY
includes both signal-level-dependent and signal-level-independent noise at the
ATU-R.
F) ATU R signals attainable cutback implicitly in downstream Gis based on a
signal-
level-independent SNR or noise%rror signal estimate
In yet an alternate embodiment, the ATU-R determines the amount of downstream
power
cutback that can be tolerated while meeting the target downstream rate and SNR
margin.
In order to determine the SNR impact of a reduced downstream transmit signal
where
mis-equalization error may be dominant, the ATU-R requires several parameters.
These
parameters include an estimate of the receiver noise floor over the downstream
frequency
n
CA 02354483 2001-07-30
band in the absence of the downstream signal, an estimate of the channel
attenuation in
the downstream band, and knowledge of the provisioned downstream rates and
margins
(i.e. maximums and targets). The ATU-R calculates the achievable downstream
cutback
from the this information and communicates the cutback back to the ATU-C,
which
implements the cutback in time for the ATU-R to re-adjust its receiver AGC
prior to the
SNR estimation during C-MEDLEY. A sample method for estimating the above-
described parameters is described as follows.
The ATU-R measures the received downstream power (during C-REVERB 1 ) and
estimates the average loop attenuation, as per the current standards, across
carriers 7-18
based on a measured per-carrier received signal level and known per-carrier
transmit
signal levels in C-REVERB1. C-REVERB1 may be sent with a politeness cutback.
The
ATU-R is informed of the extent of the cutback via a message from the ATU-C. C-
REVERB1 may also be replaced by another standard-specific line probing signal
that
permits extrapolation of the channel attenuation over the downstream frequency
band but
has different spectral characteristics than those for C-REVERB1.
The ATU-R measures the received downstream noise level during C-QUIET. A
minimum noise floor measurement resolution capability may be specified to
ensure that
the ATU-R is capable of supporting a specific downstream power cutback for a
given
downstream loop loss and crosstalk noise environment. Any receiver AGC should
be set
to a value sufficient to resolve the noise level at the loop interface.
However, it should
not be set so high as to ignore noise contributions from its own front end.
The receiver
may need to budget for those noise sources if they become dominant at low AGC
gain
settings.
The ATU-C communicates the maximum, target and minimum downstream transmission
rates and SNR margins to the ATU-R. This can be achieved, for example, during
C-
MSG1 in the current standard.
12
CA 02354483 2001-07-30
The ATU-R calculates the amount of downstream cutback possible without
reducing
estimated downstream capacity below the provisioned maximum downstream rate at
the
provisioned maximum margin. The capacity calculations are based on a received
signal
(transmit signal less channel attenuation) and receiver input referenced noise
levels as
measured at each upstream Garner frequency and mapped to an equivalent per-
carrier
SNR, SNRi. Note that this capacity formulation is based on the measured signal
and noise
levels, independent of near end echo, timing fitter or mis-equalization
effects. An
additional 6dB of margin is included in the SNR gap figure used in the
capacity
calculation to cover these impairments, measurement tolerances, and for
providing a
hedge against introduction of more NEXT sources into the same binder group.
The line
rate capacity estimates are calculation using equations 2 and 3 as previously
described.
The ATU-R communicates the required cutback to the ATU-C, for example in a
field of
R-MSGS1. Allowing for a power cutback range of 0-lSdB in 1dB steps, a 4-bit
message
field is required.
After receiving and decoding the message from the ATU-R, the ATU-C applies the
power cutback approximately 500-3500 symbols into C-MEDLEY. The ATU-R is able
to adapt its receiver gain to the new ATU-C transmit level over this interval
prior to its
final SNR estimation during C-MEDLEY.
Although the invention has been described with reference to certain specific
embodiments, various modifications thereof will be apparent to those skilled
in the art
without departing from the spirit and scope of the invention as outlined in
the claims
appended hereto.
13
