Emanuel Wahlqvist - DiVA portal

Pre-study of new electrical coupling between train cars
Emanuel Wahlqvist
Teknisk- naturvetenskaplig fakultet UTH-enheten Besöksadress: Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0 Postadress: Box 536 751 21 Uppsala Telefon: 018 – 471 30 03 Telefax: 018 – 471 30 00 Hemsida: http://www.teknat.uu.se/student
Abstract
Emanuel Wahlqvist
This study is meant to be an initial study of the possibility to replace the discrete control signal wires over the electrical coupler between train cars with a data bus system. The reason for this is that the electrical coupler is large and heavy due to the high amount of contacts it contains. It is also a problem for manufacturers who are upgrading an existing fleet and need to transfer more signals through a coupler with no spare contacts to use. Except the control signals there are also Ethernet and power signals in the electrical coupler. Some trains also use a bus system for control signals and/or signals containing a large amount of data such as passenger information. This report gives a presentation of some common ways to distribute electrical signals throughout a train used by most manufacturers. It also presents some design recommendations for a system that would collect existing signals to a bus system and two different design proposals that should be considered if such a system is to be developed. The study shows that there are already systems on the market for transferring control signals over a bus but they are more aimed for trains under construction. For this reason a new bus system would only be suitable for upgrading existing couplers to free up space in the electrical coupler unit.
ISRN-UTH-INGUTB-EX-E-2011/06-SE Examinator: Nora Masszi Ämnesgranskare: Tadeusz Stepinski Handledare: Lars Möttönen
Sammanfattning
Syftet med examensarbetet är att undersöka möjligheten att använda en databuss för
överföring av diskreta styrsignaler över den elektriska kopplingen på ett tågkoppel. Detta
skulle drastiskt minska ned storleken och således även vikten på kopplet. Dessutom skulle
de bli mycket enklare att uppgradera ett befintligt elkoppel så att det kan bära flera
styrsignaler utan större modifieringar av kopplet.
I kopplet går inte bara de diskreta styrsignalerna utan även Ethernet trafik och
strömförsörjning. En del tåg har även någon form av bussystem för styrsignaler och/eller
datakrävande signaler som till exempel passagerarinformation.
Rapporten ger en presentation av de vanligaste metoderna att överföra dessa signaler, några
rekommendationer för hur ett nytt bussystem bör utformas samt två förslag på hur ett
sådant system skulle kunna se ut hårdvarumässigt. Resultatet av examensarbetet är att ett
bussystem är ett bra alternativ vid uppgradering av befintliga elkoppel, men vid
nykonstruktion så är de befintliga bussystemen det bästa alternativet då det inte bara
minimerar antalet kablar genom elkopplingen utan genom hela tåget.
Thesis work: Pre-study of new electrical coupling between train cars
I
Preface
This report gives a proposal for Dellner Couplers AB (DCAB) how to design an electrical
coupler that is smaller than the ones used today. This is not so much for new designs but
mostly for upgrading older couplers to carry more signals without replacing the whole
electrical coupler unit.
The thesis was initiated and supervised by Lars Möttönen at DCAB. The approval and
review was done by Professor Tadeusz Stepinski from Uppsala University.
Thanks to everybody at DCAB for their support with special thanks to Lars Möttönen and
the others at the electrical department.
II
2 How signals are transferred today .................................................................................. 2
2.1 Dedicated wires ....................................................................................................... 2
2.2.1 Functionality .................................................................................................... 2
3.1 Standards ................................................................................................................. 5
3.2.2 Additional guidelines for systems including software and processors ............ 7
4 Concept design proposals ............................................................................................... 8
4.1 Multiplexer with microprocessors ........................................................................... 8
4.1.1 General description .......................................................................................... 8
4.2 Multiplexer without microprocessors .................................................................... 11
4.2.1 General description ........................................................................................ 11
4.2.2 Input sampling ................................................................................................ 11
7.1 Based on OP-amp .................................................................................................. 17
7.1.1 Simulations ..................................................................................................... 18
III
7.2.1 Simulations ..................................................................................................... 22
7.3 Conclusions ........................................................................................................... 24
7.3.2 Based on non-linear circuits ........................................................................... 24
7.3.3 Comparison .................................................................................................... 24
8.1 Input mux .............................................................................................................. 25
8.2 Output mux ............................................................................................................ 26
Table of figures Figure 1-1: Arrangement of coupler contacts, every signal represented twice. ..................... 1
Figure 2-1: The CCU and the TCU ........................................................................................ 4
Figure 4-1: Schematic of the data flow in the units ............................................................... 9
Figure 4-2: Schematic of the integration of the units on a train........................................... 10
Figure 4-3: Schematic of the integration of the units on a train........................................... 11
Figure 7-1: Schematic of the linear logic converter ............................................................. 17
Figure 7-2: With calculated values according to the equations, R3=8.461k, R4=7.857k 18
Figure 7-3: With resistor values from E12 series, R3=8.4k, R4=6.8k ............................ 19
Figure 7-4: With calculated values according to the equations, R3=57.894k, R4=10k .. 19
Figure 7-5: With resistor values from E12 series, R3=56k, R4=10k .............................. 20
Figure 7-6: Schematic of non-linear logic converter ........................................................... 20
Figure 7-7: With calculated values according to the equations USWITCH=40V, R1=7.02k,
R2=180, R3=125 ............................................................................................................. 22
Figure 7-8: With resistor values from E12 series USWITCH=40V, R1=8.2k, R2=180,
R3=120 .............................................................................................................................. 23
R2=180, R3=88.7 ............................................................................................................ 23
R3=120 .............................................................................................................................. 24
IV
Table of tables Table 4-1: Risk analysis of multiplexer without microprocessors ....................................... 12
Abbreviations CRC – Cyclic Redundancy Check
CTR – Current Transfer Ratio
I/O – Input/Output
Thesis work: Pre-study of new electrical coupling
1 Introduction to train couplers A train coupler is the mechanism
function of a coupler is uninteresting as it is irrelevant to the electr
The electrical part of the coupler
shaped as a big rectangular contact
in a train set is identical and one half contains pins
one, but with sleeves instead.
cable inside the coupler.
any direction. The coupler also transfers electrical power and Ethernet data traffic
signals are not considered in this thesis
Figure 1-1: Arrangement of coupler contacts, every signal represented twice.
The size of the coupler varies with the application and depends on the amount and type of
signals being transferred, and
study of new electrical coupling between train cars
1
Introduction to train couplers A train coupler is the mechanism that connects two train cars. In this thesis the mechanical
function of a coupler is uninteresting as it is irrelevant to the electr
The electrical part of the coupler transfers all electrical signals that control
big rectangular contact and mounted on the mechanical coupler
in a train set is identical and one half contains pins. The other half is a mirror of the first
one, but with sleeves instead. The signals is wired to one half and linked to the
cable inside the coupler. This is to provide the functionality that any car can be coupled in
The coupler also transfers electrical power and Ethernet data traffic
considered in this thesis, as they cannot be multiplexed
: Arrangement of coupler contacts, every signal represented twice.
signals being transferred, and since 140 or so signals with a four mm pin/sleeve
not uncommon, they can get very big.
between train cars
that connects two train cars. In this thesis the mechanical
function of a coupler is uninteresting as it is irrelevant to the electrical coupling.
electrical signals that control the train. It is
and mounted on the mechanical coupler. Every coupler
he other half is a mirror of the first
The signals is wired to one half and linked to the other with a
This is to provide the functionality that any car can be coupled in
The coupler also transfers electrical power and Ethernet data traffic but these
multiplexed.
: Arrangement of coupler contacts, every signal represented twice.
four mm pin/sleeve pair each is
2
2 How signals are transferred today Today there are several different ways of transferring an electrical signal throughout the
train. The most common way is to combine two different types of signal transfer. One type
is to use a dedicated wire that stretches throughout the train, and sometimes even back
again. The other type is called Train Communication Network (TCN) and is a combination
of two data buses and distributed I/O. Both the dedicated wires and the bus combination are
used for control signals, however, some train manufacturers do not entirely trust the bus
system for their critical control signals and therefore use the more service proven dedicated
wires1.
2.1 Dedicated wires This is, by some manufacturers, still considered the safest way of transmitting control
signals. Some of the signals are even wired in a loop to get a confirmation that they have
passed through the whole train. The discrete signals that are transferred by a dedicated wire
often have a voltage level of 72 or 110 volts.
Advantages with the dedicated wire transfer type:
• A malfunction is easy to locate because the only ones possible are loss of electrical
contact or short circuit.
• The signals have no common component which means that a malfunction in signal
“A” does not affect the function of signal “B”.
• The risk that a malfunction occurs because of a failure of a contact in the electrical
coupler is very small. Experience from the field says that the frequency of an
occurred error is 8*10-9 errors per hour2.
Disadvantages with this transfer type:
• The electrical coupler expands in size, cost and weight.
2.2 Train Communication Network The TCN is defined in IEC standard 61375-1 that was first published in 1999. The second
edition, the one valid today, was published in 2007. The standard defines a combination of
two different data buses for I/O management on a train. The names of these buses are MVB
and WTB.
2.2.1 Functionality
The MVB is used in separate cars or a set of multiple cars that does not change their
composition as they are configured with fixed addresses. In a complete train there will be
1Nilsson Bo (2011) Customer Service Engineer DCAB (Verbal information) 2Eriksson Conny (2011) RAMS/LCC Engineer DCAB (Verbal information)
3
several separate buses and the purpose of WTB is to connect these segments with each
other. The WTB doesn’t have fixed addresses but instead uses a negotiation called
inauguration. The inauguration automatically numbers all units on the bus sequentially
starting with the master as number 01, ascending in one direction and descending from 63
in the other. Each node on the bus has one of the following ranks:
• Strong node
• Weak node
• Slave node
In the inauguration process the ranking is used to decide which node is to become master of
the bus. If there is only one strong node present on the bus it automatically becomes the
master, but if there are more than one the bus is divided in individual segments with one
segment per master. When this occurs the user is notified that a complete bus could not be
configured and that he or she should degrade all but one strong master to weak masters as
they then will act as slaves. In the case when several segments controlled by weak masters
are connected the segment containing the most nodes renames the smaller segment. In the
case when they consist of the same amount of nodes a randomized process decide which
segment will rename the other. A slave node cannot act as master unless a user tells it to.
The other differences is the speed, WTB operates at 1Mbit/s over a shielded twisted pair
cable and MVB at 1.5Mbit/s3 over either cables that follow the RS-485 standard for
distances up to 20m, a shielded twisted pair cable for connection distances up to 200m or
optical fibers for longer distances up to 2km4.
Advantages of the TCN:
• The electrical coupler can be made small and light.
• Opens the possibility of upgrading the train in the future without redesign of the
electrical coupler.
• Still operational if a contact fails as long as redundancy is applied.
• Embedded functions for logging and diagnosis.
Disadvantages with this transfer type:
• More difficult to debug than a wire.
• In case of a bus failure, many signals are affected.
2.2.2 Commercial product example
The Siemens SIBAS 32 control system is a complete solution for controlling a train based
on modules, which communicates over the TCN. There are the Central Control Unit who
3Kirrmann, H and Zuber, Pierre A. (2001) The IEC/IEEE Train Communications Network 4The International Electrotechnical Commission (2007). International standard IEC 61375-1:2007(E), Geneva
sends the control signals from the drivers cab to the doors, brakes, tracti
other functions. There is also the Traction Control Unit which receives information from
the CCU and controls the traction motors. For interfacing to peripheral devices the SIBAS
KLIP is used. It is a module for decentralized analog signals
Figure 2-1: The CCU and the TCU
4
sends the control signals from the drivers cab to the doors, brakes, tracti
KLIP is used. It is a module for decentralized analog signals over the TCN.
: The CCU and the TCU
between train cars
sends the control signals from the drivers cab to the doors, brakes, traction control and
over the TCN.
5
3 Other transferring methods An alternative bus system for distributing signals in a train is called distributed I/O. It
consists of a PLC with its I/O units located on the other end of a bus. This bus can be
configured with the Profisafe protocol which would mean that it meets the requirements of
Safety Integrity Level (SIL) 3. SIL is a classification divided into 5 levels (0 is the lowest
and 4 the highest) of functional safety in electrical and mechanical systems and is widely
used on automation systems in the process industry and is starting to reach the railway
industry as well5. The classification is meant to be for a complete system which in this case
would mean the whole train but it can also be used for subsystems. To get a reference on
how safe the transmission is today the value of contact errors per hour can be used. 8*10-9
errors per hour conform to the hardware requirements of SIL 3 systems6. If a Profisafe bus
would be implemented on a train it would automatically reduce the size of the electrical
coupler.
This, however, is not in DCAB’s business scope and the implementation of such a system
is therefore a matter for the train manufacturers.
A possible solution for DCAB is to develop a similar system that takes use of the individual
cables and transfer their logical values over the coupler on a data bus. As this is very much
what TCN does it wouldn’t apply for newly developed trains but would be of great use on
older ones in the process of being upgraded.
3.1 Standards If a new bus system where to be accepted by the railway industry it would need to conform
to some standard. The problem is which one to choose because it differs between countries
which are applicable. To use all of them would be a very time demanding task and probably
result in an unreasonably high cost. Based on the fact that DCAB from the beginning is a
European company it seems logical to use a European standard and then adapt the system
to the demands of other countries if the need arises. Applicable standards would then be
EN-50126, EN-50128 and EN-501297. The main consequences for the development by
these standards are that the whole development process is to be supervised by an
independent firm. The product must then be analyzed and tested for its application
according to the requirements of the standards to make sure that every possibility of a
dangerous failure is taken care of.
5Strandberg Per. (2011) Consultant at DCAB, RAMS/LCC-department (Verbal information) 6Siemens AB (2009) Simatic Safety Integrated Practical application of 62061, Kap9.2 Correlation: SIL and PFHD of a SCRF Available at: http://www.nwe.siemens.com/sweden/internet/se/produkter/industry/automation/as- event/funktionssakerhet/Maskinsakerhet_2009/Documents/Maskinsakerhet_Simatic_Safety_Integrated_Practi cal_application_of_62061.pdf (2011-05-10). 7Strandberg Per. (2011) Consultant at DCAB, RAMS/LCC-department (Verbal information)
6
Before the initialization of such a project it would be a good idea to develop a system with
the standards in mind but without the certification. This would work as a proof of concept
to present to customers and get a better feeling if they are interested. Together with an
extensive market research this would provide a base for the decision if a certified system
should be developed or not. It also produces a great opportunity for the possible customers
to express their opinions what functions or demands a system like this should have.
3.2 Design guidelines Based on the three standards mentioned above and discussions with electrical engineers at
DCAB the following guidelines should be considered when developing such a concept.
One thing that also must be considered is the additional demands on a system that includes
software and processors. The main guidelines for such a system are listed in a separate
paragraph below.
3.2.1 Guidelines
• To be able to cope with the high voltage logic found on some trains:
o The outputs should be controlling the external components via a relay. The
relay should be of solid state type to minimize fault probability and
maintenance as they do not include any mechanical parts.
o The inputs could also be equipped with relays to convert the voltage down to
a level suitable for the controlling equipment. However a relay capable of
handling about 110VDC control voltage is very large. Instead, a small
circuit could be used for the conversion. A design proposal of such a circuit
can be seen in annex A.
• A confirmation that the relays are working.
• Dual contactors on all mechanical relays.
o This is a demand to reach SIL 3 and 4 because of the mechanical parts that
can jam or be worn out.
• Any bus communication should be redundant and by being so the risk of a contact
failure disrupting the communication is minimized.
• Cables used for bus communication should be shielded to reduce interference
caused by electromagnetic interference (EMI).
• Bus communication shall be made with differential lines.
o The ground level can differ a lot between cars. In a logical system the
highest common mode voltage allowed would be less than the smallest
difference between:
7
Otherwise a signal transferred over a coupler could be misinterpreted.
Any differential receivers should therefore have a common mode
rejection greater than this voltage.
• Bidirectional bus communication should be configured with full duplex to eliminate
the risk of packet collisions.
• If there is a difference between units, they should be able to communicate between
each other in every possible combination.
• In the case of an error in the communication the devices should try to reinitiate it.
• The units in every car should be connected together with a bus that continues
throughout the train. This minimizes the signal delay between the ends of the train.
• The front switch on the coupler could be used by the unit to sense if it is an end unit
or not.
• In the case where a synchronization process takes place:
o It should be continuous to make sure the units never are out of sync.
o The units shouldn’t change the state of any outputs until synchronization has
taken place.
• If more than one unit with the ability to change the logical value of a signal is
mounted on the same signal cable, it must be done in a manner that prevents short
circuiting.
3.2.2 Additional guidelines for systems including software and processors
• Redundant processors shall agree before changing the state of an output.
• The information sent over the coupler should be checked for errors.
o This minimizes the risk of a bit error resulting in a misinterpreted signal
value.
o The profisafe protocol uses a 24 bit cyclic redundancy check8 for a 96 bit
transfer. As this is enough for SIL 3 it can be used as a good reference.
• The software should be written in a way that makes it easy to understand and
maintain.
o The program should be divided in small sub-programs and then connected
together in a main program.
o Every sub program should contain information of the author, version history
and a description of its function.
8 Cyclic redundancy check (CRC) is a method for error checking in data transfers. A CRC value is calculated by the transmitter on the data that is to be sent with a specific algorithm and sends it along with the data. When the data arrives, the receiver calculates the CRC value of the received data and compares it with the received CRC value. If the CRC values are identical, there is a very small probability that a transfer error occurred and the data can be considered correct.
8
4 Concept design proposals This chapter presents two designs made by the author that follow the previously stated
guidelines. The main functionality is to compress the signals in the train and send them
over the coupler via differential signaling. On the receiving end the signals are
decompressed via relays. The main difference between the designs is that one includes
microprocessors.
4.1.1 General description
Two units, one master and one slave, are to be mounted in every car. Both units are wired
on to the already existing signal cables via a logic converter but only the master is equipped
with output relays. This eliminates the risk of short circuiting. In case of a configuration
where not only the train computer controls the signals but also switches in each car, the
switches can be connected to the master’s inputs instead of the signal cables.
4.1.2 Communication
The units communicate with each other both in the car and over the coupler via redundant
serial buses. The buses are identically configured and utilizes CRC to prevent that a signal
value changes unintentionally due to a communication error. After every message, the
sender will await an acknowledgement that it was received properly. If an error is detected,
the unit will send the message again. Whether an error flag is to be set or not, and how a
communication error should be affecting the signal outputs is implementation specific and
is therefore handled separately in each case.
4.1.3 Change of a signal value
When the train computer changes the value of a signal the change is relayed through the
train. The units are configured so that if it receives a message from the bus over the coupler
it transmits it on the car-bus and vice versa. By sending the message as soon as it is
received the system makes sure that the delay is as short as possible. After receiving and
transmitting of a message the master unit changes the values of the outputs to correspond to
the ones stated in the message. Because the slave unit has received the same message as the
master it knows what changes to expect and either sends an acknowledgement if the
changes takes effect or an error message if the master fail to set its outputs.
Figure 4-1: Schematic
9
between train cars
Figure 4-2: Schematic
10
: Schematic of the integration of the units on a train
between train cars
4.2 Multiplexer
4.2.1 General description
This design is also separated in two different units. One unit called
called Output mux. The
connected to the existing signal cables, the
mux via output relays. The
via two separate differential lines.
user need to decide which of them should be active.
Output mux can be found in annex B.
Figure 4-3: Schematic of the integration of the units on a train
4.2.2 Input sampling
The Input mux consists of
multiplexers signal inputs are wired to the existing signal cables via Logic converter
counter controlling the multiplexer
control bits on the multiplexer. This is because the MSB of the counter controls the RESET
signal. As the counter value increases the multiplexer sequentially forward the signal values
to the signal data line. When all signals have been transmitted, the counter resets and the
sequence restart.
4.2.3 Clocks
All units have their own internal
however, the reference clock is divided by 16 before it is used by any components. The
skew between the local clocks is therefore at most a sixteenth of a period of the clock after
11
General description
This design is also separated in two different units. One unit called
. The Input mux is mounted in the car that contains the train computer
and therefore controls the train. In the other cars an Output mux is mounted. Both units are
connected to the existing signal cables, the Input mux via a logic converter
ays. The Input mux transmits data and reset signals to the
separate differential lines. In the case where there are multiple Input muxes, the
user need to decide which of them should be active. Schematics of both the
Output mux can be found in annex B.
: Schematic of the integration of the units on a train
Input sampling
consists of a multiplexer with a counter on the control inputs. The
signal inputs are wired to the existing signal cables via Logic converter
controlling the multiplexer, counter 1, has one more output bit than there are
multiplexer. This is because the MSB of the counter controls the RESET
As the counter value increases the multiplexer sequentially forward the signal values
All units have their own internal reference clock. They are all independent of each
the reference clock is divided by 16 before it is used by any components. The
between train cars
This design is also separated in two different units. One unit called Input mux and another
is mounted in the car that contains the train computer
is mounted. Both units are
via a logic converter and the Output
signals to the Output muxes
In the case where there are multiple Input muxes, the
Schematics of both the Input and
a multiplexer with a counter on the control inputs. The
signal inputs are wired to the existing signal cables via Logic converters. The
output bit than there are
multiplexer. This is because the MSB of the counter controls the RESET
As the counter value increases the multiplexer sequentially forward the signal values
clock. They are all independent of each other;
the reference clock is divided by 16 before it is used by any components. The
12
a RESET. This skew is small enough not to affect data acquisition. Before the clock is used
in the Output muxes they are delayed a fourth of a period to make sure that the signal data
from the Input mux has arrived.
4.2.4 RESET command
Every time the MSB of counter 1 goes high it sends a RESET command synchronized on a
delayed clock to the Output muxes and resets itself. The clock is delayed a fourth of a
period in relation to the clock controlling counter 1 to make sure that the right data from the
counter is available to the D flip-flop when the triggering flank arrives. The RESET
command is received by the Output muxes asynchronously and therefore no timing errors
will occur due to the delay. In the case of a contact error in the coupler the reset is wired via
buffers and pull-resistors to ensure that it always has a specific value. Pull up or down
resistors according to schematic is chosen to ensure that no outputs is changed when no
reset signal is present. Also a watchdog timer, counter 5, disables the outputs if a RESET
command hasn’t arrived one clock cycle after it is supposed to.
4.2.5 Output generation
Counter 4 controls the demultiplexer in the Output mux. The demultiplexer input is
constantly logical high and as the counter increases, it is distributed to the different output
flip-flops clock input. When a flip-flop senses a rising edge it reads the data signal from the
Input mux and sets its output accordingly. The demultiplexer’s enable input is wired to
counter 6 which together with an inverter and an AND gate enables it after a specified
number of RESET commands has arrived.
4.2.6 Risk analysis
Identified risk Probability Counter action
Contact failure in coupler 8*10-9 failures per hour Watchdog timer on RESET. Redundant transmission lines possible for both RESET and DATA.
Electromagnetic interference (EMI)
Different ground levels between cars
Common Differential receivers capable of handling a high common mode voltage (specific value depending on implementation)
Table 4-1: Risk analysis of multiplexer without microprocessors
4.3 Implementation The designs have been constructed with simplicity and modularity in focus. This means
that everything is based on standard components that will easily be found on the market,
with a very high probability to be replaceable with other components in the future. The
microprocessor multiplexer is intended to utilize a microprocessor with as many of the
needed functions already integrated as possible. The multiplexer without microprocessor is
intended to be implemented with logical circuits from the 74xx series or equivalent on a
13
PCB. It would also be possible to use an FPGA, however that would introduce the
additional work of creating document and validate a software.
14
5 Conclusion It is definitively possible to develop a multiplexing system to replace the wires in an
electrical coupler. Electrically it is not a very difficult task; the main obstacle is instead to
get the system certified to current standards. Therefore, before any actual design work is
done, it is recommended to conduct a thorough investigation of the intended applications
and the specific demands on standards. The two main possible application categories are
• On an existing train under the progress of being updated
• On a train currently under development
If the multiplexing system is to be applied to a train under development the
recommendation would be to look into TCN or another system with distributed I/O as it
would further minimize the cabling throughout the whole train. On the other hand, if the
multiplexing system is to be applied on an already existing train the development of a
system similar to one of those mentioned above should be developed. This can drastically
decrease the size of the coupler, in extreme cases, from 280 contacts to four at the same
time as more signals can be transferred.
5.1 Continuing work The next action to be taken is to decide if a prototype is to be built and in that case, which
of the above proposals to follow. The one most suited for a concept is the one without
microprocessor as it is a simpler design and the need of documentation is much smaller as it
doesn’t contain any software. This shortens the development time and the elimination of
software also makes it easy for any electrical engineer to maintain and operate the device.
However, if the purpose of a concept is to show the full ability of a multiplexing system a
system based on a microprocessor is more suited, as it can utilize a lot more functions,
which also can be added after the concept is produced simply by changing the software.
Even though wireless communication is not accepted as a safe transfer mode today, the
research in that area will most probably come up with a method that can be accepted. This
would completely minimize the need of signal pins in the coupler, and can be easily
implemented in any of the proposals given in this thesis.
There is also a need to take the decision if a thorough market research should be conducted
to utterly clarify the need of a new bus system and what standards it should conform to.
Two standards that have been mentioned when the idea of a multiplexing system was first
raised by Dellner Australia on behalf of a train manufacturer in Singapore are the
environmental conditions in the IEC standard 60077 and the EMC requirements in
EN50121. As these standards do not apply to the theoretical construction part they have not
been considered during this thesis but during the development process they can prove
relevant. The study concluded in two possible solutions, one using a PLC and serial
15
communication and one similar to the multiplexer with microprocessor presented in this
thesis. However, the customer in Singapore dropped the issue after seeing the proposal, so
there was no deeper research or development in this issue until this thesis.
At the end of the thesis, a very similar thesis report regarding this very issue was found in
the archives at DCAB. That thesis report concludes in a multiplexing system with a single
microprocessor, in and output relays, and the transmission over the couplers where
conducted over an inductive link without any error correction. The thesis where conducted
in 1996, so some of the techniques and materials used is a bit old, but the basic idea is still
applicable, and should definitively be studied in any continuing work with this issue.
16
6 Bibliography Nilsson Bo (2011) Customer Service Engineer DCAB (Verbal information)
Eriksson Conny (2011) RAMS/LCC Engineer DCAB (Verbal information)
Strandberg Per. (2011) Consultant at DCAB, RAMS/LCC-department (Verbal information)
Kirrmann, H and Zuber, Pierre A. (2001) The IEC/IEEE Train Communications Network
The International Electrotechnical Commission (2007). International standard IEC 61375-
1:2007(E), Geneva
Siemens AB (2009) Simatic Safety Integrated Practical application of 62061, Kap9.2
Correlation: SIL and PFHD of a SCRF Available at:
http://www.nwe.siemens.com/sweden/internet/se/produkter/industry/automation/as-
event/funktionssakerhet/Maskinsakerhet_2009/Documents/Maskinsakerhet_Simatic_Safety
7 Annex A
Figure 7-1: Schematic of the
This Logic converter uses voltage division to scale down the input voltage level. The
operational amplifier assures that
scales down the input signal to a level manageable for the subsequent electro
R4provide the operational amplifier with a voltage that act as reference for the switching
point between logical high and low. The resistance of R
equations below to suit the application. To eliminate the risk of
calculations simpler the value of R
the power supply has the same voltage level as the subsequent electronics logical high
level.

"
####
17
: Schematic of the linear logic converter
Logic converter uses voltage division to scale down the input voltage level. The
operational amplifier assures that the output is a distinct discrete signal.
scales down the input signal to a level manageable for the subsequent electro
provide the operational amplifier with a voltage that act as reference for the switching
point between logical high and low. The resistance of R2 and R4 is calculated with the
equations below to suit the application. To eliminate the risk of high currents and make the
calculations simpler the value of R1 is set to 220k and R3 to 10k
→ $ ∗
%
$& ∗
%
Calculating the highest possible effect in R1 at UIN = 135V and R2
614*+
between train cars
Logic converter uses voltage division to scale down the input voltage level. The
signal. Resistors R1 and R2
scales down the input signal to a level manageable for the subsequent electronics. R3 and
provide the operational amplifier with a voltage that act as reference for the switching
is calculated with the
. It is also important that
the power supply has the same voltage level as the subsequent electronics logical high
( 7-1 )
18
, -. ∗ 135 ∗ 0.000614 8345 ( 7-5 )

7.1.1 Simulations
7.1.1.1 UIN=0-135V sweep, USWITCH=60V
Figure 7-2: With calculated values according to the equations, R3=8.461k, R4=7.857k
0V 10V 20V 30V 40V 50V 60V 70V 80V 90V 100V 110V 120V 130V
0V
1V
2V
3V
4V
5V
0V
20V
40V
60V
80V
100V
120V
140V
0µA
100µA
200µA
300µA
400µA
500µA
600µA
700µA
V(uout)
19
Figure 7-3: With resistor values from E12 series, R3=8.4k, R4=6.8k
7.1.1.2 UIN=0-24V sweep, USWITCH=12V
Figure 7-4: With calculated values according to the equations, R3=57.894k, R4=10k
0V 10V 20V 30V 40V 50V 60V 70V 80V 90V 100V 110V 120V 130V
0V
1V
2V
3V
4V
5V
0V
20V
40V
60V
80V
100V
120V
140V
0µA
100µA
200µA
300µA
400µA
500µA
600µA
700µA
V(uout)
V(uin) I(R2)
0V 2V 4V 6V 8V 10V 12V 14V 16V 18V 20V 22V 24V
0V
1V
2V
3V
4V
5V
0V
4V
8V
12V
16V
20V
24V
0µA
10µA
20µA
30µA
40µA
50µA
60µA
70µA
80µA
90µA
100µA
V(uout)
Figure 7-5: With resistor values from E12 series,
7.2 Based on non
Figure 7-6: Schematic of non
This circuit uses an optocoupler for sensing the high voltage signal and converting it to low
voltage logic. In this circuit the resistors R
application. R1 and R
levels, the desired voltage level for switching from 0 to 1
The logical levels of U
• U0,L – Lowest voltage for a logical 0
• U0,H – Highest voltage for a logical 0
• U1,L – Lowest voltage for a logical 1
0V 2V 4V
20
With resistor values from E12 series, R3=56k, R4=10k
n non-linear circuits
Schematic of non-linear logic converter
is circuit the resistors R1, R2 and R3 needs to be matched for the
and R2 will depend on the input voltage UIN and its corresponding logical
, the desired voltage level for switching from 0 to 1, USWITCH
The logical levels of UIN are:
Lowest voltage for a logical 0
Highest voltage for a logical 0
Lowest voltage for a logical 1
6V 8V 10V 12V 14V 16V
V(uout)
V(uin)
=10k
needs to be matched for the
and its corresponding logical
21
• U1,H – Highest voltage for a logical 1
The voltage required to make the optocoupler switch from 0 to 1,UF and IF the
corresponding current to hold a logical 1. R3 will depend on the CTR of the optocoupler
and the current that flows through the diode at U1,L, IT. A buffer is wired on the output of
the optocoupler to enable a high output current and a distinct output switch and thus R3 will
also be affected by the voltage when the buffer switches from low to high, UB.
To find the resistor values choose a value for R2, typically around 200 and set a value for
$ ∗(% )
( 7-8 )
= ,%
( 7-11 )
Make sure that the current through the optocoupler, IF, is within its recommended values
= − ( 7-12 )
<= = <= − ( 7-13 )
If it fails, reiterate with different values on R2 and/or USWITCH until the demands are met.
Calculate IT
@ = AB$ ∗ ( 7-14 )
@ <= = AB$ ∗ <= ( 7-15 )
Calculate R3
( 7-16 )
Make sure that the buffer can handle the voltage and current at @
Calculating the effect produced in R1
, = ,C − ∗ ( 7-17 )
22
, = ∗ ( 7-18 )
Calculating the effect produced in R3 (UT is the voltage drop over the transistor at @ and
IB is the current drawn by the buffer)
, = (67 − @) ∗ (@ − D) ( 7-19 )
In simulations, IF is shown as Ix(Optocoupler:A).
7.2.1.1 UIN=0-135V sweep,U0,H=20V, U1,L=80V, UF=1V, CTR=75%, 5mA<IF<15mA
R2=180, R3=125
0V 10V 20V 30V 40V 50V 60V 70V 80V 90V 100V 110V 120V 130V
0.0V
0.1V
0.2V
0.3V
0.4V
0.5V
0.6V
0.7V
0.8V
0.9V
1.0V
0.0V
0.2V
0.4V
0.6V
0.8V
1.0V
1.2V
0mA
2mA
4mA
6mA
8mA
10mA
12mA
14mA
V(uout)
23
R3=120
7.2.1.2 UIN=0-24V sweep,U0,H=5V, U1,L=18V, UF=1V, CTR=71%, 5mA<IF<15mA
R2=180, R3=88.7
0V 10V 20V 30V 40V 50V 60V 70V 80V 90V 100V 110V 120V 130V
0.0V
0.1V
0.2V
0.3V
0.4V
0.5V
0.6V
0.7V
0.8V
0.9V
1.0V
0.0V
0.2V
0.4V
0.6V
0.8V
1.0V
1.2V
0mA
2mA
4mA
6mA
8mA
10mA
12mA
14mA
V(uout)
V(f) Ix (Optocoupler:A)
0V 2V 4V 6V 8V 10V 12V 14V 16V 18V 20V 22V 24V
0.0V
0.1V
0.2V
0.3V
0.4V
0.5V
0.6V
0.7V
0.8V
0.9V
1.0V
0.0V
0.2V
0.4V
0.6V
0.8V
1.0V
1.2V
0mA
2mA
4mA
6mA
8mA
10mA
12mA
14mA
V(uout)
24
R3=120
7.3 Conclusions
7.3.1 Based on OP-amp
The simulations show that changing the calculated resistor values to values present in the
E12 series has no significant impact on the system behavior. The biggest difference was
noted when using UIN=24V where the USWITCH is changed from 12V to 13V.
7.3.2 Based on non-linear circuits
The simulations show that this design is sensitive to changes in the resistor values. The
biggest difference is noted when using UIN=135V where the output does not change until
U1,L + 15V.
7.3.3 Comparison
The sensitivity of resistor values in combination with the trickier calculations is a con for
the non-linear design. However, the advantage with galvanic isolation provided by the
optocoupler makes it the better choice.
0V 2V 4V 6V 8V 10V 12V 14V 16V 18V 20V 22V 24V
0.0V
0.1V
0.2V
0.3V
0.4V
0.5V
0.6V
0.7V
0.8V
0.9V
1.0V
0.0V
0.2V
0.4V
0.6V
0.8V
1.0V
1.2V
0mA
2mA
4mA
6mA
8mA
10mA
12mA
14mA
V(uout)
25
26