J. T. O. Phase II (Switching Specialisation) : axe-10 axe-10 Contents Page

Signalling terminals (ST) for signalling according to CCITT No. 7 are connected to the group switch via a PCD-D. Since the signalling terminals are digital devices, the PCD-D equipment includes no conversion function but merely serves as an adaptation device towards the group switch.

The signalling information from a signalling terminal is sent through the group switch to a certain channel in an ETC.

T
his channel is then used exclusively for signalling. The advantage of connecting the signalling terminals via the group switch is that some devices can be kept in reserve and automatically replace inoperative devices.

S
ignalling Terminals for CCITT No. 7

CCITT No. 6 is a signalling system used for international connections. The basic principle is the same as for CCITT No. 7, but the system design is adapted to suit analog signalling links. This means that the transmission rate is somewhat lower (2400 bit/s), that is in comparison to 56 or 64 kbit/s when CCITT No.7 is used.

F
igure 2.3.14 shows the hardware used for CCITT No. 6.

GSS = Group Switching Subsystem

PCD = Pulse Code Modulation Device

ST-6 = Signalling Terminal for CCITT No. 6
Figure 2.3.14

Signalling Terminals for CCITT No. 6

2.4 The Digital Group Switch

Before studying the structure of the digital group switch in AXE we will touch upon some of the basic principles of digital switching.

The introduction of digital switching gave birth to a new concept:

TIME SWITCH

L
et us first see what a time switch is made up of and how it operates.

A/D = Analog/Digital converter

A time switch is made up of:

• 
  a Speech Store for temporary storage of the speech samples. Each channel in the time switch has a position of its own in the Speech Store.
  
• 
  a Control Store which controls the read-out from the Speech Store.
  

A
ssume that we are going to read out samples from the speech store in the following order: 3, 2, 1, 4 (the read-in order is 1, 2, 3, 4). The control store would then have the following contents (see Figure 2.4.2).

= Digital Signal

This small-size time switch has only 4 inputs. How, then, do we go about designing a digital group switch with tens of thousands of inputs?

In theory we could use a single time switch having the required number of inputs. But then the following question arises: “How often would we have to ‘empty’ a given position in the speech store?” The answer is 8,000 times every second for each position (the sampling frequency is 8,000 Hz). Consequently, for a 20,000 input switch the read-in/read-out rate would be 20,000 x 8,000 Hz = 160 MHz.

Today’s market does not offer any circuits that can cope with these speeds. The solution to the problem is to divide the time switch into suitable sub-units. To set-up connections from one time switch to another we use a SPACE SWITCH.

The capacity of each time switch in AXE is 512 inputs. A maximum of 32 time switches can be connected to one space switch.

Terminology : Time Switch Module (TSM)

A connection will pass through a TSM - via SPM - to the same or another TSM.

All calls are set-up via SPM, including those which return to the original TSM. We say that the switch has a T-S-T (Time-Space-Time) structure.

TIME SWITCH MODULE (TSM)

Since a TSM handles samples in both directions, we need two speech stores: one for samples entering the TSM [Speech Store A (SSA)] and another for samples leaving the TSM [Speech Store B (SSB)]. Each speech store has a separate control store: CSA and CSB, respectively (in this case, CS stands for Control Store).

T
SM also has a control store for SPM called CSC.

CSA = Control Store A

CSB = Control Store B

CSC = Control Store C

SPM = Space Switch Module

SSA = Speech Store A

SSB = Speech Store B

TSM = Time Switch Module
Figure 2.4.4

Speech Stores and Control Stores in TSM

SPACE SWITCH MODULE (SPM)

The SPM structure is very simple and can be drawn as an ordinary matrix with cross points.

Of course, in reality, the cross points represent logic gates that open and close very rapidly.

CSC = Control Store C

SPM = Space Switch Module

TSM = Time Switch Module
Figure 2.4.5

Space Switch Module (SPM)

As appears from Figure 2.4.5, the CSC of each TSM controls a row of “cross points”. Thus, CSC in TSM-0 controls all “cross points” leading to TSM-0.

When a call is to be set-up in the switch, it is the central software of the GS block (Group Switch) that selects the path through the switch. In this case, path selection refers to the moment when a sample is to be transferred. This is called “selection of an internal time slot”.

After the central software (GSU) of the GS block has selected a path, the regional software (GSR) is ordered to write information to this effect in the control stores of the TSMs concerned.

From now on, GSU will not pay any attention to the connection until the call is to be cleared.

64K GROUP SWITCH

As we know, 32 TSMs can be connected to each SPM, providing a total capacity of 32 x 512 = 16,384 inputs (This type of group switch is often called 16K).

What can we do, then, to build a larger switch?

W
e can interconnect several SPMs to form a large matrix as illustrated in Figure 2.4.6.

This gives a total switch capacity of 128 x 512 = 65,536 inputs (This type is often called 64K).

All types of digital equipment require some form of clocking. The clock rate determines the rate at which samples are read from or written into the speech stores.

The accuracy of this clock is of great importance in networks containing several interconnected digital exchanges. The whole network must be synchronized.

It is also important that the clock does not stop, as this would stop the whole group switch.

T
o prevent this happening, the group switch has three clocks, or Clock Modules (CLM).
CLM = Clock Module

ETC = Exchange Terminal Circuit

SPM = Space Switch Module

TSM = Time Switch Module

= Digital Signal
Figure 2.4.7

Clock Modules to Synchronize the Group Switch

The operation of the group switch will be trouble-free even if only one clock is used, i.e. in emergency situations.

As has been said, the whole network must be synchronized if it contains several digital exchanges.

There are various ways of doing this. The simplest method is perhaps the MASTER-SLAVE configuration, which means that one of the exchanges has a control (master) function, while the others (the slave exchanges) try to follow the operating pattern of the master.

F
igure 2.4.8

T
he master exchange has a number (usually 3) of more sophisticated and accurate clocks called Reference Clock Modules (RCM). Figure 2.4.9 shows the hardware included in the master and slave exchanges.

= Digital Signal

The photograph in Figure 2.4.10 shows an RCM magazine (left) and a CLM magazine.

T
he CLM magazine has hardware for operating a switch containing 8 TSMs (4,000 inputs, often written as 4K). For larger switches, a larger version of the CLM magazine is available.
CLM = Clock Module

RCM = Reference Clock Module
Figure 2.4.10

RCM and CLM for 4K Switch

There is also another way of synchronizing a network, Mutual Synchronization. This method is to be preferred in national transit networks.

The basic principle of mutual synchronization is that one of the exchanges operates according to a mean value based on all incoming frequencies. Consequently, the network has no “master”. In order to prevent the whole network from “drifting” as a result of frequency displacement, one of the exchanges is locked to a fixed frequency value. This reference exchange is called a SINK and has three highly stable clocks called CCMs (Cesium Clock Modules) which are connected in the same way as RCM in Figure 2.4.9.

I
t is thus common practice to use two types of synchronization in a network. A fully built-up digital network may use the configuration shown in Figure 2.4.11.

Since the digital group switch is only capable of interconnecting two inputs, external equipment must be used to set up a three-party call (for example operator intervention or “Add-on conference”). This equipment is called Multi-Junctor Circuit (MJC).

An MJC magazine can handle 10 simultaneous three-party calls.

M
JC, which also has regional and central software, forms part of the GSS subsystem.

Since the group switch forms a vital part of an AXE exchange, exacting demands are, of course, made on its functional reliability.

What would happen if, for instance, an SPM broke down? Well, as many as 16,000 calls would “collapse”. And, of course, this must not happen.

To solve this problem, AXE is equipped with two complete group switches: one called the A-plane and the other the B-plane.

A speech sample is always sent through both planes but it is only fetched from one of them, usually the A-plane.

To supervise the hardware, a number of parity check functions are provided for checking the speech samples sent through the switch. A hardware fault will immediately be detected by these functions. The faulty equipment is blocked, and corresponding equipment in the other plane takes over the traffic handling. All these measures are taken without disturbing calls in progress.

As mentioned before, there is a subsystem for handling the traffic between subscribers: the Subscriber Switching Subsystem (SSS). The subscriber stage in AXE is digital, which means that the analog signal from the subscriber line is converted into digital form. This is done in the subscriber’s Line Interface Circuit (LIC) and all switching is digital. To be able to understand the structure of the subscriber stage we will first discuss its tasks.

A subscriber stage includes the following functions:

• 
  Feed current to the subscriber line.
  
• 
  Concentrate the traffic towards the group switch.
  
• 
  Receive digits from dial telephones (pulses).
  
• 
  Receive digits from keyset telephones (tones).
  
• 
  Send ring signals to the subscriber.
  
• 
  Send different tones to the subscriber.
  
• 
  Carry out measurements on the subscriber line.
  

These functions are: current feed, polarity reversal, reception of dial pulses, relay for connecting ring signals, relay for connecting test equipment, and analog-to-digital conversion.

Each printed board assembly has 8 line interface circuits; see Figure 2.5.1.

Figure 2.5.1

B
oard with 8 Line Interface Circuits (LIC)

The board is equipped with components of special Ericsson design called SLIC and SLAC (Subscriber Line Interface Circuit and Subscriber Line Audio processing Circuit, respectively).

The flexibility of the circuits makes it easy to adapt them to varying requirements in different countries. This goes in particular for power supply, speech levels and balance.

As we have seen, the line interface circuit has no equipment for the reception of digits from keyset telephones (tones). The equipment, for this receiving function is common to several subscribers and is called Keyset code Reception Circuit (KRC).

This device is digital, and each printed board assembly can accommodate 8 KRCs. To connect the KRCs to calling subscribers we need a switch- the Extension Module Time Switch (EMTS).

All three equipment units dealt with above (LIC, KRC and EMTS) have both regional and central software; see Figure 2.5.2.

KRC = Keyset Code Reception Circuit

Additional equipment is required to connect subscribers to the group switch. This equipment, which handles the 32 digital channels to the group switch, is called the Exchange Terminal Board (ETB).

ETB is the hardware of a function block called the Remote Terminal (RT). It is the central software of the RT block which reserves channels to the exchange.

CJ, A CO-ORDINATING FUNCTION BLOCK

A function block called Combined Junctor (CJ) is provided to co-ordinate all functions in the SSS subsystem.

In addition to co-ordinating the set-up and clearing phases, CJ serves as an interface with TCS and, in particular, with the RE block. See Figure 2.5.3.

CJU = Central software of block CJ


EMTS = Extension Module Time Switch

ETB = Exchange Terminal Board

KRC = Keyset Code Reception Circuit

KRR = Regional software of block KR

KRU = Central software of block KR

LIC = Line Interface Circuit

LIR = Regional software of block LI

LIU = Central software of block LI

RTR = Regional software of block RT

RTU = Central software of block RT

TCS = Traffic Control Subsystem

TSR = Regional software of block TS

TSU = Central software of block TS
Figure 2.5.3

CJ - The Central Block of SSS

How many subscribers can be connected to an EMTS?

The answer is 128 subscribers, 8 KRCs and one 32-channel ETB. All this is referred to as an Extension Module (EM) or an LSM (Line and Switch Module).

REGIONAL SOFTWARE

The regional software for the subscriber stage is stored and executed in a processor incorporated in the magazine: the Extension Module Regional Processor (EMRP).

The routine scanning of the hardware is done by small, simple microprocessors located in different parts of the hardware. These are called Device Processors (DP) and are in their turn scanned by an EMRP.

T
he program in DP has no decision-making functions; it just reports hardware changes to EMRP.

L
SM is illustrated in Figure 2.5.5.

The primary advantage of using a digital subscriber stage is that it can be detached from the exchange and installed closer to the subscribers. This will imply less cost and less maintenance.

But before this can be done, two problems must be solved:

1. 
   The 128-subscriber capacity is too small. It must be possible to combine several LSMs to obtain the required size.
   
2. 
   How can EMRP communicate with the central processor over distances of tens of kilometres?
   

EMRP = Extension Module Regional Processor

As appears from the figure, the topmost LSM has no direct contact with the parent exchange, and calls coming from this LSM must therefore use the bus which interconnects all the LSMs. This bus is called Time Switch Bus (TSB) and is thus used for speech data. The bus is duplicated for reliability reasons.

At first sight, TSB may seem “unnecessary”, but a closer study will reveal three very important advantages:

1. 
   The number of PCM links to the parent exchange can be adapted to the traffic volume. Thus, all LSMs do not need a separate PCM link.
   
2. 
   If the “own” PCM link has no free channels, another PCM link can be used instead. This makes the subscriber stage immune to situations with unbalanced traffic load (full availability).
   
3. 
   If the contact with the parent exchange is broken, this will not affect the internal traffic within the subscriber stage.
   

Let us study the example in Figure 2.5.6. Obviously, the traffic is handled by 3 PCM links, and channel 16 of the first two links is used for signalling. For reasons of reliability, we normally have two signalling channels, which means that channels 0 and 16 cannot be used for speech transmission over these two links. In the third link, on the other hand, channel 16 is available for speech. Consequently, a maximum of 91 simultaneous calls are possible in this example.

Up to 16 LSMs can be interconnected.

In this way, the number of subscribers served by a detached unit can be varied between 128 and 2048.