Chapter 2

ISDN’s Physical Layer - BRI

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Table of Contents
Purpose ................................................................................................................................2
The Physical Layer of ISDN ...............................................................................................3
Multiple Physical Layer Technologies ...............................................................................5
Wiring Scenarios (S/T) .......................................................................................................7
Extended Passive Bus .........................................................................................................9
NT1 Star Wiring ...............................................................................................................11
An ISDN Residence ..........................................................................................................13
Electrical Configuration at S/T .........................................................................................15
Power Feeding ..................................................................................................................17
Power Source 1 Consumption Values ...............................................................................19
Pseudo Ternary Code Example at S/T ..............................................................................21
Frame Structure at Reference Points S/T ..........................................................................23
Frame Usage Example ......................................................................................................25
How Multipoint Synchronous Works ...............................................................................27
NT1 Derivation of “Echo” Channel ..................................................................................29
TE Access to D-Channel (S/T) .........................................................................................31
D-Channel Contention (S/T) .............................................................................................33
The Physical Layer Multiplexor .......................................................................................35
Derivation of Q and S bits (S/T) .......................................................................................37
Location and Characteristics of Loopbacks ......................................................................39
Codes for Q and S1 Channel Messages ............................................................................41
Codes for S2 Channel Messages .......................................................................................43
Layer 1 - NT1 to Exchange ...............................................................................................45
Echo Cancellation Theory .................................................................................................47
U-Interface Line Code ......................................................................................................49
Basic Loop Frame at U .....................................................................................................51
Superframe Format - (U) ..................................................................................................53
DSL Framer - Functional Description ..............................................................................55
DSL Framing Relationships ..............................................................................................57
Activation Initiated by the Exchange ................................................................................59
Activation Initiated by the Terminal Equipment ..............................................................61
Worst Case Basic Access Configuration ..........................................................................63

Page 2 - 1  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Purpose

T he purpose of this chapter is to completely cover the Layer 1 - Physical Layer - aspects of ISDN.
Layer 1 is common to all channels of the ISDN user-to-network interface and provides the
multiplexing functions for both interfaces.

Since there are two separate technologies involved in Layer 1; one for the subscriber side of the
network terminating equipment, and one for the network side of the network terminating
equipment, they will be covered separately.

The user side is a full duplex, synchronous system operating at a rate of 192 kbit/s. In its point-to-
multipoint mode of operation, it presents some unique features.

The network side is a full-duplex, synchronous system operating at 160 kbit/s in a straight point-
to-point mode. However, the fact that it does this at 160 kbit/s on a single pair also presents some
innovative technology.

This chapter is devoted to the Basic Rate Interface, the so-called 2B + D interface. The next chapter
is devoted to the Primary Rate Interface.

 1998 - R. E. Blackshaw Page 2 - 2

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

The Physical Layer of ISDN

F or anyone used to regarding the Physical Layer as just a stream of bits, ISDN will come as a
shock. ISDN’s Physical Layer deals in 48 bit synchronized frames. It provides basic time division
multiplexing to derive the two B-Channels and the D-Channel (in both BRI and PRI) and what
might be called “data sense multiple access with collision resolution” to share the D-Channel
equitably among several terminal devices.

There are two Physical Layer protocols; one on the user side of the network terminating device
(NT1), and one on the network side (for BRI).

Finally, ISDN has reduced the number of terminal-network interfaces to just two, and this for a
network that carries voice, data, and video.

The ITU-T is continuing work on a Broadband ISDN which will create at least one more interface,
and possibly two. Broadband ISDN has user-network interface (UNI) data rate of 155 Mbit/s user
to network, and a rate of 600 Mbit/s network to user. Work by the ATM Forum has defined other
UNI rates as well as Network - Network Interface (NNI) and even a Private Network - Network
Interface (PNNI).

The two ISDN interfaces are described in detail in this chapter. B-ISDN will be dealt with later in
Chapter 9.

Page 2 - 3  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

The Physical Layer of ISDN

New protocol

Does multiplexing

- time division multiplexing

- a unique data sense multiple access

Two protocols

- one for user side of NT1

- one for digital subscriber line

Reduced interface set

- maximum of three (maybe four)

 1998 - R. E. Blackshaw Page 2 - 4

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Multiple Physical Layer Technologies

T he Physical Layer has three different technologies. However, the one internal to the ISDN
switch is of no concern to us. There was a move in the ITU-T to define another reference point
between the Line Termination (LT) and the switch itself (usually referred to as ET in the diagrams),
called the V reference point. The reason was that network operators would have had the switch
manufacturers in the position where they could change the loop technology using any
manufacturer’s LT and NT1 devices.

The Digital Subscriber Line (DSL) is a 2-wire, synchronous, full-duplex point-to-point system.
The ANSI-T1 Standard (T1.601) has specified the use of Echo Cancellation (EC) with decision
feedback equalization and a quaternary symbol line code. More about this later.

Recommendation I.430 (ANSI T1.605 in the United States) specifies the user premises technology,
which is a 4-wire (minimum) full-duplex, point-to-multipoint synchronous system operating at a
rate of 192 kbit/s. We will deal with this protocol first.

The ITU-T has a Recommendation containing three annexes which describes three different
methods of operating the local loop between the NT and the exchange switch, since there has not
been world-wide agreement. In the United States ANSI T1.601 is the operational standard and is
the basis for discussion in this workbook.

Page 2 - 5  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Multiple Physical Layer Technologies

User premises Digital subscriber line Switch

technology internal

TE NT LT

4-wire 2-wire ?

Point-to- Point-to
multipoint point

 1998 - R. E. Blackshaw Page 2 - 6

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Wiring Scenarios (S/T)

First let us examine the wiring scenarios for the user premises supported by ISDN.
There is of course the usual point-to-point scenario of a single Terminal Equipment (TE) attached
to the Network Termination (NT). In this configuration, the wiring length may be up to 1 km
(3280 ft.) with a maximum cable attenuation of 6db at 96 kHz. We say usual, since this wiring
scenario is very similar to the current analog telephone network customer premises wiring. Prior
to the advent of the telephone plug and jack most households were simply wired from the protector
to the single telephone set. Even today, PABXs are wired from the network termination point
directly to a single PABX trunk.

The next scenario is the short passive bus, expected to be a very popular configuration. This allows
attachment of up to eight terminal devices to a single network terminating device. In this
configuration, the wiring length is approximately 100-200 meters (328-656 ft.) with the TEs
connected at random points along the cable. The 100 meter length applies to low impedance cable
(Zc=75 ohms) and the 200 meter length applies to high impedance cable (Zc=150 ohms). This
multidrop scenario could apply to a private residence, being not much different than the current
situation in households that have multiple jack appearances. However, it is more likely to apply
in the business environment. A passive bus arrangement in a corporate employee’s work space,
providing connection for telephone, personal computer, facsimile machine, etc. will soon be the
norm.

At this point you may wonder how a point-to-multipoint configuration can operate in a full-duplex,
synchronous manner. Bear with us and we will cover it later in this chapter.

Page 2 - 7  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Wiring Scenarios (S/T)

Point-to-point Wiring

T T NT
R R

TE

Short Passive Bus Wiring - 100 to 200 meters

T T NT
R R

TE TE TE Up to Eight TEs

 1998 - R. E. Blackshaw Page 2 - 8

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Extended Passive Bus

A nother passive bus scenario is the extended passive bus, similar to the short form, but designed
for inter-building arrangements. This would most likely be on the terminal side of PABXs that
serve corporations that have multi-building campus-like environments.

This configuration extends from 100 to 1000 meters (328 to 3280 ft.). However, rather than
random positioning, the terminal devices must be clustered in the last 25 to 50 meters (82 to 164
ft.) of the cable. The 1984 Red Book limitation (if it is a limitation) of eight terminals maximum
was modified in the 1988 Blue Book version of Recommendation I.430. It now states that the
number of TEs connected to the cable is to be determined by individual administrations.

This far-end clustering is quite logical, as well as electrically necessary, since the greater part of
the cable length would be in the run from the PABX to the connected building. Clustering the
terminals within an 82 to 164 foot spacing is not really restrictive.

Page 2 - 9  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Extended Passive Bus (S/T)

Extended Passive Bus Wiring

T T NT
R R

TE TE TE * See text

Length is between 100 and 1000 meters,

with TEs clustered in last 25 - 50 meters

 1998 - R. E. Blackshaw Page 2 - 10

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

NT1 Star Wiring

A dded during a recent study period is the NT1 Star wiring scenario. In this configuration, it is
possible to wire more than eight terminals point-to-point to the network terminating device.
Because of the S/T reference point protocol, the network terminating device must buffer the
transmissions from the terminals. This buffering is necessary to provide proper operation of the D
echo bits.

Note: The individual TEs may connect to different ports on the NT, or you may wire some in a
passive bus arrangement to a single port.

This may seem like a needless complication and expense, because of the extra wiring and
buffering, but it does allow the network terminating device to perform a useful function.

Under protocol rules, when a B-Channel is assigned to a terminal, it has sole possession of that
channel. This does not permit a very familiar occurrence that we have on today’s telephone
network (referred to technically as bridging). If a user picks up a telephone on a given line, it is
quite possible to pick up any other telephone on that same line and converse, albeit with some
transmission degradation. Not only the protocol rules, but also digital voice transmission prevents
this from happening. Digital telephones cannot bridge as analog telephones do. The NT1 Star
permits inclusion of a digital bridge within the NT1 to permit the two-on-a-line function. The
alternative would be to attach both B-Channels to a digital bridge inside the ISDN switch, thus
occupying both B-Channels.

The NT1 Star would also find application in business, where a single NT1 located in the telephone
cabinet, or room, serves many work centers or offices and can provide a single power source for all.

As a general rule, any configuration that can meet the requirements of the S/T interface is quite
legitimate.

Page 2 - 11  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

NT1 Star Wiring (S/T)

T
TE R

T
R

Multiple
NT
TEs

T
R

T
TE R

 1998 - R. E. Blackshaw Page 2 - 12

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

An ISDN Residence

T his seems an appropriate place to introduce an impression of a possible ISDN residence wiring
scenario.

The diagram is a passive bus arrangement with a couple of additions.

First, we have a telephone, or several telephones, a personal computer, and terminal device(s) that
have not even been built yet.

The new twist that requires more standardization work is the telemetry control device, after all -
why build the technology into every device that will use telemetry when a single standard could
provide one black box that serves all devices?

The other area of standardization* is the home bus. The home bus is a power line carrier system
presently used in some forms of whole-house light and appliance control, e.g., the BSR X-10
system. Could we expand this to include the increasing number of smart appliances with VLSI
processor controls, or include some means of connection to our already processor controlled
automobiles?

This could permit some form of centralized diagnosis by a few expert service people and perhaps
reduce the costs of shotgun repairs, i.e., where the mechanic or repairman isn’t sure what part is
defective and replaces them all until the device starts working again.

* EIA has completed the standard for a Consumer Electronics Bus (CEBus) that will operate
over AC power lines, among other media.

Page 2 - 13  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

An ISDN Residence

Short Passive Bus Loop

NT1

Fax PC Tel

Telemetry
Control

“Home Bus”
To Appliances,
Automobile,
Alarm Gas Electric etc.
Controller Meter Power
Meter

 1998 - R. E. Blackshaw Page 2 - 14

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Electrical Configuration at S/T

T his drawing shows the electrical configuration specified for the S/T reference point of the basic
rate interface. The numbers refer to the pole (pin) numbers of the RJ45 plug at the device end of
the connecting cords. ISDN has specified the use of the RJ45 jack and plug which is the eight pole
version of the very familiar RJ11 (Mini-Mod) telephone jack and plug.

The four wires represented in solid black are mandatory for an ISDN connection, i.e., you must at
least have a transmit and receive pair. In this event the NT can provide power to the terminal(s)
over the center taps of the transformer windings. Only the line side of the transformers are shown
for clarity.

The network terminating device can provide a second power source, power source 2, in which case
you must provide the third pair (poles 7 & 8).

The third power source is in at least one terminal equipment and provides power source 3 over the
fourth pair (poles 1 & 2). The ITU-T says very little about this power source, merely stating that
its standardization is the province of other standards bodies. The ANSI-T1 standard shows a power
sink 3 inside the NT and states that TEs may use power source 3 to provide power to the NT or to
other TEs.

Page 2 - 15  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Electrical Configuration (S/T)

TE NT

1 1
Power Power
Source Sink
3 2 2 3
+ 3 3
+
Tra Rec
- 6 6
Power Power
Sink Source
1 1
5 5 -
Rec Tra -
4 4 +

Power 7 7 Power
Sink Source
2 2
8 8

 1998 - R. E. Blackshaw Page 2 - 16

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Power Feeding

A n interesting aspect of ISDN is that the Recommendation does not require the network to
provide power to the user-to-network interface. The reason for this may well be that many
administrations and other network operators are seriously considering the application of fiber
optics between the users and the network.

For whatever reason, the Recommendation (I.430) does not require that the network supply power
to the NT device.

In addition, there are requirements that should either the network or the NT supply the power, there
are specific restrictions regarding its consumption. For the so-called thread-of-life case, i.e.,
during commercial power outages, selected terminals will be designated to maintain some
communications. Even the designated terminal must not draw normal power, but must reduce its
consumption to specific levels.

Another point bearing upon terminal design is that polarity between the TE and NT is not
guaranteed. Power polarity could be reversed in certain instances. Also, when the NT wishes to
signal a reduced power state, it reverses the polarity of the power source. Thus, it is certain that
the terminals must be able to cope with any power polarity.

Because of these rules, designers who wish to develop truly universal terminal equipment should
plan on having the terminal provide its own power source. Such a terminal will then be capable of
operating with any ISDN.

For commercial situations, it is probable that if the NTs are grouped in a telephone room or cabinet,
they will supply power to the terminals at least for commercial power outage situations. For
residential terminals it seems only sensible that some form of backup power, e.g., nickel cadmium
cells or sealed lead acid cells, etc., will need to be provided in order to maintain the thread-of-life
aspect that has always been associated with the telephone.

Page 2 - 17  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Power Feeding

TEs NTs ISDN

Yes Probable Unlikely

Applicable Standard
Self-powered for business does not
TEs are universal situations require it

 1998 - R. E. Blackshaw Page 2 - 18

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Power Source 1 Consumption Values

R egardless of how the power is supplied, terminal devices are limited in the amount of power
that they may draw. This may seem strange, considering that the networks are unlikely to be
supplying the power in the first place. However, these design limits allow NT1 manufacturers to
know in advance the size of power supply they will need.

Some terms need definition:

Restricted power conditions refer to the situation where the local electrical power has failed and
the power is being provided by some backup means. This backup is probably battery supplied and
will not last indefinitely if wasted.

A designated terminal device refers to one terminal on the bus that maintains the thread-of-life
capability for that bus. In a residential case this would be a voice telephone, but in a business case
may be some other terminal vital to the business.

The local action state refers to a terminal that is deactivated as far as the network is concerned.
However, this terminal may be performing some local function such as modification of pre-stored
dialing numbers, etc.

The power ranges seem a bit strange for solid state devices. However, the high value of 56.5 volts
covers network provision of power. Exchange voltages are nominally 48 volts today, but do range
a little higher than that. In the normal condition, the low of 24 volts represents a large drop, but 24
volts is also a normal voltage to be found in exchange units. Under the restricted power condition,
the minimum is 32 volts. What is interesting to designers of terminal devices is the technique of
signaling that the power has gone to the restricted condition. This is done by reversing the polarity
of the power source. Some solid state devices, unless protected, will self destruct under reversed
polarity conditions.

Page 2 - 19  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Power Source 1 Consumption Values

Normal Conditions
Maximum
TE Type and State
Consumption

TE drawing power from Rated NPCU

PS1 in Active State
Note 1
TE drawing power from
PS1 in Deactivated 100 mwatt
State
Note 1: Rated NPCU not to exceed
TE drawing power from 80. For a rated NPCU > 10
Rated NPCU TE portability is not guar-
PS1 in Local Action
State Notes 1 & 2 anteed.

Note 2: Providing that PS1 can pro-

vide this power.
Restricted Conditions
Maximum Note 3: Rated RPCU not to exceed
TE Type and State 4.
Consumption
Designated TE drawing NPCU Normal Power Consump-
power from PS1 in Rated RPCU tion Unit = 100mW.
Active State Note 3
RPCU Restricted Power Consump-
Designated TE drawing tion Unit = 95mW
power from PS1 in 25mwatt
Deactivated State

Non-designated TE 0 mwatt
drawing power from PS1

Designated TE drawing
power from PS1 in Rated RPCU
Local Action State Notes 2 & 3
Locally powered TE
using connected detec- 3 mwatt
tor in any State
Locally powered TE not
using connected detec- 0 mwatt
tor in any State

 1998 - R. E. Blackshaw Page 2 - 20

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Pseudo Ternary Code Example at S/T

While the transformers provide isolation from the network, their presence in a digital direct-
current interface requires special considerations. Transformer cores will saturate magnetically in
the presence of direct current, requiring a line code that will eliminate residual direct current to the
extent possible.

The pseudo ternary line code, along with the coding rules, achieves this end. Although it is a tri-
state code, hence the name, it only conveys binary information. A no-voltage state on the line
represents binary ones, while positive or negative voltages represent binary zeroes.

The encoding rule stipulates that binary zeroes in the digital stream of information must be sent as
alternating positive or negative pulses. This has the effect of cancelling out residual direct current
on the line. This is reinforced using balancing bits in the physical layer frames, as will be seen
shortly.

The balancing bit rule is that if device transmits an even number of binary zeroes, the balancing bit
is set to binary one, i.e., no-signal. If a device transmits an unequal number of binary zeroes, the
balancing bit is set to a binary zero of the polarity required by the encoding rule. Because
alternating polarities must be used, there will be equal and opposing direct current voltages sent in
any given frame, resulting in little or no residual voltage.

Page 2 - 21  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Pseudo-Ternary Code Example (S/T)

Binary 0 1 0 0 1 1 0 0 0 1 1

Signal 0

 1998 - R. E. Blackshaw Page 2 - 22

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Frame Structure at Reference Points S/T

T his diagram illustrates the two types of Physical Layer frames used at the S/T reference points
of the basic rate interface. There are 48 bits in each frame, and each frame is transmitted in 250
microseconds, giving a transmission rate of 192 kbit/s. The pulse width of a single bit is 5.2
microseconds.

Each B-Channel receives two octets (16 bits) in each frame and the D-Channel receives 4 bits. The
frame structure groups B-Channel bits into octets and the D-Channel bits are individually placed
in the frame.

Both frames have DC balancing bits (L) and framing bits (F & Fa). There is an activation bit (A)
in the NT to TE frames, used in some countries to activate the Physical Layer. The NT to TE frame
also contains a multiframing bit (M) and a spare bit (S). The NT to TE frames also contain echoes
of the received D bits (E). These echoed D bits are used in the contention resolution scheme. The
NT is the only device transmitting in the NT to TE direction an so only needs a single balancing
bit. In addition, its D bits can be of either polarity.

Note that in the TE to NT frames the D bits are always either negative binary zeroes or binary ones,
but never positive binary zeroes. Since as many as eight terminals could be sending simultaneous
frames on the D-channel it is essential to ensure that their signals be of the same polarity.

Effectively, each E bit is an echo of the nearest D bit received by the NT from the TE(s). The E bit
in position 24 of an NT to TE frame is an echo of the D bit received in position 12 of the incoming
TE to NT frame, etc. The TE gets its clocking from incoming NT frames, and displaces its outgoing
frames two bit times from the incoming frames.

A violation of the encoding rule accomplishes framing, i.e., F bits are sent as the same polarity as
the last binary zero that precedes the framing bit, causing a line code violation. To ensure rapid
frame synchronization, the first binary zero following the FL bit pair should also cause a code
violation. For cases where the B1 Channel is idle, the Recommendation provides for a frame
auxiliary bit (Fa). This causes a second code violation which will occur in 14 or fewer bit times.
The upper diagram shows the Layer 1 frames as the ITU-T does, the lower figure puts the frames
in a more conventional representation.

Note the D-Channel echo bits (E) in the NT to TE frames. These echo bits are the solution scheme
for the D-Channel, which we will describe later. Among the eight possible terminal devices, it is
only the D-Channel that has this contention. When a B-Channel is assigned to a terminal device it
is assigned to that one device, for that call, and no other terminal device will access that channel.

You will note that since there are two octets for each B-channel in each frame that the 64 kbit/s
PCM voice band rate is maintained, i.e., 16 bits in 250 µseconds = 64 kbit/s. If you recall the RDTD
(Restricted Differential Time Delay) attribute of 50 ms, we can this comes from the LDLFaL 5-bit
sequence between B1 and B2 octets. At a nominal rate of 5.2 ms per bit, a 26 ms offset between
channels will occur at each end of a call.

Page 2 - 23  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Frame Structure at Ref. Points S/T

D L• F L• B1 B1 B1B1 B1 B1 B1 B1 E D A Fa N B2 B2 B2 B2B2B2B2 B2 E D M B1 B1 B1 B1 B1 B1 B1B1 E D S B2 B2 B2 B2 B2B2 B2 B2 E D L• F L•

48 bits in 250 microseconds

2 bit times offset

D L• F L• B1 B1 B1B1 B1 B1 B1 B1 L• D L• Fa L• B2 B2B2 B2 B2 B2B2 B2 L• D L• B1B1 B1B1 B1B1 B1 B1 L• D L• B2 B2 B2 B2B2B2 B2 B2 L• D L• F L•

F = Framing bit
L = DC balancing bit
D = D-Channel Bit
E = D (Echo) Channel bit
Fa = Auxiliary Framing bit
N = Set to binary value N ≠ Fa (NT to TE)
B1 = Bit within B-Channel 1
B2 = Bit within B-Channel 2
A = Bit used for Activation
M = Multiframing bit
S = S-Channel bit

The upper frame represents frames sent from NT to TE

The lower frame represents frames sent from TE to NT
The dots (•) demarcate parts of a frame that are
independently DC-balanced

NT TE 48 bits in 250 microseconds

FL B1 E D A FaN B2 E DM B1 EDS B2 EDL

TE NT 48 bits in 250 microseconds

FL B1 L D L FaL B2 LDL B1 LDL B2 LDL

 1998 - R. E. Blackshaw Page 2 - 24

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Frame Usage Example

T his diagram shows an example of an actual bit pattern of an INFO 2 frame. The INFO n frames
are special frames used during Physical Layer initialization. Note the first line code violation
caused by the F bit being of the same polarity as the last binary zero bit of the preceding frame.
Because an INFO 2 frame calls for all binary zeroes in both B-Channels and the D-Channel, the
second code violation occurs with the very first B1 bit. Because there have already been two code
violations, in this frame, they do not use the Fa bit to cause another code violation and it is simply
set to a normal (legitimate) binary zero.

The ITU-T uses what may seem a strange convention for diagramming these frames, but there is
meaning in every line of the frame. The center line represents a no-voltage, or binary one state. The
upper line represents a positive voltage binary zero, and the lower line a negative voltage binary
zero. Wherever there are horizontal lines in the diagram it means that the bit it represents may be
of whatever value that line represents. If there is no line, then it is impossible for the bit to assume
that value as long as it observes the coding rule. The three lower diagrams illustrate this fact.

All TEs should begin with a positive binary zero, simply because a legitimate frame will end
(somewhere) with a positive binary zero. If you count the legitimate positive binary zero positions
and the legitimate negative binary zero positions, you will see that there is one more negative
position than positive position. This means that the final balancing bit (L), if used, must be a
positive binary zero. If the balancing bit is not used, it means that the last zero bit was a positive
binary zero.

Page 2 - 25  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Frame Usage Example

First code violation

D L• F L• B1 B1 B1 B1 B1 B1 B1 B1 E D A Fa N B2 B2 B2 B2B2 B2B2 B2 E D M B1 B1 B1 B1 B1 B1 B1 B1 E D S B2 B2 B2 B2B2 B2B2 B2 E D L• F L•

Second code violation 48 bits in 250 microseconds

An INFO 2 Frame

Indicates that only a binary one is legitimate

Indicates that only binary one or negative binary zero is legitimate

Indicates that all pseudo-ternary values are legitimate

Indicates that only positive zeroes are legitimate

Indicates that only negative zeroes are legitimate

 1998 - R. E. Blackshaw Page 2 - 26

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

How Multipoint Synchronous Works

F or anyone used to modem technology, the idea of three terminals in a point-multipoint

configuration all transmitting simultaneously may seem a recipe for disaster. But this is digital
technology and not subject to the same rules and restrictions. The diagram illustrates the actual bit
patterns of three terminal devices (TEa, TEb, and TEc) transmitting simultaneously on the same
D-Channel. The lowest figure illustrates what bit pattern the NT device will observe.

All TEs must send framing bits, but as we have seen, these framing bits are always a positive binary
zero. Note that the NT is insensitive to pair reversals and could see the framing bit as a negative
zero, yet still lock on to the frames.

Terminal TEa is sending on the B1-Channel, HDLC flag characters as it happens (01111110).
Since each octet contains exactly two binary zeroes, there is no requirement for TEa to use the
balancing bits (L). You will note that this is why the TE-to-NT frames have more balancing bits
because the rules require each terminal to balance the portion of the frame that it is using.

TEb is using the B2-Channel, and because it sends nothing in the B1 bit positions, it requires use
of the frame auxiliary bit (Fa). As it happens, the B2 bit pattern contains equal numbers of zeroes,
so there is no use of balancing bits.

TEc is sending on the D-Channel, and as a consequence is doing quite a lot of bit balancing since
every D-Channel bit that happens to be a binary zero will require balancing.

Because of the type of signal generators required by the standard, two generators sending the same
signal, for example a positive zero, results in a signal at the NT of the same magnitude as would
be the case from a single generator. Thus, the result that the NT sees on the line is simply the three
bit patterns overlaid.

The lower diagram is simply a different representation of the frame as the NT sees it with shading
used to show the portion each TE is using. This type of representation may be more familiar to
data-processing people.

Page 2 - 27  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

How Multipoint Synchronous Works

D L• F L• B1 B1 B1 B1B1B1 B1 B1 L• D L• Fa L• B2 B2 B2 B2 B2 B2 B2B2 L• D L• B1B1B1 B1B1 B1 B1 B1 L• D L• B2B2 B2B2 B2 B2 B2B2 L• D L• F L•

TEa is sending on the B1 Channel

D L• F L• B1 B1 B1 B1B1 B1 B1B1 L• D L• Fa L• B2 B2 B2 B2B2 B2 B2 B2 L• D L• B1B1 B1 B1B1 B1 B1 B1 L• D L• B2 B2B2 B2 B2B2 B2 B2 L• D L• F L•

TEb is sending on the B2 Channel

D L• F L• B1 B1 B1B1 B1 B1 B1B1 L• D L• Fa L• B2 B2B2B2 B2B2 B2B2 L• D L• B1 B1 B1B1 B1 B1 B1B1 L• D L• B2 B2B2 B2 B2 B2B2 B2 L• D L• F L•

TEc is sending on the D Channel

D L• F L• B1 B1 B1 B1B1 B1 B1 B1 L• D L• Fa L• B2 B2 B2B2 B2B2 B2B2 L• D L• B1 B1 B1 B1B1 B1 B1B1 L• D L• B2B2 B2B2 B2B2 B2 B2 L• D L• F L•

The NT sees this bit stream

TEs NT
48 bits in 250 microseconds

FL B1 L D L Fa L B2 LDL B1 LDL B2 LDL

TE using Channel B1 All TEs

TE using Channel B2

TE using Channel D

 1998 - R. E. Blackshaw Page 2 - 28

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

NT1 Derivation of “Echo” Channel

We saw the presence of echoed D bits in the NT to TE frames, but where do they come from?
Our busy NT1 device has this chore added to its list.

As the diagram illustrates (although don’t use this as an accurate technological reference), every D
bit that arrives at the NT1 device is sent out onto the line, as are the incoming D bits sent on to the
terminal devices. However, every D bit that goes out is echoed back to the terminal devices in the
E (echoed D bit) bit position. An exact copy of every outbound D bit is sent back as an E bit.

As shown in the frame diagrams, the last D bit of a TE to NT frame is echoed in the first E bit of
the next NT to TE frame, and so on. Thus the relationship of D to E bits is always:

TE to NT Frame NT to TE Frame

D bit 47 E bit 11

D bit 12 E bit 24

D bit 25 E bit 35

D bit 36 E bit 46

D bit 47 E bit 11

and so on.

Remember that D bits are only ever ones (no voltage) or negative zeroes (voltage) so the NT1 will
either see a binary zero as long as one TE sends a binary zero, and will only see a binary one if both
TEs send a binary one.

Page 2 - 29  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

NT1 Derivation of “Echo” Channel

S/T-Ref. Pt NT1 U-Ref. Pt.

B
B
B
D
Echo B
D

Echo of D-Channel Bits

B
B
B
D
B
D

 1998 - R. E. Blackshaw Page 2 - 30

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

TE Access to D-Channel (S/T)

A s noted earlier, terminals negotiate the use of B-Channels and they are assigned to a single TE
at a time. The D-Channel is accessible to, and must be shared equally by, up to a maximum of
eight TEs. The question is “How to do it?”

As it turns out, there is a simple and elegant solution. The contention resolution is based on the
echoed D bits and the use of HDLC at Layer 2. This dependency upon another layer’s protocol is
not quite kosher from the OSI view, however, it is not that the Layer 1 looks at the Layer 2 protocol,
it simply depends upon HDLC’s bit-stuffing technique to ensure that Layer 1 will never see more
than seven contiguous one bits on a busy channel.

Each TE scans the E bits in the incoming frames from the NT. For every one bit that it sees it
increments a counter (C), and for every zero bit it sees it resets this counter (C). If the counter
reaches a count of eight, the TE may safely assume that the D-Channel is idle. This is true because
the use of HDLC at Layer 2 ensures that a stream of valid data cannot contain eight contiguous one
bits.

The TE begins sending a complete Layer 2 frame on the D-Channel and it also monitors the E bits
returned by the NT (remember they are accurately positioned by the rules). As long as the E bits
match the D bits, the TE continues sending until it has transmitted a complete Layer 2 frame.

At this point the TE relinquishes the D-Channel and increments its idle count (C) requirement to
nine. This ensures that all other TEs that want the channel will get preference over the TE that has
already had one turn. The standard is a little sketchy on when the expected count is reset to a value
of eight; presumably it is whenever a TE detects nine contiguous one bits in the echoed channel,
at which point it knows that no other terminal with an expected count of eight wanted the D-
Channel.

This expected count of eight only applies to priority traffic, i.e., signaling messages. Packet-mode
and telemetry traffic are non-priority transmissions and use an expected count of ten.

This explains how they share the D-Channel equally among the terminal devices, but how is actual
contention resolved? This is a synchronous, single clock channel, so presumably two TEs can
begin simultaneously. Turn to the next page.

Page 2 - 31  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

TE Access to D-Channel (S/T)

0011111111 D

01111110 00000001 11111110 Echo

1 TE listens for 8 contiguous 1 bits

TE
1 3
2 TE begins sending information on D-Channel

3 TE listens for matching info on “E”-Channel

2

01111110 00000001 11111110 D

Note: Although a 16 kbit/s stream of ‘echoed’ D bits is included in the framing

from NT to TE, it is not referred to as an Echo-Channel in the documents.

 1998 - R. E. Blackshaw Page 2 - 32

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

D-Channel Contention (S/T)

T his diagram illustrates contention resolution, which we must remember is based on the digital
design and the use of HDLC at Layer 2. This diagram illustrates two terminal devices (TEa and
TEb) which simultaneously seized the D-Channel and began sending information.

You will note that they both have sent flag characters (01111110) as their first octet, and since that
is what they both observed in the echoed bits, remain on the channel. Note that this bit comparison
is done bit-by-bit with no delays.

Each TE then sends its next character (the first address octet of the HDLC frame) and since they
are both sending signaling information, the addresses match so there is still no contention.

Now the TEs end the second HDLC address octet, and since these octets contain the terminal
address they will be different. As soon as TEb sees a different bit echoed than it had transmitted,
it immediately ceases sending and reverts to the bit counting state to seek an idle condition on the
D-Channel.

Because of the line encoding rules, TEa sees exactly what it sent and is completely unaware that
contention occurred because there was no corruption of its data. TEa continues sending until it has
transmitted a complete Layer 2 frame, at which point it relinquishes the D-Channel.

Page 2 - 33  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

D-Channel Contention (S/T)

1 0 0 1 0 0 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 0
TEa sends

1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 0
TEb sends

1 0 0 1 0 0 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 0
NT sees and
echoes

TEb gets incorrect echo and leaves channel, TEa continues

 1998 - R. E. Blackshaw Page 2 - 34

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

The Physical Layer Multiplexor

T his diagram shows, in very simplified form, the Physical Layer multiplexor. It delivers
B-Channel bits (octet aligned) and D-Channel bits (one at a time) to the Layer 2 entities.

This diagram should also illustrate the special D-Channel contention resolver somewhere in the
D-Channel path, but this seems to be the case of where a few words say more than a picture.
Effectively, in concert with the Layer 2 protocol, this function provides multiple logical
D-Channels over a single physical D-Channel. More about this in the next chapter.

Page 2 - 35  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

The Physical Layer Multiplexor (BRI)

B1 B2 D
Layer 2

Layer 1
B1 octets B2 octets D bits

48 bit Physical Frame (simplified)

Physical Media

 1998 - R. E. Blackshaw Page 2 - 36

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Derivation of Q and S bits (S/T)

A s if two 64 kbit/s B-Channels and a 16 kbit/s D-Channel were not enough, recent work in ITU-
T and ANSI-T1 has provided us with more channels on the S/T reference point. These are the
Q-Channel and the S-Sub-channels. They derive these extra channels through an alternative use of
the Fa bit, and use of what had been two spare bits, the M and S bits.

The Q-Channel uses a technique known as multiframing, where a multiframe consists of 5 Physical
Layer frames. The Q bit (in the frame Fa bit, position 14) only appears in every fifth frame in the
direction TE-to-NT. Identification of a multiframe is achieved by the NT setting the M bit
(position 26). This indicates that the subsequent TE to NT frame is to be a multiframe. The TE
uses four contiguous multiframes to convey a four bit code to the NT, i.e., Q1, Q2, Q3, and Q4,
and also sets the M bit to a binary one, but only at the beginning of a twenty frame multiframe
group. This is a uni-directional channel only between the TEs and the NT; it does not extend to
the exchange. This results in an 800 bit/s channel bandwidth.

They have also defined five S-Sub-channels by using the S bit of the twenty frame multiframe
group. There are four SC bit positions allowing for a four-bit code. The S-Channel is in the
direction NT-to-TE only and results in an effective rate of 4000 bit/s, or 800 bit/s per S-channel.
The four bit codes should be repeated in at least six consecutive S characters, or at least as many
times as necessary to obtain the desired response.

The use of Q- and S-Channels is for loop-back requests, error indications, alarms, generally any
maintenance feature. It is safest to consult the actual standard for bit values and uses; however, the
next two pages will deal with those covered in the latest documents. The standard specifies use of
the first sub-channel (SC1) with the other four sub-channels reserved for future standardization.

Page 2 - 37  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Derivation of Q and S bits (S/T)

Frame NT-TE Fa bit value NT-TE M bit value TE-NT Fa bit value NT-TE S bit
1 ONE ONE Q1 SC11
2 ZERO ZERO ZERO SC21
3 ZERO ZERO ZERO SC31
4 ZERO ZERO ZERO SC41
5 ZERO ZERO ZERO SC51
6 ONE ZERO Q2 SC12
7 ZERO ZERO ZERO SC22
8 ZERO ZERO ZERO SC32
9 ZERO ZERO ZERO SC42
10 ZERO ZERO ZERO SC52
11 ONE ZERO Q3 SC13
12 ZERO ZERO ZERO SC23
13 ZERO ZERO ZERO SC33
14 ZERO ZERO ZERO SC43
15 ZERO ZERO ZERO SC53
16 ONE ZERO Q4 SC14
17 ZERO ZERO ZERO SC24
18 ZERO ZERO ZERO SC34
19 ZERO ZERO ZERO SC44
20 ZERO ZERO ZERO SC54

 1998 - R. E. Blackshaw Page 2 - 38

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Location and Characteristics of Loopbacks

T he loopback points for Layer 1 are shown in the upper figure opposite. The specifications are
covered in the lower table. These are the same loopbacks as specified by the ITU-T and by
ANSI-T1.

The actual locations shown in the upper figure are not to be taken as absolute or as having any
significance as regards actual devices. This is stated in the standards and is in keeping with the fact
that in this same diagram (as used for the ISDN reference) the reference points and components are
abstractions.

Page 2 - 39  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Location, Characteristics of Loopbacks

TE1 NT2 NT1

S T U
B C

S
TE2 TA
R

Channels Control Alternate

Loopback Location Looped Point Control

Inside
C B1, B2 TE or NT2 None
NT1

Inside None
B B1, B2 TE or NT1
NT2

 1998 - R. E. Blackshaw Page 2 - 40

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Codes for Q and S1 Channel Messages

T he table opposite shows the codes assigned for Q- and S-Channel messages; the three codes in
the darker shading at the bottom are reserved for V.230. Recommendation V.230 is a new
recommendation that proposes the S/T reference point arrangement as a universal DTE/DCE
interface. Except where the standard states otherwise, the code for a message shall be repeated in
at least six consecutive Q or S characters or as many times as necessary to get the desired response,
i.e., loopback. Except where stated otherwise, a message shall be considered as received only
when the proper code is received (detected) in three consecutive Q or S characters. A character is
a 4-bit binary sequence.

The messages in the upper part (lighter shading) of the table have are in priority order except for
the idle code.

TE-to-NT messages:

Idle is transmitted during normal conditions, i.e., when no other message is being transmitted. This
is the message used to clear loopbacks. Loss of power is an indication to the NT that the TE has
lost power. It should be transmitted at least one multiframe just prior to INFO 0 before frame
alignment is lost. This means that TEs should have enough energy storage for two multiframes or
10 ms. Request self test is sent by the TE to request the NT to perform a self test. The scope of
the test is not defined nor is its use with an NT2. Request for loopback is either for channel B1
(LB1) or channel B2 (LB2). The code for LB1/2 is only received when two TEs simultaneously
request LB1 and LB2. Reverting to the idle message clears the loopback.

NT-to-TE messages:

Idle is transmitted during normal conditions. Loss of power is an indication that the NT has lost
power and the same energy storage implications apply. Detected access Transmission System
Error (DTSE) is an indication that the NT has detected an error in a block of bits, Out and In in the
table apply to the transmission direction as seen by the NT. Self test indicates that the NT is in a
self test mode. The self test report indicates the result of a self test. Loopback indicates the
loopback in effect. Loss of received signal indicates that the signal at the U interface has been lost,
e.g., loss of frame synchronization.

Page 2 - 41  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Codes for Q & S1 Channel Messages

Code Used and Direction

Message NT to TE TE to NT
S11 S12 S13 S14 Q1 Q2 Q3 Q4
Idle (normal) 0 0 0 0 1 1 1 1

Loss of Power 1 1 1 1 0 0 0 0

STP Pass 0 0 1 0 - - - -
STP Fail 0 0 0 1 - - - -
ST Request - - - - 0 0 0 1
STI Indication 0 1 1 1 - - - -

DTSE In & Out 1 1 0 0 - - - -

DTSE In 1 0 0 0 - - - -
DTSE Out 0 1 0 0 - - - -

LB1 Request - - - - 0 1 1 1
LB1I Indication 1 1 0 1 - - - -
LB2 Request - - - - 1 0 1 1
LB2I Indication 1 0 1 1 - - - -
LB1/2 Request - - - - 0 0 1 1
LB1/2I Indication 1 0 0 1 - - - -

Loss of Received
1 0 1 0 - - - -
Signal Indication

Disruptive
Operation 0 0 1 1 - - - -
indication

V-DTE Slave - - - - 1 1 0 1
V-DCE Slave - - - - 1 1 0 0
V-DCE Master 0 1 1 0 - - - -

 1998 - R. E. Blackshaw Page 2 - 42

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Codes for S2 Channel Messages

The S2 sub-channel was added in the latest version of ANSI T1.605-1990 to add more messages
for the NT to report status to attached TEs. The descriptions in the table on the facing page are self
explanatory, or al least they will be when we have studied the next part of this chapter on the Digital
Subscriber Loop (DSL) protocol.

In the first two messages, dea and act refer to two bits in the DSL frame which are the deactivation
and activation bits respectively. The crc notation in CcrcR and CcrcN refer to the cyclic
redundancy check bits in the DSL frames. The eoc in the RUeoc message refers to an embedded
operations channel in the DSL frame.

Because more than one of the conditions covered in the S2 sub-channel messages can occur
simultaneously, e.g., a B1 loopback (LB1N) and an intentionally corrupted crc (CcrcR), the NT
will cycle through all existing conditions. The four bit message codes are sent in at least six but not
more than eight repetitions. Each four bit code requires one multiframe (5 milliseconds) so that six
or eight repetitions would require between 30 and 40 milliseconds to transmit. The order of
repetition is the same as the table which, you will note, is from the highest binary value to the
lowest binary value.

Page 2 - 43  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Codes for S2 Channel Messages

Code Used
Message Description
SC21 SC22 SC23 SC24

Rdea Receiving valid dea = 0 1 1 1 1

Ract Receiving valid act = 0 1 1 1 0

NAI Network has notified NT of 1 1 0 1

a network failure

PPb Primary power bad, Secondary Good 1 1 0 0

PSb Primary power good, Secondary bad 1 0 1 1

2B+D looped back to network, or NT1 1 0 1 0

LUTI placed into Quiet or Insertion Loss Mode
-
LB1N B1 looped back to network 1 0 0 1
LB2N B2 looped back to network 1 0 0 0

CcrcR Corrupt crc requested 0 1 1 1

CcrcN Corrupt crc notified 0 1 1 0

NT receiver is in either Full Reset 0 1 0 1

Rset or Receive Reset state

Receiving valid but unrecognized 0 0 1 1

RUeoc eoc message

Introduce extended set of 0 0 1 0

Extend messages

SC2 subchannel is implemented 0 0 0 1

IBNAM but there are no active messages

SC2 subchannel is not 0 0 0 0

SNI implemented

 1998 - R. E. Blackshaw Page 2 - 44

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Layer 1 - NT1 to Exchange

We now come to the next protocol in the Layer 1 set, that used between the NT1 and the ISDN
switch’s line termination (ET).

This protocol is a point-to-point arrangement, no contention here. It is only required to carry 144
kbit/s of user-network information (2 x 64 kbit/s + 16 kbit/s = 144 kbit/s).

It is also full-duplex and synchronous, although there are techniques such as Time Compression
Multiplexing (TCM) that, while half-duplex, give the appearance of being full-duplex. TCM was
a contender for the loop technology for a time.

Finally, we must have some method of demarcation for the bit stream and synchronization of the
two end points. There is also an Embedded Operations Channel (EOC), and this accounts for 16
kbit/s so the resultant bit stream is at a 160 kbit/s rate.

Page 2 - 45  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Layer 1 - NT1 to Exchange

2-wire loop
NT ET

144 kbit/s of user information to carry

Some Overhead Req’d - typically 16 kbit/s

Full-Duplex Point-to-point synchronous

Different Protocol than S / T Reference Points

 1998 - R. E. Blackshaw Page 2 - 46

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Echo Cancellation Theory

T he final choice for the ANSI-T1 standard (ANSI T1.601) for the Digital Subscriber Line (DSL)
was echo-cancellation with decision feedback equalization. A long name for a complex technique,
but one with many advantages.

The diagram represents the major blocks of an echo-cancellation device. Beginning at the upper
left, the transmitted bit stream goes to the hybrid and to the Decision Feedback Equalizer (DFE).
In the hybrid circuit, half of the transmitted power goes out onto the loop and the balancing network
(N) absorbs the other half. This circuit ensures that none of the transmitted power reaches the
receiver.

The hybrid is designed to send all power received from the loop to the receiver. Nothing in this
world is perfect, and hybrid circuits and balancing networks are no exception. The balancing
network is adjusted to be as near a match as possible for the characteristics of the loop. Perfection
being unattainable, this match is never exact and some small amount of transmitted power reflects
back toward the hybrid from these mismatches. Also, there are bridge taps* along the loop cable
and another hybrid at the distant end of the loop. These taps and far end hybrids result in more
reflections of power back to the hybrid and thence to the receiver.

However, while the signal is being transmitted, the DFE is calculating a best guess at what part of
the signal will be reflected back. The DFE inverts this calculated reflection signal and sends it to
the summing (Σ) circuit at the same time as the reflected power (the echo) arrives. Since it is an
inversion of the echo, the two signals should cancel when summed**.

The receiver hears whatever does not get cancelled by the summation and feeds this back to the
DFE. Using these results, the DFE recalculates its guess at the echo and the process repeats. By
transmitting a known pattern and repeating the cycle, the circuit trains itself until the received echo
becomes low enough to be negligible.

This technique of self training, when repeated at both ends of the loop, permits full-duplex
transmission on a single pair of wires. In use, the training, or decision circuit, is always
operational.

This used to be expensive technology, however, recent advances in Large Scale Integration (LSI)
silicon technology has drastically reduced the costs.

* When a local cable extends from the exchange building, it does not only run down a single
path. Often there are several branches from the main cable, and individual cable pairs may
appear in all branches. These pairs, called bridge taps, at ISDN frequencies can behave like
stubs on a radio transmission line, causing signal reflections.

** This may also be a subtraction circuit, eliminating inversion by the DFE.

Page 2 - 47  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Echo Cancellation Theory

Transmitter

Decision
Feedback N Hybrid
Equalizer

Receiver ∑

 1998 - R. E. Blackshaw Page 2 - 48

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

U-Interface Line Code

A line rate of 160 kbit/s does not sound like very much to data-processing people used to short
buses and shielded environments. However, it is a fast line rate for cable pairs designed to carry
analog voice between 100 to 4000 Hz.

Anything that could reduce the effective rate would be very welcome, and experts proposed
various line codes before ANSI-T1 finally settled on the 2B1Q line code.

Two Binary One Quaternary (2B1Q) simply means that each single line code symbol represents a
binary bit pair. They use the term symbols because they do not specify the actual voltage used.
Each symbol is composed of a sign (+ or -) and a magnitude (1 or 3). The symbols +3, +1, -1, and
-3 represent the four possible bit patterns. The line voltage of the symbols is nominally 5/6th of the
symbol value, giving 2.5 volt, and 0.84 volt values for 3 and 1 respectively.

This effectively halves the actual line rate which now becomes 80 kbaud. We use bits/sec to
specify data rates, we use bauds to specify actual line coding rates. Refer to any good book on data
transmission techniques for a complete discussion.

It is estimated (by laboratory tests) that an 80 kbaud line rate will cover 99% of the standard loops
in use by the network operators. A standard loop is 18,000 feet and may be as fine a wire gauge
as #26, heavier wire gauges may well exceed the 18,000 foot limit.

Note: At least one vendor now has a product which extends this loop length considerably.

Page 2 - 49  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

U-Interface Line Code (2B1Q)

+3

+1

-1

-3

Symbol +1 +1 +3 -1 -1 +1 -3

Bits 11 11 10 01 01 11 00

Symbol Binary
+3 1 0

+1 1 1

-1 0 1

-3 0 0

 1998 - R. E. Blackshaw Page 2 - 50

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Basic Loop Frame at U

We define the basic frame structure for a loop frame in terms of quats, each quat being one line
symbol but two binary bits. The first nine quats of each frame is either a synchronization word or
an inverted synchronization word. The next 108 quats contain the information from 2 B-Channels
and the D-Channel, actually consisting of twelve octets from each B-Channel and 24 bits from the
D-Channel. There are 3 quats of maintenance information at the end of each frame.

The whole frame takes 1.5 milliseconds to transmit, and simple calculation will show that this
maintains the 8 kHz octet rate, i.e., 64 kbit/s. This is all carefully designed to maintain what ISDN
terms 8 kHz integrity, or bit/octet alignment through the network.

The Synchronization Word (SW) and Inverted Synchronization Word (ISW) provide the framing
for this bit stream (or is it a quat stream). Synchronization Words mark each basic frame and the
Inverted Synchronization Word marks the beginning of a superframe. Superframes are defined on
the next page pair.

Page 2 - 51  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Basic Loop Frame at U

1.5 milliseconds

SW / ISW 12 x (2B + D) M

9 Quats 108 Quats 3 Quats

SW = Synchronization Word

ISW = Inverted Synchronization Word

M = Maintenance or Overhead

Quat = Quaternary Symbol = 1 Baud = 2 Binary Bits

Line Rate = 120 Quats in 1.5 milliseconds = 80 kbaud / second

B1 B2 D

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2

 1998 - R. E. Blackshaw Page 2 - 52

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Superframe Format - (U)

It seems that no one is ever satisfied with single frame formats and so we have a superframe for
the Digital Subscriber Line (DSL). It allows a more structured, and therefore useful, allocation of
the 3 maintenance quats (6 bits) at the end of each frame. A DSL superframe is nothing more than
a concatenation of eight basic frames.

The ANSI T1 standard letters superframes A - Z in one direction and numbers them 1 - 26 in the
other direction. They are set apart by inverting the synchronization word in the first basic frame
of each superframe. The whole significance of a superframe is in the maintenance bits (M1 to M6).

NT to Network Chart

Use of the M1 to M3 bits is for addressing (eoca1 to eoca3) and up to seven locations are
addressed. When discussing a point-to-point loop, this may seem puzzling, but when discussing
the hypothetical worst case loop, the need for addressing will be seen. Beside the address bits there
is a one bit data message indicator (eocdm) and an eight bit information field (eoci1 to eoci8).

The M4 bit provides three functions, and the unused bits are set to one. There is an activation bit
(act), a power status bit pair (ps1 and ps2), and an NT in test mode bit (ntm).

The M5 and M6 bits are only used for error detection, using a twelve bit cyclic redundancy check
(crc1 to crc12). They use the Far End Block Error bit (FEBE) to signal receipt of errored frames.
A binary one indicates no errors and a binary zero indicates an errored frame. The main purpose is
for performance monitoring.

The significance of the designations 118s and 118m in the quat position row refers to the quat
symbol sign (s) and magnitude (m). Refer to the page describing the 2B1Q line code for
clarification.

Network to NT Chart

The diagram for the Network-to-NT direction shows that the only difference is in the use of the
M4 bit, which here is used as an activation / deactivation status, i.e., it is set when the DSL is fully
active as a status indicator rather than a command.

Page 2 - 53  1998 - R. E. Blackshaw

ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Superframe Formats
Framing 2B + D Overhead Bits
Quat Pos 1-9 10 - 117 118s 118m 119s 119m 120s 120m
Bit Pos 1 - 18 19 - 234 235 236 237 238 239 240
SuFr BaFr Sync
# # Word 2B + D M1 M2 M3 M4 M5 M6

1 ISW 2B + D eoca1 eoca2 eoca3 act 1 1

2 SW 2B + D eocdm eoci1 eoci2 ps1 1 febe
3 SW 2B + D eoci3 eoci4 eoci5 crc1 crc2
ps2
4 SW 2B + D eoci6 eoci7 eoci8 ntm crc3 crc4
1
5 SW 2B + D eoca1 eoca2 eoca3 cso crc5 crc6

6 SW 2B + D eocdm eoci1 eoci2 1 crc7 crc8

7 SW 2B + D eoci3 eoci4 eoci5 1 crc9 crc10

8 SW 2B + D eoci6 eoci7 eoci8 1 crc11 crc12

Framing 2B + D Overhead Bits

Quat Pos 1-9 10 - 117 118s 118m 119s 119m 120s 120m
Bit Pos 1 - 18 19 - 234 235 236 237 238 239 240
SuFr BaFr Sync
Word 2B + D M1 M2 M3 M4 M5 M6
# #
1 ISW 2B + D eoca1 eoca2 eoca3 act 1 1

2 SW 2B + D eocdm eoci1 eoci2 dea 1 febe

3 SW 2B + D eoci3 eoci4 eoci5 crc1 crc2
1
4 SW 2B + D eoci6 eoci7 eoci8 crc3 crc4
1
A
5 SW 2B + D eoca1 eoca2 eoca3 1 crc5 crc6

6 SW 2B + D eocdm eoci1 eoci2 crc7 crc8

1
7 SW 2B + D eoci3 eoci4 eoci5 1 crc9 crc10

8 SW 2B + D eoci6 eoci7 eoci8 1 crc11 crc12

1 = reserved for future standard crc = cyclic redundancy check

an = eoc address bits febe = far end block error bit
dm = eoc data/message indicator psn = power status bits
in = eoc information (data/message) ntm = NT in test mode bit
act = activation bit cso = cold start only bit (set = 1)
dea = deactivation bit
 1998 - R. E. Blackshaw Page 2 - 54
Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

DSL Framer - Functional Description

T his diagram, while part of the ANSI-T1 standard, is presumably used only to indicate how to
build DSL superframes, and not as a definitive method of building such devices. It illustrates how
the maintenance bits are grouped as noted on the previous page, and how they would be gradually
inserted into the frame.

The box marked scra deserves explanation. It represents a bit scrambler function that applies to
all bits except the synchronization words. Scrambling is done to present random pulses on the line,
a technique which statistically changes the bit patterns such that there is no concentration of line
codes at or around one specific symbol, i.e., to eliminate the chance of a line signal consisting of
mostly +3s or +1s or -3s, etc. This produces a better power spectrum on the loop, reducing co-
channel interference. The inverse of the scrambling function is done at the receiver to restore the
information to its original state.

Page 2 - 55  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

DSL Framer - Functional Description

eoc a

eocdm m M1, M2, M3

u
x
eoci s = sign (+ or-)

m = magnitude (1 or 3)
B1

B2 m 2B+D s
u 2B1Q m
x enc m s
m 2B+D, M4 m scra
D u m
u u
act x s x
x BOS EOS sync
ps1/dea m
crc1 crc1 gen
ps2 m M4 crc • crc m
•
u • • u
ntm • •
x gen crc12 latch crc12 x M5, M6
cso Invert
1*

febe
m M5
u
1** x
* - Reserved bits in M4 slot (included in crc check)
** - Reserved bits in M5 slot (not included in crc check)

 1998 - R. E. Blackshaw Page 2 - 56

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

DSL Framing Relationships

T his diagram illustrates the relationships of the superframes in each direction of transmission.
Superframes from the NT toward the network are offset from the incoming frames from the
network by 60 ±2 quaternary symbols, approximately 0.75 milliseconds.

This illustrates the relation between the far end block error bit (febe) and the actual frame it
references. It also illustrates that the cyclic redundancy check bits are not for the frame containing
them, but for the immediately preceding frame.

For example, the network frame marked C contains the calculated crc for the preceding network
frame B. It also contains the febe bit for the frame 1 coming from the user’s NT.

Page 2 - 57  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

DSL Framing Relationships

Z A B C D E
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

crcZ,febe-1,eocLN crcA,febe0,eocLN crcB,febe1,eocLN crcC,febe2,eocLN

12ms 24ms 36ms 48ms

0 1 2 3 4 5
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

crc0,febeY,eocNL crc1,febeZ,eocNL crc2,febeA,eocNL crc3,febeB,eocNL

0.75ms 12.75ms 24.75ms 36.75ms 48.75ms
0ms
Time

Where: Y,Z,A,B,C, = Network to NT Superframes

-1,0,1,2,3, = NT to Network Superframes

i = ith 1.5ms Basic Frame in a given Superframe

crcX = Cyclic Redundancy Check for SuperframeX

febeY = Far End Block Error bit for SuperframeY

eocij = Embedded Operations Channel for ij direction

LN = Network to NT, NL = NT to Network

 1998 - R. E. Blackshaw Page 2 - 58

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Activation Initiated by the Exchange

T his chart shows the events and actions when the exchange side activates the physical layer. The
diagonal arrows show the direction of signal flow, while the horizontal arrows show the direction
of information flow. The notation L1, L2 refers to the protocol layers.

Activation begins when the Physical Layer at the Exchange Termination (ET) receives an
Activation Request (AR), this also sets the dea bit in the network equal to 1. Timer M5 is started
and sends tone TL toward the subscriber’s network termination NT. When NT detects the tone, it
starts timer M4 and returns tone TN for the required duration.

The NT, after completing the tone signal, sends an Info 1 signal toward the exchange, followed by
an Info 0 signal (see definitions below). When the exchange detects these signals, it responds with
Info 1, followed by Info 2 frames in which the dea bit is set to 1. When the NT has acquired frame
synchronization (SW) it returns Info 2 frames. Note that these Info 2 frames are now flowing in
both directions. When the NT acquires superframe (ISW) it stops sending Info 2 frames and sends
Info 3 frames toward the exchange and stops timer M4. When the exchange acquires superframe
synchronization (ISW), it also changes from Info 2 to Info 3 frames toward the NT.

After the NT acquires superframe synchronization it forwards the Info 2 frames toward the
terminals (TE), and, at this time, an active terminal would respond with Info 3 frames. It is these
Info 3 frames that the NT forwards to the exchange.

When the NT recognizes Info 3 frames from the exchange, it sets the act bit in its outbound frames
to 1. When the exchange detects that act is set it transmits an Activation_indication (Ai) to Layer
2, and sets the act bit in its outbound frames to a 1. The NT, detecting the setting of this bit,
forwards the Info 4 frames toward the TE, which notifies its Layer 2 of the activation by means of
an Ai primitive.

Signals and frames; (SNn is the DSL equivalent of Info n)

Info 0 No signal.
Info 1 A continuous signal with the pattern of a positive zero, six ones, and a
negative zero.
Info 2 A frame with all B, D, and D-echo channel bits set to zeroes - other bits
set according to the rule.
Info 3 Synchronized frames with operational data on the B- and D-Channels.
Info 4 Frames with operational data on B- and D-Channels, and A bit set to 1.

Page 2 - 59  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Activation Initiated by the Exchange

ET/LT Digital System (U) NT1 S/T Ref. Point TE

L2 L1 L1 L1 L1 L2
Start M5
AR TL Start M4
TN
SN1
SN0
SL1
dea = 1 SL2
SW Sync
SN2
SN3 INFO 2
ISW Sync Stop M4
INFO 3
ISW Sync
SL3
Stop M5
act = 1

AI act = 1
INFO 4
AI

Note: receipt of INFO3 and SL3 at the NT can theoretically occur in either order.

 1998 - R. E. Blackshaw Page 2 - 60

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Activation Initiated by the Terminal Equipment

T his diagram illustrates the sequence of actions taken when the terminal equipment activates the
Physical Layer.

Network Termination (NT1) in response to an Activate_request (AR) from the TE management

function. From this point, a good deal of activity takes place over the digital system, i.e., subscriber
loop.

NT1 first starts timer M4 and sends the tone signal toward the exchange (ET/LT). NT1 then sends
a normal frame synch word with all other bits in that frame set to binary one (SN1). More (SN1)
frames follow until the exchange responds with a similar frame (SL1). The exchange then sends
a superframe with all binary zeroes in the 2B+D bits (SL2). When the NT1 has acquired frame
sync it responds with a superframe (SN2) and finally, when NT1 acquires superframe sync, it sends
normal superframes (SN3). At this point, NT1 stops timer M4 and an Info 2 frame is sent to the
TE. The TE responds with an Info 3 frame. When the exchange has acquired superframe sync
from the NT1, it sets the activate bit to 1 indicating that the DSL is in the active state. This causes
the NT1 to send Info 4 frame(s) to the TE. The TE responds with Activate_indication to the
management function that started the whole procedure.

Note: SNn and Info n are equivalent, but in the format of either S/T or U reference points.

Activation of the T reference point without activation of the Digital Subscriber Line (DSL) is
permitted in the ANSI standard and assumes use of a freerunning clock in the NT device. This is
required to permit maintenance operations in situations where the network interface cannot
activate.

Page 2 - 61  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Activation Initiated by Terminal Equipment

ET/LT Digital System (U) NT1 S/T Ref. Point TE

L2 L1 L1 L1 L1 L2

Start M4 INFO 1 AR
Start M5 TN
SN1
SN0

SL1
dea = 1 SL2
SW Sync
SN2
SN3 INFO 2
ISW Sync Stop M4
INFO 3
ISW Sync
SL3
Stop M5
act = 1

AI act = 1 INFO 4
AI

Note: receipt of INFO3 and SL3 at the NT can theoretically occur in either order.

 1998 - R. E. Blackshaw Page 2 - 62

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Worst Case Basic Access Configuration

It is normal in setting standards for telecommunication networks to provide a hypothetical

reference connection. The purpose of this reference connection is to specify the overall end-to-end
circuit allowable.

This diagram provides a similar service for the Digital Subscriber Line (DSL) and illustrates the
worst case DSL, or the maximum numbers of intervening equipment that could exist between the
subscriber and the exchange.

The provision of address bits within the Embedded Operations Channel (EOC) in the superframe
allows us to activate/deactivate tests within these intervening equipments. The address rule is that
if the address is in the range from 2 to 6 for messages directed toward the subscriber’s NT, the
equipment decrements the address and passes the message onward. If the address is equal to a
binary one, the equipment responds to the message.

In the other direction (toward the exchange LT), if the address is in the range from 1 to 5 the
equipment increments the address and passes the message onward.

In either direction, if the address is a binary zero no change is made to the address and the
equipment passes the message onward. If the address is a binary seven, the equipment responds to
the message, does not change the address, and passes the message onward.

The NT will only respond to an address of zero or seven.

This configuration is not simply a standards exercise, for some of the early trials used Universal
Interoffice Carriers and Universal Digital Loop Carriers to extend ISDN service from a
metropolitan digital switch to a user located in the suburbs where the local switch could not adapt
to ISDN.

Page 2 - 63  1998 - R. E. Blackshaw

 ISDN Training Course Chapter 2 - ISDN’s Physical Layer - BRI

Worst Case Basic Access Configuration

ISDN Switch (Remote) Central Office (Local)

ET D-Bank D-Bank COT RDT

DSL DSL
LT LU/NT LU/LT LU/NT LU/LT RP
MDF

OSN Universal Interoffice Carrier UDLC NT

OS

COT = Central Office Terminal OSN = Operating System Network

DSL = Digital Subscriber Line OS = Operations System

LT = Line Termination RDT = Remote Digital Terminal

NT = Network Termination RP = Loop Repeater (Network)

LU/LT = Line Unit “LT” UDLC = Universal Digital Loop Carrier

LU/NT = Line Unit “NT” = Digital Facility (Met. or Fiber)

MDF = Main Distributing Frame

 1998 - R. E. Blackshaw Page 2 - 64

Chapter 2 - ISDN’s Physical Layer - BRI ISDN Training Course

Notes:

Page 2 - 65  1998 - R. E. Blackshaw