lecture Instrumentation - Sensors and actorsPLC.pdf

Industrial Automation
Automation Industrielle
Industrielle Automation
3 Industrial Communication Systems
Field Bus: principles
3.1 Bus de terrain: principes
Feldbusse: Grundlagen
Prof. Dr. H. Kirrmann
EPFL / ABB Research Center, Baden, Switzerland
CAN, DeviceNet, SDS, ASI-bus, Interbus-S
Ethernet, ControlNet
TCP - IP
Ethernet
Sensor Busses
simple switches etc.
Plant Network
Office
network
Fieldbus
intelligent field devices
FF, PROFIBUS, MVB, LON
2005 March, HK

3.1 Field bus principles
2/25
Field bus: principles
Classes
Physical layer
3.2 Field bus operation
Centralized - Decentralized
Cyclic and Event Driven Operation
3.3 Standard field busses

3/25
Sensor/
Actor
Bus
Field bus
Field bus
Programmable
Logic Controller
Process bus
SCADA level
Process Level
Field level
File
Edit
Network
Management
Operator 2
12
2
33
23
4
Location of the field bus in the plant hierarchy
direct I/O

4/25
What is a field bus ?
A data network, interconnecting a control system, characterized by:
- transmission of numerous small data items (process variables) with bound delay (1ms..1s)
- harsh environment (temperature, vibrations, EM-disturbances, water, salt,…)
- robust and easy installation by skilled people
- high integrity (no undetected errors)
- high availability (redundant layout)
- clock synchronization (milliseconds down to a few microseconds)
- continuous supervision and diagnostics
- low attachment costs ( € 5.- / node)
- moderate data rates (50 kbit/s … 5 Mbit/s) but large distance range (10m .. 4 km)
- non-real-time traffic for commissioning (e.g. download) and diagnostics
- in some applications intrinsic safety (oil & gas, mining, chemicals,..)

5/25
Expectations
- reduce cabling
- increased modularity of plant (each object comes with its computer)
- easy fault location and maintenance
- simplify commissioning (mise en service, IBS = Inbetriebssetzung)
- simplify extension and retrofit
- large number of off-the-shelf standard products to build “Lego”-control systems
- possibility to sell one’s own developments (if based on a standard)

6/25
The original idea: save wiring
marshalling
bar
I/O
PLC
PLC
but: the number of end-points remains the same !
energy must be supplied to smart devices
dumb devices
field bus
(Rangierung,
tableau de brassage (armoire de triage)
COM
tray
capacity

7/25
Marshalling (Rangierschiene, Barre de rangement)
The marshalling is the interface between
the PLC people and the instrumentation
people.

8/25
Field busses classes
CAN, DeviceNet, SDS, ASI-bus, Interbus-S
Ethernet, ControlNet
TCP IP
Ethernet
Sensor Busses
simple switches etc.
Plant Network
Office
network
Fieldbus
intelligent field devices
FF, PROFIBUS PA, LON
The field bus depends on:
its function in the hierarchy
the distance it should cover
the data density it should gather

9/25
Geographical extension of industrial plants
The field bus suits the physical extension of the plant
Control and supervision of large distribution networks:
• water - gas - oil - electricity - ...
Out of primary energy sources:
• waterfalls - coal - gas - oil - nuclear - solar - ...
Manufacturing and transformation plants:
• cement works - steel works - food silos - printing - paper
pulp processing - glass plants - harbors - ...
• locomotives - trains - streetcars - trolley buses - vans -
buses - cars - airplanes - spacecraft - ...
• energy - air conditioning - fire - intrusion - repair - ...
Transmission & Distribution
Power Generation
Industrial Plants
Vehicles
Building Automation
Manufacturing
flexible manufacturing cells - robots
50 m .. 3 km
1 km .. 5 km
1 km .. 1000 km
1 m .. 800 m
500m .. 2 km
1 m .. 1 km

10/25
Networking busses: Electricity Network Control
houses
substation
Modicom
ICCP
control
center
Inter-Control Center Protocol
IEC 870-6
HV
MV
LV
High
Voltage
Medium
Voltage
Low
Voltage
SCADA
FSK, radio, DLC, cable, fiber,...
substation
RTU
RTU RTU
RTU
COM
RTU RTU RTU Remote Terminal Units
RTU
RTU
IEC 870-5 DNP 3.0 Conitel RP 570
control
center
control
center
low speed, long distance communication, may use power lines or telephone modems.
Problem: diversity of protocols, data format, semantics...
serial links (telephone)

11/25
Substation (air isolated)
Node in the electricity grid

12/25
Substations (indoor)
Gas Isolated high voltage medium voltage
consist of bays (départs, Abgang), interconnected by a buss bar (barre, Sammelschiene)

13/25
Substation electrical busses
G
A
B
bussbars
switch position
and commands
current, voltage,
temperature
Generator
Bay
Bay
Bay
Transformer
isolator
(Trenner
Interrupteur)
circuit
breaker
(Schalter,
Disjoncteur)
Current Transformer
(measure)

14/25
Substation data busses
IED 2
IED 1
IED 3
bay i
IED 2
IED 1
IED 3
bay 1
IED 2
IED 1
IED 3
bay n
gateway
workstation1
gateway
workstation2
logger
printer
station bus
the structure of the data busses reflects the substation structure
switch
control and
protection devices

15/25
Fieldbus Application: wastewater treatment
Pumps, gates, valves, motors, water level sensors, flow meters, temperature sensors,
gas meters (CH4), generators, … are spread over an area of several km2
Some parts of the plant have explosive atmosphere.
Wiring is traditionally 4..20 mA, resulting in long threads of cable (several 100 km).

16/25
Process Industry Application: Water treatment plant
S
M.C.C.
Control Room
Sub Station
SCADA
Bus Monitor
JB JB
Remote
Maintenance
System
Ethernet
Segment 1
Segment 2
Segment 3
Segment 4
FB Protocol
Converter
PLC
Digital Input/Output
PID
PID PID
PID
PID
H1 Speed Fieldbus
LAS
JB JB
AI AI AI AI AI
AI AI AI AI AI
AI AI AI
AI AI AI
AI
AO AO
AO
AO
AO
AO
DI
S S
S
S
AI
AO
AI
Japan
Malaysia
Numerous analog inputs (AI),
low speed (37 kbit/s) segments merged to 1 Mbit/s links.
source: Kaneka, Japan

17/25
Data density (Example: Power Plants)
Acceleration limiter and prime mover: 1 kbit in 5 ms
Burner Control: 2 kbit in 10 ms
per each 30 m of plant: 200 kbit/s
Data are transmitted from the periphery or from fast controllers to higher level, but slower links to
the control level through field busses over distances of 1-2 km.
The control stations gather data at rates of about 200 kbit/s over distances of 30 m.
Fast controllers require at least 16 Mbit/s over distances of 2 m
The control room computers are interconnected by a bus of at least 10 Mbit/s,
over distances of several 100 m.
Planning of a field bus requires to estimate the data density per unit of length (or surface)
and the requirements in response time and throughput over each link.

18/25
Distributed peripherals
Many field busses are just
extensions of the PLC’s Inputs
and Outputs,
field devices are data
concentrators.
Devices are only visible to the
PLC that controls them
relays and fuses

19/25
Application: Building Automation
Source: Echelon
low cost, low data rate (78 kbit/s), may use power lines (10 kbit/s)

20/25
Application: Field bus in locomotives
cockpit
motors
power electronics
brakes
power line
track signals
Train Bus
diagnosis
radio
data rate
delay
medium
number of stations
1.5 Mbit/second
1 ms (16 ms for skip/slip control)
twisted wire pair, optical fibers (EM disturbances)
up to 255 programmable stations, 4096 simple I/O
Vehicle Bus
cost engineering costs dominate
integrity very high (signaling tasks)

21/25
Application: automobile
- 8 nodes
- 4 electromechanical wheel brakes
- 2 redundant Vehicle Control Unit
- Pedal simulator
- Fault-tolerant 2-voltage on-board power supply
- Diagnostic System
Bordnetz
ECU
Monitoring
und
Diagnose
Bremsen
ECU
4
redundantes
Bordnetz
12V und 48V
ECU
ECU
ECU
c
ECU
Betätigungs-
einheit

22/25
Application: Avionics (Airbus 380)

23/25
requires integration of power electronics and communication at very low cost.
The ultimate sensor bus
power switch and
bus interface

24/25
Assessment
• What is a field bus ?
• How does a field bus supports modularity ?
• What is the difference between a sensor bus and a process bus ?
• Which advantages are expected from a field bus ?

lecture Instrumentation - Sensors and actorsPLC.pdf

2005 April, HK
Field Bus Operation
3.2 Bus de terrain: mode de travail
Feldbus: Arbeitsweise
ABB Research Center, Baden, Switzerland

2
Fieldbus - Operation
3.1 Field bus types
Classes
Physical layer
Data distribution
Cyclic Operation
Event Driven Operation
Real-time communication model
Networking

3
Objective of the field bus
Distribute to all interested parties process variables, consisting of:
•accurate process value and units
•source identification: requires a naming scheme
•quality indication: good, bad, substituted,
•time indication: how long ago was the value produced
•(description)
time
quality
value
source description

4
Master or peer-to-peer communication
AP
all traffic passes by the master (PLC);
adding an alternate master is difficult
(it must be both master and slave)
input output
PLCs may exchange data,
share inputs and outputs
allows redundancy
and “distributed intelligence”
devices talk directly to each other
separate bus master from application master !
input output
PLC
PLC PLC PLC
PLC
central master: hierarchical
peer-to-peer: distributed
“slaves”
“master”
“slaves”
“masters”
alternate
master
communication in a control system is evolving from hierarchical to distributed
AP
AP
AP
AP

5
application
processor
application
processor
application
processor
Broadcasts
A variable is read on the average in 1..3 different places
Broadcasting messages identified by their source (or contents) increases efficiency.
=
variable
instances
application
processor
plant
image
plant
image
plant
image
plant
image
=
distributed
data base
The bus refreshes the plant image in the background, it becomes an on-line database
Each station snoops the bus and reads the variables it is interested in.
Each device is subscribed as source or as sink for a number of process variables
Only one device may be source of a certain process data (otherwise, collision).
The replicated traffic memories can be considered as "caches" of the plant state
(similar to caches in a multiprocessor system), representing part of the plant image.
bus

6
Data format
time
quality
value
source
In principle, the bus could transmit the process variable in clear text, possibly using XML.
However, this is quite expansive and only considered when the communication network
offers some 100 Mbit/s and a powerful processor is available to parse the message
More compact ways such as ASN.1 have been used in the past with 10 Mbit/s Ethernet.
Field busses are still slow (1Mbit/s ..12 Mbits/s) and therefore more compact
encodings are used.

7
Datasets
wheel
speed
air
pressure
line
voltage
time
stamp
analog variables
Dataset
binary variables
all door closed
lights on
heat on
air condition on
bit offset
16 32 48
0 64 66 70
size
Field busses devices had a low data rate and transmit over and over the same variables.
It is economical to group variables of a device in the same frame as a dataset.
A dataset is treated as a whole for communication and access.
A variable is identified within a dataset by its offset and its size
Variables may be of different types, types can be mixed.
dataset
identifier

8
Dataset extension and quality
To allow later extension, room is left in the datasets for additional variables.
Since the type of these future data is unknown, unused fields are filled with '1".
To signal that a variable is invalid, the producer overwrites the variable with "0".
Since both an "all 1" and an "all 0" word can be a meaningful combination, each
variable can be supervised by a check variable, of type ANTIVALENT2:
0 1 0 1 1 1 0 0 0 1
check
0 0 0 0 0 0 0 0
1 1 1 1 1 1 1 1
0 0
1 1
correct variable
error
undefined
variable value
A variable and its check variable are treated indivisibly when reading or writing
The check variable may be located anywhere in the same data set.
Dataset
var_offset
chk_offset
10 = substituted
00 = network error
01 = ok
11 = data undefined

9
Decoupling Application and Bus traffic
sending: application writes data into memory
receiving: application reads data from memory
the bus controller decides when to transmit
bus and application are not synchronized
application
processor
bus
controller
traffic
memory
decoupled (asynchronous):
sending: application inserts data into queue
and triggers transmission,
bus controller fetches data from queue
receiving: bus controller inserts data into queue
and interrupts application to fetch them,
application retrieves data
application
processor
bus
controller
queues
coupled (event-driven):
events
(interrupts)

10
Traffic Memory: implementation
Bus and Application are (de)coupled by a shared memory, the Traffic Memory,
where process variables are directly accessible to the application.
Ports (holding a dataset)
Application
Processor
Bus
Controller
Traffic Memory
Associative
memory
two pages ensure that read and
write can occur at the same time
(no semaphores !)
bus
an associative memory decodes
the addresses of the subscribed
datasets

11
Freshness supervision
It is necessary to check that the data in the traffic memory is still up-to-date,
independently of a time-stamp (simple devices do not have time-stamping)
Applications tolerate an occasional loss of data, but no stale data.
To protect the application from using obsolete data, each Port in the traffic
memory has a freshness counter.
This counter is reset by writing to that port. It is incremented regularly,
either by the application processor or by the bus controller.
The application should always read the value of the counter before using
the port data and compare it with its tolerance level.
The freshness supervision is evaluated by each reader independently, some
readers may be more tolerant than others.
Bus error interrupts in case of severe disturbances are not directed to the
application, but to the device management.

12
Process Variable Interface
Access of the application to variables in a traffic memory is very easy:
ap_get (variable_name, variable value, variable_status, variable_freshness)
ap_put (variable_name, variable value)
Rather than fetch and store individual variables, access is done by clusters
(predefined groups of variables):
ap_get (cluster_name)
ap_put_cluster (cluster_name)
The cluster is a table containing the names and values of several variables.
Note: Usually, only one variable is allowed to raise an interrupt when received: the one
carrying the current time (sent by the common clock)
The clusters can correspond to "segments" in the function block programming.

13
Time-stamping and clock synchronisation
In many applications, such as disturbance logging and sequence-of-events,
the exact sampling time of a variable must be transmitted together with its value.
To this purpose, the devices are equipped with a clock that records the creation date of
the value (not the transmission time).
To reconstruct events coming from several devices, clocks must be synchronized.
considering transmission delays over the field bus (and in repeaters,....)
A field bus provides means to synchronize clocks in spite of propagation delays and
failure of individual nodes. Protocols such as IEEE 1588 can be used.
bus
input input input processing
t1 t2 t3 t4
t1 val1

14
Transmission principle
The previous operation modes made no assumption, how data are transmitted.
The actual network can transmit data
cyclically (time-driven) or
on demand (event-driven),
or a combination of both.

15
Cyclic and Event-Driven transmission
event-driven: send when value change by more than x% of range
limit update
frequency !,
limit hysteresis
cyclic: send value every xx milliseconds
nevertheless transmit:
- every xx as “I’m alive” sign
- when data is internally updated
- upon quality change (failure)
miss the peak
(Shannon!)
always the same,
why transmit ?
how much hysteresis ?
- coarse (bad accuracy)
- fine (high frequency)
time
individual
period
hysteresis

16
Fieldbus: Cyclic Operation mode
3.1 Field bus types
Classes
Physical layer
Data distribution
Cyclic Operation
Networking

17
Cyclic Data Transmission
address
devices
(slaves)
Bus
Master
Individual period
2 x Tpd
N polls
time [µs]
read transfer
time [ms]
The duration of each poll is the sum of
the transmission time of address and
data (bit-rate dependent)
and of the reply delayof the signals
(independent of bit-rate).
plant
The master polls the addresses in a fixed sequence, according to its poll list.
1 2 3 4 5 6
address
(i)
data
(i)
address
(i+1)
10 µs/km
Poll
List
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
Individual period
44 µs .. 296 µs

18
Cyclic operation principle
The delivery delay (refresh rate) is deterministic and constant.
No explicit error recovery needed since a fresh value will be transmitted in the next cycle.
Only states may be transmitted, not state changes.
To keep a low poll time, only small data items may be transmitted (< 256 bits)
Cyclic operation is used to transmit the state variables of the process.
These are called Process Data (or Periodic Data)
The bus is under control of a central master (or distributed time-triggered algorithm).
Data are transmitted at fixed intervals, whether they changed or not.
Cycle time is limited by the product of the number of data transmitted by the
duration of each poll (e.g. 100 µs / point x 100 points => 10 ms)
The bus capacity must be configured beforehand.
Determinism gets lost if the cycles are modified at run-time.

19
Source-Addressed Broadcast
The bus master broadcasts the identifier of a variable to be transmitted:
Phase1:
Process Data are transmitted by source-addressed broadcast.
The device that sources that variable responds with a slave frame
containing the value, all devices subscribed as sink receive that frame.
Phase 2:
bus.
master
bus
subscribed devices
subscribed
device
subscribed
device
source sink sink
sink
variable value
bus
variable identifier
bus
master devices
(slaves)
source sink sink
subscribed devices
sink
device
device
devices
(slaves)

20
Read And Write Transfers
turn-around
time
address
source
data
time
Most field busses operate with read cycles only.
read transfer: master
Write-No ack transfer
write transfer:
master (source)
address
next transfer
Read Transfer
Write Transfer With Ack
master (source)
arb
arb
turn-around
time
next transfer
address data address
arb
arb
data address
arb
address
arb ack
Local Area Networks operate with write-only transfers.
Their link layer or transport layer provides acknowledgements by another write-only transfer
next transfer
time
time
destination
Parallel busses use read and write-ack transfers
•
•
turn-around time may be
large compared with
data transfer time.
•

21
Round-tip Delay
The
round-trip
delay limits
the extension
of a read-only
bus
master remotest data source
repeater
repeater
closest data sink
Master Frame
access delay
propagation delay
(t_pd = 6 µs/km)
t_source
distance
next Master Frame
t_ms
Slave Frame
T_m
T_m
T_s
T_m
t_repeat
t_repeat
(t_repeat < 3 µs)
t_repeat
t_sm
t_mm

22
Optimizing Cyclic Operation
Solution: introduce sub-cycles for less urgent periodic variables:
Cyclic operation uses a fixed portion of the bus's time
The poll period increases with the number of polled items
The response time slows down accordingly
Cyclic polling need tools to configure the poll cycles.
The poll cycles should not be modified at run-time (non-determinism)
A device exports many process data (state variables) with different priorities.
If there is only one poll type per device, a device must be polled at the
frequency required by its highest-priority data.
To reduce bus load, the master polls the process data, not the devices
group with
period 1 ms
time
4a 8 16 1 4b 64
3
1 ms period
(basic period)
2 ms period
2 4a
4 ms period
1 ms 1 ms
1 1
1 2

23
Cyclic Transmission and Application
Bus and applications are decoupled by a shared memory, the traffic memory,
which acts as distributed database actualized by the network.
The bus master scans the identifiers at its own pace.
The bus traffic and the application cycles are asynchronous to each other.
Traffic
Memory
cyclic
algorithms
cyclic
algorithms
cyclic
algorithms
cyclic
algorithms
port address
application
1
Ports Ports Ports
application
2
application
4
source
port
sink
port
port data
sink
port
cyclic
poll
bus
controller
bus
master
application
3
bus
Periodic
List
Ports
bus
controller
bus
controller
bus
controller
bus
controller

24
Application Of Cyclic Bus
The principle of cyclic operation, combined with source-addressed
broadcast, has been adopted by most modern field busses
This method gives the network a deterministic behavior, at expenses of a reduced
bandwidth and geographical extension.
It is currently used for power plant control, rail vehicles, aircrafts, etc...
The poll scan list located in the central master (which may be duplicated for
availability purposes) determines the behavior of the bus.
It is configured for a specific project by a single tool, which takes into account
the transmission wishes of the applications.
This guarantees that no application can occupy more than its share of the bus
bandwidth and gives control to the project leader.

25
Example: delay requirement
Worst-case delay for transmitting all time critical variables is the sum of:
Source application cycle time
Individual period of the variable
Sink application cycle time
8 ms
16 ms
8 ms
= 32 ms
subscribers application instances
device
publisher
application instance
bus instance
device device
applications
bus

26
Example: traffic pattern in a locomotive
number of devices: 37 ( including 2 bus administrators)
37 of 16 bits
16 ms 32 ms 64 ms 128 256
49 frames of 256 bits
1024
18 of 32
period
% periodic time
occupancy is proportional to surface
total = 92%

27
Fieldbus: Event-driven operation
3.1 Field bus types
Classes
Physical layer
Data distribution
Cyclic Operation
Networking

28
Event-driven Operation
Detection of an event is an intelligent process:
• Not every change of a variable is an event, even for binary variables.
• Often, a combination of changes builds an event.
• Only the application can decide what is an event, since only the application
programmer knows the meaning of the variables.
Events cause a transmission only when an state change takes place.
Bus load is very low on the average, but peaks under exceptional situations
since transmissions are correlated by the process (christmas-tree effect).
•
•
event-
reporting
station
event-
reporting
station
event-
reporting
station
plant
Multi-master bus: uses write-only transfers
intelligent
stations
sensors/
actors

29
Bus interface for event-driven operation
Application
Processor
Bus
Controller
message (circular) queues
bus
driver
filter
application
Each transmission on the bus causes an interrupt.
The bus controller only checks the address and
stores the data in the message queues.
The driver is responsible for removing the messages
of the queue memory and prevent overflow.
The filter decides if the message can be processed.
interrupt

30
Response of Event-driven operation
Interruption of server device at any instant can disrupt a time-critical task.
Buffering of events cause unbound delays
Gateways introduce additional uncertainties
Since events can occur anytime on any device, stations communicate by
spontaneous transmission, leading to possible collisions
Caller
Application
Called
Application
Transport
software
Transport
software
interrupt
request
indication
confirm
Bus
time

31
Determinism and Medium Access In Busses
Although the moment an event occurs is not predictable, the communication
means should transmit the event in a finite time to guarantee the reaction delay.
Events are necessarily announced spontaneously: this requires a
multi-master medium like in a LAN.
The time required to transmit the event depends on the medium access
(arbitration) procedure of the bus.
Medium access control methods are either deterministic or not.
Non-deterministic
Collision
(Ethernet)
Deterministic
Central master,
Token-passing (round-robin),
Binary bisection,
Collision with winner.

32
Events and Determinism
Although a deterministic medium access is the condition to guarantee delivery
time, it is not sufficient since events messages are queued in the devices.
The average delivery time depends on the length of the queues, on the bus
traffic and on the processing time at the destination.
Often, the computers limit far more the event delay than the bus does.
Real-time Control = Measurement + Transmission + Processing + Acting
bus
F F F F
F F F F
F F
F F
data packets
acknowledgements
input and
output queues
events
producers
& consumers

33
Events Pros and Cons
In an event-driven control system, there is only a transmission or an operation
when an event occurs.
Advantages:
Drawbacks:
Can treat a large number of events - if not all at the same time
Supports a large number of stations
System idle under steady - state conditions
Better use of resources
Uses write-only transfers, suited for LANs with long propagation delays
Suited for standard (interrupt-driven) operating systems (Unix, Windows)
Requires intelligent stations (event building)
Needs shared access to resources (arbitration)
No upper limit to access time if some component not deterministic
Response time difficult to estimate, requires analysis
Limited by congestion effects: process correlates events
A background cyclic operation is needed to check liveliness

34
Fieldbus: real-time communication model
3.1 Field bus types
Classes
Physical layer
Centralized - Decentralized
Cyclic Operation
Networking

35
Mixed Data Traffic
represent the state of the plant represent state changes of the plant
-> Periodic Transmission
of Process Variables
short and urgent data items
Since variables are refreshed periodically,
no retransmission protocol is needed to
recover from transmission error.
-> Sporadic Transmission of
Process Variables and Messages
infrequent, sometimes lengthy
messages reporting events, for:
• System: initialisation, down-loading, ...
Since messages represent state
changes, a protocol must recover lost data in
case of transmission errors
• Users: set points, diagnostics, status
Process Data Message Data
... motor current, axle speed, operator's
commands, emergency stops,...
periodic
phase
periodic
phase
event
sporadic
phase
time
basic period basic period
sporadic
phase

36
Mixing Traffic is a configuration issue
Cyclic broadcast of source-addressed variables is the standard solution in field busses
for process control.
Cyclic transmission takes a large share of the bus bandwidth and should be reserved
for really critical variables.
The decision to declare a variable as cyclic or event-driven can be taken late in a
project, but cannot be changed on-the-fly in an operating device.
A message transmission scheme must exist alongside the cyclic transmission to carry
not-critical variables and long messages such as diagnostics or network management
An industrial communication system should provide both transmission kinds.

37
Real-Time communication stack
The real-time communication model uses two stacks, one for time-critical variables
and one for messages
Logical Link
Control
time-critical
process variables
Management
Interface
time-benign
messages
Physical
Link (Medium Access)
Network (connectionless)
Transport (connection-oriented)
Session
Presentation
Application
7
6
Remote Procedure Call 5
4
3
2'
1
connectionless
connectionless
connection-oriented
medium access
implicit
implicit
Logical Link Control
2"
media
common

38
Application Sight Of Communication
R4
Traffic
Memory
Periodic Tasks
R3
R2
R1
Message Data
(destination-oriented)
Process Data
(Broadcast)
E3
E2
E1
Event-driven Tasks
bus
Supervisory
Data
bus controller
Message Services
Variables Services
Queues
station

39
Field - and Process bus
Fieldbus Process Bus
controlled by a central master
(redundant for availability)
cyclic polling
call/reply in one bus transfer
(read-cycle)
("fetch principle")
number of participants limited by
maximum period
cheap connection (dumb)
only possible over a limited
geographical extension
strictly deterministic
multi-master bus (Arbitration)
event-driven
call/reply uses two different messages.
both parties must become bus master
("bring - principle")
large number of participants
costly connection (intelligent)
also suited for open systems
deterministic arbitration -> Token
non - deterministic

40
Cyclic or Event-driven Operation For Real-time ?
Data are transmitted at fixed intervals,
whether they changed or not.
Data are only transmitted when they
change or upon explicit demand.
cyclic operation event-driven operation
(aperiodic, demand-driven, sporadic)
(periodic, round-robin)
Worst Case is normal case Typical Case works most of the time
Non-deterministic: delivery time vary widely
Deterministic: delivery time is bound
All resources are pre-allocated Best use of resources
message-oriented bus
object-oriented bus
Fieldbus Foundation, MVB, FIP, .. Profibus, CAN, LON, ARCnet
The operation mode of the communication exposes the main approach to
conciliate real-time constrains and efficiency in a control systems.

41
Fieldbus: Networking
3.1 Field bus types
Classes
Physical layer
Data distribution
Cyclic Operation
Networking

42
Networking field busses
Networking field busses is not done through bridges or routers,
because normally, transition from one bus to another is associated with:
- data reduction (processing, sum building, alarm building, multiplexing)
- data marshalling (different position in the frames)
- data transformation (different formats on different busses)
Only system management messages could be threaded through from end to end,
but due to lack of standardization, data conversion is today not avoidable.

43
Networking: Printing machine (1)
B C D
E
PM
LS
LS
LS
PM
LS
LS
LS
PM
LS
LS
LS
PM
LS
LS
LS
MPS
Section Control
Line bus (AF100)
Section Busses (AF100)
Console,
Section Supervision
Reelstand bus (Arcnet)
Reelstand-Gateways
Operator bus (Ethernet)
Plant-bus (Ethernet)
Production
Reelstands
Printing Towers
RPE
RPD
RPC
RPB
SSC SSD SSE
SSB
multiplicity of field busses with different tasks, often associated with units.
main task of controllers: gateway, routing, filtering, processing data.
most of the processing power of the controllers is used to route data

44
Networking: Printing Section (2)
Falz- und
Wendeturm-
steuerung
to production preparation
(Ethernet) bridge
PM
PM
standby
GW GW
standby
Section bus D
Line bus Rollenwechsler-
koppler A
Pressmasterbus (Ethernet)
Interbus-S
ARCnet
Rollen-
wechslerkoppler I
Sektions-
steuerung
MR93
KT94
IBG
V-Sercos
IBG
Interbus
AC160 AC160
H -steuerungen
Service-Arcnet
Turmsteuerung
Section bus B
Section bus C
H-Sercos
IBG
V-Sercos
IBG
Interbus
AC160
Turmsteuerung
IBG
V-Sercos
IBG
Interbus
AC160
Turmsteuerung
IBG
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
ODC
KT94
Oxydry-Arcnet
Oxydry
Sektions-
steuerung
AC160
Auro
Tower-ARCnet
LS LS LS
V-Sercos
Section bus C

45
The worst-case delay for the transmission of all variables is the sum of 5 delays:
The actual delay is non-deterministic, but bounded
Transmission delay over a Trunk Bus (cyclic bus)
gateway
speed
stop
speed
stop
Feeder Bus Feeder Bus
Trunk Bus
gateway
copying,
filtering &
marshalling
delay
copying,
filtering &
marshalling
delay
• feeder bus delay
• gateway marshalling delay
• trunk bus delay
• gateway marshalling delay
• feeder bus delay
32 ms
16 ms
25 ms
10 ms (synchronized)
32 ms
= 100 ms

46
Assessment
• What is the difference between a centralized and a decentralized industrial bus ?
• What is the principle of source-addressed broadcast ?
• What is the difference between a time-stamp and a freshness counter ?
• Why is an associative memory needed for source-addressed broadcast ?
• What are the advantages / disadvantages of event-driven communication ?
• What are the advantages / disadvantages of cyclic communication ?
• How are field busses networked ?

2004 June, HK
Open System Interconnection (OSI) model
3.3.1 Modèle OSI d’interconnexion
OSI-Modell
Physical
Link
Network
Transport
Session
Presentation
6
5
4
3
2
1
Application
7

2
2004 June, HK 3.3.1 OSI model
EPFL - Industrial Automation
The OSI model
• was developed to structure telecommunication protocols in the ‘70
(Pouzin & Zimmermann)
• standardized by CCITT and ISO as ISO / IEC 7498
• is a model, not a standard protocol, but a suite of protocols with the same name
has been standardized by UIT / ISO / IEC for open systems data interconnection
(but with little success)
• all communication protocols (TCP/IP, Appletalk or DNA) can be mapped to the
OSI model.
• mapping of OSI to industrial communication requires some additions
The Open System Interconnection (OSI) model is a standard way to
structure communication software that is applicable to any network.

3
OSI-Model (ISO/IEC standard 7498)
Physical
Link
Network
Transport
Session
Presentation
6
5
4
3
2
1
Application
7
"Transport"
protocols
"Application"
protocols
Definition and conversion of the data
formats (e.g. ASN 1)
All services directly called by the end user
(Mail, File Transfer,...) e.g. Telnet, SMTP
Management of connections
(e.g. ISO 8326)
End-to-end flow control and error recovery
(e.g. TP4, TCP)
Routing, possibly segmenting
(e.g. IP, X25)
Error detection, Flow control and error recovery,
medium access (e.g. HDLC)
Coding, Modulation, Electrical and
mechanical coupling (e.g. RS485)

4
OSI Model with two nodes
Physical
Link
Network
Transport
Session
Presentation
Application
Physical Medium
node 1 node 2
7
6
5
4
3
2
1
7
6
5
4
3
2
1

5
Repeater
repeater
Ethernet
server
Ethernet
server
To connect a workstation of
department A to the printer of
department B, the cable becomes too
long and the messages are corrupted.
workstations
department A
department B
Physically, there is only one bus carrying
both department’s traffic, only one node
may transmit at a time.
printer
The repeater restores signal
levels and synchronization.
It introduces a signal delay of
about 1..4 bits
500m
500m
500m

6
OSI model with three nodes (bridge)
7
6
5
4
3
2
1
2
1
physical medium (0)
2
1
7
6
5
4
3
2
1
Physical
Link
Network
Transport
Session
Presentation
Application
Node 1 bridge Node 2
The subnet on both sides of a bridge have:
• the same frame format (except header),
• the same address space (different addresses on both sides of the bridge)
• the same link layer protocol (if link layer is connection-oriented)
Bridges filter the frames on the base of their link addresses
physical medium (0)
e.g. Ethernet 100 MBit/s e.g. ATM

7
Bridge example
repeater
Ethernet
server
Ethernet
server
Bridge
Ethernet 1
server
Ethernet 2
In this example, most traffic is directed from the workstations to the department
server, there is little cross-department traffic
workstations
department A
department B
There is only one Ethernet which carries
both department’s traffic
department A
There are now two Ethernets and only the
cross-department traffic burdens both busses
printer
server
department B
printer

8
Networking with bridges
port
port
LAN
port
port
port
port
LAN
port
port
LAN
LAN
LAN
port
port
port
Spanning-tree-Algorithmen
avoid loops and ensures
redundancy

9
Switch
crossbar-
switch
(or bus)
queues
full-duplex
a switch is an extension of a hub that allows store-and-forward.
nodes

10
OSI Model with three nodes (router)
physical medium (0)
Frames in transit are handled in the network layer .
The router routes the frames on the base of their network address.
The subnets may have different link layer protocols
Node 1 Router Node 2
Physical
Link
Network
Transport
Session
Presentation
Application
3
2
1
2
1
7
6
5
4
3
2
1
7
6
5
4
3
2
1

11
Repeater, Bridge, Router, Gateway: Topography
same speed
same medium
access
same frames
Bridge
Router
backbone (e.g. FDDI)
segment
Repeater
subnet (LAN, bus, extended link)
end-to-end
transport protocol
gateway
application-
dependent
connects different speed,
different medium access
by store-and-forward
same frames and addresses
initially transparent in both ways.
can limit traffic by filtering
devices (nodes, stations) have different link addresses
devices (nodes, stations) have different physical addresses
different subnetworks,
same address space
same transport protocol,
segmentation/reassembly
routers are initially opaque

12
Repeaters, Bridges, Routers and Gateways: OSI model
Net
Trp
Ses
Pre
Apl
Trp
Ses
Pre
Apl
MDS
LLC
Net
Trp
Ses
Pre
Apl
MAC
10 Mbit/s coax
MIS
MDS
Layer 1
MDS
repeater
or hub
10 Mbit/s fibre
MDS
MIS
MDS
MIS
Layer 2
100 Mbit/s Ethernet
bridge
( "switch")
(store-and-forward)
MDS
MIS
LLC
MAC
Layer 3
MDS
MIS
LLC
MAC
ATM 155 Mbit/s
MDS
MIS
LLC
MAC
Net
Trp
Ses
Pre
Apl
MAC MAC
router
MDS
LLC
IP
TCP
RPC
gateway
intelligent linking devices can
do all three functions
(if the data rate is the same)
Fibre

13
To which level does a frame element belong ?
destination source final origin
preamble
physical link
bridge
LLC NC
network
router
TRP SES PRE APL
application
(gateway)
repeater, hub
CRC
A frame is structured according to the ISO model
ED
link
LLC
Network
Control
transport
session
presentation
application
phy

14
Encapsulation
Frame
Signal
Error detection
Flag Flag
Link-address
Link control
(Acknowledge, Token,etc.)
Network address
Transport header
size
User information
CRC
LinkAdr
LinkCrt
NetAdr
INFO
TrpCrt
Each layer introduces its own header and overhead

15
Example: OSI-Stack frame structure
>48
ISO 8473
connectionless network control
5
ISO 8073
class 4 transport control
MA. frame control
MA. destination
address
(6 octets)
MA. source
address
(6 octets)
L_destination SAP
L_source SAP
L_PDU
L_PDU = UI, XID, TEST
LI
TPDU
Protocol Identifier
Header Length
Version/Protocol ID (01)
Lifetime
DT/ER Type
SP MS ER
Checksum
PDU Segment Length
Destination Address
(18 octets)
Source Address
(18 octets)
ADDRESS
PART
Segmentation
(0 or 6 octets)
Options
(priority = 3 octets)
(CDT)
N(S)
ET
MAC_header LNK_hdr NET_header TRP_header
Destination
Reference
FIXED
PART
13 3
DATA
AFI = 49
IDI, Area ID
(7 octets)
PSI
Physical Address
(6 octets)
LSAP = FE
NSAP = 00
IDP
(initial
domain
part)
DSP
(domain
specific
part)
DATA (DT) TPDU
(normal format)
LSAP = DSAP
FE = network layer
18 = Mini-MAP Object
Dictionary Client
19 = Network Management
00 = own link layer
(81)
IEEE 802.4
token bus
ISO 8802
logical link control
address length

16
Protocol Data Units and Service Data Units
Protocol
Data Unit
(PDU)
N - Layer
N+1- Layer
N-1 Layer
Protocol
Data Unit
(PDU)
Service-
Data Unit
(SDU)
Service-
Data Unit
(SDU)
Layer N provides services to Layer N+1;
Layer N relies on services of Layer n-1
(n)-layer entity
(n)-layer entity
(n+1)-layer entity
(n+1)-layer entity
(n-1)-layer entity
(n-1)-layer entity

17
Service Access Points
user of
service N
user of
service N
provider of service (N-1)
provider of service (N)
functions in layer N
Service Access Points (SAP)
Service Access Points represent the interface to a service (name, address, pointer,...)
Service Access Points (SAP)

18
Address and SAPs in a device
Link
Network
Transport
z.B. TCP/IP z.B. ISO 8073
ISO 8473
ISO-stack
Transport-SAP
Physical Physical Address
Logical Address or link address
Network-SAP
(not Network address)
TSAP
NSAP
ASAP Application
(z.B. File transfer, Email,....)
PhSAP
LSAP

19
Procedure call conventions in ISO
Service User
confirm
(network)
Service Provider
(Network Transmission)
request
confirm
(local)
time
Service User
indication
response
confirm
(user)

20
OSI implementation
OSI should be considered as a model, not as an implementation guide
Even if many claim to have "OSI"-conformant implementation, it cannot be proven.
IEC published about 300 standards which form the "OSI" stack, e.g.:
OSI stack has not been able to establish itself against TCP/IP
Former implementations, which implemented each layer by an independent process,
caused the general belief that OSI is slow and bulky.
The idea of independent layers is a useful as a way of thinking, not the best implementation.
ISO/IEC 8327-1:1996 Information technology -- Open Systems
Interconnection -- Connection-oriented Session protocol: Protocol specification
ISO/IEC 8073:1997 Information technology -- Open Systems
Interconnection -- Protocol for providing the connection-mode transport service
ISO/IEC 8473-2:1996 Information technology -- Protocol for providing
the connectionless-mode network service --
ISO 8571-2:1988 Information processing systems -- Open Systems
Interconnection -- File Transfer, Access and Management
ISO/IEC 8649:1996 Information technology -- Open Systems
Interconnection -- Service definition for the Association Control Service Element

21
OSI protocols in industry
ISO-OSI standards should be used since they reduce specification and
conformance testing work and commercial components exist
the OSI model is a general telecommunication framework -
implementations considers feasibility and economics.
industrial busses use for real-time data a fast response access and
for messages a simplified OSI communication stack
the OSI model does not consider transmission of real-time data
the overhead of the ISO-OSI protocols (8073/8074) is not bearable
with low data rates under real-time conditions.
Communication is greatly simplified by adhering to conventions
negotiating parameters at run-time is a waste in closed applications.
the OSI-conformant software is too complex:
simple devices like door control or air-condition have limited power.
•
•
•
•
•
•
Theory:
Reality:
Therefore:
the devices must be plug compatible: there are practically no options.
•

22
TCP / IP structure
TCP UDP
IP routing ICMP
FTP SMTP HTTP
Files SNMP Applications
Transport
Network
Ethernet ATM radio
modem Link & Physical
The TCP/IP stack is lighter than the OSI stack, but has about the same complexity
TCP/IP was implemented and used before being standardized.
Internet gave TCP/IP a decisive push

23
Conclusions
The OSI model is the reference for all industrial communication
Even when some layers are skipped, the concepts are generally implemented
Real-Time extensions to OSI are under consideration
TCP/IP however installs itself as a competitor to the OSI suite, although some efforts are
made to integrate it into the OSI model
For further reading: Motorola Digital Data Communication Guide
TCP/IP/UDP is becoming the backbone for all non-time critical industrial communication
Many embedded controllers come with an integrated Ethernet controller, an the
corresponding real-time operating system kernel offers TCP/IP services
TCP/IP/UDP is quickly displacing proprietary protocols.
Like OSI, TCP protocols have delays counted in tens or hundred milliseconds,
often unpredictable especially in case of disturbances.
Next generation TCP/IP (V6) is very much like the OSI standards.

24
Assessment
1) Name the layers of the OSI model and describe their function
2) What is the difference between a repeater, a bridge and a router ?
3) What is encapsulation ?
4) By which device is an Appletalk connected to TCP/IP ?
5) How successful are implementations of the OSI standard suite ?

2004 June, HK
3. Industrial Communication Systems
Physical Layer
3.3.2 Niveau physique
Physische Schicht
3.3.2

2
2004 June, HK 3.3.2 Field busses - Physical Layer
Physical Layer Outline
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
8. Encoding
4. Electric Signal coupling
7. Synchronization
9. Repeaters
1. Layering

3
OSI Model - location of the physical level
Physical
Link
Network
Transport
Session
Presentation
Application
Transport
protocols
Application
protocols
(Mail, File Transfer,...)
Definition and conversion of the
data formats (e.g. ASN 1)
(e.g. ISO 8326)
(z.B. TP4, TCP)
(e.g. IP, X25)
medium access (e.g. HDLC)
mechanical coupling (e.g. V24)
6
5
4
3
2
1
7

4
Subdivisions of the physical layer
mechanical
specifications
electrical / optical
specifications
medium-dependent signalling
medium-independent signalling same for different media
(e.g. coax, fibre, RS485)
applies to one media
(e.g. optical fibres)
defines the mechanical interface
(e.g. connector type and pin-out)
applies to one media type
(e.g. 200µm optical fibres)
Physical
Layer

5
Concepts relevant to the physical layer
Topology
Mechanical
Control Send, Receive, Collision
Interface Binary bit, Collision detection [multiple access]
Signal quality supervision, redundancy control
Modulation
Binary, NRZ, Manchester,..
Synchronisation Bit, Character, Frame
Flow Control Handshake
Medium
Channels
Coding/Decoding
Baseband, Carrier band, Broadband
Ring, Bus, Point-to-point
Connector, Pin-out, Cable, Assembly
signals, transfer rate, levels
Half-duplex, full-duplex, broadcast

6
Example: RS-232 - Mechanical-Electrical Standard
DTE DCE DTE
DCE
2
Data
Terminal
Equipment
Data Communication
Equipment (Modem)
Telephone
lines
2
modem eliminator
cable
extension
Tip: Do not use
Modem cables,
only Extension
cables
Data
Terminal
Equipment
computer terminal
2
Mechanical
2
25
7
Electrical:
+12V
-12V
+3V
-3V
transmitter receiver
"1" Mark Off
"0" Space On
Topology:
Cabling rules
Originally developed for modem communication, now serial port in IBM-PCs
cable
extension
Modem Computer
Terminal
3
1

7
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

8
Topology: Simplex, Half and Full Duplex
Full-duplex
Sender/
Receiver
Link (Point -To-Point)
Bus (Half-Duplex, except when using Carrier Frequency over multiple bands)
Ring (Half-Duplex, except double ring)
Terminator
Examples:
Ethernet, Profibus
Examples:
SERCOS, Interbus-S
Examples:
RS232
Half-duplex
Sender/
Receiver
Sender/
Receiver
Sender/
Receiver
consists of point-to-point links
Examples:
RS485

9
Bus topologies
party-line
a bus is a broadcast medium (delays come from propagation and repeaters)
radio free topology
repeater
Terminator
Terminator
advantage: little wiring disadvantages: easy to disrupt, high attenuation and reflections, no fibres
hub
star
advantage: robust point-to-point links, can use fibres disadvantage: requires hub, more wiring
point-to-point

10
500m
Repeater
repeater
Ethernet
server
Ethernet
server
To connect a workstation of
department A to the printer of
department B, the cable becomes too
long and the messages are corrupted.
workstations
department A
department B
Physically, there is only one Ethernet
carrying both department’s traffic, only one
node may transmit at a time.
printer
500m
The repeater restores signal
levels and synchronization.
It introduces a signal delay of
about 1..4 bits
500m

11
Bus: repeaters and hubs
partyline
point-to-point
link
repeaters
higher-level hub
hubs assemble point-to-point links to form a broadcast medium (bus)
partyline

12
Party-line (bus) and star wiring
I/O
PLC
I/O I/O I/O I/O
PLC
wiring length = d • n,
increases linearly with number of devices
d
wiring length = d • n • n / 2 • 2
increases with square of number of devices
hub
Up to 32 devices
(more with repeaters)
Up to 16 devices
per hub
I/O I/O I/O I/O I/O
d = average distance between devices
does it fit into the
wiring tray ?
star wiring may more than offset the advantage of field busses (reduced wiring) and leads to
more concentration of I/O on the field devices.
party-line wiring is well adapted to the varying topography of control systems

13
Rings
classical ring
ring in floor wiring
wiring
cabinet
The wiring amount is the same for a bus with hub or for a ring with wiring cabinet.
Since rings use point-to-point links, they are well adapted to fibres
a ring consists only of point-to-point links
Each node can interrupt the ring and introduce its own frames

14
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

15
twinax 8 0.9 0.2 3.5 very good
Media (bandwidth x distance)
200m 700m 2000m
twisted wire
Telephone cable
Transfer rate (Mbit/s)
0.2 0.1 0.05
Costs
(Fr/m)
0.2
good (crosstalk)
bad (foreign)
Electromagnetic
Compatibility
group shielding (UTP) 1 0.3 0.1 1 good (crosstalk)
regular (foreign)
individually
shielded (STP)
2 0.35 0.15 .5 very good
50 Ohm 20 8 1 1.2
75 Ohm TV 1/2" 12 2.5 1.0 2.2 good
93-100 Ohm 15 5 0.8 2.5
single mode 2058 516 207 5.5
multimode 196 49 20 6.5
good
very good
very good
good
coaxial cables
optical fibres
Radio bad
Infrared 0.02
1 1 1 -
-
0 0 good
others Power line carrier 1 0.05 0.01 - very bad
plastic 1 0.5 - 6.5 very good
ultrasound 0.01 -
0 0 bad
the bandwidth x distance is an important quality factor of a medium

16
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

17
Electrical: Transmission media
Cost efficient wiring:
twisted pair (without
Coaxial cable
Unshielded twisted wire
screen shield
dielectric
Telephone
Shielded twisted wire
(Twinax)
flexible, cheap,
medium attenuation
~1 MHz..12 MHz
inflexible, costly,
low losses
10 MHz..100 MHz
Zw = 85Ω..120Ω
Zw = 50Ω ... 100Ω
core
very cheap,
very high losses and disturbances,
very low speed (~10 ..100 kbit/s)
numerous branches, not terminated,
except possibly at one place
Shield
very cheap,
sensible to perturbations
Uncommitted wiring
(e.g. powerline com.)
1) Classical wiring technology,
2) Well understood by electricians in the field
3) Easy to configure in the field
4) Cheap (depends if debug costs are included)
1) low data rate
2) costly galvanic separation (transformer, optical)
3) sensible to disturbances
4) difficult to debug, find bad contacts
5) heavy
twisting compensates disturbances

18
Electrical: Twisted wire pair
characteristic impedance most used in industrial environment:
120 Ohm for bus, 150 Ohm for point-to-point.
Standard from the telecommunication world: ISO/IEC 11801
Cat 5 (class D): 100 MHz, RJ 45 connector
Cat 6 (class E): 200 MHz, RJ 45 connector
Cat 7 (class F): 600 MHz, in development
These are only for point-to-point links ! (no busses)

19
Electrical: What limits transmission distance ?
Attenuation: copper resistance, dielectric loss.
Frequency dependent losses cause dispersion (edges wash-out):
Signal reflection on discontinuities (branches, connectors) cause self-distortions
Characteristic impedance
Attenuation
Linear resistance
Linear capacitance
Cross talk
Common-mode
Shield protection
All parameters are frequency-dependent

20
Consider in cables
- characteristic impedance (Zw) (must match the source impedance)
- attenuation (limits distance and number of repeaters)
- bending radius ( layout of channels)
- weight
- fire-retardant isolation
L' R'
C'
L' R'
C'
L' R'
C'
L' R'
C'
G'
lumped line model
specific inductance (H/m)
specific resistance (W/m)
specific capacitance (F/m)
specific conductance (S/m)
Zw =
L'
C'
G' G' G'

21
Electrical: Signal Coupling Types
Resistive direct coupling
driver on line without galvanic coupling
collision possible when several transmitters active
Wired-OR combination possible
Inductive transformer-coupling
galvanic separation
retro-action free
good electromagnetic compatibility (filter)
cheap as long as no galvanic separation is required (opto-coupler)
signal may not contain DC-components
bandwidth limited
Capacitive capacitor as coupler
weak galvanic separation
signal may not contain DC components
cheap
good efficiency
good efficiency
not efficient

22
Electrical: Resistive (direct) coupling
Ru
Rd
Zw
Zw
+ Us
+ Us
- Us
Unipolar, unbalanced
Open Collector
(unbalanced)
Bipolar, unbalanced
Rt
Ut
Rt
Ut = 5 V (e.g.)
Bus line, characteristic impedance = Zw
Out In
device
Out In
device
Out In
device
Terminator and
Pull-up resistor
wired-OR behaviour
(“Low” wins over “High”
Coax

23
Electrical: Balanced Transmission
Zw
Shield
(Data Ground)
Differential transmitter and receiver
+ good rejection of disturbances on the line and common-mode
- double number of lines
Differential amplifier
(OpAmp)
Used for twisted wire pairs (e.g. RS422, RS485)
Common mode rejection: influence of a voltage which is applied simultaneously on both
lines with respect to ground.
The shield should not be used as a data ground (inductance of currents into conductors)
UA UB
symmetrical line (Twisted Wire Pair) Rt
+Ub
100 ?

24
Electrical: RS-485 as an example of balanced transmission
The most widely used transmission for busses over balanced lines (not point-to-point)
stub
tap
120?
A
B
Data-GND
A
100?
RxS
TxS
RxS
TxS
RxS
TxS
Terminator
segment length
• • •
Zw ˜ 120? , C' ˜ 100 pF/m
Ishort < 250 mA
Short-circuit limitation
needed
120?
multiple transmitter
allowed

25
Electrical: RS-485 Distance x Baudrate product
20
50
100
200
500
1000
2000
5000
10000
10KBd 100KBd 1 MBd 10 MBd
limited by: Cable quality: attenuation, capacitive loading, copper resistance
Receiver quality and decoding method
distance
Signal/Noise ratio, disturbances
12
1200
Baudrate
limited by copper resistance
100? /km -> 6dB loss limit
limited by
frequency-dependent
losses ˜ 20 dB/decade

26
Electrical: Transformer Coupling
Provides galvanic separation, freedom of retro-action and impedance matching
Sender/Receiver
Twisted Wire Pair
shield
Isolation transformer
isolation resistors
but: no DC-components may be transmitted.
cost of the transformer depends on transmitted frequency band (not center frequency)
Source: Appletalk manual
Sender/Receiver

27
Electrical: MIL 1553 as an example of transformer coupling
Twisted Wire Pair
shield
isolation resistors
Direct Coupling
(short stub: 0.3 m)
short
stub
Sender/Receiver
shield
isolation resistors
long
stub
Double-Transformer
(long stub: 0.3 .. 6m)
Extract from: MIL-STD-1553
MIL 1553 is the standard field bus used in avionics since the years '60 - it is costly and obsolete

28
Electrical: Free topology wiring
terminator
voltage
source
Free topology is used to connect scattered devices which are usually line-powered.
Main application: building wiring
Transmission medium is inhomogeneous, with many reflections and discontinuities.
Radio techniques such as echo cancellation, multiple frequency transmission
(similar to ADSL) phase modulation, etc... are used.
Speed is limited by the amount of signal processing required (typically: 10 kbit/s)

29
Electrical: Power Line Carrier technology
HF-trap
A free-topology medium using the power lines as carrier.
Used for retrofit wiring (revamping old installations) and for minimum cabling
Problems with disturbances, switches, transformers, HF-traps, EMC,..
Low data rates ( < 10 kbit/s)
Proposed for voice communication over the last mile (ASCOM)
Difficult demodulation
Capacitive or inductive coupling, sometimes over shield
Applications: remote meter reading, substation remote control
220V

30
Electrical: Mechanical Connecting devices to an electrical bus
some applications require live insertion (power plants, substations)
time-outs (causing emergency stop) limit disconnection time
short stub junction box
thread-through
2 connectors
no live insertion
1 connector
live insertion
(costly) junction box
1 connector
live insertion
Electrical wiring at high speed requires careful layout
(reflections due to device clustering or other discontinuities, crosstalk, EM disturbances)
stub
double-connector
2 connectors
live insertion
installation ?
installation or operational requirements may prohibit screws (only crimping)

31
Practical solution to live insertion
Offers life insertion
but costs a lot
(also in place)

32
Electrical: Connectors
Field busses require at the same time low cost and robust connectors.
The cheapest connectors come from the automobile industry (Faston clips)
and from telephony (RJ11, RJ 45)
However, these connectors are fragile. They fail to comply with:
- shield continuity
- protection against water, dust and dirt (IP68 standard)
- stamping-proof (during commissioning, it happens that workers and vehicles pass over cables)
The most popular connector is the sub-D 9 (the IBM PC's serial port),
which exists in diverse rugged versions.
Also popular are Weidmann
and Phoenix connectors.

33
Electrical: Water-proof Connectors
connector costs can become the dominant cost factor…

34
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

35
Fiber: Principle
Is
GaAs
LED
PIN
fotodiode
different refraction
coefficients
Cable
Transmitter
laser-diode (GaAsP, GaAlAs, InGaAsP)
Receiver
Wavelength
1300 nm-window (Monomode)
Transmitter, cable and receiver must be "tuned" to the same wavelength
850 nm (< 3,5 dB/km, > 400 MHz x km)
laser (power),
glass (up to 100 km) or plastic (up to 30 m).
PIN-diode
light does not travel faster than electricity in a fiber (refraction index).
3 components:
transmitter fibre receiver

36
Fiber: Types
Multimodefibre
N(r)
50 - 300 µm 50 - 100 µm 2-10 µm
Monomode fibre
waveguide
total reflection gradual reflection
50 µm
Core
Clad
Refraction profile
Cross-section
Longitudinal section
5dB/km 3 dB/km 2,3 dB/km
800nm
(infra-red) 1300nm 0,6 dB/km 0,4 dB/km
20MHz·km 1 GHz·km 100 GHz·km
telecom - costly
50 or 62.5 µm LAN fibre
HCS (Hard-Clad Silica)
ø 200 µm, < 500m
(red) 650nm 10 dB/km

37
Fibre: Use
Material plastic glass / plastic glass
distance 70m 400m 1km
Usage local networking remote networking telephone
Connector simple high-precision
Cost cheap medium medium
aging poor very good good
bending very good good poor
bandwidth poor good very good
Type POF HCS/PCF GOF
precision
in industry, fibers cost the same as copper - think about system costs !
POF: Plastic Optical Fibres
GOF: Glass Optical Fibres
HCS: silica fibre

38
Fiber: building an optical bus
Passive coupler
Active star coupler
electrical segment (wired-or)
fibre pair
opto-electrical
transceiver
Every branch costs a
certain percentage of light
n% coupling losses
n% coupling losses
Passive star coupler
1
2
3
4
5
6
1
2
3
4
5
6
Fused zone
costly manufacturing (100 $ branches)
costly manufacturing
(100 $ / 4 branches)

39
Fiber: building an optical ring and bridging
Powered unpowered
Double ring
Mechanical bridging is difficult
This is why optical fibers are mostly used in rings (FDDI, Sercos)
example of
solution
prism
spring

40
Fiber: advantages
1 ) high bandwidth and data rate (400 MHz x km)
2 ) small, frequency-insensitive attenuation (ca. 3 dB/km)
4 ) immune against electromagnetic disturbances (great for electrical substations)
5 ) galvanic separation and potential-free operation (great for large current environment)
6 ) tamper free
7 ) may be used in explosive environments (chemical, mining)
8 ) low cable weight (100 kg/km) and diameter, flexible, small cable duct costs
10) standardized
3 ) cover long distances without a repeater
9 ) low cost cable

41
Fiber: Why are fibres so little used ?
1) In process control, propagation time is more important than data rate
2) Attenuation is not important for most distances used in factories (200m)
3) Coaxial cables provide a sufficiently high immunity
5) Galvanic isolation can be achieved with coaxial cables and twisted pairs through opto-couplers
6) Tapping is not a problem in industrial plants
8) In explosive environments, the power requirement of the optical components hurts.
9) Installation of optical fibres is costly due to splicing
4) Reliability of optical senders and connections is insufficient (MTTF ≈ 1/power).
7) Optical busses using (cheap) passive components are limited to a few branches (16)
10) Topology is restricted by the star coupler (hub) or the ring structure

42
Radio Transmission
Radio had the reputation to be slow, highly disturbed and range limited.
Mobile radio (GSM, DECT) is able to carry only limited rate of data (9.6 kbit/s) at high costs,
distance being limited only by ground station coverage.
IEEE 802.11 standards developed for computer peripherals e.g. Apple’s AirPort
allow short-range (200m) transmission at 11 Mbit/s in the 2.4 GHz band with 100mW.
Bluetooth allow low-cost, low power (1 mW) links in the same 2.4 GHz band, at 1 Mbit/s
Modulation uses amplitude, phase and multiple frequencies (see next Section)
Higher-layer protocols (WAP, …) are tailored to packet radio communication.
Radio == mobile -> power source (batteries) and low-power technologies.
bluetooth module

43
Wireless Field busses
short distance,
limited bandwidth,
area overlap and frequency limitations
not tamper-free,
difficult to power the devices
costs of base station
but: who changes the batteries ?
no wiring,
mobile,
easy to install

44
Redundancy at the physical layer
cable come together at each device
centralized wiring
star coupler B
Star topology
Party-Line
decentralized wiring both cables can run in the same conduct where
common mode failure acceptable
Terminator
Terminator
star coupler A
common mode failures cannot be excluded since wiring has to come together at each device
star couplers should be separately powered

45
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

46
Modulation
Base band
Carrier band
Broadband
Signals may be modulated on a carrier frequency
(e.g. 300MHz-400MHz, in channel of 6 MHz)
Signal transmitted as a sequence of frequencies,
several at the same time.
Signal transmitted as a sequence of binary
states, one at a time (e.g. Manchester)
Signal transmitted as a sequence of frequencies,
one at a time
(e.g. FSK = frequency shift keying = 2-phase
Modulation.
Frequency
5-108
MHz
162-400
MHz
Backward
channel
Forward-
channel

47
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

48
Synchronisation: where does it take place ?
"determine the beginning and the end of a data stream"
Bit synchronisation Recognize individual bits
Frame synchronisation Recognize a sequence of bits transmitted as a whole
Message synchronisation Recognize a sequence of frames
Session synchronisation Recognize a sequence of messages
Clock
+NRZ Data
+Framing
Data in Manchester II
Start-sync
(Violation)
Stop-sync
(Violation)
= Line Signal
Example: Frame synchronisation using Manchester violation symbols
Character synchronisation Recognize groups of (5,7,8,9,..) bits
1 1 0 1 0 0 0 1
Data

49
Frames: Synchronization
character-synchronous
(e.g. bisync)
A character is used as synchronisation character
If this character appears in the data stream, it is duplicated
The receiver removes duplicated synchronisation characters
delimiter
(e.g. IEC 61158)
A symbol sequence is used as delimiter, which includes non-data symbols
bit-synchronous
(e.g. HDLC)
A bit sequence is used as a flag (e.g. 01111110).
To prevent this sequence in the bit-stream, the transmitter inserts a "0" after
each group of 5 consecutive "1", which the receiver removes.
Delimiter (not Manchester)
"1" "1" "0" "0" "1" "1"
Manchester symbols
1 1 1 0 0 0 1 1 1 1 1 1 1 0 0 1 1 1 1 1 0
0 1 0
Data
Signal 1 1 1 0 0 0 1 1 1 1 1 1 1 0 0 1 1 1 1 1 0 1 1 1 1 1
0 1 1 1 1 1 10 0
0
Bit-stuffing
flag
SYN A B C SYN SYN D E F G SYN
Byte-stuffing
A B C SYN D E F G
Data
Signal
flag flag
Signal

50
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

51
Encoding: popular DC-free encodings
1 1 0 1 1 0 0 0
Manchester
1: falling edge at midpoint
0: rising edge at midpoint
DC-free, memoryless*
Miller (MFM)
centre frequency halved
not completely DC-free
memory: two bits
(sequence 0110)
Differential Manchester
always transition at midpoint
1: no transition at start point
0: transition at start point
(polarity-insensitive, DC-free,
memoryless)
Xerxes
replaces “101” sequence
by DC-balanced sequence
DC-free, memory: two bits
Ethernet, FIP
IEC 61158,
MVB, MIL 1553
High-density
diskettes
LON
FlexRay
memoryless*: decoding does not depend on history
user

52
Encoding: DC-free coding for transformer coupling
DC-free encoding is a necessary, but not sufficient condition
The drivers must be carefully balanced (positive and negative excursion |+U| = |-U|)
Slopes must be symmetrical, positive and negative surfaces must be balanced
Preamble, start delimiter and end delimiter must be DC-free
(and preferably not contain lower-frequency components)
Transformer-coupling requires a lot of experience.
Many applications (ISDN…) accept not completely DC-free codes, provided that
the DC component is statistically small when transmitting random data, but have to
deal with large interframe gaps.
effect of unbalance
(magnetic discharge)

53
Decoding base-band signals
Zero-crossing detector
Sampling of signal
needs Phase-Locked Loop (PLL) and preamble (? delimiter)
Signal Frequency Analysis
requires Signal Processor, Quadrature/Phase analysis
decoding depends on the distance between edges
1 0 1 0 1 0 1 0 1 N+ N- 1 0 N-N+ 0 1 1 1
Preamble Delimiter
RxS
Uh+
Uh-
idle level
active
idle
Daten
Dynamic: 10 dB
Dynamic: 32 dB
Dynamic: 38 dB
histeresis
unipolar signal
time
Uh+
Uh-
line
bipolar signal
Dynamic: 18 dB

54
Encoding: Physical frame of IEC 61158-2
Start delimiter (8 bit times)
1 N+ N- 1 0 N- N+ 0
1 N+ N- N+ N- 1 0 1
1 0 1 0 1 0 1 0
+U
-U
end delimiter
payload
0V
defines end of frame
needed since preamble
is variable length
start
preamble
1 0 0 1 1 0 1 1
Payload (variable length)
Preamble (variable) for PLL synchronisation
End delimiter (8 bit times)

55
Encodings: Multi-frequency
frequency
54
kHz
49,5 kHz
45 kHz
40,5 kHz
36
kHz
31,5 kHz
27 kHz
22,5 kHz
90 kHz
85,5 kHz
81 kHz
76,5 kHz
72
kHz
67,5 kHz
63
kHz
58,5 kHz
"0"
"SB1"
"1" "0" "1" "0" "1" "0" "1" "0" "1" "0" "1" "0" "1" "0" "1"
"SB2" "SB3" "SB8"
"SB4" "SB5" "SB6" "SB7"
unused
power
Best use of spectrum
Adaptive scheme (frequency-hopping, avoid disturbed frequencies, overcome bursts)
Base of ADSL, internet-over-power lines, etc...
Requires digital signal processor
Limited in frequency
EMC considerations

56
Bandwidth and Manchester Encoding
" 0 " " 0 " " 0 " " 1 " " 0 "
" 1 " " 1 "
2-step
Delimiter
3-step
Non-data symbols may introduce a lower-frequency component
which must go through a transformer.
The transformer must be able to transmit frequencies in a 1:20 ratio
3-step

57
Encoding: qualities
1) Self-clocking, Explicit clocking or asynchronous
Clocked: separate clock channel
Self-clocking: clock channel multiplexed with signal
Asynchronous: requires synchronisation at next higher level.
Code such as HDB3 insert "Blind Bits" to synchronize a random sequence.
2) Spectral efficiency
Which frequency components can be found in a random data sequence ?
especially: is there a DC-component ?
(Important for transformer and transceiver coupling)
Pseudo-DC-free codes such as AMI assume that "1" and "0" are equally frequent.
3) Efficiency: relation of bit rate to Baudrate
Bitrate = Information bits per second
Baudrate = Signal changes per second
4) Decoding ease
Spectral-efficient codes are difficult to decode
This is especially the case with memory-codes (value depends on former symbols)
(e.g. Miller, differential Manchester).
5) Integrity
For error detection, the type of error which can occur is important, and especially if a single
disturbance can affect several bits at the same time (Differential Manchester).
6) Polarity
A polarity-insensitive electrical wiring simplifies installation

58
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
8. Encoding
7. Synchronization
9. Repeaters
1. Layering

59
Repeater
The repeater:
• decodes and reshapes the signal (knowing its shape)
• recognizes the transmission direction and forward the frame
• detects and propagates collisions
A repeater is used at a transition from one medium to another within the same subnet.
repeater
segment 2
decoder
encoder
decoder
encoder
segment 1
(RS 485) (transformer-coupled)
node node node node node
node

60
Repeater and time skew
Repeaters introduce an impredictable delay in the transmission since they need some time
to synchronize on the incoming signal and resolve collisions.

61
Star coupler (hub)
wired-or electrical media
fibre pair
opto-electrical
transceiver
to other device
or star coupler
to other device
or star coupler
device
device device device
A star coupler is a collection of repeaters that connect point-to-point links into a bus
(e.g. for optical fibres). it is called "hub" in the Ethernet standard.
It is a star topology, but a bus structure

62
To probe further
Henri Nussbaumer, Téléinformatique 1, Presses polytechniques romandes
Fred Halsall, Data Communications, Computer Networks and Open Systems, Addison-Wesley
EIA Standard RS-485
B. Sklar , “Digital Communications,” Prentice Hall, Englewood Cliffs, 1988

63
Assessment
What is the function of the physical layer ?
What is the difference between a bus and a ring ?
How is a bus wired ?
Which electrical media are used in industry ?
How is the signal coupled to an electrical media ?
How is the signal decoded ?
What is an open-collector (open-drain) electrical media ?
What are the advantages and disadvantages of transformer coupling ?
What limits the distance covered by electrical signals and how is this to overcome ?
What are the advantages and disadvantages of optical fibres ?
When are optical fibers of 240 mm used rather than 62.5 mm ?
What is a broadband medium ?
What is DSL ?
What is the purpose of modulation ?
What is the purpose of encoding ?
What is the difference between bit rate and Baudrate and what does it say about efficiency?
What limits the theoretical throughput of a medium ?
What is the difference between Manchester encoding, Miller and differential Manchester ?
Which are the qualities expected from an encoding scheme ?

64

65

2004 June, HK
3. Industrial Communication Systems
Link Layer and Medium Access
3.3.3 Niveau de liaison et accès au médium
Verbindungsschicht und Mediumzugriff

2
2004 June, HK 3.3.3 Field busses - Link Layer
Link Layer Outline
Link Layer in the OSI model
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Medium Access control
Connection-Oriented and connectionless
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Collision with winner
Token Passing
Comparison

3
Link Layer function
1) Data integrity
2) Medium Access
3) Logical Link Control
The link layer implements the protocols used to communicate within the same subnet.
(subnet: same medium access, same bit rate)
- but different media may be interconnected by (bit-wise) repeaters
Tasks of the link layer:
4) Link Layer Management

4
Link Layer in the OSI Model
Functions
addressing
frame format
integrity control
medium allocation
(master redundancy)
connection establishment
flow Control
error handling
Medium Access
Control
(MAC)
Logical Link
Control
(LLC)
Network
Frame
Physical
Medium
Physical
Signaling
Physical
Link
Network
Transport
Session
Presentation
6
5
4
3
2
1
Application
7
Subnet (Bus or Ring)
bridge control store-and-forward
address discovery

5
OSI model with two nodes
Physical
Link
Network
Transport
Session
Presentation
Application
Physical Medium
Node 1 Node 2
7
6
5
4
3
2
1
7
6
5
4
3
2
1

6
OSI model with repeater (physical layer connection)
Physical
Link
Network
Transport
Session
Presentation
Application 7
6
5
4
3
2
1 1
physical medium (0)
Node 1 repeater
1
7
6
5
4
3
2
1
Node 2
The two segments on each side of a repeater form a single subnet, identified by
• same speed (medium, modulation may differ)
• same frame format (except fringe effects)
• same medium access
• same address space (transparent on both side of the repeater)
Repeaters introduce a delay in the order of a few bit time.
physical medium (1)

7
OSI model with three nodes (bridge): link layer connection
Physical
Link
Network
Transport
Session
Presentation
Application 7
6
5
4
3
2
1
2
1
physical medium (0)
Node 1 bridge
2
1
7
6
5
4
3
2
1
Node 2
The subnet on both sides of a bridge have:
• the same frame format (except header),
• the same address space (different addresses on both sides of the bridge)
• the same link layer protocol (if link layer is connection-oriented)
Bridges filter the frames on the base of their link addresses
physical medium (0)

8
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

9
HDLC as example of a link layer protocol
Function
Standard
Address
Flag CRC
Data
Control
8 bit
01111110 16 bit
(n · 8)
8 bit
Flag
01111110
Integrity Check
Error recovery
Medium Access
Objects
16-bit Cyclic Redundancy Check
Master/Slave, (with slave initiative possible)
positive acknowledgement and retry
7-frames (127 frames) credit system
Flow control
Reliable transmission between devices of a subnet
HDLC (High Level Data Link)
Frame structure according to ISO 3309.

10
HDLC Topology
Primary
(Mainframe)
Secondary
(Terminal)
Secondary
(Terminal)
Secondary
(Terminal)
Secondary
(Terminal)
Secondary
(Terminal)
full-duplex
or
half-duplex
medium
The Primary (Master) is connected with the Secondaries (Slaves) over a
multidrop bus (e.g. RS 485) or over point-to-point lines
Secondary
(Terminal)
Secondary
(Terminal)
Secondary
(Terminal)
HDLC bases physically on a bus, but is logically a star network

11
HDLC - Full- and Half duplex operation
Command/Data
Sender/Receiver
Responder/
Secondary
Sender/Receiver
Requester/
Primary
Acknowledgement
Command/Acknowledgement
Sender/Receiver
Responder/
Secondary
Sender/Receiver
Requester/
Primary
Data
Sender/Receiver
Responder/
Secondary
Sender/Receiver
Requester/
Primary
Half-Duplex
Full-Duplex
The Primary switches the Secondary to send mode, the Secondary sends until it returns control

12
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

13
Frame Sublayer
The frame layer is concerned with the correct frame format and contents,
with (practically) no consideration of the medium or speed.
sublayer
Medium Access
Control
(MAC)
Logical Link
Control
(LLC)
Network
Frame
Physical
Medium
Physical
Signaling
Medium Access
Control
(MAC)
Logical Link
Control
(LLC)
Network
Frame
Physical
Medium
Physical
Signaling

14
Error Handling
Industry imposes high standards regarding data integrity
Transmission quality is not very high , but consequences are severe.
Errors not detected at the link layer are very difficult to catch at higher layers
Error detection
Frame data are protected by redundant information, such as parity bits,checksum,
CRC (Cyclic Redundancy Check)
Error recovery
Except when medium is very poor (Power Line Carrier, radio), error correcting
Erroneous frames are ignored, the potential sender of the error is not informed
(the address of the sender is unknown if the frame is damaged)
The sender is informed of the lack of response when it does not receive the
expected acknowledgement within a time-out.
Definition of the time-out has a strong impact on real-time behaviour
codes are not used.

15
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

16
Error Detection
Error detection require redundancy in the transmitted information.
Signal redundancy: Signal quality supervision (squelch, jamming,..)
Coding redundancy: Timing-violations in decoder
Data redundancy: error detecting code
k data bits r check bits
n-bit codeword = (n,k) block code
• Code efficiency: CEF = k/n
• Hamming-Distance
Quality criteria
• Residual Error Rate

17
Hamming Distance
The Hamming Distance between two words is the number of bits in which they differ:
Word 1: 01100110
Word 2: 00101001 -> Hamming-Distance = 5
The Hamming Distance of a code is the minimum number of bits which need to be tilted
in a valid codeword to produce another valid (but erroneous) codeword
00000 00001 00011 00111 01111
code word code word
m = 4
Number of detectable bit errors: ZD = HD – 1
Numbers of correctable bit errors: ZC = (HD–1)/2
Example: HD = 4: ZD = 3, ZC = 1
Example
1st error 2nd error 3rd error 4th error

18
Hamming distance in 3 dimensions: parity
000
001
101
111
110
010
011
100
legal
illegal
Odd parity: sum Modulo-2 of all "1" in the codeword (including the parity bit) is odd
1 0 1 1 0 0 0 0
par D7 D6 D5 D4 D3 D2 D1
1
D0
The parity bit is the last transmitted bit (->LSB first, a matter of convention)

19
Error Detection through CRC
The data stream consists of a sequence of n "0" and "1"
This sequence is treated as a polynomial of degree n.
This polynomial is divided by a Generator polynomial of degree m, m<n,
The rest of this division (which has (m-1) bits) is appended to the data stream.
(Cyclic Redundancy Check)
rest
/
At reception, the data stream is divided through the same generator-polynomial,
the rest of that division should be identical to the transmitted rest.
To simplify the receiver circuit, the rest is also divided by the generator polynomial,
the result should then be a known constant if the transmission was error-free.
The Generator Polynomial is chosen according to the expected disturbances:
burst or scattered errors, and the channel error model (bit inversion)
Principle
data (dividend)
GP(divisor)

20
Residual Error Rate, Parity
Hamming Distance
R
e r
= 1 - (1- E
r
)
n
- n · E
r
· (1- E
r
)
n -1
Residual error rate
exactly one error
no error
E
r
Bit error probability
Rer for two word length:
E
r
= 10
-5
n = 9 bit R
er
= 72 · 10
-10
E
r
= 10
-5 n = 513 bit R
er
= 2.6· 10
-5
2
quite useless ...
quite efficient….
Rer = Probability of occurrence of an undetected error in spite of an error detecting
code as a function of the bit error probability
Example:

21
Integrity classes and bit error rate
Residual Error Rate
10 -16
10 -14
10 -12
10 -10
10 - 8
10 - 6
10 - 4
10 - 2
10 0
Integrity class I1
I
n
t
e
g
r
i
t
y
c
l
a
s
s
I
2
Integrity class I3
FT2
FT1.2
10-5 10-4 10-3 10-2 10-1 10 0
Bit error rate
The standard IEC 61870-5 (protocols for remote control of substations)
defines several classes of robustness in function of the bit error rate (bad/good bits)

22
Synchronization Errors
01111110 01111110
flag flag
FCS
HDLC-frame
01111110 01111110
flag flag
FCS
01111110
"FCS"
discarded
false
flag
1 Chance in 65536,
that the random data
form a correct CRC
disturbance
HDLC-frame with error
A single error can falsify a frame -> HD = 1
It is uninteresting how likely this case is, the fact is, that it can occur.
The synchronization should have a higher Hamming distance than the data itself.
Because of this bug, HDLC when used in industry require additional error checks.
precisely 1111110 is the most frequent sequence in a
random bit stream because of bit-stuffing.
any data
Frame Check Sequence
(CRC)

23
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

24
Error Correcting Codes
Error correcting codes are used where the probability of disturbances is high (e.g. power
line communication) and the time to retransmit is long (e.g. space probe near Jupiter).
In industry, error correcting codes are normally directly embedded in the physical layer,
e.g. as part of a multitone transmission (ADSL) or of a radio protocol (Bluetooth).
Error correction necessarily decreases the data integrity, i.e. the probability of accepting
wrong data, since the redundancy of correction is taken from the code redundancy.
It is much more important to reject erroneous data (low residual error rate) than to try to
correct transmission.
However, when the medium is very bad (radio), error correction is necessary to
transmit even short messages.
Assigning bits for error detection and correction is a tradeoff between integrity and
availability.

25
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

26
Medium Access Control
Medium Access Control gives the right to send in a multi-master bus
Network
Logical Link
Control
(LLC)
Network
Frame
Physical
Medium
Physical
Signaling
Logical Link
Control
(LLC)
Frame
Physical
Medium
Physical
Signaling
Network
Medium Access
Control
(MAC)
Medium Access
Control
(MAC)
Medium Access
Control
(MAC)
Logical Link
Control
(LLC)
Frame
Physical
Medium
Physical
Signaling

27
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

28
Medium Access Control - quality criteria
Fairness all requesters will eventually be allowed to transmit
Timelyness all requesters will be allowed to transmit within a certain
time, depending on their priority.
Robustness communication errors or the permanent failure of
one component does not prevent the others to
access the medium.
Determinism all requesters will be allowed to transmit within a finite time

29
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

30
MAC single master (e.g. Profibus DP)
bus
bus
master
devices
(slaves)
slave
slave slave
command reply command ack command reply
the bus master allocates time slots to each slave
it may assign priorities (or no priority: round-robin, all are treated equally)
the master may be the source and the destination of data
+ strictly deterministic, complete and flexible control
- polling takes time, since devices which have nothing to transmit must be polled
improvement: “ look-at-me ” (short poll frame allowing slave to request poll of a longer frame)
= "slave initiative" used in Profibus DP
time
read write with ack
command
write no ack read & write
- the master has little knowledge about data importance

31
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

32
MAC Rings (1): register insertion principle
devices
output output
output
shift register
input input
input
master
Devices are connected by point-to-point links (no bus!), there is one sender per segment.
The operation is similar to a large shift register.
The master sends a long frame with the output data to the first device in the ring.
Each device reads the data intended for it, inserts its data instead and
passes the frame to the next device.
The last device gives the frame back to the master.
application memory
time
time
data
data

33
MAC Rings (2): pros and cons
Two major field busses use the ring topology, Interbus-S and SERCOS
and the register-insertion principle described.
the position of the bit in the frame corresponds to the position of the device in the ring,
there are no device addresses - easy to use, but prone to misconfiguration.
each device introduces a delay at least equal to a few bits
+ deterministic access, good usage of capacity, addresses are given by device sequence
on the ring, only point-to-point links
- long delays (some µs per device)

34
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

35
MAC Ethernet (1): CSMA-CD principle (stochastic)
improvement 2:
(Binary Backoff)
Every station sends as it pleases
if no acknowledgement comes, it retransmit
No upper limit to the waiting time, mean waiting time depends
on the arrival rate of frames and on their average length.
(pure Aloha)
Advantage:
retry after a random time,
doubled after each collision, max 15 times
be aware that you are jammed
improvement 1:
improvement 3:
(Carrier Sense)
(Collision Detection)
Principle
do not send when the medium is occupied
Arbitration does not depend on number or on address of the stations
Drawback:
The medium access is not deterministic,
but for light traffic (<1%) there is no noticeable delay.

36
MAC Ethernet (2): collision conditions
repeater
Station A must
detect that its
frame collided
while it is still
transmitting its
frame
at 10 Mbit/s, limits radius to about 2500 m
Ethernet minimum frame size = 64 Bytes, or 51,2 ms @ 10 Mbit/s
A B
minimum
frame size
= 64 Bytes
preamble
= 8 Bytes
collision
at 100Mbit/s, limits radius to about 250 m
(2 x 7.5 s/km+ 2
repeaters)
Tpd
Station B started
transmission just
before receiving
A’s frame.
it nevertheless
transmits its
header completely
time

37
MAC Ethernet (3): propagation conditions and bus diameter
The frame must be long enough to allow all stations to detect a collision while the frame is
being transmitted.
500 m
2 x Tpd = 2 x propagation time (@10Mbit/s Ethernet: 51,2 µs)
500 m
Collisions can only be detected reliably when the frame size is longer than the
propagation delay -> padding to a minimum size (512 bits = 64 Bytes)
The "diameter" of the network is limited to 2,5 km
Since a station which expects a collision must wait one slot time before transmitting,
the maximum frame throughput on Ethernet is limited by the slot time.

38
MAC Ethernet (4): collision probability and simultaneous transmitters
previous frame 0 1 2 3 4 5 • • • time
After the end of a frame, a transmitter chooses a slot at random from a fixed number of slots
Ethernet is not efficient for small frame size and large number of simultaneous transmitters
Ethernet is considered to enter overload when reaching 10%-18% load
0
0.2
0.4
0.6
0.8
1
1 2 3 4 5 10 32 64 128 256
4096
1024
512
48
utilisation 100%
number of transmitters
frame size

39
MAC Ethernet (5): when collisions can't be detected
A small number of simultaneous transmitters causes a high probability loss of a packet .
LON can retry up to 255 times: probability of lost message is low, but delay is long
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
without collision detection
with collision detection
number of simultaneous transmitters
probability of loosing a packet source: P. Koopman, CMU
LON uses a p-persistent MAC with 16-slot collision avoidance (p = 1/16)
It is not always possible to detect collisions.

40
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

41
MAC CAN (1): Deterministic Arbitration
Such a medium allows a bit-wise "Wired-OR" operation
When several transmitters are simultaneously active, the dominant state wins
over the recessive state if there is at least one transmitter sending the dominant state
(dominant is “Low” in open collector, "Bright" in an optical fiber, or a collision on a medium
that allows collisions).
A device is capable to know if the signal it puts on the line is disturbed (XOR).
Terminator and
pull-up resistor
Bus line
Rt
Ut
Rt
Ut
The CAN fieldbus uses media with a dominant and a recessive state
Example:
open-collector:
Terminator and
pull-up resistor
device 1 device 2 device 3 device 4
Jam
Out In
XOR
Jam
Out In
XOR
Jam
Out In
XOR
Jam
Out In
XOR

42
MAC CAN (2): Collision with Winner
• Each station has a different identity (in this example: 4 bit ID)
• Each station listens before sending and can detect collisions
• Competing stations start transmitting at the same time (1st bit is a SYNC-sign)
• Each station begins its transmission with its identity, bit by bit
• In case of collision, "1" wins over "0" ("1" = Low, bright, dominant).
• A station, whose "0" bit was transformed into a “1" retires immediately
• The winning station has not been disturbed and transmits.
• Loosing stations await the end of the ongoing transmission to start again.
Station 10 (wins)
slot time
1 0 1 0
1 0 0 (1)
Station 09
1 0 1 0
0 (1) (1) (0)
Station 06
Bus signal
Also known as "Urn" or "binary bisection" method
Advantage: deterministic arbitration (assumes fairness), good behavior under load
the size of the unique ID defines arbitration time,
transmission delay defines slot time -> 40m @ 1 Mbit/s, 400m @ 100 kbit/s
Drawback:

43
MAC CAN (3): Deterministic Arbitration
Advantage: deterministic arbitration (assumes fairness, I.e. a device only transmits
again when all losers could), good behavior under load.
the slot time (one bit time) must be long enough so that every station can
detect if it has been disturbed - I.e. twice as long as the propagation time
from one end of the bus to the other ( signal speed = 5 ns / m).
Therefore, the bit rate is dependent on the bus extension:
40m @ 1 Mbit/s, 400m @ 100 kbit/s
the size of the unique ID defines arbitration time.
Drawback:

44
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

45
MAC Profibus (1): Token principle
All stations form a logical ring
Each station knows the next station in the ring (and the overnext)
Each station delegates the right to send to the next station in the ring,
(in the form of a token frame or as an appendix to a data frame).
z.B.: Token Bus (IEEE 803.4), Profibus (IEC 61158)
Variants Token with Priority (back to the station with the highest priority)
Problems: Loss or duplication of token, initialization
do not confuse with token ring !

46
MAC Profibus (2): Token passing
active stations
(potential masters)
AS AS AS
passive stations
(slaves)
12 25 31
logical ring
current bus
master
Active Stations (AS) can become master if they own the token,
for a limited duration (one frame only).
After that time, the master passes the token to a station with a higher address
or, if it has the highest address, to the station with the lowest address.
A station must send at least one frame (data or token) when it gets its turn.
station address

47
MAC Profibus (3): Token passing algorithm
Each station holds a List of Active Stations (LAS)
Previous
Station
Next
Station
This
Station
PS TS NS AS
GAP
AS AS
When the current master has no more data to send, or when its turn expires, it
sends a token frame to the Next Station (NS) in the ring.
NS acknowledges reception of the token.
If the master does not receive an acknowledgement for two consecutive trials,
the master removes the station from the LAS and declares the overnext
active station (OS) as NS.
This station accepts the token only if it receives it twice.
The master tests at regular intervals with a "Request FDL-Status" for the
presence of further stations in the gap between itself and NS.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 .. 30 32
Active
Station
Overnext
Station

48
MAC Profibus (4): Token initialization
A starting station listens to the bus before trying to send.
If it senses traffic, a station records the token frames and builds a list of active stations (LAS).
In particular, it observes if a station with the same name as itself already exists.
If a station does not register any traffic during a certain time, it sends a token frame to itself.
It sends the first token frame to itself, to let other starting stations register it.
Only when it detects no other station does a station begin with a systematic
poll of all addresses, to build the LAS.
When a master checks the gap, the station will receive a token offer.

49
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

50
Comparison of Medium Access Control Methods
optimistic
stochastic
pessimistic
deterministic
central master
token passing
collision with winner
carrier sense
collision detection
p-persistent collision

51
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

52
Logical Link Control Sublayer
Two Link services:
- unacknowledged connectionless service and
- connection oriented services
Network
Medium Access
Control
(MAC)
Logical Link
Control
(LLC)
Network
Frame
Physical
Medium
Physical
Signaling
Medium Access
Control
(MAC)
Logical Link
Control
(LLC)
Frame
Physical
Medium
Physical
Signaling

53
Connection-Oriented and Connection-Less communication
Connectionless mode
(Datagram ≈ letter)
Each packet (Datagram) contains all
information needed to forward it to its final
destination over the network, including the
path back to the sender.
The network assumes no responsibility for
the ordering of packets and does not try to
recover damaged datagrams.
The burden of flow control and error recovery
is shifted to the application.
Connection-Oriented mode
(Virtual Circuit ≈ telephone)
A connection is first established between sender
and receiver.
Information packets are identified by their
connection identifier and by their sequence
number within that connection.
The network cares for opening and closing
connections and ensures that packets are received
in same order as they are sent, recovering lost
packets and controlling the flow.
Applications see the network as a reliable pipe.
Connection is closed after use (and reused)
Semantic of CO-transmission
Open_Channel(Node, Task, Channel_Nr);
Send_Message (Channel_Nr, Msg1);
Send_Message (Channel_Nr, Msg2);
Close(Channel_Nr);
Msg1 will be received before Msg2, sequence is maintained.
Semantic of CL-transmission
Send_Packet (source, destination, Packet1);
these considerations apply to all levels of the OSI model
Send_Packet (source, destination, Packet2);

54
Connection-Oriented Link Service
Connection establishment and disconnection
Send and receive frames
Flow Control (Buffer control)
Retry in case of error
REQUEST
INDICATION
CONFIRM
service
user
service
provider
service
user
RESPONSE
Task: Flow Control and Error Recovery

55
Flow Control
"Adapt the speed of the sender to that of the receiver"
( = synchronization at the link layer)
Methods
Use Acknowledgements: do not send until an acknowledgement is received
(acknowledgements have two purposes: error recovery and flow control !)
Credit: the receiver indicates how many frames it can accept
(sliding Window protocol). Improvement: variable window size.
Explicit braking (CTRL-S/ CRTL-Q)
•
•
•
Upper Window
Edge
Lower Window
Edge
sent but not yet
acknowledged
sent and
acknowledged
can be sent may not
be sent
12
6 7 8 9 11
10
packets
time

56
Simple transfer with window size = 1
LLC
DATA (0)
ACK (1)
DATA (1)
DATA ( last)
ACK (last)
ACK (2)
LLC Consumer
alive
time-out
connect
timer
ack
timer
late
acks
Connect Request
Connect Confirm
i
Producer
nm_message_ind
nm_connect.ind
tm_message.req
nm_message.cnf
Connection
Transfer
Disconnection
Bus
nm_connect.cnf
Every packet takes at least two propagation times
time

57
Error Recovery
General rule: Erroneous information frames are repeated
(error correcting codes belong to physical layer)
1) In cyclic transmission, information is not explicitly repeated in case of loss,
the receiver will receive a fresh information in the next cycle.
A freshness control supervises the age of the data in case communication ceases.
The sender of information frames expects acknowledgement of the
receiver, indicating which information it received.
To distinguish repetitions from new information, each packet receives
a sequence number (in the minimum odd/even).
The sender repeats the missing information a number of times, until it
receives an acknowledgement or until a time-out expires.
3) In broadcast transmission, it is relatively difficult to gather acknowledgements
from all possible receivers, so in general unacknowledged broadcast is used.
The receiver is expected to protest if it does not receive the information.
2) In event-driven transmission, no information may be lost, a repetition is explicit:
a)
c)
b)
The receiver acknowledges repetitions even if it already received the
information correctly.
d)

58
Link Layer Outline
Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Error recovery
Flow control
HDLC
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison

59
Example: HDLC
HDLC (High-level Data Link Control is derived from IBM's SDLC
(Synchronous Data Link Control)
These protocols were developed for connection of point-of-sale terminals to a
one mainframe computer.
HDLC is the most frequently used link layer protocol.
It is the base for the CCITT-standard X25 (Telenet, Datex-P, Telepac) and used in Bitnet,
Appletalk, etc...
The HDLC protocol is implemented in the hardware of numerous
microcontrollers (e.g. Zilog 80C30, Intel, Siemens 82525,... and in some
microprocessors (e.g. 68360).
HDLC is the base for the Local Area Network protocol IEEE 802.2

60
HDLC Control Field (ISO 4335)
Control Field Bits
Control Field Format for:
1 2 3 4 5 6 7 8
0 N(S) P/F N(R)
Information Transfer
Command/Response
(I-Format PDU)
1 0 S P/F N(R)
1 1 M P/F M
Supervisory
Commands/Responses
(S-Format PDUs)
Unnumbered
Commands/Response
(U-Format PDUs)
16
8
FCS
adr control
8
8
01111110 01111110
flag flag
8
any data
physical address of Secondary
(for command and response)
N(S) = Sequence number of
sender
N(R) = Sequence number of
receiver
S = Supervisory
P/F = Poll/Final (Please respond/Final in sequence)

61
HDLC Connection Types
The sender includes the sequence number in each packet.
The receiver indicates which packet it expects next, either through a special frame
( Receiver Ready N(R) ) or within its information frames (I-Frame, N(R))
At the same time, this sequence number acknowledges all previously received
frames with number N(R) -1.
Traffic is divided intopackets (= information frame)each receiving a sequence
number (Modulo 8).
LAP (link access procedure): assymetric Primary/Secondary;
NRM (normal response mode): only one station as primary;
ARM (asynchronous response Mode): spontaneous transmission of secondary;
LAPB (LAP-balanced): every station can become primary and start transmitting
(if medium access allows).
HDLC provides different connection types:

62
HDLC Exchange (NMR in ISO 4335)
set normal response
mode, poll
UA,F
SNRM, P I0,0 I1,0 I2,0P
RR3,F
I3,0
Primary
(Commander)
Secondary
(Responder)
time
information packets
receiver
ready, expects 3
please confirm
accept
connection,
final
Send Sequence
accept
connection, final
set normal response
mode, poll
UA;F
SNRM; P
I0,0 I1,0 I2,0;F
RR0;P RR3;P
time
several
information packets
receiver
ready, expects 0
last frame
I3,0
receiver
ready, expects 3
Receive Sequence
Primary
(Commander)
Secondary
(Responder)
The data transmission takes place under control of the Primary.
Therefore, both "Send Frame" and "Receive Frame" are supported

63
Clocks
In a fieldbus, devices must be synchronized to a common clock to time-stamp
their transmissions.

64
Time Distribution in a single master system
At fixed intervals, the Master broadcasts the exact time as a periodic variable.
When receiving this variable, the bus controllers generate a pulse which can
resynchronize a slave clock or generate an interrupt request.
application
processor 1
bus master
PORTS
BUS
bus
controller
pulse
master
clock
time variable
int
req
pulse
int
req
bus
controller
PORTS
application
processor 2
pulse
int
req
bus
controller
PORTS
application
processor 3
slave
clock

65
Clock compensation for transmission delays
slave clocks
master
1
other master
synchronizer
slave
clock
MVB 1 other MVB
master clock
device
with clock
The clock does not need to be generated by the Master, but the master must poll the clock
The clock can synchronize sampling within a few µs across several bus segments.

66
IEEE 1588 PTP Clock Synchronization
IEEE 1588 defines the Precision Time Protocol, a clock synchronization that assumes that
two consecutive frames have the same delay, but the moment of sending suffers jitter.
The clock device (possibly coupled to a radio signal) sends the first frame with an coarse
time stamp, but registers in hardware the exact moment when the frame was sent.
It then sends a second frame with the exact time at which the first frame was sent.
Bridges and switches are responsible to compensate for their internal delays and send a
corrected time frame.

67
High precision clock synchronization
In some application, even the PTP protocol is insufficient.
In this case, either the clock is distributed by a separate, dedicated medium
(as in railways signalling and electrical substations.
Alternatively, all devices receive a radio signal from GPS to recalibrate their internal clocks.

68
Assessment
What is the purpose of the link layer ?
Which is the role of the three sublayers in the link layer ?
What is the Hamming Distance ?
What is the Residual Error Rate ?
What is the code efficiency ?
Where are error-correcting codes used ?
What is the formal of an HDLC frame ?
What is the purpose of medium access control ?
Which medium access does not require an arbitration ?
Which kinds of arbitration exist ?
How does the CAN arbitration works and what is its assumption on the medium ?
How does the Ethernet arbitration works ?
What is the influence of collision detection in a LON arbitration ?
Which medium access are deterministic ?
What is the difference between connection oriented and connectionless transmission ?
How are error corrected by the logical link control in cyclic transmission ?
How are error corrected by the logical link control in event-driven transmission ?
How does a sliding window protocol works ?
How does a transmission in HDLC work ?
How are clocks synchronized ?

Prof. Dr. Hubert Kirrmann
ABB Ltd, Baden, Switzerland
3.3.4
OSI Upper Layers - Presentation Layer, ASN.1 and data types
Niveaux supérieurs OSI - couche de présentation, ASN.1 et types de données
Obere OSI-Schichten – Darstellungsschicht, ASN.1 und Datentypen
PersonnelRecord ::= [APPLICATION 0] IMPLICIT SET {
name Name,
title [0] VisibleString,
number EmployeeNumber,
dateOfHire [1] Date,
nameOfSpouse [2] Name,
children [3] IMPLICIT
SEQUENCE OF ChildInformation
DEFAULT {}
2005, May, HK

3.0.3 Presentation Layer
2005 May, HK
2
Presentation layer in the OSI-Model (ISO/IEC standard 7498)
Transport
protocols
Application
protocols
(Mail, File Transfer,...)
Definition and conversion of the
data formats (e.g. ASN 1)
(e.g. ISO 8326)
(e.g. TP4, TCP)
(e.g. IP, X25)
medium access (e.g. 802.2, HDLC)
mechanical coupling (e.g. Ethernet)
Physical
Link
Network
Transport
Session
Presentation
6
5
4
3
2
1
Application
7

2005 May, HK
3
Presentation Layer
The presentation layer is responsible that all communication
partners agree on the format of the data

2005 May, HK
4
Transfer Syntax: describe what is in a frame ?
Role: how to define formally the format and meaning of the transmitted data
>48
ISO 8473
connectionless network control
5
ISO 8073
class 4 transport control
L_destination SAP
L_source SAP
L_PDU
L_PDU = UI, XID, TEST
LI
TPDU
Protocol Identifier
Header Length
Version/Protocol ID (01)
Lifetime
DT/ER Type
SP MS ER
PDU Segment Length
Address
Part
(CDT)
N(S)
ET
MAC_header LNK_hdr NET_header TRP_header
Destination
Reference
FIXED
PART
13 3
DATA
AFI = 49
PSI
Physical Address
(6 octets)
LSAP = FE
NSAP = 00
IDP
(initial
domain
part)
DSP
(domain
specific
part)
DATA (DT) TPDU
(normal format)
LSAP = DSAP
FE = network layer
18 = Mini-MAP Object
Dictionary Client
19 = Network Management
00 = own link layer
(81)
IEEE 802.4
token bus
ISO 8802
logical link control
address length
IDI, Area ID
(7 octets)
MA. frame control
address
(6 octets)
MA. source
(6 octets)
MA. destination
address
Checksum
Destination Address
(18 octets)
Source Address
(18 octets)
Segmentation
(0 or 6 octets)
Options
(priority = 3 octets)

2005 May, HK
5
ASN.1 justification
Why do we need ASN.1 ( or a similar notation)

2005 May, HK
6
Semantic Levels of Data Representation
Transmission
Specification
Language and
Encoding Rules
Traffic Memory format
Application Storage format
(Assembly language)
Bus Transmission Format
Application Memory
Traffic Memory
time
Storage format
SEQUENCE {
item1 INTEGER16,
item2 INTEGER4,
count UNSIGNED8;
}
D15 D00
Bus Controller
Application Language format
Semantic format
Application
Processor
struct {
unsigned count;
int item2:8;
int item1:4;
int dummy:4;
}
154,5 Vrms
Parallel Bus Format
(8 bit, 16 bit, 32 bits)
(16-bit oriented)

2005 May, HK
7
Bit Transmission Order
time
0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0
UART, HDLC, Ethernet FDDI, FIP, CAN, MVB
LSB MSB MSB LSB
first
Most data links are byte-oriented (transmit 8 bit as an indivisible unit)
The order of bit transmission within a byte (octet) is dependent on the link.
It does not matter as long as all bus participants use the same scheme
There is no relation between the bit and the byte ordering scheme
Who says which bit is really transmitted first ?
(FDDI: multi-bit symbol transmission, multi-bit
transmission with interleaving)
first
Legacy of the old telex
octet in octet out
Convention: only consider octet streams

2005 May, HK
8
Integer representation in memory: Big-Endian vs Little Endian
x0000
address
0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 1
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1
0 0 0 0 0 0 0 0
7 6 5 4 3 2 1 0
x0001
x0002
x0003
x0004
Intel Motorola
= 2 = 2
INTEGER8
INTEGER16 = 1 = 256
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
x0005
x0006
INTEGER32 = 1 = 16777216
IBM, TCP/IP,
Unix, RISC
DEC
MSB LSB
B7B6B5B4B3B2B1B0
0 1 2 3 4 5 6 7
1 2 3 4 5 6 7 8
Processor
TCP/IP
(bit position)
(bit offset)
(bit name)
LE BE
Memory contents
Profibus
three different naming schemes !

2005 May, HK
9
How are sequence of octets transmitted ?
0 1 2 3
INTEGER32
Most Significant First ?
Transmission Order on the bus ?
how to name bits ?
time
The standard in network protocols is always:
Most Significant Octet first (Big-Endian)

2005 May, HK
10
So, how to respect network byte ordering ?
Since network protocols require a “most significant octet first” transmission
(NBO =Network Byte Ordering), all little-endian processors must convert their data before
putting them into the transmission buffer, pipe or socket, by calling the Unix functions:
Function Name Description
htons Host order to network order, short (2 bytes, 16 bits)
htonl Host order to network order, long (4 bytes, 32 bits)
ntohs Network order to host order, short (2 bytes, 16 bits)
ntohl Network order to host order, long (4 bytes, 32 bits)
All data "seen" by the sockets layer and passed on to the network must be in network order
struct sockaddr_in s;
/* WRONG */
s.sin_port = 23;
/* RIGHT */
s.sin_port = htons(23);
- Extremely error prone

2005 May, HK
11
Inconsistency: Token-bus frame (IEEE 802.4)
N N 0 N N 0 0 0
F F M M M P P P
I/G L/U lsb
msb
Destination Address
first MAC
symbol
preamble
remaining 44 bits of address
I/G L/U lsb
msb
remaining 44 bits of address
lsb
msb
Frame Check Sequence
N N 1 N N 1 I E
first octet
lsb msb
Source Address
last data octet
lsb msb
Addresses are transmitted least significant bit first
Data are transmitted least significant bit first within
an octet, but most significant octet first within a
word (imposed by higher layers)
Checksum is transmitted most significant bit first
read this picture from left to right,
then top to bottom

2005 May, HK
12
How to specify transmitted data
stationID
functionID
snu gni nodeID or groupID
protocolID
command
parameter
offset 0
CommandPDU :== SEQUENCE
{
BOOLEAN1 snu -- system
BOOLEAN1 gni -- group
CHOICE (gni)
{
Group : ENUM6 groupID;
Individual: ENUM6 nodeID;
}
ENUM8 stationID;
...
intuitive formal
time
offset
0 1 2 3 4 5 6 7
first: agree on a bit and byte ordering
scheme: {0,0} scheme recommended.
second: describe the data stream
formally (machine-readable)
0
1
2
3
4
i
i+1

2005 May, HK
13
Can a “C”-declaration serve as an encoding syntax ?
how is size given?
to which structure does it point?
typedef struct {
char location[ LOCATION_LEN ];
unsigned long object_id;
alarm_type_t alarm_type;
priority_level_t priority_level;
unsigned long index_to_SNVT;
unsigned value[ 4 ];
unsigned long year;
unsigned short month;
unsigned short day;
unsigned short hour;
unsigned short minute;
unsigned short second;
unsigned long millisecond;
unsigned alarm_limit[ 4 ];
} SNVT_alarm;
allowed values in enum ?
is "short" a byte ?
is “unsigned” 16 bits or 32 bits ?
Such a machine-dependent syntax is only valid if all applications use the same syntax on
the same machine with the same compiler, it is not suited to describe the bus traffic.

2005 May, HK
14
Abstract Syntax Notation Number 1 (ASN.1)
• The IEC/ISO define a standard notation in IEC 8824 (ASN.1), allowing to define
simple types (primitive) and constructed types.
• Data structures can take forms not usually found in programming languages.
• Each data structure is identified during transmission by a tag.
• In principle, ASN.1 only defines data structures to be transmitted, but not how
they are encoded for transmission.
• One possible coding of the primitive and constructed data types is defined in
ISO/IEC 8825 as "Basic Encoding Rules“ (BER) defined in ISO 8825.
• More efficient encodings (PER,…) also exist.
• ASN.1 can be used for defining memory contents, file contents or communication
data, and in general any exchanged information.
• ASN.1 has the same role as XML, but it is far more efficient.

2005 May, HK
15
ASN.1 Syntax Example
Name:
Title:
Employee Number:
Date of Hire:
Name of Spouse:
Number of
Children:
Child Information
Name:
Date of Birth
Child Information
Name:
Date of Birth
John P Smith
Director
51
17 September 1971
Mary T Smith
2
Ralph T Smith
11 November 1957
Susan B Jones
17 July 1959
PersonRecord ::= [APPLICATION 0] SEQUENCE {
name Name
number EmployeeNumber,
dateOfHire [1] Date,
nameOfSpouse [2] Name,
children [3] SEQUENCE OF
Childinformation DEFAULT {} }
Name ::= [APPLICATION 1] SEQUENCE {
givenName VisibleString,
initial VisibleString,
familyName VisibleString}
EmployeeNumber ::= [APPLICATION 2] INTEGER
ChildInformation ::= SEQUENCE {
name Name,
dateOfBirth [0] Date}
Date ::= [APPLICATION 3] VisibleString -- YYYYMMDD
Informal ASN.1

2005 May, HK
16
Abstract syntax and transfer syntax
An abstract syntax describes the elements of information without considering their
encoding (i.e. how they are represented in memory or on a bus)
E.g. A transfer syntax defines the name of the structures and elements, their value range
[e.g. 0..15], ….
ASN.1 is defined in the standard ISO 8824-1.
A transfer syntax describes how the structures and elements are effectively
encoded for storing and transmission, so that the receiver can fully decode the
transmitted
information. At transmission time, only the transfer syntax is visible, it cannot be
interpreted
without knowing the abstract syntax.
E.g. A transfer syntax defines that an array of 33 Unicode characters is transmitted.
The receiver knows from the abstract syntax that this is a person’s name.
The basic encoding rules for ASN.1 are defined in the standard ISO 8825-1.

2005 May, HK
17
ASN.1 encoding: TLV
ASN.1 does not say how data are stored nor transmitted, but it assumes that the
transmission format consists for each item of: a tag, a length and a value (TLV)
(recursive)
length value
tag
length
tag value
tag length
tag
length
The value may be itself a structured object:
value
the tag specifies the data type of the value that follows, implicitly or explicitly.

2005 May, HK
18
ASN.1 Data Types
BOOLEAN
INTEGER
BITSTRING
OCTETSTRING
NULL
OBJECT_ID
OBJECT_DESC
EXTERNAL
REAL
ENUMERATED
ANY
SEQUENCE ordered sequence of types (record)
SEQUENCE OF ordered sequence of same type (array)
CHOICE one of an unordered, fixed sequence of
different types.
SET unused
SET OF unused
UNIVERSAL
APPLICATION
CONTEXT_SPECIFIC
PRIVATE
Constructed Types
Basic Types
Open arrays (dynamic size) and pointers are not allowed
four Tag Types

2005 May, HK
19
ASN.1: The Type SEQUENCE
An ASN.1 SEQUENCE is similar to a “C”-struct:
Example: PersonRecord ::= SEQUENCE {
person VisibleString
chief VisibleString
title VisibleString,
number INTEGER,
dateOfHire UniversalTime }
This notation assumes that all elements in the sequence are present and are transmitted
in the specified order.
person chief title number dateOfHire
INTEGER UniversalTime
VisibleString
PersonRecord

2005 May, HK
20
Tagging
PersonRecord ::= [APPLICATION 1] SEQUENCE {
person [1] VisibleString,
chief [2] VisibleString,
number [4] INTEGER,
dateOfHire [5] UniversalTime }
When elements of a sequence may be missing, it is necessary to tag the items,
i.e. identify the items by an integer.
Of course, if all types would be different, it would be sufficient to specify the type,
but this practice (called EXPLICIT tagging) should not be followed.
Structured types need in any case a tag, otherwise they cannot be distinguished
from another structured type (note how “PersonRecord” is identified)
person title
number dateOfHire
A4 A5
A1
A1 A3
tags

2005 May, HK
21
ASN.1: The CHOICE type
A choice selects exactly one alternative of several.
There is normally a distinct tag for each choice, but the type can
also be used as tag, as long as all types are distinct:
Quantity ::= CHOICE {
[0] units INTEGER16,
[1] millimetres INTEGER32,
[2] kilograms FLOAT
}
2 4 1.2 E 03
quantity in millimeters
kilograms, IEEE format
1 2 1234
quantity in millimeters
0 2 1234 quantity in units

2005 May, HK
22
Difference between ASN.1 and “C”* or XML/XDR
ASN.1 is a format for data exchange, “C” is a compiler-dependent format for storage in RAM.
The basic data types are defined differently.
ASN.1 types may have a variable size (INTEGER may be 8, 16, 32, 64 bits).
ASN.1 SEQUENCE differs from a “C” struct since not all elements must be transmitted, nor is
their order to be maintained (if tagging is used).
ASN.1 CHOICE differs from a “C” union since the length depends on the chosen item,
and the contents differ.
ASN.1 has the same role as XML. XML is however both an abstract and a transfer syntax, it is
sent in clear text.
Unix systems use XDR (Sun's external data representation) that is 32-bit oriented, not efficient
for small data items, also a mixture of abstract and transfer syntax.

2005 May, HK
23
Explicit and implicit tagging
ASN.1 can assign automatically a tag to elements of a sequence.
This practice is dangerous, because another encoding (PER) could use other
tag numbers.
It is preferable to make tags explicit everywhere that is needed.

2005 May, HK
24
Transfer Syntax: Basic Encoding Rules (BER)
0 2 02
length = 2 octets
12 34
type = UNIVERSAL Integer
value = 1234
Example:
BER supports all ASN.1 structures.
Exception: if size is 0, there is no value field (== NULL)
A companion to ASN1, BER (ISO 8825-1) defines encoding rules for ASN.1 data types
BER tags all data, either implicitly or explicitly
Type, Tag and size are transmitted before every value or structured data
(tag included in basic types)

2005 May, HK
25
BER – Tag/Type field
(for universal class only):
00 = null (size = 0, no value)
01 = Boolean type
02 = Integer type
03 = Bitstring type
04 = Octetstring type
05 = Null type
06 = Object Identifier type
07 = Object Descriptor type
16 = Sequence and Sequence_Of types
17 = Set and Set_Of types
18-22 = Character strings (numeric, printable, …)
23-24 = Time types
25 = Graphic string
26 = VisibleString (ISO646)
> 28 = reserve and escape: use a second octet.
Tag
Primitive
{0}
or
Constructed
{1}
00 = UNIVERSAL
01 = APPLICATION
10 = CONTEXT_SPECIFIC
11 = PRIVATE
Class
Example: 00000010 = UNIVERSAL INTEGER
10100001 = [CONTEXT_SPECIFIC 1] SEQUENCE
P: after the size comes a value -
C: after the size comes a tag

2005 May, HK
26
Example ASN.1 and BER
1
high_prio
2
1
command
2
reference (MSB)
reference (LSB)
0
1
4
caller (MSB)
caller (LSB)
3
2
2
P
P
P
P
P
CS
CS
CS
CS
UN
high_prio
command
reference
caller
7
C
AP
16
8 1
(useful information)
CallerRef :== [APPLICATION 7] SEQUENCE {
priority [2] INTEGER,
command [0] INTEGER,
reference [1] INTEGER,
caller INTEGER}

2005 May, HK
27
Examples
Tag/Type:
• 1000’0000 Context Specific, not constructed, implicit, tag = [0]
• 0110’0001 Application Specific, constructed, implicit, tag = [1]
AB 1010’1101 Context Specific, constructed, implicit, tag = [11]
01 0000’0001 Basic Type, not constructed, boolean
decoding the first digit:
0,1: Universal, not constructed
2,3: Universal, constructed
4,5: Application Specific, not constructed [APPLICATION 1]
6,7: Application Specific, constructed
8,9: Context Specific, not constructed [6]
A,B: Context Specific, constructed
constructed means: the next octet is not
a value, but a type/tag !

2005 May, HK
28
Examples
A0 0E 02 01 0A A1 09 A0 03 80 01 00 A1 02 80 00
A1 67 02 01 0A A1 62 A0 5D 1A 0B 54 65 6D 70 65
72 61 74 75 72 65 1A 0C 54 65 6D 70 65 72 61 74
75 72 65 31 1A 07 61 72 72 61 79 5F 35 1A 04 62
6F 6F 6C 1A 0F 66 65 65 64 65 72 31 5F 33 5F 70
68 61 73 65 1A 05 66 6C 6F 61 74 1A 0F 68 65 72
62 73 5F 74 65 73 74 5F 74 79 70 65 1A 08 75 6E
73 69 67 6E 65 64 81 01 00
--
A0 18 02 01 0B A6 13 A0 11 80 0F 66 65 65 64 65
72 31 5F 33 5F 70 68 61 73 65
A1 34 02 01 0B A6 2F 80 01 00 A1 16 81 14 66 65
65 64 65 72 31 5F 33 5F 70 68 61 73 65 24 41 64
64 72 A2 12 A2 10 A1 0E 30 05 A1 03 85 01 10 30
05 A1 03 85 01 10
--
A0 1E 02 01 0C A4 19 A1 17
green: tag / type
red: size
black: value
null

2005 May, HK
29
Beyond ASN.1 and BER
ASN.1 / BER could not impose themselves in field busses because
of the high overhead involved (32 bits for a single boolean !)
ISO / UIT developed more efficient encodings, such as
ISO/IEC 8825-2: Packed Encoding Rules (PER), that
exists in two versions: aligned (on an 8-bit boundary) or not aligned (bit stream)
In low speed busses such as fieldbus, this is still too much overhead.
IEC 61158-6 (Fieldbus) offers 3 encodings:
Traditional Encoding Rules (Profibus)
Compact Encoding Rules (for FAL)
Buffer Encoding Rules (FIP)
100 MBit/s Ethernet has sufficient bandwidth, but the burden is shifted to the processors
(Data compression gives variables length messages, costs a lot in compression
& decompression)

2005 May, HK
30
ROSIN Compact – Retrofit Encoding
The railways operators needed to define formally the exchange rules for
data of already existing devices, of different manufacturers and vintage.
Therefore, a notation was developed (ROSIN notation) to describe any data transfer,
on a bit rather than a byte-orientation. It also allows to cope with alignment
(data should be transmitted at an offset that is a multiple of their size to reduce
processor load)
Indeed, since these devices already communicate using a proprietary protocol,
transmission must already be unambiguous.
Each data type is specified, e.g. Integer32 differs from Integer32_LE (Little Endian)
The ROSIN notation uses the ASN.1 meta-syntax, but it does not imply a TLV scheme.
- the length can be deduced from the type or position,
- typing information is inserted explicitly when:
• a choice exists among several alternative types (e.g. depends on success/failure)
• types have a (large) variable size (e.g. text strings, files)
• sequences have optional fields and out-of-sequence fields (occurs too often).

2005 May, HK
31
ROSIN - Retrofit encoding rules example
0 7
snu gni node_id
parameter1
parameter2
parameter3 par4
parameter5
Type_InfoMessage ::= RECORD {
parameter1 INTEGER8 -- octet.
parameter2 INTEGER16, -- 16-bit word, MSB first
parameter3 UNSIGNED6, -- 6- bit value
par4 ANTIVALENT2, -- 2 bits for par4, e.g. check variable.
parameter5 Parameter5, -- parameter5 has a structured type
parameter6 STRING32, -- an array of up to 32 8-bit characters.
-- trailed if shorter with “0” characters.
snu ENUM1 {
USER (0), -- 0 = user
SYSTEM (1) -- 1 = system
},
gni ENUM1 { -- could also be expressed as BOOLEAN1
INDIV (0), -- individual function addressing
GROUP (1), -- group addressing.
},
node_id UNSIGNED6, -- 6-bit unsigned integer
sta_or_func ONE_OF [snu] { -- meaning depends on ‘snu’
function [USER] UNSIGNED8, -- if snu = user, function identifier
station [SYSTEM] UNSIGNED8 -- if snu = system, station identifier
},
next_station_id UNSIGNED8, -- next station or ‘FF’H if unknown
tv BOOLEAN1, -- TRUE if 1
res1 BOOLEAN1 (=0), -- 0 (place holder)
topo_counter UNSIGNED6, -- 6-bit unsigned integer
tnm_code ENUM8 { -- has only two defined values
FIRSTCASE (‘1E’H) -- one of two defined values
SECONDCASE (‘84’H) -- the other defined value
},
action_code Action_Code, -- this type is used several times
}
parameter6
CHARACTER8
sta_or_func
next_station_id
tv d1 top_counter
tnm_code
action_code
0
1
2
3
4
5
6
7
39
40
41
42
43
44

2005 May, HK
32
Comparing Coding Efficiency
Boolean
BER
24
XDR
16
A-PER
5
U-PER
5
FER BuER
16 8 1
ROSIN
Integer8 24 16 16 12 16 16 8
Unsigned8 24 16 11 11 16 16 8
Integer16 32 24 24 20 24 24 16
Unsigned 16 32 24 24 19 24 24 16
Integer32 48 40 24..48 36 40 40 32
Unsigned32 48 40 24..48 35 40 40 32
String [32] 272 272 272 272 272 272 256
(FIP)
(Profibus)
(SUN) (UIT)
(ISO)
encoding/decoding highly packed data may cost more than is won by shorter transmission

2005 May, HK
33
Engineering Units
Many process variables represent analog, physical values of the plant.
Data presentation (e.g. integer) is insufficient to express the meaning of the variable.
Therefore, it is necessary to allocate to each variable a data type in engineering units.
"A unit of measure for use by operating/maintenance personnel usually
provided by scaling the input quantity for display (meter, stripchart or CRT)"
IEEE
Scaling (determined the possible range of a variable) is necessary for analog displays.
It requires the definition of the possible range of values that the variable may take.

2005 May, HK
34
SI Units
-2
3
3
time s
current A
angle rad
force N
torque Nm
power W
frequency s-1
angular velocity rad/s
mass kg
pressure Pa
flow m /s
mass flow kg/s
tension V
reactive power var
impedance W
temperature K
volume m
energy J
position, distance m
angular acceleration rad s
All physical variables should be restricted to SI Units (NIST 330-1991, IEEE 268A-1974)
or refered directly to them
For instance:
•
variable unit
variable unit
•
angles shall be represented in radian rather than degree or grad.
speeds shall be expressed in meter/second, not in km/h or in miles/hour

2005 May, HK
35
Why floating point ?
Floating point format is the only safe representation of a physical variable.
Exception:
In special applications (e.g. GPS data), an ASCII representation may be more
appropriate, albeit not efficient. In this case, data can also be processed as BCD
(as in pocket calculators)
Floating point format (IEEE Std 254)
- require twice the place (32 bits vs 16 bits),
- adds about 50% to traffic (analog values are only 10% of total),
- cost more to process (floating point unit)
but
- removes all ambiguities, rounding errors, overflow and underflow.
Therefore, devices shall indicate their exported and imported variables as REAL32.
Internally, devices can use other formats.
Variables, whose precision do not depend on their absolute value:
time, elapsed distance (odometer), energy, money, countable objects

2005 May, HK
36
Fractionals
0
e.g. offset = 0,0 m/s, span = 100,0 m/s, base unit = UNSIGNED16
means: 0 = 0,0 m/s, 65536 == 100,0 m/s
e.g. offset = - 32,768 V, span = 65,536 V, base unit = INTEGER16
means: 0 = -32,768 m/s, 65536 == 32,767 V
A device can indicate the format of a fractional analog variable by
specifying (as a REAL32) the span and the offset to be applied to the base unit:
offset value
physical variable
65535
-32768 +32767
span value
UNSIGNED16
INTEGER16
0

2005 May, HK
37
Scaled Variables
Process Variables are often transmitted and processed as fractionals:
Fractional format expresses analog values as integer multiples of the resolution
e.g.:
Some standards provides a bipolar or unipolar analog data format:
e.g. :
fractionals require producer and consumer to agree on range and resolution:
e.g. resolution 0.5 V, range 6553.5 V
distance = 0..65535 x 0.1 m (resolution) --> 0 .. 65535 (UNSIGNED16)
distance = -32768..+32767 x 0.1 m --> -32768 .. 32768
speed = 0.. 6553,5 m/s, resolution = 0.1 m/s --> 0 .. 65535 (UNSIGNED16)
speed = 0.. 6.5535 m/s, resolution = 0,0001 m/s --> 0 .. 65535 (UNSIGNED16)
0..200% of physical variable == 0..65536 (UNSIGNED16)
-200%..+200%-e of 10 kV == -32768 .. 32768
fractionals are easy to process but error prone (overflow, underflow)
A conversion from fractional to floating point did cost over 5 Mia € (Ariane 501 accident)
(INTEGER16)
(INTEGER16)
The resolution is often a decimal fraction of a unit !

2005 May, HK
38
Constructed Application Data Types
Circuit Breaker command:
DoubleCommand
not permitted
OFF
ON
not permitted
Persistent
Regulating Command
not permitted
Lower
Higher
not permitted
RegulatingStep
Command
not permitted
Next Step Lower
Next Step Higher
not permitted
Double-Point
Information
indeterminate
determined OFF
determined OFF
not permitted
Code
00
01
10
11
Time-Stamped Variable:
value
time
status

2005 May, HK
39
Application Data Types
Structured Text:
RTF - Microsoft Word Native Format
HTTP - Hypermedia data presentation
Some standards for video:
QuickTime -- an Apple Computer specification for video and audio.
Motion Picture Experts Group (MPEG) -- video compression and coding.
Some graphic image formats:
Graphics Interchange Format (GIF) -- compression and coding of graphic images.
Joint Photographic Experts Group (JPEG) -- compression and coding for graphic images.
Tagged Image File Format (TIFF) -- coding format for graphic images.
14

2005 May, HK
40
Presentation Layer in the application
Coding and conversion functions to application layer data cannot be done in the
presentation layer due to the lack of established rules
These functions ensure that information sent from the application layer of one system will
be readable by the application layer of another system.
Examples of presentation layer coding and conversion schemes in the application:
Common data representation formats -- The use of standard image, sound, and video
formats allow the interchange of application data between different types of computer
Conversion of character representation formats -- Conversion schemes are used to
exchange information with systems using different text and data representations (such as
ASCII and Unicode).
Common data compression schemes -- The use of standard data compression schemes
allows data that is compressed at the source device to be properly decompressed at the
destination (compression can take place at different levels)
Common data encryption schemes -- The use of standard data encryption schemes allows
data encrypted at the source device to be properly unencrypted at the destination.

2005 May, HK
41
Gateway
When devices share no common transport layer protocol, gateways act as protocol
converters and application layer protocols must ensure end-to-end control.
Protocol Conversion is costly in development and real time, since protocols are in
general insufficiently specified, custom-designed, and modified without notice.
Protocol Conversion requires at least an semantical equivalent of the objects on both
sides of the gateway, so that one command can be converted into another - if possible.
bus type 2
bus type 1
network
transport
session
presentation
gateway
application
Real-Time
Protocols
link
physical
link
PD-marshalling
link
MSG
network
transport
session
presentation
PV
link
application
physical
PV
physical physical

2005 May, HK
42
To probe further
http://www.isi.salford.ac.uk//books/osi/all.html - An overview of OSI
http://www.oss.com/ - Vendor of ASN.1 tools
http://www-sop.inria.fr/rodeo/personnel/hoschka/347.txt - List of ASN.1 tools
http://lamspeople.epfl.ch/kirrmann/mms/OSI/osi_ASN1products.htm - ASN.1 products

2005 May, HK
43
Assessment
which are the “upper layers”?
which is the function of the network level ?
what is the difference between repeater, bridge, router and gateway ?
what is the difference between hierarchical and logical addressing ?
what is the role of the transport layer ?
when is it necessary to have a flow control at both the link and the transport layer ?
what is the role of the session layer in an industrial bus ?
which service of the session layer is often used in industrial networks ?
what is the role of the presentation layer ?
what is ASN.1 ?
what is the difference between an abstract syntax and a transfer syntax ?
why is ASN.1 not often used in industrial networks and what is used instead ?
what is the ROSIN notation for and how is a data structure represented ?
how should physical (analog) variables be represented ?

1 3.4 MVB case study
2003 July, HK
3.4 MVB: a fieldbus case study

2003 July, HK
MVB Outline
1. Applications in rail vehicles
2. Physical layer
1. Electrical RS 485
4. Frames and Telegrams
5. Medium Allocation
7. Fault-tolerance concept
8. Integrity Concept
2. Middle-Distance
3. Fibre Optics
9. Summary
3. Device Classes
6. Clock Synchronization

2003 July, HK
standard communication interface for all kind of on-board equipment
data rate
delay
medium
number of stations
> 600 vehicles in service in 1998
status
up to 4095 simple sensors/actuators
1'500'000 bits/second
0,001 second
twisted wire pair, optical fibres
up to 255 programmable stations
Multifunction Vehicle Bus in Locomotives
cockpit
power line
diagnosis
radio
Train Bus
motor control
power electronics
brakes track signals
Vehicle Bus

2003 July, HK
Multifunction Vehicle Bus in Coaches
covered distance: > 50 m for a 26 m long vehicle
< 200 m for a train set
diagnostics and passenger information require relatively long, but infrequent messages
brakes
air conditioning
doors
power
light
passenger
information
seat reservation
Vehicle Bus
Train Bus

2003 July, HK
MVB Physical Media
• OGF
• EMD
• ESD
Media are directly connected by repeaters (signal regenerators)
All media operate at the same speed of 1,5 Mbit/s.
(2000 m)
(200 m)
(20 m)
optical fibres
shielded, twisted wires with transformer coupling
wires or backplane with or without galvanic isolation
twisted wire segment sensors
optical links
rack
optical links
rack
star coupler
devices

2003 July, HK
MVB Covered Distance
The MVB can span several vehicles in a multiple unit train configuration:
The number of devices under this configuration amounts to 4095.
MVB can serve as a train bus in trains with fixed configuration, up to a distance of:
> 200 m (EMD medium or ESD with galvanic isolation) or
> 2000 m (OGF medium).
Train Bus
devices
node
devices with short distance bus
repeater
MVB

2003 July, HK
MVB Topography
all MVB media operate at same speed, segments are connected by repeaters.
Device
Device
Device Device
Terminator
Train Bus
OGL link
Repeater
Repeater
Repeater
Device Device Device Device
Bus
Administrator
EMD Segment
section
ESD Segment
ESD Segment
Node
Device Device
Device

2003 July, HK
MVB Outline
1. Applications in vehicles
2. Physical layer
1. ESD (Electrical, RS 485)
2. EMD (Transformer-coupled)
3. OGF (Optical Glass Fibres)
9. Summary
3. Device Classes

2003 July, HK
ESD (Electrical Short Distance) RS485
Interconnects devices over short distances (- 20m) without galvanic separation
Based on proven RS-485 technology (Profibus)
Main application: connect devices within the same cabinet.
terminator/
biasing
+ 5 V
GND
Ru
(390W)
Rm
(150 W)
Rd
(390 W)
terminator/
biasing
segment length
device 1 device N
device 2.. n-1
RxS
TxS
RxS
TxS
RxS
TxS
• • •
equipotential line
Data_N
Data_P
Bus_GND
Ru
(390W)
Rm
(150 W)
Rd
(390 W)
+ 5 V

2003 July, HK
ESD Device with Galvanic Isolation
1
power
protection
circuit
RS 485
transceiver
Data
GND
galvanic
barrier
RxS'
TxS'
TxF'
RxS
TxS
TxF
+5V
cable
female
male
shield connected to
connector casing
shield connected to
connector casing
opto-
couplers
DC/DC
converter
1
device casing
connected to
supply ground
protective
earth
0V
el
+5V
el
+Vcc

2003 July, HK
ESD Connector for Double-Line Attachment
9
8
7
6
cable
Line_A
Line_B
Line_A
Line_B
10
4 5
2
1 3 2 1
4
5 3
female
male
9 8 7 6
cable
Line_A
Line_B
Line_A
Line_B
reserved
(optional TxE)
B.Data_P
A.Data_P
A.Data_N
B.Data_N
B.Bus_5V
A.Bus_5V
A.Bus_GND
B.Bus_GND

2003 July, HK
EMD (Electrical Medium Distance) - Single Line Attachment
• Connects up to 32 devices over distances of 200 m.
• Transformer coupling to provide a low cost, high immunity galvanic isolation.
• Standard 120 Ohm cable, IEC 1158-2 line transceivers can be used.
• Main application: street-car and mass transit
• 2 x 9-pin Sub-D connector
transceiver
bus section 2
device
bus section 1
bus
controller
shield
transformer

2003 July, HK
EMD Device with Double Line Attachment
Carrying both redundant lines in the same cable eases installation
it does not cause unconsidered common mode failures in the locomotive environment
(most probable faults are driver damage and bad contact)
1
B1. Data_P
B1. Data_N B2. Data_P
B2. Data_N
transceiver A
B2
transceiver B
Connector_2
A.Data_P A.Data_N B.Data_P B.Data_N
Bus_Controller
device
Line_B
Line_A
A1. Data_P
A1. Data_N
1
B1
A1
Connector_1
Line_B
Line_A
A1. Data_P
A1. Data_N
1
A2

2003 July, HK
EMD Connectors for Double-Line Attachment
4
5
2
1
3
7
6
9
8
terminator connector
3
4
5
Connector_1 (female)
2
1
Line_B
7
6
9
8
Line_A
B1.Data_N
B1. Data_P
A1. Data_N
A1. Data_P
A.Term_P
B.Term_N
Line_B
Line_A
B.Term_P
A.Term_N
Zt.A
female
cable
3
4
5
Connector_1 (male)
2
1
Line_B
7
6
9
8
male
Line_A
B1.Data_N
B1. Data_P
A1. Data_N
A1. Data_P
shields contacts case
Zt.B

2003 July, HK
EMD Shield Grounding Concept
Shields are connected directly to the device case
Device cases should be connected to ground whenever feasible
device device
inter-section
connectors
terminator terminator
device
device ground
shield
possible shield
discontinuity
device ground
inter-device
impedance
inter-device
impedance

2003 July, HK
OGF (Optical Glass Fibre)
Covers up to 2000 m
Proven 240µm silica clad fibre
Main application: locomotive and critical EMC environment
wired-or electrical media
fibre pair
device device
device
Rack
opto-electrical
transceiver
Star Coupler
to other device
or star coupler
to other device
or star coupler
device
device device
ESD segment

2003 July, HK
OGF to ESD adapter
Double-line ESD devices can be connected to fibre-optical links by adapters
to star coupler B
from star coupler B
A.Data_P
to star coupler A
from star coupler A
A.5V
TxE
TxD
RxDA
1
RxDB
1
B.0V
B.5V
A.Data_P
5
3 3
fibre-optical transceivers
RS-485 transceiver
MVBC
A.0V

2003 July, HK
MVB Repeater: the Key Element
(redundant)
bus
administrator
The repeater:
• decodes and reshapes the signal (knowing its shape)
• recognizes the transmission direction and forward the frame
• detects and propagates collisions
A repeater is used at a transition from one medium to
another.
repeater
EMD segment
decoder
encoder
decoder
encoder
ESD segment
(RS 485) (transformer-coupled)
bus
administrator slave slave slave slave

2003 July, HK
MVB Repeater
duplicated segment
Line_A Line_B
direction
recogniser
decoder
repeater
decoder
encoder
decoder
decoder
encoder
Line_A
(single-thread
optical link)
Line_B
(unused for single-
thread)
recognize the transmission direction and forward the frame
decode and reshape the signal (using a priori knowledge about ist shape)
jabber-halt circuit to isolate faulty segments
detect and propagate collisions
increase the inter-frame spacing to avoid overlap
can be used with all three media
appends the end delimiter in the direction fibre to transformer, remove it the opposite way
handles redundancy (transition between single-thread and double-thread)

2003 July, HK
MVB Outline
2. Physical layer
2. Middle-Distance
3. Fibre Optics
9. Summary
3. Device Classes

2003 July, HK
MVB Class 1 Device
Class 1 or field devices are simple connections to sensors or actuators.
They do not require a micro-controller.
The Bus Controller manages both the input/output and the bus.
They do not participate in message data communication.
MVB
redundant
bus pairs
(ESD)
analog
or
binary
input/
output
board bus
(monomaster)
device
address
register
RS 485
drivers/
receivers
bus
controller
device
status
A B

2003 July, HK
MVB Class 2-3 Device
Class 2 and higher devices have a processor and may exchange messages.
Class 2 devices are configurable I/O devices (but not programmable)
The Bus Controller communicates with the Application Processor through a
shared memory, the traffic store, which holds typically 256 ports.
•
•
•
MVB
redundant
bus pairs
(ESD)
application
processor
shared
local RAM
private
RAM
local
input/
output
EPROM
RS 485
drivers/
receivers
Bus
Controller
traffic store
device
status
A B

2003 July, HK
MVB Class 4-5 Device
Class 4 devices present the functionality of a Programming and Test station
To this effect, they hold additional hardware to read the device status of the
other devices and to supervise the configuration.
They also have a large number of ports, so they can supervise the process
data transmission of any other device.
Class 5 devices are gateways with several link layers (one or more MVB, WTB).
Class 4 devices are capable of becoming Bus Administrators.
The device classes are distinguished by their hardware structure.

2003 July, HK
MVBC - bus controller ASIC
12 bit device address
CPU parallel bus
to traffic store
duplicated
electrical or optical
transmitters
duplicated electrical or
optical receivers
A19..1
D15..0
address
A
B
A
B
Manchester
and CRC
encoder
16x16
Tx buffer
16x16
Rx buffer Traffic Store
Control
& Arbiter
Main
Control
Unit
Class 1
logic
data
control
Clock,
Timers &
Sink Time
Supervision
• Bus administrator functions
• Bookkeeping of communication errors
• Hardware queueing for message data
• Supports 8 and 16-bit processors
• Supports big and lirttle endians
• 24 MHz clock rate
• HCMOS 0.8 µm technology
• 100 pin QFP
• Automatic frame generation and analysis
• Adjustable reply time-out
• Up to 4096 ports for process data
• 16KByte.. 1MByte traffic store
• Freshness supervision for process data
• In Class 1 mode: up to 16 ports
• Bit-wise forcing
• Time and synchronization port
DUAL
Manchester
and CRC
decoders
JTAG
interface

2003 July, HK
MVB Bus Interface
Application
processor
2 message ports
MVB
bus
controller
Traffic Store
0..4095
Logical Ports
(256 typical)
for Process
data
6 bus
management
ports
8 physical
ports
The interface between the bus and the application is a shared memory, the
Traffic Memory , where Process Data are directly accessible to the application.
messages packets
and
bus supervision
process data
base

2003 July, HK
MVB Outline
2. Physical layer
2. Middle-Distance
3. Fibre Optics
9. Summary
3. Device Classes

2003 July, HK
MVB Manchester Encoding
1 1 0 1 0 0 0 1 0 1 1 1 1 1 0 1
data
clock
frame
signal
9-bit Start Delimiter frame data 8-bit check
sequence
The Manchester-coded frame is preceded by a Start Delimiter containing
non-Manchester signals to provide transparent synchronization.
1 2 3 4 5 6 7 8
0
end
delimiter

2003 July, HK
MVB Frame Delimiters
0 8
1 2 3 4 5 6 7
active state
idle state
active state
idle state
0
Different delimiters identify master and slave frames:
This prevents mistaking the next master frame when a slave frame is lost.
Master Frame Delimiter
Slave Frame Delimiter
8
1 2 3 4 5 6 7
start bit
start bit

2003 July, HK
MVB Frames Formats
F address
9 bits 4 12 8
9 16 bits
slave frames sent in response to master frames
8
CS
9 32 bits 8
9 64 bits 8
master frames issued by the master
MSD
16
(33)
16
(33)
32
(49)
64
(81)
MSD = Master Start Delimiter (9 bits)
CS = Check Sequence (8 bits)
SSD = Slave Start Delimiter (9
bits)
useful (total) size in bits
F = F_code (4 bits)
data CS
SSD
data
SSD CS
data
SSD CS
9 64 bits 8
data
SSD CS
128
(153)
9 64 bits 8
data
SSD CS
256
(297)
64 bits
data
8
CS
data
8
CS
64 bits
data
8
CS
64 bits
data
8
CS
64 bits
The MVB distinguishes two kinds of frames:

2003 July, HK
MVB Distance Limits
master frame
master
time
distance
next master frame
t_sm
t_ms < 42,7µs
slave frame
t_s
The reply delay time-out can be
raised up to 83,4 µs for longer
distances
(with reduced troughput).
t_source
The distance is limited by the maximum allowed reply delay of 42,7 µs
between a master frame and a slave frame.
max
repeater
delay
repeater
delay
repeater
delay
t_ms
remotest
data source
propagation delay
(6 µs/km)
repeater repeater

2003 July, HK
MVB Telegrams
message
tranport
control
final function
origin node
origin function
Process Data
Message Data
16, 32, 64, 128 or 256 bits of Process Data
4 bits 12 bits
F =
0..7
Master Frame (Request) Slave Frame (Response)
dataset
time
256 bits of Message Data
source
device
destination
device
prot
ocol size
time
FN FF ON OF MTC
Master Frame
final node
4 bits 12 bits
F =
8-15
Master Frame
16 bits
Slave Frame
Supervisory Data
time
port
address
port
address
4 bits 12 bits
F =
12
source
device
decoded
by
hardware
Telegrams are distinguished by the F_code in the Master Frame
transport data

2003 July, HK
Source-addressed broadcast
The device which sources that variable responds with a slave frame
containing the value, all devices subscribed as sink receive that frame.
The bus master broadcasts the identifier of a variable to be transmitted:
Phase1:
Phase 2:
devices
(slaves)
bus
master
bus
subscribed devices
subscribed
device
subscribed
device
source sink sink
sink
variable value
bus
bus
master devices
(slaves)
source sink sink
subscribed devices
sink
subscribed
device
subscribed
device
variable identifier

2003 July, HK
Traffic Memory
bus
The bus and the application are (de)coupled by a shared memory, the
Traffic Memory, where process variables are directly accessible to the application.
Process Data
Base
Application
Processor
Bus
Controller
Traffic Memory
Associative
memory
two pages ensure that read and
write can occur at the same time

2003 July, HK
Restriction in simultaneous access
page 1 becomes valid
t2
t1
writer
reader 1
page0
page1
(slow) reader 2
page 0 becomes valid
time
• there may be only one writer for a port, but several readers
• a reader must read the whole port before the writer overwrites it again
• there may be no semaphores to guard access to a traffic store (real-time)
traffic store
starts
ends
error !
• therefore, the processor must read ports with interrupt off.

2003 July, HK
Operation of the traffic memory
In content-addressed ("source-addressed") communication, messages are broadcast,
the receiver select the data based on a look-up table of relevant messages.
For this, an associative memory is required.
Since address size is small (12 bits), the decoder is implemented by a memory block:
0
1
2
4
5
6
7
voids
4091
4092
4093
4094
4095
0
0
0
1
2
0
0
voids
0
4
0
3
0
12-bit Address
data(4)
data(5)
data (4094)
storage
bus
processor
data(4092)
port index table
0
data(4)
data(5)
data (4094)
data(4092)
0
page 0 page 1

2003 July, HK
MVB F_code Summary
F_code
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
address
logical
all devices
device
device
device
device
group
device
device
request
Process_Data
reserved
reserved
reserved
Master_Transfer
General_Event
reserved
reserved
Message_Data
Group_Event
Single_Event
Device_Status
source
single
device
subscribed
as
source
Master
>= 1devices
-
-
single device
>= 1devices
single device
single device
size
16
32
64
128
256
-
-
-
16
16
-
-
256
16
16
16
response
Process_Data
(application
-dependent)
Master_Transfer
Event_Identifier
Message_Data
Event_Identifier
Event_Identifier
Device_Status
destination
all
devices
subscribed
as
sink
Master
Master
selected device
Master
Master
Master or monitor
Master Frame Slave Frame

2003 July, HK
MVB Outline
2. Physical layer
2. Middle-Distance
3. Fibre Optics
9. Summary
3. Device Classes

2003 July, HK
Master Operation
The Master performs four tasks:
1) Periodic Polling of the port addresses according to its Poll List
2) Attend Aperiodic Event Requests
3) Scan Devices to supervise configuration
4) Pass Mastership orderly (last period in turn)
The Administrator is loaded with a configuration file before becoming Master
SD
periodic
phase
time
event
phase
guard phase
1 2 3 4 5 6 1 2 9
8 1 2
? ? ? ? ? ? ? ? EV
7
guard phase
super-
visory
phase
SD
basic period basic period
periodic
phase
event
phase
super-
visory
phase
sporadic phase
sporadic phase

2003 July, HK
Bus Traffic
State of the Plant
Response in 1..200 ms
Spurious data losses will be
compensated at the next cycle
event
Sporadic Data
time
On-Demand Transmission
Events of the Plant
Response at human speed: > 0.5 s
• Initialisation, calibration
Flow control & error recovery
protocol for catching all events
• Diagnostics, event recorder
Basic Period Basic Period
State Variable Messages
... commands, position, speed
Periodic Transmission
Periodic Data

2003 July, HK
MVB Medium Access
Between periodic phases, the Master continuously polls the devices for events.
A basic period is divided into a periodic and a sporadic phase.
During the periodic phase, the master polls the periodic data in sequence.
Since more than one device can respond to an event poll, a resolution procedure
selects exactly one event.
Periodic data are polled at their individual period (a multiple of the basic period).
periodic
phase
?
time
sporadic
phase
1 2 3 4 5 6
guard
time
7 8 9 10
basic period
periodic
phase
?
basic period
1 2
sporadic
phase
!
events ? events ? event
data
guard
time
? ? ? ? ? 1 2 3
individual period

2003 July, HK
MVB Bus Administrator Configuration
The Poll List is built knowing:
the list of the port addresses, size and individual period
the reply delay of the bus
the list of known devices (for the device scan
the list of the bus administrators (for mastership transfer)
•
•
•
•
1 1 1 1 1
2.0 2.0 2.0
4.0
4 ms
time
8.2 4.0
period 0 period 1 period 2 period 3
begin of turn
Tspo Tspo
2.1 2.1
4.1
cycle 2
period 4
Tspo Tspo
1 ms 1 ms 1 ms 1 ms
2 ms
2 ms

2003 July, HK
MVB Poll List Configuration
The algorithm which builds the poll table spreads the cycles evenly over the macroperiod
1
8.1
1
2
3
4
5
6
7
1
7
1 1
4.2
1 1 1 1 1 1
4.0
1
basic
period
period
period
period
period
period
period
period
period
T_spo
1
2.0
8.1
0
2.1
0
2.0
2.1 2.0
2.1 2.0
2.1 2.0
2.1 2.1
4.2
4.0 4.0
period
macroperiod (8 T_bp shown, in reality 1024 T_bp)
guard
1 BP datasets
2 BP datasets
2 BP datasets
>350µs

2003 July, HK
MVB Event Resolution (1)
To scan events, the Master issues a General Event Poll (Start Poll) frame.
A device with a pending event returns an Event Identifier Response.
The Master returns that frame as an Event Read frame to read the event data
If only one device responds, the Master reads the Event Identifier (no collision).
If no device responds, the Master keeps on sending Event Polls until a device
responds or until the guard time before the next periodic phase begins.
Start Event Poll
(parameters and
setup)
Event Identifier
Response
from slave
CS
Event Identifier
returned as master
frame
Event data
12 1234
Event Poll telegram Event Read telegram
xxxx
time
SSD CS
MSD xxxx
9 EMET - CS
MSD 12 1234 CS
SSD

2003 July, HK
The devices are divided into groups on the base of their physical addresses.
The Master first asks the devices with an odd address if they request an
event.
• If only one response
comes, the master returns
that frame to poll the event.
If several devices respond to an event poll, the Master detects the collision and
starts event resolution
• If collision keeps on, the
master considers the 2nd
bit of the device address.
• If there is no response,
the master asks devices
with an even address.
C
event
reading
any? xxx1 xx11 N x101 0101 A
time
group poll
collision silence
individual
poll
valid event frame
start poll and
parameter
setup
A D
collision
arbitration round
C C
telegram

2003 July, HK
000 100 010 110 001 101 011 111
x00 x10 x01 x11
xx0 xx1
silence
collision
n = 0
n = 1
width of
group
address
no event
individual poll
collision
silence
event read
n = 2
odd devices
even devices
EA EA EA EA EA EA EA EA
time
collision
silence
collision
xxx
silence
Example with a 3-bit device address: 001 and 101 compete
general poll
start arbitration

2003 July, HK
MVB Time Distribution
At fixed intervals, the Master broadcasts the exact time as a periodic variable.
When receiving this variable, the bus controllers generate a pulse which can
resynchronize a slave clock or generate an interrupt request.
Bus controller
Sync port address
Bus master
Periodic
list
Sync port variable
Master
clock
Bus controller
Slave
clock
Ports
Int Req
Application
processor 2
Bus controller
Slave
clock
Ports
Int Req
Application
processor 3
Bus controller
Ports
Int Req
Application
processor 1
Slave
clock
MVB

2003 July, HK
MVB Slave Clock Synchronization
Slave clocks
Bus
administrator 1
Bus
administrator 2
Synchronizer
Slave clock
MVB 1
Master
clock
Slave devices
Slave clocks
The clock does not need to be generated by the Master.
The clock can synchronize sampling within 100 µs across several bus segments.
MVB 2

2003 July, HK
MVB Outline
2. Physical layer
2. Middle-Distance
3. Fibre Optics
9. Summary
3. Device Classes

2003 July, HK
MVB Fault-tolerance Concept
Transmission Integrity
MVB rather stops than provides false data.
The probability for an undetected transmission error (residual error rate)
is low enough to transmit most safety-critical data.
This is achieved through an extensive error detection scheme
Transmission Availability
MVB continues operation is spite of any single device error. In
particular, configurations without single point of failure are possible.
Graceful Degradation
The failure of a device affects only that device, but not devices which
do not depend on its data (retro-action free).
Configurability
Complete replication of the physical layer is not mandatory.
When requirements are slackened, single-thread connections may
be used and mixed with dual-thread ones.
This is achieved through a complete duplication of the physical layer.

2003 July, HK
MVB Basic Medium Redundancy
The bus is duplicated for availability (not for integrity)
One frame may go lost during switchover
A frame is transmitted over both channels simultaneously.
The receiver receives from one channel and monitors the other.
Switchover is controlled by signal quality and frame overlap.
decoder
receivers
transmitters
bus line A
bus line B
bus controller
encoder selector
address data
parallel bus logic
send register receive register
A B A B
decoder
control
signal quality report

2003 July, HK
MVB Medium Redundancy
The physical medium may be fully duplicated to increase availability.
Duplicated and non-duplicated segments may be connected
Principle: send on both, receive on one, supervise the other
repeater
repeater
A
B
device
electrical segment X
optical link A
optical link B
electrical segment Y
device
A B
device
device
repeater
repeater

2003 July, HK
MVB Double-Line Fibre Layout
A
B
A
B
star coupler B
Bus Administrator
opto links A
opto links B
star coupler A
copper bus A
copper bus B
redundant
Bus
Administrator
The failure of one device cannot prevent other devices from communicating.
Optical Fibres do not retro-act.
device
rack

2003 July, HK
MVB Master Redundancy
To increase availability, the task of the bus master may be assumed by one of
several Bus Administrators
If a bus administrator detects no activity, it enters an arbitration procedure. If
it wins, it takes over the master's role and creates a token.
token passing
Bus
current bus
master
bus
administrator
1
slave
device
slave
device
slave
device
slave
device
slave
device
slave
device
slave
device
bus
administrator
2
bus
administrator
3
A centralized bus master is a single point of failure.
The current master is selected by token passing:
To check the good function of all administrators, the current master offers
mastership to the next administrator in the list every 4 seconds.

2003 July, HK
MVB Outline
2. Physical layer
2. Middle-Distance
3. Fibre Optics
9. Summary
3. Device Classes

2003 July, HK
BT0.5
MVB Transmission Integrity (1)
Manchester II encoding
Double signal inversion necessary to cause an undetected error, memoryless code
Clock
Data
Frame
Manchester II symbols
Line Signal
1 1 0 1 0 0 0 1
violations
2) Signal quality supervision
Adding to the high signal-to-noise ratio of the transmission, signal quality
supervision rejects suspect frames.
time
BT = bit time = 666
ns
reference
edge
125ns
125ns 125ns
1)
Start Delimiter
BT1.0
BT1.5

2003 July, HK
F address
9 4 12 8
9 16
2 bytes
8
CS
size in bits
repeat 1, 2 or 4 x
CS
9 32
4 bytes
8
CS
9 64
8 bytes DATA
64
8
CS
Master Frame
MSD
SSD
SSD
SSD
16 (33)
16 (33)
32 (49)
64 (81)
128 (153)
256 (297)
MD = Master frame Delimiter
CS = Check Sequence 8 bits
SD = Slave frame Delimiter
useful (total)
size in bits
3) A check octet according to TC57 class FT2 for each group of up to 64 bits,
provides a Hamming Distance of 4 (8 if Manchester coding is considered):
Slave Frame
(Residual Error Rate < 10 under standard disturbances)
-15

2003 July, HK
mm
MSD ADDRESS a ADDRESS b DATA (b)
MSD SSD
accept if 0.5µs < t_mm < 42.7 µs
time
CS DATA (a)
SSD CS CS CS
5) Response time supervision against double frame loss:
MSD ADDRESS a DATA (a)
SSD
respond within
1.3 µs < t < 4.0 µs
ms
CS CS MSD ADDRESS b CS
respond within
4 µs < t <1.3 ms
sm
time
1,3 ms
4) Different delimiters for address and data against single frame loss:
6) Configuration check: size at source and sink ports must be same as frame size.
> 22 µs > 22 µs
t

2003 July, HK
MVB Safety Concept
Very high data integrity, but nevertheless insufficient for safety applications
(signalling)
Increasing the Hamming Distance further is of no use since data falsification
becomes more likely in a device than on the bus.
• critical data transmitted periodically to guarantee timely delivery.
Data Transfer
Redundant plant inputs A and B transmitted by two independent devices.
Device Redundancy
Availability
Data Integrity
Availability is increased by letting the receiving devices receive both A and
B. The application is responsible to process the results and switchover to the
healthy device in case of discrepancy.
Diverse A and B data received by two independent devices and compared.
The output is disabled if A and B do not agree within a specified time.
• obsolete data are discarded by sink time supervision.
• error in the poll scan list do not affect safety.

2003 July, HK
MVB Integer Set-up
redundant vehicle bus
(for availability only)
input
devices
redundant input
fail-safe
comparator
and enabling
logic
redundant, integer output
°
A B A B
poll
time
individual period
spreader device
(application
dependent)
output
devices
confinement
A B
A B
application
responsibility
Bus
Administrator

2003 July, HK
MVB Integer and Available Set-up
redundant vehicle bus
(for availability)
redundant input
B
A
comparator
and enabling
logic
A B
available and integer output
switchover logic or
comparator
(application
dependent)
A B A B
poll
time
individual period
spreader device
(application
dependent)
output
devices
confinement
B
A B
A
input
devices
A B
C C
redundant
bus
administrator
redundant
bus
administrator

2003 July, HK
MVB Outline
2. Physical layer
2. Middle-Distance
3. Fibre Optics
9. Summary
3. Device Classes

2003 July, HK
MVB Summary
Topography:
Medium:
Covered distance: OGF: 2000 m, total 4096 devices
Communication chip
Processor participation none (class 1), class 2 uses minor processor capacity
Interface area on board
Additional logic RAM, EPROM , drivers.
Medium redundancy: fully duplicated for availability
Signalling: Manchester II + delimiters
Gross data rate
Response Time
Address space
Frame size (useful data)
bus (copper), active star (optical fibre)
copper: twisted wire pair
optical: fibres and active star coupler
EMD: 200 m copper with transformer-coupling
dedicated IC available
20 cm2 (class 1), 50 cm2 (class 2)
1,5 Mb/s
typical 10 µs (<43 µs)
4096 physical devices, 4096 logical ports per bus
16, 32, 64, 128, 256 bits
Integrity CRC8 per 64 bits, HD = 8, protected against sync slip
ESD: 20 m copper (RS485)

2003 July, HK
MVB Link Layer Interface
telegram
handling
Lower Link
Layer
message
data
supervisory
data
Traffic Store
process
data
LP LM LS
Upper Link Layer
Real-Time Protocols
Physical Layer
master
polling
arbitration
mastership transfer
station
management
Link Layer Interface
slave
Process Data
Message Data
Supervisory Data
frame coding

2003 July, HK
MVB Components
Bus Controllers:
BAP 15 (Texas Instruments, obsolete)
MVBC01 (VLSI, in production, includes master logic
MVBC02 (E2S, in production, includes transformer coupling)
Medium Attachment Unit:
ESD: fully operational and field tested (with DC/DC/opto galvanic separation)
OGF: fully operational and field tested (8 years experience)
EMD: lab tested, first vehicles equipped
Stack:
Link Layer stack for Intel 186, i196, i960, 166, 167, Motorola 68332, under
DOS, Windows, VRTX,...
Bus Administrator configurator
Tools:
Bus Monitor, Download, Upload, remote settings
Repeaters:
REGA (in production)
MVBD (in production, includes transformer coupling)

2003 July, HK
MVB Throughput (raw data)
32
16 48 64 80 96 112 128 144 160 176 192 208 224 240 256
MVB @ 1,5 Mbit/s
IEC Fieldbus @ 1,0 Mbit/s
IEC Fieldbus @ 2,5 Mbit/s
dataset size in bits
transmission delay [ms]
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.1

2003 July, HK
MVB & IEC 61158-2 Frames
Preamble Start Delimiter
Data
1 N+N- 1 0 N+ N- 0 1 N+N- N+N- 1 0 1
0 N-N+ 1 N- N+ 1 1 1
1 0 1 0 1 0 1 0
End Delimiter
Data
v v v v
Spacing
v v
0 0 0 0 N+ N- 0 N+N- Data v v
Master Frame
Slave Frame
PhSDU
FCS
FCS
FCS
IEC 61158-2 frame
MVB frame
8 bits
16 bits
IEC65 frames have a lesser efficiency (-48%) then MVB frames
To compensate it, a higher speed (2,5 Mbit/s) would be needed.
End Delimiter
Start Delimiter

2003 July, HK

lecture Instrumentation - Sensors and actorsPLC.pdf

More Related Content

Similar to lecture Instrumentation - Sensors and actorsPLC.pdf (20)

More from kghanem2 (20)

Recently uploaded (20)

lecture Instrumentation - Sensors and actorsPLC.pdf