T Rec G.991.2 200312 I!!pdf e

INTERNATIONAL TELECOMMUNICATION UNION
ITU-T G.991.2
TELECOMMUNICATION (12/2003)
STANDARDIZATION SECTOR
OF ITU
SERIES G: TRANSMISSION SYSTEMS AND MEDIA,

DIGITAL SYSTEMS AND NETWORKS
Digital sections and digital line system – Access networks
Single-pair high-speed digital subscriber line

(SHDSL) transceivers
ITU-T Recommendation G.991.2

ITU-T G-SERIES RECOMMENDATIONS
TRANSMISSION SYSTEMS AND MEDIA, DIGITAL SYSTEMS AND NETWORKS
INTERNATIONAL TELEPHONE CONNECTIONS AND CIRCUITS G.100–G.199

GENERAL CHARACTERISTICS COMMON TO ALL ANALOGUE CARRIER- G.200–G.299
TRANSMISSION SYSTEMS
INDIVIDUAL CHARACTERISTICS OF INTERNATIONAL CARRIER TELEPHONE G.300–G.399
SYSTEMS ON METALLIC LINES
GENERAL CHARACTERISTICS OF INTERNATIONAL CARRIER TELEPHONE G.400–G.449
SYSTEMS ON RADIO-RELAY OR SATELLITE LINKS AND INTERCONNECTION WITH
METALLIC LINES
COORDINATION OF RADIOTELEPHONY AND LINE TELEPHONY G.450–G.499
TESTING EQUIPMENTS G.500–G.599
TRANSMISSION MEDIA CHARACTERISTICS G.600–G.699
DIGITAL TERMINAL EQUIPMENTS G.700–G.799
DIGITAL NETWORKS G.800–G.899
DIGITAL SECTIONS AND DIGITAL LINE SYSTEM G.900–G.999
General G.900–G.909
Parameters for optical fibre cable systems G.910–G.919
Digital sections at hierarchical bit rates based on a bit rate of 2048 kbit/s G.920–G.929
Digital line transmission systems on cable at non-hierarchical bit rates G.930–G.939
Digital line systems provided by FDM transmission bearers G.940–G.949
Digital line systems G.950–G.959
Digital section and digital transmission systems for customer access to ISDN G.960–G.969
Optical fibre submarine cable systems G.970–G.979
Optical line systems for local and access networks G.980–G.989
Access networks G.990–G.999
QUALITY OF SERVICE AND PERFORMANCE – GENERIC AND USER-RELATED G.1000–G.1999
ASPECTS
TRANSMISSION MEDIA CHARACTERISTICS G.6000–G.6999
DIGITAL TERMINAL EQUIPMENTS G.7000–G.7999
DIGITAL NETWORKS G.8000–G.8999
For further details, please refer to the list of ITU-T Recommendations.

Single-pair high-speed digital subscriber line (SHDSL) transceivers
Summary
This Recommendation describes a transmission method for data transport in telecommunications
access networks. SHDSL transceivers are designed primarily for duplex operation over mixed gauge
two-wire twisted metallic pairs. Optional multi-pair operation is supported for extended reach
applications. Optional signal regenerators for both single-pair and multi-pair operation are specified,
as well. SHDSL transceivers are capable of supporting selected symmetric user data rates in the
range of 192 kbit/s to 2312 kbit/s using a Trellis Coded Pulse Amplitude Modulation (TCPAM) line
code. Optional extensions described in Annex F allow user data rates up to 5696 kbit/s. SHDSL
transceivers are designed to be spectrally compatible with other transmission technologies deployed
in the access network, including other DSL technologies. SHDSL transceivers do not support the use
of analogue splitting technology for coexistence with either POTS or ISDN. Regional requirements,
including both operational differences and performance requirements, are specified in Annexes A, B
and C. Requirements for signal regenerators are specified in Annex D. Annex E describes
application-specific framing modes that may be supported by SHDSL transceivers.
See Annex H/G.992.1 [1] for specifications of transceivers for use in networks with existing
TCM-ISDN service (as specified in Appendix III/G.961, in Bibliography [B1]).
History
With respect to the previous version 1 (2001), this version 2 introduces the following additions and
modifications:
• The optional four-wire mode has been extended to a more general multi-pair mode which
provides optional support for up to four-pair connections. See 7.2.1.5. Note that the integrity
of the optional 4-wire mode in revision 1 is preserved. The four-wire mode is identical to
M-pair mode with M = 2, except for the method of assigning ordinal numbers to the wire
pairs. In four-wire mode, the ordinal numbers (the wire pair identification number) are
assigned as described in 6.3, while in M-pair mode the ordinal numbers are assigned to wire
pairs as described in 7.2.1.5.
• The loops and test conditions specified in Annex B have been updated (see B.3.3), and
Appendix IV, Tabulation of Annex B Noise Profiles, has been added.
• Optional extensions, described in Annex F, allow user data rates up to 5696 kbit/s.
• Deactivation and warm-start, as specified in Annex H, have been added.
• Support for Dynamic Rate Repartitioning has been added to Dual-Bearer mode. See E.10.3.
• TPS-TC definitions have been added for Packet Transfer Mode (E.11), Synchronous
Transfer Mode with a Dedicated Signalling Channel (E.12), and V5 Encapsulated ISDN or
POTS (E.13).
Source
ITU-T Recommendation G.991.2 was approved on 14 December 2003 by ITU-T Study Group 15
(2001-2004) under the ITU-T Recommendation A.8 procedure.
ITU-T Rec. G.991.2 (12/2003) i

FOREWORD
The International Telecommunication Union (ITU) is the United Nations specialized agency in the field of
telecommunications. The ITU Telecommunication Standardization Sector (ITU-T) is a permanent organ of
ITU. ITU-T is responsible for studying technical, operating and tariff questions and issuing
Recommendations on them with a view to standardizing telecommunications on a worldwide basis.
The World Telecommunication Standardization Assembly (WTSA), which meets every four years,
establishes the topics for study by the ITU-T study groups which, in turn, produce Recommendations on
these topics.
The approval of ITU-T Recommendations is covered by the procedure laid down in WTSA Resolution 1.
In some areas of information technology which fall within ITU-T's purview, the necessary standards are
prepared on a collaborative basis with ISO and IEC.
NOTE
In this Recommendation, the expression "Administration" is used for conciseness to indicate both a
telecommunication administration and a recognized operating agency.
Compliance with this Recommendation is voluntary. However, the Recommendation may contain certain
mandatory provisions (to ensure e.g. interoperability or applicability) and compliance with the
Recommendation is achieved when all of these mandatory provisions are met. The words "shall" or some
other obligatory language such as "must" and the negative equivalents are used to express requirements. The
use of such words does not suggest that compliance with the Recommendation is required of any party.
INTELLECTUAL PROPERTY RIGHTS

ITU draws attention to the possibility that the practice or implementation of this Recommendation may
involve the use of a claimed Intellectual Property Right. ITU takes no position concerning the evidence,
validity or applicability of claimed Intellectual Property Rights, whether asserted by ITU members or others
outside of the Recommendation development process.
As of the date of approval of this Recommendation, ITU had received notice of intellectual property,
protected by patents, which may be required to implement this Recommendation. However, implementors
are cautioned that this may not represent the latest information and are therefore strongly urged to consult the
TSB patent database.
 ITU 2004
All rights reserved. No part of this publication may be reproduced, by any means whatsoever, without the
prior written permission of ITU.
ii ITU-T Rec. G.991.2 (12/2003)

CONTENTS
Page
1 Scope ............................................................................................................................ 1
2 References..................................................................................................................... 1
3 Definitions and abbreviations ....................................................................................... 2
3.1 Definitions ...................................................................................................... 2
3.2 Abbreviations ................................................................................................. 3
4 Reference models.......................................................................................................... 6
4.1 STU-x functional model ................................................................................. 6
4.2 User plane protocol reference model.............................................................. 7
4.3 Application models......................................................................................... 7
5 Transport capacity ........................................................................................................ 8
6 PMD layer functional characteristics............................................................................ 8
6.1 Data mode operation....................................................................................... 8
6.2 PMD activation sequence ............................................................................... 12
6.3 PMD pre-activation sequence......................................................................... 16
6.4 G.994.1 pre-activation sequence .................................................................... 20
7 PMS-TC layer functional characteristics...................................................................... 23
7.1 Data mode operation....................................................................................... 23
7.2 PMS-TC activation......................................................................................... 28
8 TPS-TC layer functional characteristics....................................................................... 30
8.1 Payload block data structure........................................................................... 30
8.2 Data interleaving in M-pair mode................................................................... 31
9 Management ................................................................................................................. 32
9.1 Management reference model ........................................................................ 32
9.2 SHDSL performance primitives ..................................................................... 33
9.3 SHDSL line related performance parameters................................................. 34
9.4 Performance data storage ............................................................................... 35
9.5 Embedded operations channel........................................................................ 35
10 Clock architecture......................................................................................................... 58
10.1 Reference clock architecture .......................................................................... 58
10.2 Clock accuracy ............................................................................................... 59
10.3 Definitions of clock sources ........................................................................... 60
10.4 Synchronization to clock sources ................................................................... 60
11 Electrical characteristics ............................................................................................... 60
11.1 Longitudinal balance ...................................................................................... 61
11.2 Longitudinal output voltage ........................................................................... 62
11.3 Return loss ...................................................................................................... 62
11.4 Transmit power testing ................................................................................... 64
ITU-T Rec. G.991.2 (12/2003) iii

Page
11.5 Signal transfer delay ....................................................................................... 65
12 Conformance testing..................................................................................................... 66
12.1 Micro-interruptions......................................................................................... 66
Annex A – Regional requirements – Region 1 ........................................................................ 67
A.1 Scope .............................................................................................................. 67
A.2 Test loops........................................................................................................ 67
A.3 Performance Tests .......................................................................................... 68
A.4 PSD masks...................................................................................................... 82
A.5 Region-specific functional characteristics...................................................... 89
Annex B – Regional requirements – Region 2 ........................................................................ 96
B.1 Scope .............................................................................................................. 96
B.2 Test loops........................................................................................................ 96
B.3 Performance testing ........................................................................................ 99
B.4 PSD masks...................................................................................................... 117
B.5 Region-specific functional characteristics...................................................... 122
Annex C – Regional requirements – Region 3 ........................................................................ 125
Annex D – Signal regenerator operation.................................................................................. 125
D.1 Reference diagram.......................................................................................... 125
D.2 Start-up procedures......................................................................................... 126
D.3 Symbol rates ................................................................................................... 130
D.4 PSD masks...................................................................................................... 130
Annex E – Application-specific TPS-TC framing................................................................... 130
E.1 TPS-TC for clear channel data ....................................................................... 130
E.2 TPS-TC for clear channel byte-oriented data................................................. 131
E.3 TPS-TC for unaligned DS1 transport ............................................................. 132
E.4 TPS-TC for aligned DS1/fractional DS1 transport ........................................ 133
E.5 TPS-TC for European 2048 kbit/s digital unstructured leased line
(D2048U)........................................................................................................ 134
E.6 TPS-TC for unaligned European 2048 kbit/s digital structured leased line
(D2048S) ........................................................................................................ 134
E.7 TPS-TC for aligned European 2048 kbit/s digital structured leased line
(D2048S) and fractional ................................................................................. 135
E.8 TPS-TC for synchronous ISDN basic access ................................................. 136
E.9 TPS-TC for ATM transport ............................................................................ 154
E.10 Dual-bearer TPS-TC mode............................................................................. 161
E.11 TPS-TC for PTM transport............................................................................. 176
E.12 TPS-TC for STM with a Dedicated Signalling Channel (DSC)..................... 182
E.13 TPS-TC for LAPV5 enveloped POTS or ISDN............................................. 184
iv ITU-T Rec. G.991.2 (12/2003)

Page
Annex F – Region 1 requirements for payload data rates up to 5696 kbit/s............................ 188
F.1 Scope .............................................................................................................. 188
F.2 Data rate.......................................................................................................... 189
F.3 Mapper............................................................................................................ 191
F.4 PSD masks...................................................................................................... 192
F.5 Crosstalk interference requirements ............................................................... 194
F.6 Functional characteristics ............................................................................... 195
Annex G – Reserved for Region 2 requirements for data rates between 2320 kbit/s and
**max rate** ................................................................................................................ 196
Annex H – Deactivation and warm-start procedure................................................................. 196
H.1 Deactivation to reduced power mode ............................................................. 196
H.2 Warm-start activation ..................................................................................... 198
Appendix I – Test circuit examples ......................................................................................... 203
I.1 Example crosstalk injection test circuit.......................................................... 203
I.2 Example coupling circuits for longitudinal balance and longitudinal
output voltage ................................................................................................. 203
I.3 Return loss test circuit .................................................................................... 204
I.4 Transmit PSD/total power measurement test circuit ...................................... 205
Appendix II – Typical characteristics of cables....................................................................... 205
II.1 Typical characteristics of cables for Annex B................................................ 205
Appendix III – Signal regenerator start-up description ........................................................... 206
III.1 STU-R initiated Start-up................................................................................. 207
III.2 STU-C initiated start-up ................................................................................. 208
III.3 SRU initiated start-up ..................................................................................... 209
III.4 Collisions and retrains .................................................................................... 209
III.5 Diagnostic mode activation ............................................................................ 210
Appendix IV – Tabulation of Annex B noise profiles ............................................................. 210
BIBLIOGRAPHY.................................................................................................................... 223
ITU-T Rec. G.991.2 (12/2003) v

Single-pair high-speed digital subscriber line (SHDSL) transceivers
1 Scope
This Recommendation describes a transmission method for providing Single-pair High-speed
Digital Subscriber Line (SHDSL) service as a means for data transport in telecommunications
access networks. This Recommendation does not specify all the requirements for the
implementation of SHDSL transceivers. Rather, it serves only to describe the functionality needed
to assure interoperability of equipment from various manufacturers. The definitions of physical user
interfaces and other implementation-specific characteristics are beyond the scope of this
Recommendation.
For interrelationships of this Recommendation with other G.99x-series ITU-T Recommendations,
see ITU-T Rec. G.995.1 in Biblography [B2] (informative).
The principal characteristics of this Recommendation are as follows:
− provisions for duplex operation over one (or, optionally, multiple) mixed gauge two-wire
twisted metallic pairs;
− specification of the physical layer functionality, e.g., line codes and forward error
correction;
− specification of the data link layer functionality, e.g., frame synchronization and framing of
application, as well as Operations, Administration and Maintenance (OAM) data;
− provisions for optional use of repeaters for extended reach;
− provisions for spectral compatibility with other transmission technologies deployed in the
access network;
− provisions for regional requirements, including functional differences and performance
requirements.
2 References
The following ITU-T Recommendations and other references contain provisions which, through
reference in this text, constitute provisions of this Recommendation. At the time of publication, the
editions indicated were valid. All Recommendations and other references are subject to revision;
users of this Recommendation are therefore encouraged to investigate the possibility of applying the
most recent edition of the Recommendations and other references listed below. A list of the
currently valid ITU-T Recommendations is regularly published. The reference to a document within
this Recommendation does not give it, as a stand-alone document, the status of a Recommendation.
[1] ITU-T Recommendation G.992.1 (1999), Asymmetric Digital Subscriber Line (ADSL)
transceivers.
[2] ITU-T Recommendation G.994.1 (2003), Handshake procedures for digital subscriber line
(DSL) transceivers, plus Amendment 1 (2004).
[3] ITU-T Recommendation G.997.1 (2003), Physical layer management for digital subscriber
line (DSL) transceivers.
[4] IETF RFC 1662 (1994), PPP in HDLC-like Framing.
[5] ISO 8601:2000, Data elements and interchange formats – Information interchange –
Representation of dates and times.
ITU-T Rec. G.991.2 (12/2003) 1

[6] ITU-T Recommendation G.996.1 (2001), Test procedures for Digital Subscriber Line
(DSL) transceivers.
[7] IEC 60950 (1999), Information technology equipment – Safety.
[8] ITU-T Recommendation I.432.1 (1999), B-ISDN user-network interface – Physical layer
specification: General characteristics.
[9] ETSI EN 300 324-1 (1994), V Interfaces at the digital Local Exchange (LE); V5.1 interface
for the support of Access Network (AN); Part 1: V5.1 Interface specification.
[10] ETSI EG 201 900-1 V1.1.1 (2001), Services and Protocols for Advanced Networks
(SPAN); Narrowband Services over ATM; Loop Emulation Service (LES) using AAL2;
Part 1: LES interface specification [ATM Forum Specification af-vmoa-0145.000 (2000),
modified].
[11] Coded Identification of Equipment Entities of the North American Telecommunications
System for Information Exchange [Revision of T1.213-1990 (R1996)], May 2001.
3 Definitions and abbreviations
3.1 Definitions
This Recommendation defines the following terms:
3.1.1 bit-error ratio: The ratio of the number of bits in error to the number of bits sent over a
period of time.
3.1.2 downstream: STU-C to STU-R direction (central office to remote terminal).
3.1.3 loopback: A reversal in the direction of the payload (i.e., the user data) at a specified
SHDSL network element.
3.1.4 mapper: A device for associating a grouping of bits with a transmission symbol.
3.1.5 micro-interruption: A temporary line interruption.
3.1.6 modulo: A device having limited value outputs (not the same as the mathematical modulo
operation).
3.1.7 payload block: One of the sections of a frame containing user data.
3.1.8 plesiochronous: A clocking scheme in which the SHDSL frame is based on the input
transmit clock but the symbol clock is based on another independent clock source.
3.1.9 precoder: A device in the transmitter for equalizing some of the channel impairments.
3.1.10 precoder coefficients: Coefficients of the filter in the precoder that are generated in the
receiver and transferred to the transmitter.
3.1.11 remote terminal: A terminal located downstream from a central office switching system.
3.1.12 scrambler: A device to randomize a data stream.
3.1.13 segment: The portion of a span between two terminations (either STUs or SRUs).
3.1.14 SHDSL network element: An STU-R, STU-C or SRU.
3.1.15 span: The link between STU-C and STU-R, including regenerators.
3.1.16 spectral shaper: A device that reshapes the frequency characteristics of a signal.
3.1.17 stuff bits: Bits added to synchronize independent data streams.
3.1.18 synchronous: A clocking scheme in which the SHDSL frame and symbol clocks are based
on the STU-C input transmit clock or a related network timing source.
2 ITU-T Rec. G.991.2 (12/2003)

3.1.19 upstream: STU-R to STU-C direction (remote terminal to central office).
3.2 Abbreviations
This Recommendation uses the following abbreviations:
α The interface between the PMS-TC and TPS-TC layers in an STU-C
ß The interface between the PMS-TC and TPS-TC layers in an STU-R
γC The interface between the TPS-TC layer and the application specific section in an
STU-C
γR The interface between the TPS-TC layer and the application specific section in an
STU-R
ak Convolutional Encoder Coefficients
AFE Analogue Front End
AGC Automatic Gain Control
bk Convolutional Encoder Coefficients
BER Bit Error Ratio
bit/s Bits per Second
Ck The kth Precoder Coefficient
CLEITM Common Language Equipment Identifier
CMRR Common Mode Rejection Ratio
CO Central Office
CPE Customer Premises Equipment
CRC Cyclic Redundancy Check
CRC-6 CRC of Order 6 (used in SHDSL frame)
crc(X) CRC Check Polynomial
DAC Digital-to-Analogue Converter
dBm dB reference to 1 mW, i.e., 0 dBm = 1 mW
DC Direct Current
DLL Digital Local Line
DRR Dynamic Rate Repartitioning
DS Downstream
DSC Dedicated Signalling Channel
DSL Digital Subscriber Line
DUT Device Under Test
EOC Embedded Operations Channel
ES Errored Second
fs Sampling rate
fsym Symbol rate
FCS Frame Check Sequence
ITU-T Rec. G.991.2 (12/2003) 3

FEC Forward Error Correction
FEXT Far-End CrossTalk
FSW Frame Synchronization Word
g(X) Generating Polynomial for CRC
HDLC High-level Data Link Control
HW Hardware
I/F Interface
kbit/s Kilobits per second
LB Longitudinal Balance
LCL Longitudinal Conversion Loss
losd Bit indicating Loss of signal at the application interface
LOSW Loss Of Sync Word failure
LSB Least Significant Bit
LT Line Termination
m(X) Message Polynomial for CRC
Mbit/s Megabits per second
MSB Most Significant Bit
MTU Maintenance Termination Unit
NEXT Near-End CrossTalk
NT Network Termination
OAM Operations, Administration and Maintenance
OH Overhead
PAM Pulse Amplitude Modulation
2-PAM PAM having two levels (used at startup)
PBO Power Back-Off
PL-OAM Physical Layer – OAM
PMD Physical Medium Dependent
PMMS Power Measurement Modulation Session (Line Probe)
PMS-TC Physical Medium-Specific TC Layer
ppm Parts Per Million
PPP Point-to-Point Protocol
ps Power status bit
PSD Power Spectral Density
PTD Path Terminating Device (CO side terminating equipment)
PTM Packet Transfer Mode
REG Signal Regenerator
rms Root mean square
4 ITU-T Rec. G.991.2 (12/2003)

RSP Regenerator Silent Period bit
RX Receiver
S/T Logical interface between the STU-R and attached user terminal equipment
sb stuff bit
sbid stuff bit identified indicator bit
sega segment anomaly indicator bit
segd segment defect indicator bit
SES Severely Errored Second
SHDSL Single-Pair High-Speed DSL
SNR Signal-to-Noise Ratio
SRU SHDSL Regenerator Unit
STU SHDSL Transceiver Unit
STU-C STU at the Central Office
STU-R STU at the Remote End
TBD To Be Determined
TC Transmission Convergence layer
TCM Trellis Coded Modulation
TCM-ISDN Time-Compression Multiplexed ISDN (specified in Appendix III/G.961 [B1])
TCPAM Trellis Coded PAM (used in data mode)
TPS-TC Transmission Protocol-Specific TC Layer
TX Transmitter
U-C Loop Interface – Central Office end
U-R Loop Interface – Remote Terminal end
UAS Unavailable Second
US Upstream
UTC Unable to Comply
V Logical interface between STU-C and a digital network element such as one or more
switching systems
xDSL a collective term referring to any of the various types of DSL technologies
ITU-T Rec. G.991.2 (12/2003) 5

4 Reference models
4.1 STU-x functional model
Figure 4-1/G.991.2 – STU-x functional model
Figure 4-1 is a block diagram of an SHDSL Transceiver Unit (STU) transmitter showing the
functional blocks and interfaces that are referenced in this Recommendation. It illustrates the basic
functionality of the STU-R and the STU-C. Each STU contains both an application invariant section
and an application specific section. The application invariant section consists of the PMD and
PMS-TC layers, while the application specific aspects are confined to the TPS-TC layer and device
interfaces. As shown in the figure, one or more optional signal regenerators may also be included in
an SHDSL span. Management functions, which are typically controlled by the operator's network
management system, are not shown in the figure. See clause 9 for details on management. Remote
power feeding, which is optionally provided across the span by the STU-C, is not illustrated in the
figure.
The functions at the central office side constitute the STU-C (or Line Termination (LT)). The
STU-C acts as the master both to the customer side functions of the STU-R (or Network
Termination (NT)) and to any regenerators.
The STU-C and STU-R, along with the DLL (Digital Local Line) and any regenerators, make up an
SHDSL span. The DLL may consist of a single copper twisted pair, or, in optional configurations,
multiple copper twisted pairs. In the multi-pair cases, each STU contains multiple separate PMD
layers, interfacing to a common PMS-TC layer. If enhanced transmission range is required, one or
more signal regenerators may be inserted into the loop at intermediate points. These points shall be
chosen to meet applicable criteria for insertion loss and loop transmission characteristics.
The principal functions of the PMD layer are:
• symbol timing generation and recovery;
• coding and decoding;
• modulation and demodulation;
• echo cancellation;
• line equalization;
• link startup.
The PMD layer functionality is described in detail in clause 6.
6 ITU-T Rec. G.991.2 (12/2003)

The PMS-TC layer contains the framing and frame synchronization functions, as well as the
scrambler and descrambler. The PMS-TC layer is described in clause 7.
The PMS-TC is connected across the α and ß interfaces in the STU-C and the STU-R, respectively,
to the TPS-TC layer. The TPS-TC is application specific and consists largely of the packaging of
user data within the SHDSL frame. See clause 8 for details. This may include multiplexing,
demultiplexing, and timing alignment of multiple user data channels. Supported TPS-TC user data
framing formats are described in Annex E.
The TPS-TC layer communicates with the Interface blocks across the γR and γC interfaces.
Depending upon the specific application, the TPS-TC layer may be required to support one or more
channels of user data and associated interfaces. The definition of these interfaces is beyond the
scope of this Recommendation.
Note that the α, ß, γR and γC interfaces are only intended as logical separations and need not be
physically accessible.
4.2 User plane protocol reference model

NT1, NT2/1 LT
STU-R STU-C
Transport protocol (e.g., SDH) Transport protocol (e.g., SDH)
TPS-TC TPS-TC
Not Not
specified PMS-TC PMS-TC specified
PMD PMD
User data interface Physical transmission media Internal interface
S/T U LT internal
interface
Figure 4-2/G.991.2 – User plane protocol reference model
The User Plane Protocol Reference Model, shown in Figure 4-2, is an alternate representation of the
information shown in Figure 4-1. This figure is included to emphasize the layered nature of this
Recommendation and to provide a view that is consistent with the generic xDSL models shown
in ITU-T Rec. G.995.1 [B2].
4.3 Application models
Figure 4-3/G.991.2 – Application model
ITU-T Rec. G.991.2 (12/2003) 7

Figure 4-3 is an application model for a typical SHDSL system, showing reference points and
attached equipment. In such an application, an STU-R will typically connect to one or more user
terminals, which may include data terminals, telecommunications equipment, or other devices.
These connections to these pieces of terminal equipment are designated S/T reference points. The
connection between STU-R and STU-C may optionally contain one or more SHDSL signal
regenerators (SRUs). The connections to the DLLs that interconnect STUs and SRUs are designated
U reference points. For each STU-x and SRU, the Network side connection is termed the U-R
interface and the Customer side connection is termed the U-C interface. The STU-C typically
connects to a Central Office network at the V reference point.
5 Transport capacity
This Recommendation specifies a two-wire operational mode for SHDSL transceivers that is
capable of supporting user (payload) data rates from 192 kbit/s to 2.312 Mbit/s in increments of
8 kbit/s. The allowed rates are given by n × 64 + i × 8 kbit/s, where 3 ≤ n ≤ 36 and 0 ≤ i ≤ 7. For
n = 36, i is restricted to the values of 0 or 1. See Annexes A and B for details of specific regional
requirements. Note that optional extensions described in Annex F allow user data rates up to
5696 bit/s.
This Recommendation also specifies an optional M-pair operational mode that is capable of
supporting user (payload) data rates from M × 192 kbit/s to M × 2.312 Mbit/s in increments of
M × 8 kbit/s, where 1 ≤ M ≤ 4. Note that optional extensions described in Annex F allow user data
rates up to M × 5696 kbit/s. Four-wire mode is identical to M-pair mode with M = 2, except for the
method of assigning ordinal numbers to the wire pairs. In four-wire mode the ordinal numbers (the
wire pair identification number) are assigned as described in 6.3, while in M-pair mode the ordinal
numbers are assigned to wire pairs as described in 7.2.1.5. Again, see Annexes A and B for details
of specific regional requirements and Annex F for extended data rates.
6 PMD layer functional characteristics
6.1 Data mode operation

6.1.1 STU data mode PMD reference model
A reference model of the data mode PMD layer of an STU-C or STU-R transmitter is shown in
Figure 6-1.
TCM Spectral
Scrambler Precoder shaper
encoder output at
loop
interface
Figure 6-1/G.991.2 – Data mode PMD reference model
The time index n represents the bit time, the time index m represents the symbol time, and t
represents analogue time. The input from the framer is f(n), and s(n) is the output of the scrambler.
Both the framer and the scrambler are contained within the PMS-TC layer and are shown here for
clarity. x(m) is the output of the TCM (Trellis Coded Modulation) encoder, y(m) is the output of the
channel precoder, and z(t) is the analogue output of the spectral shaper at the loop interface. When
transferring K information bits per one-dimensional PAM symbol, the symbol duration is K times
the bit duration, so the K values of n for a given value of m are {mK + 0 , mK + 1, … , mK + K – 1}.
8 ITU-T Rec. G.991.2 (12/2003)

In the optional M-pair mode, M separate PMD sublayers are active – one for each wire pair. In this
case, n represents the bit time for each wire pair rather than the aggregate system line rate.
6.1.1.1 PMD rates
The operation of the PMD layer at the specified information rate shall be as specified in A.5.1 or
B.5.1. The operation of the PMD layer at the optional extended rates specified in Annex F shall be
as specified in F.2.
6.1.2 TCM encoder
The block diagram of the TCM encoder is shown in Figure 6-2. The serial bit stream from the
scrambler, s(n), shall be converted to a K-bit parallel word at the mth symbol time, then processed
by the convolutional encoder. The resulting K + 1-bit word shall be mapped to one of 2K + 1
pre-determined levels forming x(m).
.
.
.
Serial to
parallel Mapper
Convolutional
encoder
Figure 6-2/G.991.2 – Block diagram of the TCM encoder
6.1.2.1 Serial-to-parallel converter

The serial bit stream from the scrambler, s(n), shall be converted to a K-bit parallel word
{X1(m) = s(mK + 0), X2(m) = s(mK + 1), … , XK(m) = s(mK + K – 1)} at the mth symbol time, where
X1(m) is the first input bit in time.
6.1.2.2 Convolutional encoder
Figure 6-3 shows the feedforward non-systematic convolutional encoder, where Ts is a delay of one
symbol time, "⊕" is binary exclusive-OR, and "⊗" is binary AND. X1(m) shall be applied to the
convolutional encoder, Y1(m) and Y0(m) shall be computed, then X1(m) shall be shifted into the shift
register.
ITU-T Rec. G.991.2 (12/2003) 9

Figure 6-3/G.991.2 – Block diagram of the convolutional encoder
The binary coefficients ai and bi shall be passed to the encoder from the receiver during the
activation phase specified in 7.2.1.3. A numerical representation of these coefficients is A and B
where:
A = a20 ⋅ 220 + a19 ⋅ 219 + a18 ⋅ 218 + ... + a0 ⋅ 20; and
B = b20 ⋅ 220 + b19 ⋅ 219 + b18 ⋅ 218 + ... + b0 ⋅ 20
The choice of encoder coefficients is vendor specific. They shall be chosen such that the system
performance requirements are satisfied (see Annex A and/or Annex B for performance
requirements).
6.1.2.3 Mapper
The K + 1 bits YK(m), … , Y1(m), and Y0(m) shall be mapped to a level x(m). Table 6-1 gives the bit
to level mapping for 16-level mapping.
Table 6-1/G.991.2 – Mapping of bits to PAM levels

Y3(m) Y2(m) Y1(m) Y0(m) x(m) for 16-PAM
0 0 0 0 −15/16
0 0 0 1 −13/16
0 0 1 0 −11/16
0 0 1 1 −9/16
0 1 0 0 −7/16
0 1 0 1 −5/16
0 1 1 0 −3/16
0 1 1 1 −1/16
1 1 0 0 1/16
1 1 0 1 3/16
1 1 1 0 5/16
1 1 1 1 7/16
1 0 0 0 9/16
1 0 0 1 11/16
1 0 1 0 13/16
1 0 1 1 15/16
10 ITU-T Rec. G.991.2 (12/2003)

6.1.3 Channel precoder
The block diagram of channel precoder is shown in Figure 6-4, where Ts is a delay of one symbol
time.
Modulo
–
Ts Ts Ts Ts
Precoder filter
Figure 6-4/G.991.2 – Block diagram of the channel precoder
The coefficients of the precoder filter, Ck, shall be transferred to the channel precoder as described
in 7.2.1.2. The output of the precoder filter, v(m), shall be computed as follows:
N
v ( m) = ∑ Ck y ( m − k )
k =1
Where 128 ≤ N ≤ 180. The function of the modulo block shall be to determine y(m) as follows: for
each value of u(m), find an integer, d(m), such that:
− 1 ≤ u ( m) + 2d ( m) < 1
and then
y ( m) = u ( m) + 2d ( m)
6.1.4 Spectral shaper

The choice of spectral shape shall be region-specific. The details of PSDs for Regions A and B are
given in A.4 and B.4. The details of PSDs for the optional extended rates in Annex F are given
in F.4.
6.1.5 Power backoff
SHDSL devices shall implement Power Backoff, as specified in this clause. The selected power
backoff value shall be communicated during pre-activation through the use of G.994.1 parameter
selections.
The power backoff value shall be selected to meet the requirements shown in Table 6-2. The power
backoff calculations are based on Estimated Power Loss (EPL), which is defined as:
Estimated Power Loss (dB) = TX Power (dBm) – Estimated RX Power (dBm),
evaluated for the data mode PSD.
No explicit specification is given herein for the method of calculating Estimated RX Power.
Depending upon the application, this value may be determined based on line probe results, a priori
knowledge, or G.994.1 tone levels.
ITU-T Rec. G.991.2 (12/2003) 11

The Power Backoff that is applied shall be no less than the Default Power Backoff, and it shall not
exceed the Maximum Power Backoff Value.
Table 6-2/G.991.2 – Required power backoff values

Estimated power loss (dB) Maximum power backoff (dB) Default power backoff (dB)
EPL > 6 31 0
6 ≥ EPL > 5 31 1
5 ≥ EPL > 4 31 2
4 ≥ EPL > 3 31 3
3 ≥ EPL > 2 31 4
2 ≥ EPL > 1 31 5
1 ≥ EPL > 0 31 6
6.2 PMD activation sequence

This clause describes waveforms at the loop interface and associated procedures during activation
mode. The direct specification of the performance of individual receiver elements is avoided when
possible. Instead, the transmitter characteristics are specified on an individual basis and the receiver
performance is specified on a general basis as the aggregate performance of all receiver elements.
Exceptions are made for cases where the performance of an individual receiver element is crucial to
interoperability. In 6.2.2, "convergence" refers to the state where all adaptive elements have reached
steady-state. The declaration of convergence by a transceiver is therefore vendor dependent.
Nevertheless, actions based on the state of convergence are specified to improve interoperability.
6.2.1 PMD activation reference model
The reference model of the activation mode of an STU-C or STU-R transmitter is shown in
Figure 6-5.
Figure 6-5/G.991.2 – Activation reference model
The time index m represents the symbol time, and t represents analogue time. Startup uses 2-PAM
modulation, so the bit time is equivalent to the symbol time. The output of the activation framer is
f(m), the framed information bits. The output of the scrambler is s(m). Both the framer and the
scrambler are contained within the PMS-TC layer and are shown here only for clarity. The output of
the mapper is y(m), and the output of the spectral shaper at the loop interface is z(t). d(m) is an
initialization signal that shall be logical ones for all m. The modulation format shall be uncoded
2-PAM, at the symbol rate selected for data mode operation.
In devices supporting the optional M-pair mode, the core activation procedure shall be considered
as an independent procedure for each pair. Such devices shall be capable of detecting the
completion of activation for all pairs and upon completion shall initiate the transmission of user
data over all pairs.
12 ITU-T Rec. G.991.2 (12/2003)

6.2.2 PMD activation sequence description
The timing diagram for the activation sequence is given in Figure 6-6. The state transition diagram
for the startup sequence is given in Figure 6-7. Each signal in the activation sequence shall satisfy
the tolerance values listed in Table 6-3.
tAct
Tc Datac
STU-C
tcrsc
tcr
STU-R
Cr Tr Datar
Figure 6-6/G.991.2 – Timing diagram for activation sequence

Figure 6-6a shows the total activation sequence at a high level for G.991.2, which includes
pre-activation and core activation. Included as an example in the pre-activation phase are two
sessions of handshake per ITU-T Rec. G.994.1 and line probe.
Figure 6-6a/G.991.2 – G.991.2 total activation sequence

The global activation time is the sum of the pre-activation and core activation times. Therefore,
from Figure 6-6a,
t pre _ activation + tcore _ activation ≤ tact _ global
where tpre_activation is the combined duration of the G.994.1 sessions (see 6.4) and line probing
(see 6.3), tcore_activation is the core activation duration (see 6.2). The values for tact and tact_global are
defined in Table 6-3. The value for tp-total is given in Table 6-5.
Table 6-3/G.991.2 – Timing for activation signals
Time Parameter Reference Nominal value Tolerance
a)
tcr Duration of Cr 6.2.2.1 1×βs ±20 ms
tcrsc Time from end of Cr to beginning of Sc 6.2.2.2 500 ms ±20 ms
a)
tcrsr Time from end of Cr to beginning of Sr 6.2.2.3 1.5 × β s ±20 ms
a)
tact Maximum time from start of Cr to Datar 15 × β s
tpayloadValid Maximum time from start of Datac or 1s
Datar to valid SHDSL payload data
tsilence Minimum silence time from exception 2s
ITU-T Rec. G.991.2 (12/2003) 13

Table 6-3/G.991.2 – Timing for activation signals
Time Parameter Reference Nominal value Tolerance
condition to start of train
tPLL Maximum time from start of Sc to 5s
STU-R PLL lock
tact_global Time from start of initial pre-activation 30 s
session (6.3) to Datarb)
a)
β is dependent on bit rate. β = 1 for n > 12, β = 2 for n ≤ 12, where n is defined in clause 5.
b)
In the majority of the cases, tact_global will be less than 30 seconds. However, since the definition
of the handshake mechanism in ITU-T Rec. G.994.1 is outside the scope of this
Recommendation, a maximum value tact_global cannot be assured.
Figure 6-7/G.991.2 – STU-C and STU-R transmitter activation state transition diagram
6.2.2.1 Signal Cr
After exiting the pre-activation sequence (per ITU-T Rec. G.994.1 [2], see 6.3 for details), the
STU-R shall send Cr. Waveform Cr shall be generated by connecting the signal d(m) to the input of
the STU-R scrambler as shown in Figure 6-5. The PSD mask for Cr shall be the upstream PSD
mask, as negotiated during pre-activation sequence. Cr shall have a duration of tcr and shall be sent
0.3 s after the end of pre-activation.
14 ITU-T Rec. G.991.2 (12/2003)

NOTE – The end of pre-activation can be defined in two ways according to ITU-T Rec. G.994.1. For the
purpose of this Recommendation, the end of pre-activation will be from the end of the ACK(1) message
transmission plus the required timers. The minimum and maximum values of those timers are 0.04 and 1.0
second. Therefore, the total time between the end of the ACK(1) message and the beginning of Cr should be
between 0.34 and 1.3 seconds.
6.2.2.2 Signal Sc
After detecting Cr, the STU-C shall send Sc. Waveform Sc shall be generated by connecting the
signal d(m) to the input of the STU-C scrambler as shown in Figure 6-5. The PSD mask for Sc shall
be the downstream PSD mask, as negotiated during pre-activation sequence. Sc shall be sent tcrsc
after the end of Cr. If the STU-C does not converge while Sc is transmitted, it shall enter the
exception state (6.2.2.8).
6.2.2.3 Signal Sr
The STU-R shall send Sr, beginning tcrsr after the end of Cr. Waveform Sr shall be generated by
connecting the signal d(m) to the input of the STU-R scrambler as shown in Figure 6-5. The PSD
mask for Sr shall be the same as for Cr. If the STU-R does not converge and detect Tc while Sr is
transmitted, it shall enter the exception state (6.2.2.8). The method used to detect Tc is vendor
dependent. In timing modes supporting loop timing, waveform Sr and all subsequent signals
transmitted from the STU-R shall be loop timed, i.e., the STU-R symbol clock shall be locked to the
STU-C symbol clock.
6.2.2.4 Signal Tc
Once the STU-C has converged and has been sending Sc for at least tPLL (Table 6-3), it shall send
Tc. Waveform Tc contains the precoder coefficients and other system information. Tc shall be
generated by connecting the signal f(m) to the input of the STU-C scrambler as shown in Figure 6-5.
The PSD mask for Tc shall be the same as for Sc. The signal f(m) is the activation frame information
as described in 7.2.1. If the STU-C does not detect Tr while sending Tc, it shall enter the exception
state (6.2.2.8). The method used to detect Tr is vendor dependent.
6.2.2.5 Signal Tr
Once the STU-R has converged and has detected the Tc signal, it shall send Tr. Waveform Tr
contains the precoder coefficients and other system information. Tr shall be generated by connecting
the signal f(m) to the input of the STU-R scrambler as shown in Figure 6-5. The PSD mask for Tr
shall be the same as for Cr. The signal f(m) is the activation frame information as described in 7.2.1.
If the STU-R does not detect Fc while sending Tr, it shall enter the exception state (6.2.2.8). The
method used to detect Fc is vendor dependent.
6.2.2.6 Signal Fc
Once the STU-C has detected Tr and completed sending the current Tc frame, then it shall send Fc.
The first bit of the first Fc frame shall follow contiguously the last bit of the last Tc frame. Signal Fc
shall be generated by connecting the signal f(m) to the input of the STU-C scrambler as shown in
Figure 6-5. The PSD mask for Fc shall be the same as for Sc. The signal f(m) is the activation frame
information as described in 7.2.1 with the following exceptions: the frame sync word shall be
reversed in time, and the payload information bits shall be set to arbitrary values. The CRC shall be
calculated on this arbitrary-valued payload. The signal Fc shall be transmitted for exactly two
activation frames. As soon as the first bit of Fc is transmitted, the payload data in the Tr signal shall
be ignored.
6.2.2.7 Datac and Datar
Within 200 symbols after the end of the second frame of Fc, the STU-C shall enter data mode and
send Datac, and the STU-R shall enter data mode and send Datar. These TCPAM signals are
described in 6.1. The PSD mask for Datar and for Datac shall be according to A.4 or B.4, as
ITU-T Rec. G.991.2 (12/2003) 15

negotiated during the pre-activation sequence. There is no required relationship between the end of
the activation frame and any bit within the SHDSL data-mode frame. tpayloadValid (Table 6-3) after the
end of Fc, the SHDSL payload data shall be valid at the α or ß interface.
6.2.2.8 Exception state
If activation is not achieved within tact (Table 6-3) or if any exception condition occurs, then the
exception state shall be invoked. During the exception state the STU shall be silent for at least tsilence
(Table 6-3), then wait for transmission from the far end to cease, then return to the corresponding
initial startup state; the STU-R and STU-C shall begin pre-activation, as per 6.3.
6.2.2.9 Exception condition
An exception condition shall be declared during activation if any of the timeouts in Table 6-3 expire
or if any vendor-defined abnormal event occurs. An exception condition shall be declared during
data mode if a vendor-defined abnormal event occurs. A vendor-defined abnormal event shall be
defined as any event that requires a loop restart for recovery.
6.2.3 Framer and scrambler
The activation mode framer and scrambler are described in 7.2.
6.2.4 Mapper
The output bits from the scrambler, s(m), shall be mapped to the output level, y(m), as follows:
Table 6-4/G.991.2 – Bit-to-level mapping

Scrambler output s(m) Mapper output level, y(m) Data mode index
0 −9/16 0011
1 +9/16 1000
These levels, corresponding to the scrambler outputs 0 and 1, shall be identical to the levels in the
16-TCPAM constellation (Table 6-1) corresponding to indexes 0011 and 1000, respectively.
The same spectral shaper shall be used for data mode and activation mode as described in A.4
or B.4. For the optional extended rates in Annex F, the same spectral shaper shall be used for data
mode and activation mode as described in F.4.
6.2.6 Timeouts
Table 6-3 shows the system timeouts and their values. tact shall be the maximum time from the start
of Cr to the start of Datar. It controls the overall time of the train. tpayloadValid is the time between the
start of data mode and the instant at which the SHDSL payload data is valid (this accounts for
settling time, data flushing, frame synchronization, etc). tsilence shall be the minimum time in the
exception state in which the STU-C or STU-R is silent before returning to pre-activation (per ITU-T
Rec. G.994.1 [2], see 6.3 for details). tPLL shall be the time allocated for the STU-R to pull in the
STU-C timing. The STU-C shall transmit Sc for at least tPLL.
6.3 PMD pre-activation sequence

This clause describes waveforms at the loop interface and associated procedures during
pre-activation mode. The direct specification of the performance of individual receiver elements is
avoided when possible. Instead, the transmitter characteristics are specified on an individual basis
and the receiver performance is specified on a general basis as the aggregate performance of all
receiver elements. Exceptions are made for cases where the performance of an individual receiver
element is crucial to interoperability.
16 ITU-T Rec. G.991.2 (12/2003)

In the optional 4-wire mode, Pair 1 and Pair 2 shall be determined during the pre-activation
sequence per the procedures defined in Annex B/G.994.1 entitled, "Operation over multiple wire
pairs". Pair 1 shall be defined as the pair on which the final G.994.1 transaction is conducted.
Four-wire mode is identical to M-pair mode with M = 2, except for the method of assigning ordinal
numbers to the wire pairs. In the optional M-pair mode, the ordering of wire pairs shall be
determined as per 7.2.1.5.
6.3.1 PMD pre-activation reference model
The reference model of the pre-activation mode of an STU-C or STU-R transmitter is shown in
Figure 6-8.
Figure 6-8/G.991.2 – Pre-activation reference model
The time index m represents the symbol time, and t represents analogue time. Since the probe signal
uses 2-PAM modulation, the bit time is equivalent to the symbol time. The output of the scrambler
is s(m). The scrambler used in the PMD pre-activation may differ from the PMS-TC scrambler used
in activation and data modes. See 6.3.3 for details of the pre-activation scrambler. The output of the
mapper is y(m), and the output of the spectral shaper at the loop interface is z(t). d(m) is an
initialization signal that shall be logical ones for all m. The probe modulation format shall be
uncoded 2-PAM, with the symbol rate, spectral shape, duration and power backoff selected by
ITU-T Rec. G.994.1. Probe results shall be exchanged by ITU-T Rec. G.994.1.
In the optional M-pair mode, the G.994.1 exchange shall follow the defined procedures for
multi-pair operation. In this case, Signals Pri and Pci, as described below, shall be sent in parallel on
all wire pairs.
6.3.2 PMD pre-activation sequence description
A typical timing diagram for the pre-activation sequence is given in Figure 6-9. Each signal in the
pre-activation sequence shall satisfy the tolerance values listed in Table 6-5.
STU-C
G.994.1 tps tph

tprd tprd tprd tpcd tpcd G.994.1
thp tps tps tprc
STU-R
Figure 6-9/G.991.2 – Typical timing diagram for pre-activation sequence
ITU-T Rec. G.991.2 (12/2003) 17

Table 6-5/G.991.2 – Timing for pre-activation signals (Note)
Time Parameter Nominal value Tolerance
thp Time from end of handshake to start 0.2 s ±10 ms
of remote probe
tprd Duration of remote probe Selectable from 50 ms to 3.1 s ±10 ms
tps Time separating two probe sequences 0.2 s ±10 ms
tprc Time separating last remote and first 0.2 s ±10 ms
central probe sequences
tpcd Duration of central probe Selectable from 50 ms to 3.1 s ±10 ms
tph Time from end of central probe to 0.2 s ±10 ms
start of handshake
tp-total Total probe duration, from end of the 10 s maximum
first G.994.1 session to the start of the
second G.994.1 session
NOTE – Tolerances are relative to the nominal or ideal value. They are not cumulative across the
pre-activation sequence.
6.3.2.1 Signal Pri

If the optional line probe is selected during the G.994.1 session (see ITU-T Rec. G.994.1 [2] for
details), the STU-R shall send the remote probe signal. The symbol rate for the remote probe signal
shall be negotiated during the G.994.1 session, and shall correspond to the symbol rate used during
activation for the specified data rate. If multiple remote probe symbol rates are negotiated during
the G.994.1 session, then multiple probe signals will be generated, starting with the lowest symbol
rate negotiated and ending with the highest symbol rate negotiated. If both symmetric and
asymmetric PSD probe signals are selected, the symmetric PSD probe signals shall be sent first, in
order of ascending symbol rate, followed by the asymmetric PSD probe signals in order of
ascending symbol rate. If symmetric PSD probe signals are selected from both Annexes A and F,
then the symmetric PSD probe signals from Annex A will be sent first, followed by the symmetric
PSD probe signals from Annex F, all in order of ascending symbol rate. If "transmit silence" is
negotiated, then a probe signal consisting of transmitted silence will precede all other probe signals.
Pri is the ith probe signal (corresponding to the ith symbol rate negotiated or silence). Waveform Pri
shall be generated by connecting the signal d(m) to the input of the STU-R scrambler as shown in
Figure 6-8. The PSD mask for Pri shall be the upstream PSD mask used for signal Cr at the same
symbol rate, and shall be selectable between the PSDs for activating at data rates of 192 kbit/s to
2304 kbit/s in steps of 64 kbit/s. Note that optional extensions described in Annex F allow the
selection of Pri masks corresponding to data rates up to 5696 kbit/s. Alternatively, waveform Pri can
be selected to transmit silence. The duration (tprd) and power backoff shall be the same for all Pri,
and shall be negotiated during the G.994.1 session. The duration shall be selectable between 50 ms
and 3.1 s in steps of 50 ms, and the power backoff shall be selectable between 0 dB and 15 dB in
steps of 1 dB. The probe signal power backoff can be selected using either the received G.994.1
signal power or a priori knowledge. If no information is available, implementors are encouraged to
select a probe power backoff of at least 6 dB. The first remote probe signal shall begin thp after the
end of the G.994.1 session. There shall be a tps second silent interval between successive remote
probe signals.
In the optional M-pair mode, Pri shall be sent in parallel on all wire pairs.
6.3.2.2 Signal Pci
The STU-C shall send the central probe signal tprc after the end of the last remote probe signal. The
symbol rate for the central probe signal shall be negotiated during the G.994.1 session, and shall
18 ITU-T Rec. G.991.2 (12/2003)

correspond to the symbol rate used during activation for the specified data rate. If multiple central
probe symbol rates are negotiated during the G.994.1 session, then multiple probe signals will be
generated, starting with the lowest symbol rate negotiated and ending with the highest symbol rate
negotiated. If both symmetric and asymmetric PSD probe signals are selected, the symmetric PSD
probe signals shall be sent first, in order of ascending symbol rate, followed by the asymmetric PSD
probe signals in order of ascending symbol rate. If symmetric PSD probe signals are selected from
both Annexes A and F, then the symmetric PSD probe signals from Annex A will be sent first,
followed by the symmetric PSD probe signals from Annex F, all in order of ascending symbol rate.
If "transmit silence" is negotiated, then a probe signal consisting of transmitted silence will precede
all other probe signals. Pci is the ith probe signal (corresponding to the ith symbol rate negotiated or
silence). Waveform Pci shall be generated by connecting the signal d(m) to the input of the STU-C
scrambler as shown in Figure 6-8. The PSD mask for Pci shall be the downstream PSD mask used
for signal Sc at the same symbol rate, and shall be selectable between the PSDs for activating at data
rates of 192 kbit/s to 2304 kbit/s in steps of 64 kbit/s. Note that optional extensions described in
Annex F allow the selection of Pci masks corresponding to data rates up to 5696 kbit/s.
Alternatively, waveform Pci can be selected to transmit silence. The duration (tpcd) and power
backoff shall be the same for all Pci, and shall be negotiated during the G.994.1 session. The
duration shall be selectable between 50 ms and 3.1 s in steps of 50 ms, and the power backoff shall
be selectable between 0 dB and 15 dB in steps of 1 dB. The probe signal power backoff can be
selected using either the received G.994.1 signal power or a priori knowledge. If no information is
available, implementors are encouraged to select a probe power backoff of at least 6 dB. There shall
be a tps silent interval between successive central probe signals, and there shall be a tph second silent
interval between the last central probe signal and the start of the following G.994.1 session.
In the optional M-pair mode, Pci shall be sent in parallel on all wire pairs.
6.3.3 Scrambler
The pre-activation mode scrambler shall have the same basic structure as the data mode scrambler,
but may employ a different scrambler polynomial. During the G.994.1 session, the scrambler
polynomial for the line probe sequence shall be selected by the receiver from the set of allowed
scrambler polynomials listed in Table 6-6. The transmitter shall support all the polynomials in
Table 6-6. During the line probe sequence, the transmit scrambler shall use the scrambler
polynomial selected by the receiver during the G.994.1 session. The scrambler shall be initialized to
all zeros.
Table 6-6/G.991.2 – Pre-activation scrambler polynomials

Polynomial
index STU-C polynomial STU-R polynomial
(i2, i1, i0)
000 s (m) = s (m − 5) ⊕ s (m − 23) ⊕ d (m) s (m) = s (m − 18) ⊕ s (m − 23) ⊕ d (m)
001 s (m) = s (m − 1) ⊕ d (m) s (m) = s (m − 1) ⊕ d (m)
101 s (m) = s (m − 2) ⊕ s (m − 3) s (m) = s (m − 4) ⊕ s (m − 5)
⊕ s (m − 4) ⊕ s (m − 8) ⊕ d (m) ⊕ s (m − 6) ⊕ s (m − 8) ⊕ d (m)
110 Reserved Reserved
111 Not Allowed Not Allowed
ITU-T Rec. G.991.2 (12/2003) 19

6.3.4 Mapper
The output bits from the scrambler, s(m), shall be mapped to the output level, y(m), as described
in 6.2.4.
The same spectral shaper shall be used for data mode and activation mode as described in 6.1.4.
6.3.6 PMMS target margin
PMMS target margin is used by the receiver to determine if a data rate can be supported with this
margin under current noise and/or reference worst-case noise specified in Annexes A and B. A data
rate may be included in the capabilities list resulting from line probe only if the estimated SNR
associated with that data rate minus the SNR required for BER = 10–7 is greater than or equal to
target margin in dB. If both worst-case target margin and current-condition target margin are
specified, then the capabilities exchanged shall be the intersection of data rates calculated using
each noise condition separately.
The use of negative target margins with respect to reference worst-case noise corresponds to
reference noise with fewer disturbers. This may be applicable when the number of disturbers is
known to be substantially fewer than specified by the reference worst-case noise. Use of negative
target margins with respect to current-conditions is not advised. Use of the current-condition target
margin mode may result in retrains if the noise environment changes significantly.
The negotiation of the target margins is done as follows:
The target margins to be used by both the STU-C and the STU-R for determining the supported data
rates are under the control of the STU-C. In the PMMS parameter exchange, the STU-C shall set
the upstream and downstream PMMS target margins to identical values. This does not imply that
the worst-case and current-conditions target margins are the same.
To determine which data rates the STU-C can support, the STU-C can choose to use the upstream
PMMS target margin transmitted by the STU-R in the PMMS parameter exchange, or the STU-C
may choose to use an alternative internal value for the PMMS target margins. The STU-R shall use
the downstream PMMS target margin parameters sent by the STU-C for determining which data
rates the STU-R can support.
This procedure is applicable to both the current-condition and the worst-case target margins.
6.4 G.994.1 pre-activation sequence

As noted in 6.3, ITU-T Rec. G.994.1 [2] shall be used to begin the pre-activation sequence. A
second G.994.1 sequence shall follow the pre-activation line probe, as described in that clause. The
G.994.1 protocol shall be the mechanism for exchanging capabilities and negotiating the operational
parameters for each SHDSL connection. The use of a line probe sequence, as described in 6.3, is
optional. If each STU has sufficient a priori knowledge of the line characteristics and the
capabilities of the other STU, either from a previous connection or from user programming, the line
probe sequence may be bypassed. In this case, the G.994.1 sequence will be followed by SHDSL
activation, as described in 6.2.
6.4.1 G.994.1 code point definitions
The following definitions shall be applied to the SHDSL parameters specified in ITU-T
Rec. G.994.1:
6.4.1.1 base data rate/PSD: These octets are used as follows:
– for PMMS, they indicate rates for line probing segments;
– for training, they indicate payload data rates;
20 ITU-T Rec. G.991.2 (12/2003)

Separate bits are provided for symmetric and asymmetric PSDs.
NOTE – In CLR, upstream training parameters indicate what data mode rates the STU-R is capable of
transmitting, and downstream training parameters indicate what data mode rates the STU-R is capable of
receiving. In CL, downstream training parameters indicate what data mode rates the STU-C is capable of
transmitting, and upstream training parameters indicate what data mode rates the STU-C is capable of
receiving. If optional line probe is used, the receiver training parameters will be further limited by the probe
results. If repeaters are used, the training parameters of the SRU-R will be further limited by the training
parameters of all downstream SRUs and the STU-R.
6.4.1.2 clock modes: Set to indicate clock mode, as defined in Table 10-1.
6.4.1.3 diagnostic mode: Set to indicate a diagnostic mode train (for use with SRUs).
6.4.1.4 DRR support: Indicates whether DRR is supported. See E.10.3.
6.4.1.5 four-wire: Set to indicate four-wire operation.
6.4.1.6 lead time: Indicates the lead time of DRR protocol responses, measured as a number of
SHDSL frames. The range of supported values is from 1 to 15. See E.10.3.5.
6.4.1.7 low latency: Set to indicate that low latency operation, as defined in 11.5, is required. If not
set, an STU may choose a higher latency encoding scheme.
6.4.1.8 M-pair count: Indicates the number of pairs used in the optional M-pair mode.
6.4.1.9 multiple-pair operation: Set to indicate M-pair mode. Four-wire mode is identical to
M-pair mode with M = 2, except for the method of assigning ordinal numbers to the wire pairs. In
four-wire mode, the ordinal numbers (i.e., the wire pair identification number) are assigned as
described in 6.3; while in M-pair mode, the ordinal numbers are assigned to wire pairs as described
in 7.2.1.5.
6.4.1.10 PBO: Power Backoff (in 1.0 dB increments).
6.4.1.11 PMMS duration: The length of each line probe (PMMS) segment (in 50 ms
increments).
6.4.1.12 PMMS mode: An indication that an STU (or SRU) is prepared to begin a PMMS
("Power Measurement Modulation Session", or Line Probe) using the associated parameters.
6.4.1.13 PMMS scrambler: The scrambler polynomial used during line probe (PMMS).
See 6.3.3.
6.4.1.14 PMMS target margin: If worst-case target margin is selected, target margin is relative
to reference worst-case crosstalk specified in Tables A.13 and B.14. If current-condition target
margin is selected, specified target margin is relative to noise measured during line probe. The 5-bit
target margin is specified by (bits 5-1 × 1.0 dB) – 10 dB. For example, 1011112 in the worst-case
PMMS target margin octet corresponds to 15 dB – 10 dB = 5 dB target margin relative to reference
worst-case noise.
If the capability for PMMS mode is indicated in a G.994.1 CLR/CL capabilities exchange, both
target margin octets shall be sent. The specific values for target margin shall be ignored during the
capabilities exchange, as all STUs (and SRUs) shall be capable of evaluating the results of PMMS
using both types of target margin.
6.4.1.15 Regenerator Silent Period (RSP): A bit used to force an STU or SRU into a 1-minute
silent interval to facilitate startup of spans including regenerators.
6.4.1.16 SRU: Set to indicate that the unit is a Signal Regenerator and not an STU.
6.4.1.17 stuff bits: Indicates the value that the upstream and downstream stb1 – stb4 bits shall
take on. See 7.1.2.7 for details.
ITU-T Rec. G.991.2 (12/2003) 21

6.4.1.18 sub-data rate: For symmetric PSDs, the data rate octets indicate the base data rate in
64 kbit/s increments (n × 64 kbit/s). The sub-data rate bits indicate additional 8 kbit/s increments
(i × 8 kbit/s) of data. The total payload data rate is set by: base data rate + sub-data rate. The
sub-data rate bits do not apply to the asymmetric 2.048 Mbit/s, and 2.304 Mbit/s PSDs (from
Annex B). For the asymmetric 768 or 776 kbit/s and asymmetric 1.536 or 1.544 Mbit/s PSDs (from
Annex A), the base data rate bits indicate 768 kbit/s or 1.536 Mbit/s, and the sub-data rate bits
for 0 and 8 kbit/s are valid for selecting the total payload data rate.
6.4.1.19 sync word: Indicates the value that the upstream and downstream sw1 – sw14 bits shall
take on. See 7.1.2.1 for details.
6.4.1.20 TPS-TC: The TPS-TC mode is selected from the set of modes specified in Annex E.
6.4.1.21 training mode: An indication that an STU (or SRU) is prepared to begin SHDSL
activation using the associated parameters.
6.4.1.22 warm-start enable: Set to indicate that warm-start is available. See Annex H.
6.4.2 G.994.1 tone support
SHDSL devices shall support half-duplex mode G.994.1 operation using the A4 carrier set from the
4 kHz signalling family. Manufacturers are encouraged to support additional carrier sets, the
4.3125 kHz signalling family, and full-duplex operation of G.994.1 to provide interoperable
handshake sequences with other types of DSL equipment.
6.4.3 G.994.1 transactions
If no a priori capabilities information is available to the STU-R, it should begin the G.994.1 session
by initiating Transaction C (CLR/CL). Otherwise, it may begin immediately with one of the mode
selection transactions (e.g., A or B). In this capabilities exchange (CLR/CL sequence), each unit
shall indicate the functions that it is currently capable of performing. This means that user options
that have been disabled shall not be indicated as capabilities of the unit. If a unit's capabilities
change due to user option settings or other causes, that unit shall cause a capabilities exchange to
occur during the next G.994.1 session.
If both the STU-R and STU-C indicate the capability for line probing and no a priori information
exists concerning the characteristics of the loop, the STU-R should initiate Transaction D
(MP/MS/Ack(1)) by sending an MP with the G.991.2 line probe mode selected. This MP message
shall include parameters for the downstream line probe sequence. The STU-C shall then issue a
corresponding MS message containing the upstream line probe parameters and an echo of the
downstream line probe parameters. Following an Ack(1) from the STU-R, the units shall exit
G.994.1 and enter the G.991.2 line probe mode, as described in 6.3. Following the completion of
line probing, the STU-C shall initiate a new G.994.1 session. The STU-R shall then initiate a
Transaction C (CLR/CL) capabilities exchange to indicate the results of the line probe. Each unit
shall, in this exchange, indicate the intersection of its capabilities and the capabilities of the loop, as
determined during the line probe sequence. The PBO octet shall be used to indicate the desired
received Power Backoff. Following this second capabilities exchange, the units may use any valid
transaction to select operational SHDSL parameters.
Following the selection of the G.991.2 parameter set, G.994.1 shall terminate and the SHDSL
activation sequence (6.2) shall begin.
6.4.4 Operation with signal regenerators
In general, SRUs will act as STUs during G.994.1, as described in 6.4.3. In some situations,
however, they are required to issue "Regenerator Silent Period" (via the G.994.1 RSP bit) mode
selections rather than selecting a G.991.2 operational mode, as described in Annex D and
Appendix II. The parameters that SRUs report during capabilities exchanges are also slightly
different. The advertised capabilities of an SRU-R shall be the intersection of its own capabilities
22 ITU-T Rec. G.991.2 (12/2003)

and those reported across the regenerator's internal interface as indicative of the capabilities of the
downstream units and line segments. The lone exception to this rule shall be the PBO octet, which
shall be considered as a local parameter for each segment.
7 PMS-TC layer functional characteristics
7.1 Data mode operation

7.1.1 Frame structure
Table 7-1 summarizes the SHDSL frame structure. Complete bit definitions may be found in 7.1.2.
The size of each payload block is defined as k bits, where k = 12 (i + n × 8). The payload data rate is
set by: n × 64 + i × 8 kbit/s, where 3 ≤ n ≤ 36 and 0 ≤ i ≤ 7. For n = 36, i is restricted to the values
of 0 or 1. Note that optional extensions described in Annex F allow values of n up to 89. The value
of i shall be negotiated during startup, and shall apply to all values of n. The selected value of i
applies to all values of n, will be negotiated during pre-activation, and does not include the 8 kbit/s
framing overhead.
In the optional M-pair mode, M separate PMS-TC sublayers are active – one for each wire pair. In
this case, the above formula represents the payload data rate for each pair rather than the aggregate
payload rate. All pairs shall operate at the same payload rate, and the transmitters for all pairs shall
maintain frame alignment within specified limits. In the STU-C, the symbol clocks for each pair
shall be derived from a common source. The maximum differential delay between the start of
STU-C frames shall be no greater than four (4) symbols at the line side of each SHDSL transmitter.
In the STU-R, symbol clocks may be derived from loop timing on each pair, so these clocks shall be
locked in frequency but shall have an arbitrary phase relationship. The maximum differential delay
between the start of STU-R frames shall be no greater than six (6) symbols at the line side of each
SHDSL transmitter.
Table 7-1/G.991.2 – SHDSL frame structure

Over-
Frame
Time head Name Description Notes
bit #
bit #
0 ms 1-14 1-14 sw1-sw14 Frame Sync Word
15 15 fbit1/losd Fixed Indicator bit #1 (Loss
of Signal)
16 16 fbit2/sega Fixed Indicator bit #2
(Segment Anomaly)
17 → -------- b1 Payload block #1
k + 16
k + 17 17 eoc01 EOC bit #1
k + 18 18 eoc02 EOC bit #2
k + 19 19 eoc03 EOC bit #3
k + 20 20 eoc04 EOC bit #4
k + 21 21 crc1 Cyclic Redundancy Check #1 CRC-6
k + 22 22 crc2 Cyclic Redundancy Check #2 CRC-6
k + 23 23 fbit3/ps Fixed Indicator bit #3
(Power Status)
k + 24 24 sbid1 Stuff bit ID #1 Spare in
synchronous mode
ITU-T Rec. G.991.2 (12/2003) 23

Table 7-1/G.991.2 – SHDSL frame structure
Over-
Frame
Time head Name Description Notes
bit #
bit #
k + 25 25 eoc05 EOC bit #5
k + 26 26 eoc06 EOC bit #6
k + 27 → -------- b2 Payload block #2
2k + 26
2k + 27 27 eoc07 EOC bit #7
2k + 28 28 eoc08 EOC bit #8
2k + 29 29 eoc09 EOC bit #9
2k + 30 30 eoc10 EOC bit #10
2k + 31 31 crc3 Cyclic Redundancy Check #3 CRC-6
2k + 33 33 fbit4/segd Fixed Indicator bit #4
(Segment Defect)
2k + 34 34 eoc11 EOC bit #11
2k + 35 35 eoc12 EOC bit #12
2k + 36 36 sbid2 Stuff bit ID #2 Spare in
synchronous mode
2k + 37 → -------- b3 Payload block #3
3k + 36
3k + 37 37 eoc13 EOC bit #13
3k + 38 38 eoc14 EOC bit #14
3k + 39 39 eoc15 EOC bit #15
3k + 40 40 eoc16 EOC bit #16
3k + 43 43 eoc17 EOC bit #17
3k + 44 44 eoc18 EOC bit #18
3k + 45 45 eoc19 EOC bit #19
3k + 46 46 eoc20 EOC bit #20
6 – 3/ 3k + 47 → -------- b4 Payload block #4
(k + 12) ms 4k + 46
4k + 47 47 stb1 Stuff bit #1 Vendor dependent
in synchronous
mode
6 ms 4k + 48 48 stb2 Stuff bit #2 Vendor dependent
nominal in synchronous
mode
4k + 49 49 stb3 Stuff bit #3 Not present in
synchronous mode
6 + 3/ 4k + 50 50 stb4 Stuff bit #4 Not present in
(k + 12) ms synchronous mode
24 ITU-T Rec. G.991.2 (12/2003)

7.1.2 Frame bit definitions
In Table 7-1, the bit sequence of the SHDSL frame (prior to scrambling at the transmit side and
after descrambling at the receive side) is presented. The frame structures are identical in both
upstream and downstream directions of transmission. Spare bits in either direction shall be set to 1.
The following frame bit definitions are used:
7.1.2.1 sw1 – sw14 (Frame Sync Word)
The frame synchronization word (FSW) enables SHDSL receivers to acquire frame alignment. The
FSW (bits sw1 – sw14) is present in every frame and is specified independently for the upstream
and downstream directions.
7.1.2.2 b1 – b4 (Payload Blocks)
Used to carry user data. The internal structure of the payload blocks is defined in 8.1.
7.1.2.3 eoc01 – eoc20 (Embedded Operations Channel)
20 bits (eoc01... eoc20) are provided as a separate maintenance channel. See 9.5 for details. In
M-pair mode, eoc01 – eoc20 on Pair 1 shall carry the primary EOC data. The corresponding Pair 2
to Pair M eoc bits shall be duplicates of the Pair 1 eoc bits.
7.1.2.4 crc1 – crc6 (Cyclic Redundancy Check code)
Six bits assigned to a cyclic redundancy check (CRC) code (see 7.1.3).
7.1.2.5 fbit1 – fbit4 (Fixed Indicator bits)
Used for the indication of time-critical framing information. Specific bit definitions are given
below.
7.1.2.5.1 fbit1 = losd (Loss of Signal)
Used to indicate the loss of signal from the application interface. Loss of Signal = 0, Normal = 1.
Definition of the conditions causing the indication of losd is vendor specific and beyond the scope
of this Recommendation. In M-pair mode, losd on Pair 1 shall carry the primary losd indication.
The losd bit on all other pairs shall be duplicates of the Pair 1 bit.
7.1.2.5.2 fbit2 = sega (Segment Anomaly)
Used to indicate a CRC error on the incoming SHDSL frame. A segment anomaly indicates that a
regenerator operating on a segment has received corrupted data and therefore the regenerated data is
unreliable. The purpose of segment anomaly is to ensure internal performance monitoring integrity;
it is not intended to be reported to an external management entity. CRC Error = 0, Normal = 1.
7.1.2.5.2.1 STU operation
The STU shall set the sega bit to 1.
7.1.2.5.2.2 SRU operation
If a CRC error is declared for an incoming frame, an SRU shall set the sega bit to 0 in the next
available outgoing frame in the forward direction, i.e., in the direction of the data over which the
CRC error was observed. If no CRC error is declared, then an SRU shall pass the sega bit without
modification.
7.1.2.5.3 fbit3 = ps (Power Status)
The power status bit ps is used to indicate the status of the local power supply in the STU-R. The
power status bit is set to 1 if power is normal and to 0 if the power has failed. On loss of power at
the STU-R, there shall be enough power left to send the ps bit in at least one and preferably
ITU-T Rec. G.991.2 (12/2003) 25

three consecutive frames towards the STU-C. Note that, in the event of a power failure, the ps bit
should be set to 0 for as many frames as possible before deactivation. If the ps bit is set for less than
three frames, it is up to the application at the STU-C to determine the validity of the message.
Regenerators shall pass this bit transparently. In M-pair mode, ps on Pair 1 shall carry the primary
power status indication. The ps bit on all other pairs shall be duplicates of the Pair 1 ps bit.
7.1.2.5.4 fbit4 = segd (Segment Defect)
Used to indicate a loss of sync on the incoming SHDSL frame. A segment defect indicates that a
regenerator has lost synchronization and therefore the regenerated data is unavailable. This bit is
typically reported to an external management entity and is used to ensure timely protection
switching, alarm filtering, etc. Loss of Sync = 0, Normal = 1.
7.1.2.5.4.1 STU operation
The STU shall set the segd bit to 1.
7.1.2.5.4.2 SRU operation
If a LOSW-Defect is declared, an SRU shall set the segd bit to 0 in the next available outgoing
frame in the forward direction, i.e., in the direction of the data over which the LOSW-Defect was
observed. If no LOSW-Defect is declared, then an SRU shall pass the segd bit without modification.
7.1.2.6 sbid1, sbid2 (Stuff Indicator bits)
In plesiochronous mode, the stuff indicator bits indicate whether or not a stuffing event occurs in
the frame. Both bits shall be set to 1 if the four stuff bits are present at the end of the current frame.
Both bits shall be set to 0 if there are no stuff bits at the end of the current frame. In synchronous
mode, sbid1 and sbid2 are spare bits.
7.1.2.7 stb1 – stb4 (Stuffing Bits)
In plesiochronous mode, these bits are used together. Either zero or four stuffing bits are inserted,
depending on the relation of the timing between the upstream and downstream channels. In
synchronous framing mode, stb1 and stb2 are present in every frame, and stb3 and stb4 are not
present. The values of stb1 – stb4 are specified independently for the upstream and downstream
directions.
7.1.3 CRC Generation (crc1 … crc6)
A cyclic redundancy check (CRC) shall be generated for each frame and transmitted on the
following frame. The six CRC bits (crc1 to crc6) shall be the coefficients of the remainder
polynomial after the message polynomial, multiplied by D6, is divided by the generating
polynomial. The message polynomial shall consist of all bits in the frame except for the
synchronization word, CRC bits, and the stuff bits. (There are thus 4k + 26 message bits in a frame
that are covered by the CRC check.) The message bits shall be ordered as in the frame itself, i.e., m0
is the first bit, m1 is the second bit, etc. The CRC check bits shall be calculated according to the
equation:
crc(D) = m(D)D6 mod g(D)
where:
m(D) = m0D4k + 25 ⊕ m1D4k + 24 ⊕ … ⊕ m4k + 24D ⊕ m4k + 25
is the message polynomial,
g(D) = D6 ⊕ D ⊕ 1
is the generating polynomial,
crc(D) = crc1D5 ⊕ crc2D4 ⊕ … ⊕ crc5D ⊕ crc6
26 ITU-T Rec. G.991.2 (12/2003)

is the CRC check polynomial, ⊕ indicates modulo-2 addition (exclusive OR), and D is the delay
operator.
7.1.4 Frame synchronization
In plesiochronous clocking mode, SHDSL uses a variable length PMS-TC frame and bit stuffing to
synchronize the PMS-TC frame rate with the incoming payload rate. Quick acquisition of frame
synchronization and the ability to maintain frame synchronization in the presence of errors are
important properties of the frame structure.
Three types of bit fields are provided for use in frame synchronization: Frame Sync Word, Stuff
Bits, and Stuff Bit IDs. The Frame Sync Word is 14 bits long and is present on every frame. The
stuff bits are four contiguous bits which are present only at the end of long frames. Stuff Bit IDs are
two bits distributed within the frame which indicate whether the current frame contains the four
stuffing bits. These distributed bits provide improved immunity to frame alignment errors caused by
burst errors.
The precise manner in which this information is used to acquire or maintain frame synchronization
is the choice of the receiver designer. Since different frame synchronization algorithms may require
different values for the bits of the FSW and Stuff Bits, a provision has been made to allow the
receiver to inform the far end transmitter of the particular values that are to be used for these fields
in the transmitted PMS-TC frame.
7.1.5 Scrambler
The scrambler in the STU-C and the STU-R transmitters shall operate as shown in Figures 7-1
and 7-2, respectively. In these figures, Tb indicates a delay of one bit duration and ⊕ is the binary
exclusive-OR operation. The frame sync word bits and the stuff bits in the SHDSL data mode frame
(Table 7-1) shall not be scrambled. While the frame sync word bits and stuff bits are present at f(n),
the scrambler shall not be clocked, and f(n) shall be directly connected to s(n).
7.1.5.1 STU-C scrambler
The block diagram of the STU-C scrambler is shown in Figure 7-1.
Figure 7-1/G.991.2 – Block diagram of the STU-C scrambler
ITU-T Rec. G.991.2 (12/2003) 27

7.1.5.2 STU-R scrambler
The block diagram of the STU-R scrambler is shown in Figure 7-2.
Figure 7-2/G.991.2 – Block diagram of the STU-R scrambler
7.1.6 Differential delay buffer

In the optional M-pair mode, it is understood that the characteristics of the M wire pairs may differ.
Differences in wire diameter, insulation type, length, number and length of bridged taps and
exposure to impairments may result in differences in transmission time between pairs. It is
recommended that such differences in signal transfer delay between any two pairs be limited to a
maximum of 50 µs at 150 kHz, corresponding to about 10 km difference in line length between
STU-R and STU-C.
In transceivers supporting M-pair mode, a delay difference buffer shall be implemented to
compensate for any difference in total transmission time of the SHDSL frames on different pairs.
Such delay differences may be due to the pair differences described above, as well as to delays due
to signal processing in the SHDSL transceivers in the STU-C, STU-R and possible signal
regenerators. The function of this delay difference buffer is to align the SHDSL frames so that
frames can be correctly reassembled. This buffer shall be capable of absorbing a delay difference of
at least 6 symbols + 50 µs at the line side of each SHDSL receiver.
7.2 PMS-TC activation

7.2.1 Activation frame
The format of the activation frame is shown in Table 7-2. A Tc or Tr signal shall be generated by
repetitively applying the activation frame information shown in Table 7-2 to the STU scrambler as
shown in Figure 6-5. The activation frame contents shall be constant during the transmission of Tc
and Tr. The activation frame sync bits are not scrambled, so they shall be applied directly to the
uncoded 2-PAM constellation. The total number of bits in the activation frame is 4227. The
activation frame shall be sent starting with bit 1 and ending with bit 4227.
In the optional M-pair mode, activation shall proceed in parallel on each of the M wire pairs.
28 ITU-T Rec. G.991.2 (12/2003)

Table 7-2/G.991.2 – Activation frame format
Activation
frame bit Definition
LSB:MSB
Frame Sync for Tc and Tr: 111110011010112, where the left-most bit is sent first in
1:14 time
Frame Sync for Fc: 110101100111112, where the left-most bit is sent first in time
Precoder Coefficient 1: 22 bit signed two's complement format with 17 bits after the
15:36
binary point, where the LSB is sent first in time
37:58 Precoder Coefficient 2
59:3952 Precoder Coefficients 3-179
3953:3974 Precoder Coefficient 180
3975:3995 Encoder Coefficient A: 21 bits where the LSB is sent first in time
3996:4016 Encoder Coefficient B: 21 bits where the LSB is sent first in time
4017:4144 Vendor Data: 128 bits of proprietary information
4145:4146 M-pair mode: STU-C: Number of wire pairs/STU-R: Ordering of wire pairs
4147:4211 Reserved: 65 bits set to logical zeros
4212:4227 CRC: c1 sent first in time, c16 sent last in time
7.2.1.1 Frame sync

The frame sync for Tc and Tr is a 14-bit code. In binary, the code shall be 11111001101011, and
shall be sent from left to right. For Fc, the frame sync shall be 11010110011111, or the reverse of
the frame sync for Tc and Tr.
7.2.1.2 Precoder coefficients
The precoder coefficients are represented as 22-bit two's complement numbers, with the five most
significant bits representing integer numbers from −16 (10000) to +15 (01111), and the remaining
17 bits are the fractional bits. The coefficients are sent sequentially, starting with coefficient C1 and
ending with coefficient CN (from Figure 6-4), and the least significant bit of each coefficient is sent
first in time. The minimum number of precoder coefficients shall be 128 and the maximum number
shall be 180. If fewer than 180 precoder coefficients are used, the remaining bits in the field shall be
set to zero.
7.2.1.3 Encoder coefficients
Referring to Figure 6-3, the coefficients for the programmable encoder are sent in the following
order: a0 is sent first in time, followed by a1, a2, … , and b20 is sent last in time.
7.2.1.4 Vendor data
These 128 bits are reserved for vendor-specific data.
7.2.1.5 M-pair mode: Ordering of wire pairs
In the optional M-pair mode these two bits are used to define the order of the M wire pairs. They are
used to determine how user data is split into M loops at the transmitter and combined in the receiver
as specified in 7.1.1. The assignment of loop 1 to loop M is vendor-specific.
Bits 4145 to 4146 in the activation frame of the STU-C device are used to specify the number M of
wire pairs. LSB first. M = 1: 002; M = 2: 102; M = 3: 012; M = 4: 112. This activation frame entry is
identical on all M wire pairs.
ITU-T Rec. G.991.2 (12/2003) 29

Bits 4145 to 4146 of the activation frame of the STU-R device are used to identify the ordinal
number of each of the M wire pairs. LSB first. Wire pair 1: 002; wire pair 2: 102; wire pair 3: 012;
wire pair 4: 112. This activation frame entry is different on each of the M wire pairs.
If the system is not operating in M-pair mode, these two bits shall be set to logical zeros. In four-
wire mode, the ordinal numbers are assigned as described in 6.3, and bits 4145 to 4146 of the
activation frame shall be set to zero.
7.2.1.6 Reserved
These 65 bits are reserved for future use and shall be set to logical zeros.
7.2.1.7 CRC
The sixteen CRC bits (c1 to c16) shall be the coefficients of the remainder polynomial after the
message polynomial, multiplied by D16, is divided by the generating polynomial. The message
polynomial shall be composed of the bits of the activation frame, where m0 is bit 15 and m4196 is bit
4211 of the activation frame, such that:
crc(D) = m(D)D16 mod g(D)
where:
m(D) = m0D4196 ⊕ m1D4195 ⊕ … ⊕ m4195D ⊕ m4196
is the message polynomial,
g(D) = D16 ⊕ D12 ⊕ D5 ⊕ 1
is the generating polynomial,
crc(D) = c1D15 ⊕ c2D14 ⊕ … ⊕ c15D ⊕ c16
is the CRC check polynomial, ⊕ indicates modulo-2 addition (exclusive OR), and D is the delay
operator.
7.2.2 Activation scrambler
The scrambler in the STU-C and the STU-R transmitters (see Figure 6-5) shall operate as shown in
Figures 7-1 and 7-2, where Tb is a delay of one bit duration, and ⊕ is binary exclusive-OR. The
frame sync bits in the activation frame shall not be scrambled. While the frame sync bits are present
at f(n), the scrambler shall not be clocked, and f(n) shall be directly connected to s(n).
8 TPS-TC layer functional characteristics
8.1 Payload block data structure

Each payload block shall consist of 12 Sub-blocks, and shown in Figure 8-1. The size of each
payload sub-block is defined as ks, where ks = i + n × 8 [bits]. As stated in 7.1, the payload data rate
is set by: n × 64 + i × 8 kbit/s, where 3 ≤ n ≤ 36 and 0 ≤ i ≤ 7. For n = 36, i is restricted to the values
of 0 or 1. Note that optional extensions described in Annex F allow values of n up to 89. All
structure of data within payload sub-blocks (i.e., support for clear broadband channels, subchannels,
and region-specific services) is specified in Annex E.
30 ITU-T Rec. G.991.2 (12/2003)

Figure 8-1/G.991.2 – Structure of payload blocks
8.2 Data interleaving in M-pair mode

In the optional M-pair mode, interleaving of payload data between pairs is necessary. This shall be
accomplished by interleaving within payload sub-blocks among all pairs. ks bits in each sub-block
shall be carried on Pair 1, and an additional ks bits shall be carried on each of the other pairs, as
shown in Figure 8-2 for the case of M = 2. The size of each payload sub-block is defined as M × ks,
where ks = i + n × 8. As stated in 7.1, the payload data rate per pair is set by: n × 64 + i × 8 kbit/s,
where 3 ≤ n ≤ 36 and 0 ≤ i ≤ 7. For n = 36, i is restricted to the values of 0 or 1. Note that optional
extensions described in Annex F allow values of n up to 89. All structure of data within payload
sub-blocks (i.e., support for clear broadband channels, subchannels, and region-specific services) is
specified in Annex E.
ITU-T Rec. G.991.2 (12/2003) 31

Figure 8-2/G.991.2 – Data interleaving within payload blocks
9 Management
9.1 Management reference model
Figure 9-1/G.991.2 – Management reference model
32 ITU-T Rec. G.991.2 (12/2003)

Figure 9-1 shows the Management Reference Model for user data transport over SHDSL. This
example includes two regenerator units for informative purposes. The presence of two regenerators
is not intended to be a requirement or limit. An SHDSL segment is characterized by a metallic
transmission medium utilizing an analogue coding algorithm, which provides both analogue and
digital performance monitoring at the segment entity. An SHDSL segment is delimited by its two
end points, known as segment terminations. An SHDSL segment termination is the point at which
the analogue coding algorithms end and the subsequent digital signal is monitored for integrity.
All SHDSL performance monitoring data is transported over the EOC. The fixed indicator bits in
the SHDSL frame are used for rapid communication of interface or SHDSL segment defects, which
may lead to protection switching. In addition, the fixed indicator bits may be used for rapid alarm
filtering SHDSL segment failures.
9.2 SHDSL performance primitives

9.2.1 Cyclical Redundancy Check Anomaly (CRC Anomaly)
A CRC anomaly shall be declared when the CRC bits generated locally on the data in the received
SHDSL frame do not match the CRC bits (crc1 – crc6) received from the transmitter. A CRC
anomaly only pertains to the frame over which it was declared.
9.2.2 Segment Anomaly (SEGA)
An upstream segment anomaly shall be declared when any SRU declares a CRC anomaly for an
SHDSL frame moving in the direction from STU-R to STU-C. A downstream segment anomaly
shall be declared when any SRU declares a CRC anomaly for an SHDSL frame moving in the
direction from STU-C to STU-R. A segment anomaly indicates that a regenerator operating on a
segment has received corrupted data and therefore the regenerated data is unreliable. The purpose of
segment anomaly is to ensure internal SHDSL PMD integrity; it is not intended to be reported to an
external management entity. A segment anomaly is indicated via the sega bit in the SHDSL frame
(7.1.2.5.2).
9.2.3 Loss of Sync Defect (LOSW defect)
In plesiochronous mode, an LOSW defect shall be declared when at least three consecutive received
frames contain one or more errors in the framing bits. The term framing bits shall refer to that
portion of Frame Sync Word, Stuff Bits and Stuff Bit Ids, which are used for frame
synchronization. An LOSW defect shall be cleared when at least two consecutive received frames
contain no errors in the framing bits.
In synchronous mode, an LOSW defect shall be declared when at least three consecutive received
frames contain one or more bit errors in the Frame Sync Word. An LOSW defect shall be cleared
when at least two consecutive received frames contain no errors in the Frame Sync Word.
9.2.4 Segment Defect (SEGD)
An upstream segment defect shall be declared when any SRU declares a LOSW defect for data
moving in the direction from STU-R to STU-C. A downstream segment defect shall be declared
when any SRU declares a LOSW defect for data moving in the direction from STU-C to STU-R.
A segment defect indicates that a regenerator has lost SHDSL synchronization and therefore the
regenerated data is unavailable. A segment defect shall be cleared when all SRUs have no LOSW
defects. This primitive is typically reported to an external management entity and is used to ensure
timely protection switching, alarm filtering, etc. A segment defect is indicated via the segd bit in the
SHDSL frame (7.1.2.5.4).
ITU-T Rec. G.991.2 (12/2003) 33

9.2.5 Loop attenuation defect
A Loop Attenuation Defect shall be declared when the observed Loop Attenuation is at a level
higher than the configured threshold (9.5.5.7.5).
9.2.6 SNR margin defect
An SNR Margin Defect shall be declared when the observed SNR Margin is at a level lower than
the configured threshold (9.5.5.7.5). SNR Margin is defined as the maximum dB increase in
equalized noise or the maximum dB decrease in equalized signal that a system can tolerate and
maintain a BER of 10−7.
9.2.7 Loss of Sync Word Failure (LOSW failure)
An LOSW failure shall be declared after 2.5 ± 0.5 s of contiguous LOSW defect. The LOSW failure
shall be cleared when the LOSW defect is absent between 2 and 20 s. The minimum hold time for
indication of LOSW failure shall be 2 s.
9.3 SHDSL line related performance parameters

9.3.1 Code Violation (CV)
The SHDSL parameter Code Violation is defined as a count of the SHDSL CRC anomalies
occurring during the accumulation period. This parameter is subject to inhibiting – see 9.3.6.
9.3.2 Errored Second (ES)
The SHDSL parameter Errored Second is defined as a count of 1-second intervals during which one
or more CRC anomalies are declared and/or one or more LOSW defects are declared. This
parameter is subject to inhibiting – see 9.3.6.
9.3.3 Severely Errored Second (SES)
The SHDSL parameter Severely Errored Second is defined as a count of 1-second intervals during
which at least 50 CRC anomalies are declared or one or more LOSW defects are declared. (50 CRC
anomalies during a 1-second interval is equivalent to a 30% errored frame rate for a nominal frame
length.) This parameter is subject to inhibiting – see 9.3.6.
9.3.4 LOSW Second (LOSWS)
The SHDSL parameter LOSW Second is defined as a count of 1-second intervals during which one
or more SHDSL LOSW defects are declared.
9.3.5 Unavailable Second (UAS)
The SHDSL parameter Unavailable Second is a count of 1-second intervals for which the SHDSL
line is unavailable. The SHDSL line becomes unavailable at the onset of 10 contiguous SESs. The
10 SESs are included in the unavailable time. Once unavailable, the SHDSL line becomes available
at the onset of 10 contiguous seconds with no SESs. The 10 s with no SESs are excluded from
unavailable time.
9.3.6 Inhibiting rules
– UAS parameter counts shall not be inhibited.
– ES and SES shall be inhibited during UAS. Inhibiting shall be retroactive to the onset of
unavailable time and shall end retroactively to the end of unavailable time.
– The CV parameter shall be inhibited during SES.
Further information on inhibiting rules and how ES and SES are decremented can be found in
IETF RFC 2495: Definitions of Managed Objects for the DS1, E1, DS2 and E2 Interface Types
[B9].
34 ITU-T Rec. G.991.2 (12/2003)

9.4 Performance data storage
In order to support SHDSL performance history storage at the STU-C, each SHDSL network
element shall monitor performance and maintain a modulo counter for each performance parameter
that is specified in 9.5.5.7.14 and 9.5.5.7.15, as appropriate. No initialization of these modulo
counters is specified or necessary. By comparing the current reading of the modulo counter with the
previous reading stored in memory, the data base manager in the STU-C can determine the number
of counts to add to the appropriate performance history bin. (Note that the number of counts may
decrease under some fault conditions – see 9.3 for additional information.) The modulo counters are
reported in the SHDSL Performance Status Messages (9.5.5.7.14 and 9.5.5.7.15).
The STU-C shall collect performance history by polling each SHDSL network element with a time
interval that precludes overflow of the modulo counter. For example, the modulo counter for
Errored Seconds is 8 bits which allows a maximum of 255 s between polls before overflow may
occur. Note that the polling that is referred to herein is implemented by the internal database
manager in the STU-C rather than an external network manager.
The STU-C shall maintain performance history bins for each SHDSL segment endpoint. The
performance history bins shall include the total collected counts for the current 15-minute period,
32 previous 15-minute periods, current 24-hour period, and 7 previous 24-hour periods.
9.5 Embedded operations channel

9.5.1 Management reference model
The STU-C shall maintain a management information database for external access by network
management or via craft interface.
Optionally, the STU-R may maintain a management information database, which can be locally
accessed (through a craft interface). This is particularly useful when the STU-C, due to fault
conditions, is unreachable via the EOC.
Access to the management information database from craft interfaces on attached units shall be
provided through a virtual-terminal interface.
9.5.2 EOC overview and reference model
The EOC allows terminal units to maintain information about the span. There are two basic flows of
data, differentiated by which terminal unit initiates the data flow (and subsequently stores the
information for external access). The data flow initiating from the STU-C is mandatory. The data
flow initiating from the STU-R is optional, but all units must respond to requests in either direction
of data flow. In all cases the "master database" shall be stored at the STU-C and all conflicts shall
be resolved in favour of the STU-C (i.e., the information at the STU-C takes precedence). The data
flows are illustrated in Table 9-1 for a two regenerator link (Q denotes a query or command
message, R denotes a response message). Up to eight regenerators are supported by the protocol
definition. Asterisks denote optional message transmissions. A block diagram example of a link
with two regenerators is shown in Figure 9-1.
ITU-T Rec. G.991.2 (12/2003) 35

Table 9-1/G.991.2 – Illustration of EOC flow with two regenerators
Messages from Messages from Messages from Messages from
STU-C Msg(src,dest) SRU1 Msg(src,dest) SRU2 Msg(src,dest) STU-R Msg(src,dest)
Q(1,3) Æ Æ Process
Process Å Å R(3,1)
Q(1,4) Æ Æ Forward Æ Æ Process
Process Å Å Forward Å Å R(4,1)
Q(1,2) Æ Æ Forward Æ Æ Forward Æ Æ Process
Process Å Å Forward Å Å Forward Å Å R(2,1)
Process Å Å Q(2,3)*
R(3,2) Æ Æ Process
Process Å Å Forward Å Å Q(2,4)*
R(4,2) Æ Æ Forward Æ Æ Process
Process Å Å Forward Å Å Forward Å Å Q(2,1)*
R(1,2) Æ Æ Forward Æ Æ Forward Æ Æ Process
* Indicates optional messages.
The data link layer of SHDSL EOC checks the FCS and if valid passes the packet to the network
layer. If the CRC is invalid the entire packet is ignored. The network layer consists of three possible
actions: Process, Forward, and Ignore/Terminate. Process means that the source address and HDLC
information field are passed on to the application layer. Forward means that the packet is sent
onward to the next SHDSL element. (Note that only SRUs will forward packets.) Ignore/Terminate
means that the HDLC packet is ignored and is not forwarded. An SRU may both process and
forward a packet in the case of a broadcast message. If the segment is not active in the forwarding
direction, the SRU shall discard the packet instead. When the segment is active in the forwarding
direction, the maximum forwarding delay in an SRU shall be 300 ms. All retransmission and flow
control is administered by the endpoints, the STUs.
To accommodate the dual data flows, SHDSL regenerators have dual addresses as shown in
Table 9-1. One address is for communication with the STU-C and the other address is for
communication with the STU-R. During Discovery, the STU-C and optionally the STU-R send
discovery probe messages, which propagate across the span and allow the SRUs to be numbered via
a hop count field in the message. This process is explained in detail below.
The SHDSL terminal units communicate unidirectionally and thus have only one address. The
STU-C is assigned a fixed address of 1 and the STU-R is assigned a fixed address of 2. At
power-up, each SRU is assigned the address of 0 for each direction. Under a LOSW failure
condition, the SRU shall reset its source address to 0 for the direction in which the LOSW failure
exists. The SRU source address shall be changed from 0 if and only if a discovery probe message is
received and processed. In this way, a regenerator will only communicate in the direction of a
database. For instance, if a regenerator receives a probe message from the STU-C and not from the
STU-R, then its address will remain 0 in the direction towards the remote.
9.5.3 EOC startup
After loop activation, the SHDSL EOC goes through three initialization stages: Discovery,
Inventory and Configuration. During Discovery, the STU-C and optionally the STU-R will learn if
any mid-span regenerators exist and their addresses will be determined. During Inventory, the
STU-C will poll each SRU and the STU-R to establish inventory information on each element for
36 ITU-T Rec. G.991.2 (12/2003)

the terminal unit's database. (Similarly, the STU-R may poll each SRU and the STU-C to establish
its own database, although this is optional.) During Configuration, the STU-C configures the
STU-R and any SRUs for alarm thresholds, signal characteristics, etc. There is no enforcement of
the order or time of the Inventory and Configuration phases; the initiating STU is in control.
Table 9-12 is an example of Discovery starting from the STU-C and then followed by an optional
Discovery initiated by the STU-R. Although these are shown sequentially in this example, they are
actually independent; it is not necessary for the STU-R to wait until it received the probe from the
STU-C before initiating its own Discovery phase. The STU-R may send its probe as soon as its
EOC is active. The Discovery Response contains the current hop count, the vendor ID, EOC version
and an indication of LOSW in the forward direction (i.e., in the direction of EOC flow that is
opposite to the direction that the Discovery Response is sent).
Table 9-2/G.991.2 – Illustration of EOC discovery phase

Messages from STU-C Messages from SRU1 Messages from SRU2 Messages from STU-R
Msg(src,dest,h) Msg(src,dest,h) Msg(src,dest,h) Msg(src,dest,h)
DP(1,0,0) Æ
Å DR(3,1,1)
DP(0,0,1) Æ
Å Forward Å Å DR(4,1,2)
DP(0,0,2) Æ
Å Forward Å Å Forward Å Å DR(2,1,3)
Å DP(2,0,0)
DR(3,2,1)Æ
Å DP(4,0,1)
DR(4,2,2) Æ Æ Forward Æ
Å DP(3,0,2)
DR(1,2,3) Æ Æ Forward Æ Æ Forward Æ
NOTE – h = hop count, DP = Discovery Probe, DR = Discovery Response.
After the Initiator (STU-C and optionally STU-R) has received a Discovery Response message
from an element, it shall then begin the Inventory phase for that particular element. This is
accomplished by polling that particular element for its inventory information. After the Initiator has
received the inventory information for a unit, it shall then begin the Configuration phase by sending
the appropriate configuration information to the corresponding element. The Inventory and
Configuration Phases operate independently for each responding terminal/regenerator unit.
To ensure interoperability, the behaviour of slave or responding units is carefully specified by this
Recommendation. The particular method for handling dropped packets or no response is left to the
discretion of the initiating STU.
Table 9-3 shows the EOC state table for the network side of an SRU. Note that an identical, but
independent, state machine exists for the customer side of an SRU to support messages originating
from the STU-R.
The state machine consists of three states: Offline, Discovery and EOC Online. The Offline state is
characterized by LOSW failure (a loss of SHDSL sync). The Discovery state is characterized by an
unknown address. Once the address is learned through the Discovery message, the SRU enters the
EOC online or active state. At this point, the SRU will respond to inventory, configuration,
maintenance, or other messages from the STU-C.
ITU-T Rec. G.991.2 (12/2003) 37

Table 9-3/G.991.2 – SRU network EOC state table
Offline state
Event Action
Network LOSW = 0 EOC State = Discovery Ready
Discovery ready state
Event Action
Network LOSW = 1 Network EOC Address = 0
Network EOC State = Offline
Discovery probe message received from Increment Hop Count
the Network side Set Network EOC address to Hop Count +2
Compose and present Discovery message to Customer
side application layer
Send Discovery Response to STU-C
Network EOC State = EOC Online
Message with address not equal to unit's Request forwarding of the message from the Customer
address received from the Network side. side network layer
Message Forwarding Requested from Send requested message toward Network if EOC not
Customer side offline
EOC online state
Event Action
Network LOSW = 1 Network EOC Address = 0
Network EOC State = Offline
Discovery message received from the Increment Hop Count
Network side Set Network EOC address to Hop Count +2
Compose and present Discovery message to Customer
side application layer
Send Discovery Response to STU-C
Message with broadcast destination Process the message
address received from the Network side Request the Customer side EOC network layer to
forward the message
Message with unit's destination address Process the message
or address 0 received from the Network
side
Message with address not equal to unit's Request forwarding of message from the Customer
address received from the Network side side network layer
Message forwarding requested from Send requested message toward Network if EOC not
Customer side network layer offline
38 ITU-T Rec. G.991.2 (12/2003)

9.5.4 Remote management access
The STU-C shall maintain the master management database for the entire SHDSL span. (An
optional second database is maintained at the STU-R.) Other units are only required to store enough
information to accurately send information via the EOC. The information contained in the master
database shall be accessible from any SHDSL unit that has a craft port and from network
management if it is available. The craft access is in the form of a virtual-terminal interface (or
virtual-craft-port interface). This interface is defined so that it can be used by any attached unit to
access the terminal screen of another unit on the same SHDSL span. Support for this feature is
optional, with the exception of the STU-C, which shall support the "host" side of at least one remote
terminal connection. (Whether this interface can be active simultaneously with local craft access to
the STU-C is a vendor decision and beyond the scope of this Recommendation.) The
virtual-terminal interface consists of connect, disconnect, keyboard, and screen messages. After a
connection has been established, input characters from the craft port are sent in Keyboard data
messages to the "host" unit. The host unit, in turn, shall send information in the form of ASCII text,
ASCII control codes, and screen control functions in Screen messages, whose contents are
transmitted back to the craft port. The host unit shall echo characters.
The method for determining that remote access through the local craft port is desired or should be
terminated is vendor specific, and beyond the scope of this Recommendation. Whatever method is
used, capability for transmitting all valid key sequences (ASCII characters and control codes) shall
be provided.
9.5.5 EOC transport
The EOC shall be transported in the SHDSL frame in bits eoc1 through eoc20. Five octets are
contained in each two SHDSL frames, with specified alignment. The least significant bit (LSB) of
the octets are located in bits 1, 9, and 17 of the EOC bits in the first frame and bits 5 and 13 of the
second frame; each octet is transmitted LSB first. Octet alignment across frames is achieved
through detection of the alignment of the HDLC Sync pattern (7E16).
For optional M-pair operation, each EOC message shall be sent in parallel such that redundant and
identical messages are sent over all M loops.
9.5.5.1 EOC data format
Numerical data and strings are placed in the EOC with octet alignment. Data items that are not an
integral number of octets have been packed together to minimize message sizes.
Numerical Fields shall be transmitted most significant octet first, least significant bit first within an
octet. (This is consistent with "network octet ordering" as in IETF RFC 1662: PPP in HDLC-like
Framing [4].)
Strings shall be represented in the data stream with their first character (octet) transmitted first.
Strings shall be padded with spaces or terminated with a NULL (0016) to fill the allocated field size.
String fields are fixed length so characters after a NULL in a string data field are "don't care".
ITU-T Rec. G.991.2 (12/2003) 39

9.5.5.2 EOC frame format
The EOC channel shall carry messages in an HDLC-like format as defined in 6.3/G.997.1 [3]. The
channel shall be treated as a stream of octets; all messages shall be an integral number of octets.
The frame format uses a compressed form of the HDLC header, as illustrated in Table 9-4. The
destination address field shall be the least significant 4 bits of octet 1; the source address field shall
occupy the most significant 4 bits of the same octet (the address field). There is no control field.
One or more sync octets (7E16) shall be present between each frame. Inter-frame fill shall be
accomplished by inserting sync octets as needed. Discovery probe messages shall be preceded by at
least 5 sync octets to assure proper detection of octet alignment. The Information Field contains
exactly one Message as defined below. The maximum length of a frame shall be 75 octets, not
including the sync pattern or any octets inserted for data transparency.
Table 9-4/G.991.2 – Frame format for SHDSL EOC

MSB LSB
Octet # Contents
Sync pattern 7E16
Source address bits 7..4 Destination address bits 3..0
1 Message ID per Table 9-6. Information
2 Message Content – Octet 2 field
… …
L Message Content – Octet L …
FCS octet 1
FCS octet 2
Sync pattern 7E16
9.5.5.3 Data transparency

Transparency for the information payload to the sync pattern (7E16) and the control escape pattern
7D16 shall be achieved by octet stuffing.
Before transmission:
• octet pattern 7E16 is encoded as two octets 7D16, 5E16;
• octet pattern 7D16 is encoded as two octets 7D16, 5D16.
At reception:
• octet sequence 7D16, 5E16 is replaced by octet 7E16;
• octet sequence 7D16, 5D16 is replaced by octet 7D16;
• any other two-octet sequence beginning with 7D16 aborts the frame.
9.5.5.4 Frame check sequence
The frame check sequence (FCS) shall be calculated as specified in IETF RFC 1662 [4]. (Note that
the FCS is calculated before data transparency.) The FCS shall be transmitted as specified in
IETF RFC 1662.
40 ITU-T Rec. G.991.2 (12/2003)

9.5.5.5 Unit addresses
Each unit uses one source and destination address when communicating with upstream units and a
separate, independent source and destination address when communicating with downstream units.
Each address shall have a value between 016 and F16. Units shall be addressed in accordance with
Table 9-5. Address F16 may only be used as a destination address and shall specify that the message
is addressed to all units. Address 016 is used to address the next attached or adjacent unit.
Table 9-5/G.991.2 – Device addresses

Address (Base16) Device
0 Adjacent device
1 STU-C
2 STU-R
3-A Regenerators 1-8
B-E Reserved (D and E not allowed)
F Broadcast message, to all stations
NOTE – This Recommendation is not intended to indicate how many regenerators can or should be
supported by a product; only how to identify them if they exist.
9.5.5.6 Message IDs
Table 9-6 summarizes message ID. Message IDs are listed as decimal numbers. Messages 0-64
represent request messages. Messages 128-192 represent messages that are sent in response to
request messages. Each request message is acknowledged with the corresponding response.
Request/Response Message IDs usually differ by an offset of 128.
Table 9-6/G.991.2 – Summary of message IDs

Message ID
Message type Initiating unit Reference
(decimal)
0 Reserved
1 Discovery Probe STU-C, STU-R*, SRU 9.5.5.7.1
2 Inventory Request STU-C, STU-R* 9.5.5.7.3
3 Configuration Request – SHDSL STU-C 9.5.5.7.5
4 Reserved for Application Interface
Configuration
5 Configuration Request – Loopback STU-C, STU-R* 9.5.5.7.6
Timeout
6 Virtual Term. Connect Req. STU-R*, SRU* 9.5.5.7.16
7 Virtual Terminal Disc. Req. STU-R*, SRU* 9.5.5.7.16
8 Keyboard data message STU-R*, SRU* 9.5.5.7.17
9 Maintenance request – System STU-C, STU-R* 9.5.5.7.18
Loopback
10 Maintenance request – Element STU-C, STU-R* 9.5.5.7.19
Loopback
11 Status Request STU-C, STU-R* 9.5.5.7.11
12 Full Status Request STU-C, STU-R* 9.5.5.7.12
ITU-T Rec. G.991.2 (12/2003) 41

Message ID
(decimal)
13-14 Reserved
15 Soft restart/Power backoff disable STU-C 9.5.5.7.21
Request
16 Reserved (Future)
17 ATM Cell Status Request STU-C, STU-R* E.9.4.7
18 STU-R Configuration Request – STU-C 9.5.5.7.9
Management
19 Reserved for Voice Transport Request Undefined
(Future)
20 ISDN Request STU-C, STU-R E.8.7.1,
E.13.3
21 LAPV5 POTS and ISDN Setup Request STU-C E.13.6
22 Deactivation Request STU-C, STU-R H.1.3.1
23 Mapping Request STU-C, STU-R* 9.5.5.7.27
23-63 Reserved (Future)
64-88 Reserved for Line management Request Undefined 9.5.5.7.22
89-111 Reserved
112-119 Proprietary Message Undefined 9.5.5.7.23
120 External Message Undefined 9.5.5.7.24
121 G.997.1 Message STU-C*, STU-R* 9.5.5.7.25
122-124 Reserved
125-127 Excluded (7D16, 7E16, 7F16)
128 Reserved
129 Discovery Response All 9.5.5.7.2
130 Inventory Response All 9.5.5.7.4
131 Configuration Response – SHDSL STU-R, SRU 9.5.5.7.7
Configuration
133 Configuration Response – Loopback All 9.5.5.7.8
Timeout
134 Virtual Terminal Connect Response STU-C, SRU*, 9.5.5.7.16
STU-R*
135 Reserved
136 Screen data message STU-C, SRU*, 9.5.5.7.17
STU-R*
137 Maintenance Status All 9.5.5.7.20
138 Reserved
139 Status/SNR All 9.5.5.7.13
140 Performance Status SHDSL Network SRU, STU-R 9.5.5.7.14
Side
42 ITU-T Rec. G.991.2 (12/2003)

Message ID
(decimal)
141 Performance Status SHDSL Customer STU-C, SRU 9.5.5.7.15
Side
Performance
143 Reserved (Future)
144 Generic Unable to Comply (UTC) 9.5.5.7.26
145 ATM Cell Status Information All E.9.4.8
146 Configuration Response – Management STU-R, SRU 9.5.5.7.10
147 Reserved for Voice Transport Response Undefined
(Future)
148 ISDN Response STU-C, STU-R E.8.7.1,
E.13.3
149 LAPV5 POTS and ISDN Set-up STU-R E.13.6
150 Deactivation Response STU-C, STU-R H.1.3.2
151 Mapping Response All 9.5.5.7.28
192-216 Segment Management Response Undefined 9.5.5.7.22
(reserved)
240-247 Proprietary message Response Undefined 9.5.5.7.23
248-252 Reserved
253-255 Excluded (FD16, FE16, FF16)
* Denotes optional support. A unit may initiate this message.
9.5.5.7 Message contents

Each message shall have the contents in the format specified in Tables 9-7 through 9-31. If any
message has a message length longer than expected and is received in a frame with a valid FCS,
then the known portion of the message shall be used and the extra octets discarded. This will permit
addition of new fields to existing messages and maintain backward compatibility. New data fields
shall only be placed in reserved bits after the last previously defined data octet. Reserved bits and
octets shall be filled with the value 0016 for forward compatibility.
Response messages may indicate UTC (Unable to Comply). Note that this is not an indication of
non-compliance. UTC indicates that the responding unit was unable to implement the request.
9.5.5.7.1 Discovery probe – Message ID 1
The discovery probe message shall be assigned Message ID 1, and is used to allow an STU to
determine how many devices are present and assign addresses to those units.
ITU-T Rec. G.991.2 (12/2003) 43

Table 9-7/G.991.2 – Discovery probe information field
Octet # Contents Data type Reference
1 1 Message ID
2 Hop Count Unsigned char 9.5.3
9.5.5.7.2 Discovery response – Message ID 129

The discovery response message shall be assigned Message ID 129. This message shall be sent in
response to a discovery probe message. The hop count field shall be set to 1 larger than the value
received in the discovery probe message causing the response. (The full receive state machine is
described in Table 9-3.) Forward LOSW indication means that the segment is down in the forward
direction from the SRU. In the case of optional M-pair operation, Forward LOSW indication means
that all M loops are down in the forward direction from the SRU. In either case, the SRU is unable
to forward the Discovery Probe message to the adjacent unit and it reports this fact to the initiating
STU. The Forward LOSW octet field shall be set to 0016 for responses from an STU.
The Vendor ID field is used to identify the system integrator, as specified in 9.5.5.7.4.
The SHDSL version number indicates the SHDSL standard to which the system was built. For the
present version of this Recommendation (12/2003), these bits shall be set to 00001000.
The vendor EOC software version number shall be assigned by the system vendor as identified by
the vendor ID. Software version numbers are to be incremented for each new SHDSL standard.
Table 9-8/G.991.2 – Discovery response information field

1 129 Message ID
2 Hop Count Unsigned char 9.5.3
3 Reserved
4-11 Vendor ID (ordered identically to bits
in G.994.1 Vendor ID)
12 Vendor EOC Software Version Unsigned char
13 SHDSL Version # Unsigned char
14 bits 7..1 Reserved
14 bit 0 Forward LOSW indication, EOC Bit 1 = Unavailable
unavailable 0 = Available
9.5.5.7.3 Inventory request – Message ID 2

The inventory request message shall be assigned Message ID 2. This message is used to request an
inventory response from a particular unit. It shall only be transmitted by STU devices. There shall
be no octets of content for this message.
Table 9-9/G.991.2 – Inventory request information field

1 2 Message ID
9.5.5.7.4 Inventory response – Message ID 130

The inventory response message shall be assigned Message ID 130. This message shall be sent in
response to an inventory request message.
44 ITU-T Rec. G.991.2 (12/2003)

The SHDSL version number indicates the SHDSL standard to which the system was built. For
G.shdsl.bis (draft), these bits shall be set to 00001000.
The Vendor ID field is used to specify the system integrator. In this context, the system integrator
usually refers vendor of the smallest field-replaceable unit. This typically is also the entity pointed
to by the Unit Identification Code (CLEITM) Field. As such, the Vendor ID field contents may not
be the same as the Vendor ID indicated within ITU-T Rec. G.994.1, which relates to the
manufacturer of the physical layer interface. The serial number, model number, issue number, list
number, and software revision number shall all be assigned with respect to the system integrator.
Special unit identification codes are used in North America. These CLEI (Common Language
Equipment Identifier) codes are used by network service providers for inventory, spare-part
ordering, provisioning and maintenance operations. In North America, CLEI codes are used as a
vendor's product ID. CLEI codes conform to ANSI T1.213, Coded Identification of Equipment
Entities of the North American Telecommunications System for Information Exchange. In regions
outside North America, these fields may be set to zero.
Information on the modem software version, the vendor list (modem hardware version), the modem
issue, the model number, and the modem serial number are specific to the system. Therefore, this
information shall be assigned by the system vendor as identified by the vendor ID.
The vendor software version indicates the version of the software of the SHDSL system. The
vendor software version is not necessarily identical to the EOC software version in 9.5.5.7.2. The
vendor list number indicates the version number of the system hardware. The vendor issue field
indicates the particular usage of the unit. The vendor model number is a unique number for the
particular type of unit. The vendor serial number is a number that identifies every unit individually.
Table 9-10/G.991.2 – Inventory response information field

1 130 Message ID
2 SHDSL Version # Unsigned char
3-5 Vendor List # 3-octet string
6-7 Vendor Issue # 2-octet string
8-13 Vendor Software Version 6-octet string
14-23 Unit Identification Code (CLEI) 10-octet string
24 Reserved
25-32 Vendor ID (ordered identically to bits
in G.994.1 Vendor ID)
33-44 Vendor model # 12-octet string
45-56 Vendor serial # 12-octet string
57-68 Other vendor information 12-octet string
ITU-T Rec. G.991.2 (12/2003) 45

9.5.5.7.5 Configuration request – SHDSL: Message ID 3
The configuration request – SHDSL message is transmitted by the STU-C to configure the SHDSL
interface(s) of attached units. This message may be broadcast or addressed to specific units. It is
acknowledged with a configuration response – SHDSL message. For SHDSL, SNR is measured
internal to the transceiver decision device as opposed to the external segment termination. The
"Off" setting indicates that threshold crossings are not reported. Loop attenuation and SNR margin
are local alarms that are reported in Messages 140 and 141. In addition, these alarms may be
physically indicated on the equipment. SHDSL loop attenuation shall be defined as follows:
 f sym 
 2 10 × log  S f − nf ) df −
1
 ∫0 10  ∑ ( sym


2  n =0  
LoopAttenSHDSL ( H ) =
f sym  f sym  1 2

 ∫0
(
 2 10 × log10  ∑ S f − nf sym ) H ( f − nf sym )  
df
 n =0  
1
where fsym is the symbol rate, is the insertion loss of the loop, and S ( f ) is the nominal
H( f )
transmit PSD.
Table 9-11/G.991.2 – Configuration request – SHDSL information field

1 3 Message ID
2 bit 7 Config Type Bit 0-normal,
1-Read only
2 bits 6..0 SHDSL Loop Attenuation threshold (dB) Enumerated 0 = off, 1 to 127
3 bits 7..4 SHDSL SNR Margin threshold (dB) Enumerated 0 = off, 1 to 15
3 bits 3..0 Reserved Set to 0
9.5.5.7.6 Configuration request – loopback timeout: Message ID 5

The configuration request – loopback timeout message is transmitted by the STU-C (and optionally
the STU-R) to set loopback timeouts for individual elements. If a loopback is not cleared before the
expiration of the timeout, then the element shall revert to normal operation. This message may be
broadcast or addressed to specific units. It is acknowledged with a configure response – loopback
timeout message. If date and time information is sent in octets 4-21, then these strings shall conform
to ISO 8601 [5]. If date and time information is not sent, then these fields shall be filled with zeros.
Table 9-12/G.991.2 – Configuration request – loopback timeout information field

1 5 Message ID
2 bit 7 Config Type Bit 0 = normal,
1 = Read-only.
2 bits 3..0-3 Loopback timeout 12-bit unsigned integer In minutes, 0 = no timeout
4-13 YYYY-MM-DD 10-octet date string ISO 8601
14-21 HH:MM:SS 8-octet time string ISO 8601
46 ITU-T Rec. G.991.2 (12/2003)

9.5.5.7.7 Configuration response – SHDSL: Message ID 131
The configuration response – SHDSL message is transmitted to the STU-C in response to a
configuration request – SHDSL message. This response is sent after the applicable configuration
changes have been made. The values of the response shall be set to the new values, after they have
been applied. If a transceiver unit is unable to comply with the request, the bit in the compliance
octet is set and the current settings are reported. If the config request message was received with a
config type of "Read-Only," then no changes are made to the current configuration and the current
values are reported.
Table 9-13/G.991.2 – Configuration response – SHDSL information field

1 131 Message ID
2 bit 0 UTC (Unable to Comply) Bit 0 = OK, 1 = UTC
3 SHDSL Loop Attenuation threshold (dB) Char 0 = off, 1 to 127
4 bits 7..4 SHDSL SNR Margin threshold (dB) Enumerated 0 = off, 1 to 15
9.5.5.7.8 Configuration response – loopback timeout: Message ID 133

The configuration response – loopback timeout message is transmitted to acknowledge the
configuration request – loopback timeout message. This response is sent after the applicable
configuration changes have been made. The values of the response shall be set to the new values,
after they have been applied. If a transceiver unit is unable to comply with the request, the bit in the
compliance octet is set and the current settings are reported. If the config request message was
received with a config type of "Read-Only," then no changes are made to the current configuration
and the current values are reported.
Table 9-14/G.991.2 – System Loopback Timeout Response information field

Octet # Information Field Data type Reference
1 133 Message ID
2 bit 0 UTC (Unable to Comply) Bit 0 = OK, 1 = UTC
3 bits 3..0-4 Loopback timeout 12-bit unsigned integer In minutes,
0 = no timeout
5-14 YYYY-MM-DD 10-octet date string ISO 8601 [5]
15-22 HH:MM:SS 8-octet time string ISO 8601
9.5.5.7.9 STU-R config – management: Message ID 18

The config request – management message is transmitted by the STU-C to enable or disable STU-R
initiated management flow. The destination address shall be F16 to indicate this is a broadcast
message. STU-R initiated management flow is enabled by default. When disabled, an SRU shall not
respond to any STU-R-initiated request messages, and the STU-R shall not issue any such messages
(messages 2-12). Config type of Read-Only indicates that the addressed unit ignore the subsequent
values in the message and report back its current configuration.
ITU-T Rec. G.991.2 (12/2003) 47

Table 9-14a/G.991.2 – Configuration request – management information field
1 Message ID 18 Message ID
2 bit 7 ConfigType Bit 0-normal, 1-Read-Only
2 bit 0 STU-R Initiated Management Flow Bit 0-Enable, 1-Disabled
9.5.5.7.10 Config response – management message: Message ID 146

Config response – management message is sent by all units to acknowledge to the config request –
management message.
Table 9-14b/G.991.2 – Configuration response – management information field

2 bit 0 UTC (Unable to Comply) Bit 0-OK, 1-UTC
3 bit 0 STU-R Initiated Management Flow Bit 0-Enabled, 1-Disabled
Status
9.5.5.7.11 Status request – Message ID 11

The status request message is used to poll an element for alarm and general performance status.
The polled unit will respond with one or more of the following status response messages:
• Status/SNR response – 139 (9.5.5.7.13).
• SHDSL network side performance status – 140 (9.5.5.7.14).
• SHDSL customer side performance status – 141 (9.5.5.7.15).
• Maintenance status – 137 (9.5.5.7.20).
In the optional M-pair mode, messages 139, 140, and 141 contain status information that is specific
to a particular pair. In this case, M messages each (one corresponding to each pair) of types 139,
140, and 141 may be sent by the polled unit in response to a status request message. The responding
element shall provide the loop ID information in EOC messages 139, 140, and 141. The responding
element shall first provide the information relating to loop 1, followed shortly thereafter with the
requested information for loop 2 (if M ≥ 2), then loop 3 (if M ≥ 3), then loop 4 (if M = 4).
If active alarm, fault or maintenance conditions exist then the polled unit shall respond with the
messages that correspond to the active conditions.
If there has been any change in performance status other than SNR margin since the last time a unit
was polled, then the unit shall respond with the messages which contain the change in performance
status.
Otherwise, the polled unit shall respond with the status/SNR response – 139 (9.5.5.7.13).
Table 9-15/G.991.2 – Status request information field

Octet # Information field Data type
48 ITU-T Rec. G.991.2 (12/2003)

9.5.5.7.12 Full status request – Message ID 12
The full status request message is used to poll an element for its complete current status. The
following messages shall be sent in response to the full status request:
• SHDSL network side performance status (9.5.5.7.14).
• SHDSL customer side performance status (9.5.5.7.15).
• Maintenance status (9.5.5.7.20).
In the optional M-pair mode, the following messages shall be sent in response to the full status
request:
• SHDSL network side performance status (9.5.5.7.14) – related to Loop 1.
• SHDSL network side performance status – for Loop 2 to Loop M (one message per loop).
• SHDSL customer side performance status (9.5.5.7.15) – related to Loop 1.
• SHDSL customer side performance status – for Loop 2 to Loop M (one message per loop).
• Maintenance status (9.5.5.7.20).
Table 9-16/G.991.2 – Full Status Request information field

9.5.5.7.13 Status response/SNR – Message ID 139

The performance status/SNR message shall be sent in response to the status request message under
the conditions specified in 9.5.5.7.9. The reported integer represents dB SNR noise margin values
rounded up. Because each STU only connects to one SHDSL segment, the application interface side
SNR margin data shall be 0 (i.e., the network side SNR margin shall be 0 at the STU-C and the
customer side SNR shall be 0 at the STU-R).
Table 9-17/G.991.2 – Status response OK/SNR information field

2 Network Side SNR Margin (dB) Signed char (127 = Not Available)
3 Customer Side SNR Margin (dB) Signed char (127 = Not Available)
4 Loop ID Unsigned char (1 = Loop 1, 2 = Loop 2,
3 = Loop 3, 4 = Loop 4)
9.5.5.7.14 SHDSL network side performance status – Message ID 140

This message provides the SHDSL network side performance status. Device fault shall be used to
indicate hardware or software problems on the addressed unit. The definition of device fault is
vendor dependent but is intended to indicate diagnostic or self-test results. DC continuity fault shall
be used to indicate conditions that interfere with span powering such as short and open circuits. The
definition of DC continuity fault is vendor dependent.
In octet 11, bits 7..4 are used to indicate that an overflow or reset has occurred in one or more of the
modulo counters. Bits 7 and 5 shall indicate that an overflow has occurred since the last SHDSL
network side status response. For example, if more than 256 errored seconds occur between SHDSL
network side status responses, then the ES modulo counter will overflow. Bits 6 and 4 shall be used
to indicate that one or more of the modulo counters have been reset for any reason (e.g., system
powerup or a non service-affecting reset.) Bits 7 and 6 shall be cleared to 0 after a SHDSL network
ITU-T Rec. G.991.2 (12/2003) 49

side status response is sent to the STU-C. Bits 5 and 4 shall be cleared to 0 after a SHDSL network
side status response is sent to the STU-R.
Table 9-18/G.991.2 – SHDSL-network side performance status information field

2 bit 7 Reserved
Bit 6 N – Power Backoff Status Bit 0 = default
1 = selected
Bit 5 Device Fault Bit 0 = OK, 1 = Fault
Bit 4 N – DC Continuity Fault Bit 0 = OK, 1 = Fault
Bit 3 N – SNR Margin alarm Bit 0 = OK, 1 = alarm
Bit 2 N – Loop Attenuation Alarm Bit 0 = OK, 1 = alarm
Bit 1 N – SHDSL LOSW Failure Alarm Bit 0 = OK, 1 = alarm
Bit 0 Reserved Set to 0
3 N – SHDSL SNR Margin (dB) Signed char (127 = NA)
4 N – SHDSL Loop Attenuation (dB) Signed char (–128 = NA)
5 N – SHDSL ES Count modulo 256 Unsigned char
6 N – SHDSL SES Count modulo 256 Unsigned char
7-8 N – SHDSL CRC Anomaly Count Unsigned int
modulo 65536
9 N – SHDSL LOSW Defect Second Unsigned char
Count modulo 256
10 N – SHDSL UAS Count modulo 256 Unsigned char
11 bit 7 N – Counter Overflow Indication to 0 = OK
STU-C 1 = Overflow
11 bit 6 N – Counter Reset Indication to 0 = OK
STU-C 1 = Reset
11 bit 5 N – Counter Overflow Indication to 0 = OK
STU-R 1 = Overflow
11 bit 4 N – Counter Reset Indication to 0 = OK
STU-R 1 = Reset
11 N-Power Back-Off Base Value (dB) Unsigned char 0 .. 15
bits 3..0
12 bit 7 N-Power Back-Off Extension (dB) Bit 0→ PBO = Base
Value +0 dB
1→ PBO = Base
Value +16 dB
12 Reserved
bits 6..3
12 Loop ID Unsigned char 1 = Loop 1
bits 2..0 2 = Loop 2
3 = Loop 3
4 = Loop 4
50 ITU-T Rec. G.991.2 (12/2003)

9.5.5.7.15 SHDSL customer side performance status – Message ID 141
This message provides the SHDSL Customer Side Performance Status. Device Fault shall be used
to indicate hardware or software problems on the addressed unit. The definition of Device Fault is
vendor dependent but is intended to indicate diagnostic or self-test results. DC Continuity Fault
shall be used to indicate conditions that interfere with span powering such as short and open
circuits. The definition of DC Continuity Fault is vendor dependent.
In octet 11, bits 7..4 are used to indicate that an overflow or reset has occurred in one or more of the
modulo counters. Bits 7 and 5 shall indicate that an overflow has occurred since the last SHDSL
Customer Side status response. For example, if more than 256 Errored Seconds occur between
SHDSL Customer Side status responses, then the ES modulo counter will overflow. Bits 6 and 4
shall be used to indicate that one or more of the modulo counters have been reset for any reason
(e.g., system powerup or a non-service-affecting reset). Bits 7 and 6 shall be cleared to 0 after a
SHDSL Customer Side status response is sent to the STU-C. Bits 5 and 4 shall be cleared to 0 after
a SHDSL Customer Side status response is sent to the STU-R.
Table 9-19/G.991.2 – SHDSL-customer side performance status information field

2 bit 7 Reserved
Bit 6 C – Power Backoff Status Bit 0 = default
1 = selected
Bit 5 Device Fault Bit 0 = OK, 1 = Fault
Bit 4 C – DC Continuity Fault Bit 0 = OK, 1 = Fault
Bit 3 C – SNR Margin alarm Bit 0 = OK, 1 = alarm
Bit 2 C – Loop Attenuation Alarm Bit 0 = OK, 1 = alarm
Bit 1 C – SHDSL LOSW Failure Alarm Bit 0 = OK, 1 = alarm
Bit 0 Reserved Set to 0
3 C – SHDSL SNR Margin (dB) Signed char (127 = NA)
4 C – SHDSL Loop Attenuation (dB) Signed char (128 = NA)
5 C – SHDSL ES Count modulo 256 Unsigned char
6 C – SHDSL SES Count modulo 256 Unsigned char
7-8 C – SHDSL CRC Anomaly Count Unsigned int
modulo 65536
9 C – SHDSL LOSW Defect Second Unsigned char
Count modulo 256
10 C – SHDSL UAS Count modulo 256 Unsigned char
11 bit 7 C – Counter Overflow Indication to 0 = OK
STU-C 1 = Overflow
11 bit 6 C – Counter Reset Indication to 0 = OK
STU-C 1 = Reset
11 bit 5 C – Counter Overflow Indication to 0 = OK
STU-R 1 = Overflow
11 bit 4 C – Counter Reset Indication to 0 = OK
STU-R 1 = Reset
ITU-T Rec. G.991.2 (12/2003) 51

Table 9-19/G.991.2 – SHDSL-customer side performance status information field
11 C-Power Back-Off Base Value (dB) Unsigned char 0 .. 15
bits 3..0
12 bit 7 C-Power Back-Off Extension (dB) Bit 0→ PBO = Base
Value + 0 dB
1→ PBO = Base
Value + 16 dB
12 Reserved
bits 6..3
12 Loop ID Unsigned char 1 = Loop 1
bits 2..0 2 = Loop 2
3 = Loop 3
4 = Loop 4
9.5.5.7.16 Virtual terminal connect/disconnect request/response (Msg. IDs 6, 7, 134)

Three messages are used to maintain (establish, tear down) virtual terminal sessions between units.
A unit may request a connection but must wait for "connect" status response before using the
connection. The connection shall remain until a disconnect request is processed or, if implemented,
a timeout occurs. At least one session shall be supported by the STU-C. STU-R and SRU may
silently ignore the connect request or may respond with a "no connect" status if terminal screens are
not supported.
The connect/disconnect process is necessary for handling the case where keyboard messages are
received from more than one device. If a unit cannot accommodate another connect request, it shall
send the "no connect" response.
The connect request message can be sent to cause a refresh of the current screen. When a connect
request is accepted the "connect" response shall be transmitted, followed by screen messages with
the current screen. If this is a new connection then the first screen shall be sent. The end unit that
issues the connect request (Message 6) shall issue the corresponding request (Message 7) to
terminate the virtual terminal session. A far-end unit shall respond with Virtual Terminal "no
connect" status (Message 134) when it receives a keyboard message from the near-end unit that
terminates a virtual terminal session. (This lets the near-end know that the far-end has terminated
the connection.) If the far-end unit has dropped the terminal session and a keyboard message is
received, the far-end unit shall respond with Message 134 – "no connect".
For any keyboard character that has special meaning at the near-end (e.g., and escape command),
means shall be provided to send that keyboard character to the far end. (This is analogous to the
CTRL-] escape command used in TELNET sessions. When one has escaped in a TELNET session
to the local terminal, one can typically issue a "send escape" command or similar to send the CTRL-
] character to the far end.)
Table 9-20/G.991.2 – Virtual terminal connect

1 Message ID 6 – Virtual Terminal Connect Message ID
52 ITU-T Rec. G.991.2 (12/2003)

Table 9-21/G.991.2 – Virtual terminal disconnect
1 Message ID 7 – Virtual Terminal Disconnect Message ID
Table 9-22/G.991.2 – Virtual terminal connect response

1 Message ID 134 – Virtual Terminal Connect Response Message ID
2 Connection status 1 = connected
0 = no connect
9.5.5.7.17 Screen message/keyboard message (Msg. IDs 8, 136)

Keyboard and screen messages are only sent over an active connection between units. Keyboard
messages shall be 1 to 8 data octets per message. Queuing of keystrokes from the customer may
affect user response times and should be done with care. Screen messages shall be 1 to 24 data
octets per message, and their contents are vendor defined. See 9.5.6 for more information on
Screen/Keyboard messages.
Table 9-23/G.991.2 – Keyboard information field

1 Message ID 8 – Keyboard Message ID
2.. (L + 1) ASCII character(s) and escape sequences char array
Table 9-24/G.991.2 – Screen information field

1 Message ID 136 – Screen Message ID
2.. (L + 1) ASCII characters and escape sequences char array
9.5.5.7.18 Maintenance Request – System Loopback Messages (9)

The Maintenance Request – System Loopback Message contains loopback commands for all of the
elements on the span. The contents of the Maintenance Request – System Loopback message are
shown in Table 9-25. The System Loopback message shall have a broadcast destination address
when sent from the STU-C. When optionally sent from the STU-R, the System Loopback message
shall have the STU-C as its destination address. Upon reception of this message, each SRU and
STU shall comply with its corresponding command field and respond to the sender with the
Maintenance Status message. Note that the SRUs are numbered consecutively beginning with
closest SRU to the STU-C. Each SRU shall determine its number by subtracting 2 from its network
side EOC address. Since the network side EOC addresses must be known, the STU-R shall not use
the System Loopback Message if the STU-C is offline. To invoke SRU loopbacks while the STU-C
is offline, the STU-R shall use the Maintenance Request-Element Loopback message. (Maintenance
request messages may also be used by the STU devices to poll for current loopback status, using the
unchanged bit flags.)
ITU-T Rec. G.991.2 (12/2003) 53

Table 9-25/G.991.2 – Maintenance Request – System Loopback information field
1 Message ID 9 – Maintenance Request-
System Loopback
2 STU-C Loopback Commands Bit flags Table 9-26
3 STU-R Loopback Commands Bit flags Table 9-26
4 SRU #1 Loopback Commands Bit flags Table 9-26
Table 9-26/G.991.2 – Loopback command bit flag definitions

Bit positions Definition
7 Reserved
6 Clear All Maintenance States (including any proprietary states)
5 Initiate Special Loopback
4 Terminate Special Loopback
3 Initiate Loopback toward the Network
2 Initiate Loopback toward the Customer
1 Terminate Loopback toward the Network
0 Terminate Loopback toward the Customer
NOTE – Bit set to 1 – perform action, Bit Set to 0 – no action taken, report current status.
9.5.5.7.19 Maintenance Request – Element Loopback Message ID 10

The Maintenance Request – Element Loopback Message contains loopback commands for an
individual element. The contents of the Maintenance Request – Element Loopback message are
shown in Table 9-27. The Element Loopback message shall have an individual unit's destination
address according to the data flow addresses described in 9.5.2. Upon reception of the Element
Loopback message, the addressed unit shall comply with the loopback commands and reply with
the Maintenance Status Response message.
Table 9-27/G.991.2 – Maintenance Request – Element Loopback information field

1 Message ID 10 – Maintenance Request Message ID
2 Loopback Commands Bit flags Table 9-26
54 ITU-T Rec. G.991.2 (12/2003)

9.5.5.7.20 Maintenance Status Response Message ID 137
Maintenance status is sent in response to the Maintenance Request – System Loopback,
Maintenance Request – Element Loopback, Status Request, and Full Status Request query
messages. The "Special loopback" is defined for the STU-R as a Maintenance Termination Unit
(MTU) loopback; it is not defined at other units.
Table 9-28/G.991.2 – Maintenance Status information field

1 Message ID 137 – Message ID
Maintenance Status-Loopback
2 bit 7 Loopback Timeout Status Bit 0 = unchanged,
1 = changed
2 bit 6 Proprietary Maintenance State active Bit 0 = off, 1 = on
2 bit 5 Special loopback active Bit 0 = off, 1 = on
2 bit 4 Loopback active toward STU-R Bit 0 = off, 1 = on
2 bit 3 Loopback active toward STU-C Bit 0 = off, 1 = on
2 bit 2 Local or span-powered unit Bit 0 = span powered
1 = local powered
2 bit 1 Customer Tip/Ring Reversal Bit 0 = normal
1 = reversed
2 bit 0 Network Tip/Ring Reversal Bit 0 = normal
1 = reversed
9.5.5.7.21 Soft restart/power backoff disable Message ID 15

The purpose of this message is to switch a receiver between the default and selected modes of
power backoff. If default mode is set, PBO shall be set to the default value. Otherwise, in selected
mode, PBO may be negotiated through G.994.1 to another value. In order for a change in power
backoff mode to take effect, the receiver must reactivate. The Soft Restart request shall cause the
receiving unit to terminate the corresponding SHDSL connection and enter the Exception State
(Figure 6-7). The connection shall not be terminated unless the corresponding Soft Restart bit is set
in this message. The receiving unit shall wait 5 ± 1 s before terminating the SHDSL connection.
This message carries the command to set the power backoff mode. The power backoff mode
received in this message shall be maintained as long as power is applied to the unit. Maintaining the
power backoff mode in non-volatile storage is optional. Note that the configuration of power
backoff mode applies to the receiver; i.e., the receiver requests a PSD mask based on both the
received power and the configuration of its power backoff mode.
ITU-T Rec. G.991.2 (12/2003) 55

Table 9-29/G.991.2 – Soft restart information field
1 Message ID 15 – Soft Restart/Backoff Message ID
2 bit 1 Network Side Power Backoff Setting Bit 0 = default
1 = selected
2 bit 0 Network Side Soft Restart (after 5 s) Bit 0 = no Restart
1 = Restart
3 bit 1 Customer Side Power Backoff Setting Bit 0 = default
1 = selected
3 bit 0 Customer Side Soft Restart (after 5 s) Bit 0 = no Restart
1 = Restart
9.5.5.7.22 Segment management message – (IDs 64-88, 192-216)

A range of Message IDs is reserved for segment management (e.g., continuous precoder update).
9.5.5.7.23 Proprietary messages (IDs 112-119, 240-247)
A range of Message IDs is reserved for proprietary messages. It is the responsibility of the STU to
address Proprietary Messages to the appropriate destination. An SRU shall either process or forward
a proprietary message. A proprietary message shall not be broadcast.
9.5.5.7.24 Proprietary external message (ID 120)
Support for external data ports is optional. No interface for an external data port is specified in this
Recommendation. If an STU does not have an external data port, then it shall ignore any received
Proprietary External Messages.
Table 9-30/G.991.2 – External information field

1 Message ID 120 – External Message ID
2 Logical Port Number Unsigned char
3.. (N + 2) External message data (N octets)
9.5.5.7.25 G.997.1 external message (ID 121)

Support for G.997.1 [3] external messaging is optional. The interface for G.997.1 messages is
beyond the scope of this Recommendation. If an STU does not have an interface for G.997.1
messaging, it shall ignore any received G.997.1 External Messages.
Logical port number FF16 is reserved for indicating the transport of SNMP packets, as described
in 6.4/G.997.1. SNMP packets may be transmitted using one or more such messages.
Table 9-31/G.991.2 – G.997.1 external information field

2 Logical Port Number Unsigned char
3.. (N + 2) G.997.1 External message data (N octets)
56 ITU-T Rec. G.991.2 (12/2003)

9.5.5.7.26 Generic Unable to Comply (UTC) Message (ID 144)
The Generic UTC message should be sent back to the source unit in the event that the destination
unit is unable to comply with the request. In this case, the definition of UTC is vendor dependent.
Note that this message is not meant to replace the UTC bit in those response messages that contain a
UTC bit.
Table 9-32/G.991.2 – Generic Unable to Comply (UTC) information field

1 Message ID 144 – Generic UTC Message ID
2 Message ID of request message Unsigned char
9.5.5.7.27 Mapping Request – Message ID 23

The Mapping Request Message is used to determine the mapping between the physical pair (or
loop) number labelled on the equipment and the logical wire pair (or loop) ordinal number (7.2.1.5).
While this mapping is vendor specific, this information is useful for troubleshooting circuits. The
response to this request shall be message ID 151.
Table 9-32a/G.991.2 – Mapping Request information field

1 23 Message ID
9.5.5.7.28 Mapping Response – Message ID 151

The Mapping Response Message is sent in response to a Mapping Request Message (message
ID 23). It is used to determine the mapping between the physical pair (or loop) number and the
logical wire pair (or loop) ordinal number. The physical pair number is the number labelled
externally on the equipment. The logical wire pair number is determined from bits 4145 to 4146 of
the activation frame from the STU-R device as specified in 7.2.1.5. The physical pair number is
composed of two octets, with the first octet containing the most significant byte, and the second
octet containing the least significant byte. For example, if the 16-bit number in octets 3/4 contains
the value 4, then logical wire pair 1 from 7.2.1.5 is transported over the equipment's physical pair
labelled number 4. If the responding unit is a repeater, then the mapping response information for
the network side of the repeater will be sent first with bit 3 of octet two set to zero, followed
immediately by the mapping response information for the customer side of the repeater with bit 3 of
octet two set to one.
ITU-T Rec. G.991.2 (12/2003) 57

Table 9-32b/G.991.2 – Mapping Response information field
1 151 Message ID
2 bit 7 Response Side Bit 0 = network side
information
1 = customer side
information
2 bits 6-3 Reserved
2 bits 2-0 Number of Wire Pairs, M Unsigned 1 = 1 pair
2 = 2 pair
3 = 3 pair
4 = 4 pair
3-4 Physical Pair Number Corresponding Unsigned Char
to Logical Wire Pair 1 (7.2.1.5)
5-6 Physical Pair Number Corresponding Unsigned Char
to Logical Wire Pair 2 (7.2.1.5)
. .
. .
. .
2×M+1– Physical Pair Number Corresponding Unsigned Char
2×M+2 to Logical Wire Pair M (7.2.1.5)
9.5.6 Examples of virtual terminal control functions

This informative note gives examples of some common ANSI X3.4-1986 (R1997) [B3] escape
sequences.
Table 9-33/G.991.2 – Examples of ANSI X3.4-1986 (R1997) control functions

Description Format Comments
Erase entire screen (ED) ESC [ 2 J
Position cursor (CUP) ESC [ RR;CCH (Note)
Position cursor (in column 1) ESC [ RRH Subset of Position cursor
Home cursor ESC [ H Subset of Position cursor
NOTE – ESC has the value of 1B16. RR is the row number; CC is the column number expressed
as ASCII digits. As an example, row 4 column 12 would encode as ESC [ 4;12H. The
hexadecimal equivalent of this sequence is 1B16 5B16 3416 3B16 3116 3216 4816. The screen starts
with row 1, column 1.
10 Clock architecture
10.1 Reference clock architecture

Due to the multiple applications and variable bit rates called for in SHDSL, a flexible clocking
architecture is required. The STU-C and STU-R symbol clocks are described in terms of their
allowed synchronization references.
58 ITU-T Rec. G.991.2 (12/2003)

The SHDSL reference configuration permits the flexibility to provide a symbol clock reference
based on the sources shown in Figure 10-1. It illustrates the clock reference options in the context of
a simplified SHDSL reference model. Table 10-1 lists the normative synchronization configurations
as well as example applications.
STU-C STU-R
Transmit clk Receive clk
Transmit data Receive data

SHDSL SHDSL
transceiver transceiver Transmit clk
Receive clk
Receive data Transmit data
Local oscillator Local oscillator

Network reference clk
(Start-up mode only)
Receive symbol clk
Figure 10-1/G.991.2 – Reference clock architecture
Table 10-1/G.991.2 – Clock synchronization configurations

Mode STU-C symbol STU-R symbol Example
Mode
number clock reference clock reference application
1 Local oscillator Received symbol "Classic" HDSL Plesiochronous
clock
2 Network reference Received symbol "Classic" HDSL Plesiochronous with
clock clock with embedded timing reference
timing reference
3a Transmit data clock Received symbol Main application is Synchronous
or network clock synchronous
reference clock transport in both
directions
3b Transmit data clock Received symbol Synchronous Hybrid:
clock downstream downstream:
transport and bit- synchronous
stuffed upstream is upstream:
possible plesiochronous
10.2 Clock accuracy

At all rates, the transmit symbol clock during data mode from any SHDSL device shall be accurate
to within ±32 ppm of the nominal frequency. During activation, the STU-C shall maintain ±32 ppm
accuracy of its transmit symbol clock, but the STU-R transmit symbol clock may vary up to
±100 ppm.
ITU-T Rec. G.991.2 (12/2003) 59

10.3 Definitions of clock sources
The following definitions shall apply to the clock sources shown in Figure 10-1.
10.3.1 Transmit symbol clock reference
A reference clock from which the actual transmit symbol clock is derived (i.e., the STU's transmit
symbol clock is synchronized to this reference).
10.3.2 Local oscillator
A clock derived from an independent local crystal oscillator.
10.3.3 Network reference clock
A primary reference clock derived from the network.
10.3.4 Transmit data clock
A clock that is synchronous with the transmitted data at the application interface.
10.3.5 Receive symbol clock
A clock that is synchronous with the downstream received symbols at the SHDSL line interface.
This clock is used as the transmit symbol clock reference in the STU-R.
10.3.6 Receive clock
A clock that is synchronous with the received data at the application interface.
10.4 Synchronization to clock sources

In synchronous mode, the STU-C can be synchronized to the transmit data clock or to a network
reference clock. If a network reference clock is used, the transmit data clock must be synchronized
to the network reference clock. (The various transmit data rates are independent of the reference
clock frequency.)
When available, the network reference clock shall be either a fundamental 8 kHz network clock or a
related reference clock at some multiple of 8 kHz. Such reference clocks are typically 1544 MHz or
2048 MHz, although in some applications other frequencies, such as 64 kHz, may be available.
These related clocks include implicit 8 kHz1 timing signals. Selection of a specific network clock
reference frequency shall be application dependent.
11 Electrical characteristics
This clause specifies conformance tests for SHDSL equipment. These out-of-service tests verify the
electrical characteristics of SHDSL metallic interfaces.
____________________
1 The 6 ms SHDSL frame for synchronous data transport and the network 8 kHz clock have a fixed
relationship. Each SHDSL frame contains 48(1 + i + n × 8) bits (i = 0 … 7 and n = 3 … 36, or, optionally,
n = 37 … 89, as described in Annex F). The relationship can be calculated with: T = 6 ms/48 = 125 µs and
f = 1/T = 8 kHz. At the STU-R, an 8 kHz clock signal can be derived from the synchronous 6 ms frame.
60 ITU-T Rec. G.991.2 (12/2003)

11.1 Longitudinal balance
Longitudinal balance or longitudinal conversion loss (LCL) is a figure of merit describing the
coupling between longitudinal VL (common mode) and metallic VM (normal mode) signal
components. This term is equivalent to the familiar common mode rejection ratio (CMRR) and
defined as follows:
V
Longitudinal Balance (dB) = 20 log L
VM
Longitudinal balance at the SHDSL loop interface shall be measured with a coupling circuit having
a metallic termination of 135 Ω and a longitudinal termination of 33.8 Ω (Figure 11-1). Example
coupling circuits are shown in Appendix I. This test shall be performed with the DUT transmitter
turned off (quiet mode) and with span power circuitry (in either CO and RT units) activated by an
appropriate external DC current source/sink. The active power feed requirement may be waived for
locally powered systems.
Figure 11-1/G.991.2 – Longitudinal balance measurement
The measured longitudinal balance at the SHDSL loop interface shall lie above the specified limit
mask defined in Figure 11-2. The values of the parameters in the figure are region-specific and are
specified in A.5.4 and B.5.4. The longitudinal test circuit shall be calibrated such that when a 135 Ω
resistor (placed across tip and ring) is substituted for the device under test and the DC current
source/sink is disconnected, the measured longitudinal balance shall be at least 20 dB above the
limit mask. The longitudinal balance shall be measured over the frequency range of 20 kHz to
2 MHz.
Figure 11-2/G.991.2 – Longitudinal balance limit mask
ITU-T Rec. G.991.2 (12/2003) 61

11.2 Longitudinal output voltage
Longitudinal output voltage at the SHDSL loop interface shall be measured with a coupling circuit
having a metallic termination of 135 Ω and a longitudinal termination of 33.8 Ω as shown in
Figure 11-3. Example coupling circuits are shown in Appendix I. This test shall be performed with
the transmitter active (sending random data) and the span power circuitry (in either CO and
RT units) activated by an appropriate external DC current source/sink. The active power feed
requirement may be waived for locally powered systems.
Figure 11-3/G.991.2 – Longitudinal output voltage measurement
The measured longitudinal output rms voltage at the SHDSL loop interface shall be less than
–50 dBV over any 4 kHz frequency band when averaged over one second periods. The
measurement frequency range is region-specific and is specified in A.5.5 and B.5.5.
11.3 Return loss

This test measures return loss at the SHDSL loop interface with respect to a 135 Ω reference (line)
impedance. In SHDSL applications, return loss is generally used as a measure of termination
impedance distortion (deviation in both magnitude and phase from the reference impedance value).
Return loss limits are necessary to prevent large termination mismatches between equipment from
compliant vendors. Return loss may be measured directly using an impedance analyser or indirectly
as a voltage output in a bridge circuit. For either method, care must be taken to prevent
measurement errors from possible unintentional circuit paths between the common ground of the
measuring instrument(s) and the DUT power feed circuitry. In addition, when measuring under span
powered conditions, the test instrument must be galvanically isolated from the loop interface to
prevent damaging the test equipment with the high voltage DC power feed. For measurements
performed with an impedance analyser return loss is defined as follows:
Z ( f ) + Z REF
Return Loss(f) = 20 log TEST
ZTEST ( f ) − Z REF
where:
ZTEST ( f ) measured complex impedance at frequency f at the DUT loop interface
Z REF reference impedance (135 Ω).
62 ITU-T Rec. G.991.2 (12/2003)

Figure 11-4/G.991.2 – Return loss impedance analyser test method
For measurements performed with a test bridge the return loss is defined as follows:
VIN ( f )
Return Loss(f) = 20 log
V OUT ( f )
An example return loss test bridge is shown in Appendix I.
Figure 11-5/G.991.2 – Return loss bridge test method
The return loss test shall be performed with the DUT transmitter turned off (quiet mode). The DUT
may be tested span powered or locally powered as required by the intended application of the DUT.
For span powered applications, if the DUT is an STU-C the test shall be performed with the span
power supply activated and an appropriate DC current sink (with high AC impedance) attached to
the test circuit. If the DUT is an STU-R, the test shall be performed with power (DC voltage)
applied at the loop interface (TIP/RING) by an external voltage source feeding through an AC
blocking impedance. Note that the DC current source/sink must present a high impedance (at signal
frequencies) to common ground.
The nominal driving point impedance of the SHDSL loop interface shall be 135 Ω. Return loss shall
be measured with either the impedance analyser method of Figure 11-4 or the bridge method of
Figure 11-5. The measured return loss values relative to 135 Ω shall lie above the limit mask
specified in Figure 11-6. The values of the parameters are region-specific, and are specified in A.5.2
and B.5.2. The loop interface return loss shall be measured over the frequency range of 1 kHz to
2 MHz.
ITU-T Rec. G.991.2 (12/2003) 63

G.991.2_F11-6
Figure 11-6/G.991.2 – Return loss limit mask
11.4 Transmit power testing

The total average transmit power may be tested while span powered or locally powered as required
by the intended application of the DUT. For span powered applications, if the DUT is an STU-C the
test shall be performed with the span power supply activated and an appropriate DC current sink
(with high AC impedance) attached to the test circuit. If the DUT is an STU-R, the test shall be
performed with power (DC voltage) applied at the loop interface (TIP/RING) by an external voltage
source feeding through an AC blocking impedance. The test circuit must contain provisions for DC
power feed and possibly transformer isolation for the measurement instrumentation. Note that the
DC current source/sink must present a high impedance (at signal frequencies) to common ground.
Figure 11-7/G.991.2 – PSD/total power measurement set-up
11.4.1 Test circuit

The test circuit must contain provisions for DC power feed and possibly transformer isolation for
the measurement instrumentation. Transformer isolation of the instrumentation input prevents
measurement errors from unintentional circuit paths through the common ground of the
instrumentation and the DUT power feed circuitry. When the driving point impedance of the test
circuit meets the calibration requirements defined in 11.4.2, the test circuit will not introduce more
than ±0.25 dB error with respect to a perfect 135 Ω test load. An example test circuit is shown in
Appendix I. Note that the same circuit may be used for measuring total transmit power and
transmit PSD.
64 ITU-T Rec. G.991.2 (12/2003)

11.4.2 Test circuit calibration
The nominal driving point impedance of the test circuit shall be 135 Ω. The minimum return loss
with respect to 135 Ω over the frequency band of 3 kHz to 3 MHz shall be 35 dB from 10 kHz to
500 kHz with a slope of 20 dB/decade below and above these corner frequencies.
NOTE – 35 dB return loss will allow ±0.20 dB measurement error with respect to the nominal 135 Ω value.
11.4.3 Total transmit power requirement
The average transmit power of the STU-C shall be measured while continuously sending either
signal Sc (6.2.2.2) or signal Datac (6.2.2.7). If Datac is used, the total power measured into 135 Ω
shall fall in the range (PSHDSL ± 0.5 dB) as specified in A.4 and B.4. If Sc is used, the total power
measured into 135 Ω shall fall in the range (PSHDSL – 0.2 dB ± 0.5 dB). The average transmit power
of the STU-R shall be measured while continuously sending either signal Sr (6.2.2.3) or signal Datar
(6.2.2.7). If Datar is used, the total power measured into 135 Ω shall fall in the range
(PSHDSL ± 0.5 dB) as specified in A.4 and B.4. If Sr is used, the total power measured into 135 Ω
shall fall in the range (PSHDSL – 0.2 dB ± 0.5 dB). This power measurement in activation mode will
be 0.2 dB lower than the associated data mode transmit power due to the 2-PAM constellation
definition.
The transmit power spectral density of the STU-C shall be measured while continuously sending
either signal Sc (6.2.2.2) or signal Datac (6.2.2.7). The transmit power spectral density of the STU-R
shall be measured while continuously sending either signal Sr (6.2.2.3) or signal Datar (6.2.2.7). If
Datac or Datar is used, the measured transmit PSD into 135 Ω shall remain below the corresponding
PSDMask(f) from A.4 and B.4. If Sc or Sr is used, the measured transmit PSD into 135 Ω shall
remain below the corresponding PSDMask(f) from A.4 and B.4 reduced by 0.2 dB in the passband
(i.e., PSDMask(f) with PBO increased by 0.2 dB).
11.4.3.1 Transmit power spectral density test procedure
The transmit power spectral density (PSD) may be tested span powered or locally powered as
required by the intended application of the DUT. For span powered applications, if the DUT is an
STU-C the test shall be performed with the span power supply activated and an appropriate DC
current sink (with high AC impedance) attached to the test circuit. If the DUT is an STU-R, the test
shall be performed with power (DC voltage) applied at the loop interface (TIP/RING) by an
external voltage source feeding through an AC blocking impedance.
The transmit power spectral density for the STU-C and STU-R shall be measured with signals as
defined in 11.4.3. The transmit power spectral density shall be measured over the frequency range
of 1 kHz to 3 MHz. The STU-C transmit signal shall be compliant with the appropriate A.4 or B.4
PSD requirements. The STU-R transmit signal shall be compliant with the appropriate A.4 or B.4
PSD requirements.
11.4.3.2 PSD test circuit and calibration
The test circuit must contain provisions for DC power feed and possibly transformer isolation for
the measurement instrumentation. Transformer isolation of the instrumentation input prevents
instrumentation and the DUT power feed circuitry. The test circuit shall meet the requirements
of 11.4.2.
11.5 Signal transfer delay

The STU shall be capable of providing PMD-layer one-way, single-span latency of 500 µs or less
for user data rates of 1.5 Mbit/s and above, and 1.25 ms or less for user data rates below 1.5 Mbit/s
as measured between the α and ß interfaces.
ITU-T Rec. G.991.2 (12/2003) 65

12 Conformance testing
12.1 Micro-interruptions
A micro-interruption is a temporary interruption due to external mechanical action on the copper
wires constituting the transmission segment, for example, at a cable splice. Splices can be
hand-made wire-to-wire junctions, and during cable life oxidation phenomena and mechanical
vibrations can induce micro-interruptions at these critical points. Example causes of this impairment
include a large motor vehicle driving over a buried cable installation or an aerial cable movement
from wind forces.
The effect of a micro-interruption on the transmission system can be a failure of the digital
transmission link, together with a failure of the span power feeding (if provided) for the duration of
the micro-interruption. The operating objective is that in the presence of a micro-interruption of
specified maximum length the system shall not reset, and the system shall automatically reactivate
with a complete start-up procedure if a reset occurs due to an interruption.
The configuration for micro-interruption susceptibility testing is shown in Figure 12-1. In this
arrangement, a periodic trigger signal S stimulates a normally closed micro-relay device inducing
periodic micro-interruptions on the transmission link. Note that the micro-interruptions are induced
on one termination at a time. The test loops shall be composed of 1.5 km of 0.4 mm (or 5000' of
26 AWG) copper wire, and the tests shall be conducted at the maximum supported data rate. Using
the test arrangement as described in Figure 12-1 with local powering on, the SHDSL transceivers
shall not be reset by a micro-interruption of at least t = 10 ms when stimulated with a signal of
period T = 5 s for a test interval of 60 s at a single termination. The micro-interruptions shall be
induced at both the STU-C and STU-R terminations. This test shall be repeated with span-powering
on and a micro-interruption of at least t = 1 ms.
STU-C Test loop STU-R
Figure 12-1/G.991.2 – Micro-interruption test circuit
66 ITU-T Rec. G.991.2 (12/2003)

Annex A
Regional requirements – Region 1

A.1 Scope
This annex describes those specifications that are unique to SHDSL systems operating under
conditions such as those typically encountered within the North American network. The clauses in
this annex provide the additions and modifications to the corresponding clauses in the main body.
A.2 Test loops

The primary constants for the following test loops are listed in Annex A/G.996.1 [6]. Note that the
test loops shown in Figure A.1 are PIC and specified at 70° F (21.1º C). Loop 0 is the null loop:
≤ 10' and ≤ 26 AWG.
Figure A.1/G.991.2 – Test loops
ITU-T Rec. G.991.2 (12/2003) 67

A.3 Performance Tests
This clause specifies performance tests for SHDSL equipment. These out-of-service tests verify the
performance of SHDSL in impaired environments.
Figure A.2 shows the test set-up for measuring the performance of SHDSL systems in the presence
of noise impairments. The test system consists of an SHDSL central office transceiver (STU-C) and
a remote end transceiver (STU-R). The SHDSL transceivers are connected by a test loop. Simulated
noise is locally injected into the test loop through the specified coupling circuit at the receiving
transceiver.
Bit error ratio (BER) measurement is performed by applying a pseudo-random binary sequence
(PRBS) test signal at one transceiver input and detecting errors in the received PRBS data stream of
the other transceiver. The PRBS signal shall have a minimum period of 223 – 1. BER measurement
shall be performed for both directions of transmission and the tests in each direction shall be
performed in full-duplex mode with both SHDSL transceivers simultaneously transmitting data. In
all cases these noise impairment tests shall be performed one unit at a time (i.e., the STU-C and
STU-R are not impaired simultaneously) and with noise from only one impairment source active at
a time.
Figure A.2/G.991.2 – Crosstalk margin and impulse noise test set-up
68 ITU-T Rec. G.991.2 (12/2003)

A.3.1 Crosstalk margin tests
A.3.1.1 Crosstalk noise injection
Simulated crosstalk (NEXT and FEXT) is introduced by injecting a calibrated, filtered Gaussian
noise source into the test circuit. The crosstalk shall be locally injected into the test loop at the
receiving transceiver through a balanced high-impedance parallel-connected feed network. The
high-impedance parallel-connected feed network allows injection of the desired crosstalk power
level without disturbing the transmission characteristics or driving point impedance of the test loop.
The injection circuit shall have a Thevenin output impedance of at least 4 kΩ. An example crosstalk
signal injection circuit is shown in Figure I.1.
A.3.1.2 Calibration accuracy of crosstalk generator
The simulated crosstalk shall have the total power and the power spectral density (PSD) defined
in A.3.3. However, if the method of generating simulated crosstalk is as defined in Figure A.2, then
the power level and PSD accuracy will depend on the accuracy of the filters designed to shape the
white noise for each injected crosstalk source. The highest level of accuracy is required within the
frequency band (or bands) corresponding to the largest values of the PSD for each crosstalk source.
For each specified crosstalk source, the accuracy of the simulated PSD obtained shall be ±1.0 dB
within the ideal PSD template (defined by the equations in A.3.3) over the frequency band(s) where
the ideal PSD template is within 30 dB of its maximum value. The measured average power
(integral of the crosstalk PSD function) for each specified crosstalk source shall be within ±0.25 dB
of the integrated power of the ideal specified crosstalk PSD template (A.3.3).
The white noise source of Figure A.2 shall cover the frequency band from DC to 1.5 MHz and have
a Gaussian amplitude distribution with a crest factor of at least 5.0.
A.3.1.3 Calibration measurement of crosstalk generator
The PSD and average power for each crosstalk test scenario shall be calibrated by measuring the
output of the crosstalk injection circuit with the test loop replaced by a load of two parallel 135 Ω
resistors (67.5 Ω) and no connected terminal equipment. The two parallel 135 Ω resistors simulate
the terminating load of a zero-length loop. The crosstalk signal shall be measured as a voltage by a
high-impedance frequency-selective voltmeter (i.e., spectrum analyser) and converted into a power
level assuming a 135 Ω reference impedance. This procedure effectively measures the crosstalk
power fed into a single resistor (one side of the loop only). The measured crosstalk PSD(s) and
average crosstalk power(s) coupled into the calibration load must remain within the limits defined
in A.3.1.2 for each specified crosstalk scenario defined in A.3.1.6.
NOTE – The injected noise is intended to match the theoretical noise PSD when the transceiver under test is
connected to the loop. On Loop S for payload rates of 1024 kbit/s and below, and on all loops for payload
rate of 192 kbit/s, it has been found that impedance mismatch could generate an increased noise PSD at low
frequencies. One method of compensation is to modify the factor, ∆, defined in A.3.1.4 by replacing the
theoretical noise, N(f), in step 3 of A.3.1.4 with the noise PSD measured when connected to the loop under
test. A second method is to place a passive circuit, consisting of a resistor R in parallel with a capacitor C, in
series with each wire of the noise generator output pair. The RC values of R = 1.2 kΩ and C = 1 µF are
suggested and should be adjusted for each noise generator such that the injected noise matches the theoretical
noise PSD. A third method is to calibrate the noise generator waveform into the loop under test such that
when connected to the loop under test, the theoretical noise waveform is present at the transceiver terminals.
A.3.1.4 Calibration of loop simulator
There is significant variation in loop insertion loss for the same loop model on loop simulators from
both different and identical manufacturers. Typical loop simulators may exhibit insertion loss
variations greater than ±1.0 dB of the ideal loop model over the SHDSL signal band. Insertion loss
variation of loop simulators may cause significant variation of measured system noise margin. To
ITU-T Rec. G.991.2 (12/2003) 69

minimize measurement variation caused by the loop simulator, the crosstalk generator output power
may be adjusted to maintain a consistent SNR at the receiver input. The calibration procedure is as
follows:
1) Given the discrete form of the DFE-based SNR formula, SNRdB, given below:

( ) (
 S f sym − f k H f sym − f k )2
S ( fk ) H ( fk )
2 

1+ + +
1 M

 (
N f sym − f k ) N ( fk )


SNRdB = ∑ 10 log10
 
M k =1 ( ) (
 S 2 f sym − f k H 2 f sym − f k)2
( ) ( )2
S f sym − f k H f sym − f k 
+

 ( )
N 2 f sym − f k ( )
N f sym − f k 

calculate SNR1, the ideal receive signal-to-noise ratio, by setting SNR1 equal to SNRdB
where S(f) shall be the nominal far-end transmit signal power spectral density
(NominalPSD(f) from A.4), |H(f)|2 shall be the magnitude squared of the ideal loop
insertion gain function, N(f) shall be the injected crosstalk noise power spectral density
(PSDCase-n(f) from A.3.3.9), and fsym shall be the transmit symbol rate. For this application
use fk = k × 1000, k = 1…M, where M is the maximum value of k such that M × 1000 < fsym
≤ (M + 1) × 1000. The ideal loop insertion gain function shall be calculated from the
primary constants of twisted pair copper as defined in Annex A/G.996.1 [6].
2) Measure the insertion loss of the loop simulator with 135 Ω terminations at points f k
defined in step 1. Note that the termination return loss with respect to 135 Ω should be
greater than 35 dB from 20 kHz to fsym to ensure insertion loss measurement accuracy
within 0.25 dB over the main part of the SHDSL signal band. An example insertion loss
measurement set-up is shown in Figure A.3. The measured loss in dB of the loop at each
frequency shall be within 5% (in dB) of the theoretical loop insertion loss function as
calculated in step 1. As the measurements for the calibration procedure are easily made
with error, the return loss measurement set used to verify the 35 dB return loss of the test
fixture terminations shall be calibrated with a known return loss test load of at least 55 dB
over the range of 20 kHz to 500 kHz. In addition, the line simulator should exhibit a
longitudinal balance of 35 dB or better for frequencies in the range of 0 to fsym.
Figure A.3/G.991.2 – Example loop insertion loss measurement set-up
70 ITU-T Rec. G.991.2 (12/2003)

3) Calculate SNR2, the measured receive signal-to-noise ratio, by setting SNR2 equal to SNRdB
from step 1 where |H(f)|2 is the magnitude squared of the measured loop insertion gain
function from step 2 above, and S(f), N(f), fsym, and fk are the same as in step 1 above.
4) Adjust the noise margin target in Table A.1 by ∆ = (SNR2 – SNR1) dB. Note that a
negative difference corresponds to a decrease in crosstalk generator power. Note that this
procedure assumes the crosstalk generator was previously calibrated as per A.3.1.2
and A.3.1.3. All crosstalk power adjustments shall be limited to 3.0 dB maximum. Test
set-ups requiring greater than 3.0 dB crosstalk power adjustment shall not be valid.
A.3.1.5 Crosstalk margin compliance procedure
The SHDSL transceivers shall have noise margins that meet or exceed the values listed in Table A.1
for the specified test loop and crosstalk combinations. The definitions of the test loops are given in
Figure A.1, and specifications for the crosstalk PSDs are given in A.3.3. The test for noise margin
compliance shall be defined as follows:
1) Calibrate the crosstalk injection circuit (using the calibration load of 67.5 Ω) to the
corresponding PSD and total power value specified in A.3.3.
2) Increase the injected crosstalk power by the corresponding noise margin value specified in
Table A.1.
3) Using the test set-up from Figure A.2, activate the SHDSL transceivers and allow a
minimum 5-minute fine-tuning period.
4) Measure the BER over a minimum of 109 bits.
5) The measured BER at each end shall be less than 10–7.
A.3.1.6 Crosstalk interference requirements
Table A.1 shows the minimum set of test loops and crosstalk combinations required for testing
SHDSL margins. A compliant unit shall pass the BER test described in A.3.1.5 for all crosstalk
scenarios and test loops defined in Table A.1. 0 dB Power Backoff shall be used for both the STU-C
and STU-R.
Table A.1/G.991.2 – Crosstalk scenarios & required SHDSL noise margins (Note)
Test
Payload
loop Required
L Test data Interferer
Test (from PSD margin
× 1000')
(× unit rate combination
Figure (dB)
(kbit/s)
A.1)
1 C4 – STU-C 1544 Asymmetric 24T1 + 24 SHDSL 5 + ∆*
2 C4 – STU-C 1544 Asymmetric 39 SHDSL 5 + ∆*
3 C4 – STU-C 1544 Asymmetric 24 FDD ADSL + 5 + ∆*
24 HDSL
4 S 9.0 STU-C 1544 Asymmetric 24T1 + 24 SHDSL 5 + ∆*
5 S 9.0 STU-C 1544 Asymmetric 39 SHDSL 5 + ∆*
6 S 9.0 STU-C 1544 Asymmetric 24 FDD ADSL + 5 + ∆*
24 HDSL
7 C4 – STU-R 1544 Asymmetric 24T1 + 24 SHDSL 5 + ∆*
8 S 9.0 STU-R 1544 Asymmetric 24T1 + 24 SHDSL 5 + ∆*
ITU-T Rec. G.991.2 (12/2003) 71

Test
Payload
loop Required
× 1000')
Figure (dB)
(kbit/s)
A.1)
9 S 6.3 STU-C 2304 Symmetric 24-T1 + 24 SHDSL 5 + ∆*
asym 1544
10 BT1-C 5.2 STU-C 2304 Symmetric 24-T1 + 5 + ∆*
24 SHDSL
asym 1544
11 BT1-C 5.2 STU-C 2304 Symmetric 49-SHDSL 5 + ∆*
12 S 6.3 STU-R 2304 Symmetric 49-SHDSL 5 + ∆*
13 BT1-R 5.2 STU-R 2304 Symmetric 49-SHDSL 5 + ∆*
14 BT1-R 5.2 STU-R 2304 Symmetric 24-T1 + 5 + ∆*
24 SHDSL
asym 1544
15 S 6.8 STU-C 2048 Symmetric 24-SHDSL + 5 + ∆*
24-FDD ADSL
17 BT1-C 5.6 STU-C 2048 Symmetric 24-T1 + 5 + ∆*
24 SHDSL
asym 1544
24 SHDSL
asym 1544
21 S 7.9 STU-C 1544 Symmetric 39-SHDSL 5 + ∆*
asym 1544
22 BT1-C 6.4 STU-C 1544 Symmetric 24-FDD ADSL + 5 + ∆*
24 SHDSL
asym 1544
23 BT1-C 6.4 STU-C 1544 Symmetric 24-SHDSL + 5 + ∆*
24-FDD ADSL
24 SHDSL
asym 1544
27 S 11.0 STU-C 768 Symmetric 49-HDSL 5 + ∆*
29 BT1-C 10.2 STU-C 768 Symmetric 49-HDSL 5 + ∆*
30 S 11.0 STU-R 768 Symmetric 49-HDSL 5 + ∆*
32 BT1-R 10.2 STU-R 768 Symmetric 49-HDSL 5 + ∆*
72 ITU-T Rec. G.991.2 (12/2003)

Test
Payload
loop Required
× 1000')
Figure (dB)
(kbit/s)
A.1)
33 S 11.2 STU-C 768 Asymmetric 49-HDSL 5 + ∆*
34 BT1-C 10.4 STU-C 768 Asymmetric 49-HDSL 5 + ∆*
35 BT1-C 10.4 STU-C 768 Asymmetric 24-FDD ADSL + 5 + ∆*
24-HDSL
36 S 11.2 STU-R 768 Asymmetric 24-T1 + 24 HDSL 5 + ∆*
37 BT1-R 10.4 STU-R 768 Asymmetric 24-T1 + 5 + ∆*
24-SHDSL
38 BT1-R 10.4 STU-R 768 Asymmetric 39-FDD ADSL 5 + ∆*
39 S 14.8 STU-C 384 Symmetric 24-SHDSL + 5 + ∆*
24-DSL
24-DSL
42 S 14.8 STU-R 384 Symmetric 24-SHDSL + 5 + ∆*
24-DSL
43 BT2-R 13.8 STU-R 384 Symmetric 24-SHDSL + 5 + ∆*
24-DSL
45 S 17.2 STU-C 256 Symmetric 49-DSL 5 + ∆*
46 BT2-C 16.4 STU-C 256 Symmetric 49-DSL 5 + ∆*
24-DSL
48 S 17.2 STU-R 256 Symmetric 49-DSL 5 + ∆*
49 BT2-R 16.4 STU-R 256 Symmetric 49-DSL 5 + ∆*
50 BT2-R 16.4 STU-R 256 Symmetric 24-SHDSL + 5 + ∆*
24-DSL
51 S 19.8 STU-C 192 Symmetric 49-DSL 5 + ∆*
52 BT2-C 19.1 STU-C 192 Symmetric 49-DSL 5 + ∆*
53 BT2-C 19.1 STU-C 192 Symmetric 24-DSL + 5 + ∆*
24 SHDSL
54 S 19.8 STU-R 192 Symmetric 49-DSL 5 + ∆*
55 BT2-R 19.1 STU-R 192 Symmetric 49-DSL 5 + ∆*
56 BT2-R 19.1 STU-R 192 Symmetric 24-DSL + 5 + ∆*
24 SHDSL
NOTE – The crosstalk scenarios listed in this table were developed under the assumption of a
50-pair cable binder. Cable binders of other sizes are for further study.
*
The indicated noise margins in Table A.1 shall have a tolerance of 1.25 dB due to the aggregate
effect of crosstalk generator tolerance and calibrated loop simulator tolerance. The offset ∆ is
defined in A.3.1.4.
ITU-T Rec. G.991.2 (12/2003) 73

All interferers are assumed to be co-located. The notation 24 or 49 SHDSL refers to SHDSL at the
same rate and PSD as the system under test. All interferer PSDs are described in A.3.3.9.
The process for selecting which tests to perform for a specific G.991.2 device under test (DUT) is
determined by following each of these 6 steps in order:
1) Determine the set of rates which are in common between the set of supported payload data
rates and the following set of payload data rates: (symmetric PSD: 192, 256, 384, 768,
1544, 2048, 2304 kbit/s; asymmetric PSD: 768, 1544 kbit/s). Call the resulting list of
common rates the intersection list.
2) If 1544 kbit/s asymmetric is in the intersection list, then test the DUT with test cases 1-8 in
Table A.1.
3) If 768 kbit/s asymmetric is in the intersection list, then test the DUT with test cases 33-38
in Table A.1.
4) If 1544 kbit/s symmetric is in the intersection list, then test the DUT with test cases 21-26.
5) For the highest and the lowest symmetric PSD rate in the intersection list, test the DUT
with all six cases associated with that rate. For example, if 192 kbit/s symmetric is the
lowest rate and 2304 kbit/s symmetric is the highest rate, then test with test cases 51-56 and
9-14 in Table A.1.
6) For all remaining rates in the intersection list that have not been tested, test using the cases
involving only Loop S. For example, if 256, 384, 768 and 2048 kbit/s symmetric are the
remaining rates, then test with the additional test cases 48, 45, 42, 39, 30, 27, 18 and 15.
If all rates are implemented by the DUT, there will be a total of 40 tests.
A.3.2 Impulse noise tests
A.3.2.1 Impulse noise test procedure
The impulse noise waveform V(t) (hereafter called the "test impulse") is defined as:
 K t −3 / 4 t > 0
 
V (t ) = 0 t = 0
 −3 / 4 
− K t t < 0
where t is time given in units of seconds and K is a constant defined numerically in Table A.2. If the
T
pulse is realized using discrete samples of V(t), the waveform should be sampled at t = (2n − 1) ,
2
where T is the sampling period and (1/T) should be at least twice the symbol rate of the system
under test. The sampled peak-to-peak amplitude will vary with sampling rate. It can be calculated
3
−
T 4
using the following formula: V p − p = 2 K .
2
Table A.2/G.991.2 – Impulse noise peak-to-peak voltage requirement

K VP-P of the test impulse sampled at 2 Msamples/s
–6
1.775 × 10 320 mV
For a sampling rate of 2 Msamples/s, a minimum of 8000 samples is required with an amplitude
accuracy of at least 12 bits. Figure A.4 shows the test impulse sampled at 2 Msamples/s. The
injection circuit shall be identical to that described in A.3.1.
74 ITU-T Rec. G.991.2 (12/2003)

Figure A.4/G.991.2 – Time domain representation of the test pulse sampled at 2 Msamples/s
A.3.2.2 Impulse noise test performance

A compliant unit shall pass the impulse noise test specified in Table A.3. The minimum test period
shall be 10 s. Each SHSDL termination shall be tested independently, i.e., the impulse noise
waveform is not injected at both terminations simultaneously.
Table A.3/G.991.2 – Impulse noise test criteria

Test pulse VP-P when
Test pulse repetition Bit error ratio
Test loop sampled at
rate upper limit
2 Msamples/s
Loop C4 320 mV 10 Hz −4
5.0 × 10
Loop S, L = 9000' 320 mV 10 Hz −4
5.0 × 10
NOTE – The entries in this table only correctly apply to the 1544 kbit/s asymmetric case.
Appropriate values for other rates and PSDs are for further study.
ITU-T Rec. G.991.2 (12/2003) 75

A.3.3 Power spectral density of crosstalk disturbers
A.3.3.1 Simulated HDSL PSD
The PSD of HDSL disturbers shall be expressed as:
2
  πf  
 sin   
2   f 0   1
PSDHDSL = K HDSL × × × , f 3dB = 196 kHz, 0 ≤ f < ∞
f 0   πf    f 
8
    1 +  
  f 0   f
 3dB 
where
2
5 Vp
f 0 = 392 kHz, K HDSL = × ,V p = 2.70V, and R = 135 Ω
9 R
This equation gives the single-sided PSD; that is, the integral of PSD, with respect to f, from 0 to
infinity, gives the power in Watts. PSDHDSL is the PSD of a 392 ksymbol/s 2B1Q signal with
random equiprobable levels, with full-band square-topped pulses and with 4th order Butterworth
filtering (f3 dB = 196 kHz).
A.3.3.2 Simulated T1 line PSD
The PSD of the T1 line disturber is assumed to be the 50% duty-cycle random Alternate Mark
Inversion (AMI) code at 1.544 Mbit/s. The single-sided PSD shall be expressed as:
2
  πf  
 sin   
V p 2   f 0  
2
 πf  1 f2
PSDT 1 = × × sin 2   × × ,0 ≤ f < ∞
RL f 0   πf    2 f 0   f 
6
f 2
+ f 2
    1 +  
c
f
  0    f 3 dB 
 
where
Vp = 3.6 V, RL = 100 Ω, and f0 = 1.544 MHz.
The formula assumes that transmitted pulses are passed through a low-pass shaping filter. The
shaping filter is chosen as a 3rd order low-pass Butterworth filter with 3 dB point at 3.0 MHz. The
filter magnitude squared transfer function is:
2 1
H shaping ( f ) = 6
 f 
1+  
 f 3 dB 
 
The formula also models the coupling transformer as a high-pass filter with 3 dB point at 40 kHz
using:
2 f2
H Transformer ( f ) =
f 2 + f c2
A.3.3.3 Simulated ADSL downstream Frequency Division Duplex (FDD) PSD

The ADSL Downstream FDD PSD is based on the ATU-C transmitter PSD mask for reduced
NEXT defined in Figure A.2/G.992.1 [1]. The simulated PSD used for SHDSL performance testing
shall be defined as this G.992.1 mask reduced by 3.5 dBm/Hz over all frequencies.
76 ITU-T Rec. G.991.2 (12/2003)

A.3.3.4 Simulated ADSL upstream PSD
The ADSL Upstream PSD is based on the ATU-R transmitter PSD mask defined in
Figure A.3/G.992.1 [1]. The simulated PSD used for SHDSL performance testing shall be defined
as this G.992.1 mask reduced by 3.5 dBm/Hz over all frequencies.
A.3.3.5 Simulated SHDSL upstream PSD
The SHDSL Upstream PSD masks are defined in A.4. The simulated PSD used for SHDSL
performance testing shall be the worst-case ensemble summation of the nominal upstream PSDs
from A.4, with PBO set to 0 dB. The nominal PSD is given by the expression NominalPSD(f)
in A.4.1, A.4.2 and A.4.3.
A.3.3.6 Simulated SHDSL downstream PSD
The SHDSL Downstream PSD masks are defined in A.4. The simulated PSD used for SHDSL
performance testing shall be the worst-case ensemble summation of the nominal downstream PSDs
from A.4, with PBO set to 0 dB. The nominal PSD is given by the expression NominalPSD(f)
in A.4.1, A.4.2 and A.4.3.
A.3.3.7 Simulated DSL PSD
The power spectral density (PSD) of basic access DSL disturbers is expressed as:
2
  πf  
 sin   
2   f 0   1
PSDDSL-Disturber = K DSL × × × , f 3 dB = 80 kHz, 0 ≤ f < ∞
f 0   πf   4
    1 +  f 
  f 0    f 3 dB 
 
where
2
5 Vp
f 0 = 80 kHz, K DSL = × , V p = 2.50 V, and R = 135 Ω
9 R
This equation gives the single-sided PSD; that is, the integral of PSD, with respect to f, from 0 to
infinity, gives the power in Watts. PSDDSL-Disturber is the PSD of an 80 ksymbol/s 2B1Q signal with
random equiprobable levels, with full-band square-topped pulses and with 2nd order Butterworth
filtering (f3 dB = 80 kHz).
A.3.3.8 NEXT
The NEXT power transfer function uses the two-piece Unger model which has a slope of
14 dB/decade for frequencies greater than 20 kHz and a slope of 4 dB/decade for frequencies less
than or equal to 20 kHz. This is defined as follows where N is the total number of NEXT disturbers:
2 
−10 0. 4 0. 6
4.6288 × 10 × f × N , f ≤ 20 kHz
H NEXT −2− Piece ( f , N ) = 
2.3144 × 10 −14 × f 1.4 × N 0.6 , f > 20 kHz
The two-piece Unger model shall be used to model crosstalk when evaluating performance of the
1.536 or 1.544 Mbps asymmetric PSD.
The one-piece model for NEXT power transfer function is defined as follows where N is the total
number of NEXT disturbers:
2
H NEXT −1− Piece ( f , N ) = 0.8536 × 10−14 × f 1.5 × N 0.6
ITU-T Rec. G.991.2 (12/2003) 77

The one-piece model shall be used to model crosstalk when evaluating performance for all rates and
PSDs except the 1.536 or 1.544 Mbit/s asymmetric PSD.
The model for FEXT power transfer function is defined as follows where N is the total number of
FEXT disturbers:
2 2
H FEXT ( f , N , L, D) = L( f ) × D × 7.744 × 10 − 21 × f 2 × N 0.6
where L(f) is the insertion loss of the loop through which the interferer passes while the interferer
and the signal under test are adjacent in the same binder, and D is the length of the loop in feet. The
FEXT model shall be used to model crosstalk from asymmetric interferers (specifically
1.544 Mbit/s asymmetric and ADSL).
A.3.3.9 Crosstalk PSD definitions
The following PSD definitions are to be used to generate the crosstalk interferer combinations used
for performance testing in Table A.1.
24 × PSDT 1 ( f ) + 24 × PSDSHDSL −1544− Asym− Down ( f ) 2
PSD Case−1= × H NEXT −2− Piece ( f , 48) +
48
PSDSHDSL−1544− Asym−Up ( f ) × H FEXT ( f , 24, LC 4 , 7600) 2
PSD Case−2 = PSDSHDSL−1544− Asym− Down ( f ) × H NEXT −2− Piece ( f , 39) 2 +
PSDSHDSL−1544− Asym−Up ( f ) × H FEXT ( f , 39, LC 4 , 7600) 2
24 × PSD ADSL − Down ( f ) + 24 × PSDHDSL ( f ) 2

PSD Case −3 = × H NEXT − 2 − Piece ( f , 48) +
48
PSD ADSL −Up ( f ) × H FEXT ( f ,24, LC 4 ,7600) 2
24 × PSDT 1 ( f ) + 24 × PSDSHDSL −1544 − Asym − Down ( f ) 2

PSD Case − 4 = × H NEXT − 2 − Piece ( f , 48) +
48
PSD ADSL −1544 − Asym −Up ( f ) × H FEXT ( f ,24, LS 9.0 ,9000) 2
PSD Case −5 = PSDSHDSL −1544 − Asym − Down ( f ) × H NEXT − 2 − Piece ( f , 39) 2 +
PSDSHDSL −1544 − Asym −Up ( f ) × H FEXT ( f , 39, LS 9.0 , 9000) 2
24 × PSD ADSL − Down ( f ) + 24 × PSDHDSL ( f ) 2

48
PSD ADSL −Up ( f ) × H FEXT ( f ,24, LS 9.0 ,9000) 2
24 × PSDT 1 ( f ) + 24 × PSDSHDSL −1544 − Asym −Up ( f ) 2

48
PSDSHDSL −1544 − Asym − Down ( f ) × H FEXT ( f , 24, LC 4 , 7600) 2

PSD Case −8 = × H NEXT − 2 − Piece ( f , 48) +
48
PSDSHDSL −1544 − Asym − Down ( f ) × H FEXT ( f , 24, LS 9.0 , 9000) 2
78 ITU-T Rec. G.991.2 (12/2003)

PSD Case −9 = × H NEXT −1− Piece ( f , 48) +
48

48
PSDSHDSL −1544 − Asym −Up ( f ) × H FEXT ( f , 24, LBT 1− C 5.2 , 5200) 2
PSD Case −11= PSDSHDSL − 2304 − Sym ( f ) × H NEXT −1− Piece ( f , 49) 2
PSD Case −12 = PSDSHDSL − 2304 − Sym ( f ) × H NEXT −1− Piece ( f , 49) 2

48
PSDSHDSL −1544 − Asym − Down ( f ) × H FEXT ( f , 24, LBT 1− R5.2 , 5200) 2
24 × PSD ADSL − Down ( f ) + 24 × PSDSHDSL − 2048− Sym ( f ) 2

48
PSD ADSL −Up ( f ) × H FEXT ( f ,24, LS 6.8 ,6800) 2

48
PSDSHDSL −1544 − Asym −Up ( f ) × H FEXT ( f , 24, LBT 1−C 5.6 , 5600) 2
PSD Case −18 = PSDSHDSL − 2048− Sym ( f ) × H NEXT −1− Piece ( f , 49) 2

PSD Case − 20 = × H NEXT −1− Piece ( f , 48) +
48
PSDSHDSL −1544 − Asym − Down ( f ) × H FEXT ( f , 24, LBT 1− R5.6 , 5600) 2
2
PSD Case − 21= PSDSHDSL −1544 − Asym − Down ( f ) × H NEXT −1− Piece ( f , 39) +
ITU-T Rec. G.991.2 (12/2003) 79

24 × PSD ADSL − Down ( f ) + 24 × PSDSHDSL −1544 − Asym − Down ( f ) 2
48
24 × PSD ADSL −Up ( f ) + 24 × PSDSHDSL −1544 − Asym −Up ( f ) 2
× H FEXT ( f , 48, LBT 1−C 6.4 , 6400)
48
24 × PSD ADSL − Down ( f ) + 24 × PSDSHDSL −1544 − Sym ( f ) 2

48
PSD ADSL −Up ( f ) × H FEXT ( f , 24, LBT 1−C 6.4 , 6400) 2
PSD Case − 24 = PSDSHDSL −1544 − Sym ( f ) × H NEXT −1− Piece ( f , 49) 2

48
PSDSHDSL −1544 − Asym − Down ( f ) × H FEXT ( f , 24, LBT 1− R 6.4 , 6400) 2
PSD Case − 26 = PSDSHDSL −1544 − Sym ( f ) × H NEXT −1− Piece ( f , 49) 2
PSD Case − 27 = PSDHDSL ( f ) × H NEXT −1− Piece ( f , 49) 2
PSD Case − 28 = PSDSHDSL − 768− Sym ( f ) × H NEXT −1− Piece ( f , 49) 2
PSD Case − 29 = PSDHDSL ( f ) × H NEXT −1− Piece ( f , 49) 2
PSD Case −30 = PSDHDSL ( f ) × H NEXT −1− Piece ( f , 49) 2
PSD Case −31= PSDSHDSL − 768− Sym ( f ) × H NEXT −1− Piece ( f , 49) 2
24 × PSD ADSL− Down ( f ) + 24 × PSDHDSL ( f ) 2

PSD Case−35 = × H NEXT −1− Piece ( f , 48) +
48
PSD ADSL−Up ( f ) × H FEXT ( f , 24, LBT1 −C10.4 , 10400) 2
80 ITU-T Rec. G.991.2 (12/2003)

24 × PSDT 1 ( f ) + 24 × PSDHDSL ( f ) 2
PSD Case −36 = × H NEXT −1− Piece ( f , 48)
48
24 × PSDT 1 ( f ) + 24 × PSDSHDSL−768− Asym−Up ( f ) 2
PSD Case−37 = × H NEXT −1− Piece ( f , 48) +
48
PSDSHDSL−768− Asym− Down ( f ) × H FEXT ( f , 24, LBT1 − R10.4 , 10400) 2
PSD Case −38 = PSD ADSL −Up ( f ) × H NEXT −1− Piece ( f , 39) 2 +
PSD ADSL − Down ( f ) × H FEXT ( f , 39, LBT 1− R10.4 , 10400) 2
24 × PSDDSL ( f ) + 24 × PSDSHDSL −384− Sym ( f ) 2

48
24 × PSDDSL ( f ) + 24 × PSDSHDSL −384 − Sym ( f ) 2
PSD Case − 40 = × H NEXT −1− Piece ( f , 48)
48
PSDCase − 41= PSDSHDSL −384 − Sym ( f ) × H NEXT −1− Piece ( f ,49) 2
24 × PSDDSL ( f ) + 24 × PSDSHDSL − 384 − Sym ( f ) 2

48
48
PSDCase − 44 = PSDSHDSL −384 − Sym ( f ) × H NEXT −1− Piece ( f , 49) 2
PSDCase − 45 = PSDDSL ( f ) × H NEXT −1− Piece ( f , 49) 2

48

48
PSDCase −51= PSDDSL ( f ) × H NEXT −1− Piece ( f , 49) 2
ITU-T Rec. G.991.2 (12/2003) 81

PSDCase −52 = PSDDSL ( f ) × H NEXT −1− Piece ( f , 49) 2

48

48
A.4 PSD masks

For all data rates, the measured transmit PSD of each STU shall not exceed the PSD masks
specified in this clause (PSDMASKSHDSL(f)), and the measured total power into 135 Ω shall fall
within the range specified in this clause (PSHDSL ± 0.5 dB).
The inband PSD for 0 < f < 1.5 MHz shall be measured with a 10 kHz resolution bandwidth.
NOTE – Large PSD variations over narrow frequency intervals (for example near the junction of the main
lobe with the noise floor) might require a smaller resolution bandwidth (RBW) to be used. A good rule of
thumb is to choose RBW such that there is no more than 1 dB change in the signal PSD across the RBW.
Support for the symmetric PSDs specified in A.4.1 shall be mandatory for all supported data rates.
Support for the asymmetric PSDs specified in A.4.2 and A.4.3 shall be optional.
A.4.1 Symmetric PSD masks
For all values of framed data rate available in the STU, the following set of PSD masks
(PSDMASKSHDSL(f)) shall be selectable:
   πf  
2

 − PBO  sin    MaskedOffsetdB ( f ) 
 
10 10 × K SHDSL × 1 ×   sym   × 1
Nf
× 10 10 , f < f int

 135 f  πf 
2
 
2 × Order 
PSDMASK SHDSL ( f ) =  sym  
 f  

 Nf sym  1 + 
   f 3 dB 
   
0.5683 × 10 −4 × f −1.5 , f ≤ f ≤ 1.1 MHz 
 int 
where MaskOffsetdB(f) is defined as:
 f 3 dB − f
1 + 0.4 × , f < f 3 dB
MaskOffsetdB( f ) =  f 3 dB
1 , f ≥ f 3 dB

fint is the frequency where the two functions governing PSDMASKSHDSL(f) intersect in the range 0 to
fsym. PBO is the power backoff value in dB. KSHDSL, Order, N, fsym, f3dB, and PSHDSL are defined in
Table A.4. PSHDSL is the range of power in the transmit PSD with 0 dB power backoff. R is the
payload data rate.
82 ITU-T Rec. G.991.2 (12/2003)

Table A.4/G.991.2 – Symmetric PSD parameters
Payload data fsym
KSHDSL Order N f3dB PSHDSL (dBm)
rate, R (kbit/s) (ksymbol/s)
R < 1536 7.86 6 1 (R + 8)/3 1.0 × fsym/2 P1(R) ≤ PSHDSL ≤ 13.5
1536 or 1544 8.32 6 1 (R + 8)/3 0.9 × fsym/2 13.5
R > 1544 7.86 6 1 (R + 8)/3 1.0 × fsym/2 13.5
P1(R) is defined as follows:

P1( R) = 0.3486 log 2 ( R × 1000 + 8000) + 6.06 dBm
For 0 dB power backoff, the measured transmit power into 135 Ω shall fall within the range
PSHDSL ± 0.5 dB. For power backoff values other than 0 dB, the measured transmit power into 135 Ω
shall fall within the range PSHDSL ± 0.5 dB minus the power backoff value in dB. The measured
transmit PSD into 135 Ω shall remain below PSDMASKSHDSL(f).
Figure A.5 shows the PSD masks with 0 dB power backoff for payload data rates of 256, 512, 768,
1536, 2048 and 2304 kbit/s.
Figure A.5/G.991.2 – PSD masks for 0 dB power backoff
The equation for the nominal PSD measured at the terminals is:
   πf  
2

 − PBO  sin    
  2
10 10 × K SHDSL × 1 ×   sym   × 1 f
Nf
× , f < f int

 135 f  πf 
2
 
2 × Order
f 2
+ f 2 
NominalPSD ( f ) =  sym  
 f 
c 

 Nf sym  1 + 
   f 3 dB 
   
0.5683 × 10 −4 × f −1.5 , f ≤ f ≤ 1.1 MHz 
 int 
ITU-T Rec. G.991.2 (12/2003) 83

where fc is the transformer cut-off frequency, assumed to be 5 kHz. Figure A.6 shows the nominal
transmit PSDs with 13.5 dBm power for payload data rates of 256, 512, 768, 1536, 2048 and
2304 kbit/s.
NOTE 1 – The nominal PSD is intended to be informative in nature; however, it is used for purposes of
crosstalk calculations (see A.3.3.5 and A.3.3.6) as representative of typical implementations.
Figure A.6/G.991.2 – Nominal PSDs for 0 dB power backoff

NOTE 2 – In this clause, PSDMASK(f) and NominalPSD(f) are in units of W/Hz, and f is in units of Hz.
A.4.2 Asymmetric 1.536 or 1.544 PSD mask
The asymmetric PSD mask set specified in A.4.2.1 and A.4.2.2 shall optionally be supported for
1.536 and 1.544 Mbit/s payload data rates (1.544 and 1.552 Mbit/s framed data rates) in
North America. The PSD masks are described for the 0 dB power backoff case. For other values of
power backoff, the passband PSD masks shall shift, but the out-of-band mask shall remain constant.
Power and power spectral density is measured into a load impedance of 135 Ω.
A.4.2.1 PSD mask for STU-C
For 0 dB power backoff, the output power of the STU-C during data mode shall be
(16.8 ± 0.5) dBm in the frequency band from 0 to 440 kHz and shall be limited by the mask of
Figure A.7. Table A.5 provides the numerical values for the mask of Figure A.7. The PSD mask is
created by linear interpolation of the frequency and power (dBm/Hz) entries of Table A.5.
84 ITU-T Rec. G.991.2 (12/2003)

Figure A.7/G.991.2 – STU-C PSD mask for 1.536 or 1.544 Mbit/s
with 0 dB power backoff
Table A.5/G.991.2 – STU-C PSD mask values for 1.536 or 1.544 Mbit/s
Maximum Maximum Maximum
Frequency Frequency Frequency
power power power
(kHz) (kHz) (kHz)
(dBm/Hz) (dBm/Hz) (dBm/Hz)
≤1 –54.2 – PBO 280 –35.7 – PBO 1000 –89.2
2 –42.2 – PBO 375 –35.7 – PBO 2000 –99.7
12 –39.2 – PBO 400 –40.2 – PBO ≥3000 –108
190 –39.2 – PBO 440 –68.2
236 –46.2 – PBO 600 –76.2
The STU-C PSD mask shall be calculated by subtracting PBO (the Power Backoff value, in dB)
from each PSD value in Table A.5 for frequencies less than or equal to 400 kHz, then by linear
interpolation of the frequency and power (dBm/Hz) over all frequencies. The output power of the
STU-C during data mode shall be (16.8 – PBO ± 0.5) dBm in the frequency band from 0 to
440 kHz. The power level during start-up shall be (16.6 – PBO ± 0.5) dBm. The nominal PSD
(NominalPSD(f)) is defined as the PSD mask with PBO set to 1 dB.
NOTE – The nominal PSD is intended to be informative in nature; however, it is used for purposes of
ITU-T Rec. G.991.2 (12/2003) 85

A.4.2.2 PSD mask for STU-R
For 0 dB power backoff, the output power of the STU-R during data mode shall be
Figure A.8/G.991.2 – STU-R PSD Mask for 1.536 or 1.544 Mbit/s with 0 dB power backoff
Table A.6/G.991.2 – STU-R PSD mask values for 1.536 or 1.544 Mbit/s
power power power
(kHz) (kHz) (kHz)
≤1 –54.2 – PBO 220 –34.4 – PBO 555 −102.6
2 –42.1 – PBO 255 –34.4 – PBO 800 −105.6
10 –37.8 – PBO 276 –41.1 – PBO 1400 −108
175 –37.8 – PBO 300 –77.6 ≥2000 −108
The STU-R PSD mask shall be calculated by subtracting PBO (the Power Backoff value, in dB)
STU-R during data mode shall be (16.5 – PBO ± 0.5) dBm in the frequency band from 0 to
(NominalPSD(f)) is defined as the PSD mask with PBO set to 1 dB.
86 ITU-T Rec. G.991.2 (12/2003)

A.4.3 Asymmetric PSD masks for 768 or 776 kbit/s data rates
The asymmetric PSD mask set specified in A.4.3.1 and A.4.3.2 shall optionally be supported for the
768 kbit/s and 776 kbit/s payload data rates (776 and 784 kbit/s framed data rates) in North
America. The PSD masks are described for the 0 dB power backoff case. For other values of power
backoff, the passband PSD masks shall shift, but the out-of-band mask shall remain constant. Power
and power spectral density is measured into a load impedance of 135 Ω.
A.4.3.1 PSD mask for STU-C
For 0 dB power backoff, the output power of the STU-C during data mode shall be
Figure A.9/G.991.2 – STU-C PSD mask for 768 or 776 kbit/s with 0 dB power backoff
Table A.7/G.991.2 – STU-C PSD mask values for 768 or 776 kbit/s
power power power
(kHz) (kHz) (kHz)
≤50 –36.5 – PBO 135 –45.5 – PBO 250 –50.5 – PBO
80 –39.5 – PBO 145 –39.5 – PBO 400 –45.5 – PBO
90 –44 – PBO 150 –37.5 – PBO 600 –70
105 –57 – PBO 155 –36.5 – PBO 1000 –89.2
110 –57 – PBO 200 –39.25 – PBO 2000 –99.7
210 –42 – PBO ≥3000 –108
ITU-T Rec. G.991.2 (12/2003) 87

The STU-C PSD mask shall be calculated by subtracting PBO (the Power Backoff value, in dB)
STU-C during data mode shall be (14.1 – PBO ± 0.5) dBm in the frequency band from 0 to
(NominalPSD(f)) is defined as the PSD mask with PBO set to 1 dB, multiplied by f 2/(f 2 + fc2) where
f is the frequency in Hz and fc is 5000 Hz, the nominal transformer cut-off frequency.
A.4.3.2 PSD mask for STU-R
For 0 dB power backoff, the output power of the STU-R during data mode shall be
Figure A.10. Table A.8 provides the equations for the mask of Figure A.10.
Figure A.10/G.991.2 – STU-R PSD mask for 768 or 776 kbit/s with 0 dB power backoff
88 ITU-T Rec. G.991.2 (12/2003)

Table A.8/G.991.2 – STU-R PSD mask values for 768 or 776 kbit/s
Frequency, f Maximum power
(Hz) (dBm/Hz)
0 < f ≤ 50 000 –36 – PBO
50 000 < f ≤ 125 000 –36 – PBO – ((f – 50 000)/75 000)
125 000 < f ≤ 130 000 –37 – PBO
130 000 < f ≤ 307 000 –37 – PBO – 142 log10(f/130 000)
307 000 < f ≤ 1 221 000 –90
1 221 000 < f ≤ 1 630 000 –90 peak, with max power in the [f, f + 1 MHz] window of
(–90 – 48 log2(f/1 221 000) + 60) dBm
f > 1 630 000 –90 peak, with max power in the [f, f + 1 MHz] window of –50 dBm
The STU-R PSD mask shall be calculated by subtracting PBO (the Power Backoff value, in dB)
from each PSD value in Table A.8 for frequencies less than or equal to 307 kHz, then by evaluation
of the equations for power (dBm/Hz) over all frequencies. The output power of the STU-R during
data mode shall be (14.1 – PBO ± 0.5) dBm in the frequency band from 0 to 307 kHz. The power
level during start-up shall be (13.9 – PBO ± 0.5) dBm. The nominal PSD (NominalPSD(f)) is
defined as the PSD mask with PBO set to 1 dB, multiplied by f 2/(f 2 + fc2) where f is the frequency
in Hz and fc is 5000 Hz, the nominal transformer cut-off frequency.
A.5 Region-specific functional characteristics

A.5.1 Data rate
The operation of the STU in data mode at the specified information rate shall be as specified in
Table A.9.
Table A.9/G.991.2 – Framed data mode rates

Payload data rate, Symbol rate K
Modulation
R (kbit/s) (ksymbol/s) (Bits per symbol)
R = n × 64 + (i) × 8 16-TCPAM (R + 8) ÷ 3 3
For devices supporting Annex A functionality, no additional limitation on data rates shall be placed
beyond the limitations stated in clause 5 and reiterated in 7.1.1, 8.1 and 8.2.
A.5.2 Return loss
For devices supporting Annex A functionality, return loss shall be specified based on the
methodology of 11.3 and the limitations of Figure 11-6. The following definitions shall be applied
to the quantities shown in Figure 11-6:
RLMIN = 12 dB
f0 = 12.56 kHz
f1 = 50 kHz
ITU-T Rec. G.991.2 (12/2003) 89

f2 = fsym/2
f3 = 1.99 fsym
where fsym is the symbol rate.
A.5.3 Span powering
The capability for an STU-C to provide power over a span to an STU-R is optional. However, if this
capability is provided, the STU-C shall meet the requirements of A.5.3.1. The capability for an
STU-R (or an SRU) to be remotely powered over the span is optional. However, if this capability is
provided, the STU-R or SRU shall meet the requirements of A.5.3.2. Segments that do not support
span powering or that have it disabled may optionally provide wetting (sealing) current, as defined
in A.5.3.3.
The STU-C, STU-R and SRU shall comply with all applicable industry safety standards that are
consistent with their deployment. In particular, it is highly desirable that SHDSL equipment comply
with ITU-T Rec. K.50 [B4].
If an STU-R is deployed as CPE (i.e., it is part of a subscriber's installation), then span powering
shall be disabled at the STU-C. The STU-C may optionally provide wetting current, as specified
in A.5.3.3.
When implemented, SHDSL span powering shall support DC powering of remote terminal units
over single-span loop resistances from 0 to 1800 Ω. The maximum span resistance shall include the
worst-case loop resistance plus the wiring inside the central office and remote site. The STU-C span
power supply shall be designed as a voltage source and shall be considered a voltage-limited circuit
in the application of all referenced standards.
The span powering requirements defined herein are intended for use across a single segment from
an STU-C to either an STU-R or an SRU. Application of these requirements in the STU-C to SRU
case shall result in the termination of span powering voltages at the SRU. Succeeding segments may
optionally support wetting current. Powering across multiple spans is not prohibited; however, the
requirements are for further study. Wetting current may optionally be supported across any segment
(STU-R to STU-C, STU-C to SRU, SRU to STU-R or SRU to SRU-R).
To ensure interoperability and reliable operation, the STU-C and STU-R (or SRU) shall meet the
following requirements when span powering is implemented:
A.5.3.1 STU-C span powering source
A.5.3.1.1 Output voltage
The maximum potential between tip and ring shall be 200 V. The minimum potential between tip
and ring shall be 160 V.
A.5.3.1.2 Power
The minimum steady-state power output capability shall be 15 W.
A.5.3.1.3 Polarity
The negative potential shall be applied to the terminal designated "ring" or "R". The potential from
tip-to-ground should be zero or negative.
A.5.3.1.4 Slew Rate
The supply voltage power-up slew rate at the STU-C loop interface (rise time of VTEST) shall be at
least 1 V/ms but no greater than 30 V/ms when measured in the test circuit of Figure A.11 under all
test conditions defined in Table A.10.
90 ITU-T Rec. G.991.2 (12/2003)

RTEST
VSTU-C VTEST
CTEST
STU-C Device Under Test Load test circuit
Figure A.11/G.991.2 – STU-C power-up slew rate test circuit
Table A.10/G.991.2 – Test conditions for

STU-C slew rate
CTEST (µF) Ω)
RTEST (Ω
1.0 100
1.0 1800
15 100
15 1800
NOTE (informative) – On a 900 Ω loop, the STU-C output voltage specification results in a maximum
remote power load of 7.1 W.
A.5.3.1.5 Power feeding oscillation
The STU-C power supply should be designed to ensure that power feeding oscillation (a condition
that could result in excessive noise coupling into other wire pairs in the cable) does not occur for the
electrical characteristics shown for the protection circuit in Figure A.12.
STU-C STU-R
Tip
RL 25 ٛ
VS
RL 25 ٛ
75 V
breakover
Ring
NOTE (informative) – With appropriate current (to ground) restrictions, these requirements are
not in conflict with the criteria of the Class 2 Voltage limits contained in [B5].
Figure A.12/G.991.2 – Power oscillation example circuit
A.5.3.2 STU-R (and SRU) powering

A.5.3.2.1 Input voltage
The STU-R (or SRU) shall operate properly over the range of input voltages from 80 V to 200 V.
The STU-R (or SRU) may operate with input voltages less than 80 V.
ITU-T Rec. G.991.2 (12/2003) 91

A.5.3.2.2 Polarity
An STU-R (or SRU) shall function normally independent of the polarity of the line power input
voltage. Note that tip/ring reversal is indicated via the EOC by the Maintenance Status Response
message (9.5.5.7.20).
A.5.3.2.3 Capacitance
The capacitance of the STU-R (or SRU) shall be less than or equal to 15 µF.
A.5.3.2.4 Load characteristic
In order to guarantee power system stability during power-up and steady-operation, STU-R (or
SRU) shall present a load characteristic which produces the following observable measurements
when inserted in the test circuit shown in Figure A.13.
While VLINE is ramped up from 0 V to the specified maximum voltage at the specified slew rate, the
values of VLINE and VLOAD shall be observed and recorded. Set t0 as the recorded time point during
the power-up sequence when VLOAD = VLINE/2. The load characteristic of the STU-R (or SRU)
device under test (DUT) shall be such that for all time t > t0, VLOAD > VLINE/2. This criteria shall be
met for all test conditions defined in Table A.11.
RSPAN
STU-R
VLINE Device
VLOAD Under Test
Figure A.13/G.991.2 – Test circuit for STU-R turn-on load characteristic
Table A.11/G.991.2 – Test conditions for STU-R turn-on load characteristic

VLINE slew rate (V/ms) VLINE maximum voltage Ω)
RSPAN (Ω
1.0 200 100
1.0 160 1800
30.0 200 100
30.0 160 1800
The test power supply used to generate VLINE should have a minimum load capacity of 20 W at all
output voltages up to 200 V. The test power supply should use linear voltage regulation to minimize
transient output voltage effects (observed at VLINE) in the presence of test load variations.
A.5.3.3 Wetting current
The STU-R (or SRU-R) shall be capable of drawing between 1.0 and 20 mA of wetting (sealing)
current from the remote feeding circuit when span powering is disabled or is not supported. The
maximum rate of change of the wetting current shall be no more than 20 mA per second.
The STU-C (or SRU-C) may optionally supply power to support wetting current if span powering is
disabled or is not supported. When enabled, this power source should produce a nominal –48 V
potential measured at ring with respect to tip. The maximum voltage of the power source (if
provided) should be limited to –56.5 V. The minimum voltage should be high enough to ensure a
92 ITU-T Rec. G.991.2 (12/2003)

voltage of at least –39 V at the inputs of the STU-R (or SRU-R), measured at ring with respect to
tip, to guarantee that the STU-R (or SRU-R) metallic termination will turn on and allow wetting
current to flow. In no case shall the wetting current source apply a potential greater than –72 V
between ring and tip. The potential at tip with respect to ground should be zero or negative.
A.5.3.4 Metallic termination
A metallic termination at the STU-R shall be provided in conjunction with the use of wetting
current (A.5.3.3). The SRU-R shall meet the same requirements specified in this clause for an
STU-R.
Table A.12 and Figure A.14 give characteristics that apply to the DC metallic termination of the
STU-R. The metallic termination provides a direct current path from tip to ring at the STU-R,
providing a path for sealing current. By exercising the non-linear functions of the metallic
termination, a network-side test system may identify the presence of a conforming STU-R on the
customer side of the interface. The characteristics of the metallic termination shall not be affected
by whether the STU-R is powered in any state, or unpowered.
There are two operational states of the DC metallic termination:
a) the ON or conductive state; and
b) the OFF or non-conductive state.
A.5.3.4.1 ON state
The application of a voltage across the metallic termination greater than VAN, the
activate/non-activate voltage, for a duration greater than the activate time shall cause the
termination to transition to the ON state. The activate/non-activate voltage shall be in the range of
30.0 to 39.0 V. The activate time shall be in the range of 3.0 to 50.0 ms. If a change of state is to
occur, the transition shall be completed within 50 ms from the point where the applied voltage
across the termination first exceeds VAN. Application of a voltage greater than VAN for a duration
less than 3.0 ms shall not cause the termination to transition to the ON state. See Table A.12 and
Figure A.14.
While in the ON state, when the voltage across the termination is 15 V, the current shall be greater
than or equal to 20 mA. The metallic termination shall remain in the ON state as long as the current
is greater than the threshold IHR (see Table A.12 and Figure A.14) whose value shall be in the range
of 0.1 to 1.0 mA. Application of 90.0 V through 200 to 4000 Ω (for a maximum duration of 2 s)
shall result in a current greater than 9.0 mA.
A.5.3.4.2 OFF state
The metallic termination shall transition to the OFF state if the current falls below the threshold IHR
whose value shall be in the range of 0.1 to 1.0 mA for a duration greater than the "guaranteed
release" time (100 ms) (see Table A.12 and Figure A.14). If a change of state is to occur, the
transition shall be completed within 100 ms from the point where the current first falls below IHR. If
the current falls below IHR for a duration less than 3.0 ms, the termination shall not transition to the
OFF state. While in the OFF state, the current shall be less than 5.0 µA whenever the voltage is less
than 20.0 V. The current shall not exceed 1.0 mA while the voltage across the termination remains
less than the activate voltage.
Descriptive material can be found in Table A.12 and Figure A.14.
A.5.3.4.3 STU-R capacitance
While the metallic termination is OFF, the tip-to-ring capacitance of the STU-R when measured at a
frequency of less than 100 Hz shall be 1.0 µF ± 10%.
ITU-T Rec. G.991.2 (12/2003) 93

A.5.3.4.4 Behaviour of the STU-R during metallic testing
During metallic testing, the STU-R shall behave as follows:
a) when a test voltage of up to 90 V2 is applied across the loop under test, the STU-R shall
present its DC metallic termination as defined in A.5.3.4, Table A.12 and Figure A.14, and
not trigger any protective device that will mask this signature. The series resistance (test
system + test trunk + loop + margin) can be from 200 to 4000 Ω (balanced between the two
conductors);
b) the STU-R may optionally limit current in excess of 25 mA (20 mA maximum sealing
current + 5 mA implementation margin).
Table A.12/G.991.2 – Characteristics of DC metallic termination at the STU-R

Type of operation Normally OFF DC termination. Turned ON by
application of metallic voltage. Held ON by loop
current flow. Turned OFF by cessation of loop
current flow.
Current in the ON state and at 15 V ≥ 20 mA
DC voltage drop (when ON) at 20 mA current ≤ 15 V
DC current with application of 90 V through min 9 mA (Note). See Figure A.14.
4000 Ω for up to 2 s.
DC leakage current (when OFF) at 20 V ≤ 5.0 µA
Activate/non-activate voltage 30.0 V DC ≤ VAN ≤ 39.0 V DC
Activate (breakover) current at VAN ≤ 1.0 mA
Activate time for voltage ≥ VAN 3 ms to 50 ms
Hold/release current 0.1 mA ≤ IHR ≤ 1.0 mA
Release/non-release time for current ≤ IHR 3 ms to 100 ms
NOTE – This requirement is intended to ensure a termination consistent with test system operation.
____________________
2 One test system in common use today applies 70 V DC plus 10 Vrms AC (84.4 V peak) to one conductor
of the loop while grounding the other conductor.
94 ITU-T Rec. G.991.2 (12/2003)

Figure A.14/G.991.2 – Illustration of DC characteristics of the STU-R
(Bilateral switch and holding current)
A.5.4 Longitudinal Balance

For devices supporting Annex A functionality, longitudinal balance shall be specified based on the
to the quantities in Figure 11-2.
LBMIN = 40 dB
f1 = 20 kHz
f2 = fsym/2
ITU-T Rec. G.991.2 (12/2003) 95

A.5.5 Longitudinal output voltage
For devices supporting Annex A functionality, longitudinal output voltage shall be specified based
on the methodology of 11.2. The measurement frequency range shall be between 20 kHz and
450 kHz.
A.5.6 PMMS target margin
If the optional line probe is selected during the G.994.1 session, the receiver shall use the negotiated
target margin. If worst-case PMMS target margin is selected, then the receiver shall assume the
disturbers of Table A.13 to determine if a particular rate can be supported. Reference crosstalk shall
be computed as defined in A.3.3 with the FEXT components in A.3.3.9 ignored. The reference
crosstalk specified in this clause may not be representative of worst-case conditions in all networks.
Differences between crosstalk environments may be compensated by adjusting the target margin.
Table A.13/G.991.2 – Reference disturbers used during PMMS

for worst-case target margin
Rate (kbit/s) PSD (direction) Reference disturber
All Symmetric (US/DS) 49 SHDSL
768/776 Asymmetric (US) 49 HDSL
768/776 Asymmetric (DS) 24 T1 + 24 HDSL
1536/1544 Asymmetric (US) 39 SHDSL (NEXT only)
1536/1544 Asymmetric (DS) 24 T1 + 24 SHDSL (NEXT only)
A.5.7 Span powering in M-pair mode

In the optional M-pair mode, the requirements for remote power feeding or wetting current for each
of the M pairs shall be identical to the requirements for a single pair specified in A.5.3.
NOTE – This implies that the powering/wetting current is provided by a potential difference between tip and
ring on each of the M pairs.
Annex B

B.1 Scope
This annex describes those specifications that are unique to SHDSL systems operating under
conditions such as those typically encountered within European networks. The clauses in this annex
provide the additions and modifications to the corresponding clauses in the main body.
B.2 Test loops

B.2.1 Functional description
The test loops in Figure B.1 are based on the existing HDSL test loops. The length of the individual
loops are chosen such that the transmission characteristics of all loops are comparable. The purpose
is to stress the equalizer of the SHDSL unit under test similarly over all loops when testing SHDSL
at a specific bit rate. The total length of each loop is described in terms of physical length, and the
length of the individual sections as a fixed fraction of this total. If implementation tolerances of one
test loop causes its resulting electrical length to be out of specification, then its total physical length
shall be scaled accordingly to correct this error. One test loop includes bridged taps to achieve rapid
96 ITU-T Rec. G.991.2 (12/2003)

variations in amplitude and phase characteristics of the cable transfer function. In some access
networks, these bridge taps have been implemented in the past, which stresses the SHDSL modem
under test differently.
Loop #1 is a symbolic name for a loop with zero (or near zero) length, to prove that the SHDSL
transceiver under test can handle the potentially high signal levels when two transceivers are
directly interconnected.
B.2.2 Test loop topology
The topology of the test loops is specified in Figure B.1. The basic test cable characteristics, the
transfer function of the test loops specified using these cables and the variation of input impedance
of the test loops are shown in Appendix II.
Figure B.1/G.991.2 – Test loop topology
ITU-T Rec. G.991.2 (12/2003) 97

B.2.3 Test loop length
The length of each test loop for SHDSL transmission systems is specified in Table B.1. The
specified insertion loss Y at the specified test frequency measured with a 135 Ω termination
(electrical length) is mandatory. If implementation tolerances of one test loop causes that its
resulting electrical length is out of specification, then its total physical length shall be scaled
accordingly to adjust this error.
The test frequency fT is chosen to be a typical mid-band frequency in the spectrum of long range
SHDSL systems. The length is chosen to be a typical maximum value that can be handled correctly
by the SHDSL transceiver under test. This value is bit rate dependent; the higher the payload bit
rate, the lower is the insertion loss that can be handled in practice.
Table B.1/G.991.2 – Values of the electrical length Y of the SHDSL noise test loops,
when testing SHDSL at noise model A
Y Y
Payload
fT [dB] L1 L2 L3 L4 L5 L7 fT [dB] L6
bit rate
[kHz] @fT, [m] [m] [m] [m] [m] [m] [kHz] @fT, [m]
[kbit/s]
@135 Ω @135 Ω
384 150 43.0 <3 4106 5563 5568 11 064 4698 115 40.5 3165
512 150 37.0 <3 3535 4787 4789 9387 3996 115 35.0 2646
768 150 29.0 <3 2773 3747 3753 7153 3062 275 34.5 1904
1024 150 25.5 <3 2439 3285 3291 6174 2668 275 30.0 1547
1280 150 22.0 <3 2105 2829 2837 5193 2266 275 26.0 1284
1536 150 19.0 <3 1820 2453 2455 4357 1900 250 21.5 1052
2048 (s) 200 17.5 <3 1558 2046 2052 3285 1550 250 18.5 748
2304 (s) 200 15.5 <3 1381 1815 1820 2789 1331 250 16.5 583
2048 (a) 250 21.0 <3 1743 2264 2272 3618 1726 250 21.0 1001
2304 (a) 250 18.0 <3 1494 1927 1937 2915 1402 250 18.0 702
NOTE – The electrical length Y (insertion loss at specified frequency fT) is mandatory, the (estimated)
physical lengths L1-L7 are informative.
(s) those electrical lengths apply to the symmetric PSD.
(a) those electrical lengths apply to the asymmetric PSD.
when testing SHDSL at noise model B, C, or D
Y Y
Payload
bit rate
[kbit/s]
@135 Ω @135 Ω
384 150 50.0 <3 4773 6471 6477 13 021 5508 115 47.5 3859
512 150 44.0 <3 4202 5692 5698 11 344 4814 115 41.5 3261
768 150 35.5 <3 3392 4592 4596 8970 3815 275 42.0 2536
1024 150 32.0 <3 3058 4135 4141 7990 3403 275 38.0 2223
1280 150 28.5 <3 2725 3678 3684 7011 3006 275 33.5 1816
1536 150 25.5 <3 2439 3285 3291 6174 2673 250 29.0 1680
2048 (s) 200 24.0 <3 2135 2812 2820 4886 2271 250 25.5 1426
2304 (s) 200 21.5 <3 1913 2509 2518 4257 2010 250 23.0 1208
98 ITU-T Rec. G.991.2 (12/2003)

when testing SHDSL at noise model B, C, or D
Y Y
Payload
bit rate
[kbit/s]
@135 Ω @135 Ω
2048 (a) 250 28.0 <3 2323 3030 3034 5189 2389 250 28.0 1607
2304 (a) 250 25.0 <3 2075 2699 2705 4514 2102 250 25.0 1387
NOTE – The electrical length Y (insertion loss at specified frequency fT) is mandatory, the (estimated) physical
lengths L1-L7 are informative.
(s) those electrical lengths apply to the symmetric PSD.
(a) those electrical lengths apply to the asymmetric PSD.
B.3 Performance testing

The purpose of transmission performance tests is to stress SHDSL transceivers in a way that is
representative to a high penetration of systems scenario in operational access networks. This high
penetration approach enables operators to define deployment rules that apply to most operational
situations. It means also that in individual operational cases, characterized by lower noise levels
and/or insertion loss values, the SHDSL system under test may perform better than tested.
The design impedance RV is 135 Ω. All spectra are representing single sided power spectral
densities (PSD).
B.3.1 Test procedure
The purpose of this clause is to provide an unambiguous specification of the test set-up, the
insertion path and the way signal and noise levels are defined. The tests are focused on the noise
margin, with respect to the crosstalk noise or impulse noise levels when SHDSL signals under test
are attenuated by standard test-loops and interfered with standard crosstalk noise or impulse noise.
This noise margin indicates what increase of crosstalk noise or impulse noise level is allowed under
specific operational conditions to ensure sufficient transmission quality.
B.3.2 Test set-up definition
Figure B.2 illustrates the functional description of the test set-up. It includes:
• A bit error ratio test set (BERTS) applies a 215 – 1 pseudo-random bit sequence (PRBS) test
signal to the transmitter in the direction under test at the bit rate required. The transmitter in
the opposing direction shall be fed with a similar PRBS signal, although the reconstructed
signal in this path need not be monitored.
• The test loops, as specified in B.2.
• An adding element to add the (common mode and differential mode) impairment noise (a
mix of random, impulsive and harmonic noise), as specified in B.3.5.
• An impairment generator, as specified in B.3.5, to generate both the differential mode and
common mode impairment noise, that are fed to the adding element.
• A high impedance, and well-balanced differential voltage probe (e.g., better than 60 dB
across the whole band of the SHDSL system under test) connected with level detectors such
as a spectrum analyser or a true RMS voltmeter.
• A high impedance, and well-balanced common mode voltage probe (e.g., better than 60 dB
across the whole band of the SHDSL system under test) connected with level detectors such
as a spectrum analyser or a true RMS voltmeter.
ITU-T Rec. G.991.2 (12/2003) 99

Application Application
interface interface
PRBS
BERTS
pattern
receiver
generator
Tx Rx
Test loop
Adding
Test "Cable"
element
Differential Voltage Voltage

voltage probe probe
probe U1 U2
Level Level Level

detector detector detector
GND GND
NOTE – To allow test reproducibility, the testing equipment and the Termination Units (STU-C and STU-R) should
refer to an artificial earth. If the Termination Units have no earth terminal, the test should be performed while the
Termination Units are placed on a metal plate (of sufficient large size) connected to earth.
Figure B.2/G.991.2 – Functional description of the set-up of the performance tests
The two-port characteristics (transfer function, impedance) of the test-loop, as specified in B.2, are
defined between port TX (node pairs A1, B1) and port RX (node pair A2, B2). The consequence is
that the two-port characteristics of the test "cable" in Figure B.2 must be properly adjusted to take
full account of non-zero insertion loss and non-infinite shunt impedance of the adding element and
impairment generator. This is to ensure that the insertion of the generated impairment signals does
not appreciably load the line.
The balance about earth, observed at port TX, at port RX, and at the tips of the voltage probe shall
exhibit a value that is 10 dB greater than the transceiver under test. This is to ensure that the
impairment generator and monitor function do not appreciably deteriorate the balance about earth of
the transceiver under test.
The signal flow through the test set-up is from port TX to port RX, which means that measuring
upstream and downstream performance requires an interchange of transceiver position and test
"cable" ends.
The received signal level at port RX is the level, measured between node A2 and B2, when port TX
as well as port RX are terminated with the SHDSL transceivers under test. The impairment
generator is switched off during this measurement.
Test Loop #1, as specified in B.2, shall always be used for calibrating and verifying the correct
settings of generators G1-G7, as specified in B.3.5, when performing performance tests.
100 ITU-T Rec. G.991.2 (12/2003)

The transmitted signal level at port TX is the level, measured between node A1 and B1, under the
same conditions.
The impairment noise shall be a mix of random, impulsive and harmonic noise, as defined in B.3.5.
The level that is specified in B.3.5 is the level at port RX, measured between node A2 and B2 (and
includes both differential mode and common mode impairments), while port TX as well as port RX
are terminated with the design impedance RV. These impedances shall be passive when the
transceiver impedance in the switched-off mode is different from this value.
NOTE – The injected noise is intended to match the theoretical noise PSD when the transceiver under test is
connected to the loop. On loop #2 and #3 for payload rates of 1024 kbit/s and below, it has been found that
impedance mismatch could generate an increased noise PSD at low frequencies. One method of
compensation is to modify the factor, ∆, defined in A.3.1.4, by replacing the theoretical noise, N(f), in step 3
of A.3.1.4 with the noise PSD measured when connected to the loop under test. A second method is to place
a passive circuit, consisting of a resistor R in parallel with a capacitor C, in series with each wire of the noise
generator output pair. The RC values of R = 1.2 Kohms and C = 1 µF are suggested and should be adjusted
for each noise generator such that the injected noise matches the theoretical noise PSD. A third method is to
calibrate the noise generator waveform into the loop under test such that when connected to the loop under
test, the theoretical noise waveform is present at the transceiver terminals.
B.3.3 Signal and noise level definitions
The signal and noise levels are probed with a well-balanced differential voltage probe, and the
differential impedance between the tips of the probe shall be higher than the shunt impedance of
100 kΩ in parallel with 10 pF. Figure B.2 shows the probe position when measuring the RX signal
level at the STU-C or STU-R receiver. Measuring the TX signal level requires the connection of the
tips to node pair [A1, B1].
The various PSDs of signals and noises specified in this Recommendation are defined at the TX or
RX side of the set-up. The levels are defined when the set-up is terminated, as described above,
with design impedance RV or with transceivers under test.
Probing an rms-voltage Urms [V] in this set-up, over the full signal band, means a power level of
P[dBm] that equals:
U 2 
P = 10 × log10  rms ×1000  [dBm]
 R 
 V 
Probing an rms-voltage Urms [V] in this set-up, within a small frequency band of ∆f (in Hertz),
corresponds to an average spectral density level of P[dBm/Hz] within that filtered band that equals:
U 2 1000 
P = 10 × log10  rms × [dBm/Hz]
 R ∆f 
 V 
The bandwidth ∆f identifies the noise bandwidth of the filter, and not the –3 dB bandwidth.
B.3.3.1 Noise injection network
B.3.3.1.1 Differential mode injection
The noise injector for differential mode noise (which is the portion of the Adding Element shown in
Figure B.2 that is used to couple differential impairments to the test cable) is a two-port network in
nature, and may have additional ports connected to the impairment generator. The Norton
equivalent circuit diagram is shown in Figure B.2a. The current source Ix is controlled by the
impairment generator. The parasitic shunt impedance Zinj shall have a value of |Zinj| > 4 kΩ in the
frequency range from 100 Hz to 2 MHz.
ITU-T Rec. G.991.2 (12/2003) 101

LX-port IX RX-port
Figure B.2a/G.991.2 – Norton equivalent circuit diagram for

differential mode noise injection
B.3.3.1.2 Common mode injection

The specification of this injection network is for further study.
B.3.3.2 Noise levels calibration
B.3.3.2.1 Differential mode noise calibration
The differential mode noise injection is calibrated using the configuration shown in Figure B.2b.
During calibration the RX side of the noise injector is terminated by the design impedance RV
(= 135 Ω) and the LX (Test Loop interface) side of the noise injector is terminated by an impedance
ZLX. The noise levels given in B.3.5 specify the PSD dissipated in RV on the RX side when ZLX on
the LX side is equal to the calibration impedance Zcal. The impedance Zcal is defined in Figure B.2c.
Differential mode
ZLX LX-port noise injection RX-port RV
network
Impairment
generator
Figure B.2b/G.991.2 – Configuration for noise level calibration
Figure B.2c/G.991.2 – Calibration impedance Zcal
The impedance ZLX on the LX side of the noise injection circuit is equal to the calibration
impedance Zcal as given in Figure B.2c. The PSD dissipated in the impedance RV shall be equal to
the differential noise PSD Pxn(f) defined in B.3.5.1.
NOTE – This is theoretically equivalent to the following: For an arbitrary value of the impedance ZLX, the
PSD dissipated in RV is equal to:
Pcal ( f ) = G ( f , Z LX ) Pxn ( f )
where G(f, ZLX) is the impedance dependent correction factor, which is specified as:
2
1 1 1
++
Z LX Z inj Rv
G ( f , Z LX ) =
1 1 1
+ +
Z cal Z inj Rv
102 ITU-T Rec. G.991.2 (12/2003)

where Zcal is the calibration impedance given in Figure B.2c, Zinj is the Norton equivalent impedance of the
noise injection circuit (see Figure B.2a), and RV = 135 Ω is the SHDSL design impedance.
The noise generator gain settings determined during calibration shall be used during performance
testing. During performance testing the noise injection circuit will be configured as shown in
Figure B.2. Because the loop impedance and the impedance of the modem under test may differ
from the impedance's ZLX and RV used during calibration, the voltage over the RX port of the
modem may differ from the voltage UX observed during calibration.
B.3.3.2.2 Common mode noise calibration
This calibration method is for further study.
B.3.4 Performance test procedure
The test performance of the SHDSL transceiver shall be such that the bit error ratio (BER) on the
disturbed system is less than 10−7, while transmitting a pseudo-random bit sequence. The BER
should be measured after at least 109 bits have been transmitted.
The tests are carried out with a margin which indicates what increase of noise is allowed to ensure
sufficient transmission quality. Network operators may calculate their own margins for planning
purposes based on a knowledge of the relationship between this standard test set and their network
characteristics.
A test sequence as specified in Table B.3 shall be concluded. The test loops referred to are specified
in Figure B.1. The test loops are characterized by the insertion loss Y and/or the cable length L,
which depend on the data rate to be transported and have to be scaled adequately.
In Table B.3, upstream and downstream only determine the topology of the test loop. The STU-C
must pass all tests 1 through 12. The STU-R must pass all tests 1 through 12.
A test is defined as the measurement of a given BER associated with a single test path, direction,
test noise, rate and margin. The ensemble of tests associated with a particular value of N in
Table B.3 is defined as a test set.
Table B.3/G.991.2 – Test sequence for performance testing

Direction
N Test path Comments
(Note 6)
1 #1 (Note 1) Upstream Y = 0 dB; Test noise A (Notes 5, 7)
2 #2 Upstream Y = Y1 (Note 2); Test noise A, C and D (Notes 7, 8)
3 #3 Upstream Y = Y1; Test noise D (Notes 5, 7, 8)
4 #4 Downstream Y = Y1; Test noise A and C (Notes 5, 7, 8)
5 #5 Upstream Y = Y1; Test noise B (Notes 5, 7, 8)
6 #6 Downstream Y = Y1; Test noise A and C (Notes 5, 7, 8)
7 #7 Downstream Y = Y1; Test noise A, B, C and D (Notes 5, 7, 8)
8 Common mode rejection test (Note 4)
9 (Note 3) (Note 3) Y = Y2; Test noise is the noise corresponding to the test
with the highest BER in test sets 1 through 7 (Note 7)
10 (Note 3) (Note 3) Y = Y3; No added impairment; Worst path of tests 1 to 7;
BER < 10−8
11 #2 Upstream Y = Y1; Impulse test as described in B.3.5.3.7
12 As <TBD> <TBD> Micro-interruption test as described in 12.1
ITU-T Rec. G.991.2 (12/2003) 103

Table B.3/G.991.2 – Test sequence for performance testing
NOTE 1 – Test Path = #1 means that the path under test shall be connected with test loop #1
as defined in Figure B.1.
NOTE 2 – Y1 = Y dB (as specified in Table B.2 for noise models B, C and D and in Table
B.1 for noise model A), Y2 = Y1 – 10 dB, Y3 = Y1 + 3 dB.
NOTE 3 – The tests (for any data rate) are carried out on the loop that gives the highest BER
(for that data rate) in test sets 1 through 7, when the test noise is increased by 6 dB. If no
errors in 109 bits are recorded for all the tests in test sets 1 through 7, then loop #3 upstream
is used for this test set by default.
NOTE 4 – The measuring arrangement for this test is specified in ITU-T Rec. O.9 [B8].
NOTE 5 – Only tested for lowest and highest data rate in Table B.1 or Table B.2 (that the
equipment supports) and for asymmetric PSDs when supported.
NOTE 6 – Upstream means that the unit under test is connected to the STU-C end of the test
loop and downstream means that the unit under test is connected to the STU-R end of the test
loop. For example, test set 5 for an STU-C would connect the STU-C under test to the STU-C
end of the loop as shown in Figure B.1 and apply noise model X.C.B to the STU-C end of the
loop. The same test for an STU-R would connect the STU-R under test to the STU-C end of
the loop as shown in Figure B.1 and apply noise model X.R.B to the STU-C end of the loop.
NOTE 7 – The BER shall be less than 10–7 when the test noise is increased by 6 dB (this is
equivalent to 6 dB of margin).
NOTE 8 – In order to reduce the number of noise shapes used, a mandatory noise shape
substitution rule is given in B.3.5.5.
NOTE 9 – To test the M-pair mode, while one path is under test, the other path(s) must be
connected to loop(s). The characteristics of the other loop(s) must not be worse than the ones
of the path under test. Furthermore, the differential delay between the path under test and the
path(s) connected to the second loop should not exceed the value of the differential delay
buffer specified in 7.1.6.
B.3.5 Impairment generator

The noise that the impairment generator injects into the test set-up is frequency dependent, is
dependent on the length of the test loop and is also different for downstream performance tests and
upstream performance tests. Figure B.3 illustrates this for the alien noise (other than the SHDSL
modem under test), as described in B.3.5.4.1, for the case that the length of test loop #1 is fixed at
3 km, using the crosstalk models described in B.3.5.2. Figure B.4 illustrates this for various loop
lengths for the case that the alien noise of model "B" is applied. These figures are restricted to alien
noise only. The self noise (of SHDSL) shall be combined with this alien noise.
104 ITU-T Rec. G.991.2 (12/2003)

A A
NOTE – This is the noise resulting from three of the four noise models for SHDSL in the case that the length of test loop #2
is fixed at 3 km.
Figure B.3/G.991.2 – Examples of alien noise spectra that are to be injected

into the test set-up, while testing SHDSL systems
Figure B.4/G.991.2 – Examples of alien noise spectra that are to be injected

into the test set-up, while testing SHDSL systems
The definition of the impairment noise for SHDSL performance tests is very complex and for the
purposes of this Recommendation it has been broken down into smaller, more easily specified
components. These separate, and uncorrelated, impairment "generators" may therefore be isolated
and summed to form the impairment generator for the SHDSL system under test. The detailed
specifications for the components of the noise model(s) are given in this clause, together with a
brief explanation.
ITU-T Rec. G.991.2 (12/2003) 105

B.3.5.1 Functional description
Figure B.5 defines a functional diagram of the composite impairment noise. It defines a functional
description of the combined impairment noise as it must be probed at the receiver input of the
SHDSL transceiver under test. The probing is described in B.3.3.
The functional diagram has the following elements:
• The seven impairment "generators" G1 to G7 generate noise as defined in B.3.5.3.1
to B.3.5.3.7. Their noise characteristics are independent from the test loops and bit rates.
• The transfer function H1(f, L) models the length and frequency dependency of the NEXT
impairment, as specified in B.3.5.3.1. The transfer function is independent of the test loops,
but changes with the electrical length of the test loop. Its transfer function changes with the
frequency f, roughly according to f 0.75.
• The transfer function H2(f, L) models the length and frequency dependency of the FEXT
impairment, as specified in B.3.5.3.2. Its transfer function is independent of the test loops,
but changes with the electrical length of the test loop. Its transfer function changes with the
frequency f, roughly according to f times the cable transfer function.
• Switches S1-S7 determine whether or not a specific impairment generator contributes to the
total impairment during a test.
• Amplifier A1 models the property to increase the level of some generators simultaneously
to perform the noise margin tests. A value of x dB means a frequency independent increase
of the level by x dB over the full band of the SHDSL system under test, from fL to fH.
Unless otherwise specified, its gain is fixed at 0 dB.
In a practical implementation of the test set-up, there is no need to give access to any of the internal
signals of the diagram in Figure B.5. These functional blocks may be incorporated with the test loop
and the adding element as one integrated construction.
106 ITU-T Rec. G.991.2 (12/2003)

G1 H1(f, L ) S1
NEXT noise
G2 H2(f, L) S2
FEXT noise A1
Background noise G3 S3
Cable independent
Probe
G4 S4 level
Differential
Broadcast RF noise G5 S5d mode
Cable independent
Fixed powers, fixed freq
Amateur RF noise G6 S6d

Cable independent
Fixed power, variable freq
Impulsive noise G7 S7
Cable independent
Bursty in nature
S5c
Probe
level
S6c Common
mode
Figure B.5/G.991.2 – Functional diagram of the composition of the impairment noise
This functional diagram will be used for impairment tests in downstream and upstream direction.
Several scenarios have been identified to be applied to SHDSL testing. These scenarios are intended
to be representative of the impairments found in metallic access networks.
Each scenario (or noise model) results in a length-dependent and test loop-dependent PSD
description of noise. Each noise model is subdivided into two parts: one to be injected at the STU-C
side, and another to be injected at the STU-R side of the SHDSL modem link under test. Therefore,
seven individual impairment generators G1 to G7 can represent different values for each noise
model they are used in. Specifically, G1 and G2 are dependent on which unit, STU-R or STU-C, is
under test.
Generators G1-G4 represent crosstalk noise. The spectral power Pxn(f) for crosstalk noise is
characterized by the sum:
Pxn(f) = |A1|2 × { |H1(f, L)|2 × PG1(f) + |H2(f, L)|2 × PG2(f) + PG3(f) } + PG4(f)
Each component of this sum is specified in the following clauses. Only the noise generators that are
active during testing should be included during calibration. This combined impairment noise is
applied to the receiver under test, at either the STU-C (for upstream) or STU-R (for downstream)
ends of the test-loop.
Generators G5 and G6 represent ingress noise.
ITU-T Rec. G.991.2 (12/2003) 107

B.3.5.2 Cable crosstalk models
The purpose of the cable crosstalk models is to model both the length and frequency dependency of
crosstalk measured in real cables. These crosstalk transfer functions adjust the level of the noise
generators in Figure B.5 when the electrical length of the test loops is changed. The frequency and
length dependency of these functions is in accordance with observations from real cables. The
specification is based on the following constants, parameters and functions:
• Variable f identifies the frequency in Hz.
• Constant f0 identifies a chosen reference frequency, which was set to 1 MHz.
• Variable L identifies the physical length of the actual test loop in metres. This physical
length is calculated from the cable models in Appendix II from the specified electrical
length. Values are summarized in Tables B.1 and B.2 for each combination of payload bit
rate, noise model, and test loop.
• Constant L0 identifies a chosen reference length, which was set to 1 km.
• Function sT0(f, L) represents the frequency and length dependent amplitude of the insertion
loss of the actual test loop terminated into 135 Ω.
• Constant Kxn identifies an empirically obtained number that scales the NEXT transfer
function H1(f, L). The resulting transfer function represents a power summed crosstalk
model of the NEXT as it was observed in a test cable. Although several disturbers and wire
pairs were used, this function H1(f, L) is scaled down as if it originates from a single
disturber in a single wire pair.
• Constant Kxf identifies an empirically obtained number that scales the FEXT transfer
function H2(f, L). The resulting transfer function represents a power summed crosstalk
model of the FEXT as it was observed in a test cable. Although several disturbers and wire
pairs were used, this function H2(f, L) is scaled down as if it originates from a single
disturber in a single wire pair.
The transfer functions in Table B.4 shall be used as crosstalk transfer functions in the impairment
generator.
Table B.4/G.991.2 – Definition of the crosstalk transfer functions

0.75
 f  4
H1 ( f , L) = K xn ×   × 1 − sT 0 ( f , L)
 f0 
 f  L
H 2 ( f , L) = K xf ×   × × ST 0 ( f , L)
 f0  L0
(–50/20)
Kxn = 10 ≈ 0.0032, f0 = 1 MHz
(–45/20)
Kxf = 10 ≈ 0.0056, L0 = 1 km
ST 0 ( f , L) = test loop insertion loss
B.3.5.3 Individual impairment generators

B.3.5.3.1 Equivalent NEXT disturbance generator [G1.xx]
The NEXT noise generator represents the equivalent disturbance of all impairment that is identified
as crosstalk noise from a predominantly near end origin. This noise, filtered by the NEXT crosstalk
coupling function of B.3.5.2, will represent the contribution of all NEXT to the composite
impairment noise of the test.
108 ITU-T Rec. G.991.2 (12/2003)

The PSD of this noise generator is one of the PSD profiles, as specified in B.3.5.4. For testing
upstream and downstream performance, different PSD profiles shall be used, as specified below.
G1.C.# = X.C.#
G1.R.# = X.R.#
The symbols in this expression, refer to the following:
• Symbol "#" is a placeholder for noise model "A", "B" , "C" or "D".
• Symbols "X.C.#" and "X.R.#" refer to the crosstalk profiles, as defined in B.3.5.4.
This PSD is not related to the cable because the cable portion is modelled separately as transfer
function H1(f, L), as specified in B.2.2.
The noise of this noise generator shall be uncorrelated with all the other noise sources in the
impairment generator, and uncorrelated with the SHDSL system under test. The noise shall be
random in nature and near Gaussian distributed, as specified in B.3.5.4.2.
B.3.5.3.2 Equivalent FEXT disturbance generator [G2.xx]
The FEXT noise generator represents the equivalent disturbance of all impairment that is identified
as crosstalk noise from a predominantly far end origin. This noise, filtered by the FEXT crosstalk
coupling function of B.3.5.2, will represent the contribution of all FEXT to the composite
impairment noise of the test.
The PSD of this noise generator is one of the PSD profiles, as specified in B.3.5.4.1. For testing
upstream and downstream performance, different PSD profiles shall be used, as specified below.
G2.C.# = X.R.#
G2.R.# = X.C.#
The symbols in this expression, refer to the following:
• Symbol "#" is a placeholder for noise model "A", "B", "C" or "D".
• Symbols "X.C.#" and "X.R.#" refer to the crosstalk profiles, as defined in B.3.5.4.
This PSD is not related to the cable because the cable portion is modelled separately as transfer
function H2(f, L), as specified in B.2.2.
The noise of this noise generator shall be uncorrelated with all the other noise sources in the
impairment generator, and uncorrelated with the SHDSL system under test. The noise shall be
random in nature and near Gaussian distributed, as specified in B.3.5.4.2.
B.3.5.3.3 Background noise generator [G3]
The background noise generator is inactive and set to zero.
B.3.5.3.4 White noise generator [G4]
The white noise generator has a fixed, frequency independent value, and is set to a level between
–140 and –120 dBm/Hz, into 135 Ω. The output signal of this noise generator shall be uncorrelated
with all the other noise sources in the impairment generator, and uncorrelated with the SHDSL
system under test. The noise shall be random in nature and near Gaussian distributed, as specified
in B.3.5.4.2.
B.3.5.3.5 Broadcast RF noise generator [G5]
NOTE 1 – Work on a specification dealing with generic RFI testing methods is ongoing. It is expected that a
future version of this Recommendation will contain a complete RFI testing specification, which will be
mandatory. This clause is currently for information only.
ITU-T Rec. G.991.2 (12/2003) 109

The broadcast RF noise generator represents the discrete-tone line interference caused by amplitude
modulated broadcast transmissions in the SW, MW and LW bands, which ingress into the cable.
These interference sources have more temporal stability than the amateur (ham) interference
(see B.3.5.3.6) because their carriers are not suppressed. Ingress causes differential mode as well as
common mode interference.
The ingress noise signal for differential mode impairment (or common mode impairment) is a
superposition of random modulated carriers (AM). The total voltage U(t) of this signal is defined as:
U(t) = Σk Uk × cos(2π·fk × t + ϕk) × (1 + m × αk(t))
The individual components of this ingress noise signal U(t) are defined as follows:
– Uk – The voltage Uk of each individual carrier should be as specified in Table B.5 as
power level P (dBm) into a resistive load R, equal to the design impedance RV = 135 Ω.
Note that spectrum analysers will detect levels that are slightly higher than the values
specified in Table B.5 when their resolution bandwidths are set to 10 kHz or more, since
they will detect the modulation power as well.
– fk – The frequency fk of each individual carrier should be as specified in Table B.5. The
frequency values in Table B.5 do not represent actual broadcast frequencies but are chosen
such that they cover the frequency range that is relevant for SHDSL modems. Note that the
harmonic relation between the carriers in Table B.5 is minimal.
– ϕk – The phase offset ϕk of each individual carrier shall have a random value that is
uncorrelated with the phase offset of every other carrier in the ingress noise signal.
– m – The modulation depth m of each individually modulated carrier shall be m = 0.32, to
enable a modulation index of at least 80 % during the peak levels of the modulation signal
m × αk(t).
– αk(t) – The normalized modulation noise αk(t) of each individually modulated carrier shall
be random in nature, shall be Gaussian distributed in nature, shall have an rms value of
αrms = 1, shall have a crest factor of 2.5 or more, and shall be uncorrelated with the
modulation noise of each other modulated carrier in the ingress noise signal.
– ∆b – The modulation width ∆b of each modulated carrier shall be at least 2 × 5 kHz. This is
equivalent to creating αk(t) from white noise, filtered by a low-pass filter with a cut-off
frequency at ∆b/2 = 5 kHz. This modulation width covers the full modulation band used by
AM broadcast stations.
NOTE 2 – The precise specification of the spectral shape requirements of the modulation signal is for further
study.
Table B.5/G.991.2 – Average minimum RFI noise power versus frequency

frequency
153 207 270 531 603 711 801 909 981 1296
(kHz)
power
–70 –44 –70 –70 –49 –70 –70 –44 –70 –49
(dBm)
B.3.5.3.6 Amateur RF noise generator [G6]

The amateur radio noise generator is identical to the broadcast RF noise generator with different
frequency and power values. These values are for further study.
110 ITU-T Rec. G.991.2 (12/2003)

B.3.5.3.7 Impulse noise generator [G7]
A test with this noise generator is required to prove the burst noise immunity of the SHDSL
transceiver. This immunity shall be demonstrated on short and long loops and noise to model
crosstalk and RFI.
B.3.5.4 Profiles of the individual impairment generators
Crosstalk noise represents all impairment that originates from systems connected to adjacent wire
pairs that are bundled in the same cable. Their wires are coupled to the wires of the xDSL system
under test, causing this spectrum of crosstalk noise to vary with the electrical length of the test loop.
To simplify matters, the definition of crosstalk noise has been broken down into smaller, more
easily specified components. The two generators G1 and G2 represent the "equivalent disturbance".
Their noise level originate from a mixture of many disturbers in a real scenario, as if all disturbers
are collocated at the ends of the test loops.
This equivalent disturbance, filtered by the NEXT and FEXT coupling functions, will represent the
crosstalk noise that is to be injected in the test set-up. This approach has isolated their definition
from the NEXT and FEXT coupling functions of the cable. The noise generated by these two
equivalent disturbers is specified in this clause in the frequency domain as well as in the time
domain.
The frequency domain characteristics of each generator G1 and G2 is defined by a spectral profile,
so each noise model has its own pair of spectral profiles.
• The profiles X.C.# in this clause describe the total equivalent disturbance of a technology
mix that is virtually colocated at the STU-C end of the test loop. This noise is represented
by equivalent disturbance generator G1, when stressing upstream signals, and by equivalent
disturbance generator G2 when stressing downstream signals.
• The profiles X.R.# in this clause describe the total equivalent disturbance of a technology
mix that is virtually colocated at the STU-R end of the test loop. This noise is represented
by equivalent disturbance generator G2, when stressing upstream signals, and by equivalent
disturbance generator G1 when stressing downstream signals.
Note that the PSD levels of equivalent disturbance generator G1 and G2 are interchanged when
changing from upstream testing to downstream testing.
B.3.5.4.1 Frequency domain profiles for SHDSL
This subclause specifies the PSD profiles X.R.# and X.C.# that apply for the equivalent disturbers
G1 and G2 when testing SHDSL systems. In this nomenclature, "#" is used as a placeholder for
noise model "A", "B" ,"C" and "D".
Four noise models have been defined for SHDSL:
• Type "A" models are intended to represent a high penetration scenario where the SHDSL
system under test is placed in a distribution cable (up to hundreds of wire pairs) that is filled
with many other (potentially incompatible) transmission systems.
• Type "B" models are intended to represent a medium penetration scenario where the
SHDSL system under test is placed in a distribution cable (up to tens of wire pairs) that is
filled with many other (potentially incompatible) transmission systems.
• Type "C" models are intended to represent a legacy scenario that accounts for systems
such as ISDN-PRI (HDB3), in addition to the medium penetration scenario of model "B".
• Type "D" models are intended as reference scenario to demonstrate the difference between
a cable filled with SHDSL only, or filled with a mixture of SHDSL techniques.
ITU-T Rec. G.991.2 (12/2003) 111

The PSD profiles for each noise model are build up by a weighed sum of two individually defined
profiles: self and alien crosstalk profiles.
X.C.# = (XS.C.# ♦ XA.C.#)
X.R.# = (XS.R.# ♦ XA.R.#)
The symbols in this expression refer to the following:
• Symbols "#" is used as a placeholder for noise model "A", "B", "C" or "D".
• Symbols "XS.C.#" and "XS.R.#" refer to the self crosstalk profiles, as defined in
B.3.5.4.1.1.
• Symbol "XA.C.#" and "XA.R.#" refer to the alien crosstalk profiles, as defined in
B.3.5.4.1.2.
• Symbol "♦" refers to the crosstalk sum of two PSDs, defined as
(
PX = PXS K n + PXA K n ) 1/ Kn
where P denotes the PSDs in W/Hz, and Kn = 1/0.6.
These profiles shall be met for all frequencies between 1 kHz to 1 MHz.
B.3.5.4.1.1 Self crosstalk profiles
The noise profiles XS.C.# and XS.R.#, representing the equivalent disturbance of self crosstalk, are
specific to the PSD parameters of the SHDSL system under test, defined by the specific payload,
symmetry and power-back-off features. For compliance with the requirements of this
Recommendation, the appropriate nominal PSD from B.4 shall be used.
For testing SHDSL, four noise models for self crosstalk have been defined. The STU-R and STU-C
profiles are specified in Table B.6.
In this nomenclature, "#" is a placeholder for model "A", "B" ,"C" or "D". "SHDSL.dn" is the signal
spectrum that SHDSL transmits in downstream direction, and "SHDSL.up" in upstream direction.
Table B.6/G.991.2 – Definition of the self crosstalk for SHDSL testing

Model A (XS.#.A) Model B (XS.#.B) Model C (XS.#.C) Model D (XS.#.D)
XS.C.#: "SHDSL.dn" + 11.7 dB "SHDSL.dn" + 7.1 dB "SHDSL.dn" + 7.1 dB "SHDSL.dn" + 10.1 dB
XS.R.#: "SHDSL.up" + 11.7 dB "SHDSL.up" + 7.1 dB "SHDSL.up"+ 7.1 dB "SHDSL.up" + 10.1 dB
NOTE – The different noise models use different Gain factors.
B.3.5.4.1.2 Alien crosstalk profiles

The noise profiles XA.C.# and XA.R.#, representing the equivalent disturbance of alien crosstalk,
are implementation specific for the SHDSL system under test. For testing SHDSL, four noise
models for alien crosstalk have been defined. The STU-C profiles are specified in Table B.7 and the
STU-R profiles in Table B.8. Each PSD profile originates from a mix of disturbers. The alien noise
in model D is made inactive, to achieve one pure self crosstalk scenario.
112 ITU-T Rec. G.991.2 (12/2003)

Table B.7/G.991.2 – Break frequencies of the "XA.C.#" PSD profiles that specify
the equivalent disturbance spectra of alien disturbers
XA.C.A 135 Ω XA.C.B 135 Ω XA.C.C 135 Ω XA.C.D 135 Ω

[Hz] [dBm/Hz] [Hz] [dBm/Hz] [Hz] [dBm/Hz] [Hz] [dBm/Hz]
1 −20.0 1 −25.7 1 −25.7
15 k −20.0 15 k −25.7 15 k −25.7
30 k −21.5 30 k −27.4 30 k −27.4 ALL –∞
67 k −27.0 45 k −30.3 45 k −30.3
125 k −27.0 70 k −36.3 70 k −36.3
138 k −25.7 127 k −36.3 127 k −36.3
400 k −26.1 138 k −32.1 138 k −32.1
1104 k −26.1 400 k −32.5 400 k −32.5
2.5 M −66.2 550 k −32.5 550 k −32.5
4.55 M −96.5 610 k −34.8 610 k −34.8
30 M −96.5 700 k −35.4 700 k −35.3
1104 k −35.4 1104 k −35.3
4.55 M −03.0 1.85 M −58.5
30 M −103.0 22.4 M −103.0
30 M −103.0
NOTE – The PSD profiles are constructed with straight lines between these break frequencies, when
plotted against a logarithmic frequency scale and a linear dBm scale. The levels are defined with a
135 Ω resistive load.
Table B.8/G.991.2 – Break frequencies of the "XA.R.#" PSD profiles that specify
XA.R.A 135 Ω XA.R.B 135 Ω XA.R.C 135 Ω XA.R.D 135 Ω

1 −20.0 1 −25.7 1 −25.7

15 k −20.0 15 k −25.7 15 k −25.7
60 k −25.2 30 k −26.8 30 k −26.8 ALL –∞
276 k −25.8 67 k −31.2 67 k −31.2
500 k −51.9 142 k −31.2 142 k −31.2
570 k −69.5 156 k −32.7 156 k −32.7
600 k −69.9 276 k −33.2 276 k −33.2
650 k −62.4 400 k −46.0 335 k −42.0
763 k −62.4 500 k −57.9 450 k −47.9
1.0 M −71.5 570 k −75.7 750 k −45.4
2.75 M −96.5 600 k −76.0 1040 k −45.5
30 M −96.5 650 k −68.3 2.46 M −63.6
763 k −68.3 23.44 M −103.0
ITU-T Rec. G.991.2 (12/2003) 113

Table B.8/G.991.2 – Break frequencies of the "XA.R.#" PSD profiles that specify
XA.R.A 135 Ω XA.R.B 135 Ω XA.R.C 135 Ω XA.R.D 135 Ω

1.0 M −77.5 30 M −103.0

2.8 M −103.0
30 M −103.0
NOTE – The PSD profiles are constructed with straight lines between these break frequencies, when
plotted against a logarithmic frequency scale and a linear dBm scale. The levels are defined with a
135 Ω resistive load.
B.3.5.4.2 Time domain profiles of generators G1-G4

The noise, as specified in the frequency domain in B.3.5.3.1 to B.3.5.3.4, shall be random in nature
and near Gaussian distributed. This means that the amplitude distribution function of the combined
impairment noise injected at the adding element shall lie between the two boundaries as illustrated
in Figure B.6, where the non-shaded area is the allowed region. The boundaries of the mask are
specified in Table B.9.
It is expected that noise generators will generate signals that are approximately Gaussian.
Therefore, the upper bound of Figure B.6 is lost. PDFs of signals generated by noise generators are
expected to be well below the upper bound allowed by the PDF mask shown in Table B.9.
The amplitude distribution function F(a) of noise u(t) is the fraction of the time that the absolute
value of u(t) exceeds the value "a". From this definition, it can be concluded that F(0) = 1 and that
F(a) monotonically decreases up to the point where "a" equals the peak value of the signal. From
there on, F(a) vanishes:
F (a) = 0 , for a ≥ u peak
The boundaries on the amplitude distribution ensure that the noise is characterized by peak values
that are occasionally significantly higher than the rms-value of that noise (up to 5 times the
rms-value).
114 ITU-T Rec. G.991.2 (12/2003)

Figure B.6/G.991.2 – Mask for the amplitude distribution function
Table B.9/G.991.2 – Upper and lower boundaries of the

amplitude distribution function of the noise
σ = rms value of noise)

Boundary (σ Interval Parameter Value
Flower(a) = (1 – ε) · {1 – erf((a/σ)/√2) } 0 ≤ a/σ < CF Crest factor CF = 5
Flower(a) = 0 CF ≤ a/σ < ∞ Gaussian gap ε = 0.1
Fupper(a) = (1 + ε) · {1 – erf((a/σ)/√2) } 0 ≤ a/σ < A A = CF/2 = 2.5
Fupper(a) = (1 + ε) · {1 – erf(A/√2) } A ≤ a/σ < ∞
The meaning of the parameters in Table B.9 is as follows:

− CF denotes the minimum crest factor of the noise, that characterizes the ratio between the
absolute peak value and rms value (CF = |upeak|/urms).
− ε denotes the Gaussian gap that indicates how "close" near Gaussian noise approximates
true Gaussian noise.
− A denotes the point beyond which the upper limit is alleviated to allow the use of noise
signals of practical repetition length.
B.3.5.5 Mandatory noise shape substitution rule
The strict application of the test procedure requires a different noise shape for each test although
some of the noise shapes are very similar. In order to reduce the number of possible noise shapes,
the following substitution rule is mandatory. It reduces the number of noise shapes from 280 to 22.
ITU-T Rec. G.991.2 (12/2003) 115

Table B.9a tabulates the noise substitution rule. The following nomenclature is used to describe a
shape:
"Side (C or R) Rate (384 to 2304) PSDType(s for symmetric) NoiseModel (A to D)"
Example 1: C384sA2 represents the noise shape on the STU-C side for the 384 kbit/s rate using the
symmetric PSD corresponding to noise model A and loop 2.
Example 2: C384sAX represents the noise shape on the STU-C side for the 384 kbit/s rate using the
symmetric PSD corresponding to noise model A and any loop.
Example 3: Rule 7 requires that the following noise shapes: R384sA1, R384sA2, R384sA3,
R384sA4, R384sA5, R384sA6, R384sA7, R512sA1, R512sA2, R512sA3, R512sA4, R512sA5,
R512sA6, R512sA7 be replaced by the single noise shape R768sA2.
Example 4: Conducting Test Set 3 of Table B.3 for 384 kbit/s at the STU-C end. The loop and
transceiver would be set up as per the test description (Loop #3 upstream set to 43 dB @ 150 kHz,
which is equivalent to a length of 5563 m). The transceiver would be set to 384 kbit/s. The noise
shape injected would be 'R768sC2' rather than 'C384sD3' (rule 9).
Table B.9a/G.991.2 – Noise shape substitution rule

Rule This
Replaces those shapes (on a row by row basis)
# shape
1 'C768sA2' 'C384sAX' 'C512sAX'
2 'C768sC2' 'C384sBX' 'C512sBX' 'C384sCX' 'C512sCX'
3 'C1536sA2' 'C768sAX' 'C1024sAX' 'C1280sAX'
4 'C1536sC2' 'C768sBX' 'C1024sBX' 'C1280sBX' 'C768sCX' 'C1024sCX' 'C1280sCX'
5 'C2304sA2' 'C1536sAX' 'C2048sAX' 'C2304sAX' 'C1536sAX'
6 'C2304sC2' 'C1536sBX' 'C2048sBX' 'C2304sBX' 'C1536sCX' 'C2048sCX' 'C2304sCX'
7 'R768sA2' 'R384sAX' 'R512sAX'
8 'R768sB2' 'R384sBX' 'R512sBX'
9 'R768sC2' 'R384sCX' 'R512sCX' 'C384sDX' 'R384sDX' 'C512sDX' 'R512sDX'
10 'R1536sA2' 'R768sAX' 'R1024sAX' 'R1280sAX' 'R1536sAX'
11 'R1536sB2' 'R768sBX' 'R1024sBX' 'R1280sBX' 'R1536sBX'
12 'R1536sC2' 'R768sCX' 'R1024sCX' 'R1280sCX' 'R1536sCX'
13 'R2048sA2' 'R2048sAX'
14 'R2048sB2' 'R2048sBX'
15 'R2048sC2' 'R2048sCX'
16 'R2304sA2' 'R2304sAX'
17 'R2304sB2' 'R2304sBX'
18 'R2304sC2' 'R2304sCX'
19 'C1280sD2' 'C768sDX' 'R768sDX' 'C1280sDX' 'R1280sDX'
20 'C1536sD2' 'C1024sDX' 'R1024sDX' 'C1536sDX' 'R1536sDX'
21 'C2048sD2' 'C2048sDX' 'R2048sDX'
22 'C2304sD2' 'C2304sDX' 'R2304sD'
116 ITU-T Rec. G.991.2 (12/2003)

B.3.5.6 Measurement of noise margin
At start-up, the level and shape of crosstalk noise or impulse noise are adjusted, while their level is
probed at port RX to meet the impairment level specification in B.3.4. This relative level is referred
to as 0 dB. The transceiver link is subsequently activated, and the bit error ratio of the link is
monitored.
B.3.5.6.1 Measurement of crosstalk noise margin
For measuring the crosstalk margin, the crosstalk noise level of the impairment generator as defined
in B.3.5.4.1 shall be increased by adjusting the gain of amplifier A1 in Figure B.5, equally over the
full frequency band of the SHDSL system under test, until the bit error ratio is higher than 10−7.
This BER will be achieved at an increase of noise of x dB, with a small uncertainty of ∆ dB. This
value x is defined as the crosstalk noise margin with respect to a standard noise model. The
indicated noise margins shall have a tolerance of 1.25 dB due to the aggregate effect of crosstalk
generator tolerance and calibrated loop simulator tolerance. The offset ∆ is defined using the same
procedure as in A.3.1.4.
The noise margins shall be measured (after allowing a minimum 5-minute fine tuning period) using
the test loops specified in Figure B.1 and scaled according to Tables B.1 and B.2.
NOTE – Currently, the injected noise, for the purpose of crosstalk noise margin measurement, consists of the
sum of generators G1, G2 and G4 as described in B.3.5.1. Appendix IV tabulates the values of the injected
noise corresponding to 0 dB margin and a white noise generator value of –140 dBm/Hz. The injected noise
should be measured as per B.3.3. The mandatory test cases are described in B.3.4. A mandatory noise
substitution rule is described in B.3.5.5.
B.3.5.6.2 Measurement of impulse noise margin
For further study.
B.4 PSD masks

For all data rates, the measured transmit PSD of each STU shall not exceed the PSD masks
specified in this clause (PSDMASKSHDSL(f)), and the measured total power into 135 Ω shall fall
within the range specified in this clause (PSHDSL ± 0.5 dB).
Support for the symmetric PSDs specified in B.4.1 shall be mandatory for all supported data rates.
Support for the asymmetric PSDs specified in B.4.2 shall be optional.
Table B.10 lists the supported PSDs and the associated constellation sizes.
Table B.10/G.991.2 – PSD and constellation size

Symmetric PSDs Asymmetric PSDs
DS US DS US DS US
16-TCPAM 16-TCPAM 16-TCPAM 16-TCPAM 8-TCPAM 16-TCPAM
Mandatory Optional For further study
For the 16-TCPAM upstream and downstream constellations shown in Table B.10, the details of
payload data rate, the associated symbol rate, and the mapping of bits per symbol are specified in
Table B.11.
ITU-T Rec. G.991.2 (12/2003) 117

Table B.11/G.991.2 – Framed data mode rates
Payload data rate, Symbol rate K
Modulation
R (kbit/s) (ksymbol/s) (Bits per symbol)
R = n × 64 + (i) × 8 16-TCPAM (R + 8) ÷ 3 3
As specified in clause 5, the allowed rates are given by n × 64 + i × 8 kbit/s, where 3 ≤ n ≤ 36 and
0 ≤ i ≤ 7. For n = 36, i is restricted to the values of 0 or 1.
B.4.1 Symmetric PSD masks
   πf 
2
 sin   
 − PBO K   Nf sym 
MaskedOffsetdB ( f )
10 10 × SHDSL × 1 ×   ×
1
×10 10 , f < f int
 135 f sym 2 2 × Order
 πf   f 
   1 +  
PSDMASK SHDSL ( f ) =   Nf sym   f 3dB 
  
0.5683 × 10 −4 × f −1.5 , f int ≤ f ≤ 1.5 MHz

− 90 dBm/Hz peak with maximum power in a [ f , f + 1 MHz ] window of
− 50 dBm, 1.5 MHz < f ≤ 11.04 MHz


 f 3 dB − f
1 + 0.4 × , f < f 3 dB
1 , f ≥ f 3 dB

NOTE 1 – Large PSD variations over narrow frequency intervals (for example near the junction of the main
fint is the frequency where the two functions governing PSDMASKSHDSL(f) intersect in the frequency
range from 0 to Nfsym. PBO is the power backoff value in dB. KSHDSL, Order, N, fsym, f3dB, and PSHDSL
are defined in Table B.12. PSHDSL is the range of power in the transmit PSD with 0 dB power
backoff. R is the payload data rate.
Table B.12/G.991.2 – Symmetric PSD parameters

Symbol rate
Payload data
KSHDSL Order N fsym f3 dB PSHDSL (dBm)
rate, R (kbit/s)
(ksymbol/s)
R < 2048 7.86 6 1 (R + 8)/3 1.0 × fsym/2 P1(R) ≤ PSHDSL ≤ 13.5
R ≥ 2048 9.90 6 1 (R + 8)/3 1.0 × fsym/2 14.5
P1(R) is defined as follows:

P1( R) = 0.3486 log 2 ( R × 1000 + 8000) + 6.06 dBm
118 ITU-T Rec. G.991.2 (12/2003)

Figure B.7 shows the PSD masks with 0 dB power backoff for payload data rates of 256, 512, 768,
1536, 2048 and 2304 kbit/s.
Figure B.7/G.991.2 – PSD masks for 0 dB power backoff
   πf  
2
 sin   
 − PBO K 1   Nf sym  1 f2
10 10 × SHDSL
× × × × , f < f int
2 2 × Order
 135 f sym  πf   f  f 2 + f c2
   1 +  
Nominal PSD ( f ) =   Nf sym   f 3dB 
  
0.5683 × 10 −4 × f −1.5 , f int ≤ f ≤ 1.5 MHz

− 50 dBm , 1.5MHz < f ≤ 11.04 MHz

where fc is the transformer cut-off frequency, assumed to be 5 kHz. Figure B.8 shows the nominal
transmit PSDs with 13.5 dBm power for payload data rates of 256, 512, 768, 1536, 2048 and
2304 kbit/s.
NOTE 2 – The nominal PSD is given for information only.
ITU-T Rec. G.991.2 (12/2003) 119

Figure B.8/G.991.2 – Nominal symmetric PSDs for 0 dB power backoff
NOTE 3 – In this clause, PSDMASK(f) and NominalPSD(f) are in units of W/Hz unless otherwise specified,
and f is in units of Hz.
B.4.2 Asymmetric 2.048 Mbit/s and 2.304 Mbit/s PSD masks
The asymmetric PSD mask set specified in this clause shall optionally be supported for the
2.048 Mbit/s and the 2.304 Mbit/s payload data rate. Power and power spectral density is measured
into a load impedance of 135 Ω.
For the 2.048 Mbit/s and the 2.304 Mbit/s payload data rates available in the STU, the following set
of PSD masks (PSDMASKSHDSL(f)) shall be selectable:
   πf 
2
 − PBO 
sin    MaskedOffsetdB ( f )
  f x 
10 10 × K SHDSL × 1 ×  ×
1
× 10 10 , f < f int
 2 2 × Order
 135 fx  πf   f 
   1 +  
PSDMASK SHDSL ( f ) =   fx   f 3dB 

0.5683 × 10 −4 × f −1.5 , f int ≤ f ≤ 1.5 MHz

− 50 dBm , 1.5 MHz < f ≤ 11.04 MHz

 f 3 dB − f
1 + 0.4 × , f < f 3 dB
1 , f ≥ f 3 dB

120 ITU-T Rec. G.991.2 (12/2003)

fint is the frequency where the two functions governing PSDMASKSHDSL(f) intersect in the frequency
range from 0 to fx. PBO is the power backoff value in dB. KSHDSL, Order, fx, f3dB and PSHDSL are
defined in Table B.13. PSHDSL is the range of power in the transmit PSD with 0 dB power backoff.
R is the payload data rate.
Table B.13/G.991.2 – Asymmetric PSD parameters

Payload date PSHDSL
Transmitter KSHDSL Order fx (Hz) f3dB (Hz)
rate (kbit/s) (dBm)
2048 STU-C 16.86 7 1 370 667 548 267 16.25
2048 STU-R 15.66 7 685 333 342 667 16.50
2304 STU-C 12.48 7 1 541 333 578 000 14.75
2304 STU-R 11.74 7 770 667 385 333 15.25
Figure B.9 shows the PSD masks with 0 dB power backoff for payload data rates of 2048 and
2304 kbit/s.
Figure B.9/G.991.2 – PSD masks for 0 dB power backoff
ITU-T Rec. G.991.2 (12/2003) 121

   πf 
2
 − PBO sin
   
  f x  f2
10 10 × K SHDSL × 1 ×  ×
1
× , f < f int
 2 2 × Order
 135 fx  πf   f  f 2 + f c2
   1 +  
NominalPSD ( f ) =   fx   f 3dB 

0.5683 × 10 −4 × f −1.5 , f int ≤ f ≤ 1.5 MHz

− 50 dBm, 1.5 MHz < f ≤ 11.04 MHz
where fc is the transformer cut off frequency, assumed to be 5 kHz. Figure B.10 shows the nominal
transmit PSDs with 0 dB power backoff for payload data rates of 2048 and 2304 kbit/s.
NOTE 2 – The nominal PSD is given for information only.
Figure B.10/G.991.2 – Nominal asymmetric PSDs for 0 dB power backoff

NOTE 3 – In this clause, PSDMASK(f) and NominalPSD(f) are in units of W/Hz unless otherwise specified,
and f is in units of Hz.
B.5 Region-specific functional characteristics

B.5.1 Data rate
For devices supporting Annex B functionality, no additional limitation on data rates shall be placed
beyond the limitations stated in clause 5 and reiterated in 7.1.1, 8.1 and 8.2. For details of the
supported symbol rates and their association with PSDs, see B.4.
122 ITU-T Rec. G.991.2 (12/2003)

B.5.2 Return loss
For devices supporting Annex B functionality, return loss shall be specified based on the
to the quantities shown in Figure 11-6:
RLMIN = 12 dB
f0 = 12.56 kHz
f1 = 50 kHz
f2 = fsym/2
f3 = 1.99fsym
NOTE – The intention of the return loss specification is to maintain some power constraint, even under
severe mismatched conditions, when SHDSL modems are connected to real cables. A minimum return loss
bounds the (complex) output impedance Zs within a restricted range around the design impedance
Rv = 135 Ω, and thus the maximum available power from that source. Therefore it is expected that the power
dissipated into a complex load impedance ZL should never exceed the appropriate PSD masks and maximum
aggregate powers for all values ZL in the range of 10 Ω < |ZL| < 2000 Ω, as specified for Rv = 135 Ω in B.4
and Tables B.12 and B.13. The extension of the existing power constraints to the severely mismatched case
is for further study.
B.5.3 Span powering
B.5.3.1 General
This clause deals with power feeding of the STU-R, regenerators (if required) and the provision of
power to the application interface for narrow-band services under restricted conditions (life line
circuit). The requirements given in this clause imply compliance to IEC 60950 [7].
B.5.3.2 Power feeding of the STU-R
The STU-R shall be able to consume power from the remote power feeding circuit when the local
power supply fails.
NOTE – The remote feeding strategy may not be applicable for extremely long lines or lines including
regenerators. In those cases specific feeding methods may be applied, which are for further study.
The STU-R shall be able to draw up to a maximum of 10 mA as wetting current from the remote
feeding circuit when the STU-R is being powered locally. When the local power fails, the
maximum current drawn by the STU-R from the remote feeding circuit shall be limited to the value
specified in IEC 60950 [7].
It is optional for the STU-C to provide wetting current.
B.5.3.3 Power feeding of the interface for narrow-band services
When simultaneous telephone service is provided by the STU-R, feeding of restricted mode power
for life line service has to be provided for at least one telephone set in case of local power failure.
NOTE – The remote feeding strategy may not be applicable for extremely long lines or lines including
regenerators. In those cases, specific feeding methods may be applied which are for further study.
ITU-T Rec. G.991.2 (12/2003) 123

B.5.3.4 Feeding power from the STU-C
The feeding power shall be limited to the values specified by the TNV requirements in
IEC 60950 [7].
NOTE – This means that the sum of the DC- and AC-voltage at the STU-R may not exceed 120 V. The
safety standards may for extraordinary cases with long lines or regenerators allow higher power to be
supplied from the STU-C. This is left for further study. It is likely that supporting long lines and/or
regenerators may imply floating (not connected to ground) power feeding circuits.
B.5.3.5 Power available at the STU-R
The STU-R shall be able to deal with any polarity. With a minimum voltage of 45 V (see Note) at
the input of the STU-R, it shall enter a full operational state.
NOTE – This value depends on the supply voltage and is for further study.
When remote power feeding is provided by the network, the STU-R and the side of the SRU
directed towards the STU-C shall enter a high impedance state within 2 s after interruption of the
remote current fed towards the STU-R or the SRU respectively. This state shall be maintained as
long as the voltage on the line stays below 18 V (DC + AC peak). In this state the leakage current
shall be less than 10 µA and the capacitance shall be greater than 2 µF. A guard time of at least 2 s
between removing the remote power and applying a test voltage is necessary.
B.5.4 Longitudinal balance
For devices supporting Annex B functionality, longitudinal balance shall be specified based on the
to the quantities in Figure 11-2.
LBMIN = 40 dB
f1 = 5 kHz
f2 = fsym/2
B.5.5 Longitudinal output voltage
For devices supporting Annex B functionality, longitudinal output voltage shall be specified based
on the methodology of 11.2. The measurement frequency range shall be between 100 Hz and
400 kHz.
B.5.6 PMMS target margin
If the optional line probe is selected during the G.994.1 session, the receiver shall use the negotiated
target margin. If worst-case PMMS target margin is selected, then the receiver shall assume the
disturbers of Table B.14 to determine if a particular rate can be supported. Reference crosstalk shall
be computed using the cable crosstalk models of B.3.5.2, assuming infinite loop length so that
FEXT components are ignored and NEXT is independent of loop length. The reference crosstalk
specified in this clause may not be representative of worst-case conditions in all networks.
Differences between crosstalk environments may be compensated by adjusting the target margin.
124 ITU-T Rec. G.991.2 (12/2003)

Table B.14/G.991.2 – Reference disturbers used during PMMS
for worst-case target margin
Rate (kbit/s) PSD (direction) Reference disturber
All Symmetric (US/DS) 49 SHDSL
2048 Asymmetric (US) 49 SHDSL-SYM with fsym = 685 333 Hz
2048 Asymmetric (DS) 49 SHDSL-SYM with fsym = 685 333 Hz
2304 Asymmetric (US) 49 SHDSL-SYM with fsym = 770 667 Hz
2304 Asymmetric (DS) 49 SHDSL-SYM with fsym = 770 667 Hz
B.5.7 Span powering in M-pair mode

In the optional M-pair mode, the requirements for remote power feeding or wetting current for each
of the M pairs shall be identical to the requirements for a single pair specified in B.5.3.
NOTE – This implies that the powering/wetting current is provided by a potential difference between tip and
ring on each of the M pairs.
Annex C
See Annex H/G.992.1 [1] for specifications of transceivers for use in networks with existing
TCM-ISDN service (as specified in Appendix III/G.961 [B1]).
Annex D
Signal regenerator operation
In order to achieve data transmission over greater distances than are achievable over a single
SHDSL segment, one or more signal regenerators (SRUs) may be employed. In the optional M-pair
mode, M-pair regenerators may be used when this reach extension is required. This annex specifies
operational characteristics of signal regenerators and the start-up sequence for SHDSL spans
containing signal regenerators. Additional explanatory text is included in Appendix III.
D.1 Reference diagram

Figure D.1 is a reference diagram of a SHDSL span containing two regenerators. Up to eight (8)
regenerators per span are supported within the EOC addressing scheme (9.5.5.5), and no further
limitation is intended herein. Each SRU shall consist of two parts: an SRU-R for interfacing with
the STU-C (or a separate SRU-C), and an SRU-C for interfacing with the STU-R (or a separate
SRU-R). An internal connection between the SRU-R and SRU-C shall provide the communication
between the two parts during start-up and normal operation. An SHDSL span containing
X regenerators shall contain X + 1 separated SHDSL segments, designated TR1 (STU-C to SRU1),
TR2 (SRUX-C to STU-R), and RRn (SRUn-C to SRUn + 1 − R, where 1 ≤ n ≤ X – 1). Each segment
shall follow the general principles described in 6.2, 6.3 and 7.2 for the pre-activation and activation
procedures. Additional requirements specific to spans containing regenerators are described in this
annex.
ITU-T Rec. G.991.2 (12/2003) 125

Figure D.1/G.991.2 – Block diagram of a SHDSL span with two signal regenerators
D.2 Start-up procedures

D.2.1 SRU-C
Figure D.2 shows the State Transition Diagram for SRU-C start-up and operation. The SRU-C
begins in the "Idle" state and, in the case of an STU-R initiated start-up, transitions first to the "Wait
for STU-C" state. For an STU-C initiated start-up, the SRU-C moves from "Idle" to the "G.994.1
Session 1" state. An SRU initiated start-up shall function identically to an STU-C initiated start-up
from the perspective of the SRU-C.
The SRU-C shall communicate "Capabilities Available" status and transfer a list of its capabilities
to the SRU-R across the regenerator's internal interface upon entering the "Wait for STU-C" state.
The SRU-C's capabilities list, as transferred to the SRU-R, shall be the intersection of its own
capabilities, the capabilities list it received from the STU-R (or SRU-R) in its G.994.1 session, and
the segment capabilities determined by the line probe, if used.
The SRU-C shall receive mode selection information from the SRU-R in association with the
"SRU-R Active" indication. In the subsequent G.994.1 session, the SRU-C shall select the same
mode and parameter settings for the SHDSL session.
The timer TSRUC shall be set to 4 minutes. If TSRUC expires before the SRU-C reaches the "Active"
state, the SRU-C shall return to the "Idle" state and shall indicate link failure to the SRU-R across
the internal interface. The SRU-C shall also indicate failure and return to the "Idle" state if a
G.994.1 initiation is unsuccessful after 30 s.
The "Diagnostic Mode" bit, if set in the G.994.1 Capabilities Exchange, shall cause an SRU-C to
function as an STU-C if the subsequent segment fails. This implies that an internal failure indication
received while in the "Wait for STU-C" state shall cause the SRU-C to select an operational mode,
initiate a G.994.1 session, and transition to state "G.994.1 Session 2".
126 ITU-T Rec. G.991.2 (12/2003)

Idle G.994.1
Receive internal failure
"Link Initiation" indication
Initiate G.994.1 session G.994.1
failure
Receive MS
with "Silent Period" = 1
Complete G.994.1 session TSRUC timeout
Internal failure
Start TSRUC Indication and Receive G.994.1 initiation
not diagnostic
G.994.1 session
mode
Capabilities exchange
(Line probe)
(Second G.994.1 session)
Receive G.994.1 initiation TSRUC timeout (Capabilities exchange)
G.994.1 session Send MS with "Silent Period" = 1
(Line probe) Start TSRUC
(Capabilities exchange)
Send MS with "Silent Period" = 1 Receive G.994.1 initiation
Restart TSRUC G.994.1 session
Internal failure MS with "Silent Period" = 1
Indication and
Receive capabilities exchange Diagnostic Mode Receive internal
"SRU-R Active" indication
Capabilities exchange Initiate G.994.1
(Line probe) Session Initiate G.994.1 session
Send MS with "Silent Period" = 1 G.994.1 Session 2
Restart TSRUC
Complete G.994.1 session
with mode selection
Receive internal
Disconnect, "SRU-R Active" indication SHDSL Activation
failure, or retrain
Initiate G.994.1 session
Active
Figure D.2/G.991.2 – SRU-C state transition diagram
D.2.2 SRU-R
Figure D.3 shows the State Transition Diagram for SRU-R start-up and operation. The SRU-R
begins in the "Idle" state and, in the case of an STU-R initiated train, transitions first to the "G.994.1
Session 1" state. For an STU-C initiated train, the SRU-C moves from "Idle" to the "G.994.1
Session 2" state.
ITU-T Rec. G.991.2 (12/2003) 127

The SRU-R shall communicate "Link Initiation" status to the SRU-C across the regenerator's
internal interface upon entering the "Wait for STU-R" state. Upon entering the "Active" state, it
shall communicate "SRU-R Active" status to the SRU-C. If plesiochronous operation (Clock
Mode 1; see clause 10) is selected, the SRU-R may optionally indicate its entry into the "Active"
state to the SRU-C prior to the completion of the SHDSL activation sequence. If synchronous or
network referenced plesiochronous clocking is selected (Clock Modes 2, 3a or 3b; see clause 10),
the SRU-R shall not indicate entry into the "Active" state until the SHDSL activation sequence has
been completed.
The SRU-R shall receive a list of capabilities from the SRU-C across the regenerator's internal
interface in association with the "Capabilities Available" indication. The SRU-R's capabilities list,
as indicated in the subsequent G.994.1 session, shall be the intersection of its own capabilities with
the capabilities list it received from the SRU-C.
The SRU-R shall provide mode selection information to the SRU-C in association with the "SRU-R
Active" indication, based on the selections it has received in the G.994.1 session.
The timer TSRUR shall be set to 4 minutes. If TSRUR expires before the SRU-R reaches the "Active"
state, the SRU-R shall return to the "Idle" state and shall indicate link failure to the SRU-C across
the internal interface. The SRU-R shall also indicate failure and return to the "Idle" state if a
G.994.1 initiation is unsuccessful after 30 s.
The "Diagnostic Mode" bit, if set in the G.994.1 Capabilities Exchange, shall cause an SRU-R to
function as an STU-R if the subsequent segment fails. This implies that an internal failure indication
received while in the "Wait for STU-R" state shall cause the SRU-R to initiate a G.994.1 session
and transition to state "G.994.1 Session 2".
128 ITU-T Rec. G.991.2 (12/2003)

Idle
Receive G.994.1 initiation
TSRUR timeout
Receive internal
"Capabilities Available"
G.994.1 session 1 indication
TSRUR timeout
Internal failure G.994.1
Receive Ack(1) failure
Indication and
Complete G.994.1 session not diagnostic
mode
Start TSRUR
Receive internal
"Capabilities Available"
indication
G.994.1 session
Send MS with "Silent Period" = 1 G.994.1 session 2
Internal failure
Indication and
diagnostic mode
Initiate G.994.1
session
Receive REQ-MR
Start G.994.1 session normally
Select mode
SHDSL activation Capabilities exchange
(Line probe)
Receive MS with
"Silent Period" = 1
Start TSRUR
Receive mode selection

Disconnect, SHDSL activation Wait for STU-C
failure or retrain

Select mode
SHDSL activation
Active
Figure D.3/G.991.2 – SRU-R state transition diagram
D.2.3 STU-C
In order to support operation with regenerators, each STU-C shall support the Regenerator Silent
Period (RSP) bit, as specified in ITU-T Rec. G.994.1. Second, the STU-C shall not indicate a
training failure or error until it has been forced into "silent" mode for at least 5 consecutive minutes.
ITU-T Rec. G.991.2 (12/2003) 129

D.2.4 STU-R
In order to support operation with regenerators, each STU-R shall support the Regenerator Silent
Period (RSP) bit, as specified in ITU-T Rec. G.994.1. The STU-R shall not indicate a training
failure or error until it has been forced into "silent" mode for at least 5 consecutive minutes.
D.2.5 Segment failures and retrains
In the case of a segment failure or a retrain, each segment of the span shall be deactivated, with
each SRU-C and each SRU-R returning to its "Idle" state. The restart may then be initiated by the
SRU, the STU-R, or the STU-C.
D.3 Symbol rates

For Annex A operational modes, signal regenerators may transmit at symbol rates up to and
including 280 ksymbol/s in either two-wire or the optional M-pair mode. This corresponds, for
16-TCPAM, to maximum user data rates (not including framing overhead) of 832 kbit/s and
M × 832 kbit/s for two-wire and M-pair operation, respectively. Operation at higher symbol rates is
for further study.
For Annex B operational modes, signal regenerators may transmit at symbol rates up to and
including 685.33 ksymbol/s in either two-wire or the optional M-pair mode. This corresponds, for
16-TCPAM, to maximum user data rates (not including framing overhead) of 2.048 Mbit/s and
M × 2.048 Mbit/s for two-wire and M-pair operation, respectively. Operation at higher symbol rates
is for further study.
In either case, each STU and SRU on a span shall select the same operational data rate.
D.4 PSD masks

Any of the PSDs from Annex A or Annex B may be used for the TR1 segment (STU-C to SRU1-R),
as appropriate to the given region. All other segments shall employ one of the appropriate
symmetric PSDs, as described in either A.4.1 or B.4.1. The selection of PSD shall be limited by the
symbol rate considerations of D.3.
Annex E
Application-specific TPS-TC framing
This annex provides implementation details for the various types of TPS-TC framing that may be
supported by SHDSL transceivers. The TPS-TC framing mode is selected during pre-activation, but
the criteria for selecting a particular TPS-TC mode are application-specific and are beyond the
E.1 TPS-TC for clear channel data

In Clear Channel mode, there shall be no specified relationship between the structure of the user
data and its positioning within the Payload Sub-Blocks. ks bits of contiguous user data shall be
contained within each Sub-Block, as specified in 8.1. The temporal relationship between the user
data stream and the data within the Sub-Blocks shall be maintained such that the order of bits in
time from the user data stream shall match the order of transmission within the SHDSL Payload
Sub-Blocks. Any additional structure within the user data shall be maintained by an unspecified
higher layer protocol and is outside the scope of this Recommendation.
130 ITU-T Rec. G.991.2 (12/2003)

In the optional M-pair mode, clear channel data will be carried over all pairs using interleaving, as
described in 8.2. The bitstream of the user data consisting of M × ks bits is mapped to the M pairs by
placing alternating bitstreams consisting of ks bits of contiguous user data in each of the M SHDSL
channels. ks bits of contiguous user data shall be contained within a Sub-Block on Pair 1, and the
following sets of ks bits of contiguous user data shall be contained within the corresponding Sub-
Blocks of subsequent pairs. As noted above, any additional structure within the user data shall be
maintained by an unspecified higher layer protocol and is outside the scope of this
Recommendation.
E.2 TPS-TC for clear channel byte-oriented data

In the byte-oriented clear channel mode, the input byte stream shall be aligned within the SHDSL
Payload Sub-Block such that the byte boundaries are preserved. Each Payload Sub-Block is treated
as containing n 8-bit time slots. Each byte from the input data stream is mapped LSB-first into the
next available time slot. The first time slot begins at the first bit position within the Payload Sub-
Block, followed by time slot 2, time slot 3, … , time slot n. A total of ks bits (or n bytes) of
contiguous data shall be contained within each Sub-Block, as specified in 8.1, where ks = i + n × 8,
and, in this mode, i = 0 and 3 ≤ n ≤ 36. Note that optional extensions described in Annex F allow
values of n up to 89. See Figure E.1 for additional details.
...
Figure E.1/G.991.2 – Clear channel byte-oriented framing
In the optional M-pair mode, byte-oriented data is carried over all M pairs using interleaving, as
described in 8.2. A total of M × ks bits (M × n bytes) of byte-oriented data shall be transported per
SHDSL Payload Sub-Block. ks = i + n × 8, and, in this mode, i = 0 and 3 ≤ n ≤ 36. Note that
optional extensions described in Annex F allow values of n up to 89. Only numbers of time slots
divisible by M may be supported in M-pair mode. The input byte stream shall be aligned within the
SHDSL Payload Sub-Block such that the byte boundaries are preserved. Each Payload Sub-Block is
treated as containing M × n 8-bit time slots. Each byte from the input data stream is mapped LSB-
first into the next available time slot. The first time slot begins at the first bit position within the
Payload Sub-Block, followed by time slot 2, time slot 3, … , time slot n. A total of M × ks bits (or
ITU-T Rec. G.991.2 (12/2003) 131

M × n bytes) of contiguous data shall be contained within each Sub-Block, as specified in 8.1,
where ks = i + n × 8, and, in this mode, i = 0 and 3 ≤ n ≤ 36. The bytes from the input data stream
shall be interleaved among all M pairs, such that pair M carries the mth byte out of every block of
M bytes. See Figure E.2 for additional details.
...
...
Figure E.2/G.991.2 – M-pair framing for byte-oriented clear channel (for the M = 2 case)
E.3 TPS-TC for unaligned DS1 transport

Much of the data within the North American network is structured as "DS1" data streams, which,
for purposes of this Recommendation, can be described as 1.544 Mbit/s data streams containing
8 kHz framing, with each frame containing 24 8-bit time slots and one framing bit. Details of DS1
framing and associated data structure can be found in 2.1/G.704 [B6].
In Unaligned DS1 mode, there shall be no specified relationship between the DS1 frames and their
positioning within the Payload Sub-Blocks. A total of ks bits of contiguous data shall be contained
within each Sub-Block, as specified in 8.1, where ks = i + n × 8, and, in this mode, n = 24 and i = 1.
The DS1 framing clocks shall be synchronized to the SHDSL clocks such that the DS1 frame
always appears in the same position within each SHDSL Payload Sub-Block; however, no
particular alignment is specified. The temporal relationship between the DS1 data stream and the
data within the Sub-Blocks shall be maintained, such that the order of bits in time from the DS1
data stream shall match the order of transmission within the SHDSL Payload Sub-Blocks. The
optional M-pair mode will not support Unaligned DS1 transport.
132 ITU-T Rec. G.991.2 (12/2003)

E.4 TPS-TC for aligned DS1/fractional DS1 transport
As noted in E.3, "DS1" data streams consist of 1.544 Mbit/s data streams containing 8 kHz framing,
with each frame containing 24 8-bit time slots and one framing bit. In some cases, "Fractional DS1"
data streams are used, where DS1 frames contain less than the normal 24 8-bit time slots. Aligned
DS1/Fractional DS1 mode is also applicable to 1.544 Mbit/s PRI (Primary Rate ISDN), as described
in 4.2/I.431 [B10].
In Aligned DS1/Fractional DS1 mode, each DS1 frame shall be aligned within the SHDSL Payload
Sub-Block such that the DS1 framing bit occupies the first bit position within the Payload Sub-
Block, followed by time slot 1, time slot 2, … , time slot n. A total of ks bits of contiguous data shall
be contained within each Sub-Block, as specified in 8.1, where ks = i + n × 8, and, in this mode,
i = 1. In DS1 applications, n = 24, and, in Fractional DS1 applications, 3 ≤ n < 24. The DS1 framing
clocks shall be synchronized to the SHDSL clocks such that the DS1 frame always appears in the
defined position within each SHDSL Payload Sub-Block. See Figure E.3 for additional details.
...
Figure E.3/G.991.2 – Aligned DS1/fractional DS1 framing
In the optional M-pair mode, DS1/Fractional DS1 data will be carried over all M pairs using
interleaving, as described in 8.2. A total of M × (ks – 1) + 1 bits of DS1/Fractional DS1 data shall be
transported per SHDSL Payload Sub-Block. ks = i + n × 8, and, in this mode, i = 1. In DS1
applications, n = 24/M, and in Fractional DS1 applications, 3 ≤ n < 24/M. In M-pair mode only
multiples of M DS1 time slots may be supported. Each DS1 frame shall be aligned within the
SHDSL Payload Sub-Block such that the DS1 framing bit occupies the first bit position within the
Payload Sub-Block on each of the M wire pairs. The time slots of the DS1 frame shall be
interleaved among all M wire pairs, such that pair M carries the mth time slot out of every block of
M slots. See Figure E.4 for additional details.
ITU-T Rec. G.991.2 (12/2003) 133

...
...
Figure E.4/G.991.2 – M-Pair framing for DS1/fractional DS1 (for the M = 2 case)
E.5 TPS-TC for European 2048 kbit/s digital unstructured leased line (D2048U)
D2048U data streams contain unstructured 2.048 Mbit/s data with no specified framing. These data
streams shall be carried using the Clear Channel TPS-TC described in E.1.
E.6 TPS-TC for unaligned European 2048 kbit/s digital structured leased line (D2048S)
Much of the data within the European network is structured as D2048S data streams, which, for
purposes of this Recommendation, can be described as 2.048 Mbit/s data streams containing 8 kHz
framing, with each frame containing 32 8-bit time slots. Details of D2048S framing and associated
data structure can be found in 2.3/G.704 [B6].
In Unaligned D2048S mode, there shall be no specified relationship between the D2048S frames
and their positioning within the Payload Sub-Blocks. A total of ks bits of contiguous data shall be
contained within each Sub-Block, as specified in 8.1, where ks = i + n × 8, and, in this mode, n = 32
and i = 0. The D2048S framing clocks shall be synchronized to the SHDSL clocks such that the
D2048S frame always appears in the same position within each SHDSL Payload Sub-Block;
however, no particular alignment is specified. The temporal relationship between the D2048S data
stream and the data within the Sub-Blocks shall be maintained, such that that the order of bits in
time from the D2048S data stream shall match the order of transmission within the SHDSL Payload
Sub-Blocks. The optional M-pair mode will not support Unaligned D2048S transport.
134 ITU-T Rec. G.991.2 (12/2003)

E.7 TPS-TC for aligned European 2048 kbit/s digital structured leased line (D2048S) and
fractional
As noted in E.6, D2048S data streams consist of 2048 Mbit/s data streams containing 8 kHz
framing, with each frame containing 32 8-bit time slots. In some cases, Fractional D2048S data
streams are used, where frames contain less than the normal 32 8-bit time slots. Aligned D2048S
mode is also applicable to 2.048 Mbit/s PRI (Primary Rate ISDN), as described in 5.2/I.431 [B10].
In the aligned D2048S mode, each D2048S frame shall be aligned within the SHDSL Payload
Sub-Block such that the first time slot begins at the first bit position within the Payload Sub-Block,
followed by time slot 2, time slot 3, … , time slot n. A total of ks bits of contiguous data shall be
contained within each Sub-Block, as specified in 8.1, where ks = i + n × 8, and, in this mode, i = 0.
In D2048S applications, n = 32, and, in Fractional D2048S applications, 3 ≤ n < 32. The D2048S
framing clocks shall be synchronized to the SHDSL clocks such that the D2048S frame always
appears in the defined position within each SHDSL Payload Sub-Block. See Figure E.5 for
additional details.
...
Figure E.5/G.991.2 – Aligned D2048S/fractional D2048S framing
In the optional M-pair mode, D2048S/Fractional D2048S data will be carried over all M pairs using
interleaving, as described in 8.2. A total of M × ks bits of D2048S/Fractional D2048S data shall be
transported per SHDSL Payload Sub-Block. ks = i + n × 8, and, in this mode, i = 0. In D2048S
applications, n = 32/M, and in Fractional DS1 applications, 3 ≤ n < 32/M. In M-pair mode, only
multiples of M D2048S time slots may be supported. The time slots of the D2048S frame shall be
interleaved among all M wire pairs, such that pair M carries the mth time slot out of every block of
M slots. See Figure E.6 for additional details.
ITU-T Rec. G.991.2 (12/2003) 135

...
...
Figure E.6/G.991.2 – M-Pair framing for aligned D2048S/fractional D2048S

(for the M = 2 case)
E.8 TPS-TC for synchronous ISDN basic access

In this TPS-TC mode, the mapping of the ISDN customer data channels to SHDSL payload
channels is specified for synchronous transport of multiple ISDN BAs (Basic Access) using clock
mode 3a (see 10.1).
The ISDN customer data channels are embedded into the payload data within the SHDSL frames.
ISDN channels and SHDSL frames (and any other TPS-TC if Dual-Bearer mode is utilized –
see E.10) are synchronized to the same clock domain.
E.8.1 ISDN BA over SHDSL frames
Figure E.7 illustrates typical transport of ISDN BAs within the SHDSL frames. The basic
characteristics of this transport are as follows:
• B-channels and D-channels are mapped on SHDSL payload channels.
• The ISDN BA does not need a separate synchronization since the SHDSL frames are
synchronized to the same clock domain. Therefore, the ISDN frame word (12 kbit/s) is not
needed.
• The ISDN M-channel transports ISDN line status bits, transmission control information as
well as signalling to control the ISDN connection. Only the ISDN M-channel functions
which are needed to control the interface to the ISDN terminal equipment are transported
over a messaging channel (SHDSL EOC or fast signalling channel).
136 ITU-T Rec. G.991.2 (12/2003)

E.8.2 Mapping of ISDN B- and D-channels on SHDSL payload channels
The ISDN B- and D- channels are transported within the SHDSL payload sub-blocks. The SHDSL
payload data is structured within the SHDSL frames as follows:
• Each payload sub-block contains ks = i + n × 8 bits (i = 0..7 and n = 3..36, or, optionally,
n = 37…89, as described in Annex F).
• Each sub-block is ordered in the following way: i 1-bit time slots followed by n 8-bit time
slots.
• 1-bit time slots are referred to as Z-bits, and 8-bit time slots are referred to as TS1 ... TSn.
ISDN BA1 ISDN BAm
Figure E.7/G.991.2 – Mapping of ISDN B- and D-channels
The payload sub-blocks are composed of combinations of n × 8 bit-TS time slots and i × 1-bit
Z-time slots:
• n corresponds to the number of 64 kbit/s payload channels;
• i corresponds to the number of 8 kbit/s channels
This payload structure allows efficient mapping of ISDN BA channels on SHDSL frames.
• Data channels (64 kbit/s each, designated B1 – By) are mapped onto 64 kbit/s TS-channels.
• Signalling channels (16 kbit/s each, designated D1 – Dx) are mapped onto two 8 kbit/s
Z-channels each.3
____________________
3 If four or more ISDN BAs are transported, four D16 channels are mapped on one 64 kbit/s B-channel.
ITU-T Rec. G.991.2 (12/2003) 137

A general example of this mapping technique is shown in Figure E.7.
E.8.3 Multi-ISDN BAs
The transport of up to six ISDN BAs is described in detail in the next paragraphs. Figure E.8 shows
a mapping example for two ISDN BAs.
ISDN BA1 ISDN BAm
Figure E.8/G.991.2 – Framing example: 2 × ISDN BA
The transport of the customer data channels of each ISDN BA requires 144 kbit/s bandwidth.
Table E.1 shows the number of required TS- and Z-channels.
Table E.1/G.991.2 – K × ISDN BA

Number of TS-channels Z-channels
Payload bit rate
ISDN BA Application (64 kbit/s) (8 kbit/s)
K × (128 kbit/s + 16 kbit/s)
K n i
1 144 1 ISDN BA 2 2
2 288 2 ISDN BA 4 4
3 432 3 ISDN BA 6 6
4 576 4 ISDN BA 9 0
5 720 5 ISDN BA 11 2
6 864 6 ISDN BA 13 4
138 ITU-T Rec. G.991.2 (12/2003)

E.8.4 ISDN BA for lifeline service
Lifeline service in case of local power failure can be provided by one ISDN BA. The lifeline BA
always is that one which is transported over the first time slots of each payload sub-block (e.g., Z1,
Z2, TS1, TS2). Remote power feeding is provided by the central office such that the transceiver can
operate in a reduced power mode.
E.8.5 Time slot positions of ISDN B- and D16-channels (EOC signalling)
If multiple ISDN BAs are transported over SHDSL, certain data channels in the SHDSL payload
blocks must be assigned to each ISDN BA. Tables E.2 to E.5 show the allocation of the ISDN data
channels of up to 4 BAs. The signalling is transmitted over the SHDSL EOC.
In order to avoid unnecessary shifting of ISDN D- and B-bits, the respective D-bits are transmitted
after their B-bits in the subsequent SHDSL payload sub-block (B-bits in Nth payload sub-block and
D-bits in N + 1th payload sub-block; if the B-bits are transmitted in the last payload sub-block of an
SHDSL frame, the D-bits are transmitted in the first payload sub-block of the next SHDSL frame).
Table E.2/G.991.2 – Time slot allocation for 1 ISDN BA

ISDN BA number ISDN B1 time slot ISDN B2 time slot ISDN D16 time slots
1 TS1 TS2 Z1 + Z2
Table E.3/G.991.2 – Time slot allocation for 2 ISDN BAs

1 TS1 TS2 Z1 + Z2
2 TS3 TS4 Z3 + Z4

1 TS1 TS2 Z1 + Z2
2 TS3 TS4 Z3 + Z4
3 TS5 TS6 Z5 + Z6

1 TS2 TS3 TS1 (Bits 1 and 2)
ITU-T Rec. G.991.2 (12/2003) 139

E.8.5.1 Time slot positions of ISDN B- and D16-channels (EOC signalling) in M-pair mode
In the optional M-pair mode, the allocation of up to 3 ISDN BAs to time slots and Z-bits shall be as
shown in Tables E.2 to E.4. The allocation for 4 ISDN BAs is shown in Table E.5a.
Table E.5a/G.991.2 – Time slot allocation for 4 ISDN BAs

1 TS1 TS2 Z1 + Z2
2 TS3 TS4 Z3 + Z4
3 TS5 TS6 Z5 + Z6
4 TS7 TS8 Z7 + Z8
The Z-bits and time slots shall be interleaved among all M wire pairs. See Figure E.8a for additional
details.
Figure E.8a/G.991.2 – M-pair framing for ISDN BA (for the M = 2 case)
E.8.6 Time slot positions of ISDN B- and D16-channels and the optional fast signalling
channel
The optional 8 kbit/s fast signalling channel is always conveyed in Z1, as shown in Figure E.9. If
this fast signalling channel is used, up to 6 ISDN BAs can be transported over SHDSL.
140 ITU-T Rec. G.991.2 (12/2003)

In order to avoid unnecessary shifting of ISDN D- and B-bits, the respective D-bits are transmitted
after their B-bits in the subsequent SHDSL payload sub-block (B-bits in Nth payload sub-block and
D-bits in N + 1th payload sub-block; if the B-bits are transmitted in the last payload sub-block of an
SHDSL frame, the D-bits are transmitted in the first payload sub-block of the next SHDSL frame).
ISDN
fast signalling
channel
ISDN BA1 ISDN BAm
Figure E.9/G.991.2 – Mapping of ISDN B- and D-channels with a fast signalling channel
Table E.6/G.991.2 – Time slot allocation for 1 ISDN BA using the fast signalling channel
1 TS1 TS2 Z2 + Z3
Table E.7/G.991.2 – Time slot allocation for 2 ISDN BAs using the fast signalling channel
1 TS1 TS2 Z2 + Z3
2 TS3 TS4 Z4 + Z5
ITU-T Rec. G.991.2 (12/2003) 141

1 TS1 TS2 Z2 + Z3
2 TS3 TS4 Z4 + Z5
3 TS5 TS6 Z6 + Z7

using the fast signalling channel
1 TS2 TS3 Z2 + Z3

1 TS2 TS3 Z2 + Z3
2 TS4 TS5 Z4 + Z5
E.8.6.1 Time slot positions of ISDN B- and D16-channels (fast signalling) in M-pair mode
In the optional M-pair mode, the allocation of up to 3 ISDN BAs to Time Slots and Z-bits shall be
as shown in Tables E.6 to E.8. The allocation for 4 to 6 ISDN BAs is shown in Tables E.11a to
Table E.11c.
142 ITU-T Rec. G.991.2 (12/2003)

Table E.11a/G.991.2 – Time slot allocation for 4 ISDN BAs
1 TS1 TS2 Z2 + Z3
2 TS3 TS4 Z4 + Z5
3 TS5 TS6 Z6 + Z7
4 TS7 TS8 Z8 + Z9
Table E.11b/G.991.2 – Time slot allocation for 5 ISDN BAs

1 TS1 TS2 Z2 + Z3
2 TS3 TS4 Z4 + Z5
3 TS5 TS6 Z6 + Z7
4 TS7 TS8 Z8 + Z9
5 TS9 TS10 Z10 + Z11
Table E.11c/G.991.2 – Time slot allocation for 6 ISDN BAs

1 TS1 TS2 Z2 + Z3
2 TS3 TS4 Z4 + Z5
3 TS5 TS6 Z6 + Z7
4 TS7 TS8 Z8 + Z9
5 TS9 TS10 Z10 + Z11
6 TS11 TS12 Z12 + Z13
In fast signalling mode, the time slots and Z-bits frame shall be aligned within the SHDSL Payload
Sub-Block such that the Z1 fast signalling bit occupies the first bit position within the Payload Sub-
Block on each of the M pairs. The remaining Z-bits and time slots shall be interleaved alternating
among all M pairs. See Figure E.9a for additional details.
ITU-T Rec. G.991.2 (12/2003) 143

Figure E.9a/G.991.2 – M-pair framing for ISDN BA (for the M = 2 case)
E.8.7 Signalling over the SHDSL EOC or the fast signalling channel
The ISDN status signalling information can be optionally transmitted over two different channels:
• SHDSL EOC.
• Fast signalling channel.
In both cases, SHDSL EOC messages with their HDLC-like format are used to transport the ISDN
message code. The STU-C as well as the STU-R unit can initiate EOC messages. Generally, the
ISDN related EOC messages are transported over the SHDSL EOC. In some applications, it is
necessary to set up an additional fast signalling channel with 8 kbit/s bandwidth for these ISDN
related EOC messages. This is the case when more than four ISDN BAs are used. It may also be
used when low latency signalling is required or when another TPS-TC's signalling (e.g., ATM) has
substantially restricted the use of the SHDSL EOC channel.
E.8.7.1 SHDSL EOC messages
The EOC messages number 20 and 148 are used to transmit the ISDN maintenance and control
functions as well as the other ISDN EOC messages.
144 ITU-T Rec. G.991.2 (12/2003)

Table E.12/G.991.2 – ISDN Request – Message ID 20
2 bits 4-7 ISDN BA Number Unsigned char
2 bits 0-3 Unused Set to 00002
3 ISDN message code
Table E.13/G.991.2 – ISDN Response – Message ID 148

2 bits 4-7 ISDN BA Number Unsigned char
2 bits 0-3 Unused Set to 00002
3 ISDN message code
ISDN BA Number: Each ISDN BA can be addressed independently. To each ISDN BA, a
four-digit number is assigned (BA 1 = 0000, ... BA 6 = 0101).
E.8.7.2 ISDN message codes
The message codes which are contained as an octet in the SHDSL EOC message "ISDN Requests"
are listed in Table E.14. The message codes which are contained as an octet in the SHDSL EOC
message "ISDN Response" are listed in Table E.15.
Table E.14/G.991.2 – ISDN message codes commands

Function Message EOC message code Comment
SIA 0001 0000 S-interface activate
(STU-C → STU-R)
SID 0001 0001 S-interface deactivate
(STU-C → STU-R)
S-Bus Control
SAI 0001 0010 S-interface activated
(STU-R → STU-C)
SDI 0001 0011 S-interface deactivated
(STU-R → STU-C)
ACT 0000 0001 Readiness for layer 2
communication
(STU-C → STU-R)
ISDN Transceiver (STU-R → STU-C)
Status DEA 0000 0010 Intention to deactivate
(STU-C → STU-R)
CSO 0000 0011 Cold start only
(STU-R → STU-C)
S reset 0000 0000 Reset of ISDN control
BA Termination Reset unit at STU-R
(STU-C → STU-R)
ITU-T Rec. G.991.2 (12/2003) 145

Table E.14/G.991.2 – ISDN message codes commands
Operate 2B + D 0011 0001 (STU-C → STU-R)
loopback
Operate B1-channel 0011 0010 (STU-C → STU-R)
loopback (Note)
ISDN EOC Messages
Operate B2-channel 0011 0011 (STU-C → STU-R)
loopback (Note)
Return to normal 0011 1111 (STU-C → STU-R)
Hold state 0011 0000 (STU-C → STU-R)
NOTE – The use of B1- and B2-channel loopbacks is optional. However, the loopback codes are
reserved for these functions.
Table E.15/G.991.2 – ISDN Message codes responses

SIA 1001 0000 S-interface activated
SIAF 1101 0000 S-interface activation
failed
SID 1001 0001 S-interface deactivated
S-Bus Control
SIDF 1101 0001 S-interface deactivation
failed
SAI 1001 0010 S-interface activated
SDI 1001 0011 S-interface deactivated
ACT 1000 0001 Readiness for layer 2
ISDN Transceiver communication
Status DEA 1000 0010 Intention to deactivate
CSO 1000 0011 Cold start only
S reset ack 1000 0000 Reset of ISDN control
BA Termination Reset
unit at STU-R
146 ITU-T Rec. G.991.2 (12/2003)

Table E.15/G.991.2 – ISDN Message codes responses
Operate 2B + D 1011 0001 S-interface activate with
loopback (success) loop2
Operate 2B + D 1111 0001
loopback (failure)
Operate B1-channel 1011 0010 Operate B1-channel
loopback (success) loop can be requested
whenever the SHDSL
link is activated
Operate B1-channel 1111 0010
loopback (failure)
Operate B2-channel 1011 0011 Operate B2-channel
ISDN EOC Messages loopback (success) loop can be requested
whenever the SHDSL
link is activated
Operate B2-channel 1111 0011
loopback (failure)
Return to normal 1011 1111
(success)
Return to normal 1111 1111
(failure)
Hold state 1011 0000
Unable to comply 1111 0100
acknowledgement
E.8.8 S-bus control

The ISDN S-buses which connect the ISDN terminals with the STU-R can be controlled
independently with the respective message codes (SIA, SID, SAI, SDI) for each S-bus. The STU-C
side can activate and deactivate the S bus and gets status information. These messages are
transmitted as SHDSL EOC messages.
The S-interfaces of each ISDN BA can be addressed independently. To each ISDN BA a four-digit
number is (BA 1 = 0000, ... BA 6 = 0101) contained in the ISDN related SHDSL EOC messages.
SIA: In STU-C to STU-R direction, this function is used to request the STU-R to activate the
interface at the S reference point. If the interface at the S reference point is to be activated, this
message may be sent. In STU-R to STU-C direction the respective response is SIA (S-Interface
Activated).
SID: In STU-C to STU-R direction, this function is used to request the STU-R to deactivate the
interface at the S reference point. If the interface at the S reference point is to be deactivated, this
message may be sent. In STU-R to STU-C direction the respective response is SID (S-Interface
Deactivated).
SAI: In STU-R to STU-C direction, this message is used to inform the STU-C that the S-interface
and S-bus have been activated.
SDI: In STU-R to STU-C direction, this message is used to inform the STU-C that the S-interface
and S-bus have been deactivated.
ITU-T Rec. G.991.2 (12/2003) 147

E.8.9 BA termination reset
The status and condition of each ISDN BA and its S-interface at the STU-R side can be individually
monitored from the STU-C side. If a failure or blocking at one ISDN BA is detected, this situation
can be resolved by a reset. "BA termination reset" puts the control unit of the S-interface to its
default state (the deactivated state). Other BAs or other services are not affected.
LT (STU-C)
SIA (Request)
FE1
INFO2
SIA (Response)
INFO3
ACT (Request)
ACT (Response)
ACT (Request)
INFO4
FE4 ACT (Response)
Figure E.9b/G.991.2 – ISDN BA activation initiated by the exchange
LT (STU-C) TE
INFO1
SAI (Request)
FE2
INFO2
SAI (Response)
INFO3
ACT (Request)
ACT (Response)
ACT (Request)
INFO4
FE4 ACT (Response)
Figure E.9c/G.991.2 – ISDN BA activation initiated by the terminal equipment
LT (STU-C) TE
SID (Request)
FE5
INFO0
FE6
SID (Response)
INFO0
Figure E.9d/G.991.2 – ISDN BA activation initiated by the terminal equipment
148 ITU-T Rec. G.991.2 (12/2003)

Table E.16/G.991.2 – State transition table for the NT
State
NT1.1 NT1.2 NT1.3 NT1.4 NT1.5 NT1.5A NT1.6 NT1.7 NT1.8 NT1.9 NT2.0 NT2.0A NT2.1
number
ISDN service activation ISDN service activated Loopback 2
ISDN
State T interface LOS/LFA Loopback
Reset service T interface Active LOS/LFA Deactivation Loopback Loopback
name Initiated activated Active at T activate
deactivated activated pending at T initiated pending operated
ack pending ack
INFO sent
(CP-IWF INFO0 INFO0 INFO0 INFO2 INFO2 INFO2 INFO4 INFO2 INFO2 INFO0 INFO2 INFO2 INFO4
→ ΤΕ)
Internal
state G1 G1 G1 G2 G2 G2 G3 G2 G2 G4 G4 G4 G4
Event
Receiving NT1.7 SDI
– – – – – – – – NT1.2 – – –
INFO0 (Request)
Receiving NT1.3 SAI
– – – – – – – – – – – –
INFO1 (Request)
Receiving NT1.5 ACT NT1.5 ACT NT2.1 ACT
– – – – – – – – – –
INFO3 (Request) (Request) (Request)
LOS/LFA NT1.7 SDI
– – – – – – – – – – – –
at T (Request)
SIA NT1.4 SIA NT1.4 SIA
– – – – – – – – – –
(Request) (Response) (Response)
SAI
– – NT1.4 – – – – – – – – – –
(Response)
SID NT1.9 SID NT1.9 SID NT1.9 SID NT1.9 SID NT1.9 SID NT1.9 SID NT1.9 SID NT1.9 SID NT1.9 SID NT1.9 SID
– – –
(Request) (Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response)
ACT
– – – NT1.5 – – – – – – NT2.1 –
(Response)
ACT NT1.6 ACT
– – – – – – – – – – – –
(Request) (Response)
149 ITU-T Rec. G.991.2 (12/2003

Table E.16/G.991.2 – State transition table for the NT
State
NT1.1 NT1.2 NT1.3 NT1.4 NT1.5 NT1.5A NT1.6 NT1.7 NT1.8 NT1.9 NT2.0 NT2.0A NT2.1
number
ISDN service activation ISDN service activated Loopback 2
ISDN
State T interface LOS/LFA Loopback
Reset service T interface Active LOS/LFA Deactivation Loopback Loopback
name Initiated activated Active at T activate
deactivated activated pending at T initiated pending operated
ack pending ack
NT2.0
Operate
Operate 2B+D
2B+D
– loopback – – – – – – – – – – –
loopback
(success)
(Request)
(Response)
NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1
S reset
– S reset ack S reset ack S reset ack S reset ack S reset ack S reset ack S reset ack S reset ack S reset ack S reset ack S reset ack S reset ack
(Request)
(Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response) (Response)
SDI
– – – – – – – NT1.8 – –
(Response)
SHDSL:
Datar not – NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1 NT1.1
reached
SHDSL:
Datar NT1.2 – – – – – – – – – – –
reached
150 ITU-T Rec. G.991.2 (12/2003)

Table E.16a/G.991.2 – State transition table for the LT
State
LT1.1 LT1.2 LT1.3 LT1.4 LT1.5 LT1.6 LT1.7 LT1.8 LT2.0 LT2.1 LT2.2
number
ISDN ISDN service activation ISDN service activated Loopback 2

State
Reset service T interface Active LOS/LFA Deactivation Loopback Loopback Loopback
name Initiated Active
deactivated activated pending at T initiated requested pending operated
FE sent
(CO-IWF
→ ΕΤ) FE7 FE6 FE2 FE2 FE3 FE4 FE12 (Note) FE3 FE3 FE4
Event
LT1.3 SIA LT1.3 SIA
FE1 – – – – – – – –
(Request) (Request)
Start T2 Start T2 Start T2 Start T2 Start T2 Start T2 Start T2 Start T2
FE5 – – LT1.8 SID LT1.8 SID LT1.8 SID LT1.8 SID LT1.8 SID – LT1.8 SID LT1.8 SID LT1.8 SID
(Request) (Request) (Request) (Request) (Request) (Request) (Request) (Request)
LT2.0 LT2.0 LT2.0
Operate Operate Operate
FE8 – 2B+D 2B+D 2B+D
loopback loopback loopback
(Request) (Request) (Request)
S reset LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1
SAI LT1.4 SAI LT1.3 SAI
– – – – – – – – –
(Request) (Response) (Response)
LT1.5 ACT LT1.5 ACT
LT2.2
ACT (Response) (Response)
– – – – – – – ACT –
(Request) ACT ACT
(Response)
(Request) (Request)
SDI LT1.7 SDI
– – – – – – – – – –
(Request) (Response)
SIA
– LT1.4 – – – – –
(Response)
151 ITU-T Rec. G.991.2 (12/2003

Table E.16a/G.991.2 – State transition table for the LT
State
LT1.1 LT1.2 LT1.3 LT1.4 LT1.5 LT1.6 LT1.7 LT1.8 LT2.0 LT2.1 LT2.2
number
ISDN ISDN service activation ISDN service activated Loopback 2

State
Reset service T interface Active LOS/LFA Deactivation Loopback Loopback Loopback
name Initiated Active
deactivated activated pending at T initiated requested pending operated
SID
– – – – – – – LT1.2 – – –
(Response)
ACT
– – – – LT1.6 – – – –
(Response)
Operate
2B+D
loopback – – – – – – LT2.1 – –
(success)
(Response)
S reset ack
LT1.2 – – – – – – – – –
(Response)
SHDSL:
/ LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1 LT1.1
Datac failed
SHDSL:
Datac LT1.2 – – – – – – – – – –
reached
Expiry
– – – – – LT1.3 – – –
Timer 2
152 ITU-T Rec. G.991.2 (12/2003)

Table E.16b/G.991.2 – Legend for the State transition tables
Name Description
– No state change
/ Impossible by definition of peer-to-peer physical layer procedures or system internal
reasons.
| Impossible by definition of the physical layer service.
Start T2 Start Timer T2
A description of timer T2 can be found in Note 2 to Table 6 of ETSI ETS 300 012: "Timer
2 (T2) prevents unintentional reactivation. Ist value is 25 ms ≤ value ≤ 100 ms. This
implies that the TE has to recognize INFO0 and to react on it within 25 ms. If the NT is
able to unambiguously recognize INFO1, then the value of timer 2 may be 0, and an
MPH-DEACTIVATE REQUEST would cause a direct transition from state G2 or G3
to G1. It should be noted that the unambiguous detection of INFO1 may not be possible in
passive bus configurations, considering all possible implementations."
Note The FE sent to the network is identical to the FE sent prior to the issue of FE5 from the
network.
FE1 (LT ← ET) Activate access
FE2 (LT → ET) Access activation initiated
FE3 (LT → ET) Access digital section activated
FE4 (LT → ET) Access or loopback activated
FE5 (LT ← ET) Deactivate access
FE6 (LT → ET) Access deactivated
FE7 (LT → ET) LOS/LFA in DS or loss of power in NT1
FE8 (LT ← ET) Activate loopback 2
FE12 (LT → ET) LOS/LFA at T reference point
Table E.17/G.991.2 – Reset Request

Message EOC message code Comment
S reset 0000 0000
Table E.18/G.991.2 – Reset Response

Message EOC message code Comment
S reset acknowledge 1000 0000
E.8.10 Transport of ISDN EOC messages over SHDSL EOC

Table E.19 shows the six of the eight codes of the EOC functions which are defined in the ISDN
standard. (The two messages concerning the corrupted CRC are not required.)
ITU-T Rec. G.991.2 (12/2003) 153

Table E.19/G.991.2 – ISDN EOC message codes
Message Message code Network STU-R1 REG
Operate 2B + D 0011 0001 o d t/d
loopback
Operate B1-channel 0011 0010 o d t/d
loopback (Note)
Operate B2-channel 0011 0011 o d t/d
loopback (Note)
Return to normal 0011 1111 o d t/d
Hold state 0011 0000 d/o o/d o/d/t
Origin (o) & destination (d) & transfer (t)
NOTE – The use of B1- and B2-channel loopbacks is optional. However, the loopback codes are
reserved for these functions.
E.9 TPS-TC for ATM transport

E.9.1 Definitions
ATM Asynchronous Transfer Mode
HEC Header Error Check
E.9.2 Reference model for ATM transport
The ATM TC layer for SHDSL is consistent with ITU-T Rec. I.432.1 [8]. It shall provide the
following functions, as defined in ITU-T Rec. I.432.1:
• Rate decoupling between ATM layer and the synchronous (or plesiochronous) PMS-TC
layer.
• Insertion/Extraction4 of Idle cells.
• Insertion/Extraction5 of ATM Header Error Check (HEC) byte.
• Cell payload scrambling/descrambling for SDH-based systems.
• Cell delineation in the receive channel.
• Bit timing and ordering (MSB sent first with bit timing synchronous to the STU-C
downstream timing base).
The HEC covers the entire cell header. The code used for this function is capable of either:
• single bit error correction; or
• multiple bit error detection.
Error detection shall be implemented as defined in ITU-T Rec. I.432.1 [8] with the exception that
any HEC error shall be considered as a multiple bit error, and therefore, HEC error correction shall
not be performed.
Figure E.10 shows the logical interface between the ATM Layer, the ATM-TC and the SHDSL
PMS-TC function.
____________________
4 An idle cell inserted at the transmit side has to be extracted at the remote side.
5 A HEC byte inserted at the transmit side has to be extracted at the remote side.
154 ITU-T Rec. G.991.2 (12/2003)

Figure E.10/G.991.2 – ATM-TC logical interface to PMS-TC and TPS-TC ATM layer
An ATM Utopia level 2 interface connects the ATM-TC to the ATM Layer. This interface may also
be realized logically. Byte boundaries, at the ATM Utopia interface, shall be preserved in the
SHDSL payload. Bytes are transmitted MSB first, in accordance with ITU-T Rec. I.432.1 [8].
E.9.2.1 Framing
The PMS-TC provides a clear channel to the ATM-TC and cells are mapped into the SHDSL
payload on a byte-by-byte basis. At the STU-C, cells are mapped across the logical α interface
while at the STU-R, cells cross the logical β interface, as identified in 4.1. At the α and β interface,
logical data and clock lines are present. Cell alignment to the frame is optional. The ATM stream
shall be aligned within the SHDSL Payload Sub-Block such that the byte boundaries are preserved.
Each Payload Sub-Block is treated as containing n 8-bit time slots. Each byte from the input ATM
data stream is mapped MSB-first into the next available time slot. The first time slot begins at the
first bit position within the Payload Sub-Block, followed by time slot 2, time slot 3, … , time slot n.
A total of ks bits (or n bytes) of contiguous data shall be contained within each Sub-Block, as
specified in 8.1, where ks = i + n × 8, and, in this mode, i = 0 and 3 ≤ n ≤ 36. Note that optional
extensions described in Annex F allow values of n up to 89. See Figure E.10a for additional details.
ITU-T Rec. G.991.2 (12/2003) 155

...
Figure E.10a/G.991.2 – ATM framing
In the optional M-pair mode, ATM data is carried over all pairs using interleaving, as described
in 8.2. In M-pair mode only multiples of M time slots may be supported. The input ATM stream
shall be aligned within the SHDSL Payload Sub-Block such that the byte boundaries are preserved.
Each Payload Sub-Block is treated as containing M × n 8-bit time slots. Each byte from the input
ATM data stream is mapped MSB-first into the next available time slot. The first time slot begins at
the first bit position within the Payload Sub-Block, followed by time slot 2, time slot 3, … , time
slot n. A total of M × ks bits (or M × n bytes) of contiguous data shall be contained within each
Sub-Block, as specified in 8.1, where ks = i + n × 8, and, in this mode, i = 0 and 3 ≤ n ≤ 36. Note
that optional extensions described in Annex F allow values of n up to 89. The bytes from the input
ATM data stream shall be interleaved among all M pairs, such, where byte bm is carried on Pair m,
byte bm + 1 is carried in the corresponding time slot on Pair m + 1. Thus, pair M carries the mth time
slot out of every block of M time slots. See Figure E.10b for additional details.
156 ITU-T Rec. G.991.2 (12/2003)

...
...
Figure E.10b/G.991.2 – M-pair framing for ATM (for the M = 2 case)
E.9.2.2 Timing
STUs shall be operated in either synchronous or plesiochronous mode; however, in most
applications synchronous operation is preferred. In either case, the STU-C frame clock is locked to
network timing.
The provision of Network Timing Reference from the STU-C to the STU-R for ATM is optional;
however, if an NTR is provided, the SHDSL PMS-TC shall operate in clock synchronization
mode 3a (see 10.1). The network timing reference shall be an 8 kHz marker from which clocks at
other frequencies could easily be derived. In this clock mode, both the frame and symbol clocks at
the STU-C are locked to the NTR. The STU-R may extract the NTR from the received Frame
Synchronization Word (FSW). Referring to Figure E.10, the TxRef (in the STU-C) lines carries
NTR directly to the PMS-TC, while RxRef (in the STU-R) carries the NTR to the ATM Layer from
PMS-TC. Synchronization to the NTR shall be as described in 10.4.
E.9.2.3 IMA using the ATM TPS-TC (Informative)
The ATM TPS-TC, as defined in E.9, is intended to be compatible with Inverse Multiplexing for
ATM (IMA) Specification, as defined in af-phy-0086.001 [B12]. IMA is a protocol that provides
for inverse multiplexing of an ATM cell stream over multiple physical layer transmission links. It
operates by multiplexing the ATM cell stream between the links on a cell-by-cell basis and then
inserting special IMA Control Protocol (ICP) cells into each of the individual ATM cell streams.
Since the IMA cell stream for each link is structurally identical to a stream of normal ATM cells,
IMA cell streams may be carried without modification using the SHDSL ATM TPS-TC. Note that
the IMA Specification assumes that the ATM TPS-TC will be compatible with the IMA exceptions
to the Interface Specific Transmission Convergence Sublayer, as defined in the IMA Specification,
5.2.1 (specifically, items R-3 and R-4).
ITU-T Rec. G.991.2 (12/2003) 157

The IMA Specification (9.1) indicates that the differential delay from the IMA transmitter to the
loop interface (U-R or U-C) is to be no greater than 2.5 cells. Clause 7.1.6 recommends a maximum
differential signal transfer delay between non-repeatered SHDSL wire pairs of no more than 50 µs
at 150 kHz. With regard to repeaters, note that this Recommendation (9.5.5.5) allows up to
8 repeaters in an access link; however, it does not define the delay though the repeater. Also note
that the number of repeaters deployed in a loop is dependent on network-specific conditions.
Implementers are encouraged to take into account the various sources of differential delay,
including differential latencies introduced by repeaters (if present), in the design of IMA systems.
E.9.3 Transport capacity and flow control
An STU transporting ATM shall support N × 64 kbit/s data rates. The payload data rate shall be:
n × 64 + i × 8 kbit/s, where 3 ≤ n ≤ 36 and i = 0. This restriction applies to the data rate and payload
block size, as specified in 7.1.1, 8.1 and 8.2. Note that optional extensions described in Annex F
allow values of n up to 89.
In the optional M-pair mode, the rates specified shall apply per pair.
The ATM-TC shall provide flow control, allowing the STU-C and STU-R to control the cell flow
from the ATM layer. This functionality is implemented through the TX_Cell Handshake and
RX_Cell handshake at the ATM Utopia bus interface. A cell may be transferred to the ATM-TC
layer only after the completion of a TX_Cell Handshake. Similarly, a cell may be transferred from
the ATM-TC to the ATM Layer only after the STU has completed an RX_Cell_Handshake. This
functionality is important to avoid cell overflow and underflow at the TU layer.
E.9.4 Operations and maintenance
The ATM-TC requires Operations, Administration and Maintenance (OAM) functionality. The
messaging protocol and format should be handled in accordance with clause 9. The OAM functions
notify the OAM entity at the opposite end of the line upon the status of the cell delineation process
(e.g., Header Error Check (HEC) anomalies and Loss of Cell Delineation defects (LCD)).
Performance parameters are derived from anomalies and defects.
E.9.4.1 ATM data path related near-end anomalies
Near-end No Cell Delineation (nncd) anomaly: An nncd anomaly occurs immediately after ATM-
TC start-up, when ATM data is received and the cell delineation process is in HUNT or PRESYNC
state. Once cell delineation is acquired, subsequent losses of cell delineation shall be considered
nocd anomalies.
Near-end Out of Cell Delineation (nocd) anomaly: An nocd anomaly occurs when the cell
delineation process in operation transitions from the SYNC state to HUNT state. An nocd anomaly
terminates when the cell delineation process transition from PRESYNC to SYNC state or when nlcd
defect maintenance status is entered.
Near-end Header Error Control (nhec) anomaly: An nhec anomaly occurs when an ATM cell
header error control fails.
E.9.4.2 ATM data path related near-end defects
Near-end Loss of Cell Delineation (nlcd) defect: An nlcd defect occurs when at least one nocd is
present in 9 consecutive SHDSL frames and no losw defect (loss of synchronization word) is
detected. An nlcd defect terminates after the cell delineation process has entered and remained in
the SYNC state in 9 consecutive SHDSL frames.
158 ITU-T Rec. G.991.2 (12/2003)

E.9.4.3 ATM data path related far-end anomalies
Far-end No Cell Delineation (fncd) anomaly: An fncd anomaly is an nncd anomaly that is reported
from the far end by the NCD indicator in the EOC ATM Cell Status Information message. An fncd
anomaly occurs immediately after start-up and terminates if the received NCD indicator is coded 0.
Note that, since the far end reports the NCD indicator only on request, the fncd anomaly may be
inaccurate for derivation of the far-end NCD failure. Therefore, the NCD failure is autonomously
reported from the far end.
Far-end Out of Cell Delineation (focd) anomaly: A focd anomaly is a nocd anomaly, that is
reported from the far end by the OCD indicator in the EOC ATM Cell Status Information message.
The OCD indicator shall be coded 0 to indicate no nocd anomaly has occurred since last reporting
and shall be coded 1 to indicate that at least one nocd anomaly has occurred since last reporting. An
focd anomaly occurs if no fncd anomaly is present and a received OCD indicator is coded 1. An
focd anomaly terminates if a received OCD indicator is coded 0.
Far-end Header Error Control (fhec) anomaly: An fhec anomaly is an nhec anomaly, that is
reported from the far end by the HEC indicator in the EOC ATM Cell Status Information message.
The HEC indicator shall be coded 0 to indicate no nhec anomaly has occurred since last reporting
and shall be coded 1 to indicate that at least one nhec anomaly has occurred since last reporting. An
fhec anomaly occurs if a received HEC indicator is coded 1. An fhec anomaly terminates if a
received HEC indicator is coded 0.
E.9.4.4 ATM data path related far-end defects
Far-end Loss of Cell Delineation (flcd) defect: An flcd defect is an nlcd that is reported from the far
end of the line by the LCD indicator in the EOC ATM Cell Status Information message. The LCD
indicator shall be coded 0 to indicate no nlcd defect has occurred since last reporting and shall be
coded 1 to indicate that at least one nlcd defect has occurred since last reporting. An flcd defect
occurs when the LCD indicator is coded 1. An flcd defect terminates when the LCD indicator is
coded 0.
Note that, since the far end reports the LCD indicator only on request, the flcd defect may be
inaccurate for derivation of the far-end LCD failure. Therefore, the LCD failure is autonomously
reported from the far end.
E.9.4.5 ATM cell level protocol performance information collection
HEC violation count (hvc): An hvc performance parameter is the count of the number of nhec
anomalies modulo 65536.
HEC total count (htc): An htc performance parameter is the count of the total number of cells
passed through the cell delineation process, while operating in the SYNC state, since the last
reporting.
These values shall be counted, such that the Management system is able to retrieve current counts
on a 15-minute and 24-hour basis.
E.9.4.6 Failures and performance parameters
nncd failures and nlcd failures relate to persistent nncd anomalies and persistent nlcd defects,
respectively. The definitions below are derived from 7.1.2/G.997.1 [3]. These failures are reported
in the ATM Cell Status Information message.
E.9.4.6.1 ATM data path related near-end failures
The following near-end failure indications shall be provided by the STU-C and the STU-R:
ITU-T Rec. G.991.2 (12/2003) 159

E.9.4.6.1.1 Near-End No Cell Delineation (nncd) failure
An nncd failure is declared when an nncd anomaly persists for more than 2.5 ± 0.5 s after the start
of Data Mode. An nncd failure terminates when no nncd anomaly is present for more than
10 ± 0.5 s.
E.9.4.6.1.2 Near-End Loss of Cell Delineation (nlcd) failure
An nlcd failure is declared when an nlcd defect persists for more than 2.5 ± 0.5 s. An nlcd failure
terminates when no nlcd defect is present for more than 10 ± 0.5 s.
E.9.4.6.2 ATM data path related far-end failures
The following far-end failure indications shall be provided at the STU-C (the STU-R is at the far
end), and are optional at the STU-R (the STU-C is at the far end).
E.9.4.6.2.1 Far-End No Cell Delineation (fncd) failure
An fncd failure is declared when an fncd anomaly persists for more than 2.5 ± 0.5 s after the start of
Data Mode. An fncd failure terminates when no fncd anomaly is present for more than 10 ± 0.5 s.
E.9.4.6.2.2 Far-End Loss of Cell Delineation (flcd) failure
An flcd failure is declared when an flcd defect persists for more than 2.5 ± 0.5 s. An flcd failure
terminates when no flcd defect is present for more than 10 ± 0.5 s.
E.9.4.7 EOC ATM Cell Status Request message format – Message ID 17
The ATM Cell Status Request/Confirmation message is used for two purposes. This message is
used as ATM Cell Status Request message to get the STU-R ATM Status. For this purpose, the
whole information of EOC ATM Cell Status Information message – Message ID 145 shall be sent
in response to this message. If an unexpected receipt of ATM Cell Status message – Message ID
145 is received including NCD or LCD failure indication, this message may be used to confirm the
reception and stop future autonomous transmission of the ATM Cell Status message –
Message ID 145 due to the current failure condition.
Table E.20/G.991.2 – ATM Cell Status Request information field

E.9.4.8 EOC ATM Cell Status Information message format – Message ID 145
The ATM Cell Status Information message shall be sent in response to the ATM Cell Status
Request message and shall be sent autonomously upon the occurrence of an nlcd Failure or an nncd
Failure. Table E.21 shows the OAM message bit encoding for an ATM Cell Status Information
message. The HEC Indicator is implicitly defined as set to 1 if the HEC violation count has changed
since last reporting and set to 0 otherwise. If sent autonomously, Message ID 145 is sent once every
second until a Message ID 17 is received from the STU-C or the failure is cleared.
The NCD, OCD, and LCD Indicator bits shall indicate the state of nncd anomaly, nocd anomaly,
and nlcd defect, respectively. NCD Failure and LCD Failure bits shall serve as indications of nncd
failure and nlcd failure, respectively.
160 ITU-T Rec. G.991.2 (12/2003)

Table E.21/G.991.2 – ATM cell status information message
2, bit 7 NCD Indicator (Note) Bit 0 = OK, 1 = alarm
2, bit 6 OCD Indicator (Note) Bit 0 = OK, 1 = alarm
2, bit 5 LCD Indicator (Note) Bit 0 = OK, 1 = alarm
2, bits 4-2 Reserved
2, bit 1 NCD Failure Bit 0 = OK, 1 = alarm
2, bit 0 LCD Failure Bit 0 = OK, 1 = alarm
3 HEC violation count (hvc) MS Byte 16-bit counter, modulo 65536
4 HEC violation count (hvc) LS Byte 16-bit counter, modulo 65536
NOTE – Only one of the NCD, OCD, and LCD Indicators can be set to 1 at any time.
E.10 Dual-bearer TPS-TC mode

The TPS-TC modes in E.1 through E.9 and E.11 through E.13 are described as operating in Single-
Bearer Mode; i.e., the payload is treated as a single data stream, and the TPS-TC uses all of the bits
in each Payload Sub-Block. In some applications, however, it is desirable to split the payload into
separate data streams supporting multiple user interfaces or different data types. Dual-Bearer Mode
provides support for these cases.
Support for Dual-Bearer Mode is optional, as is support for each of the Dual-Bearer TPS-TC
combinations specified in Table E.22.
In Dual-Bearer Mode, each Payload Sub-Block is split between two separate TPS-TC instances.
The TPS-TC modes are negotiated independently in ITU-T Rec. G.994.1, and there is no direct
interaction between them. TPS-TCa is assigned the first ksa bits of each Payload Sub-Block, and
TPS-TCb is assigned the last ksb bits of each Payload Sub-Block (see Figure E.11). For each of the
two TPS-TCs, the ks bits assigned to it are treated as if they constituted a complete Payload
Sub-Block, and appropriate framing is applied, as described in E.1 to E.9 and E.11 to E.13
associated with the selected TPS-TC.
ITU-T Rec. G.991.2 (12/2003) 161

TPS-TCa TPS-TCb
Figure E.11/G.991.2 – Dual-bearer mode TPS-TC framing
Figure E.12 shows an example of a Dual-Bearer mode in which Fractional DS1 is TPS-TCa and
ATM is TPS-TCb.
TPS-TCa TPS-TCb
Payload data, bits 1a to ksa Payload data, bits 1b to ksb
Fa TS1a TS2a ... TSna TS1b TS2b ... TSnb
Figure E.12/G.991.2 – Example of dual-bearer mode TPS-TC framing
162 ITU-T Rec. G.991.2 (12/2003)

In the optional M-pair mode, the same procedure is followed for Dual-Bearer Mode. The first ksa
bits on each pair are assigned to TPS-TCa, and the last ksb bits on each pair are assigned to
TPS-TCb. The appropriate M-pair TPS-TC framing is then applied, as described in E.1 through E.9
and E.11 through E.13.
E.10.1 Dual-bearer clock synchronization
In Dual-Bearer Mode, it is assumed that timing for the two Bearer Channels is derived from a
common source and that the two data streams thus have a definite clocking relationship. As such, no
mechanism is provided within the payload blocks to maintain synchronization between the Bearer
Channels, regardless of the clock mode that is selected (10.1).
Note that some TPS-TCs have limitations on the clock modes that are supported. Specifically, ATM
using NTR (E.9.2) and Synchronous ISDN BA (E.8) are only defined for Clock Mode 3a (see 10.1).
When either of these TPS-TCs is used as part of a Dual-Bearer Mode, the system shall operate in
Clock Mode 3a.
E.10.2 Dual-bearer mode types
The following three types of dual-bearer modes are supported within SHDSL:
Type 1 – STM + Broadband.
Type 2 – STM + Cell/Packet.
Type 3 – STM + Clear Channel.
For each type of dual-bearer mode, separate specification bits are provided within ITU-T Rec.
G.994.1 for the selection of the two TPS-TCs to be used. Table E.22 lists the combinations that are
supported. Other supported types are for further study.
Table E.22/G.991.2 – Supported TPS-TCs in dual-bearer mode

Type Description TPS-TCa TPS-TCb
1 STM + Synchronous ISDN BA (E.8) Clear Channel (E.1)
Broadband LAPV5 Enveloped POTS or ISDN Clear Channel Byte-Oriented (E.2)
(E.13) (Note 2)
Unaligned DS1 (E.3) (Note 1)
STM with DSC (E.12)
Aligned DS1/Fractional DS1
(E.4) (Note 1)
Unaligned D2048U (E.5) (Note 2)
Unaligned D2048S (E.6) (Note 2)
Aligned D2048S/Fractional D2048S
(E.7) (Note 2)
ATM (E.9)
PTM (E.11)
2 STM + Unaligned DS1 (E.3) (Note 1) ATM (E.9)
Cell/Packet Aligned DS1/Fractional DS1 PTM (E.11)
(E.4) (Note 1)
(E.7) (Note 2)
ITU-T Rec. G.991.2 (12/2003) 163

Table E.22/G.991.2 – Supported TPS-TCs in dual-bearer mode
3 STM + Clear Unaligned DS1 (E.3) (Note 1) Clear Channel (E.1)
Channel Aligned DS1/Fractional DS1 (E.4) Clear Channel Byte-Oriented (E.2)
(Note 1)
(E.7) (Note 2)
NOTE 1 – Denotes TPS-TC modes that typically apply only in North American networks.
NOTE 2 – Denotes TPS-TC modes that typically apply only in European networks.
E.10.3 Dynamic rate repartitioning

Dynamic Rate Repartitioning (DRR) is the procedure for temporarily allocating time slots between
the STM Bearer and the Broadband Bearer. The DRR protocol is a master/slave protocol based on
messaging, at a rate of one message per superframe. Either the STU-C or the STU-R may be the
DRR master; this is configured through ITU-T Rec. G.994.1 [2] during pre-activation. The DRR
protocol will be triggered and controlled by a higher layer management entity, denoted in this
clause as a supervisory entity.
Figure E.13/G.991.2 – Dual-bearer mode framing with DRR
Figure E.13 shows an example of a dual-bearer mode with a dedicated DRR control channel, for
transport of the DRR protocol messages. These messages control the activation and de-activation of
time slots in the STM Bearer, and the corresponding de-allocation/allocation to the Broadband
Bearer. The Dedicated Signalling Channel (DSC) carries signalling information for telephony. Its
164 ITU-T Rec. G.991.2 (12/2003)

bandwidth depends on the application, and can be 0. This example shows 1 bit dedicated to DRR in
each Sub-Block, which corresponds to 8 kbit/s capacity. Adding more DRR bits increases the
capacity of the DRR control channel.
E.10.3.1 Message structure
The DRR message structure is shown in Figure E.14. These messages will be sent between the DRR
master and the DRR slave. The messages consist of one leading Control octet, followed by
Channel-ID octet(s). There is one Channel-ID octet for every 8 time slots to be managed by the
DRR procedure. The Control octet has 4 bits for the message type, followed by 4 bits for the
sequence number. Each bit of the Channel-ID octets corresponds to one time slot, the time slots
following, in the frame, the same order as the Channel-ID bits:
– "1": The corresponding time slot is currently active as part of the STM bearer channel, or is
in the process of being activated.
– "0": The corresponding time slot is not in use, and is thus available for broadband data.
Octet #1 Octet #2 Octet #3 Octet #4

(Control) (Channel ID) (Channel ID) (Channel ID)
b b b b b b b b b b b b b b b b b b b b b b b b b b b b bb b b
Message Sequence 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
type No. Time slots
Figure E.14/G.991.2 – DRR message structure

NOTE – This example assumes the SHDSL system is managing 24 time slots under DRR.
Each message has a sequence number that is used to control the DRR protocol. The exact usage is
given in the description of each state; however, in general it serves to indicate either how many
times a particular message has been sent in a sequence; or, in a responding message, to which
message number it is responding. Particularly, in an environment in which line disturbance can
cause protocol delays, the sequence number can be used to ensure synchronization of framing
change.
The complete set of DRR messages is shown in Table E.23.
Table E.23/G.991.2 – Messages used in DRR protocol

DRR message type Code Direction
MONITOR 1111 Master-to-Slave, Slave-to-Master
DEMAND 1110 Master-to-Slave
DEMAND ACK 1101 Slave-to-Master
DEMAND NAK 1011 Slave-to-Master
EXEC 0001 Master-to-Slave
EXEC ACK 0100 Slave-to-Master
REQUEST 1100 Slave-to-Master
ITU-T Rec. G.991.2 (12/2003) 165

E.10.3.2 Message flow for DRR
Figure E.15 shows a typical message flow for a DRR event.
Downstream Upstream
Message sent by Sequence Message received Message sent by Message received
Sequence No.
Master No. by Slave Slave by Master
Monitor <0> Monitor Monitor <0> Monitor
Demand <1> Demand Monitor <0> Monitor
Demand <2> Demand Demand Ack <1> Demand Ack
Demand <3> Demand Demand Ack <1> Demand Ack
Exec <1> Exec Demand Ack <1> Demand Ack
Exec <2> Exec Exec Ack <1> Exec Ack
Exec <3> Exec Exec Ack <2> Exec Ack
Monitor <0> Monitor Exec Ack <3> Exec Ack
NOTE – Shading change indicates change of framing.
Figure E.15/G.991.2 – Message flow, assuming STU-C is DRR Master, j = 2
E.10.3.3 Error protection

Each DRR message is stated 3 times within the same SHDSL superframe, and the correct message
is determined by a 2-out-of-3 majority decision at the recipient's end.
E.10.3.4 DRR control channel
The DRR messages are carried by a DRR control channel, a dedicated channel made up of one or
more Z-bits (8 kbit/s channel). Each Z-bit provides 48 bits (6 octets) per superframe. Since each
message is sent 3 times in the same superframe, each Z-bit provides for 2 octets of message. A
1 Z-bit channel can manage up to 8 time slots, while a 2 Z-bit channel, with 4 octets of message,
can manage up to 24 time slots. Messages sent from the DRR master to the DRR slave are referred
to as "downstream", and messages from the DRR slave to the DRR master are referred to as
"upstream". The number of Z-bits to be used must be configured during pre-activation through
ITU-T Rec. G.994.1 [2]. Channel-ID bits that are in excess of the number of managed time slots
will not be used.
E.10.3.5 Lead time
The lead time j used in the countdown is the number of downstream superframes starting with
EXEC <1> and ending just before the first downstream superframe with the new framing. This will
be the same as the number of upstream superframes starting with EXEC ACK <1> and ending just
before the first upstream superframe with the new framing. The value of j is to be negotiated during
pre-activation through ITU-T Rec. G.994.1 [2].
E.10.3.6 The DRR Protocol – Finite state machine description
The state diagrams for master and slave are given in Figures E.16 and E.17, respectively. The states
are shown as bubbles. The name of the state is given in the upper half of the bubble in italic font.
The message which is transmitted during the state is given in the lower half of the bubble in
CAPITAL letters. Incoming messages which trigger state transition are given in CAPITAL letters
as well. Information, commands and notifications to/from the supervisory entity are underlined.
Logical operation (i.e., and, or) are given in bold letters as well. These rules also apply to the
textual description. Notifications to/from the supervisory entity are primitives and are used for
illustrative purposes only. Supervisory actions are out of the scope of this Recommendation.
166 ITU-T Rec. G.991.2 (12/2003)

Reset
Idle
MONITOR
External DRR
Initiation or REQUEST
Initiation DEMAND NAK

DEMAND
DEMAND ACK
Go Ahead-1
EXEC
EXEC ACK
Go Ahead-2
EXEC
Receive Framer Ready
Figure E.16/G.991.2 – State diagram of the master, showing state,

outgoing message and trigger conditions
ITU-T Rec. G.991.2 (12/2003) 167

Reset
MONITOR Ready for new DRR

Not Ready Idle
DEMAND MONITOR Slave Request and

NAK Ready for new DRR
DEMAND and
Not Ready for new DRR
DEMAND and Slave Req

MONITOR Ready for
new DRR REQUEST
DEMAND
Confirmation
DEMAND
ACK
EXEC
Go Ahead
EXEC ACK
Framer Ready
MONITOR or DEMAND
Wait for Wait for

Monitor Framer
EXEC ACK EXEC ACK
MONITOR Framer Ready

or DEMAND
Figure E.17/G.991.2 – State diagram of the slave, showing state,

outgoing message and trigger conditions
168 ITU-T Rec. G.991.2 (12/2003)

E.10.3.7 DRR Master state machine
Table E.24/G.991.2 – Idle state of the Master

Entrance:
From state Trigger condition
Any Reset from supervisory entity
Go Ahead-2 Receiver Framer Ready from master
Initiation DEMAND NAK
Action:
Transmission of MONITOR <0>
Exit:
Trigger conditions Target state Notification
External DRR initiation, or Initiation
REQUEST
Fail-safe precaution: In the event of a mismatch in the time-slot settings in the Channel-ID octets of
the MONITOR upstream and downstream messages, the notification Time-Slot Alarm is issued.
Table E.25/G.991.2 – Initiation state of the Master

Entrance:
Idle External DRR initiation, or
REQUEST
+
Action:
Transmission of DEMAND nD begins with 1, and increments
<nD> until the first trigger condition.
Exit:
Trigger conditions Target state Notification
DEMAND ACK Go Ahead-1 Initiation of Transmit Framer
DEMAND NAK Idle Slave not ready for DRR
Fail-safe precaution: If nD reaches 15, it no longer increments. This could happen if recognition of
DEMAND ACK or DEMAND NAK is delayed, due to disturbance on the line. The notification
Sequence Number Overflow is issued, and the message DEMAND <15> continues to be
transmitted. The master stays in this state until a valid slave response is received, unless there is
supervisory intervention.
ITU-T Rec. G.991.2 (12/2003) 169

Table E.26/G.991.2 – Go Ahead-1 state of the Master
Entrance:
Initiation DEMAND ACK
Action:
Transmission of EXEC <nE> nE begins with 1, and increments
until the first trigger condition.
Exit:
Trigger condition Target state Notification
EXEC ACK Go Ahead-2 Initiation of Receiver Framer
Fail-safe precaution: If nE reaches 15, it no longer increments. This could happen if recognition of
the first EXEC ACK is delayed, due to disturbance on the line. The notification Sequence Number
Overflow is issued, and the message EXEC <15> continues to be transmitted. The master stays in
this state until a valid slave response is received, unless there is supervisory intervention.
Table E.27/G.991.2 – Go Ahead-2 state of the Master

Entrance:
Go Ahead State-1 EXEC ACK
Action:
Transmission of EXEC <nE> nE is fixed at the value it had when
exiting Go Ahead-1 State.
Exit:
Receive Framer Ready Idle DRR complete
E.10.3.8 DRR Slave state machine

An upper-layer supervisory entity also controls the DRR procedure at the DRR slave side. This
entity continually asserts a notification, stating whether the slave is ready to accept a new DRR or
not (Ready for new DRR, Not Ready for new DRR).
170 ITU-T Rec. G.991.2 (12/2003)

Table E.28/G.991.2 – Idle state of the Slave
Entrance:
Any Reset from supervisory entity
Confirmation MONITOR
Not Ready MONITOR
Wait for Monitor MONITOR, or DEMAND
Wait for Framer Framer Ready
Action:
Transmission of MONITOR <0>
Exit:
DEMAND and Ready for new DRR Confirmation
Slave Request and Ready for new Slave Request
DRR
DEMAND and Not Ready for new Not Ready
DRR
Table E.29/G.991.2 – Slave Request state of the Slave

Entrance:
Idle Slave Request and Ready for new
DRR
Action:
Transmission of REQUEST <nR> nR begins with 1, and increments
Exit:
DEMAND Confirmation
NOTE – In applications with tight timing requirements, it is recommended that the Slave Request
State not be used. Instead, the system should be configured with the Dedicated Signalling Channel
(DSC, see E.10.3) to allow normal telephony signalling to inform the master of the need for a DRR.
ITU-T Rec. G.991.2 (12/2003) 171

Table E.30/G.991.2 – Confirmation state of the Slave
Entrance:
Idle DEMAND <nD> and
Ready for new DRR
Slave Request DEMAND <nD>
Action:
Transmission of DEMAND nDA is fixed at the sequence
ACK <nDA> number nD of the triggering
DEMAND.
Exit:
Trigger condition Target state Notifications
EXEC Go Ahead Send both:
– Initiation of Receive and
Transmit Framer
– Sequence Number of First
Received EXEC (for
synchronization purposes)
Table E.31/G.991.2 – Not Ready state of the Slave

Entrance:
Idle DEMAND <nD> and Not Ready
for new DRR
Action:
Transmission of DEMAND nDN is fixed at the sequence number
NAK <nDN> nD of the triggering DEMAND
Exit:
MONITOR Idle DRR aborted
Table E.32/G.991.2 – Go Ahead state of the Slave

Entrance:
Confirmation EXEC
Action:
Transmission of EXEC ACK <nEA> nEA begins with 1, and increments
Exit:
Framer Ready Wait for Monitor
MONITOR, or DEMAND Wait for Framer
172 ITU-T Rec. G.991.2 (12/2003)

Fail-safe precaution: If nEA reaches 15, it no longer increments. This could happen if recognition of
the first MONITOR or DEMAND is delayed, due to disturbance on the line. The notification
Sequence Number Overflow is issued, and the message EXEC ACK <15> continues to be
transmitted. The slave stays in this state until a valid master message is received, unless there is
supervisory intervention.
Table E.33/G.991.2 – Wait for Monitor state of the Slave

Entrance:
Go Ahead Framer Ready
Action:
Transmission of EXEC ACK <nEA> nEA is fixed at the value it had
when exiting Go Ahead State.
Exit:
MONITOR, or DEMAND Idle DRR complete
Table E.34/G.991.2 – Wait for Framer state of the Slave

Entrance:
Go Ahead MONITOR, or DEMAND
Action:
Transmission of EXEC ACK nEA is fixed at the value it had when
<nEA> exiting Go Ahead State.
Exit:
Framer Ready Idle DRR complete
E.10.3.9 Result of DRR procedure

Figure E.11 shows the TPS-TC framing for the Dual-Bearer mode. Figure E.18 demonstrates how
the mapping of the payload sub-block will be changed by the DRR procedure, in a typical
application example. In the initial configuration of this example, the eight 8-bit time slots TSxa that
belong to TPC-TCa carry STM (voice) and the n 8-bit time slots TSxb that belong to TPS-TCb carry
ATM. When the supervisory entity recognizes that time slot TS2a is not currently carrying voice
samples, it instigates a DRR procedure which temporarily repartitions TS2a to the ATM bearer: then
seven time slots are carrying STM data, and (n + 1) are carrying ATM data.
Also shown is the DRR control channel and the Dedicated Signalling Channel (DSC) in
Figure E.18. In this example, the DRR control channel uses only 1 Z-bit, which is enough to
manage eight time slots (see E.10.3.4). The Dedicated Signalling Channel (DSC) carries the
higher-layer telephony signalling for the STM time slots (e.g., per Telcordia GR-303 [B14] or
ETSI V5 [9] & [B16]); in applications using channel associated signalling (CAS), and without tight
timing constraints, the DSC is optional.
ITU-T Rec. G.991.2 (12/2003) 173

TPS-TCa TPS-TCb
D D
R S TS1a TS2a TS3a ... TS8a TS1b TS2b .... Ts2n
R C
1 M 8 bits 8 bits 8 bits 8 bits 8 bits 8 bits 8 bits
bit bits
TPS-TCa TPS-TCb
D D
R S TS1a TS2a TS3a ... TS8a TS1b TS2b .... Ts2n
R C
1 M 8 bits 8 bits 8 bits 8 bits 8 bits 8 bits 8 bits
bit bits
After the DRR, time slot TS2a carries ATM data for TPS-TCb
Figure E.18/G.991.2 – DRR repartitions TS2a from STM bearer to ATM bearer (example)
E.10.3.10 Dual-bearer mode types for DRR

DRR is appropriate for use with some Dual-Bearer Type 1 and Type 2 TPS-TC combinations, as
specified in E.10.2 and Table E.22. In particular, DRR may be used with the set of TPS-TCa and
TPS-TCb combinations shown in Table E.34a.
Table E.34a/G.991.2 – TPS-TCs from dual-bearer mode Types 1 and 2

for which DRR is supported
1 STM + Synchronous ISDN BA (E.8) Clear Channel Byte-Oriented (E.2)
Broadband LAPV5 Enveloped POTS or ISDN Aligned DS1/Fractional DS1 (E.4)
(E.13) Aligned D2048S/Fractional D2048S
STM with DSC (E.12) (E.7)
ATM (E.9)
PTM (E.11)
2 STM + Aligned DS1/Fractional DS1 (E.4) ATM (E.9)
Cell/Packet Aligned D2048S/Fractional D2048S PTM (E.11)
(E.7)
NOTE – See Table E.22 for the complete definitions of TPS-TC Types for Dual-Bearer Mode.
E.10.3.11 Payload block ordering with DRR

The clauses describing each of the TPS-TCs define the arrangement of bits within each TPS-TC. As
noted in E.10.3.4, the DRR control channel occupies 1 to 3 single-bit time slots (referred to as
Z-bits). In addition, an ISDN or LAPV5 TPS-TC may use one or more Z-bits (E.8 and E.13), and a
DSC (Dedicated Signalling Channel), if used, may occupy 1 to 7 Z-bits or may be mapped into the
first 8-bit time slot (referred to as a B-channel). This clause defines how the different channels are
mapped into the TPS-TCs.
174 ITU-T Rec. G.991.2 (12/2003)

48*(1+i+n*8) bits
14 2 12 * (i+n*8) 10 12 * (i+n*8) 10 12 * (i+n*8) 10 12 * (i+n*8) 2 [bits]
S
P P P P P P P P P P p
Sync O O O O
0 0 0 ...... 1 1 ......... 2 2 ......... 3 3 ......... 4 a
word H H H H
1 2 3 2 3 4 5 6 7 8 r
e
6 ms
0 ms 6 ms
n * 8, n = 3....36
TPS-TCb channel [bits]
Z1 Z2 Z3 Z4 Z5 Z6 Z7 B1 Bn B (n+1) Bm B5 ...... ...... Bn
Signalling Channel(s) POTS ISDN BA Broadband

DRR, eoc, D16 channel(s)) (B-channel(s))
Z1 Z2 Z3 Z4 Z5 Z6 Z7 Z8 Z9 Z10 Z11 Z12 Z13 Z14
DRR Fast DSC ISDN D16

control EOC channel(s)
channel
Counting direction
Z1 Z2 Z3 Z4 Z5 Z6 B1
Required Z-bits are >7 result in combining the bits to one B-channel
Figure E.18a/G.991.2 – SHDSL payload block ordering with DRR
Figure E.18a shows how to combine the Z-bit time slots if their number exceeds 7. The formula is
based on the number of required Z-bits modulo 8.
ITU-T Rec. G.991.2 (12/2003) 175

E.11 TPS-TC for PTM transport
E.11.1 Packetized data transport
E.11.1.1 Functional model
The functional mode of packetized data transport is presented in Figure E.19. In the transmit
direction, the PTM entity obtains data packets to be transported over SHDSL from the application
layer interface. The PTM entity processes each packet and applies it to the γ-interface for
packetized data transport. The PTM TPS-TC receives the packet from γ-interface, encapsulates it
into a special frame (PTM-TC frame) and maps it into PMS-TC frame (transmission frame) for
transmission over the SHDSL link.
In the receive direction, the PTM-TC frame extracted from the received PMS-TC frame is directed
into the PTM-TC. The PTM-TC recovers the transported packet and delivers it to the PTM entity
via the γ-interface.
The PTM path-related OAM data, including information on errored packets, shall be presented to
the TPS-TC management entity providing all necessary OAM functions to support the PTM-TC.
Figure E.19/G.991.2 – Functional model of PTM transport
The γ-interface is described in E.11.3.1. The α/β-interfaces are application independent and thus
have the same format as for other TPS-TCs (see E.11.3.2).
E.11.2 Transport of PTM data
The bit rates of PTM data transport in the RX and TX direction on the SHDSL link are identical and
may be set to any eligible value which is less than (Dual-Bearer application) or equal to the
assigned maximum payload bit rate. This bit rate is set during the system configuration.
176 ITU-T Rec. G.991.2 (12/2003)

The PTM-TC shall provide full transparent data transfer between γSTU-C and γSTU-R interfaces
(except non-correctable errors in the PMD sublayer due to the noise in the loop). The PTM-TC shall
provide packet integrity over the assigned bearer channel.
E.11.3 Interface description
E.11.3.1 γ-Interface
The γC and γR reference points define the interfaces between the PTM entity and the PTM-TC at the
STU-C and STU-R, respectively, as shown in Figure E.19. These interfaces are functionally
identical and are independent of the contents of the transported packets. The interfaces are defined
by the following flows of signals between the PTM entity and the PTM-TC sublayer:
• data flow;
• synchronization flow;
• control flow;
• OAM flow.
E.11.3.1.1 Data flow
The data flow shall consist of two contra-directional octet-based streams of packets: transmit
packets (Tx_PTM) and receive packets (Rx_PTM). The packets transported in either direction over
the γ-interface may be of variable length. Bits within an octet are labelled a1 through a8, with a1
being the LSB and a8 being the MSB. If either of data streams is transmitted serially, the first octet
of the packet shall be transmitted first and bit a1 of each octet shall be transmitted first as shown in
Figure E.21. The Data Flow signal description is presented in Table E.35.
Table E.35/G.991.2 – PTM-TC: γ-interface data, synchronization and

control flows signal summary
Flow Signal Description Direction
Transmit signals
Data Tx_PTM Transmit Data PTM → PTM-TC
Control Tx_Enbl Asserted by PTM-TC, indicates that PTM may PTM ← PTM-TC
push packets to PTM-TC
Control Tx_Err Errored transmit packet (request to abort) PTM → PTM-TC
Sync Tx_Avbl Asserted by the PTM entity if data is available PTM → PTM-TC
for transmission
Sync Tx_Clk Clock signal asserted by the PTM entity PTM → PTM-TC
Sync Tx_SoP Start of the Transmit Packet PTM → PTM-TC
Sync Tx_EoP End of the Transmit Packet PTM → PTM-TC
Receive signals
Data Rx_PTM Receive Data PTM ← PTM-TC
Control Rx_Enbl Asserted by PTM-TC, indicates that PTM may PTM ← PTM-TC
pull packets from PTM-TC
Control Rx_Err Received error signals including FCS error, PTM ← PTM-TC
Invalid frame and OK
Sync Rx_Clk Clock Signal asserted by PTM entity PTM → PTM-TC
Sync Rx_SoP Start of the Receive Packet PTM ← PTM-TC
Sync Rx_EoP End of the Receive Packet PTM ← PTM-TC
ITU-T Rec. G.991.2 (12/2003) 177

E.11.3.1.2 Synchronization flow
This flow provides synchronization between the PTM entity and the PTM-TC sublayer and contains
the necessary timing to provide packet integrity during the transport. The synchronization flow shall
consist of the following signals as presented in Table E.35:
• Transmit and receive timing signals (Tx_Clk, Rx_Clk), both asserted by PTM entity.
• Start of packet signals (Tx_SoP, Rx_SoP): asserted by PTM entity and by PTM-TC
respectively and intended to identify the beginning of the transported packet in the
corresponding direction of transmission.
• End of packet signals (Tx_EoP, Rx_EoP), asserted by PTM entity and by PTM-TC
respectively and intended to identify the end of the transported packet in the corresponding
direction of transmission.
• Transmit Packet Available Signal (Tx_Avbl), asserted by PTM entity to indicate that data
for transmission in Tx direction is ready.
E.11.3.1.3 Control flow
Control signals are used to improve robustness of data transport between the PTM entity and the
PTM-TC and are presented in Table E.35:
• Enable Signals (Tx_Enbl, Rx_Enbl): asserted by PTM-TC and indicates that data may be
respectively sent from PTM entity to PTM-TC or pulled from PTM-TC to PTM entity.
• Transmit Error (Tx_Err): asserted by PTM entity and indicates that the packet or part of the
packet already transported from PTM entity to PTM-TC is errored or undesirable for
transmission (abort of transmitted packet).
• Receive Error (Rx_Err): asserted by PTM-TC to indicate that an errored packet is
transported from PTM-TC to PTM entity.
Handling of packet errors is described in E.11.4.2.
E.11.3.1.4 OAM flow
The OAM Flow across the γ-interface exchanges OAM information between the OAM entity and
its PTM related TPS-TC management functions. OAM flow is bidirectional.
The OAM flow primitives are for further study.
E.11.3.2 α/β interface
The α and β reference points define interfaces between the PTM-TC and PMS-TC at the STU-C
and STU-R respectively. Both interfaces are functional, application independent, and should
comply with the generic definition for all TPS-TCs as specified in clause 8.
E.11.4 PTM TPS-TC functionality
The following PTM TPS-TC functionality should be applied to both Rx and Tx direction.
E.11.4.1 Packet encapsulation
For packet encapsulation an HDLC type mechanism shall be used with detailed characteristics as
specified in the following clauses.
E.11.4.1.1 Frame structure
The PTM-TC frame format shall be as shown in Figure E.20. The opening and the closing Flag
Sequences shall be set to 7E16. They identify the start and the end of the frame. Only one Flag
Sequence is required between two consecutive frames.
178 ITU-T Rec. G.991.2 (12/2003)

8 7 6 5 4 3 2 1 Bits
7E16 Opening flag sequence
Default = FF16 Address field
Default = 0316 Control field
Data Information field
FCS-1 First octet of FCS
FCS-2 Second octet of FCS
7E16 Closing flag sequence
Figure E.20/G.991.2 – PTM-TC frame format
The Address and Control octets are intended for auxiliary information. They shall be set to their
default values of hexadecimal FF16 and 0316 respectively if not used.
NOTE 1 – The address and Control fields may be used for different auxiliary OAM functions. The usage of
these fields is for further study.
The information field shall be filled with the transported packet data. Prior to encapsulation the
octets of the data shall be numbered sequentially. Octets shall be transmitted in ascending numerical
order.
The frame check sequence (FCS) octets are used for packet level error monitoring, and shall be set
as described in E.11.4.1.3.
After encapsulation, bits within an octet are labelled b1 through b8, as defined in Figure E.21. If the
α(β) interface is serial by implementation, bit b8 of each octet shall be transmitted first.
NOTE 2 – In keeping with existing labelling convention for the α(β) interface, bit b8 (MSB) is transmitted
first. The PTM-TC functionality defines a correspondence between a1 and b8, a2 and b7, etc., in order to
conform to the HDLC convention of transmitting bit a1 first.
ITU-T Rec. G.991.2 (12/2003) 179

Figure E.21/G.991.2 – PTM-TC data flow
E.11.4.1.2 Octet transparency

To prevent failures due to false frame synchronization, any octet inside the PTM-TC frame that is
equal to 7E16 (the Flag Sequence) or 7D16 (the Control Escape) shall be escaped as described below.
After FCS computation, the transmitter examines the entire frame between the opening and the
closing Flag Sequences. Any data octets which are equal to Flag Sequence or the Control Escape
shall be replaced by a two-octet sequence consisting of the Control Escape octet followed by the
original octet exclusive-OR'ed with 2016. In summary, the following substitutions shall be made.
• Any data octet of 7E16 – encoded as two octets 7D16, 5E16.
• Any data octet of 7D16 – encoded as two octets 7D16, 5D16.
On reception, prior to FCS computation, each Control Escape octet shall be removed and the
following octet shall be exclusive OR'ed with 2016 (unless the following octet is 7E16 which is the
flag and indicates the end of the frame, and therefore an abort has occurred). In summary, the
following substitutions are made:
• any sequence of 7D16, 5E16 – replaced by the data octet 7E16.
• any sequence of 7D16, 5D16 – replaced by the data octet 7D16.
• a sequence of 7D16, 7E16 aborts the frame.
NOTE – Since octet stuffing is used, the PTM-TC frame is guaranteed to have an integer number of octets.
E.11.4.1.3 Frame check sequence
The FCS shall be calculated over all bits of the address, control, and information fields of the PTM-
TC frame as defined in ISO/IEC 13239 [B13], i.e., it shall be the one's complement of the sum
(modulo 2) of:
180 ITU-T Rec. G.991.2 (12/2003)

• the remainder of xk(x15 + x 14 + x13 + x12 + x11 + x10 + x9 + x8 + x7 + x6 + x5 + x4 + x3 +
x2 + x + 1) divided (modulo 2) by the generator polynomial x16 + x12 + x5 + 1, where k is the
number of bits in the frame existing between, but not including, the last bit of the opening
flag and the first bit of the FCS, excluding octets inserted for transparency (E.11.4.1.2); and
• the remainder of the division (modulo 2) by the generator polynomial x16 + x12 + x5 + 1, of
the product of x16 by the content of the frame existing between, but not including, the last
bit of the opening flag and the first bit of the FCS, excluding octets inserted for
transparency.
The FCS is 16 bits (2 octets) in length and occupies fields FCS-1, FCS-2 of the PTM-TC frame.
The FCS shall be mapped into the frame so that bit a1 (b8) of FCS-1 shall be the MSB of the
calculated FCS, and bit a8 (b1) of the FCS-2 shall be the LSB of the calculated FCS (Figure E.21).
The register used to calculate the FCS at the transmitter shall be initialized to the value FFFF16.
NOTE – As a typical implementation at the transmitter, the initial content of the register of the device
computing the remainder of the division is preset to all binary ONEs and is then modified by division by the
generator polynomial, as described above, on the information field. The one's complement of the resulting
remainder is transmitted as the 16-bit FCS.
As a typical implementation at the receiver, the initial content of the register of the device
computing the remainder of the division is preset to all binary ONEs. The final remainder, after
multiplication by x16 and then division (modulo 2) by the generator polynomial x16 + x12 + x5 + 1 of
the serial incoming protected bits after removal of the transparency octets and the FCS, will be
00011101000011112 (x15 through x0, respectively) in the absence of transmission errors.
E.11.4.2 Packet error monitoring
Packet error monitoring includes detection of invalid and errored frames at receive side.
E.11.4.2.1 Invalid frames
The following conditions result in an invalid frame:
• Frames which are less than 4 octets in between flags not including transparency octets (Flag
Sequence and Control Escape). These frames shall be discarded.
• Frames which contain a Control Escape octet followed immediately by a Flag (i.e., 7D16
followed by 7E16). These frames shall be passed across the γ-interface to the PTM entity.
• Frames which contain control escape sequences other than 7D16, 5E16 and 7D16, 5D16.
These frames shall be passed across the γ-interface to the PTM entity.
All invalid frames shall not be counted as FCS errors. The receiver shall immediately start looking
for the opening flag of a subsequent frame upon detection of an invalid frame. A corresponding
receive error message (Rx_Err – E.11.3.1.3) shall be sent across the γ-interface to the PTM entity.
E.11.4.2.2 Errored frames
A received frame shall be qualified as an errored frame (FCS-errored) if the CRC calculation result
for this frame is different from the one described in E.11.4.1.3. Errored frames shall be passed
across the γ-Interface. A corresponding receive error message (Rx_Err – E.11.3.1.3) shall be sent
across the γ-interface to the PTM entity.
E.11.4.3 Data rate decoupling
Data rate decoupling is accomplished by filling the time gaps between transmitted PTM-TC frames
with additional Flag Sequences (7E16). Additional Flag Sequences shall be inserted at the transmit
side between the closing Flag Sequence of the last transmitted PTM-TC frame and the subsequent
opening Flag Sequence of the next PTM-TC frame, and discarded at the receive side respectively.
ITU-T Rec. G.991.2 (12/2003) 181

E.11.4.4 Frame delineation
The PTM-TC frames should be delineated by detecting of Flag Sequence. The incoming stream is
examined on an octet-by-octet basis for the value 7E16. Two (or more) consecutive flag sequences
constitute an empty frame (frames), which shall be discarded, and not counted as a FCS error.
E.11.4.5 Mapping to the SHDSL framing
The PMS-TC provides a clear channel to the PTM-TC and packets are mapped into the SHDSL
payload on a byte-by-byte basis. At the STU-C, packets are mapped across the logical α interface
while at the STU-R, packets cross the logical β interface. At the alpha and beta interface, logical
data and clock lines are present. Packet alignment to the SHDSL frame is optional. The provided
bandwidth by the PMS-TC is ks = i + n × 8 with 0 ≤ i < 7 and 3 ≤ n ≤ 36. For n = 36, i is restricted
to values of 0 and 1. Note that optional extensions described in Annex F allow values of n up to 89.
In the optional M-pair mode, PTM data is carried over all pairs using interleaving, as described
in 8.2. In M-pair mode only multiples of M time slots may be supported. Each Payload Sub-Block is
treated as containing M × n 8-bit time slots. Each byte from the input PTM data stream is mapped
MSB-first into the next available time slot. The first time slot begins at the first bit position within
the Payload Sub-Block, followed by time slot 2, time slot 3, … , time slot n. A total of M × ks bits
(or M × n bytes) of contiguous data shall be contained within each Sub-Block, as specified in 8.1,
where ks = i + n × 8, and, in this mode, i = 0 and 3 ≤ n ≤ 36. Note that optional extensions described
in Annex F allow values of n up to 89. The bytes from the input PTM data stream shall be
interleaved among all M pairs, such, where byte bk is carried on Pair 1, byte bk+1 is carried in the
corresponding time slot on Pair 2, etc. Byte bk+M–1 is carried in the corresponding time slot on
Pair M.
E.12 TPS-TC for STM with a Dedicated Signalling Channel (DSC)

In certain STM applications, including some channelized voice and data applications, a dedicated
channel is desired to carry higher-layer telephony signalling for the STM time slots (e.g., per
Telcordia GR-303 [B14] or ETSI V5 [9] and [B16]). This TPS-TC defines a transport format for
channelized STM with a Dedicated Signalling Channel (DSC).
Figure E.22 shows the alignment of STM time slots and the DSC within the SHDSL frame. Each
Payload Sub-Block contains a DSC (i bits in length, where 1 ≤ i ≤ 7), followed by n 8-bit time slots
referred to as TS1 ... TSn. Note that the details of the protocols used over the DSC are beyond the
182 ITU-T Rec. G.991.2 (12/2003)

Figure E.22/G.991.2 – STM framing with a dedicated signalling channel
In the optional M-pair mode, both the STM and the DSC are carried over all M pairs using
interleaving, as described in 8.2. A total of M × n time slots shall be transported per SHDSL
Payload Sub-Block. The STM time slots shall be interleaved among all M wire pairs, such that pair
m carries the mth time slot out of every block of M time slots. The DSC is interleaved among the M
pairs such that it occupies the first i bit positions within each Payload Sub-Block on each of the M
wire pairs. i may take any value in the range 1 ≤ i ≤ 7, so a total of M × i bits make up the DSC.
i bits of contiguous DSC data shall be contained within a Sub-Block on Pair 1, and the following
sets of i bits of contiguous DSC data shall be contained within the corresponding Sub-Blocks of
subsequent pairs. See Figure E.23 for additional details.
ITU-T Rec. G.991.2 (12/2003) 183

Figure E.23/G.991.2 – M-pair STM framing with a dedicated signalling channel
(for the case where M = 2 and i = 1)
E.13 TPS-TC for LAPV5 enveloped POTS or ISDN

The mapping and time slot allocation of STM based, LAPV5 controlled, PSTN and ISDN-BA
transport is specified, which is for ISDN an alternative procedure to the simple use of D-channel
messages as described in E.8. It is not expected that the TPS-TC described in this clause will be
used simultaneously with the ISDN transport described in E.8 or the POTS transport described
in E.13.
This clause describes the transport of POTS and ISDN over a combination of the SHDSL EOC,
Z-channels, and B-channels. Control and signalling information is transported over either the EOC,
Z, or first B-channels using frame-based V5 wrappings. The POTS voice and ISDN B-channels are
transported over STM based pre-assigned SHDSL B-channels.
E.13.1 Signalling channel
Signalling as well as the other POTS or ISDN related messages are transported over a common
signalling channel. Depending on the required amount of signalling and port control information,
either a portion of the SHDSL EOC or a portion of the payload sub-block may be used for this
signalling transport. If the SHDSL EOC is used for the signalling transport, then the V5 signalling
messages are wrapped using SHDSL EOC Message IDs. If the SHDSL EOC is not used for this
transport, then within the SHDSL frame, the signalling bits are either mapped into 1 to 7
Z-channel(s), or are mapped into the first B-channel time slot of each sub-block.
184 ITU-T Rec. G.991.2 (12/2003)

In order to transport signalling information, the STU-C and STU-R must agree on the particular
signalling channel to be used. The signalling channel is identified using parameter (Nsig) with a
range of 0 to 8 plus the value 16. The value 0 indicates that the signalling is on the SHDSL EOC.
The values 1 through 7 indicate that there are 1 through 7 Z-channel bits present and that the
signalling is to be transported there. A value of 8/16 indicates that the signalling is transported in
the first one/two B-channel time slots of each sub-block. Other values of Nsig, such as 24 and 32, are
for future study.
E.13.2 Mapping of 64 kbit/s payload channels
One or multiple 64 kbit/s POTS voice channels and/or one or multiple ISDN B-channel pairs are
mapped onto B-channels in the SHDSL sub-frame. The POTS channels are mapped sequentially
into the first B-channels of each sub-frame after any signalling B-channels. The ISDN B-channel
pairs are mapped into the first B-channels of each sub-frame after any signalling or POTS
B-channels. These mappings are similar to those in E.8 and E.12.
In order to transport payload information, both the STU-C and STU-R have to agree as to how
many POTS and ISDN BA circuits to allocate B-channels for. The number of channels shall be the
same for both directions. The number of POTS circuits shall be specified as an integer (Npots) with a
range of 0 to 35. The number of ISDN circuits shall be specified as an integer (Nisdn) with a range of
0 to 17. (Other values are for future study.)
The total number of B-channels consumed for the control and payload transport is (1 or 2 if Nsig = 8
or 16, else 0) + Npots + (2 × Nisdn). The remaining B-channels are available for the underlying
application.
E.13.3 Signalling and port control
In the case where the common signalling channel is carried over the SHDSL EOC, (that is Nsig = 0),
the TPS-TC is addressed by the ISDN Message IDs within the EOC (IDs 20 and 148, see 9.5.5.6).
Octet 2 is not used and octets 3 through n contain the LAPV5 message. The message content is
enveloped by LAPV5-EF. Envelope functions and message contents are specified in ETSI EN 300
324-1 [9] and ETSI EG 201 900-1 [10]. See Tables E.36 and E.37 for details.
Table E.36/G.991.2 – ISDN Request – Message ID 20

1 20 Message ID
2 Not used
3 to n LAPV5 message code
Table E.37/G.991.2 – ISDN Response – Message ID 148

1 148 Message ID
2 Not used
3 to n LAPV5 message code
In ETSI EN 300 324-1 [9] clause 9.1.5, the maximum frame size is specified as 533 octets. In the
SHDSL EOC, the limit is 75 octets. Applications which require control and signalling frames larger
than 76 octets should choose Nsig > 0.
ITU-T Rec. G.991.2 (12/2003) 185

In the case where the common signalling channel is carried over the Z- or B-channel, (that is
Nsig > 0), the message format is as specified in ETSI EN 300 324-1 [9], clause 9. This mode shall
use all of clause 9, including subclauses for the flag sequence, interframe fill time, transparency,
frame check sequence, format conversion, and invalid frames which are not used in the EOC mode
above.
E.13.4 Protocol architecture for LAPV5 enveloped POTS and ISDN
Table E.38 shows the layered structure for LAPV5 enveloped POTS and ISDN services. Note that
the left lower column is for EOC signalling transport and the right lower column is for Z- or
B-channel signalling transport.
Table E.38/G.991.2 – Protocol architecture

POTS signalling POTS/ISDN port control ISDN signalling
ETSI EN 300 324-1 [9], ETSI EN 300 324-1 [9],
clause 13 clause 14
LAPV5-DL LAPD
ETSI EN 300 324-1 [9], clause 10
LAPV5-EF address LAPV5-EF

ETSI EN 300 324-1 [9], clause 9.1.4 ETSI EN 300 324-1 [9], clause 9
TPS-TC Signalling Z- or B-channels
Message ID: LAPV5 enveloped POTS and ISDN
PMD-TC, PMS-TC
NOTE – The ISDN signalling (LAPD and layer 3) is part of the ISDN-TE functionality and outside
the scope of this Recommendation.
The LAPV5-EF envelope address (ETSI EN 300 324-1 [9], clause 9) envelopes the frames for
signalling of an individual ISDN access, or for POTS signalling or for POTS/ISDN port control.
For the reliable transport of POTS signalling and POTS/ISDN port control messages, the data link
protocol LAPV5-DL is used which is a simplified version of LAPD. The LAPV5-DL protocol is
specified as in ETSI EN 300 324-1 [9], clause 10.
As in ETSI EG 201 900-1 [10] (Loop Emulation Service using AAL2), the following differences
with respect to ETSI EN 300 324-1 [9] exist:
• Only one common instance of LAPV5-DL is used for both the POTS signalling and the
POTS/ISDN port control.
• The LAPV5-DL address takes the value of all zeros.
• POTS signalling messages and POTS/ISDN port control messages are distinguished by
means of the Message type information element.
• A common error handling procedure for "unrecognized message type" errors is used for
both the PSTN and Control protocol: Whenever an unrecognized message is received, the
protocol entity shall generate an internal error indication and ignore the message.
• ISDN signalling is conveyed via frame relay as described in ETSI EN 300 324-1 [9],
clause 11. This means that the customer's D-channel data link layer protocol is not fully
terminated.
NOTE – The existing TPS-TC for ISDN as described in E.8 remains unchanged. It provides a lean
alternative for networks where no POTS, but only ISDN is provided.
186 ITU-T Rec. G.991.2 (12/2003)

E.13.5 System procedures
E.13.5.1 System start-up
With regard to the remainder of this subclause, actions required for any items that are not
provisioned shall be ignored.
NOTE – The procedures are derived from 5.4.4.1 and 5.4.4.2 of af-vmoa-0145.000 [B15].
E.13.5.1.1 Preconditions
The initial states of the various Finite State Machines (FSM) involved in the start-up are as follows:
Table E.39/G.991.2 – Initial states of finite state machines

FSM Initial state
Port Control Protocol FSM Out of Service (AN0/LE0)
PSTN Port Status FSM Blocked (AN1.0/LE1.0)
ISDN BA Port Status FSM Blocked (AN1.0/LE1.0)
PSTN Protocol FSM Port Blocked (AN6/LE6)
NOTE – These FSMs are defined in the V5 specifications ETSI EN 300 324-1 [9]. The "LE"
states relate to the STU-C side and the "AN" states relate to the STU-R side of the connection.
E.13.5.1.2 Normal procedure

a) Activation of LAPV5-DL: MDL-Establish-Request shall be sent to the LAPV5-DL.
b) When MDL-ESTABLISH-CONFIRM or MDL-ESTABLISH-INDICATION is received
from the LAPV5-DL, START-TRAFFIC shall be sent to the port control protocol FSMs.
c) Entering the normal state.
d) Post-processing: The STU-C side shall initiate the coordinated unblock procedure for all
relevant user ports. The STU-R side shall not initiate unblocking at this time.
E.13.5.1.3 Exceptional procedures in case of failure in system start-up
When the system start-up cannot be continued for some reason (e.g., LAPV5-DL failure) and is
unable to enter the normal state, system restart shall be performed.
E.13.5.2 System restart
System restart refers to the re-starting of a single LAPV5-DL protocol instance between a STU-C
side and a STU-R side. Under system restart the following actions apply:
1) The interface shall be brought into a state in which no established LAPV5-DL exists.
NOTE 1 – The remote side takes this as a trigger for system restart.
2) Timer TL1 shall be started.
3) On expiry of TL1 system start-up shall be performed.
Timer TL1 shall have a predefined value of 20 seconds.
NOTE 2 – Timer TL1 triggers system start-up. It is needed to guarantee that the release of the LAPV5-DL is
recognized at the remote side and hence both the STU-R side and STU-C side undergo system start-up. This
timer is started when the system has been stopped for any reason during the system start-up or normal
operation. It shall also be run prior to invoking the system start-up when performing a cold start.
Situations where system restart shall be applied:
a) Reception of Release-Indication of LAPV5-DL.
b) Under request by the Management System.
ITU-T Rec. G.991.2 (12/2003) 187

E.13.6 Nsig, Npots and Nisdn
In order to support interoperability, the STU-C and STU-R need to agree on the values of the
parameters Nsig, Npots and Nisdn. This agreement may be by prior agreement outside the scope of this
annex.
Alternatively, the STU-C may configure the STU-R via the SHDSL EOC. To support this, there is a
Message ID for LAPV5 POTS and ISDN Set-up. The purpose is to specify the values for Nsig, Npots
and Nisdn. The Message ID 21 is the Request from the STU-C to the STU-R and Message ID 149 is
the Response from the STU-R to the STU-C. See Tables E.40 and E.41 for details.
The request message allows the STU-C to configure the STU-R with the values of Nsig, Npots and
Nisdn. The response message is an acknowledge from the STU-R to STU-C. If octets 2, 3, and 4 of
the response match those in the request, then the response indicates that the STU-R accepts the
values sent by the STU-C. If the STU-R does not accept the values proposed by the STU-C, it may
respond with the octets 2, 3, and/or 4 modified to contain an acceptable value and also the MSB of
each octet in question set. The STU-R should respond to a request within 500 ms. In the event that
the STU-R does not respond, the STU-C will try at least three times before concluding that the
option cannot be supported.
Table E.40/G.991.2 – LAPV5 POTS and ISDN Set-up Request – Message ID 21

Octet # Contents Data type Note
1 21 Message ID
2 Nsig Unsigned char
3 Npots Unsigned char
4 Nisdn Unsigned char
Table E.41/G.991.2 – LAPV5 POTS and ISDN Set-up Response – Message ID 149
Octet # Contents Data type Note
1 149 Message ID
2 Nsig Unsigned char
3 Npots Unsigned char
4 Nisdn Unsigned char
Annex F
Region 1 requirements for payload data rates up to 5696 kbit/s

F.1 Scope
The clauses in this annex provide the additions and modifications to the corresponding clauses in
the main body and Annex A for payload data rates up to 5696 kbit/s. Support for this annex is
optional.
NOTE – Some countries have standards for spectrum management requirements that limit the length of the
lines for transmission of certain signal levels in this annex, for example Spectrum Management Standard
T1.417 applies in the United States access network.
188 ITU-T Rec. G.991.2 (12/2003)

F.2 Data rate
The operation of the STU in data mode at the specified information rate shall be as specified in
Table F.1.
Table F.1/G.991.2 – Framed data mode rates

Payload data rate, R Symbol rate K
Modulation
(kbit/s) (ksymbol/s) (Bits per symbol)
R = n × 64 + (i) × 8 16-TCPAM (R + 8) ÷ 3 3
R = n × 64 + (i) × 8 32-TCPAM (R + 8) ÷ 4 4
As specified in the main body (per clause 5, reiterated in 7.1.1, 8.1 and 8.2), the allowed single-pair
rates are given by n × 64 + i × 8 kbit/s, where 3 ≤ n ≤ 36 and 0 ≤ i ≤ 7. In these clauses, the allowed
values i are further restricted to 0 or 1 for n = 36. These definitions correspond to (payload) data
rates from 192 kbit/s to 2.312 Mbit/s in increments of 8 kbit/s.
This annex extends those rates. It is applicable for single-pair rates given by n × 64 + i × 8 kbit/s.
For 16-TCPAM, 36 ≤ n ≤ 60 and 0 ≤ i ≤ 7. For 16-TCPAM and n = 36, the applicable values of i
are 2 ≤ i ≤ 7. For 16-TCPAM and n = 60, the applicable value of i is 0. This corresponds to
(payload) data rates from 2320 kbit/s to 3840 kbit/s in increments of 8 kbit/s for 16-TCPAM. For
32-TCPAM, 12 ≤ n ≤ 89 and 0 ≤ i ≤ 7. For 32-TCPAM and n = 89, the applicable value of i is 0.
This corresponds to (payload) data rates from 768 kbit/s to 5696 kbit/s in increments of 8 kbit/s for
32-TCPAM.
This annex is also applicable for optional operation on more than one pair (M-pair mode).
F.2.1 Support for multiple encodings
Support for the data rates specified in this annex is optional, and, as such, an STU supporting this
annex is not required to support all specified data rates. For each rate that an STU-R supports, it
shall support all available encodings (i.e., both 16- and 32-TCPAM for rates where both encodings
are specified). Support for multiple encodings is optional at the STU-C.
F.2.2 G.994.1 Pre-activation sequence
As specified in 6.4, ITU-T Rec. G.994.1 is used to begin the pre-activation sequence.
To support a wide range of data rates and multiple encodings, this clause introduces a new way to
encode data rates in G.994.1 code points. This method of encoding rates is used for both the PMMS
rates and the training rates. Data rates are encoded as a set of ranges, where each range is expressed
as a 3-tuple (minimum, maximum, step). The 3-tuple represents all rates of the form
(m + k × s) × (64 kbit/s) where m is the minimum value, s is the step value, and k is the set of all
integers greater than or equal to zero such that m + k × s is less than or equal to the maximum value.
Thus, for example, the 3-tuple (40, 70, 10) represents the rates 40 × 64 kbit/s, 50 × 64 kbit/s,
60 × 64 kbit/s, and 70 × 64 kbit/s.
ITU-T Rec. G.991.2 (12/2003) 189

Each data rate parameter in this annex can be expressed as a set of between 1 to 8 ranges, where the
supported rates are the union of those supported by the individual ranges. Thus, for example, the
3-tuples (20, 30, 4), (40, 70, 10) represent the rates 20 × 64 kbit/s, 24 × 64 kbit/s, 28 × 64 kbit/s,
40 × 64 kbit/s, 50 × 64 kbit/s, 60 × 64 kbit/s, and 70 × 64 kbit/s. If all bits of the extended base data
rate minimum and maximum are set to zero, then those rates are not supported for line probe. If
only one range of rates is required, then only the octets associated with (min1, max1, step1) shall be
sent.
Also, in many cases, the values in the data range 3-tuple can be less than or equal to 89
(representing the maximum payload data rate of 5696 supported in this annex). When using G.994.1
code point representation, only 6 bits are available for the value of an NPar(3). To support numbers
greater than 63, the value must be split across multiple octets. When encoding a data range using
G.994.1, 4 octets are used, where the first octet contains the highest order bit from each of the
values in the 3-tuple. This is illustrated in Table 11.16.10/G.994.1.
The complete set of rate capabilities shall be the union of the extended rates specified in Annex F
(G.994.1 Table 11.16.0.1 bits 4-6 and Table 11.16.0.2 bits 1-3) with the non-extended rates
specified in Annex A (G.994.1 Table 11.16 bits 1-4).
Ranges of rates may overlap and may contain some rates which are identical. For example, the
3-tuples (40, 60, 10) and (50, 70, 5) would be a valid set of ranges. In this case, the union of these
two 3-tuples would be the rates 40 × 64 kbit/s, 50 × 64 kbit/s, 55 × 64 kbit/s, 60 × 64 kbit/s,
65 × 64 kbit/s, and 70×64 kbit/s. Note that, for PMMS, if two ranges contain some rates which are
identical, the probe waveforms associated with these identical rates are only sent once.
The following definition is added to the G.994.1 code point definitions in 6.4.1 for the support of
the extended data rates specified in this annex.
Extended Base Data Rate: These octets are used to specify payload rates for this annex, as
follows:
– The PMMS octets indicate rates for line probing segments. Note that while PMMS uses
2-PAM modulation, the PMMS symbol rates are specified assuming 32 TCPAM encoding,
so the PMMS symbol rate (in ksymbol/s) would be equal to the (payload data
rate (kbit/s) + 8 kbit/s)/4. If both symmetric and asymmetric PSDs are indicated, then all of
the indicated symmetric PSDs shall be sent first, followed by all of the indicated
asymmetric PSDs. Valid values for min and max shall be between 49 and 89, inclusive, and
valid values for step shall be between 1 and 40, inclusive. The variables j5 and j6 associated
with the PMMS rates shall be independent, and shall range from 1 to 8, inclusive. If only
one range of rates is required, then only the octets associated with (min1, max1, step1) shall
be sent.
– The training parameter octets indicate extended payload data rates supported.
– In CLR, upstream training parameters indicate which data mode rates the STU-R is capable
of transmitting and downstream training parameters indicate which data mode rates the
STU-R is capable of receiving. If the optional line probe is used, the receiver training
parameters will be further limited by the probe results. Valid values for minimum and
maximum shall be between 36 and 60, inclusive, for 16-TCPAM and between 12 and 89,
inclusive, for 32-TCPAM. Valid values for step shall be between 1 and 89, inclusive. The
variables j1, j2, j3 and j4 associated with the training rates shall be independent, and shall
range from 1 to 8, inclusive. The STU-R shall indicate support for both 16- and 32-TCPAM
for all supported rates for which both encodings are defined in this annex.
190 ITU-T Rec. G.991.2 (12/2003)

– In CL, downstream training parameters indicate which data mode rates the STU-C is
capable of transmitting and upstream training parameters indicate which data mode rates
the STU-C is capable of receiving. Valid values for minimum and maximum shall be
between 36 and 60, inclusive, for 16-TCPAM and between 12 and 89, inclusive, for
32-TCPAM. Valid values for step shall be between 1 and 89, inclusive. The variables j1, j2,
j3 and j4 associated with the training rates shall be independent, and shall range from 1 to 8,
inclusive. If optional line probe is used, the receiver training parameters will be further
limited by the probe results.
– Data rate selections shall be specified in MP and MS messages by setting the maximum and
minimum rates to the same value.
F.3 Mapper
The K + 1 bits YK(m), … , Y1(m), and Y0(m) shall be mapped to a level x(m). In 6.1.2.3, the mapper
function is specified for 16-TCPAM. This annex extends that mapping to include both 16- and
32-TCPAM encodings. Table F.2 shows the bit to level mapping for 16- and 32-level mapping.
Table F.2/G.991.2 – Mapping of bits to PAM levels

Y4(m) Y3(m) Y2(m) Y1(m) Y0(m) 32-PAM (5 bits) 16-PAM (4 bits)
0 0 0 0 0 –31/32 –15/16
0 0 0 0 1 –29/32 –13/16
0 0 0 1 0 –27/32 –11/16
0 0 0 1 1 –25/32 –9/16
0 0 1 0 0 –23/32 –7/16
0 0 1 0 1 –21/32 –5/16
0 0 1 1 0 –19/32 –3/16
0 0 1 1 1 –17/32 –1/16
0 1 1 0 0 –15/32 1/16
0 1 1 0 1 –13/32 3/16
0 1 1 1 0 –11/32 5/16
0 1 1 1 1 –9/32 7/16
0 1 0 0 0 –7/32 9/16
0 1 0 0 1 –5/32 11/16
0 1 0 1 0 –3/32 13/16
0 1 0 1 1 –1/32 15/16
1 1 0 0 0 1/32 –
1 1 0 0 1 3/32 –
1 1 0 1 0 5/32 –
1 1 0 1 1 7/32 –
1 1 1 0 0 9/32 –
1 1 1 0 1 11/32 –
1 1 1 1 0 13/32 –
1 1 1 1 1 15/32 –
1 0 1 0 0 17/32 –
1 0 1 0 1 19/32 –
ITU-T Rec. G.991.2 (12/2003) 191

Table F.2/G.991.2 – Mapping of bits to PAM levels
Y4(m) Y3(m) Y2(m) Y1(m) Y0(m) 32-PAM (5 bits) 16-PAM (4 bits)
1 0 1 1 0 21/32 –
1 0 1 1 1 23/32 –
1 0 0 0 0 25/32 –
1 0 0 0 1 27/32 –
1 0 0 1 0 29/32 –
1 0 0 1 1 31/32 –
F.4 PSD masks

For symmetric PSDs using 16-TCPAM payload data rates greater than or equal to 2320 kbit/s, and
for symmetric PSDs using 32-TCPAM payload data rates greater than or equal to 768 kbit/s, the
measured transmit PSD of each STU shall not exceed the PSD masks specified in this clause
(PSDMASKSHDSL(f)), and the measured total power into 135 Ω shall fall within the range specified
in this clause (PSHDSL ± 0.5 dB).
PSDMASK SHDSL ( f ) =
   πf 
2
 sin  
 − PBO K 
  sym 
Nf MaskedOffsetdB ( f )
10 SHDSL 1 1 10
10 × × × 2
× 2 × Order
×10 W/Hz, f < f int
 135 f sym  πf   f 
   1 +  
  Nf sym   f 3dB 
  
− 90 dBm/Hz peak, with max power in the [ f , f + 1 MHz] window of
 −4 −1.5
[10 log10 (0.5683 ×10 × f ) + 90] dBm , f int ≤ f ≤ 3.184 MHz
− 90 dBm/Hz peak, with max power in the [ f , f + 1 MHz] window of

− 50 dBm , 3.184 MHz ≤ f ≤ 12 MHz
 f 3 dB − f
1 + 0.4 × , f < f 3 dB
1 dB , f ≥ f 3 dB

fint is the frequency where the two functions governing PSDMASKSHDSL(f) intersect in the range 0 to
fsym. PBO is the power backoff value in dB. KSHDSL, Order, N, fsym, f3dB, and PSHDSL are defined in
Table F.3. PSHDSL is the range of power in the transmit PSD with 0 dB power backoff. R is the
payload bit rate. The variables f, fsym, fint, and f3dB in the equations are in units of Hz.
192 ITU-T Rec. G.991.2 (12/2003)

Table F.3/G.991.2 (Part 1) – Symmetric PSD parameters, 16-TCPAM
Payload bit rate, fsym
R (kbit/s) (ksymbol/s)
2320 ≤ R ≤ 3840 7.86 6 1 (R + 8)/3 1.0 × fsym/2 13.5
Table F.3/G.991.2 (Part 2) – Symmetric PSD parameters, 32-TCPAM

Payload bit rate, fsym
R (kbit/s) (ksymbol/s)
768 ≤ R ≤ 5696 7.86 6 1 (R + 8)/4 1.0 × fsym/2 13.5
Figure F.1 shows the PSD masks with 0 dB power backoff for payload data rates of 3840
(16-TCPAM) and 5696 (32-TCPAM) kbit/s.
Figure F.1/G.991.2 – PSD masks for 0 dB power backoff
   πf 
2
 sin  
 − PBO K  Nf sym 
1    1 f2
10 10 × SHDSL × × 2
× 2 × Order
× 2 2
W/Hz , f < f int
 135 f sym    f  f + f
 πf 
NominalPSD( f ) c
1 +  
  Nf sym   f 3dB 
  
−4 −1.5
0.5683 ×10 × f W/Hz , f int ≤ f ≤ 3.184 MHz
− 100 dBm/Hz , 3.184 MHz ≤ f ≤ 12 MHz

ITU-T Rec. G.991.2 (12/2003) 193

where fc is the transformer cut-off frequency, assumed to be 5 kHz. The variables f, fsym, fint, and f3dB
in the equations are in units of Hz. Figure F.2 shows the nominal transmit PSDs with 13.5 dBm
power for payload data rates of 3840 (16-TCPAM) and 5696 (32-TCPAM) kbit/s.
NOTE 2 – The nominal PSD is intended to be informative in nature; however, it is used for purposes of
crosstalk calculations as representative of typical implementations.
Figure F.2/G.991.2 – Nominal PSDs for 0 dB power backoff
F.5 Crosstalk interference requirements

Table F.4 shows the minimum set of test loops and crosstalk combinations required for testing
SHDSL margins. A compliant unit shall pass the BER test described in A.3.1 for all crosstalk
scenarios and test loops defined in Table F.4 for all supported data rates and modulation type (e.g.,
16-TCPAM or 32-TCPAM). 0 dB Power Backoff shall be used for both the STU-C and STU-R.
The calibration procedure and testing methods used shall be identical to those used for Annex A.
The test loops and disturbers are identical to the corresponding cases in Annex A.
Table F.4/G.991.2 – Crosstalk scenarios and required SHDSL noise margins (Note)
Test
Payload
loop Required
Test (from Modulation PSD margin
(× 1000') unit rate combination
Figure (dB)
(kbit/s)
A.1)
1 S 4.5 STU-C 3840 16-TCPAM Symmetric 24 HDSL2 + 24 T1 5 + ∆*
(Case 4)
2 S 4.5 STU-R 3840 16-TCPAM Symmetric 24 HDSL2 + 24 T1 5 + ∆*
(Case 14)
(Case 4)
(Case 14)
5 S 5.7 STU-C 2560 16-TCPAM Symmetric 49 SHDSL sym 2304 5 + ∆*
( )
194 ITU-T Rec. G.991.2 (12/2003)

Table F.4/G.991.2 – Crosstalk scenarios and required SHDSL noise margins (Note)
Test
Payload
loop Required
Test (from Modulation PSD margin
(× 1000') unit rate combination
Figure (dB)
(kbit/s)
A.1)
(Case 11)
6 S 5.7 STU-R 2560 16-TCPAM Symmetric 49 SHDSL sym 2304 5 + ∆*
(Case 11)
(Case 4)
(Case 14)
(Case 4)
(Case 14)
(Case 11)
(Case 11)
(Case 16)
(Case 16)
15 S 2.3 STU-C 5696 32-TCPAM Symmetric 24 FDD ADSL + 5 + ∆*
24 HDSL (Case 6)
16 BT1-C 1.9 STU-C 5696 32-TCPAM Symmetric 24 HDSL2 + 24 T1 5 + ∆*
(Case 4)
17 BT1-R 1.9 STU-R 5696 32-TCPAM Symmetric 24 HDSL2 + 24 T1 5 + ∆*
(Case 14)
18 BT2-C 3.9 STU-C 2560 32-TCPAM Symmetric 49 HDSL sym 2048 5 + ∆*
(Case 16)
19 BT2-R 3.9 STU-R 2560 32-TCPAM Symmetric 49 HDSL sym 2048 5 + ∆*
(Case 16)
NOTE – The crosstalk scenarios listed in this table were developed under the assumption of a 50 pair cable binder.
Cable binders of other sizes are for further study.
* The indicated noise margins in Table F.4 shall have a tolerance of 1.25 dB due to the aggregate effect of crosstalk
generator tolerance and calibrated loop simulator tolerance. The offset ∆ is defined in A.3.1.4.
All interferers are assumed to be co-located. All interferer PSDs are described in A.3.3.9. The
disturbers used for these tests are identical to those used in Annex A. For example, test 1 in
Table F.4 uses the identical disturber shape as case 4 from Annex A, exactly as described by
PSDCase-4 in A.3.3.9.
F.6 Functional characteristics

Functional characteristics return loss, Span Powering, Longitudinal Balance, and Longitudinal
Output Voltage shall be as described in A.5.
ITU-T Rec. G.991.2 (12/2003) 195

Annex G
Reserved for Region 2 requirements for data rates

between 2320 kbit/s and **max rate**
Annex H
Deactivation and warm-start procedure
Support of the reduced power mode, the deactivation and the warm-start is optional.
NOTE – Frequent transitions to/from the reduced power mode introduce a non-stationary noise environment,
the effect of which on deployed xDSL systems is not fully known. Because of this, regional access
restrictions regarding this procedure might apply.
H.1 Deactivation to reduced power mode

This clause describes waveforms at the loop interface and associated procedures during
deactivation. Figure H.1 illustrates the deactivation sequence.
H.1.1 Deactivation sequence
The deactivation can be initiated by the STU-R or by the STU-C. EOC signalling is used to initiate
the deactivation. The initiating side is called unit A, the other side is called unit B.
The standard sequence is as follows: Upon receiving the EOC message "Deactivation Request",
unit B responds by the EOC message "Deactivation Response" or by "Generic Unable To Comply
(UTC)". After sending the "Deactivation Response" containing an acceptance to the deactivation
(Deactivation Acknowledge bit = "1"), unit B continues transmitting and waits for the deactivation
of unit A. After receiving the acceptance to the deactivation request, unit A stops transmitting and
enters the reduced power mode. After detecting that unit A has stopped transmitting, e.g., by
detecting an LOSW error, unit B stops transmitting and enters the reduced power mode as well.
The EOC messages "Deactivation Request" and "Deactivation Response" indicate the ability of the
sender for the deactivation to the reduced power mode and a subsequent warm-start.
196 ITU-T Rec. G.991.2 (12/2003)

Figure H.1/G.991.2 – Deactivation sequence
H.1.2 Deactivation inhibiting

With messages "Deactivation Request" and "Deactivation Acknowledge", however, each
transceiver can also inhibit or stop an initiated deactivation process by setting bit OK to "0" in the
relevant EOC message. This is useful when during or after the transmission of the "Deactivation
Request" it becomes apparent that the data link is about to be used.
In warm-start the transmission shall be active for at least time tactive to minimize effects of
non-stationary crosstalk to systems sharing the same binder.
H.1.3 Deactivation EOC messages
H.1.3.1 Deactivation Request message: Message ID 22
The Deactivation Request message is transmitted to request a deactivation or to withdraw an issued
deactivation request. The destination address shall be F16 to indicate this is a broadcast message.
ITU-T Rec. G.991.2 (12/2003) 197

Table H.1/G.991.2 – Deactivation Request
1 22 Message ID
2 bit 0 Deactivation Request Bit 0 = Deactivation request
1 = Deactivation request cancelled
3 Reserved Set to 0
H.1.3.2 Deactivation Response Message: Message ID 150

The deactivation response message is used to confirm the deactivation command or to refuse a
deactivation request.
Table H.2/G.991.2 – Deactivation Acknowledge
1 150 Message ID
2 bit 0 Deactivation Bit 0 = Deactivation OK
Acknowledge 1 = Deactivation not possible
3 Reserved Set to 0
H.2 Warm-start activation

The warm-start can be initiated by the STU-R or the STU-C. This clause describes waveforms at
the loop interface and associated procedures during warm-start. The direct specification of the
performance of individual receiver elements is avoided when possible. Instead, the transmitter
characteristics are specified on an individual basis and the receiver performance is specified on a
general basis as the aggregate performance of all receiver elements. Exceptions are made for cases
where the performance of an individual receiver element is crucial to inter-operability.
In contrast to the activation described in 6.2, a warm-start makes use of all settings stored in a
previous successful activation to achieve a minimum start-up time. An activation is successful if
convergence has been achieved and the data mode has been reached (see 6.1). All settings (i.e.,
negotiated configuration in the pre-activation, all data in the activation frame, and all values in
adaptive filters) have to be stored before deactivating the transmission. The warm-start relies on the
fact that all previously stored settings such as the transfer characteristics of the receive and transmit
path and the timing relation between receive and transmit signals are still relevant. Small changes
(e.g., due to variations of ambient temperature) should not inhibit the warm-start activation;
however, if the equipment or the loop characteristics have changed significantly, the warm-start
activation may fail, and a cold-start will be performed instead.
H.2.1 Warm-start activation PMD reference model
The block diagram of the warm-start activation PMD layer of an STU-C and STU-R transmitter is
shown in Figure H.2.
198 ITU-T Rec. G.991.2 (12/2003)

logical ones –d1(m)
s(m) x(m) y(m) Spectral
Scrambler Mapper Channel
precoder shaper
logical zeros –d0(m) z(t)
Output at
loop
G.994.1 interface
Figure H.2/G.991.2 – Warm-start activation PMD reference model
The time index m represents the symbol time, and t represents analogue time. Since activation uses
2-PAM modulation, the bit time is equivalent to the symbol time. The output of the scrambler is s(m).
The output of the mapper is y(m), and the output of the spectral shaper at the loop interface is z(t).
d1(m) is an initialization signal that shall be logical ones for all m. d0(m) is an initialization signal that
shall be logical zeros for all m. The modulation format shall be Tomlinson-coded 2-level signal, with
the full symbol rate selected for data mode operation. During activation, the timing reference for the
activation signals have a tolerance of ±32 ppm at the STU-C and ±100 ppm at the STU-R.
The output bits from the scrambler s(m) shall be mapped to an output level y(m) as follows:
Table H.3/G.991.2 – Bit-to-level mapping

Scrambler output s(m) Mapper output level y(m) Data mode index
0 –9/16 0011
1 +9/16 1000
The levels corresponding to a 0 and 1 at the output of the scrambler shall be identical to the levels
of the 16-TCPAM constellation corresponding to indexes 0011 and 1000 respectively.
H.2.2 Warm-start activation sequence
The sequence and timing diagram for the warm-start activation sequence is given in Figure H.3.
tWSact
tWUN tWS tWUL tWS tECN tWS tSYN
STU-R WWUN WECN WSN WOKN Data
STU-R in
operational status
STU-C WWUL WSL WOKL Data
Half-duplex segments STU-C and STU-R

only if activation in operational status
initiated by STU-R
Figure H.3/G.991.2 – Timing diagram for the warm-start activation sequence
ITU-T Rec. G.991.2 (12/2003) 199

Table H.4/G.991.2 – Durations and tolerances for activation signals
Nominal
Signal Parameter Reference Tolerance
value
tWUN Duration of WWUN H.2.4.1 12 ms ±2 ms
tWS Guard time to prevent overlapping signals 6 ms ±2 ms
tWUL Duration of WWUL H.2.4.2 20 ms ±2 ms
tECN Duration of the half-duplex segment of the STU-R H.2.4.3 40 ms ±2 ms
tSYN Minimum Duration of the half-duplex segment of the STU-C 100 ms ±2 ms
tWSact Maximum activation time 500 ms
tactive Minimum time the link has to remain active 5 min
NOTE – The maximum time for activation, occurring after a deactivation without any intervening
loopback or powering action and without any change in cable characteristic for a metallic pair cable
transmission system is tWSact. This value for activation time is understood as a 95%-value when testing
with line models specified for the digital transmission system.
H.2.3 State Transition Diagram

The state transition diagram for the warm-start activation of the STU-R and the STU-C is given in
Figure H.4.
200 ITU-T Rec. G.991.2 (12/2003)

Figure H.4/G.991.2 – STU-C and STU-R transmitter warm-start state transition diagram
H.2.4 Signals used in warm-start activation

H.2.4.1 Signal WWUN
The STU-R initiated warm-start shall start with the STU-R sending the warm-start wake up signal,
WWUN for a duration of tWUN. The waveform and the transmit power of WWUN is the same as of the
12 kHz R-Tone used in ITU-T Rec. G.994.1 [2].
ITU-T Rec. G.991.2 (12/2003) 201

H.2.4.2 Signal WWUL
The wake-up signal for the STU-C initiated warm-start shall be the WWUL. If the warm-start is
initiated by the STU-R, the STU-C shall send the signal WWUL after detecting the signal WWUN.
WWUL shall have a duration of tWUL. The waveform and the transmit power of WWUL is the same as
of the 20 kHz C-Tone used in ITU-T Rec. G.994.1 [2].
H.2.4.3 Signal WECN
The STU-R shall send WECN, beginning tWS after the end of WWUL. Waveform WECN shall be
generated by connecting logical ones to the input of the STU-R scrambler as shown in Figure H.2.
The transmit power, symbol rate and PSD mask for WECN shall be as for signal WSL.
Half-duplex signal WECN shall be sent for time tECN.
H.2.4.4 Signal WSL
The STU-C shall send WSL beginning tWS after the end of WECN. Waveform WSL shall be generated
by connecting logical ones to the input of the STU-C scrambler as shown in Figure H.2. The
transmit power, symbol rate and PSD mask for WSL shall be as negotiated during the pre-activation
sequence.
H.2.4.5 Signal WSN
The STU-R shall start transmitting WSN beginning tWS + tSYN after the end of WECN. Waveform WSN
shall be generated by connecting logical ones to the input of the STU-R scrambler as shown in
Figure H.2. The transmit power, symbol rate and PSD mask for WSN shall be as negotiated during
the pre-activation sequence.
H.2.4.6 Signal WOKN
The STU-R shall start transmitting WOKN when the STU-R achieves full operational status. Full
operational status of the STU-R means that the STU-R is ready to enter data mode. Waveform
WOKN shall be generated by connecting logical zeros to the input of the STU-R scrambler as shown
in Figure H.2. The transmit power, symbol rate and PSD mask for WOKN shall be as for signal WSN.
H.2.4.7 Signal WOKL
The STU-C shall send WOKL when the STU-C has both detected WOKL and achieves full operational
status. Full operational status of the STU-C means that the STU-C is ready to enter data mode.
Waveform WOKL shall be generated by connecting logical zeros signal to the input of the STU-C
scrambler as shown in Figure H.2. The transmit power, symbol rate and PSD mask for WOKL shall
be the same as for WSL. WOKL shall be sent for exactly 256 symbols.
H.2.4.8 Datac and Datar
Within 200 symbols after the end of WOKL, the STU-C shall send Datac and STU-R shall send
Datar. These signals are described in 6.2.2.7. There is no required relationship between the end of
WOKL and any bit within the SHDSL data-mode frame. The SHDSL payload data shall be valid
TpayloadValid (see Table H.5) after the end of WOKL.
H.2.4.9 Warm-start exception condition
An exception condition shall be declared during warm-start if the timeout values given in Table H.5
expire or if any vendor-defined abnormal event occurs.
202 ITU-T Rec. G.991.2 (12/2003)

H.2.4.10 Warm-start exception state
If an exception condition is declared during warm-start, the STU-C or STU-R enters the exception
state and warm-start is aborted. During the exception state, the STU shall be silent for at least Tsilence
(see Table H.5), wait for transmission from the far end to cease, then return to the corresponding
initial start-up state. The STU-R and STU-C shall begin pre-activation, as per 6.3.
H.2.4.11 Timeouts
Table H.5 shows the system timeouts and their values.
Table H.5/G.991.2 – Timeout values

Name Parameter Value
Tsilence Minimum time in the warm-start exception state where the See Table 6-3
STU-C or STU-R are silent before the start of pre-activation.
TpayloadValid Time from start of Datac or Datar to valid SHDSL payload data See Table 6-3
Appendix I
Test circuit examples

I.1 Example crosstalk injection test circuit
Figure I.1 is an example of a high-impedance crosstalk injection circuit.
Figure I.1/G.991.2 – Example high-impedance crosstalk injection circuit
I.2 Example coupling circuits for longitudinal balance and longitudinal output voltage
Longitudinal balance and longitudinal output voltage may be measured using the coupling circuits
described in ANSI/IEEE Standard 455-1985 [B7] and ITU-T Rec. O.9 [B8]. The coupling circuit in
Figure I.2 is based upon the measurement method defined in ANSI/IEEE Standard 455-1985. In
order to provide sufficient measurement resolution the resistors must be matched within 0.05 %
tolerance. The coupling circuit in Figure I.3 is based on the measurement method described in
ITU-T Rec. O.9. This test circuit uses precision balanced (bifilar wound) transformers/baluns and
ITU-T Rec. G.991.2 (12/2003) 203

does not require precision matched resistors. The balun circuit is often more convenient for high-
frequency measurements.
Figure I.2/G.991.2 – Example resistive coupling circuit
Figure I.3/G.991.2 – Example balun coupling circuit
I.3 Return loss test circuit

The test circuit in Figure I.4 is based upon the traditional return loss bridge with added components
to accommodate the DC power feed voltage and provide transformer isolation for the measurement
instrumentation. Transformer isolation of both test signal source and meter load prevent
instrumentation and the DUT power feed circuitry. Input VIN is connected to a sweeping sine wave
generator (50 Ω source) and VOUT is connected to a high-impedance frequency selective voltmeter
(or spectrum analyser). For this test circuit, the return loss is defined as follows:
Z ( f ) + Z REF
Return Loss(f) = 20 log TEST
ZTEST ( f ) − Z REF
Figure I.4/G.991.2 – Example return loss bridge test circuit (ground isolated)
204 ITU-T Rec. G.991.2 (12/2003)

I.4 Transmit PSD/total power measurement test circuit
The test circuit in Figure I.5 is designed to measure total transmit power and transmit PSD. The test
contains provisions for DC power feed and transformer isolation for the measurement
instrumentation. Transformer isolation of the instrumentation input prevents measurement errors
from unintentional circuit paths through the common ground of the instrumentation and the DUT
power feed circuitry. VOUT is connected to a high-impedance wideband rms voltmeter (or spectrum
analyser).
Figure I.5/G.991.2 – Example ground-isolated power/PSD measurement test circuit
Appendix II
Typical characteristics of cables

II.1 Typical characteristics of cables for Annex B
NOTE – Parameters in this appendix differ from those specified in ITU-T Rec. G.996.1 [B11] for PE 04
and PE 05 cable.
Table II.1/G.991.2 – PE cable constants
PE04 PE05 PE06 PE08
freq Rs Ls Cp Rs Ls Cp Rs Ls Cp Rs Ls Cp
[Hz] [Ω/m] [H/m] [F/m] [Ω/m] [H/m] [F/m] [Ω/m] [H/m] [F/m] [Ω/m] [H/m] [F/m]
× 103 × 10–3 × 10–9 × 10-12 × 10–3 × 10–9 × 10–12 × 10–3 × 10–9 × 10–12 × 10–3 × 10–9 × 10–12
0 268 680 45.5 172 680 25 119 700 56 67 700 37.8
10 268 678 45.5 172 678 25 120 695 56 70.0 700 37.8
20 269 675 45.5 173 675 25 121 693 56 72.5 687 37.8
40 271 669 45.5 175 667 25 125 680 56 75.0 665 37.8
100 282 650 45.5 190 646 25 146 655 56 91.7 628 37.8
150 295 642 45.5 207 637 25 167 641 56 105 609 37.8
200 312 635 45.5 227 629 25 189 633 56 117 595 37.8
400 390 619 45.5 302 603 25 260 601 56 159 568 37.8
500 425 608 45.5 334 592 25 288 590 56 177.5 560 37.8
700 493 593 45.5 392 577 25 340 576 56 209 553 37.8
1000 582 582 45.5 466 572 25 405 570 56 250 547 37.8
2000 816 571 45.5 655 565 25 571 560 56 353 540 37.8
ITU-T Rec. G.991.2 (12/2003) 205

Table II.2/G.991.2 – PVC cable constants
PVC032 PVC04 PVC063
freq Rs Ls Cp Rs Ls Cp Rs Ls Cp
[Hz] [Ω/m] [H/m] [F/m] [Ω/m] [H/m] [F/m] [Ω/m] [H/m] [F/m]
× 103 × 10–3 × 10–9 × 10–12 × 10–3 × 10–9 × 10–12 × 10–3 × 10–9 × 10–12
0 419 650 120 268 650 120 108 635 120
10 419 650 120 268 650 120 108 635 120
20 419 650 120 268 650 120 108 635 120
40 419 650 120 268 650 120 111 630 120
100 427 647 120 281 635 120 141 604 120
150 453 635 120 295 627 120 173 584 120
200 493 621 120 311 619 120 207 560 120
400 679 577 120 391 592 120 319 492 120
500 750 560 120 426 579 120 361 469 120
700 877 546 120 494 566 120 427 450 120
1000 1041 545 120 584 559 120 510 442 120
2000 1463 540 120 817 550 120 720 434 120
Appendix III
Signal regenerator start-up description
This appendix describes the start-up sequence used on spans employing regenerators. The sequence
applies to spans with an arbitrary number of regenerators (up to 8), but for simplicity, the
description here assumes a two-regenerator link. The use of line probing is optional, but its use is
assumed for the purpose of this description.
The basic premise is that capability lists and line probe results propagate from the STU-R toward
the STU-C and that the SHDSL training begins at the STU-C and propagates in the direction toward
the STU-R. The Regenerator Silent Period (RSP) bit in ITU-T Rec. G.994.1 is used to hold off
segments while the start-up process propagates across the span.
The block diagram in Figure III.1 shows a typical SHDSL span with two regenerators as a reference
for the start-up sequences described below.
Figure III.1/G.991.2 – Block diagram of a SHDSL span with two signal regenerators
206 ITU-T Rec. G.991.2 (12/2003)

III.1 STU-R initiated Start-up
In most typical SHDSL installations, the STU-R can be expected to initiate the start-up process. The
proposed SHDSL start-up process for STU-R initiation is described in the text below and shown
graphically in Table III.1.
In this mode, the STU-R triggers the start-up process by initiating a G.994.1 session with the
regenerator closest to it (over segment TR2). The STU-R and the SRU2-C then exchange
capabilities and optionally perform a line probe and a second capabilities exchange. The units do
not have enough information to begin SHDSL activation at this point, so the SRU2-C issues an MS
with the RSP bit set to hold off the STU-R while the start-up process propagates across the span.
The G.994.1 session terminates normally, and the STU-R begins its waiting period.
Next, the SRU2-C conveys the capabilities from Segment TR2 to the SRU2-R across the
regenerator's internal interface. The SRU2-R then initiates a G.994.1 session with the SRU1-C and
performs the same capabilities exchange and line probing sequence described above for the first
segment. The capabilities expressed by the SRU2-R are the intersection of its own capabilities with
the capabilities it has received for Segment TR2. The units still do not have sufficient information
to begin SHDSL activation, so, again, the SRU1-R issues an MS with the RSP bit set. The G.994.1
session terminates normally, and the SRU2-R begins its waiting period.
As before, the SRU1-C then conveys the capabilities from Segment RR1 (including the information
from Segment TR2) to the SRU1-R across the regenerator's internal interface. The SRU1-R initiates
a G.994.1 session with the STU-C and performs a capabilities exchange. Optionally, a line probe
and a second capabilities exchange may be used. As before, the capabilities expressed by the
SRU1-R are the intersection of its own capabilities with the capabilities it has received for Segments
RR1 and TR2. At this point, the STU-C possesses all of the required information to select the span's
operational parameters. The data rate and other parameters are selected, just as in a normal (non-
regenerator) pre-activation sequence and then the SHDSL activation begins for Segment TR1.
When the STU-C/SRU1-R link (over Segment TR1) has completed the SHDSL activation sequence
(or the G.994.1 session, if clock mode 1 is selected), the SRU1-R communicates the selected
operational parameters to the SRU1-C across the regenerator's internal interface. At this point, the
SRU1-C initiates a G.994.1 session with the SRU2-R over Segment RR1. Parameters are selected –
there should be no need for another CLR-CL exchange at this point – and the units perform the
normal SHDSL activation. If clock mode 1 is selected (classic plesiochronous), there is no need to
lock symbol timing to a network clock reference. In this case, the SRU1-C/SRU2-R G.994.1 session
and activation should begin as soon as the STU-C/SRU1-R G.994.1 sessions complete. In clock
modes 2, 3a, and 3b, such a network or data clock reference is necessary for establishing symbol
timing. In these modes, the SRU1-C will delay the initiation of its G.994.1 session until the
STU-C/SRU1-R activation is complete. In this way, the required reference clock will be available
for symbol timing on the SRU1-C/SRU2-R segment.
When the SRU1-C/SRU2-R link (over Segment RR1) has completed the SHDSL activation
sequence (or the G.994.1 session, if clock mode 1 is selected), the SRU2-R communicates the
selected operational parameters to the SRU2-C across the regenerator's internal interface. The
SRU2-C initiates a G.994.1 session with the STU-R over Segment TR2. Parameters are selected and
the units perform the normal SHDSL activation. When this activation sequence is complete, the
span can become fully operational.
ITU-T Rec. G.991.2 (12/2003) 207

Table III.1/G.991.2 – STU-R initiated start-up sequence
Segment TR2 Segment RR1 Segment TR1
(STU-R/SRU2-C) (SRU2-R/SRU1-C) (SRU1-R/STU-C)
G.994.1 Start →
Line probe
← MS (RSP)
G.994.1 Start →
Line probe
← MS (RSP)
G.994.1 Start →
Line probe
Mode Selection
SHDSL activation
← G.994.1 Start
Mode Selection
SHDSL activation
← G.994.1 Start
Mode Selection
SHDSL activation
III.2 STU-C initiated start-up

In some cases, it may be desirable for the STU-C to initiate the start-up process. The proposed
SHDSL start-up process for STU-C initiation is described in the text below and shown graphically
in Table III.2.
In this mode, the STU-C triggers the start-up process by initiating a G.994.1 session with the
regenerator closest to it (over segment TR1). The SRU2-C issues an MS with the RSP bit set to hold
off the STU-C while the start-up process propagates across the span. The G.994.1 session
terminates normally, and the STU-C begins its wait period. Next, the SRU1-C initiates a G.994.1
session with the SRU2-R, which, again is terminated following an MS from the SRU2-R with the
RSP bit set.
The SRU2-C next initiates a G.994.1 session with the STU-R. From this point on, the start sequence
is as described in III.1 for the STU-R initiated start-up.
208 ITU-T Rec. G.991.2 (12/2003)

Table III.2/G.991.2 – STU-C initiated start-up sequence
Segment TR2 Segment RR1 Segment TR1
(STU-R/SRU2-C) (SRU2-R/SRU1-C) (SRU1-R/STU-C)
← G.994.1 Start
MS (RSP) →
← G.994.1 Start
MS (RSP) →
← G.994.1 Start
Line probe
← MS (RSP)
G.994.1 Start →
Line probe
← MS (RSP)
G.994.1 Start →
Line probe
Mode Selection
SHDSL activation
← G.994.1 Start
Mode Selection
SHDSL activation
← G.994.1 Start
Mode Selection
SHDSL activation
III.3 SRU initiated start-up

In some limited applications (including some maintenance and retrain scenarios), it may be
desirable for a regenerator to initiate the start sequence. In this mode, the SRU will initiate the train
in the downstream direction – i.e., toward the STU-R in the same manner that it would have for the
corresponding segment of the STU-C Start-up Procedure (as described in III.2). The STU-R will
then initiate the capabilities exchange and line probing procedure toward the STU-C, as in a normal
STU-C initiated start-up. The start-up sequence begins with the initiating SRU-C and propagating
toward the STU-R.
III.4 Collisions and retrains

Collisions (equivalent to "glare" conditions in voice applications) can occur in cases where both the
STU-C and the STU-R attempt to initiate connections simultaneously. Using the process described
above, these collisions are resolved by specifying that R-to-C capabilities exchanges and probes
will always take precedence over C-to-R train requests. G.994.1 sessions inherently resolve
collisions on individual segments.
ITU-T Rec. G.991.2 (12/2003) 209

In ITU-T Rec. G.994.1, the RSP timeout is specified as approximately 1 minute. For spans with no
more than one regenerator, this is ideal. For multi-regenerator spans, however, an STU may time
out and initiate a new G.994.1 session before the SRU is prepared to begin the next phase of the
train. In such cases, the SRU should respond to the G.994.1 initiation and issue an MS message
with the RSP bit set to hold off the STU once again. For its part, the SRU should implement an
internal timer and should not consider a start-up to have failed until that timer has expired. The
timer should be started when the SRU receives a RSP bit in an MS message and should not expire
for at least 4 minutes.
If any segment must retrain due to line conditions or other causes, each segment of the span shall be
deactivated and the full start-up procedure shall be reinitiated.
III.5 Diagnostic mode activation

If a segment fails, the start-up procedure will also fail for the entire span. This would normally be
characterized at the STU by being told to enter a silent interval via the RSP bit and never receiving
another G.994.1 request. Without some diagnostic information, the service provider would have no
easy way to test the integrity of the various segments.
This concern is resolved by the use of the "Diagnostic Mode" in ITU-T Rec. G.994.1 to trigger a
diagnostic training mode. This bit, when set, causes an SRU connected to a failed segment to act as
an STU and allow the start-up procedure to finish. In this way, all of the segments before the failed
segment may be tested using loopbacks and EOC-initiated tests, allowing network operators to
quickly isolate the segment where the failure has occurred.
Appendix IV
Tabulation of Annex B noise profiles
Appendix IV tabulates the total noise profile (sum of self and alien) corresponding to 0 dB of
margin for all the Annex B test cases. Those noise PSDs were used during the theoretical
computation of the margin. The tabulated noise profiles should be measured into the calibrating
impedance (see B.3.3.1).
Noise Profile Nomenclature: ABBBCDE
A: Side (either C or R)
BBB: Rate
C: PSD type (either s for symmetric or a for asymmetric)
D: Noise Type (A ,B, C or D)
E: Loop Number (from 2 to 7).
The noise shapes used for test #1 will be identical to Noise A of test #2
210 ITU-T Rec. G.991.2 (12/2003)

Table IV.1/G.991.2 – STU-C side/symmetric PSDs
Noise
Magnitude of the noise in dBm per Hz (sign is always negative) as a function of frequency in kHz
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
C384sA2 114.9 99.2 95.0 93.1 92.5 92.3 92.9 93.9 93.4 92.7 92.0 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C384sC2 120.6 104.6 100.4 98.4 97.7 97.6 98.7 101.8 102.6 102.0 101.3 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C384sD2 131.8 104.4 99.5 97.1 95.8 95.3 96.4 100.5 107.0 114.2 121.6 138.0 138.0 138.0 138.0 138.0 138.0 138.0 138.0
C512sA2 114.9 99.4 95.3 93.4 92.8 92.3 92.0 91.9 92.0 92.2 91.9 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C512sC2 120.6 104.9 100.8 98.8 98.1 97.7 97.4 97.5 98.3 100.0 100.9 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C512sD2 132.8 105.6 100.6 98.0 96.4 95.4 94.8 94.8 95.8 98.7 103.2 131.4 138.0 138.0 138.0 138.0 138.0 138.0 138.0
C768sA2 114.9 99.6 95.6 93.7 93.3 92.8 92.4 91.9 91.2 90.7 90.3 88.1 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C768sC2 120.6 105.2 101.2 99.3 98.8 98.4 98.0 97.5 97.0 96.6 96.4 94.4 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C768sD2 134.2 107.3 102.2 99.5 97.7 96.5 95.5 94.8 94.3 93.9 93.8 102.6 120.9 138.0 138.0 138.0 138.0 138.0 138.0
C1024sA2 114.9 99.7 95.7 93.9 93.6 93.2 92.8 92.3 91.5 90.8 90.3 87.5 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C1024sC2 120.6 105.3 101.4 99.6 99.2 99.0 98.7 98.2 97.5 96.9 96.4 93.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C1024sD2 135.0 108.5 103.3 100.6 98.8 97.5 96.4 95.5 94.9 94.3 93.9 93.6 102.1 115.8 130.8 138.0 138.0 138.0 138.0
C1280sA2 114.9 99.7 95.8 93.9 93.8 93.5 93.1 92.6 91.8 91.1 90.4 87.3 86.0 84.9 83.8 82.9 82.1 79.4 77.6
C1280sC2 120.6 105.4 101.5 99.7 99.5 99.4 99.3 98.8 98.0 97.4 96.8 93.2 92.2 91.3 90.2 89.3 88.5 87.8 86.8
C1280sD2 135.7 109.4 104.3 101.5 99.7 98.3 97.2 96.3 95.5 94.9 94.4 92.8 94.0 101.7 112.6 124.2 136.9 138.0 138.0
C1536sA2 115.0 99.7 95.8 94.0 93.8 93.6 93.3 92.8 92.0 91.2 90.6 87.3 85.8 84.7 83.8 82.9 82.1 79.4 77.6
C1536sC2 120.6 105.4 101.5 99.8 99.6 99.7 99.7 99.3 98.5 97.8 97.2 93.3 91.9 91.1 90.2 89.3 88.5 87.8 86.8
C1536sD2 136.1 110.2 105.0 102.3 100.4 99.0 97.9 96.9 96.1 95.5 94.9 92.9 92.3 94.4 101.4 110.4 119.9 138.0 138.0
C2048sA2 115.0 99.7 95.7 93.9 93.8 93.6 93.3 92.8 91.9 91.2 90.5 87.2 85.5 84.3 83.5 82.8 82.1 79.4 77.6
C2048sC2 120.6 105.4 101.5 99.8 99.6 99.7 99.7 99.4 98.5 97.8 97.2 93.1 91.6 90.4 89.7 89.2 88.5 87.8 86.8
C2048sD2 136.3 110.4 105.2 102.5 100.6 99.1 98.0 97.0 96.2 95.5 94.8 92.6 91.3 90.7 91.2 94.1 99.8 128.9 138.0
C2304sA2 115.0 99.7 95.8 94.0 93.8 93.6 93.4 92.9 92.0 91.2 90.6 87.2 85.5 84.3 83.4 82.7 82.0 79.4 77.6
C2304sC2 120.6 105.4 101.5 99.8 99.7 99.9 100.0 99.7 98.8 98.1 97.4 93.2 91.6 90.4 89.5 88.9 88.4 87.8 86.8
C2304sD2 136.6 110.9 105.7 102.9 101.0 99.6 98.4 97.4 96.6 95.9 95.3 92.9 91.5 90.7 90.4 91.3 94.4 118.1 138.0
211 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
C384sD3 131.8 104.4 99.5 97.1 95.8 95.3 96.4 100.5 107.0 114.2 121.6 138.0 138.0 138.0 138.0 138.0 138.0 138.0 138.0
C512sD3 132.8 105.6 100.6 98.0 96.4 95.4 94.8 94.8 95.8 98.7 103.2 131.4 138.0 138.0 138.0 138.0 138.0 138.0 138.0
C768sD3 134.1 107.3 102.2 99.5 97.7 96.5 95.5 94.8 94.3 93.9 93.8 102.6 120.9 138.0 138.0 138.0 138.0 138.0 138.0
C1024sD3 135.0 108.5 103.3 100.6 98.8 97.4 96.4 95.5 94.8 94.3 93.9 93.6 102.1 115.8 130.8 138.0 138.0 138.0 138.0
C1280sD3 135.6 109.4 104.2 101.5 99.7 98.3 97.2 96.2 95.5 94.9 94.3 92.8 94.0 101.7 112.6 124.2 136.9 138.0 138.0
C1536sD3 136.1 110.1 105.0 102.3 100.4 99.0 97.8 96.9 96.1 95.4 94.8 92.9 92.3 94.4 101.4 110.4 119.9 138.0 138.0
C2048sD3 136.3 110.3 105.2 102.4 100.5 99.1 97.9 96.9 96.1 95.4 94.8 92.6 91.3 90.7 91.2 94.1 99.8 128.9 138.0
C2304sD3 136.6 110.8 105.6 102.9 101.0 99.5 98.3 97.4 96.5 95.8 95.2 92.9 91.5 90.7 90.4 91.3 94.4 118.1 138.0
C384sA4 114.9 99.2 95.0 93.1 92.5 92.3 92.9 93.9 93.4 92.7 92.0 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C384sC4 120.6 104.6 100.4 98.4 97.7 97.6 98.7 101.8 102.6 102.0 101.3 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C512sA4 114.9 99.4 95.3 93.4 92.8 92.3 92.0 91.9 92.0 92.2 91.9 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C512sC4 120.6 104.9 100.8 98.8 98.1 97.7 97.4 97.5 98.3 100.0 100.9 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C768sA4 114.9 99.5 95.6 93.7 93.3 92.8 92.4 91.8 91.2 90.6 90.3 88.0 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C768sC4 120.6 105.2 101.2 99.3 98.7 98.4 98.0 97.5 97.0 96.6 96.4 94.4 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C1024sA4 114.9 99.6 95.7 93.8 93.6 93.2 92.8 92.2 91.5 90.8 90.2 87.5 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C1024sC4 120.6 105.3 101.4 99.6 99.2 99.0 98.7 98.2 97.5 96.9 96.4 93.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C1280sA4 114.9 99.6 95.7 93.9 93.7 93.4 93.1 92.5 91.7 91.0 90.4 87.3 86.0 84.9 83.8 82.9 82.1 79.4 77.6
C1280sC4 120.6 105.3 101.5 99.7 99.4 99.4 99.2 98.8 98.0 97.3 96.8 93.2 92.2 91.3 90.2 89.3 88.5 87.8 86.8
C1536sA4 114.9 99.6 95.7 93.9 93.8 93.5 93.2 92.7 91.9 91.2 90.5 87.3 85.8 84.7 83.8 82.9 82.1 79.4 77.6
C1536sC4 120.6 105.3 101.5 99.8 99.6 99.7 99.7 99.3 98.5 97.8 97.2 93.3 91.9 91.1 90.2 89.3 88.5 87.8 86.8
C2048sA4 115.0 99.6 95.6 93.9 93.7 93.5 93.2 92.6 91.8 91.0 90.4 87.1 85.5 84.3 83.5 82.8 82.1 79.4 77.6
C2048sC4 120.6 105.3 101.4 99.7 99.6 99.7 99.7 99.3 98.4 97.7 97.1 93.1 91.5 90.4 89.7 89.2 88.5 87.8 86.8
C2304sA4 115.0 99.6 95.6 93.9 93.7 93.5 93.2 92.7 91.8 91.1 90.4 87.1 85.4 84.2 83.4 82.7 82.0 79.4 77.6
C2304sC4 120.6 105.3 101.4 99.7 99.6 99.8 99.9 99.5 98.7 98.0 97.3 93.2 91.6 90.4 89.5 88.9 88.4 87.8 86.8
212 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
C384sB5 120.4 104.6 100.3 98.4 97.7 97.6 98.7 101.7 102.6 102.0 101.3 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C512sB5 120.4 104.9 100.7 98.8 98.0 97.6 97.4 97.5 98.3 100.0 100.9 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C768sB5 120.4 105.1 101.1 99.2 98.6 98.3 97.9 97.4 96.9 96.5 96.3 94.4 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C1024sB5 120.4 105.2 101.2 99.4 99.0 98.8 98.5 98.0 97.4 96.8 96.3 93.5 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C1280sB5 120.4 105.2 101.3 99.5 99.2 99.1 99.0 98.6 97.8 97.2 96.7 93.2 92.2 91.3 90.2 89.3 88.5 87.8 86.9
C1536sB5 120.4 105.1 101.2 99.5 99.3 99.3 99.3 98.9 98.2 97.5 97.0 93.2 91.9 91.1 90.2 89.3 88.5 87.8 86.9
C2048sB5 120.4 105.0 101.1 99.3 99.1 99.1 99.1 98.7 97.9 97.3 96.7 93.0 91.5 90.4 89.7 89.2 88.5 87.8 86.9
C2304sB5 120.5 104.9 101.0 99.2 99.0 99.1 99.1 98.8 98.1 97.4 96.8 93.0 91.5 90.4 89.5 88.9 88.4 87.8 86.9
C384sA6 114.9 99.2 95.0 93.1 92.5 92.3 92.9 93.9 93.4 92.7 92.0 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C384sC6 120.6 104.6 100.4 98.4 97.7 97.6 98.7 101.8 102.6 102.0 101.3 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C512sA6 114.9 99.4 95.3 93.4 92.8 92.3 92.0 91.9 92.0 92.2 91.9 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C512sC6 120.6 104.9 100.8 98.8 98.1 97.7 97.4 97.5 98.3 100.0 100.9 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C768sA6 114.9 99.6 95.6 93.7 93.3 92.9 92.4 91.9 91.2 90.7 90.3 88.0 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C768sC6 120.6 105.2 101.2 99.3 98.8 98.4 98.0 97.5 97.0 96.6 96.4 94.4 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C1024sA6 115.0 99.7 95.7 93.9 93.6 93.2 92.8 92.3 91.5 90.9 90.3 87.5 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C1024sC6 120.6 105.3 101.4 99.6 99.2 99.0 98.7 98.2 97.5 96.9 96.4 93.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C1280sA6 115.0 99.7 95.8 94.0 93.8 93.5 93.2 92.7 91.9 91.2 90.5 87.3 86.0 84.9 83.8 82.9 82.1 79.4 77.6
C1280sC6 120.6 105.4 101.5 99.7 99.5 99.4 99.3 98.8 98.1 97.4 96.8 93.2 92.2 91.3 90.2 89.3 88.5 87.8 86.8
C1536sA6 115.1 99.8 95.8 94.0 93.9 93.7 93.4 92.9 92.1 91.4 90.7 87.3 85.7 84.8 83.8 82.9 82.1 79.4 77.6
C1536sC6 120.6 105.4 101.5 99.8 99.6 99.8 99.8 99.4 98.6 97.9 97.3 93.3 91.9 91.1 90.2 89.3 88.5 87.8 86.8
C2048sA6 115.4 100.0 95.9 94.0 93.9 93.7 93.4 93.0 92.1 91.4 90.7 87.2 85.5 84.4 83.5 82.8 82.1 79.4 77.6
C2048sC6 120.7 105.4 101.5 99.8 99.7 99.8 99.8 99.5 98.6 97.9 97.3 93.1 91.5 90.4 89.7 89.2 88.5 87.8 86.8
C2304sA6 115.6 100.2 96.0 94.1 94.0 93.8 93.6 93.1 92.3 91.5 90.9 87.3 85.5 84.4 83.4 82.7 82.0 79.4 77.6
C2304sC6 120.7 105.4 101.5 99.8 99.7 100.0 100.1 99.8 99.0 98.3 97.6 93.2 91.6 90.4 89.6 88.9 88.4 87.8 86.8
213 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
C384sA7 114.9 99.2 95.0 93.1 92.5 92.3 92.9 93.9 93.4 92.7 92.0 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C384sB7 120.6 104.6 100.4 98.4 97.7 97.6 98.7 101.8 102.6 102.0 101.3 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C384sC7 120.6 104.6 100.4 98.4 97.7 97.6 98.7 101.8 102.6 102.0 101.3 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C384sD7 131.8 104.4 99.5 97.1 95.8 95.3 96.4 100.5 107.0 114.2 121.6 138.0 138.0 138.0 138.0 138.0 138.0 138.0 138.0
C512sA7 114.9 99.4 95.3 93.4 92.8 92.3 92.0 91.9 92.0 92.2 91.9 88.1 86.3 85.0 83.8 82.9 82.1 79.4 77.6
C512sB7 120.6 104.9 100.8 98.8 98.1 97.7 97.4 97.5 98.3 100.0 100.9 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C512sC7 120.6 104.9 100.8 98.8 98.1 97.7 97.4 97.5 98.3 100.0 100.9 94.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C512sD7 132.8 105.6 100.6 98.0 96.4 95.4 94.8 94.8 95.8 98.7 103.2 131.4 138.0 138.0 138.0 138.0 138.0 138.0 138.0
C768sA7 114.9 99.5 95.6 93.7 93.3 92.8 92.4 91.8 91.2 90.7 90.3 88.1 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C768sB7 120.6 105.2 101.2 99.3 98.7 98.4 98.0 97.5 97.0 96.6 96.4 94.4 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C768sC7 120.6 105.2 101.2 99.3 98.7 98.4 98.0 97.5 97.0 96.6 96.4 94.4 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C768sD7 134.1 107.2 102.1 99.5 97.7 96.5 95.5 94.8 94.3 93.9 93.8 102.6 120.9 138.0 138.0 138.0 138.0 138.0 138.0
C1024sA7 114.9 99.6 95.7 93.8 93.6 93.2 92.8 92.2 91.5 90.8 90.3 87.5 86.3 84.9 83.8 82.9 82.1 79.4 77.6
C1024sB7 120.6 105.3 101.4 99.6 99.2 99.0 98.7 98.2 97.5 96.9 96.4 93.5 92.7 91.4 90.2 89.3 88.5 87.8 86.9
C1024sC7 120.6 105.3 101.4 99.6 99.2 99.0 98.7 98.2 97.5 96.9 96.4 93.5 92.7 91.4 90.2 89.3 88.5 87.8 86.8
C1024sD7 135.0 108.4 103.3 100.6 98.8 97.4 96.4 95.5 94.9 94.3 93.9 93.6 102.1 115.8 130.8 138.0 138.0 138.0 138.0
C1280sA7 114.9 99.6 95.7 93.9 93.7 93.4 93.1 92.6 91.8 91.1 90.4 87.3 86.0 84.9 83.8 82.9 82.1 79.4 77.6
C1280sB7 120.6 105.3 101.5 99.7 99.4 99.4 99.2 98.8 98.0 97.4 96.8 93.2 92.2 91.3 90.2 89.3 88.5 87.8 86.9
C1280sC7 120.6 105.3 101.5 99.7 99.4 99.4 99.2 98.8 98.0 97.4 96.8 93.2 92.2 91.3 90.2 89.3 88.5 87.8 86.8
C1280sD7 135.7 109.4 104.2 101.5 99.7 98.3 97.2 96.3 95.5 94.9 94.3 92.8 94.0 101.7 112.6 124.2 136.9 138.0 138.0
C1536sA7 115.0 99.7 95.7 93.9 93.8 93.6 93.3 92.8 91.9 91.2 90.6 87.3 85.8 84.8 83.8 82.9 82.1 79.4 77.6
C1536sB7 120.6 105.3 101.5 99.8 99.6 99.7 99.7 99.3 98.5 97.8 97.2 93.3 91.9 91.1 90.2 89.3 88.5 87.8 86.9
C1536sC7 120.6 105.3 101.5 99.8 99.6 99.7 99.7 99.3 98.5 97.8 97.2 93.3 91.9 91.1 90.2 89.3 88.5 87.8 86.8
C1536sD7 136.1 110.1 105.0 102.2 100.4 99.0 97.8 96.9 96.1 95.5 94.9 92.9 92.3 94.4 101.4 110.4 119.9 138.0 138.0
214 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
C2048sA7 115.0 99.7 95.7 93.9 93.7 93.5 93.2 92.7 91.9 91.2 90.5 87.2 85.5 84.4 83.5 82.8 82.1 79.4 77.6
C2048sB7 120.6 105.3 101.4 99.7 99.6 99.7 99.7 99.3 98.5 97.8 97.2 93.1 91.6 90.4 89.7 89.2 88.5 87.8 86.9
C2048sC7 120.6 105.3 101.4 99.7 99.6 99.7 99.7 99.3 98.5 97.8 97.2 93.1 91.6 90.4 89.7 89.2 88.5 87.8 86.8
C2048sD7 136.3 110.3 105.1 102.4 100.5 99.1 98.0 97.0 96.2 95.5 94.8 92.6 91.3 90.7 91.2 94.1 99.8 128.9 138.0
C2304sA7 115.1 99.7 95.7 93.9 93.8 93.6 93.3 92.8 92.0 91.2 90.6 87.2 85.5 84.3 83.4 82.7 82.0 79.4 77.6
C2304sB7 120.6 105.3 101.4 99.7 99.6 99.8 99.9 99.6 98.8 98.1 97.4 93.2 91.6 90.4 89.6 88.9 88.4 87.8 86.9
C2304sC7 120.6 105.3 101.4 99.7 99.6 99.8 99.9 99.6 98.8 98.1 97.4 93.2 91.6 90.4 89.6 88.9 88.4 87.8 86.8
C2304sD7 136.6 110.8 105.6 102.9 101.0 99.5 98.4 97.4 96.6 95.9 95.3 92.9 91.5 90.7 90.4 91.3 94.4 118.1 138.0
Table IV.2/G.991.2 – STU-C side/asymmetric PSDs
Noise
profile
1 20 40 60 80 100 150 200 250 300 350 400 500 600 700 800 1000 1200 1400
C2048aA2 115.0 95.8 93.9 93.5 92.1 90.7 87.2 85.5 84.2 83.2 82.3 81.5 80.3 79.4 78.4 77.6 76.1 79.0 85.6
C2048aC2 120.6 101.6 99.8 100.1 99.0 97.5 93.2 91.5 90.2 89.1 88.2 87.5 86.5 87.7 87.6 86.8 85.3 87.9 93.8
C2048aD2 136.6 105.8 101.2 98.6 96.7 95.3 92.8 91.0 89.8 88.8 88.1 87.6 87.9 93.1 101.7 110.3 126.3 138.0 138.0
C2304aA2 115.0 95.8 94.0 93.8 92.5 91.1 87.5 85.7 84.4 83.4 82.5 81.8 80.4 79.4 78.4 77.6 76.1 79.0 85.6
C2304aC2 120.6 101.6 100.0 100.9 100.1 98.7 93.7 92.0 90.6 89.6 88.7 87.9 86.6 87.6 87.6 86.8 85.3 87.9 93.8
C2304aD2 137.5 107.5 102.9 100.3 98.5 97.0 94.5 92.8 91.5 90.5 89.7 89.1 88.7 92.0 99.6 107.6 122.5 135.2 138.0
C2048aD3 136.6 105.7 101.1 98.5 96.6 95.2 92.7 91.0 89.8 88.8 88.1 87.6 87.9 93.1 101.7 110.3 126.3 138.0 138.0
C2304aD3 137.5 107.4 102.8 100.2 98.3 96.9 94.4 92.7 91.5 90.5 89.7 89.1 88.7 92.0 99.6 107.6 122.5 135.2 138.0
C2048aA4 114.9 95.7 93.8 93.4 92.0 90.6 87.2 85.5 84.2 83.2 82.3 81.5 80.3 79.4 78.4 77.6 76.1 79.0 85.6
215 ITU-T Rec. G.991.2 (12/2003)

Table IV.2/G.991.2 – STU-C side/asymmetric PSDs
Noise
profile
1 20 40 60 80 100 150 200 250 300 350 400 500 600 700 800 1000 1200 1400
C2048aC4 120.6 101.5 99.7 100.0 98.8 97.4 93.2 91.5 90.1 89.1 88.2 87.5 86.5 87.7 87.6 86.8 85.3 87.9 93.8
C2304aA4 115.0 95.7 93.9 93.6 92.3 90.9 87.4 85.7 84.4 83.4 82.5 81.8 80.4 79.4 78.4 77.6 76.1 79.0 85.6
C2304aC4 120.6 101.5 99.9 100.8 99.9 98.5 93.7 92.0 90.6 89.6 88.7 87.9 86.6 87.6 87.6 86.8 85.3 87.9 93.8
C2048aB5 120.4 101.1 99.2 99.3 98.2 97.0 93.1 91.4 90.1 89.1 88.2 87.5 86.5 87.7 87.7 86.9 85.4 88.2 94.5
C2304aB5 120.4 101.1 99.3 99.9 99.1 97.9 93.5 91.9 90.6 89.5 88.7 87.9 86.6 87.6 87.7 86.9 85.4 88.2 94.5
C2048aA6 115.2 95.8 94.0 93.6 92.3 90.9 87.3 85.5 84.2 83.2 82.3 81.5 80.3 79.4 78.4 77.6 76.1 79.0 85.6
C2048aC6 120.6 101.6 99.8 100.2 99.1 97.6 93.2 91.5 90.2 89.1 88.2 87.5 86.5 87.7 87.6 86.8 85.3 87.9 93.8
C2304aA6 115.4 96.0 94.1 94.0 92.8 91.3 87.5 85.7 84.5 83.4 82.5 81.8 80.4 79.4 78.4 77.6 76.1 79.0 85.6
C2304aC6 120.7 101.6 100.0 101.1 100.3 98.8 93.7 92.0 90.6 89.6 88.7 87.9 86.6 87.6 87.6 86.8 85.3 87.9 93.8
C2048aA7 115.0 95.7 93.8 93.4 92.1 90.7 87.2 85.5 84.2 83.2 82.3 81.5 80.3 79.4 78.4 77.6 76.1 79.0 85.6
C2048aB7 120.6 101.5 99.7 100.1 98.9 97.5 93.2 91.5 90.2 89.1 88.2 87.5 86.5 87.7 87.7 86.9 85.4 88.2 94.5
C2048aC7 120.6 101.5 99.7 100.1 98.9 97.5 93.2 91.5 90.2 89.1 88.2 87.5 86.5 87.7 87.6 86.8 85.3 87.9 93.8
C2048aD7 136.6 105.7 101.1 98.5 96.7 95.3 92.7 91.0 89.8 88.8 88.1 87.6 87.9 93.1 101.7 110.3 126.3 138.0 138.0
C2304aA7 115.1 95.8 93.9 93.7 92.4 91.0 87.5 85.7 84.4 83.4 82.5 81.8 80.4 79.4 78.4 77.6 76.1 79.0 85.6
C2304aB7 120.6 101.5 99.9 100.9 100.0 98.6 93.7 92.0 90.6 89.6 88.7 87.9 86.6 87.6 87.7 86.9 85.4 88.2 94.5
C2304aC7 120.6 101.5 99.9 100.9 100.0 98.6 93.7 92.0 90.6 89.6 88.7 87.9 86.6 87.6 87.6 86.8 85.3 87.9 93.8
C2304aD7 137.5 107.4 102.8 100.2 98.4 97.0 94.5 92.7 91.5 90.5 89.7 89.1 88.7 92.0 99.6 107.6 122.5 135.2 138.0
216 ITU-T Rec. G.991.2 (12/2003)

Table IV.3/G.991.2 – STU-R side/symmetric PSDs
Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
R384sA2 114.9 99.2 95.3 93.7 92.6 92.1 92.3 92.4 91.8 91.0 90.4 87.9 86.2 84.8 87.3 93.1 98.1 123.1 115.4
R384sC2 120.6 104.6 100.2 98.1 97.3 96.9 97.3 98.2 97.6 96.9 96.2 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R384sD2 131.8 104.4 99.5 97.1 95.8 95.3 96.4 100.5 107.0 114.2 121.6 138.0 138.0 138.0 138.0 138.0 138.0 138.0 138.0
R512sA2 114.9 99.4 95.7 94.1 92.9 92.1 91.6 91.1 90.9 90.8 90.4 87.9 86.2 84.8 87.3 93.1 98.1 122.5 115.4
R512sC2 120.6 104.9 100.6 98.4 97.6 96.9 96.5 96.3 96.4 96.5 96.1 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R512sD2 132.8 105.6 100.6 98.0 96.4 95.4 94.8 94.8 95.8 98.7 103.2 131.4 138.0 138.0 138.0 138.0 138.0 138.0 138.0
R768sA2 114.9 99.6 96.0 94.5 93.4 92.6 91.9 91.0 90.4 89.8 89.3 87.9 86.1 84.8 87.3 93.0 98.0 117.7 114.9
R768sC2 120.6 105.2 101.0 98.9 98.1 97.5 96.9 96.3 95.6 95.1 94.7 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R768sD2 134.2 107.3 102.2 99.5 97.7 96.5 95.5 94.8 94.3 93.9 93.8 102.6 120.9 138.0 138.0 138.0 138.0 138.0 138.0
R1024sA2 114.9 99.7 96.1 94.7 93.7 92.9 92.2 91.3 90.6 89.9 89.3 87.4 86.1 84.8 87.3 93.0 97.8 113.4 113.5
R1024sC2 120.6 105.3 101.2 99.1 98.5 97.9 97.3 96.7 95.9 95.3 94.7 93.5 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R1024sD2 135.0 108.5 103.3 100.6 98.8 97.5 96.4 95.5 94.9 94.3 93.9 93.6 102.1 115.8 130.8 138.0 138.0 138.0 138.0
R1280sA2 114.9 99.7 96.1 94.8 93.9 93.1 92.5 91.6 90.8 90.1 89.5 87.2 85.9 84.7 87.2 92.8 97.4 108.6 110.3
R1280sC2 120.6 105.4 101.3 99.2 98.7 98.2 97.7 97.1 96.2 95.5 94.9 93.2 92.8 92.1 94.8 99.7 101.5 99.8 96.9
R1280sD2 135.7 109.4 104.3 101.5 99.7 98.3 97.2 96.3 95.5 94.9 94.4 92.8 94.0 101.7 112.6 124.2 136.9 138.0 138.0
R1536sA2 115.0 99.7 96.1 94.9 94.0 93.3 92.6 91.7 90.9 90.2 89.6 87.2 85.6 84.6 87.1 92.6 96.8 104.4 106.2
R1536sC2 120.6 105.4 101.3 99.3 98.8 98.3 97.9 97.3 96.5 95.7 95.1 93.3 92.4 91.8 94.7 99.6 101.4 99.8 96.9
R1536sD2 136.1 110.2 105.0 102.3 100.4 99.0 97.9 96.9 96.1 95.5 94.9 92.9 92.3 94.4 101.4 110.4 119.9 138.0 138.0
R2048sA2 115.0 99.7 96.1 94.8 94.0 93.2 92.6 91.7 90.9 90.1 89.5 87.1 85.4 84.2 86.2 90.4 94.7 100.6 102.0
R2048sC2 120.6 105.4 101.3 99.3 98.8 98.3 97.9 97.3 96.5 95.7 95.1 93.1 92.0 91.0 92.6 96.2 100.1 99.7 96.9
R2048sD2 136.3 110.4 105.2 102.5 100.6 99.1 98.0 97.0 96.2 95.5 94.8 92.6 91.3 90.7 91.2 94.1 99.8 128.9 138.0
R2304sA2 115.0 99.7 96.1 94.8 94.0 93.3 92.7 91.8 90.9 90.2 89.6 87.1 85.4 84.1 85.8 88.5 91.3 98.0 99.1
R2304sC2 120.6 105.4 101.3 99.3 98.8 98.4 98.0 97.5 96.6 95.9 95.2 93.2 92.0 90.9 92.2 93.9 96.7 99.7 96.8
R2304sD2 136.6 110.9 105.7 102.9 101.0 99.6 98.4 97.4 96.6 95.9 95.3 92.9 91.5 90.7 90.4 91.3 94.4 118.1 138.0
217 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
R384sD3 131.8 104.4 99.5 97.1 95.8 95.3 96.4 100.5 107.0 114.2 121.6 138.0 138.0 138.0 138.0 138.0 138.0 138.0 138.0
R512sD3 132.8 105.6 100.6 98.0 96.4 95.4 94.8 94.8 95.8 98.7 103.2 131.4 138.0 138.0 138.0 138.0 138.0 138.0 138.0
R768sD3 134.1 107.3 102.2 99.5 97.7 96.5 95.5 94.8 94.3 93.9 93.8 102.6 120.9 138.0 138.0 138.0 138.0 138.0 138.0
R1024sD3 135.0 108.5 103.3 100.6 98.8 97.4 96.4 95.5 94.8 94.3 93.9 93.6 102.1 115.8 130.8 138.0 138.0 138.0 138.0
R1280sD3 135.6 109.4 104.2 101.5 99.7 98.3 97.2 96.2 95.5 94.9 94.3 92.8 94.0 101.7 112.6 124.2 136.9 138.0 138.0
R1536sD3 136.1 110.1 105.0 102.3 100.4 99.0 97.8 96.9 96.1 95.4 94.8 92.9 92.3 94.4 101.4 110.4 119.9 138.0 138.0
R2048sD3 136.3 110.3 105.2 102.4 100.5 99.1 97.9 96.9 96.1 95.4 94.8 92.6 91.3 90.7 91.2 94.1 99.8 128.9 138.0
R2304sD3 136.6 110.8 105.6 102.9 101.0 99.5 98.3 97.4 96.5 95.8 95.2 92.9 91.5 90.7 90.4 91.3 94.4 118.1 138.0
R384sA4 114.9 99.2 95.3 93.7 92.6 92.1 92.3 92.4 91.8 91.0 90.4 87.9 86.2 84.8 87.3 93.1 98.1 123.1 115.4
R384sC4 120.6 104.6 100.2 98.1 97.3 96.9 97.3 98.2 97.6 96.9 96.2 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R512sA4 114.9 99.4 95.6 94.0 92.9 92.1 91.6 91.0 90.9 90.8 90.4 87.9 86.2 84.8 87.3 93.1 98.1 122.8 115.4
R512sC4 120.6 104.9 100.6 98.4 97.6 96.9 96.5 96.3 96.4 96.5 96.1 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R768sA4 114.9 99.5 95.9 94.5 93.4 92.6 91.9 91.0 90.3 89.8 89.3 87.9 86.1 84.8 87.3 93.0 98.0 119.3 115.2
R768sC4 120.6 105.2 101.0 98.9 98.1 97.5 96.9 96.3 95.6 95.1 94.7 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R1024sA4 114.9 99.6 96.1 94.7 93.7 92.9 92.2 91.3 90.5 89.9 89.3 87.4 86.1 84.8 87.3 93.0 97.9 115.1 114.3
R1024sC4 120.6 105.3 101.2 99.1 98.5 97.9 97.3 96.7 95.9 95.3 94.7 93.5 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R1280sA4 114.9 99.6 96.1 94.8 93.9 93.1 92.4 91.5 90.7 90.0 89.4 87.2 85.9 84.7 87.2 92.9 97.5 110.0 111.6
R1280sC4 120.6 105.3 101.2 99.2 98.7 98.1 97.6 97.0 96.2 95.5 94.9 93.2 92.8 92.1 94.8 99.7 101.5 99.8 96.9
R1536sA4 114.9 99.6 96.1 94.8 93.9 93.2 92.6 91.7 90.8 90.1 89.5 87.2 85.6 84.6 87.1 92.6 96.9 105.5 107.7
R1536sC4 120.6 105.3 101.3 99.3 98.8 98.3 97.9 97.3 96.4 95.7 95.1 93.2 92.4 91.8 94.7 99.6 101.4 99.8 96.9
R2048sA4 115.0 99.6 96.0 94.7 93.9 93.1 92.5 91.6 90.8 90.1 89.4 87.0 85.4 84.2 86.1 90.4 94.6 100.8 102.6
R2048sC4 120.6 105.3 101.2 99.2 98.7 98.3 97.9 97.3 96.4 95.7 95.0 93.1 92.0 91.0 92.6 96.2 100.1 99.8 96.9
R2304sA4 115.0 99.6 96.0 94.7 93.9 93.2 92.6 91.7 90.8 90.1 89.5 87.0 85.3 84.1 85.8 88.4 91.2 98.1 99.5
R2304sC4 120.6 105.3 101.2 99.2 98.8 98.4 98.0 97.4 96.6 95.8 95.2 93.2 92.0 90.9 92.1 93.9 96.7 99.7 96.8
218 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
R384sB5 120.4 104.6 100.2 98.1 97.3 96.9 97.3 98.2 97.6 96.9 96.2 94.4 93.4 92.1 93.9 98.2 102.0 129.0 121.3
R512sB5 120.4 104.9 100.6 98.4 97.6 96.9 96.5 96.3 96.3 96.5 96.1 94.4 93.4 92.1 93.9 98.2 102.0 129.0 121.3
R768sB5 120.4 105.1 100.9 98.8 98.0 97.4 96.8 96.3 95.6 95.1 94.7 94.4 93.4 92.1 93.9 98.2 102.0 128.7 121.3
R1024sB5 120.4 105.2 101.1 99.0 98.3 97.8 97.2 96.6 95.9 95.2 94.7 93.5 93.4 92.1 93.9 98.2 102.0 128.1 121.3
R1280sB5 120.4 105.2 101.1 99.0 98.5 98.0 97.5 96.9 96.1 95.5 94.8 93.2 92.8 92.1 93.9 98.2 101.9 126.1 121.2
R1536sB5 120.4 105.1 101.1 99.0 98.6 98.1 97.7 97.2 96.3 95.6 95.0 93.2 92.3 91.8 93.8 98.2 101.9 122.8 120.8
R2048sB5 120.4 105.0 100.9 98.9 98.4 98.0 97.6 97.1 96.2 95.5 94.9 93.0 91.9 90.9 92.2 95.7 100.2 115.8 118.4
R2304sB5 120.5 104.9 100.9 98.8 98.4 98.0 97.7 97.2 96.3 95.6 95.0 93.0 91.9 90.9 91.7 93.6 96.7 111.7 115.7
R384sA6 114.9 99.2 95.3 93.7 92.6 92.1 92.3 92.4 91.8 91.0 90.4 87.9 86.2 84.8 87.3 93.1 98.1 122.6 115.4
R384sC6 120.6 104.6 100.2 98.1 97.3 96.9 97.3 98.2 97.6 96.9 96.2 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R512sA6 114.9 99.4 95.6 94.1 92.9 92.1 91.6 91.1 90.9 90.8 90.4 87.9 86.2 84.8 87.3 93.1 98.0 120.2 115.2
R512sC6 120.6 104.9 100.6 98.4 97.6 96.9 96.5 96.3 96.4 96.5 96.1 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R768sA6 114.9 99.6 95.9 94.5 93.4 92.6 91.9 91.1 90.4 89.8 89.3 87.9 86.1 84.8 87.3 93.0 97.8 111.7 112.2
R768sC6 120.6 105.2 101.0 98.9 98.1 97.5 96.9 96.3 95.6 95.1 94.7 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R1024sA6 115.0 99.7 96.1 94.7 93.8 92.9 92.3 91.4 90.6 89.9 89.4 87.4 86.1 84.8 87.3 92.8 97.3 106.6 107.9
R1024sC6 120.6 105.3 101.2 99.1 98.5 97.9 97.3 96.7 96.0 95.3 94.7 93.5 93.4 92.1 94.8 99.7 101.4 99.8 96.9
R1280sA6 115.0 99.7 96.1 94.9 93.9 93.2 92.5 91.6 90.8 90.1 89.5 87.2 85.8 84.7 87.2 92.6 96.7 102.9 103.8
R1280sC6 120.6 105.4 101.2 99.2 98.7 98.2 97.7 97.1 96.3 95.5 94.9 93.2 92.8 92.1 94.8 99.6 101.3 99.8 96.9
R1536sA6 115.1 99.8 96.2 94.9 94.1 93.3 92.7 91.8 91.0 90.3 89.6 87.2 85.6 84.6 87.2 92.2 95.8 99.6 100.5
R1536sC6 120.6 105.4 101.3 99.3 98.8 98.4 97.9 97.3 96.5 95.8 95.1 93.2 92.3 91.8 94.7 99.5 101.2 99.7 96.9
R2048sA6 115.4 100.0 96.3 94.9 94.1 93.3 92.7 91.8 91.0 90.3 89.6 87.1 85.4 84.3 86.3 90.0 93.5 96.1 95.8
R2048sC6 120.7 105.4 101.3 99.3 98.8 98.4 98.0 97.4 96.5 95.8 95.1 93.1 91.9 91.0 92.6 96.1 99.9 99.6 96.8
R2304sA6 115.6 100.2 96.4 95.0 94.2 93.4 92.8 91.9 91.1 90.4 89.7 87.1 85.4 84.3 86.0 88.3 90.8 93.8 93.8
R2304sC6 120.7 105.4 101.3 99.3 98.9 98.5 98.1 97.5 96.7 95.9 95.2 93.2 92.0 91.0 92.2 93.8 96.6 99.4 96.7
219 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
R384sA7 114.9 99.2 95.3 93.7 92.6 92.1 92.3 92.4 91.8 91.0 90.4 87.9 86.2 84.8 87.3 93.1 98.1 123.1 115.4
R384sB7 120.6 104.6 100.2 98.1 97.3 96.9 97.3 98.2 97.6 96.9 96.2 94.4 93.4 92.1 93.9 98.2 102.0 129.0 121.3
R384sC7 120.6 104.6 100.2 98.1 97.3 96.9 97.3 98.2 97.6 96.9 96.2 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R384sD7 131.8 104.4 99.5 97.1 95.8 95.3 96.4 100.5 107.0 114.2 121.6 138.0 138.0 138.0 138.0 138.0 138.0 138.0 138.0
R512sA7 114.9 99.4 95.6 94.0 92.9 92.1 91.6 91.0 90.9 90.8 90.4 87.9 86.2 84.8 87.3 93.1 98.1 122.9 115.4
R512sB7 120.6 104.9 100.6 98.4 97.6 96.9 96.5 96.3 96.4 96.5 96.1 94.4 93.4 92.1 93.9 98.2 102.0 129.0 121.3
R512sC7 120.6 104.9 100.6 98.4 97.6 96.9 96.5 96.3 96.4 96.5 96.1 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R512sD7 132.8 105.6 100.6 98.0 96.4 95.4 94.8 94.8 95.8 98.7 103.2 131.4 138.0 138.0 138.0 138.0 138.0 138.0 138.0
R768sA7 114.9 99.5 95.9 94.5 93.4 92.6 91.9 91.0 90.3 89.8 89.3 87.9 86.1 84.8 87.3 93.1 98.0 120.5 115.3
R768sB7 120.6 105.2 101.0 98.9 98.1 97.5 96.9 96.3 95.6 95.1 94.7 94.4 93.4 92.1 93.9 98.2 102.0 128.8 121.3
R768sC7 120.6 105.2 101.0 98.9 98.1 97.5 96.9 96.3 95.6 95.1 94.7 94.4 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R768sD7 134.1 107.2 102.1 99.5 97.7 96.5 95.5 94.8 94.3 93.9 93.8 102.6 120.9 138.0 138.0 138.0 138.0 138.0 138.0
R1024sA7 114.9 99.6 96.0 94.7 93.7 92.9 92.2 91.3 90.6 89.9 89.3 87.4 86.1 84.8 87.3 93.0 97.9 117.3 114.8
R1024sB7 120.6 105.3 101.2 99.1 98.5 97.9 97.3 96.7 95.9 95.3 94.7 93.5 93.4 92.1 93.9 98.2 102.0 128.3 121.3
R1024sC7 120.6 105.3 101.2 99.1 98.5 97.9 97.3 96.7 95.9 95.3 94.7 93.5 93.4 92.1 94.8 99.7 101.5 99.8 96.9
R1024sD7 135.0 108.4 103.3 100.6 98.8 97.4 96.4 95.5 94.9 94.3 93.9 93.6 102.1 115.8 130.8 138.0 138.0 138.0 138.0
R1280sA7 114.9 99.6 96.1 94.8 93.9 93.1 92.5 91.5 90.8 90.1 89.5 87.2 85.9 84.8 87.3 92.9 97.7 113.2 113.3
R1280sB7 120.6 105.3 101.2 99.2 98.7 98.1 97.7 97.1 96.2 95.5 94.9 93.2 92.8 92.1 93.9 98.2 101.9 127.0 121.2
R1280sC7 120.6 105.3 101.2 99.2 98.7 98.1 97.7 97.1 96.2 95.5 94.9 93.2 92.8 92.1 94.8 99.7 101.5 99.8 96.9
R1280sD7 135.7 109.4 104.2 101.5 99.7 98.3 97.2 96.3 95.5 94.9 94.3 92.8 94.0 101.7 112.6 124.2 136.9 138.0 138.0
R1536sA7 115.0 99.7 96.1 94.8 93.9 93.2 92.6 91.7 90.9 90.2 89.6 87.2 85.6 84.6 87.2 92.8 97.3 108.8 110.3
R1536sB7 120.6 105.3 101.3 99.3 98.8 98.3 97.9 97.3 96.5 95.7 95.1 93.2 92.4 91.8 93.9 98.2 101.9 124.9 121.0
R1536sC7 120.6 105.3 101.3 99.3 98.8 98.3 97.9 97.3 96.5 95.7 95.1 93.2 92.4 91.8 94.7 99.6 101.4 99.8 96.9
R1536sD7 136.1 110.1 105.0 102.2 100.4 99.0 97.8 96.9 96.1 95.5 94.9 92.9 92.3 94.4 101.4 110.4 119.9 138.0 138.0
220 ITU-T Rec. G.991.2 (12/2003)

Noise
profile
1 10 20 30 40 50 60 70 80 90 100 150 200 250 300 350 400 600 800
R2048sA7 115.0 99.7 96.0 94.7 93.9 93.2 92.6 91.7 90.8 90.1 89.5 87.1 85.4 84.2 86.2 90.6 95.4 104.5 106.2
R2048sB7 120.6 105.3 101.2 99.2 98.8 98.3 97.9 97.3 96.5 95.7 95.1 93.1 92.0 91.0 92.2 95.8 100.4 121.1 120.2
R2048sC7 120.6 105.3 101.2 99.2 98.8 98.3 97.9 97.3 96.5 95.7 95.1 93.1 92.0 91.0 92.6 96.3 100.2 99.8 96.9
R2048sD7 136.3 110.3 105.1 102.4 100.5 99.1 98.0 97.0 96.2 95.5 94.8 92.6 91.3 90.7 91.2 94.1 99.8 128.9 138.0
R2304sA7 115.1 99.7 96.1 94.8 93.9 93.2 92.6 91.7 90.9 90.2 89.6 87.1 85.4 84.2 86.0 88.7 91.8 102.1 103.6
R2304sB7 120.6 105.3 101.2 99.2 98.8 98.4 98.0 97.5 96.6 95.9 95.2 93.2 92.0 91.0 91.8 93.7 96.9 116.6 118.9
R2304sC7 120.6 105.3 101.2 99.2 98.8 98.4 98.0 97.5 96.6 95.9 95.2 93.2 92.0 91.0 92.2 93.9 96.8 99.8 96.9
R2304sD7 136.6 110.8 105.6 102.9 101.0 99.5 98.4 97.4 96.6 95.9 95.3 92.9 91.5 90.7 90.4 91.3 94.4 118.1 138.0
Table IV.4/G.991.2 – STU-R side/asymmetric PSDs
Noise
profile
1 20 40 60 80 100 150 200 250 300 350 400 500 600 700 800 1000 1200 1400
R2048aA2 115.0 96.0 93.6 92.0 90.3 88.9 86.5 84.9 83.8 85.4 89.5 94.8 100.7 103.2 104.0 105.0 107.4 113.3 121.7
R2048aC2 120.6 101.1 98.4 97.2 95.6 94.2 92.2 91.0 90.1 91.2 94.8 99.8 101.7 99.8 98.0 96.9 95.5 97.3 99.6
R2048aD2 135.1 103.3 98.6 96.0 94.2 92.9 90.6 89.3 88.7 89.0 92.2 98.9 113.0 124.8 134.6 137.6 138.0 138.0 138.0
R2304aA2 115.0 96.1 93.9 92.6 90.8 89.4 87.0 85.3 84.1 85.6 87.9 91.2 97.7 99.6 100.2 101.0 102.6 108.2 116.8
R2304aC2 120.6 101.3 98.8 97.8 96.4 95.0 93.0 91.8 90.7 91.6 93.1 96.4 101.5 99.7 98.0 96.9 95.5 97.3 99.6
R2304aD2 136.2 105.0 100.4 97.8 95.9 94.6 92.2 90.8 89.9 89.6 90.4 93.8 106.4 118.0 129.7 136.4 138.0 138.0 138.0
R2048aD3 135.1 103.3 98.6 96.0 94.2 92.9 90.6 89.3 88.7 89.0 92.2 98.9 113.1 125.2 135.0 137.7 138.0 138.0 138.0
R2304aD3 136.2 105.0 100.3 97.7 95.9 94.5 92.2 90.8 89.9 89.6 90.4 93.8 106.4 118.2 130.1 136.6 138.0 138.0 138.0
R2048aA4 114.9 95.9 93.5 92.0 90.2 88.9 86.5 84.9 83.8 85.4 89.4 94.7 100.6 103.3 104.2 105.4 107.9 113.9 122.2
R2048aC4 120.6 101.1 98.4 97.1 95.6 94.2 92.2 91.0 90.1 91.2 94.8 99.8 101.7 99.8 98.0 96.9 95.5 97.3 99.6
221 ITU-T Rec. G.991.2 (12/2003)

Table IV.4/G.991.2 – STU-R side/asymmetric PSDs
Noise
profile
1 20 40 60 80 100 150 200 250 300 350 400 500 600 700 800 1000 1200 1400
R2304aA4 115.0 96.0 93.8 92.5 90.7 89.4 86.9 85.2 84.0 85.6 87.9 91.0 97.5 99.4 100.2 101.0 102.8 108.4 117.1
R2304aC4 120.6 101.2 98.7 97.8 96.4 94.9 92.9 91.7 90.7 91.6 93.1 96.4 101.5 99.7 98.0 96.9 95.5 97.3 99.6
R2048aB5 120.4 100.8 98.1 96.9 95.5 94.1 92.1 90.9 90.0 90.9 94.5 99.9 109.9 117.5 117.8 119.3 124.5 128.9 132.5
R2304aB5 120.4 100.9 98.4 97.6 96.2 94.8 92.8 91.7 90.6 91.3 92.9 96.3 106.7 113.3 115.1 116.9 121.1 126.2 131.8
R2048aA6 115.2 96.0 93.7 92.1 90.4 89.0 86.5 84.9 83.8 85.5 89.3 94.0 100.2 99.1 100.6 99.6 100.0 105.0 112.9
R2048aC6 120.6 101.1 98.4 97.2 95.6 94.3 92.2 91.0 90.1 91.2 94.7 99.6 101.7 99.7 98.0 96.8 95.5 97.3 99.6
R2304aA6 115.4 96.3 94.0 92.7 90.9 89.6 87.0 85.3 84.1 85.7 87.7 90.6 96.6 95.4 97.3 95.4 94.9 99.4 106.7
R2304aC6 120.7 101.3 98.8 97.9 96.4 95.0 93.0 91.7 90.7 91.7 93.1 96.3 101.4 99.6 98.0 96.8 95.5 97.3 99.6
R2048aA7 115.0 95.9 93.5 92.0 90.3 88.9 86.5 84.9 83.8 85.5 89.6 95.2 102.5 106.9 107.0 108.3 110.8 116.4 124.1
R2048aB7 120.6 101.1 98.4 97.1 95.6 94.2 92.2 91.0 90.1 91.0 94.5 100.1 111.2 122.3 119.5 120.5 125.9 129.7 132.6
R2048aC7 120.6 101.1 98.4 97.1 95.6 94.2 92.2 91.0 90.1 91.2 94.8 99.8 101.8 99.8 98.1 96.9 95.5 97.3 99.6
R2048aD7 135.1 103.3 98.6 96.0 94.2 92.9 90.6 89.3 88.7 89.0 92.2 98.9 113.6 127.0 135.9 137.8 138.0 138.0 138.0
R2304aA7 115.1 96.0 93.8 92.5 90.8 89.4 87.0 85.3 84.1 85.7 88.1 91.4 99.8 102.7 103.2 104.5 105.8 111.3 119.7
R2304aB7 120.6 101.2 98.7 97.8 96.4 95.0 92.9 91.7 90.7 91.4 93.0 96.5 108.3 118.2 118.3 119.4 124.0 128.5 132.3
R2304aC7 120.6 101.2 98.7 97.8 96.4 95.0 92.9 91.7 90.7 91.7 93.1 96.4 101.6 99.8 98.0 96.9 95.5 97.3 99.6
R2304aD7 136.2 105.0 100.3 97.7 95.9 94.6 92.2 90.8 89.9 89.6 90.4 93.9 106.7 119.4 131.9 137.2 138.0 138.0 138.0
222 ITU-T Rec. G.991.2 (12/2003)

BIBLIOGRAPHY
[B1] ITU-T Recommendation G.961 (1993), Digital transmission system on metallic local lines
for ISDN basic rate access.
[B2] ITU-T Recommendation G.995.1 (2001), Overview of digital subscriber line (DSL)
Recommendations.
[B3] ANSI X3.4-1986 (R1997), Information Systems – Coded Character Sets – 7-Bit American
National Standard Code for Information Interchange (7-Bit ASCII).
[B4] ITU-T Recommendation K.50 (2000), Safe limits of operating voltages and currents for
telecommunication systems powered over the network.
[B5] Telcordia Technologies, GR-1089-CORE: Electromagnetic Compatibility and Electrical
Safety Generic Criteria for Network Telecommunications Equipment, February 1999.
[B6] ITU-T Recommendation G.704 (1998), Synchronous frame structures used at 1544, 6312,
2048, 8448 and 44 736 kbit/s hierarchical levels.
[B7] ANSI/IEEE Std 455-1985: IEEE Standard Test Procedure for Measuring Longitudinal
Balance of Telephone Equipment Operating in the Voice Band.
[B8] ITU-T Recommendation O.9 (1999), Measuring arrangements to assess the degree of
unbalance about earth.
[B9] IETF RFC 2495 (1999), Definitions of Managed Objects for the DS1, E1, DS2 and E2
Interface Types.
[B10] ITU-T Recommendation I.431 (1993), Primary rate user-network interface – Layer 1
specification.
[B11] ITU-T Recommendation G.996.1 (2001), Test procedures for Digital Subscriber Line
(DSL) transceivers.
[B12] The ATM Technical Forum Committee, af-phy-0086.001: Inverse Multiplexing for ATM
(IMA) Specification, Version 1.1, March 1999.
[B13] ISO/IEC 13239:2002, Information technology – Telecommunications and information
exchange between systems – High-level data link control (HDLC) procedures.
[B14] Telcordia, GR-303-CORE Issue 4 – Integrated Digital Loop Carrier System Generic
Requirements, Objectives and Interface, December 2000.
[B15] The ATM Forum af-vmoa-0145.000: Voice and Multimedia over ATM – Loop Emulation
service Using AAL2, July 2000.
[B16] ETSI ETS 300 347-1: V interfaces at the digital Local Exchange (LE); V5.2 interface for
the support of Access Network (AN); Part 1: V5.2 Interface specification – September 1994.
ITU-T Rec. G.991.2 (12/2003) 223

SERIES OF ITU-T RECOMMENDATIONS
Series A Organization of the work of ITU-T
Series B Means of expression: definitions, symbols, classification
Series C General telecommunication statistics
Series D General tariff principles
Series E Overall network operation, telephone service, service operation and human factors
Series F Non-telephone telecommunication services
Series G Transmission systems and media, digital systems and networks
Series H Audiovisual and multimedia systems
Series I Integrated services digital network
Series J Cable networks and transmission of television, sound programme and other multimedia signals
Series K Protection against interference
Series L Construction, installation and protection of cables and other elements of outside plant
Series M TMN and network maintenance: international transmission systems, telephone circuits,
telegraphy, facsimile and leased circuits
Series N Maintenance: international sound programme and television transmission circuits
Series O Specifications of measuring equipment
Series P Telephone transmission quality, telephone installations, local line networks
Series Q Switching and signalling
Series R Telegraph transmission
Series S Telegraph services terminal equipment
Series T Terminals for telematic services
Series U Telegraph switching
Series V Data communication over the telephone network
Series X Data networks and open system communications
Series Y Global information infrastructure, Internet protocol aspects and Next Generation Networks
Series Z Languages and general software aspects for telecommunication systems
Printed in Switzerland
Geneva, 2004

T Rec G.991.2 200312 I!!pdf e

Uploaded by

Copyright:

Available Formats

T Rec G.991.2 200312 I!!pdf e

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

T Rec G.991.2 200312 I!!pdf e

Uploaded by

Copyright:

Available Formats

INTERNATIONAL TELECOMMUNICATION UNION

SERIES G: TRANSMISSION SYSTEMS AND MEDIA,

Single-pair high-speed digital subscriber line

ITU-T Recommendation G.991.2

INTERNATIONAL TELEPHONE CONNECTIONS AND CIRCUITS G.100–G.199

For further details, please refer to the list of ITU-T Recommendations.

Single-pair high-speed digital subscriber line (SHDSL) transceivers

ITU-T Rec. G.991.2 (12/2003) i

INTELLECTUAL PROPERTY RIGHTS

ii ITU-T Rec. G.991.2 (12/2003)

ITU-T Rec. G.991.2 (12/2003) iii

iv ITU-T Rec. G.991.2 (12/2003)

ITU-T Rec. G.991.2 (12/2003) v

Single-pair high-speed digital subscriber line (SHDSL) transceivers

ITU-T Rec. G.991.2 (12/2003) 1

3 Definitions and abbreviations

2 ITU-T Rec. G.991.2 (12/2003)

ITU-T Rec. G.991.2 (12/2003) 3

4 ITU-T Rec. G.991.2 (12/2003)

ITU-T Rec. G.991.2 (12/2003) 5

4.1 STU-x functional model

Figure 4-1/G.991.2 – STU-x functional model

6 ITU-T Rec. G.991.2 (12/2003)

4.2 User plane protocol reference model

Transport protocol (e.g., SDH) Transport protocol (e.g., SDH)

User data interface Physical transmission media Internal interface

Figure 4-2/G.991.2 – User plane protocol reference model

4.3 Application models

Figure 4-3/G.991.2 – Application model

ITU-T Rec. G.991.2 (12/2003) 7

6 PMD layer functional characteristics

6.1 Data mode operation

Figure 6-1/G.991.2 – Data mode PMD reference model

8 ITU-T Rec. G.991.2 (12/2003)

Figure 6-2/G.991.2 – Block diagram of the TCM encoder

6.1.2.1 Serial-to-parallel converter

ITU-T Rec. G.991.2 (12/2003) 9

Table 6-1/G.991.2 – Mapping of bits to PAM levels

10 ITU-T Rec. G.991.2 (12/2003)

Figure 6-4/G.991.2 – Block diagram of the channel precoder

6.1.4 Spectral shaper

ITU-T Rec. G.991.2 (12/2003) 11

Table 6-2/G.991.2 – Required power backoff values

6.2 PMD activation sequence

Figure 6-5/G.991.2 – Activation reference model

12 ITU-T Rec. G.991.2 (12/2003)

Figure 6-6/G.991.2 – Timing diagram for activation sequence

Figure 6-6a/G.991.2 – G.991.2 total activation sequence

ITU-T Rec. G.991.2 (12/2003) 13

14 ITU-T Rec. G.991.2 (12/2003)

ITU-T Rec. G.991.2 (12/2003) 15

Table 6-4/G.991.2 – Bit-to-level mapping

6.3 PMD pre-activation sequence

16 ITU-T Rec. G.991.2 (12/2003)

Figure 6-8/G.991.2 – Pre-activation reference model

G.994.1 tps tph

Figure 6-9/G.991.2 – Typical timing diagram for pre-activation sequence

ITU-T Rec. G.991.2 (12/2003) 17

6.3.2.1 Signal Pri