Curs1 RIL Stud

Industrial Automation
Automation Industrielle
Industrielle Automation
Reele industriale locale

7
Application
Presentation
Session
Transport
Network
Link
Physical
Titular curs: s.l. dr. ing. Nicoleta-Cristina G ITAN

cristinag@eed.usv.ro
2004 June, HK

Durat :
14s pt mnix3 ore
Forma de examinare:
Examen semestrial
EPFL - Industrial Automation
2004 June, HK
3.3.1 OSI model
Cum se face notarea?
Nota la examen NF = 50%NL + 40%NE + 10%PC
NF
Nota Final
NE
Nota la Examen
NL
Nota la Laborator
PC
PrezenalaCurs
Examenulsedesf oar nscris.
Sedaubonusuridepn 10%pentruactivit isuplimentare(participarealaconcursuri,

conferinestudeneti,proiecte,etc.)
2004 June, HK
3.3.1 OSI model
The OSI model
The Open System Interconnection (OSI) model is a standard way to

structure communication software that is applicable to any network.
was developed to structure telecommunication protocols in the 70
(Pouzin & Zimmermann)
standardized by CCITT and ISO as ISO / IEC 7498
all communication protocols (TCP/IP, Appletalk or DNA) can be mapped to the
OSI model.
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)
mapping of OSI to industrial communication requires some additions
2004 June, HK
3.3.1 OSI model
OSI-Model (ISO/IEC standard 7498)
"Application"
protocols
"Transport"
protocols
Application
Presentation
Session
Transport
Network
Link
Physical
2004 June, HK
All services directly called by the end user

(Mail, File Transfer,...) e.g. Telnet, SMTP
Definition and conversion of the data
formats (e.g. ASN 1)
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)
3.3.1 OSI model
OSI Model with two nodes
node 1
node 2
Application
Presentation
Session
Transport
Network
Link
Physical
Physical Medium
2004 June, HK
3.3.1 OSI model
Repeater
500m
department A
workstations
Ethernet
To connect a workstation of
department A to the printer of
department B, the cable becomes too
long and the messages are corrupted.
server
The repeater restores signal

levels and synchronization.
It introduces a signal delay of
about 1..4 bits
repeater
printer
500m
department B
server
Physically, there is only one bus carrying
both departments traffic, only one node
may transmit at a time.
Ethernet
500m
2004 June, HK
3.3.1 OSI model
OSI model with three nodes (bridge)

Node 1
bridge
Node 2
Application
Presentation
Session
Transport
Network
Link
Physical
physical medium (0)
physical medium (0)
e.g. Ethernet 100 MBit/s
e.g. ATM
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
2004 June, HK
3.3.1 OSI model
Bridge example
In this example, most traffic is directed from the workstations to the department
server, there is little cross-department traffic
department A
workstations
Ethernet
department A
server
Ethernet 1
repeater
printer
Bridge
printer
department B
department B
server
Ethernet
server
Ethernet 2
There is only one Ethernet which carries

both departments traffic
server
2004 June, HK
There are now two Ethernets and only the

cross-department traffic burdens both busses
9
3.3.1 OSI model
Networking with bridges
LAN
port
port
port
port
port
Spanning-tree-Algorithmen
avoid loops and ensures
redundancy
port
LAN
port
port
LAN
2004 June, HK
port
LAN
port
port
LAN
10
3.3.1 OSI model
Switch
nodes
queues
crossbarswitch
(or bus)
full-duplex
a switch is an extension of a hub that allows store-and-forward.
2004 June, HK
11
3.3.1 OSI model
OSI Model with three nodes (router)

Node 1
Router
Node 2
Application
Presentation
Session
Transport
Network
Link
Physical
physical medium (0)
The router routes the frames on the base of their network address.
The subnets may have different link layer protocols
Frames in transit are handled in the network layer .
2004 June, HK
12
3.3.1 OSI model
Repeater, Bridge, Router, Gateway: Topography

applicationdependent
gateway
backbone (e.g. FDDI)
different subnetworks,
same address space
same transport protocol,
segmentation/reassembly
routers are initially opaque
end-to-end
transport protocol
connects different speed,

different medium access
by store-and-forward
same frames and addresses
initially transparent in both ways.
can limit traffic by filtering
Router
Bridge
same speed
same medium
access
same frames
Repeater
segment
subnet (LAN, bus, extended link)
devices (nodes, stations) have different physical addresses
devices (nodes, stations) have different link addresses
2004 June, HK
13
3.3.1 OSI model
Repeaters, Bridges, Routers and Gateways: OSI model
intelligent linking devices can

do all three functions
(if the data rate is the same)
gateway
Apl
Apl
Apl
Pre
Pre
Pre
Pre
Ses
Ses
Ses
Ses
Trp
Trp
Trp
TCP
Net
IP
LLC
bridge
( "switch")
Trp
(store-and-forward)
Net
MAC
MAC
MIS
Layer 1
MDS
MDS
10 Mbit/s coax
Layer 3
Net
repeater
or hub
LLC
Layer 2
MDS
Apl
router
LLC
LLC
LLC
MAC
MAC
MAC
MAC
MIS
MIS
MIS
MIS
MIS
MDS
MDS
MDS
MDS
MDS
10 Mbit/s fibre
100 Mbit/s Ethernet
RPC
MDS
Fibre
ATM 155 Mbit/s
2004 June, HK
14
3.3.1 OSI model
application
presentation
session
network
link
physical
transport
To which level does a frame element belong ?
link phy
LLC
repeater, hub
destination
source
LLC
bridge
final
origin
router
NC
TRP
SES
PRE
Network Control
preamble
APL
CRC
ED
application
(gateway)
A frame is structured according to the ISO model
2004 June, HK
15
3.3.1 OSI model
Encapsulation
INFO
User information
TrpCrt
Transport header
NetAdr
Network address
Link control
(Acknowledge, Token,etc.)
LinkCrt
LinkAdr
Link-address
size
Error detection
CRC
Flag
Frame
Flag
Signal
Each layer introduces its own header and overhead
2004 June, HK
16
3.3.1 OSI model
Example: OSI-Stack frame structure

IEEE 802.4
token bus
ISO 8802
logical link control
MAC_header
13
MA. frame control
MA. destination
address
(6 octets)
ISO 8473
connectionless network control
LNK_hdr
NET_header
>48
L_destination SAP
ISO 8073
class 4 transport control
TRP_header
5
Protocol Identifier (81)
L_source SAP
Header Length
L_PDU
Version/Protocol ID (01)
LSAP = DSAP
FE = network layer
18 = Mini-MAP Object
Dictionary Client
19 = Network Management
00 = own link layer
Lifetime
SP MS ER
DT/ER Type
PDU Segment Length
L_PDU = UI, XID, TEST
Checksum
DATA
LI
TPDU
FIXED
PART
(CDT)
Destination
Reference
N(S)
ET
DATA (DT) TPDU

(normal format)
address length
IDP
(initial
domain
part)
AFI = 49
Destination Address
(18 octets)
MA. source
address
(6 octets)
ADDRESS
PART
Source Address
(18 octets)
Segmentation
(0 or 6 octets)
Options
(priority = 3 octets)
IDI, Area ID
(7 octets)
PSI
Physical Address
(6 octets)
LSAP = FE
DSP
(domain
specific
part)
NSAP = 00
2004 June, HK
17
3.3.1 OSI model
Protocol Data Units and Service Data Units
N+1- Layer
ServiceData Unit
(SDU)
N - Layer
Protocol
Data Unit
(PDU)
(n)-layer entity
(n)-layer entity
Protocol
Data Unit
(PDU)
ServiceData Unit
(SDU)
N-1 Layer
(n+1)-layer entity
(n+1)-layer entity
(n-1)-layer entity
(n-1)-layer entity
Layer N provides services to Layer N+1;

Layer N relies on services of Layer n-1
2004 June, HK
18
3.3.1 OSI model
Service Access Points
user of
service N
user of
service N
Service Access Points (SAP)
functions in layer N
Service Access Points (SAP)
provider of service (N-1)

provider of service (N)
Service Access Points represent the interface to a service (name, address, pointer,...)
2004 June, HK
19
3.3.1 OSI model
Address and SAPs in a device
ASAP
z.B. TCP/IP
z.B. ISO 8073
Transport
TSAP
Transport-SAP
NSAP
Network-SAP
(not Network address)
LSAP
Logical Address or link address
ISO 8473
Network
ISO-stack
Link
Physical
PhSAP
2004 June, HK
Application
(z.B. File transfer, Email,....)
20
Physical Address
3.3.1 OSI model
Procedure call conventions in ISO
Service User
Service Provider
(Network Transmission)
Service User
request
indication
confirm
(local)
confirm
(network)
confirm
(user)
response
time
2004 June, HK
21
3.3.1 OSI model
OSI implementation
OSI should be considered as a model, not as an implementation guide
The idea of independent layers is a useful as a way of thinking, not the best implementation.
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.:
ISO/IEC 8473-2:1996 Information technology -- Protocol for providing
the connectionless-mode network service -ISO/IEC 8073:1997 Information technology -- Open Systems
Interconnection -- Protocol for providing the connection-mode transport service
ISO/IEC 8327-1:1996 Information technology -- Open Systems
Interconnection -- Connection-oriented Session protocol: Protocol specification
ISO/IEC 8649:1996 Information technology -- Open Systems
Interconnection -- Service definition for the Association Control Service Element
ISO 8571-2:1988 Information processing systems -- Open Systems
Interconnection -- File Transfer, Access and Management
Former implementations, which implemented each layer by an independent process,

caused the general belief that OSI is slow and bulky.
OSI stack has not been able to establish itself against TCP/IP
2004 June, HK
22
3.3.1 OSI model
OSI protocols in industry

Theory:
ISO-OSI standards should be used since they reduce specification and
conformance testing work and commercial components exist
Reality:
the OSI model is a general telecommunication framework implementations considers feasibility and economics.
the overhead of the ISO-OSI protocols (8073/8074) is not bearable
with low data rates under real-time conditions.
the OSI-conformant software is too complex:
simple devices like door control or air-condition have limited power.
the OSI model does not consider transmission of real-time data

Therefore:
industrial busses use for real-time data a fast response access and
for messages a simplified OSI communication stack
the devices must be plug compatible: there are practically no options.
Communication is greatly simplified by adhering to conventions
negotiating parameters at run-time is a waste in closed applications.
2004 June, HK
23
3.3.1 OSI model
TCP / IP structure
FTP SMTP
Files
TCP
HTTP
routing
ATM
Applications
Transport
UDP
IP
Ethernet
SNMP
modem
ICMP
Network
radio
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
2004 June, HK
24
3.3.1 OSI model
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
TCP/IP/UDP is becoming the backbone for all non-time critical industrial communication
TCP/IP/UDP is quickly displacing proprietary protocols.
Next generation TCP/IP (V6) is very much like the OSI standards.
Many embedded controllers come with an integrated Ethernet controller, an the
corresponding real-time operating system kernel offers TCP/IP services
Like OSI, TCP protocols have delays counted in tens or hundred milliseconds,
often unpredictable especially in case of disturbances.
For further reading: Motorola Digital Data Communication Guide
2004 June, HK
25
3.3.1 OSI model

Titular curs: s.l. dr. ing. Nicoleta-Cristina GITAN
2004 June, HK
3.3.2 Field busses - Physical Layer
3.3.2
3. Industrial Communication Systems

Physical Layer
2004 June, HK
Physical Layer Outline

1. Layering
2. Topology
3. Physical media
4. Electric Signal coupling
5. Optical Fibres
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
To probe further
Henri Nussbaumer, Tlinformatique 1, Presses polytechniques romandes

Fred Halsall, Data Communications, Computer Networks and Open Systems, Addison-Wesley
B. Sklar , Digital Communications, Prentice Hall, Englewood Cliffs, 1988
EIA Standard RS-485
2004 June, HK
OSI Model - location of the physical level
Application
protocols
Application
Presentation
Session
Transport
Network
Link
Physical
Transport
protocols
2004 June, HK

(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)
Subdivisions of the physical layer
medium-independent signalling
medium-dependent signalling
Physical
Layer
electrical / optical
specifications
2004 June, HK
applies to one media

(e.g. optical fibres)
applies to one media type
(e.g. 200m optical fibres)
mechanical
specifications
same for different media

(e.g. coax, fibre, RS485)
defines the mechanical interface

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

Signal quality supervision, redundancy control
2004 June, HK
Example: RS-232 - Mechanical-Electrical Standard

Originally developed for modem communication, now serial port in IBM-PCs
Topology:
Data
Terminal
Equipment
DTE
DCE
Terminal
Telephone
lines
DCE
Data Communication
Equipment (Modem)
Cabling rules
Modem
Data
Terminal
Equipment
Computer
modem eliminator
extension
Tip: Do not use

Modem cables,
only Extension
cables
DTE
2
computer
extension
terminal
cable
cable
2 3
Mechanical
1
25
Electrical:
transmitter
receiver
+12V
+3V
-12V
-3V
2004 June, HK
"0" Space On
"1" Mark Off

1. Layering
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
Topology: Simplex, Half and Full Duplex

Link (Point -To-Point)
Full-duplex
Sender/
Receiver
Sender/
Receiver
Sender/
Receiver
Half-duplex
Sender/
Receiver
Examples:
RS232
Examples:
RS485
Bus (Half-Duplex, except when using Carrier Frequency over multiple bands)
Terminator
Examples:
Ethernet, Profibus
Ring (Half-Duplex, except double ring)
Examples:
SERCOS, Interbus-S
consists of point-to-point links
2004 June, HK
10
Bus topologies
party-line
Terminator
Terminator
advantage: little wiring

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

2004 June, HK
11
Repeater
500m
department A
workstations
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.
repeater
500m
printer
department B
server
Ethernet
500m
2004 June, HK
12
Bus: repeaters and hubs
higher-level hub
repeaters
point-to-point
link
partyline
partyline
2004 June, HK
13
Party-line (bus) and star wiring

d = average distance between devices
PLC
wiring length = d n,
increases linearly with number of devices
Up to 32 devices
(more with repeaters)
I/O
hub
I/O
I/O
I/O
wiring length = d n n / 2 2
increases with square of number of devices
PLC
Up to 16 devices
per hub
does it fit into the
wiring tray ?
I/O
I/O
2004 June, HK
I/O
14
I/O
I/O
I/O
Rings
classical ring
ring in floor wiring

wiring
cabinet
2004 June, HK
15

1. Layering
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
16
Media (bandwidth x distance)

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

2004 June, HK
17

1. Layering
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
18
Electrical: Transmission media
Zw = 50 ... 100
Cost
efficient
core
Coaxial
cablewiring:
twisted pair (without
dielectric
screen
inflexible, costly,
low losses
10 MHz..100 MHz
shield
Shielded twisted wire

(Twinax)
Zw = 85..120
Shield
Unshielded twisted wire

Uncommitted wiring
(e.g. powerline com.)
twisting compensates disturbances
very cheap,
sensible to perturbations
Telephone
numerous branches, not terminated,
except possibly at one place
2004 June, HK
flexible, cheap,
medium attenuation
~1 MHz..12 MHz
19
very cheap,
very high losses and disturbances,
very low speed (~10 ..100 kbit/s)
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)
2004 June, HK
20
Electrical: What limits transmission distance ?
All parameters
Characteristic impedance
Attenuation
Linear resistance
Linear capacitance
Cross talk
Common-mode
Shield protection
are frequency-dependent
Signal reflection on discontinuities (branches, connectors) cause self-distortions
2004 June, HK
21
Consider in cables
lumped line model

L'
L'
R'
C'
Zw =
G'
L'
L'
R'
C'
L'
R'
G'
specific inductance (H/m)

specific resistance (/m)
C'
G'
R'
C'
G'
specific capacitance (F/m)

specific conductance (S/m)
C'
2004 June, HK
22
Electrical: Signal Coupling Types

Resistive
direct coupling
driver on line without galvanic coupling
collision possible when several transmitters active
Wired-OR combination possible

cheap as long as no galvanic separation is required (opto-coupler)
good efficiency
Inductive
transformer-coupling
galvanic separation
good electromagnetic compatibility (filter)
retro-action free
good efficiency
signal may not contain DC-components
bandwidth limited
Capacitive
capacitor as coupler
weak galvanic separation
signal may not contain DC components
cheap
not efficient
2004 June, HK
23
Electrical: Resistive (direct) coupling
Unipolar, unbalanced
Bipolar, unbalanced
+ Us
Coax
Ru
+ Us
Zw
Zw
Rd
- Us
Open Collector
Ut
(unbalanced)
Ut = 5 V (e.g.)
Rt
Rt
Bus line, characteristic impedance = Zw
Terminator and
Pull-up resistor
wired-OR behaviour
(Low wins over High
Out
In
device
2004 June, HK
Out
In
device
24
Out
In
device
Electrical: Balanced Transmission
+Ub
Zw
100
symmetrical line (Twisted Wire Pair)
Shield
Rt
UA
Differential amplifier
(OpAmp)
UB
(Data Ground)
Used for twisted wire pairs (e.g. RS422, RS485)
2004 June, HK
25
Electrical: RS-485 as an example of balanced transmission
TxS RxS
TxS RxS
TxS RxS
100
Terminator
stub
A
tap
120
120
Data-GND
B
Zw 120, C' 100 pF/m

segment length
multiple transmitter
allowed
Short-circuit limitation
needed
Ishort < 250 mA

2004 June, HK
26
Electrical: RS-485 Distance x Baudrate product
distance
10000
5000
limited by copper resistance

100 /km -> 6dB loss limit
2000
1200
1000
500
limited by
frequency-dependent
losses 20 dB/decade
200
100
50
20
12
10KBd
100KBd
2004 June, HK
1 MBd
27
10 MBd
Baudrate
Electrical: Transformer Coupling
Sender/Receiver
Sender/Receiver
Isolation transformer
isolation resistors
shield
Twisted Wire Pair
Source: Appletalk manual
2004 June, HK
28
Electrical: MIL 1553 as an example of transformer coupling

Direct Coupling
(short stub: 0.3 m)
Double-Transformer
(long stub: 0.3 .. 6m)
Sender/Receiver
short
stub
long
stub
isolation resistors
isolation resistors
shield
shield
Twisted Wire Pair

Extract from: MIL-STD-1553
2004 June, HK
29
Electrical: Free topology wiring
voltage
source
terminator
Free topology is used to connect scattered devices which are usually line-powered.
Main application: building wiring
2004 June, HK
30
Electrical: Power Line Carrier technology
HF-trap
220V
A free-topology medium using the power lines as carrier.

Capacitive or inductive coupling, sometimes over shield
Problems with disturbances, switches, transformers, HF-traps, EMC,..
2004 June, HK
31
Electrical: Mechanical Connecting devices to an electrical bus
short stub
thread-through
double-connector
junction box
2 connectors
live insertion
installation ?
1 connector
live insertion
(costly) junction box
stub
1 connector
live insertion
2 connectors
no live insertion
2004 June, HK
32
Practical solution to live insertion
2004 June, HK
33
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)
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.
2004 June, HK
34
Electrical: Water-proof Connectors
2004 June, HK
35

1. Layering
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
36
Fiber: Principle
3 components:
transmitter
fibre
GaAs
LED
receiver
PIN
fotodiode
Is
different refraction
coefficients
Transmitter, cable and receiver must be "tuned" to the same wavelength
2004 June, HK
37
Fiber: Types
Multimodefibre
Monomode fibre
N(r)
50 m
Refraction profile
Cross-section
50 - 300 m
50 - 100 m
2-10 m
total reflection
gradual reflection
waveguide
3 dB/km
0,6 dB/km
1 GHzkm
2,3 dB/km
0,4 dB/km
100 GHzkm
Core
Clad
Longitudinal section
(red) 650nm
800nm
(infra-red) 1300nm
10 dB/km
5dB/km
20MHzkm
HCS (Hard-Clad Silica)

200 m, < 500m
2004 June, HK
50 or 62.5 m LAN fibre
38
telecom - costly
Fibre: Use
Type
POF
HCS/PCF
GOF
Material
distance
Usage
Connector
Cost
aging
bending
bandwidth
plastic
70m
local networking
simple
cheap
poor
very good
poor
glass / plastic
400m
remote networking
glass
1km
telephone
high-precision
medium
good
poor
very good
2004 June, HK
39
precision
medium
very good
good
good
Fiber: building an optical bus

n% coupling losses
Passive coupler
n% coupling losses
Passive star coupler
Active star coupler
1
2
3
4
5
6
Fused zone
electrical segment (wired-or)
1
2
3
4
5
6
opto-electrical
transceiver
fibre pair
2004 June, HK
40
Fiber: building an optical ring and bridging

spring
prism
example of
solution
Powered
unpowered
Double ring
2004 June, HK
41
Fiber: advantages
1 ) high bandwidth and data rate (400 MHz x km)
2 ) small, frequency-insensitive attenuation (ca. 3 dB/km)
3 ) cover long distances without a repeater
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
9 ) low cost cable
10) standardized
2004 June, HK
42
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
4) Reliability of optical senders and connections is insufficient (MTTF 1/power).
5) Galvanic isolation can be achieved with coaxial cables and twisted pairs through opto-couplers
6) Tapping is not a problem in industrial plants
7) Optical busses using (cheap) passive components are limited to a few branches (16)
8) In explosive environments, the power requirement of the optical components hurts.
9) Installation of optical fibres is costly due to splicing
10) Topology is restricted by the star coupler (hub) or the ring structure
2004 June, HK
43
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. Apples 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.
bluetooth module
Radio == mobile -> power source (batteries) and low-power technologies.

2004 June, HK
44
Wireless Field busses
but: who changes the batteries ?
2004 June, HK
45
Redundancy at the physical layer

Party-Line
Terminator
Terminator
decentralized wiring
Star topology
both cables can run in the same conduct where

common mode failure acceptable
star couplers should be separately powered

star coupler A
star coupler B
centralized wiring
cable come together at each device
2004 June, HK
46

1. Layering
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
47
Modulation
Base band
Carrier band
Backward
channel
Broadband
2004 June, HK
5-108
MHz
48
Forwardchannel
162-400
MHz
Frequency

1. Layering
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
49
Synchronisation: where does it take place ?
Example: Frame synchronisation using Manchester violation symbols
Data
1 1 0 1 0 0 0 1
Clock
+NRZ Data
+Framing
= Line Signal
Start-sync
(Violation)
2004 June, HK
Data in Manchester II
50
Stop-sync
(Violation)
Frames: Synchronization
character-synchronous
(e.g. bisync)
Data
A B C SYN D E F G
SYN
Signal
SYN SYN
Byte-stuffing
flag
SYN
flag
bit-synchronous
(e.g. HDLC)
Data
11 100 0 11 111
110
011 1 1 1
flag
Signal
0 1 1 1 1 1 10 1 1 1 0 0 0 1 1 1 1 1 0 1 1 0
0 1 1 1 1 1 0 0 0 1 1 1 1 11 0
Bit-stuffing
delimiter
(e.g. IEC 61158)
Signal
"1" "1" "0" "0" "1" "1"
Delimiter (not Manchester)
2004 June, HK
51
Manchester symbols

1. Layering
2. Topology
3. Physical media
5. Optical Fibers
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
52
Encoding: popular DC-free encodings

user
1
Manchester
Differential Manchester
Ethernet, FIP
IEC 61158,
MVB, MIL 1553
LON
High-density
diskettes
Miller (MFM)
Xerxes
FlexRay
memoryless*: decoding does not depend on history

2004 June, HK
53
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|)
effect of unbalance
(magnetic discharge)
2004 June, HK
54
Decoding base-band signals
Dynamic: 10 dB
line
unipolar signal
histeresis
idle level
Uh+
Uh-
time
active
RxS
Dynamic: 18 dB
idle
bipolar signal
Uh+
Uh-
Dynamic: 32 dB
Daten
1 0 1 0 1 0 1 0 1 N+ N- 1 0 N- N+ 0 1 1 1
Preamble
Delimiter
Dynamic: 38 dB
2004 June, HK
55
Encoding: Physical frame of IEC 61158-2

1
N+
N-
N-
N+
N+
N-
+U
Preamble (variable) 0V
-U
Start delimiter (8 bit times)
Payload (variable length)

N- N+
End delimiter (8 bit times)

preamble
1
defines end of frame
start
2004 June, HK
payload
56
end delimiter
Encodings: Multi-frequency
power
"SB1"
"0" "1"
unused
"SB2" "SB3"
"0" "1" "0" "1"
"SB4"
"0" "1"
"SB5"
"0" "1"
"SB6"
"SB7"
"0" "1" "0" "1"
"SB8"
"0" "1"
22,5 kHz 31,5 kHz 40,5 kHz 49,5 kHz 58,5 kHz 67,5 kHz 76,5 kHz 85,5 kHz
27 kHz 36
45 kHz
54
63
72
81 kHz
90 kHz
kHz
kHz
kHz
kHz
frequency
Best use of spectrum

Adaptive scheme (frequency-hopping, avoid disturbed frequencies, overcome bursts)
EMC considerations
2004 June, HK
57
Bandwidth and Manchester Encoding

Delimiter
" 0 " " 1 "" 0 " " 0 " " 1 " " 1 " " 0 "
3-step
2-step
3-step
2004 June, HK
58
Encoding: qualities
1) Self-clocking, Explicit clocking or asynchronous
2) Spectral efficiency
3) Efficiency: relation of bit rate to Baudrate
4) Decoding ease
5) Integrity
6) Polarity
2004 June, HK
59

1. Layering
2. Topology
3. Physical media
5. Optical Fibres
6. Modulation
7. Synchronization
8. Encoding
9. Repeaters
2004 June, HK
60
Repeater
node
node
node
repeater
2004 June, HK
encoder
decoder
(RS 485)
node
node
decoder
encoder
segment 1
node
61
segment 2
(transformer-coupled)
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.
2004 June, HK
62
Star coupler (hub)
opto-electrical
transceiver
wired-or electrical media
to other device
or star coupler
to other device
or star coupler
fibre pair
device
device
2004 June, HK
device
device
63

2004 June, HK
3.3.3 Field busses - Link Layer
3.
Industrial Communication Systems

Link Layer and Medium Access
2004 June, HK
Link Layer Outline

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

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:
1) Data integrity
2) Medium Access
3) Logical Link Control
4) Link Layer Management
2004 June, HK
Link Layer in the OSI Model

Functions
Network
Application
bridge control
Presentation
Session
Logical Link
Control
(LLC)
Medium Access
Control
(MAC)
Transport
Network
Link
Physical
Frame
store-and-forward
address discovery
connection establishment
flow Control
error handling
medium allocation
(master redundancy)
addressing
frame format
integrity control
Physical
Signaling
Physical
Medium
Subnet (Bus or Ring)
2004 June, HK
OSI model with two nodes
Node 1
Node 2
Application
Presentation
Session
Transport
Network
Link
Physical
Physical Medium
2004 June, HK
OSI model with repeater (physical layer connection)

Node 1
Application
repeater
Node 2
Presentation
Session
Transport
Network
Link
Physical
1
physical medium (0)
physical medium (1)
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.
2004 June, HK
OSI model with three nodes (bridge): link layer connection

Node 1
bridge
Node 2
Application
Presentation
Session
Transport
Network
Link
Physical
physical medium (0)
physical medium (0)
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
2004 June, HK
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
HDLC as example of a link layer protocol

Standard
HDLC (High Level Data Link)
Function
Reliable transmission between devices of a subnet
Objects
Frame structure according to ISO 3309.
Flag
01111110
Address
Control
8 bit
8 bit
Data
CRC
(n 8)
16 bit
Integrity Check
16-bit Cyclic Redundancy Check
Medium Access
Master/Slave, (with slave initiative possible)
Error recovery
positive acknowledgement and retry
Flow control
7-frames (127 frames) credit system
2004 June, HK
10
Flag
01111110
HDLC Topology
Secondary
(Terminal)
Secondary
(Terminal)
full-duplex
or
half-duplex
medium
Secondary
(Terminal)
Secondary
(Terminal)
Secondary
(Terminal)
Secondary
(Terminal)
Primary
(Mainframe)
Secondary
(Terminal)
Secondary
(Terminal)
The Primary (Master) is connected with the Secondaries (Slaves) over a

multidrop bus (e.g. RS 485) or over point-to-point lines
HDLC bases physically on a bus, but is logically a star network
2004 June, HK
11
HDLC - Full- and Half duplex operation

Full-Duplex
Command/Data
Requester/
Primary
Responder/
Secondary
Acknowledgement
Sender/Receiver
Sender/Receiver
Command/Acknowledgement
Requester/
Primary
Responder/
Secondary
Data
Sender/Receiver
Sender/Receiver
Half-Duplex
Requester/
Primary
Responder/
Secondary
Sender/Receiver
Sender/Receiver
The Primary switches the Secondary to send mode, the Secondary sends until it returns control
2004 June, HK
12
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
13
Frame Sublayer
Network
Network
Logical Link
Control
(LLC)
Logical Link
Control
(LLC)
Medium Access
Control
(MAC)
Medium Access
Control
(MAC)
Frame
Frame
Physical
Signaling
Physical
Signaling
Physical
Medium
Physical
Medium
sublayer
The frame layer is concerned with the correct frame format and contents,
with (practically) no consideration of the medium or speed.
2004 June, HK
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
codes are not used.
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
2004 June, HK
15
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
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

Quality criteria
Code efficiency: CEF = k/n
Hamming-Distance
Residual Error Rate
2004 June, HK
17
Hamming Distance
The Hamming Distance between two words is the number of bits in which they differ:
Example
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
Number of detectable bit errors:
ZD = HD 1
Numbers of correctable bit errors:
ZC = (HD1)/2
Example: HD = 4: ZD = 3, ZC = 1
00000
00001
code word
00111
2004 June, HK
2nd error
18
01111
code word
m=4
1st error
00011
3rd error
4th error
Hamming distance in 3 dimensions: parity

1
par
D7
D6
D5
D4
D3
D2
D1
D0
Odd parity: sum Modulo-2 of all "1" in the codeword (including the parity bit) is odd
The parity bit is the last transmitted bit (->LSB first, a matter of convention)
illegal
111
110
101
100
legal
011
010
001
000
2004 June, HK
19
Error Detection through CRC

(Cyclic Redundancy Check)
Principle
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.
data (dividend)
/
GP(divisor)
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)
2004 June, HK
20
Residual Error Rate, Parity

Rer = Probability of occurrence of an undetected error in spite of an error detecting
code as a function of the bit error probability
Example:
Hamming Distance
Bit error probability
E
r
Residual error rate
R
er
= 1 - (1 - E )
r
no error
n E (1 - E )
r
r
n -1
exactly one error
Rer for two word length:

E = 10
r
E = 10
r
-5
n = 9 bit
R = 72 1 0
er
-5
n = 513 bit
R = 2.6 1 0
er
2004 June, HK
21
-1 0
-5
quite efficient.
quite useless ...
Integrity classes and 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)
Residual Error Rate
10 0
10 - 2
10 - 4
10 - 6
10 - 8
10 -10
10 -12
Integrity class I3
10 -14
10 -16
10 -5
10 -4
2004 June, HK
10 -3
22
10 -2
10 -1
10 0
Bit error rate
Synchronization Errors
HDLC-frame
flag
flag
01111110
any data
disturbance
HDLC-frame with error
FCS
01111110
Frame Check Sequence

(CRC)
flag
flag
"FCS"
01111110
1 Chance in 65536,
that the random data
form a correct CRC
FCS
01111110
false
flag
01111110
discarded
precisely 1111110 is the most frequent sequence in a

random bit stream because of bit-stuffing.
A single error can falsify a frame -> HD = 1

It is uninteresting how likely this case is, the fact is, that it can occur.
Because of this bug, HDLC when used in industry require additional error checks.
The synchronization should have a higher Hamming distance than the data itself.
2004 June, HK
23
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
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.
2004 June, HK
25
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
26
Medium Access Control

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

2004 June, HK
27
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
28
Medium Access Control - quality criteria
Fairness
all requesters will eventually be allowed to transmit
Determinism
all requesters will be allowed to transmit within a finite time
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.
2004 June, HK
29
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
30
MAC single master (e.g. Profibus DP)
slave
slave
slave
bus
master
devices
(slaves)
bus
command reply
read
command ack
write with ack
command
write no ack
command
reply
read & write
time
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
- the master has little knowledge about data importance
improvement: look-at-me (short poll frame allowing slave to request poll of a longer frame)
= "slave initiative" used in Profibus DP
2004 June, HK
31
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
32
MAC Rings (1): register insertion principle
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.
data
time
output
application memory
devices
input
master
output
shift register
input
data
output
input
time
2004 June, HK
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 s per device)
2004 June, HK
34
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
35
MAC Ethernet (1): CSMA-CD principle (stochastic)
Principle
Every station sends as it pleases

if no acknowledgement comes, it retransmit
improvement 1: do not send when the medium is occupied

improvement 2: be aware that you are jammed
improvement 3: retry after a random time,
doubled after each collision, max 15 times
(pure Aloha)
(Carrier Sense)
(Collision Detection)
(Binary Backoff)
Advantage:
Arbitration does not depend on number or on address of the stations
Drawback:
No upper limit to the waiting time, mean waiting time depends

on the arrival rate of frames and on their average length.
The medium access is not deterministic,
but for light traffic (<1%) there is no noticeable delay.
2004 June, HK
36
MAC Ethernet (2): collision conditions

repeater
preamble
= 8 Bytes
(2 x 7.5 s/km+ 2
repeaters)
Tpd
minimum
frame size
= 64 Bytes
collision
Station A must
detect that its
frame collided
while it is still
transmitting its
frame
time
Station B started
transmission just
before receiving
As frame.
it nevertheless
transmits its
header completely
Ethernet minimum frame size = 64 Bytes, or 51,2 s @ 10 Mbit/s

at 10 Mbit/s, limits radius to about 2500 m
at 100Mbit/s, limits radius to about 250 m
2004 June, HK
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.
2 x Tpd = 2 x propagation time (@10Mbit/s Ethernet: 51,2 s)
500 m
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.
2004 June, HK
38
MAC Ethernet (4): collision probability and simultaneous transmitters

After the end of a frame, a transmitter chooses a slot at random from a fixed number of slots
0
previous frame
utilisation 100%
time
frame size
0.8
4096
1024
0.6
512
48
0.4
0.2
0
1
10
32
64
128
256
number of transmitters
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
2004 June, HK
39
MAC Ethernet (5): when collisions can't be detected

It is not always possible to detect collisions.
LON uses a p-persistent MAC with 16-slot collision avoidance (p = 1/16)
1
probability of loosing a packet
source: P. Koopman, CMU
0.9
0.8
0.7
without collision detection
0.6
0.5
0.4
0.3
with collision detection
0.2
0.1
0
number of simultaneous transmitters

1
10
11
12
13
14
15
16
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
2004 June, HK
40
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
41
MAC CAN (1): Deterministic Arbitration
The CAN fieldbus uses media with a dominant and a recessive state
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).
Such a medium allows a bit-wise "Wired-OR" operation

Example:
open-collector:
Terminator and
pull-up resistor
device 1
Out
Jam
device 2
In
Out
XOR
Jam
device 3
In
XOR
Out
Jam
XOR
device 4
In
Out
Jam
XOR
In
Terminator and
pull-up resistor
Ut
Ut
Rt
Rt
Bus line
A device is capable to know if the signal it puts on the line is disturbed (XOR).
2004 June, HK
42
MAC CAN (2): Collision with Winner
Also known as "Urn" or "binary bisection" method

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.
slot time
Station 10 (wins)
Station 06
(1)
(1)
(0)
Station 09
(1)
Bus signal
Advantage: deterministic arbitration (assumes fairness), good behavior under load

Drawback: the size of the unique ID defines arbitration time,
transmission delay defines slot time -> 40m @ 1 Mbit/s, 400m @ 100 kbit/s
2004 June, HK
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.
Drawback: 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.
2004 June, HK
44
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
45
MAC Profibus (1): Token principle

z.B.: Token Bus (IEEE 803.4), Profibus (IEC 61158)
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).
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 !
2004 June, HK
46
MAC Profibus (2): Token passing

logical ring
station address
12
25
31
AS
AS
AS
current bus
master
active stations
(potential masters)
passive stations
(slaves)
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.
2004 June, HK
47
MAC Profibus (3): Token passing algorithm
Each station holds a List of Active Stations (LAS)

GAP
1
AS AS
Active
Station
PS
Previous
Station
TS
10 11
NS
This
Station
12 13 14 15
..
30 32
AS
Next
Overnext
Station Station
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.
2004 June, HK
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.
When a master checks the gap, the station will receive a token offer.
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.
2004 June, HK
49
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
50
Comparison of Medium Access Control Methods
deterministic
stochastic
pessimistic
optimistic
carrier sense
central master
collision detection
collision with winner
p-persistent collision
2004 June, HK
token passing
51
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
52
Logical Link Control Sublayer

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

- connection oriented services
2004 June, HK
53
Connection-Oriented and Connection-Less communication

these considerations apply to all levels of the OSI model
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.
Semantic of CL-transmission
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);
Send_Packet (source, destination, Packet1);

Send_Packet (source, destination, Packet2);
Msg1 will be received before Msg2, sequence is maintained.
2004 June, HK
54
Connection-Oriented Link Service
Task: Flow Control and Error Recovery

Connection establishment and disconnection
Send and receive frames
Flow Control (Buffer control)
Retry in case of error
service
user
service
provider
service
user
REQUEST
INDICATION
RESPONSE
CONFIRM
2004 June, HK
55
Flow Control
( = synchronization at the link layer)
"Adapt the speed of the sender to that of the receiver"
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)
Lower Window
Edge
packets
Upper Window
Edge
8
10
11
12
time
sent and
acknowledged
sent but not yet

acknowledged
2004 June, HK
56
can be sent
may not
be sent
Simple transfer with window size = 1

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

2004 June, HK
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.
2) In event-driven transmission, no information may be lost, a repetition is explicit:
a) The sender of information frames expects acknowledgement of the
receiver, indicating which information it received.
b) The sender repeats the missing information a number of times, until it
receives an acknowledgement or until a time-out expires.
c) To distinguish repetitions from new information, each packet receives
a sequence number (in the minimum odd/even).
d) The receiver acknowledges repetitions even if it already received the
information correctly.
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.
2004 June, HK
58
Link Layer Outline

Stacks
HDLC as example
Frame sub-layer
Error detection
Error correction
Quality Criteria
Single Master
Rings
Ethernet
Token Passing
Comparison
Error recovery
Flow control
HDLC
2004 June, HK
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
2004 June, HK
60
HDLC Control Field (ISO 4335)

flag
01111110
flag
adr
control
FCS
any data
01111110
16
physical address of Secondary

(for command and response)
Control Field Bits
1
Control Field Format for:
N(S) = Sequence number of

sender
Information Transfer
Command/Response
(I-Format PDU)
Supervisory
Commands/Responses
(S-Format PDUs)
Unnumbered
Commands/Response
(U-Format PDUs)
N(S)
P/F
N(R)
N(R) = Sequence number of
receiver
P/F
N(R)
S = Supervisory
P/F
P/F = Poll/Final (Please respond/Final in sequence)
2004 June, HK
61
HDLC Connection Types

Traffic is divided into packets (= information frame) each receiving a sequence
number (Modulo 8).
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.
HDLC provides different connection types:
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).
2004 June, HK
62
HDLC Exchange (NMR in ISO 4335)
Send Sequence
set normal response
mode, poll
please confirm
information packets
SNRM, P
I0,0
I2,0P
I1,0
I3,0
Primary
(Commander)
time
UA,F
Secondary
(Responder)
RR3,F
accept
connection,
final
receiver
ready, expects 3
Receive Sequence
set normal response
mode, poll
SNRM; P
receiver
ready, expects 0
receiver
ready, expects 3
RR0;P
RR3;P
Primary
(Commander)
Secondary
(Responder)
time
UA;F
I0,0
accept
connection, final
I1,0
several
information packets
I2,0;F
I3,0
last frame
The data transmission takes place under control of the Primary.

Therefore, both "Send Frame" and "Receive Frame" are supported
2004 June, HK
63
Clocks
In a fieldbus, devices must be synchronized to a common clock to time-stamp

their transmissions.
2004 June, HK
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.
int
req
int
req
PORTS
slave
clock
master
clock
application
processor 1
pulse
application
processor 2
int
req
PORTS
pulse
bus
controller
application
processor 3
PORTS
pulse
bus
controller
bus
controller
BUS
time variable
bus master
2004 June, HK
65
Clock compensation for transmission delays
other master
with clock
slave
clock
master
1
MVB 1
other MVB
synchronizer
master clock
device
slave clocks
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 s across several bus segments.
2004 June, HK
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.
2004 June, HK
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.
2004 June, HK
68
PersonnelRecord ::= [APPLICATION 0]

SET {
name
Name,
title
[0] VisibleString,
number
EmployeeNumber,
dateOfHire [1] Date,
nameOfSpouse [2] Name,
children
[3] IMPLICIT
SEQUENCE OF ChildInformation
DEFAULT {}
3.3.4 Data presentation
2006, April HK
Prof. Dr. H. Kirrmann

ABB Research Center, Baden, Switzerland
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
2/48
3.3.4 Data Presentation
Presentation layer in the OSI-Model (ISO/IEC standard 7498)

(Mail, File Transfer,...)
Application
Definition and conversion of the

data formats (e.g. ASN 1)
Presentation
Session
Transport
Network
Link
Physical
Application
protocols
(e.g. ISO 8326)
(e.g. TP4, TCP)
Transport
protocols

(e.g. IP, X25)

medium access (e.g. 802.2, HDLC)
mechanical coupling (e.g. Ethernet)
3/48
Presentation Layer
The presentation layer is responsible that all communication

partners agree on the format of the data
4/48
Justification for a data representation standard
Why do we need ASN.1 ( or a similar notation)
5/48
Example: Transfer Syntax: describe what is in a frame ?

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

(normal format)
DSP
(domain
specific
part)
NSAP = 00
Semantic Levels of Data Representation

Semantic format
Application Language format
154,5 Vrms
struct {
unsigned count;
int
item2:8;
int
item1:4;
int
dummy:4;
}
application memory
means actually 154,5 kV on the line
application
processor
D15
D00
Application Storage format

(Assembly language)
Storage format
traffic memory
Parallel Bus Format

(8 bit, 16 bit, 32 bits)
Traffic Memory format
transmission
specification
language and
encoding rules
SEQUENCE {
item1 INTEGER16,
item2 INTEGER4,
count UNSIGNED8;
}
bus controller
(16-bit oriented)
Bus Transmission Format
time
7/48
Bit Transmission Order

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
first
first
0 1 0 0 0 0 0 0
LSB
MSB
0 0 0 0 0 0 1 0
MSB
LSB
time
FDDI, FIP, CAN, MVB
UART, HDLC, Ethernet
Legacy of the old telex
There is no relation between the bit and the byte ordering scheme
octet in
octet out
Who says which bit is really transmitted first ?
(FDDI: multi-bit symbol transmission, multi-bit
transmission with interleaving)
Convention: only consider octet streams

8/48
Integer representation in memory: Big-Endian vs Little Endian
Memory contents
MSB
address
LSB
0 0 0 0 0 0 1 0
x0001
x0002
0 0 0 0 0 0 0 1
x0003
x0004
0 0 0 0 0 0 0 1
x0005
x0006
0 0 0 0 0 0 0 0
BE
Intel
Motorola
DEC
7 6 5 4 3 2 1 0
x0000
LE
IBM, TCP/IP,
Unix, RISC
INTEGER8
=2
=2
INTEGER16
=1
= 256
INTEGER32
=1
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
= 16777216
0 0 0 0 0 0 0 0
Processor
B7 B6 B5 B4 B3 B2 B1 B0
(bit name)
TCP/IP
0 1 2 3 4 5 6 7
(bit offset)
Profibus
1 2 3 4 5 6 7 8
(bit position)
9/48
three different naming schemes !
How are sequence of octets transmitted ?
Most Significant First ?
Transmission Order on the bus ?
INTEGER32
time
how to name bits ?
The standard in network protocols is always:

Most Significant Octet first (Big-Endian)
10/48
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
htons
htonl
ntohs
ntohl
Description
Host order to
Host order to
Network order
Network order
network
network
to host
to host
order,
order,
order,
order,
short (2 bytes, 16 bits)

long (4 bytes, 32 bits)
short (2 bytes, 16 bits)
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

11/48
Inconsistency: Token-bus frame (IEEE 802.4)

first MAC
symbol
read this picture from left to right,

then top to bottom
preamble
N
N
F
F
I/G L/U
0
M
lsb
N
M
N
M
0
P
0
P
0
P
Destination Address
remaining 44 bits of address

msb
I/G L/U
Addresses are transmitted least significant bit first
lsb
Source Address
remaining 44 bits of address
msb
msb
lsb
first octet
Data are transmitted least significant bit first within

an octet, but most significant octet first within a
word (imposed by higher layers)
lsb
msb
msb
last data octet
Frame Check Sequence
Checksum is transmitted most significant bit first

I
lsb
E
12/48
How to specify transmitted data ?

offset 0
offset
time
first: agree on a bit and byte ordering

scheme: {0,0} scheme recommended.
second: describe the data stream

formally (machine-readable)
formal
intuitive
0
snu
gni
nodeID or groupID
stationID
functionID
protocolID
command
parameter
i
i+1
CommandPDU :== SEQUENCE

{
BOOLEAN1
snu -- system
BOOLEAN1
gni -- group
CHOICE (gni)
{
Group
: ENUM6 groupID;
Individual: ENUM6 nodeID;
}
ENUM8
stationID;
...
13/48
Can a C-declaration serve as an encoding syntax ?
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;
how is size given?

allowed values in enum ?
to which structure does it point?
is "short" a byte ?
is unsigned 16 bits or 32 bits ?
No!
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.
14/48
One solution: 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 was originally designed to transfer information between data bases.
ASN.1 has the same role as XML, but it is far more efficient.
see: http://asn1.elibel.tm.fr/en/index.htm
15/48
ASN.1 Syntax Example

Informal
Name:
Title:
Employee Number:
Date of Hire:
Name of Spouse:
Number of
Children:
John P Smith
Director
51
17 September 1971
Mary T Smith
2
Child Information Ralph T Smith

Name:
11 November 1957
Date of Birth
Child Information Susan B Jones
Name:
17 July 1959
Date of Birth
ASN.1
PersonRecord ::=
name
title
number
dateOfHire
nameOfSpouse
children
[APPLICATION 0] SEQUENCE {
Name
[0] VisibleString,
EmployeeNumber,
[1] Date,
[2] Name,
[3] SEQUENCE OF
Childinformation
DEFAULT {} }
ChildInformation ::= SEQUENCE {

name
Name,
dateOfBirth
[0] Date}
Name ::= [APPLICATION 1] SEQUENCE {
givenName VisibleString,
initial
VisibleString,
familyName VisibleString}
EmployeeNumber ::= [APPLICATION 2] INTEGER

Date ::= [APPLICATION 3] VisibleString -- YYYYMMDD
16/48
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 (current version is 2002).
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 persons name.
The basic encoding rules for ASN.1 are defined in the standard ISO 8825-1.
17/48
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)
tag
length
value
The value may be itself a structured object:

tag
length
tag
length
value
tag
length
value
(recursive)
the tag specifies the data type of the value that follows, implicitly or explicitly.
18/48
ASN.1 Data Types
UNIVERSAL
APPLICATION
CONTEXT_SPECIFIC
PRIVATE
four Tag Types
Basic Types
BOOLEAN
INTEGER
BITSTRING
OCTETSTRING
NULL
OBJECT_ID
OBJECT_DESC
EXTERNAL
REAL
ENUMERATED
ANY
Constructed Types
SEQUENCE
ordered sequence of types (record, struct)
SEQUENCE OF ordered sequence of same type (array)

CHOICE
one of an unordered, fixed sequence of

different types.
SET
SET OF
unused
unused
pointers are not used objects of undefined length (closing character) are treated specially
19/48
ASN.1 Universal Types

1 Boolean type
2 Integer type
3 Bitstring type
4 Octetstring type
5 Null type
6 Object identifier type
7 Object descriptor type
8 External type and Instance-of type
9 Real type
10 Enumerated type
11 Embedded-pdv type (imported type)
12 UTF8String type (unicode)
13 Relative object identifier type
14-15 Reserved for future editions
16 Sequence and Sequence-of types
17 Set and Set-of types
18-22, 25-30 Character string types
23-24 Time types
31-... Reserved for addenda
0 Reserved for use by the encoding rules
20/48
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
PersonRecord
chief
VisibleString
title
number
INTEGER
21/48
dateOfHire
UniversalTime
Tagging
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) is error-prone.
Structured types need in any case a tag, otherwise they cannot be distinguished
from another structured type (note how PersonRecord is identified as A1)
PersonRecord ::= [APPLICATION 1] SEQUENCE {
person
[1] VisibleString,
chief
[2] VisibleString,
title
[3] VisibleString,
number
[4] INTEGER,
dateOfHire
[5] UniversalTime }
person
tags
A1
[1]
number
[4]
dateOfHire
title
[3]
22/48
[5]
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
[1] millimetres
[2] kilograms
}
INTEGER16,
INTEGER32,
FLOAT
1234
quantity in units
1234
quantity in millimetres
1.2 E 03
kilograms, IEEE format
23/48
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.
24/48
Automatic, 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.
The most recent ASN.1 standard assumes that tags are implicit
25/48
Transfer Syntax: Basic Encoding Rules (BER)
BER (ISO 8825-1), a companion to ASN1, 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
Example:
type = UNIVERSAL Integer

length = 2 octets
value = 1234
02
02
12
(tag included in basic types)
34
BER supports all ASN.1 structures.

Exception: if size is 0, there is no value field (== NULL)
26/48
Class
00 = UNIVERSAL
01 = APPLICATION
10 = CONTEXT_SPECIFIC
11 = PRIVATE
Primitive {0} or
Constructed {1}
BER Tag-Type field
P: after the size comes a value C: after the size comes a tag
Example:
Tag
(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.
00000010 = UNIVERSAL INTEGER

10100001 = [CONTEXT_SPECIFIC 1] SEQUENCE
27/48
Basic Types in BER
A Boolean is represented by one octet (all zero for false)
01 01 FF = true
An Integer is represented by as many octets as necessary for the representation

of the number -> variable length encoding !
a:= 5 -> 1 octet
a:= 260 -> 2 octets
a:= 65544 -> 3 octets, etc
02 01 05
02 02 01 04
(hex representation)
02 03 01 00 08 (hex representation)
A Boolean Array is represented by as many octets as the array size requires.

The first octet indicates the number of unused bits in the trailing octet
e.g. BitString2 (Antivalent)
03 02 06 C0
bitstring
2 octets
6 unused bits
value = 11
28/48
Example ASN.1 and BER

8
AP C
CallerRef :== [APPLICATION 7] SEQUENCE {
priority
[2] INTEGER,
command
[0] INTEGER,
reference
[1] INTEGER,
caller
INTEGER}
(useful information)
29/48
7
16
CS P
1
high_prio
CS P
0
1
command
CS P
1
2
reference (MSB)
reference (LSB)
CS P
3
4
UN P
2
2
caller (MSB)
caller (LSB)
high_prio
command
reference
caller
Examples
constructed means: the octet following the size is not a value, but a type-tag !
Tag-Type:
80 10000000
Context Specific, not constructed, implicit, tag = [0]
61 01100001
Application Specific, constructed, implicit, tag = [1]
AB 10101101
Context Specific, constructed, implicit, tag = [11]
01 00000001
Basic Type, not constructed, boolean
decoding the first digit:

0,1:
2,3:
4,5:
6,7:
8,9:
A,B:
Universal, not constructed

Universal, constructed
Application Specific, not constructed
Application Specific, constructed
Context Specific, not constructed
Context Specific, constructed
30/48
[APPLICATION 1]
[6]
Examples
green: tag / type
red: size
black: value
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
31/48
null
Variable length values a tricky issue
values are transmitted with a variable length. This costs a lot of encoding and decoding that
is not justified for simple types. Example for INTEGER:
Value
Length field
1
127
128
255
32767
32768
65535
01
01
02
02
02
03
03
-1
-127
-128
-129
-255
-32767
-32768
-32769
01
01
01
02
02
02
02
03
Value Octets
00
00
7F
00 80
00 FF
01
7F
80
FF
FF
00
FF
FF
FF
80
80
FF 7F
FF
81
80
7F
01
01
00
FF
// positive numbers > 127 take one octet more
// positive numbers > 32767 take one octet more
// note how this differs from 255
// negative numbers < -128 take one octet more
// negative numbers < - 32768 take one more octet
32/48
Beyond ASN.1 and BER

ASN.1 / BER could not impose themselves in field busses because
of the high overhead involved (up to 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, in the worst case, compression takes as much place as the original)
33/48
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).
34/48
ROSIN - Retrofit encoding rules example

Type_InfoMessage ::= RECORD {
parameter1
INTEGER8
parameter2
INTEGER16,
parameter3
UNSIGNED6,
par4
ANTIVALENT2,
parameter5
Parameter5,
parameter6
STRING32,
--------
octet.
16-bit word, MSB first
6- bit value
2 bits for par4, e.g. check variable.
parameter5 has a structured type
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 FFH 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 (1EH)
-- one of two defined values
SECONDCASE (84H)
-- the other defined value
},
action_code
Action_Code,
-- this type is used several times
parameter1
parameter2
2
3
parameter3
par4
parameter5
5
6
parameter6
CHARACTER8
39 snu gni
sta_or_func
40
next_station_id
41
42
node_id
tv
d1
top_counter
43
tnm_code
44
action_code
35/48
Comparing Coding Efficiency

(ISO)
(SUN)
BER
XDR
Boolean
24
16
Integer8
24
Unsigned8
(UIT)
A-PER
(Profibus) (FIP)
U-PER
FER
BuER
ROSIN
16
16
16
12
16
16
24
16
11
11
16
16
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
encoding/decoding highly packed data may cost more than is won by shorter transmission
36/48
The semantic limit

The good news: it is possible to decode entirely the element of a BER message
(tags, length, value) provided the start or entry point is known.
The bad news: one cannot understand the values unless the receiver knows the
original ASN.1 definitions.
Even worse news: even knowing the ASN. 1 original syntax does not allow to understand
the semantics of the value. If a value of 1234 represents a variable called errorCode
or temperatureAmb, one does not yet know the physical value really it represents.
Therefore, the application must define in any case a higher-level type specification
in a human readable form.
example:
switchPos ::= SEQUENCE {
open
[1] BOOLEAN,
closed [2] BOOLEAN,
operations [3] INTEGER,
closingTime [4] INTEGER
}
-----------
true when switch is open (debouncing 2 ms)

true when switch is closed (debouncing 2 ms)
(both can be false during opening /closing)
number of operations since installation
time in milliseconds from open to close
37/48
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.
38/48
SI Units
All physical variables should be restricted to SI Units (NIST 330-1991, IEEE 268A-1974)
or referred directly to them
For instance:
speeds shall be expressed in meter/second, not in km/h or in miles/hour
angles shall be represented in radian rather than degree or grad.

variable
position, distance
mass
time
current
angle
force
torque
frequency
angular velocity
angular acceleration
unit
m
kg
s
A
rad
N
Nm
s-1
rad/s
-2
rad s
39/48
variable
power
energy
pressure
volume
flow
mass flow
tension
reactive power
impedance
temperature
unit
W
J
Pa
m3
m3 /s
kg/s
V
var
Why floating point ?

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.
Floating point format is the only safe representation of a physical variable.
Internally, devices can use other formats.
Therefore, devices shall indicate their exported and imported variables as REAL32.
Exception:
Variables, whose precision do not depend on their absolute value:
e.g. time, elapsed distance (odometer), energy, money, countable objects
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)
40/48
Fractionals
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
span value
physical variable
UNSIGNED16
0
65535
INTEGER16
-32768
+32767
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
41/48
Scaled Variables
Process Variables are often transmitted and processed as fractionals:

Fractional format expresses analog values as integer multiples of the resolution
e.g.:
caution
distance = 0..65535 x 0.1 m (resolution)

distance = -32768..+32767 x 0.1 m
speed = 0.. 6553,5 m/s, resolution = 0.1 m/s
speed = 0.. 6.5535 m/s, resolution = 0,0001 m/s
--> 0 .. 65535
--> -32768 .. 32768
--> 0 .. 65535
--> 0 .. 65535
(UNSIGNED16)
(INTEGER16)
(UNSIGNED16)
(UNSIGNED16)
The resolution is often a decimal fraction of a unit !
Some standards provides a bipolar or unipolar analog data format:

e.g. :
0%..+200-%
of 10 kV
-200%..+200-% of 10 kV
== 0 .. 65535
== -32768 .. 32768
(UNSIGNED16)
(INTEGER16)
fractionals require producer and consumer to agree on range and resolution:

e.g. resolution 0.5 V, range 6553.5 V
fractionals have some justifications in circular units

0 = 360
180
== 0
== 32768
fractionals are easy to process but error prone (overflow, underflow)

A conversion from fractional to floating point cost over 5 Mia (Ariane 501 accident)
42/48
Constructed Application Data Types

Circuit Breaker command:
Code
DoubleCommand
Persistent
Regulating Command
RegulatingStep
Command
Double-Point
Information
00
01
10
11
not permitted
OFF
ON
not permitted
not permitted
Lower
Higher
not permitted
not permitted
Next Step Lower
Next Step Higher
not permitted
indeterminate
determined OFF
determined OFF
not permitted
Time-Stamped Variable:
value
time
status
43/48
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.
44/48
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.
45/48
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.
gateway
PD-marshalling
MSG
application
Real-Time
Protocols
PV
application
presentation
presentation
session
session
transport
transport
network
network
PV
link
link
link
link
physical
physical
physical
physical
bus type 1
bus type 2
46/48
Prof. Dr. H. Kirrmann

Office
network
TCP - IP
Ethernet
Plant Network
Ethernet, ControlNet
Fieldbus
intelligent field devices
FF, PROFIBUS, MVB, LON
Sensor Busses
simple switches etc.
CAN, DeviceNet, SDS, ASI-bus, Interbus-S
3.1
2006 March, HK
Field Bus: principles

Bus de terrain: principes
Feldbusse: Grundlagen
Field bus: principles
3.1 Field bus principles

Classes
Physical layer
3.2 Field bus operation
Centralized - Decentralized
Cyclic and Event Driven Operation
3.3 Standard field busses
Location of the field bus in the plant hierarchy

File
Edit
SCADA level
Operator
23
2
4
33
12
2
Engineering
Plant bus
Programmable
Logic Controller
Plant Level
Field bus
Field level
Sensor/
Actor
Bus
direct I/O
Sensor /
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,..)
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 ones own developments (if based on a standard)
The original idea: save wiring

I/O
tray
marshalling
capacity
bar
dumb devices
PLC
(Rangierung,
PLC
COM
tableau de brassage (armoire de triage)
field bus
but: the number of end-points remains the same !

energy must be supplied to smart devices
Marshalling (Rangierschiene, Barre de rangement)
The marshalling is the interface between

the PLC people and the instrumentation
people.
Field busses classes
Office
network
TCP IP
Ethernet
Plant Network
Ethernet, ControlNet
Fieldbus
intelligent field devices
FF, PROFIBUS PA, LON
Sensor Busses
simple switches etc.
CAN, DeviceNet, SDS, ASI-bus, Interbus-S
The field bus depends on:

its function in the hierarchy
the distance it should cover
the data density it should gather
Geographical extension of industrial plants

The field bus suits the physical extension of the plant
1 km .. 1000 km
Transmission & Distribution

Control and supervision of large distribution networks:
water - gas - oil - electricity - ...
1 km .. 5 km
Power Generation
Out of primary energy sources:
waterfalls - coal - gas - oil - nuclear - solar - ...
50 m .. 3 km
Industrial Plants
Manufacturing and transformation plants:
cement works - steel works - food silos - printing - paper
pulp processing - glass plants - harbors - ...
500m .. 2 km
Building Automation
energy - air conditioning - fire - intrusion - repair - ...
1 m .. 1 km
Manufacturing
flexible manufacturing cells - robots
1 m .. 800 m
Vehicles
locomotives - trains - streetcars - trolley buses - vans buses - cars - airplanes - spacecraft - ...
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).
10
Process Industry Application: Water treatment plant

Control Room
Japan
R emote
Maintenanc e
Sy s tem
source: Kaneka, Japan

LAS
SCADA
Malaysia
Ethernet
Bus Monitor
H1 Speed Fieldbus
JB
Segment 1
Segment 3
Sub Station
JB
AO
AI
AI
P ID
AI
P ID
AI
AI
AO
AI
AI
AI
AI
PLC
AO
P ID AO
AO
M.C.C.
DI
FB Protoc ol
C onv erter
JB
Segment 2
Segment 4
D igital Input/Output
AI
AI
AI
JB
AI
AI
P ID AI P ID
AO
AO
AI AI
AI
AI
Numerous analog inputs (AI),

low speed (37 kbit/s) segments merged to 1 Mbit/s links.
11
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

Fast controllers require at least 16 Mbit/s over distances of 2 m
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.
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.
12
Distributed peripherals
Many field busses are just

extensions of the PLCs Inputs
and Outputs,
field devices are data
concentrators.
Devices are only visible to the
PLC that controls them
relays and fuses
13
Application: Building Automation

Source: Echelon
low cost, low data rate (78 kbit/s), may use power lines (10 kbit/s)
14
Application: Field bus in locomotives
power line
radio
cockpit
Train Bus
diagnosis
Vehicle Bus
brakes
power electronics
data rate
delay
medium
number of stations
integrity
cost
motors
track signals
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
very high (signaling tasks)
engineering costs dominate
15
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
16
Application: Avionics (Airbus 380)
17
Networking busses: Electricity Network Control: myriads of protocols
control
center
SCADA
control
center
IEC 870-6
Modicom
Inter-Control Center Protocol
IEC 870-5
DNP 3.0
RTU
COM
ICCP
RTU
Conitel
control
center
RP 570
RTU
HV
High
Voltage
serial links (telephone)
RTU
Remote Terminal Units
RTU
substation
substation
FSK, radio, DLC, cable, fiber,...
RTU
RTU
houses
MV
Medium
Voltage
LV
Low
Voltage
RTU
RTU
low speed, long distance communication, may use power lines or telephone modems.
Problem: diversity of protocols, data format, semantics...
18
The ultimate sensor bus
power switch and

bus interface
requires integration of power electronics and communication at very low cost.

19
3 Industrial Communication Systems
3.2
Field Bus Operation

Bus de terrain: mode de travail
Feldbus: Arbeitsweise

2006 March, HK
Fieldbus - Operation
3.1 Field bus types

Classes
Physical layer
Data distribution
Cyclic Operation
Event Driven Operation
Real-time communication model
Networking
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)
source
value
quality
time
description
Master or peer-to-peer communication
communication in a control system is evolving from hierarchical to distributed

central master: hierarchical
PLC
all traffic passes by the master (PLC);

adding an alternate master is difficult
(it must be both master and slave)
master
AP
slaves
input
peer-to-peer: distributed
PLC
PLCs may exchange data,

share inputs and outputs
allows redundancy
and distributed intelligence
devices talk directly to each other
alternate
master
PLC
AP
AP
input
slaves
PLC
AP
output
masters
PLC
AP
output
separate bus master from application master !

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
application
processor
application
processor
application
processor
plant
image
plant
image
plant
image
plant
image
=
distributed
data base
bus
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 bus refreshes the plant image in the background, it becomes an on-line database
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.
Each station snoops the bus and reads the variables it is interested in.
Data format
source
value
quality
time
In principle, the bus could transmit the process variable in clear text (even 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
(see: Presentation Layer)
Field busses are still slow (1Mbit/s ..12 Mbits/s) and therefore more compact
encodings are used.
Datasets
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
binary variables
analog variables
dataset
identifier
wheel
speed
0
air
pressure
16
bit offset
line
voltage
32
time
stamp
48
64 66
70
all door closed

lights on
heat on
air condition on
size
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:
Dataset
correct variable
error
undefined
variable value
check
0 1 0 1 1 1 0 0
0 1
0 0 0 0 0 0 0 0
0 0
1 1 1 1 1 1 1 1
1 1
00 = network error
01 = ok
10 = substituted
11 = data undefined
chk_offset
var_offset
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.
Decoupling Application and Bus traffic

coupled (event-driven):
decoupled (asynchronous):
application
processor
application
processor
events
(interrupts)
traffic
memory
queues
bus
controller
bus
controller
sending: application writes data into memory
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
receiving: application reads data from memory

the bus controller decides when to transmit
bus and application are not synchronized
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.
Application
Processor
Traffic Memory
Ports (holding a dataset)
Associative
memory
an associative memory decodes

the addresses of the subscribed
datasets
Bus
Controller
two pages ensure that read and

write can occur at the same time
(no semaphores !)
bus
10
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.
11
Process Variable Interface

Access of the application to variables in a traffic memory is very easy:
ap_put (variable_name, variable value)
ap_get (variable_name, variable value, variable_status, variable_freshness)

Rather than fetch and store individual variables, access is done by clusters
(predefined groups of variables):
ap_put_cluster (cluster_name)
ap_get (cluster_name)
The cluster is a table containing the names and values of several variables.
The clusters can correspond to "segments" in the function block programming.
Note: Usually, only one variable is allowed to raise an interrupt when received: the one
carrying the current time (sent by the common clock)
12
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.
t1
t2
processing
input
input
input
t3
t4
bus
t1 val1
13
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.
14
Cyclic versus Event-Driven transmission
cyclic: send value strictly every xx milliseconds
individual
period
misses the peak

(Shannon!)
always the same,

why transmit ?
time
hysteresis
event-driven: send when value change by more than x% of range
how much hysteresis ?

- coarse (bad accuracy)
- fine (high frequency)
limit update
frequency !,
limit hysteresis
15
nevertheless transmit:
- every xx as Im alive sign
- when data is internally updated
- upon quality change (failure)
Fieldbus: Cyclic Operation mode
3.1 Field bus types

Classes
Physical layer
Data distribution
Cyclic Operation
Networking
16
Cyclic Data Transmission

address
Bus
Master
devices
(slaves)
Poll
List
plant
The master polls the addresses in a fixed sequence, according to its poll list.
Individual period
1
Individual period
6
2 3
N polls
6
2 3
6
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 delay of the signals
(independent of bit-rate).
2 x Tpd
address
10 s/km
(i)
data
(i)
read transfer
address
(i+1)
time [s]
44 s .. 296 s
17
Cyclic operation principle

Data are transmitted at fixed intervals, whether they changed or not.
The delivery delay (refresh rate) is deterministic and constant.
The bus is under control of a central master (or distributed time-triggered algorithm).
No explicit error recovery needed since a fresh value will be transmitted in the next cycle.
Only states may be transmitted, not state changes.
Cycle time is limited by the product of the number of data transmitted by the
duration of each poll (e.g. 100 s / point x 100 points => 10 ms)
To keep a low poll time, only small data items may be transmitted (< 256 bits)
The bus capacity must be configured beforehand.
Determinism gets lost if the cycles are modified at run-time.
Cyclic operation is used to transmit the state variables of the process.
These are called Process Data (or Periodic Data)
18
Source-Addressed Broadcast
Process Data are transmitted by source-addressed broadcast.
Phase1: The bus master broadcasts the identifier of a variable to be transmitted:
bus
master
device
device
sink
source
subscribed devices
sink
devices
(slaves)
sink
bus
variable identifier
Phase 2: The device that sources that variable responds with a slave frame
containing the value, all devices subscribed as sink receive that frame.
bus.
master
subscribed
device
subscribed
device
sink
source
subscribed devices
sink
sink
devices
(slaves)
bus
variable value
19
Read And Write Transfers

master
read transfer:
arb
source
address
turn-around
time
next transfer
data
arb
address
time
Read Transfer
write transfer:
master (source)
arb
address
next transfer
data
arb
address
Write-No ack transfer
time
turn-around time may be

large compared with
data transfer time.
destination
master (source)
arb
address
data
turn-around
time
ack
next transfer
arb
address
time
Write Transfer With Ack
Local Area Networks operate with write-only transfers.
Their link layer or transport layer provides acknowledgements by another write-only transfer
Parallel busses use read and write-ack transfers

Most field busses operate with read cycles only.
20
Round-tip Delay
master
closest data sink

repeater
T_m
remotest data source

repeater
t_repeat
t_repeat
The
round-trip
delay limits
the extension
of a read-only
bus
propagation delay
(t_pd = 6 s/km)
T_m
t_source
t_mm
t_ms
(t_repeat < 3 s)
access delay
T_s
t_repeat
t_sm
T_m
distance
21
Optimizing Cyclic Operation

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
Solution: introduce sub-cycles for less urgent periodic variables:
2 ms period
4a
16
4 ms period
1 4b
64
4a
time
1 ms period
(basic period)
1 ms
1 ms
group with
period 1 ms
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
22
Cyclic Transmission and Application
cyclic
poll
cyclic
algorithms
cyclic
algorithms
cyclic
algorithms
cyclic
algorithms
application
1
application
2
application
3
application
4
bus
master
Periodic
List
source
port
Traffic
Memory
Ports
Ports
Ports
sink
port
bus
controller
bus
controller
Ports
sink
port
bus
controller
bus
controller
bus
controller
bus
port address
port data
The bus traffic and the application cycles are asynchronous to each other.
The bus master scans the identifiers at its own pace.
Bus and applications are decoupled by a shared memory, the traffic memory,
which acts as distributed database actualized by the network.
23
Application Of Cyclic Bus
The principle of cyclic operation, combined with source-addressed

broadcast, has been adopted by most modern field busses
It is currently used for power plant control, rail vehicles, aircrafts, etc...
This method gives the network a deterministic behavior, at expenses of a reduced
bandwidth and geographical extension.
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.
24
Example: delay requirement

publisher
application instance
device
subscribers application instances

device
device
applications
bus
bus instance
Worst-case delay for transmitting all time critical variables is the sum of:
Source application cycle time
8 ms
Individual period of the variable
16 ms
Sink application cycle time
8 ms
= 32 ms
25
Example: traffic pattern in a locomotive

% periodic time
number of devices: 37 ( including 2 bus administrators)
30 frames of 128 bits
37 of 16 bits
18 of 32
occupancy is proportional to surface

total = 92%
period
16 ms
32 ms
26
64 ms
128 256
1024
Fieldbus: Event-driven operation
3.1 Field bus types

Classes
Physical layer
Data distribution
Cyclic Operation
Networking
27
Event-driven Operation
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).
Multi-master bus: uses write-only transfers
intelligent
stations
eventreporting
station
eventreporting
station
eventreporting
station
sensors/
actors
plant
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.
28
Bus interface for event-driven operation

application
filter
driver
Application
Processor
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.
message (circular) queues
interrupt
Bus
Controller
bus
29
Response of Event-driven operation

Caller
Application
Transport
software
Bus
Transport
software
Called
Application
request
interrupt
indication
confirm
time
Since events can occur anytime on any device, stations communicate by

spontaneous transmission, leading to possible collisions
Interruption of server device at any instant can disrupt a time-critical task.
Buffering of events cause unbound delays
Gateways introduce additional uncertainties
30
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.
Deterministic
Non-deterministic
Central master,
Token-passing (round-robin),
Binary bisection,
Collision with winner.
Collision
(Ethernet)
31
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.
events
producers
& consumers
input and
output queues
bus
acknowledgements
data packets
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
32
Events Pros and Cons

In an event-driven control system, there is only a transmission or an operation
when an event occurs.
Advantages:
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)
Drawbacks:
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
33
Fieldbus: real-time communication model
3.1 Field bus types

Classes
Physical layer
Cyclic Operation
Networking
34
Mixed Data Traffic
Process Data
represent the state of the plant
Message Data
represent state changes of the plant
short and urgent data items
infrequent, sometimes lengthy

messages reporting events, for:
Users: set points, diagnostics, status
System: initialisation, down-loading, ...
... motor current, axle speed, operator's

commands, emergency stops,...
-> Periodic Transmission

of Process Variables
-> Sporadic Transmission of

Process Variables and Messages
Since variables are refreshed periodically,

no retransmission protocol is needed to
recover from transmission error.
Since messages represent state

changes, a protocol must recover lost data in
case of transmission errors
basic period
basic period
event
time
sporadic
phase
periodic
phase
sporadic
phase
35
periodic
phase
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.
36
Real-Time communication stack

The real-time communication model uses two stacks, one for time-critical variables
and one for messages
time-critical
process variables
time-benign
messages
Management
Interface
7
Application
implicit
Presentation
Remote Procedure Call
Session
connection-oriented
Transport (connection-oriented)
Network (connectionless)
2"
medium access
2'
Link (Medium Access)
media
Physical
implicit
Logical Link
Control
connectionless
connectionless
common
37
Application Sight Of Communication
Periodic Tasks
R1
R2
R3
Event-driven Tasks
R4
E1
Traffic
Memory
(Broadcast)
station
E3
Message Services
Variables Services
Process Data
E2
Queues
Supervisory
Data
Message Data
(destination-oriented)
bus controller
bus
38
Field - and Process bus
Fieldbus
Process Bus
strictly deterministic
non - deterministic
controlled by a central master

(redundant for availability)
multi-master bus (Arbitration)

deterministic arbitration -> Token
cyclic polling
event-driven
number of participants limited by

maximum period
large number of participants
call/reply in one bus transfer

(read-cycle)
("fetch principle")
call/reply uses two different messages.

both parties must become bus master
("bring - principle")
cheap connection (dumb)
costly connection (intelligent)
only possible over a limited

geographical extension
also suited for open systems
39
Cyclic or Event-driven Operation For Real-time ?

The operation mode of the communication exposes the main approach to
conciliate real-time constrains and efficiency in a control systems.
cyclic operation
event-driven operation
Data are transmitted at fixed intervals,

whether they changed or not.
Data are only transmitted when they

change or upon explicit demand.
Deterministic: delivery time is bound
Non-deterministic: delivery time vary widely
Worst Case is normal case
Typical Case works most of the time
All resources are pre-allocated
Best use of resources
(periodic, round-robin)
(aperiodic, demand-driven, sporadic)
object-oriented bus
message-oriented bus
Fieldbus Foundation, MVB, FIP, ..
Profibus, CAN, LON, ARCnet
40
Fieldbus: Networking
3.1 Field bus types

Classes
Physical layer
Data distribution
Cyclic Operation
Networking
41
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.
42
Networking: Printing machine (1)

MPS
Production
Plant-bus (Ethernet)
Operator bus (Ethernet)
Console,
Section Supervision
LS LS LS PM
LS LS LS PM
LS LS LS PM
LS LS LS PM
Printing Towers
Section Busses (AF100)
Section Control
D
SSB
SSC
SSD
SSE
Line bus (AF100)

Reelstand-Gateways
RPB
RPC
RPD
RPE
Reelstand bus (Arcnet)

Reelstands
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
43
Networking: Printing Section (2)

to production preparation
(Ethernet)
bridge
standby
LS
LS
LS
PM
PM
standby
GW
GW
Pressmasterbus (Ethernet)
V-Sercos
IBG
V-Sercos
IBG
Section bus C
V-Sercos
IBG
V-Sercos
IBG
H-Sercos
H -steuerungen
IBG
IBG
IBG
Interbus
Interbus
Interbus
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
KT94
Service-Arcnet
Tower-ARCnet
Section bus B
Section bus C
Line bus
Turmsteuerung
AC160
Turmsteuerung
AC160
Turmsteuerung
AC160
Falz- und
Wendeturm- AC160
steuerung
Section bus D
Rollenwechslerkoppler I KT94
Oxydry-Arcnet
MR93
Sektions- SektionsAC160
steuerung steuerung
Rollenwechslerkoppler A
Interbus-S
KT94
ODC
Oxydry
Auro
ARCnet
44
Transmission delay over a Trunk Bus (cyclic bus)

Trunk Bus
copying,
filtering &
marshalling
delay
gateway
gateway
Feeder Bus
copying,
filtering &
marshalling
delay
Feeder Bus
stop
speed
speed
stop
The actual delay is non-deterministic, but bounded

The worst-case delay for the transmission of all variables is the sum of 5 delays:
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
45
= 100 ms
2006 March, HK

3.3
2005 April, HK
Field bus: standards

Bus de terrain standard
Standard-Feldbusse
Field busses: Standard field busses
3.1 Field bus types

Classes
Physical layer
Networking
Cyclic and Event Driven Operation
3.3 Field bus standards
International standard(s)
HART
ASI
Interbus-S
CAN
Profibus
LON
Ethernet
Automotive Busses
3.3 Standard Field Busses
Which field bus ?

A-bus
Arcnet
Arinc 625
* ASI
Batibus
Bitbus
* CAN
ControlNet
DeviceNet
DIN V 43322
DIN 66348 (Mebus)
FAIS
EIB
Ethernet
Factor
Fieldbus Foundation
FIP
Hart
IEC 61158
IEEE 1118 (Bitbus)

Instabus
* Interbus-S
ISA SP50
IsiBus
IHS
ISP
J-1708
J-1850
LAC
* LON
MAP
Master FB
MB90
MIL 1553
MODBUS
* MVB
P13/42
P14
Partnerbus
P-net
* Profibus-FMS
Profibus-PA
Profibus-DP
PDV
* SERCOS
SDS
Sigma-i
Sinec H1
Sinec L1
Spabus
Suconet
VAN
WorldFIP
ZB10
...
Worldwide most popular field busses

Bus
User*
Application
Sponsor
CANs
25%
Automotive, Process control CiA, OVDA, Honeywell
Profibus (3 kinds)
26%
Process control
Siemens, ABB
Building systems
Echelon, ABB
LON
6%
Ethernet
50%
Plant bus
Interbus-S
7%
Manufacturing
Fieldbus Foundation, HART
7%
Chemical Industry
ASI
9%
Building Systems
Modbus
22%
obsolete point-to-point
ControlNet
14%
plant bus
*source: ISA, Jim Pinto (1999)
all
Phoenix Contact
Fisher-Rosemount, ABB
Siemens
many
Rockwell
Sum > 100%, since firms support more than one bus
European market in 2002: 199 Mio , 16.6 % increase (Profibus: 1/3 market share)
**source: Elektronik, Heft 7 2002
Different classes of field busses

One bus type cannot serve
all applications and all device types efficiently...
Data Networks
Workstations, robots, PCs
Higher cost
Not bus powered
Long messages (e-mail, files)
Not intrinsically safe
Coax cable, fiber
Max distance miles
10,000
1000
frame size
(bytes)
100
Sensor Bus
Simple devices
Low cost
Bus powered (?)
Short messages (bits)
Fixed configuration
Twisted pair
Max distance 500m
Honeywell
PV 6000
SP 6000
AUTO
1
High Speed Fieldbus

PLC, DCS, remote I/O,
motors
$$ Medium cost
Non bus powered
Messages: values, status
Shielded twisted pair
Max distance 800m
10
10
100
Low Speed Fieldbus

Process instruments, valves
Medium cost
Bus-powered (2 wire)
Messages: values, status
Intrinsically safe
Twisted pair (reuse 4-20 mA)
Max distance 1200m
1000
poll time, milliseconds

10,000
source: ABB
Field device: example differential pressure transducer

4..20 mA current loop
fluid
The device transmits its value by means of a current loop

4-20 mA loop - the conventional, analog standard (recall)
The 4-20 mA is the most common analog transmission standard in industry
sensor
flow
transducer
i(t) = f(v)
RL1
reader
reader
R1
i(t) = 0, 4..20 mA
RL2
RL4
R2
voltage
source
10V..24V
RL3
R3
RL4
conductor resistance
The transducer limits the current to a value between 4 mA and 20 mA,

proportional to the measured value, while 0 mA signals an error (wire break)
The voltage drop along the cable and the number of readers induces no error.
Simple devices are powered directly by the residual current (4mA), allowing to
transmit signal and power through a single pair of wires.
Remember: 4-20mA is basically a point-to-point communication (one source)
3.3.2 HART
Data over 4..20 mA loops
2005 April, HK
HART - Principle
HART (Highway Addressable Remote Transducer) was developed by Fisher-Rosemount to
retrofit 4-to-20mA current loop transducers with digital data communication.
HART modulates the 4-20mA
current with a low-level
frequency-shift-keyed (FSK)
sine-wave signal, without
affecting the average analogue
signal.
HART uses low frequencies
(1200Hz and 2200 Hz) to deal
with poor cabling, its rate is
1200 Bd - but sufficient.
HART uses Bell 202 modem
technology, ADSL technology
was not available in 1989, at
the time HART was designed
Transmission of device characteristics is normally not real-time critical
HART - Protocol
Hart communicates point-to-point, under the control of a master, e.g. a hand-held device
Master
Slave
Indication
Request
time-out
Response
Confirmation
Hart frame format (character-oriented):

preamble
start
address
5..20
(xFF)
1..5
command bytecount
1
10
[status]
data
data
[2]
0..25
(slave response) (recommended)
checksum
1
HART - Commands
Universal commands (mandatory):

identification,
primary measured variable and unit (floating point format)
loop current value (%) = same info as current loop
read current and up to four predefined process variables
write short polling address
sensor serial number
instrument manufacturer, model, tag, serial number, descriptor,
range limits,
Common practice (optional)

time constants, range,
EEPROM control, diagnostics,
total: 44 standard commands, plus user-defined commands
Transducer-specific (user-defined)
calibration data,
trimming,
11
HART - Importance
Practically all 4..20mA devices come equipped with HART today
About 40 Mio devices are sold per year.
more info:
http://www.hartcomm.org/
http://www.thehartbook.com/default.asp
12
3.3.3 ASI
Small installation bus
2005 April, HK
ASI (1) - Sensor bus Wiring

ASI = Actor-Sensor Interface
Very simple sensor bus for building automation, combining power and data on the same
wires, transmitting mostly binary signals
mechanically coded flat cable

- two wires for data and power
insulation piercing connectors
- simple & safe
- protection class up to IP67,
even after disconnecting
D0 = sensor 1
one connection
D1 = sensor 2
directly connected slaves

- sensors, actuators
- valve terminals
- electrical modules etc.
D2 = actuator 1
AS-Interface
Slave IC
1 module
enclosure
D3 = actuator 2
P0
Watchdog
energy
up to 4 sensors
or/and
4 actuators
vampire-connector
14
ASI (2) - Data sheet
master-slave principle
up to 31 slaves on one line
cycle time < 5 ms
each slave can have up to

4 digital inputs + 4 digital outputs
additional 4 parameter bits / slave
Max. 248 digital Inputs and Outputs
also possible: analogue I/O (but ..)
controller
automatic address numbering

via bus connection
master
To Slave 1
To Slave 2
Slave 1
Slave 2
To Slave 31
Slave 31
To Slave 1
Slave 1
15
master calls
slave response
ASI (3) - Topography
star
line
controller
controller
Master
controller
controller
Master
Master
Master
Slave
Slave
Slave
Slave
tree
branch lines
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
Slave
unshielded 2-wire cable
no terminating resistor necessary
data and power on one cable
free tree structure of network
extension: 100 m (300 m with extender)
protection class up to IP67
16
3.3.4 Interbus-S
Discrete Manufacturing bus
2005 April, HK
Interbus-S (2) - Topology
Master
BA
remote "bus"
(ring)
optical fibres also available

localbus (flat cable)
BA
400 m between
devices
BC IO IO IO
5-wire
loop (2 wire, includes power)
bus coupler
18
Interbus-S (4) - Analysis
+ standard in CENELEC
market centered on manufacturing
+ 1700 products, 270 manufacturers,

375.000 applications
limited number of variables (4096 bits)
+ good experience in field wiring

(intelligent wiring bar)
ring structure sensitive to disruptions
+ easy to engineer
sensitive to misplacement
+ easy to program (IEC 61131)
clumsy and slow message service
+ far extension (400m .. 13 km)
medium user community
+ good response time
few and costly tools
+ conformance test
strong ties to Phoenix Contact
19
3.3.5 CAN
Automotive bus
2005 April, HK
CAN (1) - Data Sheet
Supporters
Standard
Medium
Medium redundancy
Connector
Distance
Repeaters
Encoding
User bits in frame
Mastership
Mastership redundancy
Link layer control
Upper layers
Application Protocols
Chips
Automotive industry, Intel/Bosch, Honeywell, Allen-Bradley

SAE (automotive), ISO11898 (only drivers), IEC 61158-x (?)
dominant-recessive (fibre, open collector), ISO 11898
none
unspecified
40m @ 1 Mb/s (A); 400m @ 100kb/s (B); 1000m @ 25kb/s (B)
unspecified (useless)
NRZ, bit stuffing
64
multi-master, 12-bit bisection, bit-wise arbitration
none (use device redundancy)
connectionless (command/reply/acknowledgement)
no transport, no session, implicit presentation
CAL, SDS, DeviceNet (profiles)
comes free with processor
(Intel: 82527, 8xC196CA; Philips: 82C200, 8xC592;
Motorola: 68HC05X4, 68HC705X32; Siemens: SAB-C167
21
CAN (2) - Analysis
Unix" of the fieldbus world.
+ strong market presence, Nr 1 in USA

(> 12 Mio chips per year)
+ supported by user organisations
ODVA, Honeywell, AB.
limited product distance x rate (40 m x Mbit/s)

sluggish real-time response (2.5 ms)
non-deterministic medium access
+ numerous low cost chips, come free

with many embedded controllers
several incompatible application layers

(CiA, DeviceNet, SDS)
+ application layer definition
strongly protected by patents (Bosch)
+ application layer profiles
interoperability questionable (too many

different implementations)
+ bus analyzers and configuration tools

available
+ Market: industrial automation, automobiles
small data size and limited number of

registers in the chips.
no standard message services.
22
3.3.6 Profibus
The process bus
2005 April, HK
Profibus - Family
PROFIBUS-DP (Distributed Processing)

Designed for communication between programmable logic controllers and
decentralized I/O, basically under the control of a single master
Replaces parallel signal transmission with 24 V or 0 to 20 mA by intelligent DIN rail
PROFIBUS-PA (Process Automation)
Permits data communication and power over the bus using 2-wire
Connects sensors and actors on one common bus line even in intrinsically-safe areas.
(chemical industry)
Physical Layer according to international standard IEC 61158-2.
PROFIBUS-FMS (Field Messaging Specification)
General-purpose for peer-to-peer communication at the cell level.
Can be used for extensive and complex communication tasks.
Academic approach (layer 7 services based on MMS, ISO 9506).
Disappearing
24
Profibus - Stack
FMS
FMS
device
profiles
DP
PA
DP-profiles
PA-profiles
DP basic functions
Upper layers
Fieldbus
Messaging
Specification
Link
Phy
IEC interface
RS 485
Fibre optics
25
IEC 61158-2
Profibus - Data sheet

Topography:
Medium:
Signaling:
Integrity
Collision
Medium redundancy
Medium Access
Communication chip
Processor integration
Cycle Time
Address space
Frame size (useful data)
Link Layer Services
bus
TWP @ 31.25 kbits/s (intrinsic safety), 10 devices (PA)
RS 485 @ 19.2 kbit/s.. 500 kbit/s (FMS)
RS 485 or fibres @ 1.5 Mbit/s (12 Mbit/s) (DP)
PA: Manchester II, preamble, delimiters
DP, FMS: UART 11 bit/character
CRC8, HD = 4
none under normal conditions
not supported by the controller
DP: central master, cyclic polling (see: 3.1.2)
FMS, PA: token passing
dedicated chips for 12 Mbit/s
can use UART interface on most processors directly
depends on number of slaves (cyclic, not periodic)
8 bit device address
up to 512 bits in Process Data, 2048 bits in messages
SDN
SDA
SRD
CSRD
Send Data with No acknowledgement

Send Data with Acknowledgement
Send and Request Data with reply
Cyclic Send and Request Data with reply
26
Profibus - Analysis
MS-DOS of the fieldbus world

Standardized by CENELEC (EN 50 170-3)
Exists in four incompatible versions (FMS,
DP, PA, 12 Mbit/s), evolving specifications.
Wide support by Siemens
(Profibus DP is backbone of Simatic S7)
Most products do not implement all the
and active Profibus User Organization
Profibus functionality, interoperability is
(PNO) with >1000 companies.
questionable outside of one manufacturer
200,000 applications, > 2 Mio devices
Additional protocols exist within Siemens
Low entry price (originally simple UART
Weak physical layer (RS 485 at 1,5 Mb/s);
protocol at 500 kbit/s with RS 485 drivers)
to remedy this, a 12 Mb/s version has been
Several implementations based on most
developed (does not significantly improve
commons processors and micro controllers
response time, but limits distance).
(8051, NEC V25, 80186, 68302).
Development tools available (Softing, I-tec).
Extended Application Layer (FMS) and
Network Management (SM7, SM2)
Market: industry automation
Complex configuration - all connections

must be set up beforehand (except
network management): tools required.
Little used outside of Europe (identified in
USA / Asia with Siemens/Germany )
27
3.3.7 LonWorks
The building automation bus
2005 April, HK
LON (1) - Data sheet
Communication chip
Medium redundancy:
Signalling:
bus
STP 150 Ohm @ 1.25 Mbit/s 300m, transformer-coupling
UTP 100 Ohm, @ 78 kbit/s, 1300m, transformer-coupling
reduced to 100m with free topology
power line carrier @ 9.6 kbit/s, limited by -55dB
radio @ 4.9 kbit/s
Neuron chip (Motorola, Hitachi)
none
Differential Manchester for STP, UTP
Medium access:
Response Time
Address space
Frame size (useful data)
Integrity
Higher-level protocols
Application
Support
p-persistent CSMA/CD
3 ms (single call/reply), 400 exchanges/s @ 1.25 Mbit/s
32385 stations
up to 1824 bits
CRC16, HD = 2 against steps, =1 against sync slips)
full 7-layer stack
programmed in Neuron-C
LONMark group (www.echelon.com)
Topography:
Medium:
29
LON (2) - Stack

Application
network variable exchange,
application-specific RPC, etc..
network management
Session Layer
request-response
Transport Layer
acknowledged and unacknowledged, unicast and multicast
Authentication
server
Transaction Control Sublayer
common ordering and duplicate detection
Network Layer
connectionless, domain-wide broadcast,
no segmentation, loop-free topology, learning routers
Link Layer
connectionless frame transfer,
framing, data encoding, CRC error detection
MAC sublayer
predictive p-persistent CSMA: collision avoidance;
optional priority and collision detection
Physical Layer
multiple-media, medium-specific protocols (e.g. spread-spectrum)
30
LON (3) - Analysis
"Macintosh" of the fieldbus world
sluggish response time: > 7ms per variable.
+ several media, products, protocols,

networking, support, starter kits, tools
and documentation.
+ easy, plug-and-play access.
cannot be used in a fast control loop such

as drives or substation protection.
+ low chip costs (10$), but a LON

subprint costs about 500$.
non-deterministic medium access

(p-persistent CSMA)
+ only fieldbus in industry (except for IEC's

TCN) which supports interoperability of
networks of different speeds.
low data integrity due to the use of

differential manchester encoding and
lack of frame delimiter / size field.
+ only fieldbus to provide authentication.
no conformance testing
+ standard network variable types

definition (SNVT).
can only be accessed through Echelon tools
+ standard device description

(LonMarks), access to IEC 1131.
strong ties to Echelon

(net profit in 01Q1: 20000 $)
+ market: building automation

31
3.3.8 Ethernet
The universal bus
To probe further: "Switched LANs", John J. Roese, McGrawHill, ISBN 0-07-053413-b

"The Dawn of Fast Ethernet"
2005 April, HK
The Ethernet consortia

Ethernet/IP (Internet Protocol), Rockwell Automation
www.rockwellautomation.com
IAONA Europe (Industrial Automation Open Networking Alliance, (www.iaona-eu.com)

ODVA (Open DeviceNet Vendors Association, www.adva.org)
CIP (Control and Information Protocol) DeviceNet, ControlNet
ProfiNet
Siemens (www.ad.siemens.de), PNO (www.profibus.com)
Industrial Ethernet new cabling: 9-pin D-shell connectors
direct connection to Internet (!?)
Hirschmann (www.hirschmann.de)
M12 round IP67 connector
Fieldbus Foundation (www.fieldbus.org): HSE FS 1.0

Schneider Electric, Rockwell, Yokogawa, Fisher Rosemount, ABB
IDA (Interface for Distributed Automation, www.ida-group.org) Jetter, Kuka, AG.E, Phoenix Contact, RTI, Lenze, Schneider Electric, Sick
www.jetter.de
33
Ethernet - another philosophy

Ethernet + Fieldbus
(classical)
SCADA
switch
Ethernet
PLC
cheap field devices

decentralized I/O
cyclic operation
PLC
PLC
Fieldbus
simple
devices
Ethernet as Fieldbus
(trendy)
SCADA
switch
Ethernet
costly field devices
Soft-PLC as concentrators
Event-driven operation
Soft-PLC
Soft-PLC
Soft-PLC
Soft-PLC
This is a different wiring philosophy.

The bus must suit the control system structure, not the reverse
34
The "real-time Ethernet"

The non-determinism of Ethernet makes it little suitable for the real-time world.
Several improvement have been made, but this is not anymore a standard solution.
Method 1: Common clock synchronisation: return to cyclic.
Master clock
Method 2: IEEE 1588 (Agilent)

PTP precision time protocol
Method 3: Powerlink
B&R, Kuka, Lenze, Technikum Winterthur
www.hirschmann.de, www.br-automation.com, www.lenze.de, www.kuka.de
Method 4: Siemens Profinet V3

synchronization is in the switches
35
Ethernet and fieldbus roles

Ethernet is used for the communication among the PLCs and for communication of the
PLCs with the supervisory level and with the engineering tools
Fieldbus is in charge of the connection with the decentralized I/O and for time-critical
communication among the PLCs.
local I/O
CPU
fieldbus
Ethernet
36
Time- and safety-critical busses for cars

Contrarily to those who say fieldbus is dead, Ethernet takes it all
automobile manufacturers are developing several real-time busses for X-by-wire:
www.can.bosch.com
www.flexray-group.com
www.tttech.com
37
Car network
extreme low cost, low data rate (100 kbit/s) for general use (power slides)
extreme reliability, excellent real-time behavior for brake-by-wire or drive-by-wire
38
The automotive busses
Mbit/s
50.0
20.0
FlexRay (10 Mbit/s)
10.0
D28, MOST
Token-Ring
optical bus
byteflight
(10 Mbit/s)
5.0
TTP
TDMA, fault-tolerant
2 x 2 wire, 2 Mbit/s
2.0
1.0
CAN-A
2-wire
1 Mbit/s
0.5
MVB
CAN-B
fault-tolerant
0.2
0.1
J1850
0.05
0.02
LIN
Master-Slave
1-wire, not clocked
1
5
39
10
20 / node
Wireless fieldbus
Increasingly, fieldbus goes wireless (802.11b, 802.11g. Bluetooth, ZigBee, WiMax

Advantages: mobility, no wiring
Disadvantages:
Base stations are still costly,
work in disturbed environments and metallic structures costs
mobile = batteries
distance = 30m in factories
lifetime > 5 years ?
privacy
40
Wireless Technologies
costs
high
GPRS
medium
WLAN
low
bluetooth
0.1
10
100 Mbit/s
source: aktuelle Technik, 4/05
41
Safety bus: The organisations
www.fieldbus.org
www.iec.ch
www.interbusclub.com
www.nfpa.org
www.odva.org
www.phoenixcon.com
www.pilz.com
www.profibus.com
www.roboticsonline.com
www.rockwellautomation.com
www.safetybus.com
www.tuv.org
42
Future of field busses
Non- time critical busses are in danger of being displaced by LANs (Ethernet)
and cheap peripheral busses (Firewire, USB)
In reality, these "cheap" solutions are being adapted to the industrial environment
and become a proprietary solution (e.g. Siemens "Industrial Ethernet")
The cost objective of field busses (less than 50$ per connection) is out of reach for
LANs.
The cabling objective of field busses (more than 32 devices over 400 m) is out of reach
for the cheap peripheral busses such as Firewire and USB.
Fieldbusses tend to live very long (10-20 years), contrarily to office products.
There is no real incentive from the control system manufacturers to reduce the
fieldbus diversity, since the fieldbus binds customers.
The project of a single, interoperable field bus defined by users (Fieldbus Foundation)
failed, both in the standardisation and on the market.
43
Fieldbus Selection Criteria
Installed base, devices availability: processors, input/output

Interoperability (how likely is it to work with a product from another manufacturer
Topology and wiring technology (layout)
Power distribution and galvanic separation (power over bus, potential differences)
Connection costs per (input-output) point
Response time
Deterministic behavior
Device and network configuration tools
Bus monitor (baseline and application level) tools
Integration in development environment
44
4 Access to devices
4.3 OLE for Process Control (OPC)
4.3.1 Common elements
2006 May, HK
ABB Research Centre, Baden, Switzerland
Executive Summary
OPC is a set of standard commands collected in a software library (DLL) that can be called by
client applications, written in Visual Basic, C# or other Microsoft programming languages, that allow
to access automation devices (PLCs) in a uniform way, independently from their built or
manufacturer.
To that effect, the particularities of the automation devices are hidden by an OPC server running
either on the same machine as the client program or on another machine, by using DCOM. The
OPC Servers are supplied by the manufacturer of the PLC or by 3rd parties and can manage
several PLCs of the same type. Several servers can run in parallel.
The OPC library allows in particular to read and write process variables, read alarms and events
and acknowledge alarms, and retrieve historical data from data bases according to several criteria.
Automation platforms such as ABB's 800XA platform act as OPC clients to collect data from PLCs
or databases through third-party OPC servers. Several automation platforms act themselves as an
OPC server to publish their data, events and historical data.
OPC is the preferred connectivity for 78% of MES, 75% of HMI / SCADA, 68% of DCS / PLC and
53% or ERP /Enterprise system level applications (according to Arc Advisory Group, 2004)"
keep on reading even if you are not an executive....

2/32
4.3.1 OPC Common
OPC Common Overview
OPC Common
Overview: usage and specifications
OPC as an integration tool
Clients and Servers: configuration
Automation and Custom Interface
OPC Data Access
Overview: Browsing the server
Objects, Types and properties
Communication model
Simple Programming Example
Standard and components
OPC Alarms and Events Specification

Overview: definitions and objects
Events
Alarm Conditions
Automation Interface
OPC Historical Data Specification
Overview
3/32
4.3.1 OPC Common
What is OPC ?
OPC (formerly: "OLE1 for Process Control", now: "Open Process Control") is
an industry standard set up by the OPC Foundation
specifying the software interface (objects, methods) to a server that
collects data produced by field devices and programmable logic controllers.
interfaces
covered by the
OPC standard
application
(OPC client)
OPC server
X
OPC server
(simulator)
Ethernet
(not covered)
node
servers
OPC server
Y
Field bus
(not covered)
PLCs Brand X
PLCs Brand Y
Sensors/Actors
1) OLE (Object Linking and Embedding) is a Microsoft technology for connecting software components.
It has since been extended by the COM / DCOM technology. It corresponds to Java Beans.
4/32
4.3.1 OPC Common
Before OPC
visualization
history
data base
MasterBus
MMS driver
ABB PLCs
XWAY
driver
Profinet
driver
Tlmcanique PLCs
5/32
Siemens PLCs
4.3.1 OPC Common
With OPC: ABB Operator Workplace Connection

OperatorIT
Historian
(Information
Manager)
application software is
written independently from
the type of controller
the drivers still exist,

but the clients do not
see them anymore
ABB AC800M
AC800M
OPC server
Schneider
OPC server
Siemens
OPC server
MMS
XWAY
ProfiNet
Tlmcanique TSX
6/32
Siemens S7
4.3.1 OPC Common
Importance
OPC is the greatest improvement in automation since IEC 61131.

OPC is supported by the OPC foundation (http://www.opcfoundation.org/)
More than 150 vendors offer OPC servers to connect their PLCs, field bus devices,
displays and visualization systems.
OPC is also used for data exchange between applications and for accessing databases
OPC is available as DLL for Automation Interface (Visual Basic,..) and Custom (C++,..)
OPC consists of three major components:
1) OPC - DA = Data Access (widespread, mature)
2) OPC - AE = Alarms and Events (not yet much used)
3) OPC - HDA = Historical Data Access (seldom used)
and some profiles (batch,)
7/32
4.3.1 OPC Common
The main OPC Specifications
OPC
DX
OPC
Batch
OPC
HDA
batch
system
history
data base
OPC
Data Access
OPC
Alarms & Events
OPC
UA
8/32
4.3.1 OPC Common
Specification 1: OPC DA for Data Access

Process variables describe the plant's state, they are generated by the sensors or
calculated in the programmable logic controllers (PLCs).
Process variables can be sent upon a change, on demand or when a given time elapsed.
The OPC DA (Data Access) specification addresses collecting Process Variables.
The main clients of OPC DA are visualization and (soft-) control.
9/32
4.3.1 OPC Common
Specification 2: OPC AE for Alarms and Events
Events are changes in the process that need to be logged, such as "production start"
Alarms are abnormal states in the process that require attention, such as "low oil pressure"
OPC AE (Alarms and Events) specifies how alarms and events are subscribed, under
which conditions they are filtered and sent with their associated messages.
The main clients of OPC AE are the Alarms and Event loggers.
determine the exact time of change
(time stamping)
categorize by priorities
log for further use
acknowledge alarms
(events are not acknowledged)
link to clear text explanation
10/32
4.3.1 OPC Common
Specification 3: HDA for Historical Data Access

Historical Data are process states and events such as: process variables, operator actions,
recorded alarms,... that are stored as logs in a long-term storage for later analysis.
OPC HDA (Historical Data Access) specifies how historical data are retrieved from the logs
in the long-term storage, filtered and aggregated (e.g. compute averages, peaks).
The main client of OPC HDA are Trend Displays and Historians.
11/32
4.3.1 OPC Common
Specification 4: OPC Batch

based on:
IEC 61512-1 Batch Control Part 1: Models and Terminology
(ANSI/ISA S88.01 1995)
ISA-dS88.02-2000 draft 17 of May 2000
Procedure
consists of an
ordered set of
allows to access:
Unit
Procedure
equipment capabilities,
current operating conditions,
historical and
recipe contents
consists of an
ordered set of
Operation
consists of an
ordered set of
Phase
12/32
4.3.1 OPC Common
Beyond Microsoft: OPC UA
In a move to get more independence from Microsoft and use web technology,
a new specification called " Unified Architecture" (formerly. OPC XML) that uses web
services for all kinds of transactions: query, read, write, subscribe,...
The classical OPC DA, AE and HDA are implemented with XML / SOAP /WSDL
this allows encryption and authentication of process data.
This does not only standardize the interfaces, but also the transmitted data.
13/32
4.3.1 OPC Common

OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
14/32
4.3.1 OPC Common
OPC as a hub
OPC variables is also a convenient way to exchange data between applications

on the same machine. OPC data can be easily read in any Microsoft Office application
source: Siemens WinCC
15/32
4.3.1 OPC Common
OPC connection to databases

Tools such as LifeWires allow to build an OPC DA interface to any ODBC - equipped
database.
The database internal structure (exposed through queries) is reflected as a hierarchy
of OPC items.
This allows to give a unified access to simple items.
application
(OPC client)
OPC server
ODBC
DB
16/32
4.3.1 OPC Common
OPC for internal communication: AIP as example
ABB's Integration Platform (AIP) is at the same time an OPC server and an OPC client.
Components (aspects) within AIP expose their properties as OPC objects.
Internal (within AIP) and external communication takes place over OPC.
AIP
aspects
aspects
aspects
aspects
Asset
Optimizer
Enterprise
Historian
OPC server
process
data base
OPC
connections
OPC client
AC800M
OPC server
Schneider
OPC server
17/32
Siemens
OPC server
4.3.1 OPC Common
OPC Connection to Enterprise Resource Planning
Direct connection to SAP (BAPI) is provided by tools such as Matrikon's or Intellution's
18/32
4.3.1 OPC Common
Simulators and Explorer: which helps are available
Explorer:
Several tools are available on the market to browse OPC servers, especially:
- Matrikon OPC Explorer (no source code)

- TopServer Client (source code in VB available)
Simulator:
OPC data should be simulated before commissioning the real plant.
To this effect, commercial simulation servers allow to create, observe and change
variables by hand or according to time functions (ramp, random,).
Most PLC servers have also a simulation mode.
Freeware servers such as Matrikon have only limited number of variables
These explorers and simulators work with OPC DA, AE is yet seldom.
19/32
4.3.1 OPC Common
Client and Servers

OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
20/32
4.3.1 OPC Common
Server(s) and Client(s) in the same node

node
client application
(OPC client)
OPC server
devices
client application
(OPC client)
OPC server
devices devices
OPC server
devices
Clients and servers run as parallel processes
The OPC specification defines the interface between client and server in the form
of objects and methods.
21/32
4.3.1 OPC Common
Direct and Fieldbus access

direct connection
client application
(OPC client)
(local)
OPC server
fieldbus connection
client application
(OPC client)
clients and
servers run as
parallel
processes
(local)
OPC server
FB Manager
fieldbus
proprietary
protocol
I/O devices
The OPC server is running

all the time, as soon as at
least one client is present
can also be
a point-topoint link
22/32
fieldbus
fieldbus
FB agent
FB agent
PLC
PLC
4.3.1 OPC Common
Accessing a server in another node
client application
(OPC client)
stub
DCOM
TCP/IP
DCOM
Limitation:
does not work over firewalls.
Solution:
OPC XML (see later)
TCP/IP
TCP/IP
DCOM
DCOM
OPC
server
OPC server
FB Manager
fieldbus
23/32
4.3.1 OPC Common
Full-fledged COM/DCOM across multiple nodes

connectivity node
Ethernet
app3 TCP/IP
OPC server
for BrandZ
TCP/IP
node 1
application 2
(OPC client)
application 1
(OPC client)
OPC server
for BrandY
OPC server
for BrandX
Fieldbus
Y
Fieldbus X
panel
OPC server
for simulation
drivers
input / output to plant

The OPC servers supports multiple clients and servers on the same, or on remote nodes.
they run as separate processes (as soon as at least one client is requesting them)
24/32
4.3.1 OPC Common
Example: ABB AC800 OPC Server

Process Portal
(client)
Matrikon
client
separate
processes
IEC 61131
Programming
Environment
config
files
(Control Builder)
user panels
AC800
OPCserver
OPCServer
config
MMS
MMS Panel
TCP/IP
by FTP or internal
Ethernet
TCP/IP
MMS
PLC variables
The variables are defined in the server, not in the PLC.

25/32
4.3.1 OPC Common
OPC Technology
OPC Common
OPC Technology, client and custom interface
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
26/32
4.3.1 OPC Common
COM/DCOM quick intro

1) same process
client
2) different processes
same computer
client
3) different computers
client
library
(DLL)
stub
proxy
local
procedure
call
local
procedure
call
stub
remote
procedure
call
COM/DCOM (COM+) maintains

the same interface regardless of
the location of the components
memory
TCP/IP
27/32
proxy
network
library
library
remote
procedure
call
TCP/IP
4.3.1 OPC Common
OPC technologies
ActiveX
Object Linking and
Embedding (OLE)
OLE for Process

Control (OPC)
(Distributed) Component Object Model

(COM / DCOM)
only between
nodes
Transport
(TCP-IP, UDP, Queued)
Ethernet
OPC bases on Microsoft's COM/DCOM technology (i.e. it only works on Windows platforms).
Effort to port it to other platforms (Linux) and web transport protocols (XML) are in progress.
Advantages are the direct integration into all applications, such as Excel.
28/32
4.3.1 OPC Common
Three-tiers Active-X components
graphical interface
business logic
storage and communication
29/32
4.3.1 OPC Common
Structure of an OPC server
OPC/COM Interfaces
OPC Group & Item Management
Item Data Optimization and Monitoring

Device Specific Protocol Logic
Hardware Connection Management
30/32
4.3.1 OPC Common
Automation vs. Custom interface

The OPC specifications define two interfaces: "custom" and "automation".
custom is the native C++ interface of COM.
automation is the interface offered in Visual Basic, used in Word, Excel,..
The interface is defined by a Type Library (distributed by the OPC Foundation)
Functionality is roughly the same in both models, automation is easier to use,
but "custom" gives a more extended control.
"Automation" client
(Visual Basic, Excel,)
"Custom" client
(C++,.)
OPC Automation
wrapper
custom interface
custom interface
server
server
server
31/32
4.3.1 OPC Common
4 Access to devices
4.3.2 Data Access Specification
2005 May, HK
OPC DA: Overview

OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
2/48
4.3.2 OPC DA
To probe further.
OPC Foundation:
Specifications http://www.opcfoundation.org
SoftwareToolbox
Examples in Visual Basic
http://www.softwaretoolbox.com/Tech_Support/TechExpertiseCenter/OPC/opc.html
The Code Project

OPC and .NET
http://www.codeproject.com/useritems/opcdotnet.asp
Matrikon
Free client and server:
http://www.matrikon.com
WinTech
Toolkit for an OPC server
http://www.win-tech.com/html/opcstk.htm
NewAge Automation
http://www.newageautomation.com
3/48
4.3.2 OPC DA
OPC DA: Scope of specification
OPC DA Client
e.g. visualization
OPC DA Client
e.g. performance
indicator
OPC DA
specified
interface
OPC DA Server
An OPC DA Server
is configured using
the information
coming from the
development tools
for the controllers
Variables defined
in the controllers
are mirrored to the
OPC DA server
e.g. Ethernet
tags
controllers
controller
controller
field bus
plant
measurement
points
field devices
4/48
4.3.2 OPC DA
OPC DA: Example of access to a variable
controller development
OPC application
Reactor_1.Program2
ReadItem
("OPC:Reactor1:
Program2.MotorSpeed")
Value: 112
OPC
server
load
symbol
table
MW%1003
MotorSpeed
MW%1004
Temperature
symbols
code
Get (192.162.0.2), MW%1003)
Return (MW%1003, 112)
Network
Reactor_1
Marker: MW%1003
analog input to : IXD.11.2.1
Oven controller
5/48
4.3.2 OPC DA
OPC DA: Objects as viewed by the OPC server

An OPC server is structured as a directory with root, branches and leaves (items)
Process Line 1
Tag Name
Controller 1
Controller 2
Controller_3.Prog_1
Controller_3.Prog_2
TAG
Level_1
TAG
Level_2
TAG
Ramp4
An item is identified by its

"fully qualified ItemID",
e.g.
"Process_Line_1.Controller_2.Level_2"
Cell 1
the hierarchical position may differ from

the fully qualified ItemID
this will be detailed under browsing
Machine 2
Branches may contain other branches and items

The structure may also be flat instead of hierarchical
This structure is defined during engineering of the attached devices and sensor/actors.
(Intelligent servers could configure themselves by reading the attached devices)
6/48
4.3.2 OPC DA
OPC DA: Browsing - methods

An OPC DA server presents an interface that allows the client to explore its structure,
with the methods:
MoveDown
MoveUp
MoveToRoot
showBranches
showLeafes *
GetItemID : retrieves the fully qualified item ID (see later)
Optional:
GetAccessPath: retrieves the access path for items that can be accessed over different ways.
The Access Path is an optional information that the client may provide regarding how to get to the
data, where several possibilities exist. Its use is highly server specific. Do not confound with
hierarchical path.
(*the English error is unfortunate)
7/48
4.3.2 OPC DA
OPC DA: Browsing: Fully Qualified ItemID and hierarchy

A server has internally two ways to access the items:
1) the path shown when exploring the tree, and
2) the fully qualified ItemID, which is the internal name used by the server.
Example:
an item reached as:
needs to be accessed internally as:
Root.SimulatedItems.UserDefined.Ramp.Ramp1
UserDefined!Ramp.Ramp1
Clients usually look for an item though the hierarchical way.

They position the browser on the corresponding branch and retrieve the fully qualified
item ID, which is the name of the item as the server understands it.
The fully qualified name is only used at configuration time, afterwards, objects are
accessed over client handles and server handles (see later)
8/48
4.3.2 OPC DA
OPC DA: Objects as viewed by the OPC client

Each client structures its items by groups,
independently from the server.
clients
Initially, the client browses the server structure to

check if the items it is interested in exist.
A client registers its groups and items at the server.
The server keeps the structure of all its clients.
Server root
Client1
GroupX
Client2
GroupZ
Item1 Item2 Item3 Item1 Item2
server
Area 1
TAG
Temperature
TAG
Heat_On
TAG
Level
Area 2
TAG
Empty_Valve
Area 51
TAG
Fill_Valve
Oven_1
Tank_1
9/48
4.3.2 OPC DA
OPC DA: Client Handle and Server handle

client 1
group1
client 2
group2
1, 76584
2, 87689
3, 23443
3, 53664
4, 43222
client1
The fully qualified item
is not sufficient to identify
an item, a client may
subscribe the same item
in different groups
3, 54
2, 1201
group1
1, 1
2, 2
3, 3
client2
x
2, 2
3, 1202
1,1
2, 2
The pair { ClientHandle, ServerHandle } uniquely identifies an item.
10/48
4.3.2 OPC DA
OPC DA: Object Types and properties

OPC Common
OPC Data Access
Overview: browsing the server
Objects, types and properties
Communication model

Events
Alarm Conditions
Overview
11/48
4.3.2 OPC DA
OPC DA: Item properties

The process data are represented by three dynamic properties of an item:
value:
numerical or text
time-stamp:
the time at which this data was transmitted from the PLC to the server
this time is UTC (Greenwich Winter time), not local time.
quality:
validity of the reading (not readable, dubious data, o.k.)
(when writing, only the value is used)
12/48
4.3.2 OPC DA
OPC DA: Item types

Each item value has a type:
Boolean,
Character,
Byte,
(1 byte)
Word, (2 bytes)
Double Word, (4 bytes)
Short Integer (2 bytes)
Integer (4 bytes)
Long Integer:
Long Unsigned Integer
Single Float (4 bytes)
Double Float (8 bytes)
Currency,
Date,
String,
Array of "the above"
When accessing an item, the client may request

that it is returned with a specific type, which could
be different from the server's type.
(The server's type is returned by browsing)
Type conversion is left to the server, there are no rules
whether and how a server does the conversion.
(use with caution)
Care must be taken that the data types in the
programming language or in the database match those
of the OPC Server.
Items also may have engineering units, but this option
is not often used.
13/48
4.3.2 OPC DA
OPC DA: Communication Model

OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
14/48
4.3.2 OPC DA
OPC DA: Write Communication Models (per group)
client
myGroup.SynchWrite()
server
Call
client
myGroup.AsyncWrite()
Reply
server
Call
Reply
myGroup_AsyncWriteComplete()
The OPC interface accesses only groups, not individual items.

However, the "automation interface" allows to access individual items,
but this does not give rise to a communication
15/48
4.3.2 OPC DA
OPC DA: Read Communication Models (per group)

synchronous
client
myGroup.SynchRead()
asynchronous
server
client
Call
myGroup.AsyncRead()
Reply
server
Call
Reply
myGroup_AsyncReadComplete()
client
myGroup.IsSubscribed
on change
("subscription-based")
server
Subscribe
Notify
myGroup_DataChange()
Notify
16/48
4.3.2 OPC DA
OPC DA: Transmission by subscription (events)

client
myGroup.IsSubscribed
server
Subscribe
Notify
Notify
myGroup.Refresh()
Notify
The server notifies the client if an item changed

- in a particular group (myGroup_DataChange) or
- in any of the groups (myGroups_GlobalDataChange)
In the second case, only the group in which the item changed will be sent.
17/48
4.3.2 OPC DA
OPC DA: "Cache" or "Device" ?

"SynchRead" reads the data either from
cache (local to the PC) or
reads synchronous from the device.
client application
(OPC client)
"Write" is always to device

(DA 3.0 allows write to cache)
OPC server
synchronous call over

the field bus to the end
device (takes a while)
cache
server samples items
(at the RequestedUpdateRate)
and puts them into cache
FB manager
fieldbus
proprietary
protocol
fieldbus connection
no need for device access when

fieldbus operates cyclically
fieldbus
fieldbus
device
device
FB agent
FB agent
18/48
4.3.2 OPC DA
OPC DA: When are subscribed data transmitted ?

A group has two properties to control when a data change is to be transmitted:
myGroup.Refreshrate:
the rate at which the server samples the values, expressed in seconds ! (1/rate)
earliest interval between changes of value (throttles changes, but may miss some)
myGroup.Deadband
applied only to analog values: deadband = % the range (in Engineering Units).
value is transmitted if the difference since last transmission exceeds deadband.
Problem: applies to whole group without considering individual items, seldom implemented.
= refresh rate
max
range
deadband
min
time
19/48
4.3.2 OPC DA
OPC DA: communication paradigm

OPC DA works according to the shared memory paradigm.
This means that a newer value overwrites the older one, no queues or history are kept.
The server does not guarantee that different clients see the same snapshot of the plant.
The server does not guarantee that all changes to variables are registered,
changes may be missed if the polling period is too low.
OPC DA Client
OPC DA Client
OPC DA Server
20/48
4.3.2 OPC DA
OPC DA: Programming Example

OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
21/48
4.3.2 OPC DA
OPC DA: Object hierarchy at the client

Description
OPCServer
OPCGroups (collection)
OPCGroup
OPCGroup
OPCItems (collection)
An instance of an OPC Server. You must create an OPCServer

object before you can get references to other objects. It contains
the OPCGroups Collection and creates OPCBrowser objects.
A collection containing all of the OPCGroup objects this client has
created within the scope of the OPCServer that the Automation
Application has connected to via OPCServer.Connect()
An instance of an OPCGroup object. this object maintains state
information and provides the mechanism to access data for the
OPCItems Collection object that the OPCGroup object references.
A collection containing all of the OPCItem objects this client has
created within the scope of the OPCServer, and corresponding
OPCGroup object that the Automation Application has created.
OPCItem
OPCItem
OPCItem
An automation object that maintains the items definition, current

value, status information, last update time. Note the Custom
Interface does not provide a separate Item Object.
OPCBrowser
An object that browses item names in the servers configuration.

There exists only one instance of an OPCBrowser object per
instance of an OPC Server object.
22/48
4.3.2 OPC DA
OPC DA: Automation interface summary
23/48
4.3.2 OPC DA
OPC DA: Program - initialising a connection

find out existing OPC servers
myDummyServer.GetOPCServers
create an OPCServer object
myServer = new OPCServer
connect to that OPC server
myServer.Connect
create an OPCBrowser object

create an OPCGroups object
Set myBrowser = myServer.Browser

Set myGroups = myServer.Groups
Set myGroup1 = myGroups.Add

Set MyItems = MyGroup1.OPCItems
create an OPCGroup object

build array of items
checking with the browser that
these items exist in this server
next
group
add item array to the group
myGroup1.AddItems
myGroup1.IsActive
myGroup1.IsSubscribed
activate and subscribe group
FQItems1[1] = "Device1.Temp1"
ClientHandle1[1] = 101
ReDim ServerHandle1(nrItems)
ReDim ServerErrors1(nrItems)
ReDim Value1(nrItems)
24/48
4.3.2 OPC DA
OPC DA: Program - Declarations
Option Base 1
Dim WithEvents MyServer
Dim WithEvents MyGroups
Dim WithEvents MyGroup
' items
Dim nrItems
As
Dim MyItems
Dim MyItem
Dim ItemsID(2)
Dim ClientHandles(2)
Dim ServerHandles()
Dim ServerErrors()
As OPCServer
As OPCGroups
As OPCGroup
Integer
As OPCItems
As OPCItem
As String
As Long
As Long
As Long
'OPC
'OPC
'OPC
'OPC
arrays indices start with 1

Server Object (Events optional)
Group Collection (Events opt.)
Group Object
'OPC Item Collection Object

'OPC Item Object
'fully qualified items (see later)
' must be a dynamic array
' must be a dynamic array
Reference: "OPC Automation 2.0" must be included into Visual Basic or C#

(if missing: copy opcdaauto.dll to C:\WINNT\System32\opddaauto)
and register it: C:\>regsvr32 C:\WINNT\System32\opddaauto.
A simple way to do it: install Software Toolbox's TopServer (freeware)
25/48
4.3.2 OPC DA
OPC DA: Program - Finding the OPC servers
The GetOPCServers function applied to a dummy Server object allow to list the existing
servers on this node or on another node (over DCOM - security must be set correctly).
The information about which OPC servers exist is taken from the registry, where it has been p
by each server at its installation time
Private Sub ShowServers(netNodeName As String)
Dim dummyServer As OPCServer
Dim Servers As Variant
Dim cntServers As Integer
' this is an array of strings
Set dummyServer = New OPCServer

Servers = dummyServer.GetOPCServers(netNodeName)
' create a dummy server object

' returns all available servers
For cntServers = LBound(Servers) To UBound(Servers)

MsgBox Servers(cntServers)
Next cntServers
' display the names
Set Getserver = Nothing

Exit Sub
' delete this object (was created by New)
26/48
4.3.2 OPC DA
OPC DA: Program - Connecting to the OPC server

Set MyServer = New OPCServer
' create server object
MyServer.Connect ("Matrikon.OPC.Simulation") ' connect to Matrikon server
Before connecting, it is safe to check the name of the server from the server's list.
Also, it is preferable to include the connection in a separate routine since it can fail:
Function ServerGetCare(Name As String, ServerNode As String) As OPCServer
On Error GoTo ServerGetCareErr
Dim MyOPCServer As New OPCServer
MyOPCServer.Connect ServerName, ServerNode ' connect risky
Set ServerGetCare = MyOPCServer
Exit Function
ServerGetCare_Err:
Err.Clear
MsgBox "Could not connect"
Set MyServer = Nothing
Exit Function
' error handler if connect fails
27/48
4.3.2 OPC DA
OPC DA: Program - Browsing the server

The object OPCBrowser (of type "collection") acts as a pointer to the server's tree:
Dim MyServer As OPCServer
Dim MyBrowser As OPCBrowser
Dim vName As Variant
MyServer.Connect "Matrikon.OPC.Simulation", "Orion"
'server and node name (DCOM)
Set MyBrowser = MyServer.CreateBrowser
' create an OPC browser
MyBrowser.ShowBranches
For Each vName In MyBrowser
MsgBox "Branch: " & vName
Next vName
' show the branches
MyBrowser.ShowLeafs
For Each vName In MyBrowser
MsgBox "Leaf: " & vName
Next vName
' explore the leaves
' display the branch name
' display the leaves's name
28/48
4.3.2 OPC DA
OPC DA: Navigating
MyBrowser.MoveDown (strBranch)
' go down the selected branch tree
MyBrowser.MoveUp
' go up the selected branch tree
Server root
server
Site A
TAG
site properties
Tank
Machine Room
TAG
Level
TAG
Temperature
TAG
Valve_On
Wash Room
Grinder
Climate
Store
There may be leaves at every branch, since a branch may have properties
29/48
4.3.2 OPC DA
OPC DA: Program - get the Fully Qualified ItemID
To get the "fully qualified itemID", one positions the browser at the place where the
leaf is attached to the branch and calls GetItemID
myOPCBrowser.MoveDown("TankArea")
myOPCBrowser.MoveDown("Tank1")
FQI = myOPCBrowser.GetItemId ("Level")
e.g. FQI could be "Controller1;Tanks!Level"

Of course, one can write an Item ID directly when defining a group, but it is safer to
browse the server and get the FQI from there, since the delimiter depends on the server.
30/48
4.3.2 OPC DA
OPC DA: Program - Creating OPCGroups and OPCItems

Set MyGroups = MyServer.OPCGroups
Set MyGroup1 = MyGroups.Add("GRP1")
Set MyItems = MyGroup1.OPCItems
' create groups collection

' add group, name private
' define the OPCItems of group
FQItemIDs(1) = "Area2.Tank1.Level"
ClientHandles(1) = 5
FQItemIDs(2) = "Area2.Tank1.Temperature"
ClientHandles(2) = 6
nrItems = 2
'
'
'
'
fully qualified itemID

arbitrary
fully qualified itemID
arbitrary (but different)
MyItems.AddItems _
nrItems, _
FQItemIDs, _
ClientHandles, _
ServerHandles, _
ServerErrors
'
'
'
'
'
'
adds the items to collection

input parameter
input fully qualified ID
input ClientHandles
return parameter ServerHandles
return parameter ServerErrors
MyGroup1.ClientHandle = 1
MyGroup1.IsActive = True
MyGroup1.IsSubscribed = True
' handle of the group (no s) !

' now ready to send and receive
' and to generate events
The role of the ServerHandles and ClientHandles will be explained later

31/48
4.3.2 OPC DA
OPC DA: Data structures at the client

The client prepares data structures for its items and gives the server the corresponding
pointers so the server can update them.
Items to be written and read can be mixed in the same group.
The type of the item (Boolean, Float,) is implicit, but known at the server
communicated to server by
registering group
FullyQualifiedItemID
returned by server
when registering
dynamic changes
(refreshed on change)
ClientHandle ServerHandle ServerError
Value
Quality TimeStamp
"Channel1.Device1.Temp1"
100
34543
123.4
OK
12:09.234
"Channel1.Device1.Speed1"
102
22532
999.8
OK
12:02.214
"Channel1.PLC2.Door"
203
534676
OK
12:03.002
"Channel1.PLC2.Valve3"
204
787234
OK
12:02.345
"Channel1.PLC2.CloseDoor"
205
58432
BAD
12:02.345
..
..
..
..
..
..
Note: OPC indices start with 1 !

32/48
4.3.2 OPC DA
OPC DA: Synchronous Read of a group

Dim
Dim
Dim
Dim
Dim
Dim
thisGroup As OPCGroup
cntItems As Integer
source As Integer
serverHandles(2) As Long
values() As Variant
errors() As Long
serverHandles(1) =
serverHandles(2) =
client
myGroup.SynchRead()
server
Call
Reply
ServerHandle(11) ' copy from global variables
ServerHandle(14)
source = OPCcache
thisGroup.SyncRead
source,
nrItems,
serverHandles,
values,
errors
' could also be OPCDevice
' identifies the items to be read !

' returns be a dynamic array
' returns a dynamic array
For cntItems = LBound(serverHandles) To UBound(serverHandles) ' 1..n

MsgBox CStr(cntItems) & " : " & values(cntItems)
Next cntItems
33/48
4.3.2 OPC DA
OPC DA: Asynchronous read of single Items

...
MyGroup.AsyncRead
nrItems,
ServerHandles,
ServerErrors,
TransactionID,
CancelID
server
client
AsyncRead()
Call
Reply
AsyncReadComplete()
Private Sub Mygroup_AsyncReadComplete (
ByVal TransactionID As Long,
ByVal NumItems As Long,
ClientHandles() As Long,
ItemValues() As Variant,
Qualities() As Long,
Asynchronous read separates Call and Reply.
TimeStamps() As Date,
Errors() As Long)
Call supplies the ServerHandles
MsgBox ("Async Read Complete")

End Sub
Reply returns the corresponding ClientHandles
34/48
4.3.2 OPC DA
OPC DA: Transaction ID

server
client
Call
AsyncRead (source= device)

AsyncRead (source= cache)
AsyncRead (source = device)
Reply
AsyncReadComplete()
AsyncCancel ()
AsyncCancelComplete ()
Although the AsynchReadComplete carries the ClientHandle of each item,
it does not tell which AsynchRead caused the AsynchReadComplete event to fire.
Call and Reply are linked by the TransactionID: this ID is returned in AsynchReadComplete
It can also be used to cancel the operation
35/48
4.3.2 OPC DA
OPC DA: Reading (by events) the OPC group

...
Private Sub MyGroup_DataChange( _
ByVal TransactionID As Long, _
ByVal NrItems As Long, _
ClientHandles() As Long, _
ItemValues() As Variant, _
Qualities() As Long, _
TimeStamps() As Date)
Dim cntItems As Integer
For cntItems = LBound(ClientHandles)
TextValue(cntItems - 1).Text
TextTimeStamp(cntItems - 1).Text
TextQuality(cntItems - 1).Text
Next cntItems
' returned by the server to the client
To UBound(ClientHandles)
' index 1..n
= ItemValues(cntItems)
' display
= DateAdd("h", 9, TimeStamps(cntItems))
= Qualities(cntItems)
End Sub
This function is called each time an item in the group changes

The ClientHandles (here: 5 and 6) identifies the variables, not the fully qualified itemID
The values are displayed in the TextValue, TextTimeStamp and TextQuality fields.
The refresh rate is given in the group definition.
36/48
4.3.2 OPC DA
OPC DA: Groups Events
Although transmission by groups is more efficient than AsyncRead, it can be improved

by using Groups Events (Global Data Change)
This event is fired whenever a variable of a group changes.

If the group is subscribed also to a Group Event (DataChange), I.e. if the group is
declared WithEvents, then both Events will be fired.
The application must sort out the groups and the items.
37/48
4.3.2 OPC DA
OPC DA: GlobalDataChange

Dim WithEvents MyGroups As OPCGroups
...
Private Sub MyGroups_GlobalDataChange(
ByVal TransactionID As Long,
' =0 if called by Refresh
ByVal GroupHandle As Long,
ByVal NumItems As Long,
ClientHandles() As Long,
' identifies the items
ItemValues() As Variant,
' value of the items
Qualities() As Long,
' value of the items
TimeStamps() As Date)
' timestamps of the items
Select Case GroupHandle
Case 1
' treat group 1
Case 2
' treat group 2
' depending on the group ...
The GlobalDataChange event is fired when any item in a group changed.

(if Groups is also with events, the corresponding Group_DataChange will also be called)
38/48
4.3.2 OPC DA
OPC DA: Server Events

Dim WithEvents MyServer As OPCServer
' define the event
.. ..
Private Sub MyServer_ServerShutDown(ByVal Reason As String)
MsgBox "my OPC Server " & MyServer.ServerName & " quit"
End Sub
This event signals to the client that the server shut down.
The client must declare its server WithEvents and provide the corresponding event
Subroutine
This should stop all actions, otherwise exceptions will occur.
39/48
4.3.2 OPC DA
OPC DA: Do not forget cleanup !

To speed up connection/disconnection, an OPC server remembers its groups and clients
when a client disconnects.
To do this, an OPC server initialises its structures with a client counter of 2, instead of 1.
Therefore, it is imperative to shut down explicitly the server, otherwise links will subside
(and you will have kill the server to clear them).
Private Sub ServerShutdown

Dim dummyServer As OPCServer
Dim Servers As Variant
Dim cntServers As Integer
' this is an array of strings
Set myGroup1 = Nothing

Set myGroups = Nothing
MyServer.Remove
MyServer.RemoveAllGroups
MyServer.Disconnect
Set MyServer = Nothing
' create a dummy server object

' returns all available servers
' delete this object (was created by New)
40/48
4.3.2 OPC DA
OPC DA: Standard and components

OPC Common
OPC Data Access
Communication model
Simple programming example

Events
Alarm Conditions
Overview
41/48
4.3.2 OPC DA
OPC DA: Libraries

The OPC DA specification is not formal, conformance can hardly be checked against this
document.
To ensure that the standard is observed, the OPC foundation distributes on its website
the DLLs (opcdaauto.dll, opccomn_ps,) that contain the type libraries to access the
OPC server.
The vendors are not compelled to implement all features. For instance, the description of
the variables is seldom used. Calling unimplemented functions causes exceptions that
must be caught in Visual Basic with "On Error " statements.
There exist three versions of DA, 1.0, 2.0 and 3.0, that behave differently, however, older
servers do not have a property indicating which version they support.
42/48
4.3.2 OPC DA
OPC DA: Custom Interface

While the Automation Interface is easy to use and quite powerful, some OPC functions
are missing and special operations can only be done in Visual C++ using the custom
COM interface.
This is only recommended for experienced programmers.
IOPCCommon
IOPCServer
IConnectionPointContainer
IOPCItemProperties
[IOPCServerPublicGroups]
[IOPCBrowseServerAddressSpace]
IPersistFile
IOPCDataCallback
IOPCServerShutdown
OPC
client
OPC
server
IOPCItemMgt
IOPCGroupStateMgt
[IOPCPublicGroupSteatMgt]
IOPCSyncIO
IOPCAsyncIO2
IEnumOPCItemAttributes
IOPCSyncIO
IDataObject
43/48
OPC
group
4.3.2 OPC DA
OPC DAOPCGroup Custom Interface: comparison (1)

This checklist for experienced programmers (custom interface) shows the differences between the DA versions
Required Interfaces
DA 1.0
DA 2.0
DA 3.0
OPCGroup
IUnknown
Required
Required
Required
IOPCItemMgt
Required
Required
Required
IOPCGroupStateMgt
Required
Required
Required
IOPCGroupStateMgt2
N/A
N/A
Required
IOPCPublicGroupStateMgt
Optional
Optional
N/A
IOPCSyncIO
Required
Required
Required
IOPCSyncIO2
N/A
N/A
Required
IOPCAsyncIO2
IOPCAsyncIO3
N/A Required
N/A
Required
N/A
Required
IOPCItemDeadbandMgt
N/A
N/A
Required
IOPCItemSamplingMgt
N/A
N/A
Optional
N/A
Required
Required
IOPCAsyncIO
IDataObject
Required
Required
Optional
Optional
N/A
N/A
44/48
4.3.2 OPC DA
OPC DA OPCServer 1.0, 2.0 & 3.0 comparison (2)

This checklist for experienced programmers (custom interface) shows the differences between the DA versions
Required Interfaces
1.0
2.0
3.0
OPCServer
IOPCServer
Required
Required
Required
IOPCCommon
N/A Required
N/A
Required
Required
Required
IOPCItemProperties
IOPCBrowse
N/A Required
N/A
N/A
N/A
Required
IOPCServerPublicGroups
Optional
Optional
N/A
IOPCBrowseServerAddressSpace
Optional
Optional
N/A
IOPCItemIO
N/A
N/A
Required
The differences do not yet appear in the automation interface
45/48
4.3.2 OPC DA
4 Access to devices
4.3.3 Alarms & Events
2005 May, HK
AE: Overview
OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
2/32
4.3.3 OPC A&E
To probe further.
OPC Foundation:
SoftwareToolbox
The Code Project

OPC and .NET
Matrikon
WinTech
NewAge Automation
3/32
4.3.3 OPC A&E
AE: Configuration
OPC AE Client
e.g. event logger
OPC A&E
specified
interface
OPC AE Client
e.g. alarm printer
OPC AE Server
An OPC AE Server
is configured using
the information
coming from the
development tools
for the controllers
Events defined in
the controllers are
mirrored to the
OPC AE server
event notification
e.g. Ethernet
events
controllers
controller
controller
field bus
plant
measurement
points
field devices
4/32
4.3.3 OPC A&E
AE: Purpose
The controllers (PLC) generate events in response to changes in the plant variables.
together with their precise time of occurrence, type, severity and associated message for
the human operator.
An OPC Event server registers these events and makes them available to several clients
A particular class of events are the alarms, which are detailed events that may require
acknowledgement.
The OPC Alarms & Events Interface gives access to the OPC Event server, allowing to:
- browse the OPC A&E Server for predefined events.
- enable or disable alarms and events
- subscribe to alarms and events of interest
- receive the event and alarm notifications with the associated attributes
- acknowledge alarms
5/32
4.3.3 OPC A&E
AE: Definitions
An event is a general change of state that is relevant to the OPC server.
An event signal a change:
1) in the field device ("production started")
2) in the OPC server ("alarm acknowledged")
3) in the application ("operator action")
An alarm is a state of the process that requires attention and is relevant to the OPC server.
An alarm is represented by an alarm condition, (or short: condition), a state machine
indicating if the alarm has been enabled, triggered or acknowledged.
An event or an alarm does not transmit analogue process values,
but they transmit information about their origin, the time of their occurrence and a message
intended for a human operator.
Alarms and events may not get lost
(contrarily to OPC DA, which does not guarantee completeness)
Alarms and event are precisely time-stamped by their source,

(contrarily to process variables, which are time-stamped by the receiving OPC server).
6/32
4.3.3 OPC A&E
AE: communication paradigm

OPC AE works according to the message passing paradigm, contrarily to OPC DA, that
works according to the "shared memory" paradigm.
This means that an event is kept in a queue until all clients have read it (or timed out).
The server guarantees that different clients will see all events in the same sequence.
OPC AE Client
OPC AE
Client
OPC AE
Server
12:34 23.114
12:34 32.334
7/32
4.3.3 OPC A&E
AE: Displaying Alarms and Events

Alarms and events are usually displayed differently on an operator screen.
- Events are displayed in an event list that can become quite long (typically 1000 entries),
entries are not cleared when the source of the event returns to normal
- Alarms are displayed in a short list (typically 50 alarms)
appearance changes when the alarm is acknowledged,
an alarm line is cleared when the alarm signal is cleared.
Ack
checkbox
8/32
4.3.3 OPC A&E
AE: Server Organization

A2
branches
leaves
Areas
A23
Sources
(objects)
Events
A24
S1
E1
S2
S3
E2
E1
E2
E2
E3
simple
tracking
condition
related
condition
related
condition
related
C2
C2
C2
properties
Conditions (alarms only)
Subconditions (alarms only)
SC1
SC1
SC3
SC1 SC1 SC3
SC1
An event is identified by a source (owner object in the controller) and an event name
this combination must be unique in the AE Server.
9/32
4.3.3 OPC A&E
AE: Browsing the AE Server
root
Area 1
Source 1
Area 2
Source 2
Area 5
Source 3
Area 51
Area 52
Alarms and Event are organized by area, which themselves may contain other areas.
Contrarily to branches in OPC DA, area and sources have properties that allow to
disable or enable events or alarms by area or by source, corresponding to parts of the
plants, rooms or specific equipment of the plant.
10/32
4.3.3 OPC A&E
AE: Browsing methods
Like all other OPC Servers, an OPC A&E presents an interface that allows the client to
browse the server to explore its structure, with the methods:
BrowseOPCArea
ChangeBrowsePosition (up, down, root)
GetQualifiedAreaName
GetQualifiedSourceName
There is no "GetQualifiedItemID, since the condition name is known from the source.
11/32
4.3.3 OPC A&E
AE: Events
OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
12/32
4.3.3 OPC A&E
AE: Events kinds
OPC AE defines three kinds of events:
1.
simple: process control system related events (change of a boolean variable)
2.
condition-related: notifies a change of an alarm condition (CLEARED, ACKNOWLEDGED),

(see later)
3.
tracking-related: origin outside of the process (e.g. operator intervention)
13/32
4.3.3 OPC A&E
AE: Event- identification

An event is identified by
- its source (the object that generates the event. e.g. Tank1) and
- the event name (which can be the same as in another object, e.g. HiLevelCond)
HiLevelCond
event
HiLevelCond
event
LoLevelCond LoLevelCond
event
Tank1
Tank2
event
HiLevelCond
event
LOLevelCond
event
Tank3
14/32
4.3.3 OPC A&E
AE: Event PLC Function block

Simple Event function blocks in a controller are used to signal a simple event.
The event is identified by the concatenation of the name of the containing object
(SrcName) and the event handling function block name (here: SimpleEventDetector_1).
The source name can be that of the containing code module (owner object), assuming
that a plant object is represented by a code module.
name of the event
event signal (boolean expression)
is an external signal to be used ?
signal name for external signal (20 characters)
name of the source (30 characters)
message (60 characters)
15/32
4.3.3 OPC A&E
AE: Events - Notification
AE Client
Tank1LevelHigh_SimpleEvent
(source, timestamp, message,
severity, category)
OPC AE
Server
queue
event notification
message
timestamp
Controller
Event
FB
Tank1
Plant
Level Switch
16/32
4.3.3 OPC A&E
AE: Events - Time Stamp
There are three places where events can be time-stamped:

- at the device that originally produced the data (external event - low-level event)
allowing Sequence-Of-Events with a high resolution, down to microseconds
- at the controller, (internal event) using the controller's clock to time-stamp messages
giving precision not greater than the period of the tasks, about 1 ms.
- at the OPC Server, when an event message arrives (tracking events)
not more precise than DA, about 10 ms)
The OPC server can be configured to register the time stamp at the instant of the
event transition (positive or negative) and the instant of the acknowledgement.
17/32
4.3.3 OPC A&E
AE: Properties of an Event-object
all events
condition
-related
events
Property
Source
Time
Message
EventCategory
Severity
OPCEventAttribute
Meaning
source object (area + source)
time of occurrence
associated message for the operator
user-defined
priority (1..1000)
ConditionName
SubCondition
ChangeActiveState
ChangeAckState
ChangeEnableState
ChangeQuality
ChangeSeverity
ChangeSubCondition
ChangeMessage
ChangeAttribute
ConditionAction
ConditionAcknowleddged
Quality
AckRequired
ActiveTime
Cookie
ActorID
name of the condition within the source

name of the active subcondition (subconditions are exclusive)
server handle used for acknowledgement of alarms

identified who acknowledged the alarm (for client-side acknowledgement)
18/32
4.3.3 OPC A&E
AE: Alarm conditions

OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
19/32
4.3.3 OPC A&E
AE: Alarms - Condition Definition

An (alarm) condition is a named state machine that describes the state of an alarm
The condition state is defined by three variables:
Enabled:
the condition is allowed to send event notifications
Active:
the alarm signal is true
Acknowledged:
the alarm has been acknowledged
Alarm signal
(e.g. FIC101.PV > 100 AND FIC101.PV < 150)
Condition
Acknowledgement signal
(a positive transition of a boolean variable)
Condition state
Enable (positive transition)

Disable (positive transition)
20/32
4.3.3 OPC A&E
AE: Alarms - Acknowledgement

An alarm condition becomes active when the PLC produces an alarm signal describing
an abnormal state (e.g. the level of the tank is too high).
The operator is expected to acknowledge this condition (client ack)
Alternatively, a local operator may press a button that the PLC reads (field ack)
AE Client
Tank1Level_ConditionEvent
event notification
client ack (acknowledger ID)

OPC AE
Server
Network or field bus

message
LevelHigh
Condition
timestamp
controller
AckButton
(field ack)
Alarm Signal
21/32
4.3.3 OPC A&E
AE: Alarms - Condition states and acknowledgement

event notification
alarm signal
acknowledgement
condition state
Inactive
Acked
Active
Active
Unacked Acked
Inactive
Acked
Active
Inactive
Unacked Unacked
Inactive
Unacked
Inactive
Acked
alarm_signal
Enabled
Inactive
Acked
condition state
transitions
(here: always
enabled)
Ack
alarm_signal
alarm_signal
Enabled
Inactive
Unacked
Enabled
Active
Unacked
Ack
Enabled
Active
Acked
alarm_signal
An event is generated each time the alarm signal changes state, or is acknowledged
22/32
4.3.3 OPC A&E
AE: Alarms - Condition properties
Name
Name, unique within Server, assigned to the condition
Active
alarm expression is in the state represented by the condition
ActiveSubCondition
If condition active, name of SubCondition (see later)
Quality
quality of data upon which condition is computed
Enabled
condition may become active
Acked
alarm has been acknowledged
LastAckTime
last time that alarm was acknowledged
SubCondLastActive
last time that subcondition became active (see later)
CondLastActive
last time that condition became active
LastInactive
last time that condition became inactive
AcknowledgerID
who acknowledged the alarm
Comment
23/32
4.3.3 OPC A&E
AE: Alarms - Subconditions
A condition may be subdivided into mutually exclusive subconditions

This allows to signal an alarm identified by the object name and give details in the
subcondition.
(for instance: level high, level very high, overflow)
Name
Name, unique within the condition, assigned to the sub-condition
Definition
An expression that defines the sub-state
Severity
priority (different subconditions may have different severity levels)
Description
Text string to be included in the event notification
An alarm condition has at least one subcondition, that defines the severity.
24/32
4.3.3 OPC A&E
AE: Alarms : Example of Function Block (AC800)

name of the condition
Tank1AlarmCond
alarm signal (boolean expression)

is an external signal to be used ?
external signal name (20 characters)
active, acked,..
name of the source (30 characters)

message (60 characters)
(= Priority, 1..1000)
User defined (1..9999)
invert signal
acknowledgement method
shortest condition considered (0..3600)
enable detection (state)
field acknowledgement (positive edge)
disable condition (positive edge)
enable condition (positive edge)
This function block has only one subcondition
25/32
4.3.3 OPC A&E
AE: Summary alarms and events

Event
Alarm
AE Client
AE Client
OPC AE
Server
event notification (source, timestamp, message)
OPC AE
Server
alarm notification (source, timestamp, message,
condition, subcondition, severity, type)
message
timestamp
Event
FB
message
controller
Condition
controller
26/32
4.3.3 OPC A&E
AE: Automation Interface

OPC Common
OPC Data Access
Communication model

Events
Alarm Conditions
Overview
27/32
4.3.3 OPC A&E
AE: Object hierarchy

OPCEventServer
OPCEventSubscriptions (col)
OPCEventSubscription
OPCEventSubscription
OPCEventAreaBrowsers
OPCEventAreaBrowser
OPCEvents (col.)
OPCEvent
OPCEvent
OPCEventCondition
OPCSubConditions (col.)
OPCSubCondition
OPCSubCondition
An instance of an OPC AE Server.

A collection containing all OPCEventSubscription objects this client
has created
An object that maintains state information and provides the

mechanisms for events and alarms notification
A collection of browsers for the server
(only one instance of an OPCBrowser object per instance of an OPCServer object.)
An object that browses items in the servers configuration. It

accesses the arrays of OPCAreas and OPCAreaSources
A collection that holds the OPCEvents objects.
When the Automation Wrapper receives a callback from the AE Server, it forwards
the response as an OPCEvents collection object.
An object that represents one specific event of a subscription

An object that holds the current state of a condition instance, identified
by its Source and Condition Name
A collection that holds the subconditions associated with the event
condition
represents one subcondition associated with the event condition
28/32
4.3.3 OPC A&E
AE: Automation Interface (Summary 1/2)

Methods
Event Server
Event Subscription Mgt
29/32
4.3.3 OPC A&E
AE: Automation Interface (Summary 2/2)

Enums
Classes
Types
Constants
30/32
4.3.3 OPC A&E
4 Access to devices
4.3.4 OLE for Process Control (OPC)
Historical Data Access

2005 May, HK
HDA: Overview
OPC Common
OPC Data Access
Communication model
Events
Alarm Conditions
Overview
2/32
4.3.4 OPC HDA
To probe further.
OPC Foundation:
SoftwareToolbox
The Code Project

OPC and .NET
Matrikon
WinTech
NewAge Automation
3/32
4.3.4 OPC HDA
Historians: An introduction
The role of HDA will become clearer
after a first glance at historians
The process database

decouples the plant from the analysis tools.
It contains recent and historical data.
Invensys (Wonderware) Archestra

4/32
4.3.4 OPC HDA
Historians Example
GE Fanuc/Intellution iHistorian (iFix...)
Questions to the historian:
What was the value of FIC101 last week ?
What was the flow average during October ?
Which were the daily averages in October ?
What was the total flow in each month ?
How much fuel did we use for the batch ?
Give the answer in form of tables,

pie diagrams, spreadsheet, reports
Features:
Unlimited Point Collection
Sub-Second Data Collection Rates
Enhanced Data Compression
True Thin Client Administration
Fault Tolerant Architecture
data collection, archiving and retrieval

report generation
computations (e.g. VBScript)
secure access (FDA 21 CFR 11)
20'000 actions/ s,
up to 100'000 data points
5/32
4.3.4 OPC HDA
Historians: Market
2003 Global Information Management Market by Industry

2%
11%
21%
1%
6%
13%
13%
12%
Chemical
Food and Beverage
Petroleum
Pulp and Paper
Pharmaceuticals
Power
Textiles
Waste and Water
Other
21%
7/32
4.3.4 OPC HDA
Historians: Players
ABB
Information Manager (formerly Enterprise Historian, bases on Oracle)
AIP/PPA (embedded) Historian is another product (small player)

AspenTech
InfoPlus.21
Complete offering with wide range of accompanying applications and services

Very good user interface
OSISoft PI 3.4
Largest installed base (perceived as de facto standard)
Easy-to-use
Perceived low cost (PI 3.4: 1 Mio data points / 80'000 /s)
GE-Fanuc Intellution iHistorian

Relatively new to market. Product created with Mountain Systems
Causing a stir for OSIsoft, AspenTech on pricing
Some small players

Honeywell (POMS) Uniformance
Invensys (Wonderware) IndustrialSQL 8.0
Rockwell Automation (RSBizWare) Historian
Siemens (WinCC) Historian (bases on MS SQL Server)
8/32
4.3.4 OPC HDA
Historian implementation
application
OPC HDA
ODBC/SQL and OPC HDA are
complementary
log organisation
"SQL"
history
database
collector
corrections
operator actions
"OPC"
field data
9/32
4.3.4 OPC HDA
HDA: Historical Data Access
HDA Clients
e.g. Trend Analysis
independent
processes
e.g. Event Logger
OPC HDA
Server
hidden
calculations
history
database
raw and
ordered data
collector
proprietary
data acquisition
OPC DA Server
Field device
Field device
10/32
4.3.4 OPC HDA
HDA: Purpose
An OPC HDA server gives access to a historical data base (logs) in which data from the
process have been collected and time-stamped, possibly through an OPC DA interface.
The OPC HDA interface clients, such as trend analysis, product tracking or data mining,
that require ordered access these data logs.
The OPC HDA interface allows to:
- browse the historical data base

- retrieve data through proper filtering, e.g. by date range, by identity, by property
- build aggregates over the retrieved data, such as average, minimum, maximum.
- enter new entries, correct entries or remove entries
- enter / delete annotations in the history data base
11/32
4.3.4 OPC HDA
HDA: Raw log

12.3.02 13:40
12.3.02 13:40
12.3.02 13:40
12.3.02 13:40
12.3.02 13:41
12.3.02 13:41
12.3.02 13:41
12.3.02 13:41
12.3.02 13:41
12.3.02 13:41
12.3.02 13:41
12.3.02 13:42
12.3.02 13:42
12.3.02 13:42
Gpcpt2ofpbonne
Cpt2bac
Gpcpt2bac
Gpcptbe2
Gpcpt1bac
Gpcpt1ofpbonne
Gpcptae2
Cpt1bac
Gpdefr2
Gpvoydef
Gpr3tempscycleprd
Gpstn1e1
Gpalarme1
Gpalarme2
4824
50
70
45
151
4826
45
49
64
2
318
16
0
0
12.3.02 13:43 Gpetatmodemarche
2
1346
1
16
0
317
0
0
0
1
1992
435
1
1
0
4823
12.3.02 13:43
12.3.02 13:43
12.3.02 13:43
12.3.02 13:43
12.3.02 13:43
12.3.02 13:43
12.3.02 13:43
12.3.02 13:43
12.3.02 13:44
12.3.02 13:44
12.3.02 13:44
12.3.02 13:44
12.3.02 13:44
12.3.02 13:44
12.3.02 13:44
Gptpscycle
Gpetatmodemarche
Gpdefgene1
Gpetatmodemarche
Gptpscycle
Gpdefr2
Gpvoydef
Gpdefgene1
Gpetatmodemarche
Gpr2tempscycleprd
Gptpscycle
Gpalarme3
Gpalarme4
Gpalarme3
Gpcpt2ofpbonne
By definition, values are registered when they change.

(even if data are acquired by periodic polling)
Data in the historical database are identified by their
itemID (here, represented by their name),
value, (of the respective type)
quality (good, stale, bad), and
timestamp (UTC).
12/32
4.3.4 OPC HDA
HDA: How to reduce raw data (even before HDA comes into play)
Tag values are sent to the OPC DA server when they change (on events).
The collector records these data in a log that is organized cyclically.
Since a log has a limited storage capacity, data are reduced by several means:
1) if a variable is received more often than the log's minimal storage interval (e.g. 1s),
the log keeps the latest of all values of the interval
2) if (and only if) a received variable changed by more than the "exception deviation"
(e.g. 5%, analog values only), it is entered into the log.
3) if a variable changed by less than the exception deviation, it may nevertheless be
forced into the log after the log's maximum storage interval (e.g. 4 s) elapsed.
Two situations can occur:
- the time scale of the log is smaller than that of the trend display in this case, values have to be
interpolated to be displayed correctly
min time
max time
max time
max time
max
exception
deviation
min
time
= process value (event)
13/32
= log entry
4.3.4 OPC HDA
HDA Application: Trend Display
Parameters:
log: how were data sampled
time scale (with possible offset, zoom, pan)
amplitude scale (low range, high range, scale units)
style: smoothed, stepped, filled (several ways to display the same data)
extrapolate: how to display values not received (e.g. because they did not yet change)
14/32
4.3.4 OPC HDA
HDA: More logs
A log registers only one data point !

log 2.1
log 1
log 2.2
(log templates can be used for

several datapoints)
log 2
Data in the log may be:

Measured values (original, from the OPC DA)
Calculated values
Corrected values (replaced or inserted)
Annotations (explanation strings)
raw log
Collector
(OPC DA Client)
other servers
OPC DA Server
Field device
XML / CSV
Field device
15/32
4.3.4 OPC HDA
HDA: Hierarchical logs

long-term log
1 hour_forever
5.2 MB / Year
3 27 28 29 30
september octobre
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
october
1 mn_7 days log

6 MB
d-7
1s_24 hours log:

50 traces @ 12 B
52 MB
23
24
d-6
d-5
d-4
10
d-3
11
12
13
A hierarchical log is built on the data contained in the parent log.

To reduce the log size, several aggregations can be applied:
- record only maximum, minimum, average over a period, etc...
16/32
d-2
14
15
16
yesterday
17
18
19
today
20
21
22
Actual data
4.3.4 OPC HDA
HDA role
An HDA server allows to browse the logs and filter the logs to be read.
It can filter them according to their attributes, such as
name of the datapoints, type of the datapoint, engineering units,....
An HDA server then selects the logs to be read.

An HDA can then read the logs, defining a certain period
(Start time, End Time, and the maximum number of entries to be retrieved.
When reading the log, the HDA server can process the log by applying aggregations
The "min time", "deviation", "max time" rules used to collect raw data is also used
as an aggregation to build new logs.
Therefore, hierarchical logs can be built out of a raw log by an OPC HDA service.
Do not confound aggregates and attributes.....
17/32
4.3.4 OPC HDA
HDA: Server
The client can query the HDA server for its capabilities, and in particular:
- which LocaleID (national languages) the server supports (as in DA and AE servers)
- which attributes the server supports (the attributes of its items is a subset thereof)
- which aggregates the server support (and the client can request)
- if the server supports correction / deletion of data (synchronous and/or asynchronous)
- if the server supports insertion of annotations (synchronous and/or asynchronous)
Many capabilities of the HDA server are optional.
18/32
4.3.4 OPC HDA
HDA: Browsing: structure

Like the DA and AE interfaces, OPC HDA server offers a browser with methods allowing
to navigate up and down, and to retrieve branches and leaves at each position.
If filters are defined at its creation time, an HDA browser only browse for items with
specific attributes (e.g. Engineering Units = m/s). The filtering attributes are additive.
HDA (contrarily to DA and AE) allows several browsers, each with its different filter.
root
Leaves
Object
Log 1
Object
Log 2
Area 5
Log 3
Area 51
Area 52
19/32
4.3.4 OPC HDA
HDA Browsing and filtering

The filter criteria are based on attributes (see the following slides).
Attributes cannot be modified afterwards.
Set MyOPCHDABrowser = MyOPCHDAServer.Createbrowser
(NumCriteria, AttributeIDs, operatorCode, Filters, Errors
MyOPCHDABrowser.MoveDown (MyOPCHDABrowser.OPCHDABranches(1))
Items = MyOPCHDABrowser.OPCHDAItems
.... list these items
20/32
4.3.4 OPC HDA
HDA: General attributes of the items

Code Attributes
Description
1
2
3
4
DATA_TYPE
DESCRIPTION
ENG_UNITS
STEPPED
5
6
7
ARCHIVING
DERIVE_EQUATION
NODE_NAME
PROCESS_NAME
SOURCE_NAME
10
SOURCE_TYPE
11
NORMAL_MAXIMUM
12
NORMAL_MINIMUM
13
ITEMID
data type for the item. See the definition of a VARIANT for valid values (VT_R4, etc.)
text that describes the item.
label to use in displays to define the units for the item (e.g., kg/s).
whether data from the history repository should be displayed as interpolated (sloped lines
between points) or stepped (vertically-connected horizontal lines between points) data.
whether historian is recording data for this item (0 means no).
equation to be used by a derived item to calculate its value. This is a free-form string.
machine which is the source for the item. This is intended to be the broadest category for
defining sources. For an OPC Data Access Server source, this is the node name or IP
address of the server. For non-OPC sources, the meaning of this field is server-specific.
process which is the source for the item. This is intended to the second-broadest category
for defining sources. For an OPC DA server, this would be the registered server name.
For non-OPC sources, the meaning of this field is server-specific.
name of the item on the source. For an OPC DA server, this is the ItemID.
sort of source produces the data for the item. For an OPC DA server, this would be "OPC".
upper limit for the normal value range for the item, is used for trend display default scaling
and exception deviation limit calculations, should be the normal high value for the item.
lower limit for the normal value range for the item, is used for trend display default scaling
and exception deviation limit calculations, should be the normal low value for the item.
item identifier. This is used to allow filtering in the CreateBrowse method.
21/32
4.3.4 OPC HDA
HDA: Attributes that affect how data are archived

Code Attributes
Data Type
Description
14
Max Time Interval
VT_FILETIME
15
Min Time Interval
VT_FILETIME
16
Exception Deviation
VT_R8
17
Exception Deviation Type
VT_I2
18
High Entry Limit
VT_R8
19
Low Entry Limit
VT_R8
Specifies the maximum interval between data points in the

history repository regardless of their value change.A new value
shall be stored in history whenever OPCHDA_MAX_TIME_INT
seconds have passed since the last value stored for the item.
Specifies the minimum interval between data points in the
history repository regardless of their value change.A new value
shall be not be stored in history unless
OPCHDA_MIN_TIME_INT seconds have passed since the last
value stored for the item.
Specifies the minimum amount that the data for the item must
change in order for the change to be reported to the history
database. See OPCHDA_EXCEPTION_DEV_TYPE for the
specific meaning of this field.
Specifies whether the OPCHDA_EXCEPTION_DEV is given
as an absolute value, percent of span, or percent of value. The
span is defined as OPCHDA_HIGH_ENTRY_LIMIT OPCHDA_LOW_ENTRY_LIMIT.
Specifies the highest valid value for the item. A value for the
item that is above OPCHDA_HIGH_ENTRY_LIMIT cannot be
entered into history. This is the top of the span.
Specifies the lowest valid value for the item. A value for the
item that is below OPCHDA_LOW_ENTRY_LIMIT cannot be
entered into history. This is the zero for the span.
22/32
4.3.4 OPC HDA
HDA: Selecting the items

Each server has only one group of items, the collection OPCHDAItems.
By adding an OPCHDAItemID to that collection, the client defines its client handle and
receives the server's server handle.
The OPCHDAItemID (a string) should be retrieved by using the browser (like in DA or

AE), since entering a string directly is unsafe.
The server handle is used to access the log.
Read operations are applied to the whole collection OPCHDAItems.
....
Set colOPCHDAItems = MyOPCHDAServer.OPCHDAItems

colOPCHDAItems.AddItem OPCHDAItemID ClientHandle
colOPCHDAItems.SyncReadRaw .....
23/32
4.3.4 OPC HDA
HDA: Reading the data

When reading the data, the client has the choice to read synchronously or
asynchronously (if the server permits the latter).
The client specifies the time interval (Start time and end time) and whether the end
values of that interval shall be included.
The client has the choice to read the raw data or to apply an aggregation to the read
data. The aggregations (see next slides) allow to reduce data.
24/32
4.3.4 OPC HDA
HDA: Aggregates
Symbol
0 NoAggregate
1 Interpolative
2 Total
3 Average
4 TimeAverage
5 Count
6 StDev
7 MinimumActualTime
8 Minimum
9 MaximumActualTime
10 Maximum
11 Start
12 End
13 Delta
14 RegSlope
15 RegConst
16 RegDev
17 Variance
18 Range
19 DurationGood
20 DurationBad
21 PercentGood
22 PercentBad
23 WorstQuality
24 Annotations
Description
Do not retrieve an aggregate (use only start time - stop time and filtering).
Get interpolated values.
Get the totalized value (time integral) of the data over the resample interval.
Get the average data over the resample interval.
Get the time weighted average data over the resample interval.
Get the number of raw values over the resample interval.
Get the standard deviation over the resample interval.
Get the minimum value in the resample interval and the timestamp of the minimum value.
Get the minimum value in the resample interval.
Get the maximum value in the resample interval and the timestamp of the maximum value.
Get the maximum value in the resample interval.
Get the value at the beginning of the resample interval (time stamp).
Get the value at the end of the resample interval (time stamp of end of the interval).
Get the difference between the first and last value in the resample interval.
Get the slope of the regression line over the resample interval.
Get the intercept of the regression line over the resample interval, (time start of interval).
Get the standard deviation of the regression line over the resample interval.
Get the variance over the sample interval .
Get the difference between the minimum and maximum value over the sample interval.
Get the duration (in seconds) of time in the interval during which the data is good.
Get the duration (in seconds) of time in the interval during which the data is bad.
Get the percent of data (1 = 100 %) in the interval which has good quality.
Get the percent of data (1 = 100 %) in the interval which has bad quality.
Get the worst quality of data in the interval.
Get the number of annotations in the interval.
25/32
4.3.4 OPC HDA
HDA interpolation
interpolation is an aggregate used when the time scale of the derived log is smaller than
that of the parent log.
= parent log
= filtered log
resample interval of the new log
time
This is useful to display smoothly trend diagrams.
26/32
4.3.4 OPC HDA
HDA corrections and annotations
These features are seldom implemented.

Basically, the values are accessed using the server handles that
OPCHDAItems.AddItem returned.
27/32
4.3.4 OPC HDA
HDA Automation Model

OPC Common
OPC Data Access
Communication model
Events
Alarm Conditions
Automation Model
Overview
Automation Model
28/32
4.3.4 OPC HDA
HDA: Automation Interface summary

OPCHDAServer
An instance of an OPC HDA Server. An OPCHDAServer object must be

created to can get references to other objects.
OPCHDAItems (col)
A collection containing all the OPCHDAItem objects this client has

created within the scope of the OPCHDAServer object
OPCHDAItem
OPCHDAItem
An object that maintains the items definition. Note the Custom Interface
does not provide a separate Item Object.
OPCHDABrowsers
An object that browses item names in the servers configuration.
OPCHDAHistory
An collection containing a series of either OPCHDAValue or

OPCHDAEntry objects, representing historical values of an item.
OPCHDAValue
OPCHDAValue
An object that represents a discrete historical value for an item or an

attribute.
OPCHDAEntry
OPCHDAEntry
An object that extends the OPCHDAValue object to include additional

information
29/32
4.3.4 OPC HDA
HDA: Object Interfaces (1)
30/32
4.3.4 OPC HDA
HDA: Object Interfaces (2)
31/32
4.3.4 OPC HDA

Curs1 RIL Stud

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Curs1 RIL Stud

Uploaded by

Copyright:

Available Formats

Industrial Automation

Reele industriale locale

Titular curs: s.l. dr. ing. Nicoleta-Cristina G ITAN

Reele industriale locale

14s pt mnix3 ore

EPFL - Industrial Automation

3.3.1 OSI model

Cum se face notarea?

Nota la examen NF = 50%NL + 40%NE + 10%PC

Examenulsedesf oar nscris.

Sedaubonusuridepn 10%pentruactivit isuplimentare(participarealaconcursuri,

EPFL - Industrial Automation

3.3.1 OSI model

The OSI model

The Open System Interconnection (OSI) model is a standard way to

EPFL - Industrial Automation

3.3.1 OSI model

OSI-Model (ISO/IEC standard 7498)

EPFL - Industrial Automation

All services directly called by the end user

3.3.1 OSI model

OSI Model with two nodes

EPFL - Industrial Automation

3.3.1 OSI model

The repeater restores signal

3.3.1 OSI model

OSI model with three nodes (bridge)

physical medium (0)

physical medium (0)

e.g. Ethernet 100 MBit/s

The subnet on both sides of a bridge have:

3.3.1 OSI model

There is only one Ethernet which carries

There are now two Ethernets and only the

3.3.1 OSI model

Networking with bridges

EPFL - Industrial Automation

3.3.1 OSI model

a switch is an extension of a hub that allows store-and-forward.

EPFL - Industrial Automation

3.3.1 OSI model

OSI Model with three nodes (router)

physical medium (0)

EPFL - Industrial Automation

3.3.1 OSI model

Repeater, Bridge, Router, Gateway: Topography

backbone (e.g. FDDI)

connects different speed,

EPFL - Industrial Automation

3.3.1 OSI model

Repeaters, Bridges, Routers and Gateways: OSI model

intelligent linking devices can

100 Mbit/s Ethernet

EPFL - Industrial Automation

3.3.1 OSI model

To which level does a frame element belong ?

A frame is structured according to the ISO model

EPFL - Industrial Automation

3.3.1 OSI model

3.3.1 OSI model

Example: OSI-Stack frame structure

Protocol Identifier (81)

PDU Segment Length