CAN_DOC

June 2024
CAN Newsletter
Hardware + Software + Tools + Engineering
Functionally-safe J1939 communication

Cover story
Commercial road and off-highway vehicles and off-road

construction machines use often J1939-based application
layers. To meet the increasing demand on functional safety,
SAE has developed for CAN CC (classic) and CAN FD
dedicated protocols: J1939-76 respectively J1939-77.
www.can-cia.org
PCAN-MicroMod FD ECU
With CAN FD, a mix of digital and analog I/Os, and its tough 8 analog inputs
case, the PCAN-MicroMod FD ECU can be your solution for Resolution 16 bit
integrating custom accessories in utility and heavy duty Measuring range adjustable: ±2.5 V, ±5 V, ±10 V, ±20 V
vehicles operating under harsh conditions. 4 of the analog inputs alternatively usable as
analog output
The PCAN-MicroMod FD ECU can be configured with a Resolution 12 bit
Windows software via CAN. Besides simply mapping its Voltage range adjustable: 0 to 5 V or 0 to 10 V
I/Os to CAN messages, several function blocks for processing 2 frequency outputs
the data are available as well. Low-side switches (3 A)
Adjustable frequency range from 0 to 20 kHz
Alternatively usable as analog inputs with
voltage range from 0 to 60 V
Specifications Connections for CAN, I/O, and power supply via
High-speed CAN connection (ISO 11898-2) two 20-pole automotive connectors (Molex MX150)
Complies with CAN specifications 2.0 A/B and FD Plastic casing with increased Ingress Protection IP67
CAN FD bit rates for the data field (64 bytes max.) and flange
from 40 kbit/s up to 10 Mbit/s Operating voltage 8 to 32 V; suitable for use in
CAN bit rates from 40 kbit/s up to 1 Mbit/s 12 and 24 V vehicle electrical systems
Wake-up by CAN bus or by separate input Extended operating temperature range
4 digital inputs from -40 to +85 °C (-40 to +185 °F)
Pull-up or pull-down configurable E1 type approval in progress
8 digital outputs with High-side switches
2 outputs with 5 A and 6 outputs with 2 A
4 alternatively usable as a digital input or additionally Please note: The PCAN-MicroMod FD ECU is expected
for reading back the output level to be available in Q3 2024.
www.peak-system.com PEAK-System Technik GmbH

Take a look at our website for the Phone: +49 6151 8173-20
international sales partners. Scan the Fax: +49 6151 8173-29
QR code on the left to open that page. E-mail: info@peak-system.com
Cover story
Functionally-safe J1939 communication 4
Commercial road and off-highway vehicles and off-road

construction machines use often J1939-based application
layers. To meet the increasing demand on functional
safety, SAE has developed for CAN CC (classic) and
CAN FD dedicated protocols: J1939-76 respectively
J1939-77.
Imprint Engineering
Publishing house NMEA 2000 conformance testing and product
CAN in Automation GmbH certification 9
Kontumazgarten 3 Migrating from Flexray to CAN XL 20
DE-90429 Nuremberg AI evaluation boards with CAN connectivity 28
marketing@can-cia.org
Real-time communication: Part III - Temporal
www.can-cia.org dynamics of CAN-based systems 38
Tel.: +49-911-928819-0
In-vehicle device temperature acquisition 42
Fax: +49-911-928819-79
Reiner Zitzmann (CEO)

VAT-ID: DE812852184
CiA services
HRB: AG Nürnberg 24338 Standards and specifications 26
New CiA members by June 2024 32
Publisher
CAN in Automation e. V.
Kontumazgarten 3 Semiconductors
DE-90429 Nuremberg
Scalable processors for future software-defined
VAT-ID: DE169332292
vehicles 15
VR: AG Nürnberg 200497
CAN XL IP core 30
Editors
Olga Fischer (of)
Holger Zeltwanger (hz)
Trade shows
(responsible according Logimat: CAN deeply embedded in intralogistics
to the press law) equipment 12
pr@can-cia.org Passenger Terminal Expo: CAN in airport logistics 18
Embedded World: Edge-AI boards and computers
Layout with CAN interfaces 35
Nickel Plankermann
Media consultants
Julia Dallhammer
Feedback by readers
Tobias Kammerer
This is the second CAN Newsletter magazine issue, which integrates the CAN
Birgit Ruedel (responsible
Newsletter Online. The editors would like to receive readers’ feedback regarding these
according to the press law)
issues:
marketing@can-cia.org
◆ Which contents are you missing in general?
Table of contents
© Copyright ◆ Which topics do you like to see in next issues?

CAN in Automation GmbH ◆ How important is the “standards and specifications” column for you?
◆ How important are product category summary and trend articles for you?
The views expressed on CAN ◆ How important are new CiA member summary reports for you?
Newsletter magazine are not ◆ How important are new CANopen vendor-ID summary reports for you?
necessarily those of CiA e. V.
While every effort is made to The editors would also like to ask you to submit article proposals on new CAN-related
achieve total accuracy, neither succuss stories. In addition, we are hungry on technology-oriented articles.
CiA e. V. nor CiA GmbH can be Any response helps to improve the contents of the CAN Newsletter magazine.
held responsible for any errors or The editors depend on the discussion with readers. Please send your feedback to
omissions. pr@can-cia.org.
CAN Newsletter 2/2024 3

Functionally-safe J1939 communication
Cover story
Commercial road and off-highway vehicles as well as off-road construction

machines use often J1939-based application layers. To meet the increasing
demand on functional safety, SAE has developed for CAN CC (classic) and
CAN FD dedicated protocols: J1939-76 and J1939-77, respectively.
T his article discusses the two SAE J1939 standards for

functionally safe communications on CAN CC (SAE
J1939-76) and CAN FD (SAE J1939-77). For SAE J1939-
message is dropped and never actually transmitted.
Another example, conversely, is when a message is
received but cannot be added to a reception queue,
76, it describes the Safety Header Message (SHM) and with the result being that the message is dropped.
Safety Data Message (SDM) pairing approach used to ◆ Unacceptable delay refers to the failure to receive
communicate safety-related data from a producing safety a message within a permitted time window, thereby
application to a consuming safety application. In addition, causing a delay in the system’s response. This error
it details the features of the original version as published can occur, for example, if several messages are
in 2020 and lists the deficiencies of this version. Finally, communicated at or near the same time, causing
it details features of the revised version currently under congestion on the communication medium.
development that make up for these deficiencies. For SAE ◆ Insertion refers to the reception of a message from
J1939-77, the article describes the use of space allocated an unexpected or unknown source. This error can
for functional safety assurance information in the Multi-PG occur, for example, when two or more sources are
and FD Transport protocols to communicate safety-related transmitting the same messages.
data from a producing safety application to a consuming ◆ Masquerade refers to the inadvertent handling of a
safety application. In addition, it describes the three pro- message from a non-safety-related source as if it were
files currently under development that are tailored to meet from a safety-related source. This error can occur, for
different system needs while still meeting functional safety example, when a safety-related source, in addition to
requirements. transmitting its own messages, is forwarding messages
from a non-safety-related source. In this example, a
IEC 61784-3: Safety-relevant communication recipient inadvertently treats those messages as if they
principles were really from the safety-related source.
◆ Addressing refers to the delivery of a message to
The IEC 61784-3 standard defines various communication the wrong recipient, which nevertheless treats the
errors that can occur: reception as correct. This error can occur, for example,
◆ Corruption refers to the unexpected and undesired when a message is inadvertently addressed to a
transformation of a message such that the message multicast/broadcast address instead of to a unicast
received does not exactly match the message address.
transmitted. This error can occur, for example, when a Additionally, the IEC 61784-3 standard defines safety
device driver inadvertently swaps the byte order of a measures that can be used to detect such errors to achieve
part of the message, or when noise emissions disrupt the desired level of functional safety:
the bit patterns in communicated signals. ◆ Sequence numbers embedded in the message
◆ Unintended repetition refers to the unexpected and identify the position of the message relative to other
undesired repetition of a message. This error can messages in the same stream. They change from
occur, for example, when a device driver fails to update one message to the next in a manner such that both
its transmission queue after transmitting a message source and recipient can determine what the sequence
and so transmits the same message again. number for the next message should be.
◆ Incorrect sequence refers to the out-of-order ◆ Time expectations are when a recipient monitors
communication of messages in a sequence, e.g., the the time between two consecutively communicated
second message in a sequence gets received before messages to determine whether the period exceeds
the first message in the sequence. This error can a threshold; if it does, then the recipient assumes an
occur, for example, when messages in the sequence error.
get assigned different priorities before the messages ◆ Connection authentication is when a message has
are placed in a priority queue for transmission. a unique source and/or destination identifier for the
◆ Loss refers to the failure to receive a message that safety-related participants.
was transmitted. This error can occur, for example, ◆ Data integrity assurance adds redundant data (e.g.,
when a message is submitted for transmission to a cyclic redundancy checks, also known as CRCs) to a
queue that is already full, with the result being that the message to detect corruption in the message.
4 CAN Newsletter 2/2024

◆ Redundancy with cross-checking communicates the
safety data in separate instances, either in separate
messages or within the same message. A safety-related
recipient can then compare the data in both instances
and flag an error if differences exist.
Different data integrity assurance systems are designs,
in which communicated safety-related data use different
integrity mechanisms than those used by the communicated
non-safety-related data. This ensures that non-safety-related
messages do not affect a safety-related recipient.
Table 1 describes the coverage of various communi-
cation errors by safety measures employed in J1939-76 and
J1939-77.
Table 1: Coverage of communication errors by employed

safety measures (Source: Caterpillar, J1939-76, J1939-77)
Learn more
The adaptive machine

J1939-76: Two Safety Data Group (SDG) versions Your competitive
There are two versions of functional safety support advantage
specified in J1939-76. This support applies to Parameter
Groups (PGs) whose parameter data payloads range from
1 byte to 8 byte in length. Both versions use a Safety Data
Group (SDG), which consists of a Safety Header Message
(SHM1 or SHM2) and a Safety Data Message (SDM), to Today’s Adaptive machine
communicate safety-related data from a producer to a challenges solutions
consumer. The SDM, which is simply any PG to which an
SHM is associated, contains the safety-related parameter Mass Machines that make
customization to order
data to be used as part of a safety function. In contrast,
the SHM contains the following additional functional safety Product Instant changeovers
assurance data: proliferation on-the-fly
◆ one 32-bit CRC,
◆ one sequence number, Short product Easy reconfiguration

◆ the parameter group number (PGN), lifecycles via digital twins
◆ the destination address (DA) for point-to-point PGs,
◆ and the source address (SA) of the associated SDM.

To win in a world of mass customization, e-commerce,
Because of the need for two different messages, direct-to-consumer and omnichannel, it takes
J1939-76 specifies timing and order-of-transmission machinery that’s built to adapt. The first machinery
constraints to ensure that the right SHM instance appears concept that adapts to the products being produced
with the right SDM instance. and packaged! B&R enables adaptive manufacturing
The original publication of J1939-76 in 2020 specified through intelligent mechatronic product transport
what is now called the SHM1 version of functional safety integrated with robotics, machine vision and
support. Later analysis showed that the SHM1 version had digital twins.
br-automation.com/adaptive
some deficiencies with regards to the CRC coverage and
the size of the Sequence Number field, so SAE (Society
of Automotive Engineers) began with development of the
SHM2 version to correct those deficiencies.
B&R | A member of the ABB Group

in the SDM. The SA field is not in the payload, but it appears
Cover story
in the CAN ID and matches the value in the SDM. The six
least significant bits of the Sequence Number field are in the
first byte of the payload, while the eight most significant bits
of the field are in the second byte. The arrangement of the
CRC field is least-significant-byte-first.
The benefits of this solution are:
◆ Like in the SHM1 version, the CRC calculation in this
version covers the PG’s parameter data payload in the
SDM; however, it also covers the PGN, DA for point-to-
point PGs, SA, and Sequence Number fields in the SHM.
◆ The 14-bit Sequence Number in this version is larger
Figure 1: Format of the SHM1 payload specified in J1939-76 than that defined in the SHM1 version.
(Source: Caterpillar) ◆ Like the SHM1 version, a system can deploy this version
over either J1939-21 communications or J1939-22
Figure 1 illustrates the format of the SHM1 payload. communications.
The EDP (Extended Data Page), DP (Data Page), PF (PDU The drawbacks are:
Format), PS (PDU Specific), and SA fields are all bitwise ◆ This version employs a different CRC polynomial
inverted; the values of these fields before inversion match (labeled as CRC-32K/9 in [2]) whose Hamming distance
the values in the SDM. The arrangement of the CRC is slightly smaller than that used in the SHM1 version. A
field is least-significant-byte-first. This approach has the different polynomial was necessary to cover the larger
following benefits: amount of data.
◆ This version employs a CRC polynomial (labeled as ◆ Like the SHM1 version, this version doubles the
CRC-32K/10 in [2]) with a relatively large Hamming bandwidth needed.
distance for the expected payload size. ◆ Like the SHM1 version, this version has some relatively
◆ A system can deploy this version over either J1919-21 complicated timing requirements.
communications (based on CAN CC) or J1939-22 ◆ Like the SHM1 version, this version is not well suited for
communications (based on CAN FD). communications across routers.
◆ This version has been available for several years. ◆ This version is still under development.
But there are also some drawbacks in this version:
◆ The CRC calculation only covers the PG’s parameter J1939-77: Three profiles
data payload in the SDM; it does not cover the PGN,
DA for point-to-point PGs, SA, or Sequence Number There are three profiles specified in J1939-77 for functional
fields provided in the SHM. safety support. These profiles take advantage of the Multi-
◆ The 5-bit Sequence Number is relatively small. PG and FD Transport protocols specified in J1939-22 for
◆ This version doubles the bandwidth needed to use over CAN FD. These protocols can allocate a separate
communicate safety data. space in their messaging for cybersecurity and/or functional
◆ Some relatively complicated timing constraints due to safety assurance information for a PG’s parameter data. As
the need for two messages per SDG. a group, these profiles support PGs whose parameter data
◆ This version is not well suited for communications payloads range from 0 byte to 65526 byte in length.
across routers due to a dependence on link-local Each of the profiles provides the following functional
addresses as part of its connection authentication. safety assurance information:
◆ Either a 32-bit or a 64-bit CRC.
◆ A 32-bit Sequence Number.
◆ The length of the data over which the CRC is calculated.

In addition, two of the profiles provide a system-
specific connection authentication (DataID) that does
not depend on link-local addressing. The definition of this
authentication allows producers to communicate safety-
related messages to consumers through routers. All
three profiles have the disadvantage that they are limited
to J1939-22 communications and that they are still under
development.
The Profile #1 focuses on minimizing the amount
of functional safety assurance information required. To
Figure 2: Format of the SHM2 payload specified in J1939-76 accomplish this, the profile requires a fixed size for the PG’s
(Source: Caterpillar) parameter data payload; it also requires the incorporating
link-local address information in the data’s identification,
Figure 2 illustrates the format of the SHM2 payload. which limits its usefulness for communication through routers.
Unlike in SHM1, the EDP, DP, PF, and PS fields are not The resulting functional safety assurance data fits within
bitwise inverted; the values of these fields match the values 8 bytes.

The Profile #2 focuses on the following:
◆ Handling a PG’s parameter data payload that is of
variable length and that can be larger than 8 byte.
◆ Supporting communication across routers by not
relying on link-local addresses for connection
authentication.
The resulting functional safety assurance data fits
within 12 bytes.
Figure 4 illustrates the format of the assurance data
for Profile #2. The internal arrangement of all fields is least-
significant-byte-first. The Length field contains a count of
the bytes over which the CRC is calculated. This solution
Figure 3: Format of Profile #1 assurance data specified in offers the following advantages:
J1939-77 (Source: Caterpillar) ◆ This profile can handle a PG’s parameter data payload
whose length can range from 0 byte to 19 byte.
Figure 3 illustrates the format of the assurance data ◆ This profile is suitable for communications across
for Profile #1. The internal arrangement of both fields is routers due to its specification of a 24-bit data identifier,
least-significant-byte-first. The PGN, DA for point-to-point DataID, that provides connection authentication and
PGs, and SA fields all appear elsewhere in the Multi-PG that must be unique within a system.
messaging and so do not appear here. ◆ This profile uses the same CRC polynomial and
The benefits are: Sequence Number as that used in Profile #1.
◆ This profile has the smallest set of functional safety ◆ The CRC calculation in this profile covers the PG’s
assurance data of any profile. parameter data payload as well as the Sequence
◆ This profile consumes less space inside the trailer of a Number, DataID, and Length fields.
single C-PG (Contained Parameter Group, a part of the The drawbacks are:
Multi-PG protocol messaging) than the equivalent pair ◆ This profile cannot handle a PG whose parameter data
of C-PGs containing an SDG. payload is large enough to completely fill a CAN FD
◆ The Sequence Number in this profile is much larger data frame.
than that used in either the SHM1 or SHM2 versions. ◆ The scope of DataID definitions is specific to a system;
◆ This profile uses the same CRC polynomial as that there are no globally defined DataIDs.
used in the SHM2 version. The Profile #3 focuses on the following:
◆ The CRC calculation in this profile covers the PG’s ◆ Handling a PG’s parameter data payload that is of
parameter data payload as well as the PGN, DA for variable length and that can be much larger than that
point-to-point PGs, SA, and Sequence Number fields. supported by any other profile.
There are some trade-offs: ◆ Handling data that can only be communicated via the
◆ This profile requires that the PG’s parameter data FD Transport protocol.
payload is exactly 8 byte. ◆ Supporting communication across routers by not
◆ Like the SHM1 and SHM2 versions, this profile is relying on link-local addresses for connection
not well suited for communications across authentication.
routers. The resulting functional safety assurance information
fits within 17 bytes.
Figure 5 illustrates the format of the assurance data
for Profile #3. The internal arrangement of all fields is
least-significant-byte-first. The Length field contains a
count of the bytes over which the CRC was calculated.
The advantages are:
◆ This profile can handle a PG’s parameter data payload
whose length can range from 0 byte to 65526 byte.
◆ This profile makes use of a CRC polynomial (labeled
as CRC-64-ECMA in [2]) that results in a 64-bit
Cover story
CRC.
◆ The Sequence Number in this profile is the same as
that used in Profile #1 and Profile #2.
◆ This profile uses the same DataID as that defined in
Profile #2.
◆ The CRC calculation in this profile covers the PG’s
parameter data payload as well as the Sequence
Number, DataID, and Length fields.
Of course, there are also some drawbacks:
Figure 4: Format of Profile #2 assurance data specified in ◆ This profile has the largest set of functional safety
J1939-77 (Source: Caterpillar) assurance data of any profile.

◆ The CRC polynomial used by this profile is Conclusion
Cover story
computationally more complex than that of any other

profile. The communication errors and safety measures described
◆ Like Profile #2, the scope of DataID definitions is in IEC 61784-3:2021 serve as the basis for the functional
specific to a system. safety support as specified in SAE J1939-76 and SAE
J1939-77. These SAE J1939 specifications provide different
versions and profiles for this support over both CAN
CC and CAN FD, allowing safety-related applications to
select the appropriate version or profile that meets both
their systems’ needs and their functional safety
requirements. t
Figure 5: Format of Profile #3 assurance data specified in

J1939-77 (Source: Caterpillar)
(This article is based on the 18th international CAN

References
Conference (iCC) presentation by Travis Breitkreutz. The
complete paper is published in the 18th iCC proceedings
[1] IEC 61784-3:2021: Industrial communication networks –
2024; CiA, Nuremberg.)
Profiles – Part 3: Functional safety fieldbuses – General
rules and profile definitions
[2] P. Koopman. Best CRC Polynomials, https://users.ece.
cmu.edu/~koopman/crc/
[3] SAE J1939-21: Data Link Layer, May 2022
[4] SAE J1939-22: CAN FD Data Link Layer, September
2022 Author
[5] SAE J1939-76: SAE J1939 Functional Safety
Communications Protocol, February 2024 unpublished Travis Breitkreutz
draft
Caterpillar
[6] SAE J1939-77: SAE J1939 CAN FD Functional Safety
breitkreutz_travis_o@cat.com
Assurance Data, February 2024 unpublished draft
www.cat.com

NMEA 2000 conformance
testing and product
certification
(Source: Adobe Stock)

The NMEA 2000 product certification guarantees a certain level of quality and
interoperability for NMEA 2000 networks by ensuring that a device supports
a mandatory set of protocol services and hardware requirements. Developers
can carry out the certification themselves or be supported by engineers
experienced in the NMEA 2000 certification.
N MEA 2000 is a higher-layer protocol based on CAN

CC (classic) for marine applications. It is the de-
facto backbone for marine applications around the
◆
◆
Provide configuration services to change features such
as update rate, priority, and instance.
Operate with a supply voltage between 9 VDC and
world, integrating navigation, propulsion, lighting, power, 16 VDC.
switching, and entertainment devices. It always uses a ◆ DC and AC isolation of I/Os from the NMEA 2000
bit rate of 250 kbit/s and has a maximum of 50 physical connector.
devices on one network. Based on ISO 11783 (SAE
J1939), it specifies a set of standardized messages called
Parameter Group Number (PGN), each of them has a
unique number. There are PGNs available for a wide
variety of applications inside a vessel such as navigation,
propulsion, lighting, etc. One additional feature that NMEA
2000 requires on top of the ISO 11783 basis is the Fast
Packet Protocol. This is an additional transport protocol,
which can transfer up to 223 bytes in up to 31 CAN CC
frames. Larger vessels typically have one to three NMEA Figure 1: A typical installation for a NMEA 2000 network
2000 networks. The typical NMEA 2000 network has a (Source: Warwick Control Technologies)
backbone connected via T-pieces and drop cables down
to the devices. At each end there is a 120-Ohm termination These rules are defined in the NMEA 2000 specification
resistor which is an off-the-shelf component. Figure 1 package, which is available from the National Marine
shows a typical installation for a NMEA 2000 network. In Electronics Association (NMEA) on its website (nmea.org).
Engineering
a micro NMEA 2000 network, all cables use the 5-pin M12 To be able to state that a product is a NMEA 2000 device,
type connector. the NMEA requires that manufacturers put their products
NMEA 2000 devices have a set of rules that must be through the NMEA 2000 certification process. This is
followed to operate correctly. These include but are not broken down into two main parts. The first is concerned
limited to: with checking the requirements of the embedded software
◆ The bit rate shall be 250 kbit/s. in the device to ensure that its communications conform
◆ Provide a set of services to identify a device and its to NMEA 2000, e.g. timing, requests, responses, and
manufacturer as well as a list of transmit and receive mandatory services. The second part is a self-certification
PGNs. checklist for hardware requirements mostly e.g. con-
◆ Provide the Load Equivalency Number (LEN), which nectors, cables, PCB design, AC and DC isolation, and
describes how much current is drawn from the network. CAN signaling. These are described further in this article.

NMEA 2000 devices and implementations ◆ Declaration of storage and operational temperatures.
Engineering
◆ Details of CAN signal timing.

There is a lot of devices available on the market that claim to Companies wishing to develop NMEA 2000 devices
be NMEA 2000 compatible devices in some way. However, can carry out the certification of the product to NMEA 2000
in reality most of these devices do not actually meet all of the themselves by purchasing the NMEA 2000 certification tool
requirements of NMEA 2000. There is also a large number from the NMEA. An alternative route is to benefit from the
of open-source and private projects available online, which experience and guidance from the team at Warwick Control
aim to integrate into NMEA 2000. Again, these projects do Technologies with their NMEA 2000 conformance and
not support all of the requirements of NMEA 2000. When product certification services.
a device is a NMEA 2000 certified device, it means that it Whichever route is chosen, a Manufacturer Code
provides certain software services via the protocol and also (unique to company) and Product Code (unique to product)
the hardware meets certain requirements (e.g. DC and AC must be obtained from the NMEA.
isolation).
Certification service
Benefits of NMEA 2000 product certification
The main reason why companies use the conformance and
Product certification guarantees a certain level of quality product certification services from Warwick, is to benefit from
and interoperability for NMEA 2000 networks by ensuring the company’s expertise and experience. In general, there
that a device supports a mandatory set of NMEA 2000 are two levels of service available:
protocol services and hardware requirements. For device ◆ Pre-conformance test: This is recommended for
manufacturers, this means that they can state that their customers who need to check the main software
product is NMEA 2000 certified and use the NMEA 2000 logo conformance to the NMEA 2000 communication
in their marketing. requirements and the CAN signalling of a prototype
The NMEA 2000 product certification process partly device. This can also be extended to look at NMEA 2000
consists of automated software tests and partly of a self- hardware requirements if a customer is not confident in its
certification process to check on the hardware and other hardware design.
requirements. Both are specified in Appendix C of the NMEA ◆ Full product certification: This is recommended for
2000 specification package available from the NMEA. customers who think that their device is ready to go for
Automated tests are specified in Appendix C3 and full product certification with the NMEA. This includes a
use the bench setup shown in Figure 2. It uses a CAN CC check of both software and hardware of the device.
interface by Kvaser as the other device on the network, which The benefit of engaging Warwick earlier in the project and
connects to the Device Under Test (DUT) via a terminated using the pre-conformance test service is that the developer
bench-based CAN CC network. The NMEA 2000 certification will get early advice on issues that could delay development,
tool is used on a Windows PC and provides instructions for increase project costs, and delay the product release date.
initial conditions of each test. After setup, all of the NMEA The general benefits of the engagement include:
2000 message exchange is started by a press of a button. ◆ Engaged engineers are experienced with the NMEA
2000.
◆ Know how to analyse the test failures.
◆ Tests provided in addition to those required by the NMEA,

therefore helping to provide a high level of product for
customers.
Common issues
A common misunderstanding is that NMEA 2000 is simply

the SAE J1939 with parameters for marine applications
instead of those for trucks. From the experience of
Warwick Control Technologies, some companies simply
do this and somehow obtain the details of the PGNs
that they need from the Internet. When these are put through
Figure 2: Setup for NMEA 2000 automated software testing the software tests that are executed by the official NMEA
(Source: Warwick Control Technologies) 2000 certification tool, this often results in many failures.
There are also many commonly seen issues that lead
Appendix C2 of the NMEA 2000 specification package to a failure and the need for re-testing. These are not always
is a self-certification process and takes the user through a obvious initially and include such aspects as:
set of hardware tests and checklist items to ensure that the ◆ Sample point: The requirement is that it should be set
product meets the requirements. These include such aspects between 85 % and 90 %. However, often values of
as: 75 % or less are used. This is often due to the default
◆ Include measurements of DC or AC isolation between setting in the example source code that companies new
the NMEA 2000 connector and any other I/O that may be to CAN have obtained for implementation of their
present on the device. low-level CAN drivers.

◆ Automatic retransmission: In early implementations Final word
of CAN, the automatic retransmission of a CAN frame
upon corruption was mandatory. This became an option Whilst conformance testing and NMEA 2000 product
when Time Triggered CAN was introduced. To pass the certification ensure that a device meets the protocol
software tests for NMEA 2000, automatic retransmissions requirements, they do not ensure interoperability with other
should be enabled. This option is often disabled by devices. It is part of the testing regime required for a NMEA
default in the CAN drivers. 2000 product.
◆ AC/DC isolation of NMEA 2000 connector: The pins in The simpler devices usually are quite straightforward.
the NMEA 2000 connector need to be isolated from other However, vessels tend to have a multi-function display
I/Os. If a device is repurposed from another application, (MFD) for access to the information from NMEA 2000
often the isolation does not meet the NMEA 2000 devices and to control aspects of the vessel. This is one
requirements. of the more complex devices on the vessel. Entertainment
◆ Use of proprietary PGNs: The approach should be head units and remote controllers are other examples of
where possible to use an already available PGN from the complex devices needing to integrate with the vessel MFD
standard, even if not all fields are needed. The unneeded devices and can often be challenging. Interoperability
fields can be set to “unavailable”. Proprietary PGNs of new devices with the more complex devices is not
should only be needed when there is not a standardized guaranteed by the NMEA 2000 product certification
PGN available for the device’s purpose. The NMEA does process and therefore additional testing is recommended. t
strongly encourage companies in this situation to get
involved in the standardisation of new PGNs.
◆ Termination: Often a device is being repurposed
from another CAN application to NMEA 2000 and
may have a termination inside the device. The original
application is sometimes one that uses a PCB populated Author
termination resistor in the device. For NMEA 2000
applications, a device may not have a termination Dr. Chris Quigley
resistor. The termination resistor can only be a harness- Warwick Control Technologies
fitted component at each end of the backbone network enquiries@warwickcontrol.com
providing two terminations in total. www.warwickcontrol.com
BY ENGINEERS,
FOR ENGINEERS
We were founded by engineers for engineers, and take
pride in making CAN interfaces that work ‘out of the box’.
As connectivity in the world expands, so do we.

We are expanding our offerings beyond CAN to other
types of connectivity, creating a world of possibilities.
Join us on this journey

kvaser.com
CAN deeply embedded in intralogistics
equipment

The Logimat trade show 2024 saw nearly 70000 visitors. The international fair
for intralogistics solutions and process management took place in Stuttgart
(Germany) in March. Several CiA members and their customers exhibited
CAN-based products and equipment. Often, CAN networks were deeply
embedded and visible only on the second or even the third glance.
T he history of logistics dates back to the first ancient

civilizations. Already the Egyptian empire developed
transport and storage techniques to maintain a steady
Logistics-IQ has published a market research study
about AGVs (automated-guided vehicles) and AMRs
(autonomous mobile robots). According to the study, the
supply of food and basic commodities. Historically, installed base will exceed 2,7 million units in 2028. Around
intralogistics equipment was powered by humans and 670000 units will be sold in 2028.
animals. In the meantime, electric-powered equipment The evolution of AGVs/AMRs over the last 15 years
is state-of-the-art. The automatization of intralogistics has created a broad range of types and sub-markets on the
solutions is a relatively new trend. Autonomously driving factory floor, in the warehouse and, at home, not mentioning
vehicles and robots are increasingly in use. the special-purpose solutions for hospitals, hotels, and other
unusual applications. One of the key chasms in the world of
The intralogistics market is growing AGVs/AMRs is the safety requirement differences between
"service" AGVs/AMRs and "industrial" AGVs/AMRs. While
According to Next Move Strategy Consulting, a market- the intent is that no AGV/AMR harms a human, service
research company, the worldwide intralogistics market robots operate in the realm of the human beings. Working
2023 counted 43,13 billion US-$. It will grow to 115,92 billion environments for service robots might include operation
US-$ in 2030. This is caused by the increasing international in a grocery store, a retail store, a mall, a hospital, on the
trade. The intralogistics industry facilitates the movement, sidewalk, or in the home.
handling, and storage of materials, goods, and information
within a facility or warehouse. This includes also the Forklifts and AGVs
packing and shipping of goods. Intralogistics involves the
use of automated handling systems, conveyors, robots, The hype in intralogistics are the automated guided vehicles
sensors, and software. (AGV). They are used on the factory floor, in distribution
The Logimat trade show covers most of the centers, etc. In some cases, they have substituted classic
intralogistics solutions. More than 1600 exhibitors forklifts. Many of these forklifts use embedded CAN
presented their products and services. The Logimat networks to control the drive and lift functions. Forklifts are
organizer has established similar trade shows in China also becoming driverless. Movanis exhibited on the Logimat
(Shenzhen), India (Mumbai), and Thailand (Bangkok). a Still-based forklift using CAN-connected cameras by ifm

electronic. The boundary between forklift and AGV blurs.
Some companies name their products AGV lifters.
Most of the AGV suppliers do not talk about the used
control system and the applied communication network.
Behind closed doors, they tell often that CAN-based
networks are in the AGV. Especially, wheel-hub drives
provide CAN interfaces. Franz Morat Group offers such
products integrating CANopen drives by Dunkermotoren.
Gefeg-Neckar is another wheel-hub drive supplier. Its CAN-
connectable sub-systems are applied in minibar AGVs used
in hotels.
EBM-Papst showed at the Logimat the Argodrive
driving/steering sub-system, which comes with an
embedded CANopen network. It has been launched in
2021. In combination with drive controllers from various
manufacturers, the sub-system forms a functioning drive
system that enables free-range mobility for driverless
transport vehicles. The Argodrive sub-system comprises also
a steering angle sensor and two electrical drives featuring an
STO (safe torque off) functionality. The two motors contribute
towards steering, acceleration, movement and braking. Just
two Argodrives on opposite corners of an AGV guarantee
full omnidirectionality, two additional freely moving support
wheels ensure stability. Depending on requirements, any
number of drive systems can also be installed.
B-drives partnering with EBM-Papst offers the E-Wheel
sub-system, which features a double-axis controller coming
with an optional CANopen interface. The drives provide STO
functionality. The company has also integrated a redundant
encoder system. It is combined with the host controller from
Sick, which communicates with the two drives via CANopen.
Motors like those in the Argodrive also function as
sensors that detect a large number of conditions. Due to
condition monitoring, the necessary replacement of the CAN FD-Interfaces
wheel module can be announced before a malfunction
occurs. The sub-system can be used even for heavy loads Various Form Factors
and on inclines. The company offers the sub-system in light,
normal, and heavy versions for weight classes up to 100 kg,
PCI, PCI Express®, PCIeMini(HS), M.2,
300 kg, respectively 500 kg.
CompactPCI®(serial), XMC, PMC, USB, etc.
Highspeed FPGA Design

esdACC: most modern FPGA CAN-Controller
for up to 4 channels with DMA
Protocol Stacks
CANopen©, J1939 and ARINC 825
Software Driver Support

Windows©, Linux©, optional Realtime OS:
QNX©, RTX, VxWorks©, INtime® etc.
esd electronics gmbh

Vahrenwalder Straße 207
Figure 1: Four Agrodrive driving/steering systems, for D-30165 Hannover
example, in the heavy version allow a total vehicle weight of
up to two tons (Source: EBM-Papst)
Phone: +49(0)511 372 98-0
info@esd.eu | www.esd.eu
US Office: www.esd-electronics.us
Lafert, part of the Suitomo group, offers the Smartris The electric-powered AGVs and AMRs are often
Trade shows
AGV/AMR wheel sub-systems based on CiA 402 compliant equipped with batteries, which have a CAN CC (classic)
drives by Suitomo. The sub-system comprises a drive unit, or CANopen CC interface. In some cases, the CAN-based
a servo motor, a wheel, and a gear. STO functionality interface is also used to communicate with the charging
is provided, too. Two sub-systems integrated in an station. Multipowr located in Belgium launched at the
AGV shuttle enable an electronic differential drive Logimat the Buzzard80 charging station, which charges
system. Beginning of this year, the product comes in a wirelessly the battery. The communication between
new more compact version, measuring 120 mm in battery (for example from Varta) and charger is based on
length and weighing 0,3 kg. It is 80 % smaller than the CANopen FD. The protocol stack provider is Emotas.
predecessor.
The towing AMRs by Tractonomy are based on Save the date
CANopen drives by Nanotec. They are designed to tow
carts and waste bins, for example. They are equipped with Broken down by sector, 52 percent of Logimat 2024 visitors
360° lidar sensors. The vehicles stop, when the towed came from industry and 16 percent from wholesale and
cart is disconnected. Therefore, the drives feature an STO retail. The majority of industry professionals (57 percent)
function. were senior managers, who visited the tradshow to get a
picture of the intralogistics solutions currently available and
compare offerings. The full 38 percent of visitors came with
specific investment projects in mind. And 24 percent of the
visiting industry professionals awarded a contract during
the show or plan to do so in the near future. These figures
have been collected and evaluated by Wissler & Partner,
an independent market research institute located in Basel,
Switzerland. The next international Logimat trade show
will take place in Stuttgart (Germany) from March 11 to 13,
2025. t
hz
Figure 2: The E-Movekit is a complete CAN-connectable

sub-system with all components that are needed to drive
and control linear motion in AGVs/AMRs (Source: Ewellix)
For forklifts and AMRs with lifting functions, linear

electric-powered drives are used, in some cases.
Ewellix owned by the Schaeffler group has launched the
E-Movekit, to replace traditional hydraulic devices. The
linear drive comes with a CAN interface for integration
into the in-vehicle network. Magazino, a Jungheinrich
company, offers AMRs for material supply in assembly
lines and warehouses. The products make use of the
open-source Robot Operating System (ROS), a set
of software libraries and tools that help building robot
applications. There are libraries for CANopen drives
(CiA 402) available.
Some AGV/AMR manufacturers implement only
embedded CAN-based control networks. Safelog has
accompanied the CAN-based backbone network with
an ASi-5 bus system for I/O devices. Oceaneering does
it the other way around: Their Unimover AGV is based
on Profinet for functional-safety purposes and uses an
additional CAN-based network for other non-safety tasks
such as connecting I/O modules.
Eceon presented at the Logimat a truck-loading
and -unloading robot, looking like a pallet mover. You can
name it alternatively AGV with a lifting function. It uses
a CANopen CC network as the backbone. The supplier
claims a productivity increase, an error-rate reduction, and
lower costs compared with traditional pallet movers.

Scalable processors for future software-defined
vehicles
NXP has announced the S32N family of vehicle super-integration processors
to support the transition to future vehicles. The first family member S32N55
includes a CAN hub with four CAN XL and 24 CAN FD ports.
T he S32N family enables automakers to evolve their

vehicle electrical/electronic (E/E) architectures and
centralize cross-vehicle functions previously implemented
requirements of vehicles (see Figure 1). In current vehicles,
around 40% of the weight is electronics and wiring.
Reducing the number and weight of car components
as individual hardware boxes (ECUs: electronic control enables the automotive industry to offer lighter and more
units) into a central compute ECU. The latter can be a energy-efficient vehicles.
central vehicle controller focused on real-time applications NXP’s S32 Coreride central compute solution is
or a central vehicle computer that also provides applications based on S32N processors. The solution combines
processing. S32N compute, vehicle networking, and system power
The introduced processors can integrate dozens of management (such as the FS04 safe power management
cross-vehicle functions with mixed-criticality. The used IC) to accelerate central compute development. The
hardware isolation and virtualization technologies ensure solution was designed to adhere to the automotive industry
freedom from interference between functions. Supported standards for vehicle functional safety (ISO 26262),
by complementary system components, software tools, and security (ISO/SAE 21434), and reliability. When combined
pre-integrated software, the processors are designed to with software from company’s partner ecosystem, including
accelerate software-defined vehicle (SDV) development NXP’s pre-integrated isolation execution environments,
and reduce vehicle cost and complexity. The processors NXP will offer the holistic SDV (software defined vehicle)
offer many combinations of safe, real-time, and applications solution that can scale across the vehicle, and vehicle
processing cores. This meets central computing needs fleets.
ranging from real-time operating systems running
deterministic vehicle control to high-level operating First family member: S32N55
systems running vehicle management and OEM (original
equipment manufacturer) applications and services. NXP released the first S32N55 family member as the heart
of the S32 Coreride central compute solution. Former,
vehicle propulsion, vehicle dynamics, chassis control, body,
and other core vehicle functions have been implemented
as discrete electronic control units (ECUs), each with
their own micro-controller and wiring. Vehicle functions
can now be consolidated into an S32N55 processor with
multiple isolation execution environments, which allows
to reduce ECU hardware costs. Decreased material
and reduced weight also contribute to sustainability and
extended driving range. For carmakers, it results in lower
manufacturing complexity and time.
Figure 1: S32N processor is scalable across core vehicle

functions (Source: NXP)
Semiconductors
All S32N products integrate a CAN hub, an advanced

hardware security engine, and a multi-port TSN Ethernet
switch. Some processors also support Ethernet packet
acceleration, AI/ML acceleration, and inter-compute PCI
Express services.
Core vehicle functionality
Using S32N processors enables automakers to centrally

consolidate core vehicle functionality in a central compute Figure 2: S32N55 family targets centralized, safe, real-time
ECU. The latter can be custom-built to suit the diverse vehicle control in software-defined vehicles (Source: NXP)

Real-Time Unit 0 Real-Time Unit 1 Real-Time Unit 2 Real-Time Unit 3
(4 x Arm ® Cortex®-R52) (4 x Arm Cortex-R52) (4 x Arm Cortex-R52) (4 x Arm Cortex-R52)
Semiconductors
Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52

Split-Lock Split-Lock Split-Lock Split-Lock Split-Lock Split-Lock Split-Lock Split-Lock
Neon™ SIMD FPU Neon SIMD FPU Neon™ SIMD FPU Neon SIMD FPU Neon™ SIMD FPU Neon SIMD FPU Neon™ SIMD FPU Neon SIMD FPU
64 / 16 / 16 KB TCM 64 / 16 / 16 KB TCM 64 / 16 / 16 KB TCM 64 / 16 / 16 KB TCM 64 / 16 / 16 KB TCM 64 / 16 / 16 KB TCM 64 / 16 / 16 KB TCM 64 / 16 / 16 KB TCM
32 / 16 KB L1 Cache 32 / 16 KB L1 Cache 32 / 16 KB L1 Cache 32 / 16 KB L1 Cache 32 / 16 KB L1 Cache 32 / 16 KB L1 Cache 32 / 16 KB L1 Cache 32 / 16 KB L1 Cache
1 MB L2 Cache 1 MB L2 Cache 1 MB L2 Cache 1 MB L2 Cache
Ethernet Acceleration Memory Communication Real-Time Subsystem (CSR)
Ethernet Subsystem (NETC3) 48 MB Platform SRAM with ECC Communication Manager

TSN Ethernet Switch 256 KB Standby SRAM with ECC Cortex-M7 (lockstep) TCM 256 KB 3 MB SRAM
10 Mb / 100 Mb / 10 Mb / 100 Mb / 1x 32-bit LPDDR I / F with ECC and In-Line Encryption CAN Hub with 24x CAN / CAN FD 4x CAN XL
1 Gb / 2.5 Gb 1 Gb / 2.5 Gb (LPDDR4X Flash, LPDDR4X / 5 / 5X DRAM)
MAC MAC 8x LIN 1x 2-ch FlexRay®
2x Octal NOR Flash I / F Security Offload Timestamping
SerDes SerDes eMMC/SDXC NAND Flash I / F 1x PCIe Gen 4 Root Complex (x1 / x2)
SerDes SerDes
Foundation Subsystem (FSS) Security
System Manager Hardware Security Engine (HSE2) Control and Status
TCM 1 MB 2x PSI5 2x PSI5 Serial
Cortex-M7 (lockstep) Secure State Machines Secure Memory Secure Processing
7x I2C 3x 2-ch I2S
Power Control, Clocks, Reset Symmetric/Asymmetric Cryptos
8x SPI 5x UART
PMIC I2C Interface 2x LIN
Random Number Generators
eXtended Resource Domain Controller (HW isolation) 16-ch SENT GPIOs
System MMU (ATU) Resource Isolation Architecture
Enhanced Direct Memory Access (lockstep)

Timers
Physical Protection
FCCU, MBIST, LBIST 4x 4-ch FlexTimer
Debug Lifecycle
System and Watchdog Timers (1x per processor)
1x 1.5 V 12-bit SAR ADC (12-ch) E-Fuses Periodic Interrupt Timer
Figure 3: Block diagram of the S32N55 processor (Source: NXP)
With the “core-to-pin” hardware isolation and Debug engine for S32N55
virtualization technologies, processor’s resources can be
dynamically partitioned. Vehicle functions can be managed PLS Programmierbare Logik & Systeme (Germany) has
independently, including fault handling and reset. They can released the UDE 2024 version of the Universal Debug
receive independent software updates with the over-the-air Engine at the Embedded World 2024. This version supports
(OTA) upgradeability. the multi-core debugging and tracing for the S32N55 vehicle
The automotive-grade S32N55 processor integrates super-integration processor. The processor targets central
16 split-lock Arm Cortex-R52 processor cores running at 1,2 vehicle controllers that can consolidate dozens of electronic
GHz for real-time compute. The cores can operate in split or control units (ECUs) in software-defined vehicles (SDVs).
lockstep mode to support different functional safety levels up
to ISO 26262 ASIL D (automotive safety integrity level). Two
auxiliary pairs of lockstep Cortex-M7 cores support system
and communication management. Tightly-coupled integrated
memory and 48 MiB of system SRAM enable fast execution
with low-latency accesses. A firewalled Hardware Security
Engine provides a root of trust for secure boot, security
services, and key management.
An integrated CAN hub for internal routing of 24 CAN FD
networks and four CAN XL interfaces, time-sensitive networ-
king (TSN) 2,5-Gbit/s Ethernet switch, a and a PCI Express
Gen 4 interface help reduce wiring and system cost. Memory
can be expanded with LPDDR4X/5/5X DRAM, LPDDR4X Figure 4: UDE offers an intuitive user interface for
flash, and NAND/NOR flash interfaces. Functional safety and debugging and runtime analysis of applications when
security requirements are supported with memory error developing software for the S32N55 (Source: PLS)
correction and in-line cryptography. The S32N55 is comple-
mented with NXP’s system power management and vehicle UDE offers an intuitive user interface for debugging and
networking devices as the S32 Coreride platform’s central runtime analysis of applications when developing software
vehicle controller solution to accelerate customer designs. for the S32N55. The processor’s Cortex-R52 main cores
The FS04 safe system power management IC combined with and the Cortex-M7 auxiliary cores are all visible and can
PF53 scalable core supply for platform design, provides high- be controlled from one common, customizable debugger
efficiency power conversion, supports low-power modes, and user interface supporting multi-screen operation. This
handles power sequencing, informs the manufacturer. eliminates the need to open separate debugger instances
The S32N55 is sampling to lead customers. Enablement for each core. It is possible to define multiple views and
by NXP and a partner ecosystem, including the S32 Coreride switch between them to focus on a particular debugging
software partner ecosystem, include comprehensive board, task. Pre-defined configurations for the S32N55 also allow
enablement software, tools, and systems support to accelerate developers to quickly get started with their actual debug or
customer designs. test tasks.

Depending on the software partitioning of the appli- multi-core breakpoint feature eases debugging. The multi-
cations running on the processor and the particular debug core breakpoint is effective regardless of which core is
scenarios, the UDE’s synchronization behavior can be flexibly currently executing the specific code.
changed. The integrated run control management allows to The UDE also provides developers with a detailed
define a run control group for partial synchronization, for insight into the runtime behavior of the system and enables
example, in which only a subset of the cores is synchronized. comprehensive analysis based on the recorded trace data.
To control all cores individually, synchronization can also be These include profiling, call graph analysis, and code
deactivated. For applications where tasks are distributed coverage to verify the quality of software tests. t
across multiple cores and shared code is used, UDE's of
CAN SIC transceiver
CiA member Novosense has introduced the through the cable, which can lead to abnormal trans-
NCA1462-Q1 transceiver, which provides signal mission and even damage the transceiver. The
improvement capability (SIC) as standardized in ISO NCA1462-Q1 transceiver has anti-interference capability.
11898-2:2024 (formerly specified in CiA 601-4). According Even in extremely harsh electromagnetic environments,
to the Chinese company, the chip features a high EMC it can still maintain CAN FD communication, laying a
performance. This improves the network robustness and solid foundation for a robust communication. On the
optimizes system costs. other hand, EMI within the application system can also
radiate externally, thus affecting the transmission of
communication signals. The launched transceiver is
optimized for EMI based on an innovative patented
architecture and tested in accordance with IEC 62228-3,
as shown in the figures.
CAN FD star network multi-node communication

waveform without SIC transceiver
Figure 2: EMI test chart of NCA1462-Q1 without common

mode choke (Source: Novosense)
NCA1462-Q1 star network multi-node communication
waveform with SIC transceiver In addition, the CAN SIC transceiver achieves a
±8-kV ESD performance by optimizing the circuit
structure and layout area. Additionally, it provides a
reliable circuit protection against transient ESD threats
during automotive driving, while achieving better device
cost performance. With this EMC/ESD performance,
the transceiver can help engineers to eliminate common
mode choke or TVS tubes from peripheral circuits in
some designs. Furthermore, the VIO design as low as
Semiconductors
1,8 V can further save the use of LDO or level shifter in

the system, helping engineers to reduce overall costs.
The NCA1462-Q1 is available in SOP8 and DFN8
NCA1462-Q1 communication waveform at a bit rate of
8 Mbit/s packages and is compatible with other CAN HS (high-
speed) and CAN FD transceivers. The transceiver meets
Figure 1: Waveforms of the NCA1462-Q1 (Source: AEC-Q100, Grade 1 requirements, supports a wide
Novosense) operating temperature range of -40 °C to +125 °C, and
provides over-temperature protection. It supports TXD
In automotive applications with complex working explicit timeout protection and features remote wake-up
conditions, harsh electromagnetic interference (EMI) in standby mode. Samples of the device are available.
in the environment can be coupled to the transceiver hz

CAN in airport logistics
CAN networks are often deeply embedded and invisible. However,
(Source: Ametek)
on the Passenger Terminal Expo 2024, the tip of the CAN iceberg
became to the fore. One typical example are CANopen-connectable
drives in doors and access gates.
C iA member Dunkermotoren, a brand of Ametek,

exhibited on the Passenger Terminal Expo 2024
trade show its drive solutions for the door automation.
There are also CANopen drives by CiA member
EBM-Papst used in access gates, rotating doors, etc.
The company offers motors and drive controllers. As
They are used for sliding and swing doors. They provide the competitors, the modular drive systems are designed
CANopen interfaces compliant with the CiA 402 profile, for customer-specific solutions with generic building
internationally standardized in IEC 61800-7 series. Tobias blocks.
Johnston, key account manager building automation at
Dunkermotoren, explained: “Equipped with brushless DC The EN 17352 standard
motors and integrated or external motor control units, those
access gates provide additional safety for passengers. By publishing the EN 17352:2022 standard, suppliers of
Customized and parametrizable motion profiles contribute power-operated access control devices such as turnstiles,
to smooth and safe airport operation.” Due to its expansive swing lanes, and retractable lanes have been confronted
modular system including motors, gearboxes, encoders, with additional requirements and test methods. This
brakes, and control electronics, drive system combinations European standard covers safety in use of pedestrian
can be configured for any application. At airports or metro entrance control equipment used for normal access as well
stations, for example, entry systems such as access as in escape routes and emergency exits.
gates ensure that only authorized persons get access to Thus, manufacturers of such products are required
the airport terminal, or that the door to a platform is only to take further safety characteristics into account when
released after the metro has stopped. designing their product. Dunkermotoren helps to contribute
Different types of access gates, such as a one-way to the compliance of these requirements. Reversible
corridor or automated passport control, provide different gearboxes, for example, allow a manual opening for
installation spaces for the drive solution. The option to escape in case of emergencies, ensuring the vital safety of
select from both, planetary as well as angular gearboxes passengers. Additionally, the motor controllers can safely
such as worm or bevel gearboxes, allow Dunkermotoren switch off the drive torque so that no one is harmed on
to configure the suitable drive solution for each entrance their escape route. They are certified to EN ISO 13849-1
system no matter the space constraints. for Performance Level d (PL d) / Category 3.
When using appropriate control electronics from The EN 17352 standard applies also for revolving
the drive supplier, the motor can be set torque-free via doors. In 2021, Boon Edam has already introduced the
appropriate safety functions. Reversible gearboxes allow a Tourlock series of security doors. The Tourlock 180
manual opening for escape and rescue routes. (4-winged) and Tourlock 120 (3-winged) doors now offer
One of the customers is the Gunnebo Group compliance with the EN 17352 standard. These doors
headquartered in Gothenburg, Sweden. The group provides have been updated to provide security protection through
access gates and turnstile products. They embed several anti-tailgating and piggybacking functions. They use
drives by Dunkermotoren communicating via CANopen embedded CAN networks to communicate with sensors
with the host controller. The human machine interfaces are and actuators. t
not yet CANopen-connected, but some sensors are. hz

CAN in baggage handling
The company has set the goal of
implementing gapless baggage track-
ing hand-in-hand with customers
in line with IATA (International Air
Transport Association) Resolution
753. Roland Karch, Global Industry
Manager airports at Sick, said: “We
are market-leading in this application
with a market share of more than 70
percent.”
CiA member Nord Drivesystems
offers CANopen-connectable drives
to move the baggage conveyors.
The Logidrive product family comes
with Duodrive, Nordac-on, or IE5+
synchronous motors. The integration
(Source: Sick) of the drive into the gear unit housing reduces installation
space, the number of wear-prone parts, and thus also
CiA member Sick presented in Frankfurt on the the maintenance effort.
Passenger Terminal Expo 2024 its sensors able to There are also self-contained transport vehicles
sense baggage. The company offers different sensing moving and loading baggage carriers to aircrafts such as
technologies: laser, camera, and RFID. These products the Intrac pallet mover by Dimos, Germany. Such special
can be combined to detect and to identify the baggage. airport vehicles often use embedded CAN networks to
The communication between them is possible by means drive the vehicle and to control the lifting equipment.
of CAN networks. Multiple of such stations installed Some of them drive already autonomously.
along the conveyors allow an end-to-end tracking. hz
High-end connectivity with CAN and CAN FD
Telematics and cloud systems for IoT and Service 4.0
On-board units – from cost-saving entry telematics up to high-end modules.

Including updates-over-the-air, embedded diagnostic functionality and up to 4× CAN channels (CAN FD).
Off-board units – new high-end service VCI with Linux operating system, IoT functionality
and LTE mobile communication
Cloud solutions – IoT Device Manger | IoT Analytics Manager | Fleetmanagement.

Out of the box data mining and data logging – highly secure, comfortable and customizable.
IoT ECU – COMhawk® xt IoT VCI – COMfalcon® IoT
www.s-i-e.de 4× CAN, J1939, LTE, Integrated Multi-protocol 4× CAN, J1939, High-End Vehicle Multi-protocol
Wi-Fi, Bluetooth, flash-over-the-air support (J1939, LTE, Wi-Fi, LAN Communication support (J1939,
LAN, GPS, IOs functionality J2534, UDS, KWP, …) Interface J2534, UDS, KWP, …)
Migrating from Flexray to CAN XL
The VW Group has started to consolidate its EE architectures towards E32.0

for all types of cars, e.g. volume, premium, sports, etc. A solution is required to
have one communication technology for all vehicles, together with some other
handy features. This article describes the challenges and gained possibilities of
migrating from Flexray to CAN XL networks using the powertrain example, also
considering CAN FD and why it is only an intermediate solution.
F lexray [1] was designed to provide a high-performance

networking technology to cope with time-critical
communication demands in drive and powertrain control
loops between ECUs (electronic control units) distributed
throughout the car. Due to its complexity, limitations, and
involved costs, it only found its use in premium cars. IC
(integrated circuit) supplier support as well as OEM
(original equipment manufacturer) market acceptance was
not overwhelming.
Flexray started to enable time-triggered fast network
communication for use cases like x-by-wire. These use
cases required redundancy by concept, a synchronized
system by design to enable precise causal computation
and control loops in drive and the powertrain sub-systems
as well as the first steps for assisted driving. In order
to be safe and reliable, the system requires a fledged
specification, where almost no room for interpretation
should be given.
On the other hand, OEMs wanted design flexibility
regarding communication schedule, redundancy, topology
(e.g. active star couplers), etc. This led to vast amount of Figure 1: Asymmetric delay limitations at several topologies
configuration parameters and cross dependencies, making (Source: ISO 17458-4)
it severe complex and expensive to handle. Furthermore,
the physical layer was a challenge and had to be limited Finally, the daisy chaining of networks brought
due to some design constraints, such as the majority voting limitations in flexibility in production when it came to flashing
and configurable receiver thresholds leaving almost no and calibrating of ECUs, if not all ECUs were assembled to
room for flexible networks under generalized configuration the vehicle at the same time and simple parameters such
set-ups. as bus terminations could not be met.

The CAN FD option
One option to solve the above-mentioned cost problems

and the demand of system design flexibility is the CAN FD
[2] approach for faster bit rates and larger payloads than
CAN CC (classic) to still provide assisted driving and faster
causal computation and control loops in a reasonable way
for volume cars. But CAN FD does not feature a 256-byte
payload as Flexray provides. Additionally, the limitation to
500 kbit/s in the arbitration phase when using CAN SIC
transceivers and the maximum possible data phase bit Figure 4: Total length comparison between CAN CC,
rate of 5 Mbit/s are not sufficient for future application CAN FD, and Flexray with maximum bit rates, based on
demands. single-wire harness topology layout limits (Source:
Prior to the introduction of CAN FD networks in CARIAD SE)
VW Group cars (volume and premium) in 2019, many
investigations have been conducted to check the system Flexibility in network topology size in terms of number
behavior for networks running in different configurations of nodes and total line lengths was less compared to
between 500 kbit/s and the desired and reasonable bit CAN CC, but still better than Flexray, as production line
rate of 5 Mbit/s. It turned out, that networks in (multiple)star limitations could be omitted by star-based network topology
configuration for the most flexible use with 5 Mbit/s were layouts as well as customer-based demands to assemble
extremely sensible and prone to ringing in areas, where a ECUs in such network topologies.
system designer does not want them to be. Additional to the physical layer aspects, by the nature
It also turned out, that reducing network flexibility of the CSMA/CA bus-access method in CAN FD, to reach
down to daisy chains adds other types of issues (e.g. as less latencies and respective timing jitter, bus-loads
plateaus in the area of receiver thresholds), see Figure 2 had to be kept unusually low to achieve the required real-
showing the TXD and VDIFF signal of a high-ohmic ECU time behavior. Such bus-loads were required to be the
inside the daisy chain at a desired network topology size half of bus-loads from other “regular” networks together
about to turn over the 0,5-V threshold. with short payloads and limited number of frames. This
effectively limited the net bit rate, compared to what is
possible with Flexray by guaranteed bus access within a
given communication cycle time.
The CAN XL alternative
Considering the above-mentioned requirements and the

CAN FD limitations, it seems that CAN XL is a possible
alternative for Flexray. The maximum bit rate of up to
20 Mbit/s when using CAN SIC XL transceivers in FAST
Figure 2: Example of CAN-FD daisy-chain-based signal mode and the 2048-byte payload of the CAN XL data
plateau (Source: CARIAD SE) frames overcome the limitations of CAN FD networks, also
those using CAN SIC transceivers.
The total line length and number of nodes had to be A practical example shows, how a network system
stripped down to a very short end, which led to a refusal migration towards CAN XL can be done. This takes an
of CAN FD networks (without SIC transceivers) using already transferred communication system example from
5 Mbit/s in our systems. Thus, judging by the signal quality Flexray to CAN FD with CAN SIC transceivers. Sharpening
in those systems, being no alternative fallback or future- the focus, two example CAN FD networks (with SIC
proof solution. transceivers) are taken into account, which could represent
the split of a real-time and additional data communication
for a system design within a possible drive or power-train
sub-system.
Engineering
The goal is a real-time communication having these

two network topologies consolidated back on one bus,
which required to be split up from a Flexray system design.
This consolidated bus takes both communication
set-ups into account and is furthermore designed to work
from a physical layer perspective under all environmental
conditions and specification limits.
For this, two different analyses are done. The first
Figure 3: Node number comparison between CAN CC, CAN covers the consolidation of the communication on protocol
FD, and Flexray with maximum bit rates, based on single-wire layer by calculations and the second covers the feasibility
harness topology layout limits (Source: CARIAD SE) of the physical layer by simulations.

Protocol layer analysis 32 byte) or 16 (up to 64 byte). This limitation is not existing
Engineering
in CAN XL and byte-wise configuration is possible. The

One of these sibling networks sends real-time data for derivation of the formulas behind the calculation of the
fast control-loop precision. The other bus takes all the latencies, bus-load, and cycle jitter are based on a VW
"organizational" and "safety/security" related overhead. Group internal paper [3].
Within this context several meanings and system measures
are used: Physical layer analysis
◆ Real-time: A communication loop shall be less than
5 ms, where data is available between a given sender/ For the physical layer analysis, a reference model for mixed
receiver combination in a repetitive manner; signal (analog, digital) simulation has been developed,
◆ Message cycle: Repetitive transmission loop of a which can be configured in different manners to walk on the
message; physical layer specification limits, given in [2]. This includes
◆ Cycle jitter: Timing variation of the message cycle for exampl e output driver slopes and amplitudes, as well
based on best and worst case prioritized bus access; as transmitter and receiver asymmetries. Further PMD
◆ Latency: Time for the transmission of a message from (physical medium-dependent) sublayer and transmission
message buffer at a sending node over the network line components were taken from existing simulation
topology to the message buffer of a receiving node; ecosystems for CAN SIC and Flexray networks.
◆ Minimum latency: Latency value for the transmission System termination values have to be derived based
of a message with immediate and undisturbed bus on the number of expected ECUs and total line length in one
access; network topology to stay within physical layer specification
◆ Worst-case latency: Latency value for the limits. This is not being discussed in this article.
transmission of a message with worst case bus Within the physical layer context, several meanings
disturbance; and system measures are used:
◆ Bus-load: Average time of active message ◆ Signal integrity: Describing the necessary level of
transmission/communication time over a given time signal quality at the CAN pins analyzed by appropriate
period. qualification criteria;
All of the above-described system measures are ◆ Eye diagram: Single bit-wise qualification criterium
affected by the configuration of the active communication on the convoluted differential signal between the CAN
time on the bus and its detailed configuration, which shall pins.
be kept as low as reasonable and equally distributed. The signal integrity is evaluated by the measurement
Equal distribution at first is affected by the differing size of of the eye diagram at each sender/receiver combination in
message payloads and number of messages in conjunction. the network topology at each receiver, except the sender
Furthermore, equal distribution requires a proper scheduling itself. The eye diagram is configured by means of the
of message cycle loops in a communication set-up for a possible asymmetries, passive parasitic effects (based on
bus system inside the participating microcontrollers in the the expected network topology size in terms of number of
ECUs, meaning not all message cycles start at a defined ECUs and total line length) and tolerances in the system,
cycle start point in time. How to achieve this, will not be safety margins for EMC (electro-magnetic compatibility)
discussed in this article. robustness and bit-timing configuration.
The bus-load is affected by the number of messages The maximum eye opening is at the sampling point,
sent on the bus, as well as by the length of data to be which is configured to be in the middle of the bit plus one
transmitted in each message and the used cycle time time-quanta. The calculation of the eye-opening values
for each message. When using CAN CC or CAN FD, the mentioned above is not part of this article.
arbitration length via the use of 11-bit or 29-bit IDs is also Many analyses via simulation testbenches, lab,
affecting the bus-load. Using the CAN XL protocol, only and EMC chamber measurements have shown, that the
priority IDs with a fixed length of 11 bit are used [2]. Less switching between SLOW and FAST mode in CAN XL
messages for less arbitration overhead, lower message networks turns out to be less as critical as the pure signal
cycles and less message payload data basically lead to integrity at the FAST mode. However, the switching from
lower bus-loads. SLOW to FAST mode has still an impact on the asymmetry
The latency is affected by the number of messages on the first FAST mode bit pulse on the bus, a reasonable
sent on the bus and especially the length of data in each PCS (physical signaling sublayer) stimulus implementation
message, together with its priority identifier and message is necessary for simulation-based analysis.
cycle. Less messages for less arbitration overhead and
losses as well as smaller message payloads basically Communication set-up analysis
lead to lower latencies, respectively lower minimum and
worst-case latencies. Worst-case latencies together with Flexray as a time synchronous bus system has regularly
message cycle times directly affect the cycle jitter. a reasonable message cycle of 10 ms or less with
Message payloads using CAN FD are additionally theoretically no cycle jitter, meaning the worst-case latency
affected by the limited and not byte-wise configuration of basically being equal to the minimum latency.
the DLC. Payloads above 8 byte almost always added fill By the arbitration scheme used in CAN-based
bytes to conform the DLC configuration in case of signal asynchronous bus systems with prioritized bus access,
data being not a multiple of 4 (up to 24 byte) or 8 (up to cycle jitter is very likely to happen and worst-case latency

itself as well as cycle jitter must be limited, even when
keeping fast message cycles equal to or less than the real-
time window.
The premise for a possible consolidation shall be to
keep the quality of the real-time communication besides
the additional communication demand in one single
CAN XL network from the former sibling networks in
terms of the initial bus facts. Furthermore, there is room
for future extensions of the target CAN XL bus, if further
functionalities or ECUs are added to the network.
Figure 6: Comparison after initial merging to CAN XL

(Source: CARIAD SE)
This first comparison shows, that as expected, the

bus-load increases heavily due to the combined number of
messages from both initial CAN FD networks. This further
implies, that the bus-load overhead by arbitration increases
accordingly and the worst-case latencies for the data bus
related messages increase by some milliseconds. The
worst-case latency for the real-time data remains around
2 ms or slightly better and the cycle jitter keeps the same
Figure 5: Initial bus facts of the real-time bus (RT) and data dimension.
bus (DA) before merging to CAN XL (Source: CARIAD SE) To possibly gain more potential with CAN XL
communication, another paradigm in the communication
Figure 5 shows the initial facts from the two busses, set-up design should be analyzed in contrast to nowadays
using the CAN FD protocol with a 500-kbit/s arbitration rate parameter-based communication with probably unchanged
and a 5-Mbit/s bit rate and parameter-based communication message routings throughout a vehicle and equal cycles for
set-up. The calculations shown below ignore any non- application and message transfer. Two different optimization
cyclic messages in the communication set-up, as these approaches exemplarily show, how this can be achieved.
are defined to play no role for the computational loop Both initial busses operate with six ECUs, which plays a role
operation and repetitive exchange of vehicle state data at the discussed optimization steps.
during customer use. Examples are developer, diagnosis,
flashing, or similar messages. Optimization
The message amount sent to the data bus is around
four times the value of the real-time bus. The bus-loads The first approach focuses on reducing the bus-load by
are in the range of approximately 24 % to 29 % with a consolidation of signal data sent by one ECU into only a
approximately 18 % to 22 % overhead by arbitration. few messages. This means, that at first data with nearby
The worst-case latency at the real-time bus calculates message priority and cycle time is merged into as less
to 2 ms and at the data bus to 10 ms. The minimum used messages as possible.
message cycle at the real-time bus is 5 ms and the As a premise for the consolidation of signal data, only
maximum used message cycle is 200 ms, whereas the application-based data are merged. Other parameter data,
numbers for the data bus are 10 ms and 1000 ms. The e.g. for network management, remain untouched.
maximum used payload at the real-time bus is 24 byte or To reach a bus-load as low as possible, a good
less and at the data bus 64 byte or less. balance should be found with regard to which messages
The cycle jitter at the real-time bus reaches 30 % are consolidated into each other to:
especially at messages with low message cycle time and ◆ Not send too much data too fast and
low priority identifier due to many fast messages with very ◆ Not send too much low-priority data with too high priority
low message cycle times. In contrast, the cycle jitter at the This prevents too long messages with too high priority
Engineering
data bus is approximately the half of the value due to no on the bus and in general an equally distributed message
messages with very low message cycle time, blocking the size.
transmission of other messages with lower priority. Figure 7 (for 10 Mbit/s) and Figure 8 (for 20 Mbit/s)
The first step to one CAN XL bus is the raw merge show the comparison between the sum of both CAN FD
of the communication set-ups from both initial CAN FD networks (light grey bars) to CAN XL as raw merger (light
networks. With this raw merge, the arbitration rate will be violet bars) and as bus-load optimized set-up (dark violet
kept at 500 kbit/s and bit rates from 10 Mbit/s to 20 Mbit/s bars).
are analyzed. In Figure 6, the initial comparison of the The results after the bus-load-based optimization
raw merger from the sum of both CAN FD networks (light show, that the achieved bus-load is heavily reduced,
grey bars) to CAN XL with 10 Mbit/s (light violet bars) and compared to the raw merging of the CAN-FD networks and
20 Mbit/s (dark violet bars) is given. is even below the bus-load of one single CAN FD network.

Physical layer signal quality
Engineering
Additionally, the network topology must work under the

desired bit rates and further premises from the physical
layer perspective, such as the wire-harness topology
design needs to be as flexible as with CAN CC networks
avoiding daisy-chain or linear topologies.
The network topologies for drive and powertrain
sub-systems are by their nature mainly located in the
area around the moving axles of a vehicle, far from the
passenger cabin. Scalable electric powertrains exemplarily
have accelerating and decelerating parts at the front and
the rear axle, such as one or more electric motors and
Figure 7: Comparison at 10 Mbit/s after bus-load-based possibly further active differentials or torque distributors
optimization of merged CAN XL (Source: CARIAD SE) and braking system assembled. Finally, they are somehow
connected to a central computing instance.
Initial example topologies are built up as 5-Mbit/s
daisy-chain networks with six ECUs and a 20-m total line
length each and additional six inline connectors due to
exemplary modular part assembly.
Figure 9: Example daisy-chain network in modular set-up

Figure 8: Comparison at 20 Mbit/s after bus-load-based for CAN networks using SIC transceivers running at 5 Mbit/s
optimization of merged CAN XL (Source: CARIAD SE) (Source: CARIAD SE)
Consecutively the overall worst-case latency of the

data from the initial data bus improved to approximately
2,4 ms to 3 ms (depending on the CAN XL bit rate) and
for the data of the initial real-time bus to 1,4 ms to 0,9 ms
(depending on the CAN XL bit rate), which means that all
transmitted data has real-time latency and very stable
cycle jitter of even less than the half of the raw merged bus.
With the consolidation through the first approach,
the maximum payload reaches approximately 180 byte, Figure 10: Example star network in modular set-up for CAN
which is almost three times the maximum of CAN FD. The networks using SIC XL transceivers at 10 Mbits/s or
number of messages was reduced to nearly a quarter of 20 Mbit/s (Source: CARIAD SE)
the whole initial communication set-up.
The second approach focuses on reducing the worst- The yellow marked ECUs are high-ohmic and the red
case latencies even more. This is achieved by drastically marked ECUs are low-ohmic. As this network topology
reducing the number of sending application messages to system is a scalable powertrain example, low-ohmic
one. This implies, that all application data of each ECU termination points must still be set at the daisy-chain
is sent with the fastest message cycle and highest bus end-points, increasing wire-harness complexity, handling
priority from the original data set of the initial busses. The subsets of the complete system. Star-based topologies
results of the second approach slightly differ compared can be configured in a more flexible way keeping the
to the first approach. The overall worst-case latencies low-ohmic terminations stable at the minimum subset of
could be further reduced by approximately 0,5 ms. The network topology.
worst-case latencies for the real-time data remained the The remaining star-based network topology design
same. Due to the above-mentioned fact, sending all data for CAN XL is comparable to the CAN SIC network
with highest priorities and fastest message cycles, the topologies in terms of total line length and in-liners, due to
maximum payload increased to around 360 byte. the exposed areas and modular assemblies of the ECUs.
The bus-loads and cycle jitters remained the same at In parallel to the communication set-up analysis, the
20 Mbit/s, whereas slightly increased by 2 % to 3 %. This derived CAN XL network topology is executed in simulation
shows, that the “one message for all” approach improves set-ups from 10 Mbit/s to 20 Mbit/s and evaluated by the
best by increasing the bit rate. signal-integrity measures discussed above. The creation

of the eye-diagram signals is done directly by a simulation- References
model implementation to provide a correct synchronization
[1] ISO 17458 series: Road vehicles – FlexRay
of the given signals with an emulated PCS for worst-case
communications system. ISO, Geneve, Switzerland.
stimulus according to [2] as only a little overhead in the [2] ISO 11898 series: Road vehicles – Controller area
executable testbench without having a complex post network. ISO, Geneve, Switzerland.
processing. [3] Dr. Anna Engelmann: Method for analytical latency
Figure 11, Figure 12, and Figure 13 show the worst- evaluation of time critical messages sent of CAN,
case signal integrity found in the network topology with CAN-FD, CAN XL. CARIAD SE, Wolfsburg, Germany.
worst-case transmitter asymmetry and weakest transmitter
output driving values and an eye opening considering The physical-layer simulations show, that any of the
200 mV additional EMC safety margin on top of the bit rates up to 20 Mbit/s are visible, even under worst-case
worst-case receiver thresholds given by [2]. The differential conditions and flexible star-based network topology designs.
voltage signal is shown as the bold line, the synchronized
eye-diagram raw signals are shown as dotted lines. Conclusion
Summing up all of the previous analyses, using CAN XL makes

Flexray obsolete, reduces the number of necessary network
interfaces in central computing or zonal ECUs, reduces the
number of network interfaces in real-time sensing/acting
ECUs, reduces the number of wiring harnesses, omits
private direct connections between real-time sensing/acting
ECUs, and omits oversized automotive Ethernet-based
and switched multiple point-to-point connections in real-
Figure 11: Worst case signal integrity at CAN XL network time vehicle movement and control loops. This leads to an
topology at 10 Mbit/s (Source: CARIAD SE) overall reduction of development and system complexity,
material use, vehicle weight, and therefore to less costs. All
of the above-mentioned benefits lead to a lot of headroom in
network design in terms of number of nodes, total line length
and bit rates for being a future-proof technology.
The heavy improvement of flashing time through faster
bit rates and larger payloads just comes as a bonus factor
by the technology itself, as well as built in Ethernet frame
mapping, routing, tunneling. Adding functionalities can
Figure 12: Worst case 1 signal integrity at CAN XL network happen with a stable and powerful physical layer, without
topology at 20 Mbit/s (Source: CARIAD SE) adding more and more networks to the network architecture.
This helps even more when thinking about development
concepts such as software-defined vehicles (SDV).
Based on the analysis described above, the
communication set-up analysis in particular, focused on
the pure consolidation of parameters with the drawback
of probably sending some data more than necessary. A
PDU-based communication set-up as a further optimization
step must be analyzed. Only transmit data, when needed
Figure 13: Worst case 2 signal integrity at CAN XL network and not all at once furthermore reduces the calculated
topology at 20 Mbit/s (Source: CARIAD SE) bus-load and latencies giving even more headroom for a
stable and future-proof network design. t
According to the results it can be said, that at lower bit
rates the signal reflections implied by the network topology (This article is based on the 18th international CAN
layout can shrink bit amplitudes in the same bit after initiating Conference (iCC) keynote presentation by Marko Moch.
a slope, basically stressing the sample point. Whereas at The complete paper is published in the 18th iCC pro-
Engineering
the faster bit rates the signal reflections are jumping over ceedings 2024; CiA, Nuremberg.)
to the following bits. This further implies, that depending
on the number of sent bits of the same logical bit level
have a huge impact on the signal shapes, see Figures
19 and 20. Simulations with bit rates between 10 Mbit/s
and 20 Mbit/s show the overlapping of these effects from the Author
corner case bit rates, but still keeping outside of the eye.
Together with in-vehicle EMC measurements of similar Marko Moch
network topologies in terms of number of nodes and total line CARIAD SE (VW Group)
lengths, it can be said that the given example is capable of marko.moch@CARIAD.technology
running at any bit rate up to 20 Mbit/s. www.CARIAD.technology

Standards and specifications
This section provides news from standardization bodies and nonprofit associations
regarding CAN-related documents. Included are also recommended practices,
application notes, implementation guidelines, and technical reports.
ISO 11898-2:2024 and ISO 11898-1:2024 published

for these two standards, are withdrawn,
in order to avoid double-specifications
and not to infringe the ISO copyrights.
The withdrawn documents are CiA 601-4
(SIC transceiver), CiA 610-1 (CAN XL
protocol), and CiA 610-3 (CAN SIC XL
transceiver). Additionally, the CiA 604-1
(CAN FD Light) specification has been
withdrawn. It is now in the annex of ISO
11898-1:2024.
Unfortunately, not all final submitted
comments on ISO 11898-2:2024 have
been considered by the ISO Central
Secretariat (ISO CS). This is why CiA
has released a corrigendum (CiA 140).
This technical report proposes im-
The CiA 140 technical report provides ISO 11898-2:2024 provements and corrections, in order
corrigendum proposals (Source: CiA) to avoid misunderstandings and mis-
interpretations. In the current ISO
I SO has released new editions of the ISO 11898-1 (CAN,

Part 1: Data link layer and physical coding sublayer) and
ISO 11898-2 (CAN, Part 2: Physical medium attachment
11898-2:2024 standard, several figures and tables are not
consistent. The CiA 140 documents can be requested from
the CiA office. This service is free of charge. The technical
(PMA) sublayer) standards. These new editions contain all report is individualized by means of a watermark. The for ISO
three CAN protocol variants: CAN CC (classic), CAN FD 11898-2 responsible working group will revise the standard
(flexible data-rate), and CAN XL (extended data-field length) as soon as possible. In the meantime, the CiA 140 technical
respectively all PMA options. The PMA options cover report overcomes the weaknesses of the standard.
all kinds of CAN HS (high-speed) transceivers including There are also some minor final editorial comments
CAN FD, CAN SIC (signal improvement capability), and on the ISO 11898-1:2024 standard, which have not been
CAN SIC XL transceivers. Of course, also the low-power accepted by the ISO CS. They will be considered in the
and selective wake-up functionality is standardized in ISO next edition.
11898-2. The CiA documents, which have been submitted hz

Brief news
◆ SAE J1939DA: In April, the quarterly updated of the document to achieve compatibility with the
digital annex of J1939 has been published. It OSI model, the introduction of the T_Data abstract
contains suspect parameter (SP) and parameter service primitive interface to achieve compatibility
group (PG) specifications. with ISO 14229-2, and clarifications as well as
◆ DIN 14700: The German standard proposal for editorial corrections.
CAN-connected fire-fighting truck body application ◆ ISO TC347: In April, ISO has established a new
units has passed positively the final ballot. It is technical committee (TC) with the name data-
based on CANopen with some dedicated additional driven agrifood systems. The scope also includes
functions. This approach is also known as FireCAN. greenhouse automation. An ad-hoc task force to
◆ DIN 4630: The German standard for a CAN-based be established will evaluate the use of CAN-based
commercial vehicle body application network embedded networks. The initiative comes from
has been submitted to ISO for international South Korea and is supported by CiA.
standardization. The new work item proposal within ◆ Isobus plugfest: The nonprofit AEF association
ISO TC22 (road vehicles) is in ballot to be accepted has organized an Isobus plugfest in Houston (TX),
under the number ISO 25200. This standard needs U.S.A., end of February. Some 75 participants from
to be functionally extended regarding a secondary 20 companies tested 33 products on compatibility
display and an (electric) PTO (power take-off) to the ISO 11783 series.
manager. Additionally, there are some pending ◆ CAN XL plugfest: On May 16, CiA has organized
requests for new body application units, e.g. a CAN XL plugfest testing CAN XL protocol
hook-loaders. implementations and CAN SIC XL transceivers on
◆ ISO 15765-2:2024: ISO has released a new edition interoperability. The CAN XL implementations by
of the Road vehicles – Diagnostic communication Digital Core Design (Poland) and Peak-System
over Controller Area Network (DoCAN) – Part 2: (Germany) were tested for the first time. About 60
Transport protocol and network layer services experts participated in the plugfest.
standard. The main changes are the restructuring hz
CANopen Miniature Pressure Transmitter

CMP 8270
• Different accuracy classes i. e. 0.1 % FS typ

• Measurement of pressure and temperature
• CANopen DS301/DS404, supports CAN 2.0A/B
www.trafag.com/H72614
AI evaluation boards
with CAN connectivity

For four years, Nvidia offers several system-on-modules (SOM) with one or two
CAN FD interfaces. The CAN FD protocol controllers support the time-triggered
CAN (TTCAN) protocol extension. These SOMs are Jetson AGX Orin, Jetson Orin
NX, Jetson Orin Nano, Jetson Xavier NX series, and Jetson AGX Xavier series.
T he CAN protocol controller is part of Nvidia’s system-

on-chips. It supports CAN CC (classic) and CAN FD
data frame formats with 11-bit and 29-bit CAN-Identifiers.
◆ If the device logs a “No buffer space available”
message during transmission, enter the command
$ cangen -L 8 can0 -p 1000 to use the polling
In CAN FD mode the maximum data-phase bit rate is mechanism.
specified as 15 Mbit/s. Using the SN65HVD230 CAN The user can obtain higher data bit rates by configuring
transceiver piggy-back module by Waveshare enables the TDCR (transmission delay compensation register),
transmission rates up to 5 Mbit/s depending on the but should make sure that the transceiver being used is
used topology. These modules recommended by Nvidia able to support higher bit rates. Additionally, it is possible
are based on 3,3-V CAN FD transceiver chips by Texas to change the CAN parent clock on T194 platforms (only
Instruments. The nominal and the data-phase bit rates are on Jetson Xavier NX series and Jetson AGX Xavier series
configurable. The Jetson AGX Orin and the Jetson AGX boards). On development cards with the T194 SOC, the
Xavier board families come with two CAN ports accessible CAN parent clock is PLL_C. The PLL_C clock’s core clock
at the 40-pin header connector. frequency is set to 34 MHz. To set a higher clock frequency
To start CAN communication, the user needs to load or to obtain a bit rate of exactly 1 Mbit/s, the user can
CAN kernel drivers in the right order. To use the network, enable the PLLAON clock and make it the parent clock of
the developer first brings up CAN nodes on the network the MTTCAN protocol controller.
and installs a group of CAN utilities for testing. Then he/
she can transfer frames (loopback test) and debug the
network interface, if necessary. Several other debugging
techniques can be useful in appropriate cases:
◆ If a loopback test shows that the CAN controller is
working correctly and it is still not possible to send or
receive frames, try reconnecting the transceiver and
confirm that the connections are correctly made.
◆ Check whether all of the steps necessary to enable
CAN communication were done properly.
◆ Connect an oscilloscope and see whether the bus is Figure 1: Jetson Orin Nano evaluation board (Source:
behaving properly. Nvidia)

compatibility. It features an Ampere architecture GPU, Arm
Cortex-A78AE CPUs, deep learning and vision accelerators,
high-speed interfaces, faster memory bandwidth, and
multi-modal sensor support to feed multiple, concurrent AI
application pipelines.
“As AI transforms manufacturing, healthcare, retail,
transportation, smart cities, and other essential sectors of
the economy, demand for processing continues to surge,”
said Deepu Talla, vice president of Embedded and Edge
Computing at Nvidia. “A million developers and more
than 6000 companies have already turned to Jetson. The
availability of Jetson AGX Orin will supercharge the efforts
of the entire industry as it builds the next generation of
robotics and edge AI products.”
“With the global population expected to reach nearly
10 billion people by 2050, farmers have a steep challenge
Figure 2: CAN transceiver piggy-back (Source: Waveshare) of feeding the world and they can’t do it alone. With less
available land and labor, and many variables to work
Linux and QNX operating systems support Nvidia through, deploying and scaling advanced technology like
SOCs. Jetpack is a Linux-compliant software development autonomy is key to building a continually smart, evolving
kit (SDK) from Nvidia. Redhawk Linux is also available for and more efficient farm. Our fully autonomous tractor,
the Jetson platform, accompanied by Nightstar real-time featuring two Nvidia Jetson GPUs for quick and accurate
development tools, CUDA/GPU enhancements, and a image classification at the edge, will be on farms this year,
framework for hardware-in-the-loop and man-in-the-loop supporting farmers in overcoming challenges and providing
simulations. The QNX operating system is also suitable for for our growing world,” explained Jahmy Hindman, chief
the Jetson platform. There are success reports of installing technology officer at John Deere.
and running specific QNX packages on certain Jetson "As a recognized medical technology leader,
boards. Medtronic continues to innovate and advance solutions
to improve surgical patient care. We recognize the key
Developer kit for autonomous machines role for AI in digitization of surgery through quantitative
analytics and real-time clinical decision support systems.
Four years ago, the Jetson Xavier NX module, a platform The latest Nvidia Jetson platform brings us a new level
to develop AI applications, has been launched. Using of computational performance in the operating room and
cloud-native technologies, developers can take advantage enables us to advance intraoperative systems to better
of the module’s AI and compute performance in its credit- support surgeons, through data-enabled solutions,” said
card-sized form factor. Manufacturers of smart machines Dan Stoyanov, chief scientific officer at Medtronic Digital
and developers of AI applications can build and deploy Surgery.
software-defined features on embedded and edge devices One and a half year ago, the Jetson Orin Nano system-
targeting robotics, smart cities, healthcare, industrial IoT on-module that delivers up to eighty times the performance
(Internet of Things), and more. over the prior generation, has been introduced. It is also
Sertac Karaman, associate professor of aeronautics intended for entry-level edge AI and robotics. Now, the
and astronautics at the Massachusetts Institute of Jetson family spans six Orin-based production modules.
Technology, said: “The Jetson Xavier NX allows us to get This includes the Orin Nano — which delivers up to 40
higher performance within a tighter size and power budget, TOPS of AI performance in the smallest Jetson form factor
enabling innovative development of compact drones for a — up to the AGX Orin, delivering 275 TOPS for advanced
variety of practical applications. We believe the Jetson autonomous machines. The product supports multimodal
platform with cloud-native support is an important new sensors and addresses Robotics Operating System (ROS)
development to help build and deploy future generations of developers. ROS is a development environment, which
autonomous machines.” supports CAN-based communication including CANopen
The developer kit is smaller than a credit card (70 CC (classic).
Engineering
mm x 45 mm). It is powered by the Cuda-X computing The Orin Nano modules are available in two versions.
stack and comprises the Jetpack SDK. It combines a The Orin Nano 8GB delivers up to 40 TOPS with power
reference module and carrier board with a Linux software configurable from 7 W to 15 W, while the 4GB version
development environment. delivers up to 20 TOPS with power options as low as 5 W
Two years ago, the Jetson AGX Orin developer kit to 10 W. The Jetson Orin platform is designed to solve the
followed. It is intended for advanced robotics, autonomous robotics challenges and brings accelerated computing to
machines, and next-generation embedded as well as edge over 700000 ROS developers. Combined with the powerful
computing. The product delivers 275 trillion operations hardware capabilities of Orin Nano, enhancements in the
per second (TOPS), giving customers over eight times the latest Isaac software for ROS by Nvidia put more per-
processing power of its predecessor, Jetson AGX Xavier, formance and productivity in the hands of roboticists. t
while maintaining the same palm-sized form factor and pin hz

CAN XL IP core
(Source: Digital Core Design)

Digital Core Design (Poland) in
cooperation with DCD-Semi (Poland)
has unveiled its DCAN-XL IP Core
implemented according to the latest
ISO 11898-1:2024 standard.
T he DCAN-XL IP Core has been designed to bridge

the gap between CAN FD and 100-Mbit/s Ethernet.
It should enable faster high-speed communication and
One of the key features of the IP core is its backward
compatibility, providing support for CAN CC (classic),
CAN FD, and CAN XL protocols. This ensures seamless
reliable data transmission for automotive and industrial integration with existing systems while future-proofing
applications. The introduced IP core supports data rates designs for evolving industry standards.
up to 20 Mbit/s and provides a data field length of up to To establish a physical connection to the CAN
2048 bytes, making it suitable for demanding real-time network, external transceiver hardware is required.
applications. By maintaining the advantages and reliability DCD recommends to use CAN transceivers for bit rates
that CAN is renowned for, while offering enhanced speed below 10 Mbit/s and CAN SIC XL transceivers for bit
and flexibility, the DCAN-XL opens up different possibilities rates over 10 Mbit/s. The DCAN-XL utilizes a single or
for next-generation automotive and industrial systems. dual-ported message RAM connected via the Generic
IP core featuring CAN FD light
The CAN FD light specification is available in the ISO LLC MAC
11898-1:2024 standard (formerly in the CiA 604-1 speci- TRANSMIT

HIGH-PRIORITY
fication). Arasan offers a compliant IP core implementing BUFFER
TRANSMIT
CAN high
System SCHEDULER BIT

a CAN FD light responder node. The launched CAN FD Clock
TRANSMIT
LOW-PRIORITY
STREAM
PROCESSOR
light IP (intellectual property) core can be integrated BUFFER
in host processor using the AMBA-APB interface. The CAN

A RECIVE
configurable core supports programmable interrupts, P
RECIVE
BUFFER
ACCEPTANCE
FILTER BIT
TRANS-
CEIVER
B
frame acceptance filters, and frame buffering schemes. TIMING
The bit rate is configurable, too.

The offered IP core complies with the CAN FD light
CONTROL
CAN low
STATUS
REGISTER
responder specification in ISO 11898-1:2024. Additionally, it
BRP
supports the time stamping as specified in CiA 603 as well INTERRUPT
as TTCAN level-1 functionality as specified in ISO 11898-4.

CONTROL
According to the supplier, the IP core is optimized for Block diagram of the CAN FD light core (Source: Arasan)
Autosar and J1939 support. The provided listen-only mode
enables analysis of CAN FD traffic and automatic bit-rate tests after release. Besides other deliverables, the company
detection for performance measurement. The loopback provides synthesis scripts, exception lists, and timing
mode is intended for debugging and self-testing during reports as well as application notes. Sample firmware
integration and system set-up. with software drivers is available, too. CAN protocol and
Arasan delivers Verilog RTL source mode, a simplified trademark licenses are not included.
test bench with simulation models, to run the initial set of hz

Master Interface, enabling efficient message handling The DCAN-XL is designed in accordance to ISO
and streamlined data exchange. Integration with the host 11898-1:2024 and CiA 610-1 specifications, with support
CPU (central processing unit) is possible through the for CAN CC, CAN FD, and CAN XL frames. Its flexible data
32-bit Generic Interface, compatible with interface rates, Autosar support, and SAE J1939 compatibility make
wrappers such as AMBA, Altera Avalon Bus, and Xilinx it a versatile solution for a wide range of applications.
OPB Bus. Digital Core Design (DCD) is a provider of semi-
"The DCAN-XL represents a significant leap forward conductor IP cores, specializing in solutions for automotive,
in CAN technology, offering unparalleled speed, reliability, industrial, and consumer electronics applications. With
and flexibility," said Jacek Hanke, CEO of Digital Core a focus on innovation and quality, DCD's IP cores are
Design. "With its support for higher data rates and larger trusted by leading companies worldwide to power their
data frames, the DCAN-XL is poised to revolutionize most demanding designs. This year, DCD is celebrating
automotive and industrial applications, enabling the its 25th anniversary, operating in the IP core market since
development of advanced systems that were previously 1999. t
beyond reach."
The IP core is available in two versions: Basic and
Safety-Enhanced. The Safety-Enhanced version has been
developed as an ISO 26262-10 Safety Element out of
Context, providing additional safety mechanisms and
comprehensive safety documentation to meet the stringent
requirements of automotive safety standards. Third-party
audits validate the safety-related work products, ensuring
compliance with the ASIL-B (automotive safety integrity
level) requirements. The thorough FMEDA (failure modes
effects and diagnostic analysis) analysis provided by DCD Author
offers step-by-step instructions for seamless integration
and system-level safety analysis. Achieving ASIL-B Thomas Cwienk
readiness, this design is suitable for integration into Digital Core Design
automotive safety systems, with the option for higher ASIL tomeq@dcd.pl
level readiness. www.dcd.pl
CAN
Controller
IP Core
ISO 26262
ASIL D
New CiA members by June 2024
CiA services
The most asked CiA membership benefits are the access to CiA specifications
and the possibility to influence their development process. Currently, CAN
in Automation (CiA) has about 743 members. In the first five months of 2024,
35 companies have joined CiA.
A further important reason to be part of the CAN

community is to get in touch with other CAN product
and service providers. In particular, small and medium-
is specializing in customized software and hardware
development for single and multi-core systems.
JK Energy (United Kingdom) has been founded
sized companies need to be networked with partners to be in 2003 to develop solutions for resource-constrained
successful in the today’s world. hardware embedded in products. The specialists work with
embedded Linux and have their own-build environment.
Semiconductor suppliers The company’s network management technology has
been developed to monitor devices for automated setup
The market-leading micro-controller vendors supporting and operation. The provider also designs GUIs (graphical
CAN connectivity are members of the CiA community for user interface) for embedded systems with LCD panels.
many years. This applies also for CAN transceiver suppli- Avineas IT Consulting (the Netherlands) offers
ers. In the last years, other chipmakers joined the CiA as- software (product) development, services for change, and
sociation. This trend is continuing. There are more than 30 training for the technical implementation of processes with
members offering CAN-related integrated circuits. a (software) IT component.
Unisemipower (China) is designing, manufacturing, Megmeet Electrical (Germany) is a solution provider in
and selling power semiconductor chips with sales and the field of electrical automation, integrating software and
technical support offices in Shanghai, Beijing, Shenzhen, hardware development, production, sales, and services.
and the United States. The products are used in The business areas include industrial automation (e.g.
communication systems, information security, automotive drives, servo systems, and encoders), electric vehicles,
electronics, medical, industrial, and other application fields. railway solutions, power solutions, home appliance control,
Another new CiA member, Silvaco (U.S.A.) offers etc.
solutions spanning from simulation of material behavior A further CiA member Lachmann & Rink (Germany)
impacting semiconductor devices, to design and analysis is engaged in embedded software development for
of transistor circuits, and providing IP blocks for systems- customer’s hardware and the connection to the HMI
on-chips (SoC). The solutions are deployed by display- (human machine interface). The company is involved in
manufacturing companies, automotive OEMs (original development of Linux operating systems, firmware, device
equipment manufacturer), memory, 5G, and IoT (internet drivers, sensor solutions, embedded apps, user interfaces,
of things) providers. For example, the MultiCAN controller as well as application and business logic software. The
IP core supports up to six CAN CC interfaces supporting application areas span from the machine construction,
TTCAN (Time Triggered CAN) functionality as standardized Industry 4.0 and IoT environment, automotive, and
in ISO 11898-4. agriculture to building automation.
E.S.T.E. Technology (Italy) offers engineering
Development and engineering services solutions focusing on automotive, agricultural, heavy-duty,
and industrial sectors. Established in 2012, the company
Some 100 members provide development and engineer- deploys a group of entrepreneurs and researchers from
ing services regarding CAN-related hardware and soft- the Imamoter Institute, the Italian National Research
ware development. Additionally, many members offer sys- Council, and the University of Ferrara. The company
tem design and system integration services. operates throughout Italy and offers services ranging from
Development of circuit boards, layout changes, as hardware and software design, to the advanced on-board
well as the compilation or unbundling of circuit boards are diagnostics, functional safety, precision farming systems,
part of the service range offered by HTH Sinus Electronic etc.
(Germany). Prototypes and samples, as well as the actual Homatic Engineering (Denmark) is offering custo-
series, are manufactured by the same company. Together mized development and technological consultancy in
with joining CiA, the developer has also registered a automation, mechanics, hardware development, software
CANopen vendor-ID. development, and product approval. Application areas
Embedded Brains (Germany) offers expertise on include e.g. automation and robotics. Beside others, the
system development in the automotive engineering, tele- company offers PC applications and various drivers for
communications, and industrial automation. The company CAN.

CanBusHack (U.S.A.), established in 2010, is a are dedicated for the telecommunications, industrial,
reverse engineering services corporation specializing medical, and IT industries. Since its establishment in
in automotive embedded controller software and 2014, Shenzhen ACE Battery (China) offers solutions and
communication understanding. services for energy storage, including battery materials,
battery cells, battery modules, battery packs, storage
Sensing technologies providers systems, and recycling. To equip its products with CANopen
connectivity, the company has registered a CANopen
More than 130 CiA members develop and manufacture Vendor-ID by CiA. ChargePoint (Great Britain) is offering
CAN-connectable sensors. This includes single devices the electric vehicle (EV) charging solutions for passenger
as well as sub-systems for more complex sensing appli- cars, delivery vehicles, buses, trucks, etc. Therefore, the
cations. company has built an integrated portfolio of hardware,
A further new CiA member Matrix Elektronik cloud services, and support. Milwaukee Electric Tool
(Switzerland) is a provider of sensor technologies since (U.S.A.) is another provider of batteries, chargers, modular
1976. The manufacturer is specializing in sensor solutions storage systems, and power supplies.
used in applications with explosive and high-temperature Established in 2021, Hirschvogel E-Solutions
environments. The CANopen vendor-ID registration shows (Germany) is a part of the Hirschvogel Group, founded for
that the manufacturer is working on CANopen-based work in the electric mobility market. A separate business
products. area for bicycles and micro-mobility is being created with
Founded in 2005 Joral (U.S.A.) develops and the brand Aximo. The first series product, the Aximo drive
manufactures harsh-duty electronic devices (e.g. encoders, axle, is being used in a cargo bike from Cube. Further
inclinometers) for mobile hydraulics and industrial series for e-bike drive systems are in development. For this
applications. Its rugged position sensors provide total company, a CANopen Vendor-ID has been assigned, too.
electronic package encapsulation, LED status indicators,
and true non-contact coupling. For instance, the SGAM and
DGAM inclinometers utilize the sensor fusion technology
combining more than one sensing method. The IP67-rated
incline sensors take input by a gyroscope, accelerometer,
and magnetometer to provide a 3-axis output as well as
feedback for pitch, yaw, and roll. J1939 and CANopen are
the supported higher-layer protocols.
Sika (Germany) is producing sensors for measuring
flow, temperature, and pressure. Together with the
membership, it registered a CANopen vendor-ID, which
is cost-free for members. In addition, the Idec Alps
Technologies (Japan) has assigned a CANopen vendor-
ID. The company designs, develops, manufactures, and
sells sensors, security devices, and value-added human- Figure 1: Aximo Pedelec wheel hub motor with a 60-Nm
machine interfaces for industrial applications. peak torque (Source: Hirschvogel E-Solutions)
Novotechnik Messwertaufnehmer (Germany) pro-
vides sensors for linear and rotary measurements. The Optical Metrology Services (Great Britain) is a
devices are available with different measuring technologies measurement, inspection, and remediation company
and communication interfaces e.g., CANopen and J1939. working within the energy sector. The company’s crawlers,
Versions for redundant measurements, single-turn or multi- tools, and services support and deliver critical requirements.
turn encoders, compact-size devices, or rugged metal For example, it offers camera and laser inspection
housings are offered, too. services for hard-to-reach areas within energy production.
Klug Avalon Mechatronics (India) has joined CiA Services include cleaning and surface preparation, weld
in April. It provides load cells, limit switches, draw wire scanning, defect identification and remediation, pre-
sensors, wind speed sensors, inclinometers, as well as coating inspection, and data analysis. Pipes and welds are
HMIs mostly dedicated for use in cranes, lifting solutions, independently verified. The weld inspection services are
and material-handling equipment. mostly deployed within the oil, gas, renewable energy, and
CiA services
nuclear energy sectors.

Solutions for the energy sector
In the last years, the number of members providing batter-

ies and chargers has increased. This trend goes on. About
40 members offer CAN-connectable electric power solu-
tions. This includes also energy management systems.
For more than half a century, Benning Elektro-
technik und Elektronik (Germany) produces solutions Figure 2: The Agility.mini crawler is dedicated for pipes
for the conversion of energy in multi-purpose or energy from 200 mm to 400 mm internal diameter (Source: Optical
storage. Company’s power systems and battery chargers Metrology Services)

Another company in this sector is Megger (Sweden) remote controls, electronic control
CiA services
offering electrical testing and measurement services units, PLCs, push-button panels,
including monitoring solutions. The critical electrical operator panels, angular sensors,
measurements include insulation and ground resistance and actuators. For example, the
testing, dielectric testing, transformer diagnostics, TRS.ALMX is a safety angular incli-
dissolved gas analysis, partial discharge analysis, cable nometer sensor with two inde-
fault locating and diagnostics, smart grid testing, and pendent CAN interfaces. Another
more. The firm serves a range of industries, spanning device, the AL35R15 is an electro- Figure 4: AL35R15 is
utilities, manufacturing, maintenance, renewable energy, pneumatic distributor for moving an electro-pneumatic
heavy industry, transportation, etc. machinery and equipment. It sup- distributor (Source:
ports the CiA 401 device profile for Almec)
Welding and cutting equipment suppliers I/O modules.
Another kind of remote control is offered by telematics
Last year, CiA started the development for CANopen pro- systems. With the GPS-based telematics devices, Globtrak
files suitable for manual welding and cutting equipment. (Poland) enables monitoring of the vehicles (e.g. location
This is the reason, why several new members are related and speed) and vehicle fleets. Analysis of the collected
to this application. data reveals functions that could be managed in a more
Lorch (Germany) is one of the recently joined CiA efficient way. Globtrak offers a so-called “cloud tool” that
members participating in the SIG (special interest group) collects, processes, analyses data, and produces ready-
specifying the CiA 464 profile for manual arc welding and made reports. The company has also registered a CANopen
laser cutting. Two further companies also joined CiA to vendor-ID at CiA.
work in this group. These are EWM (Germany) developing
complete welding solutions (e.g. multi-process welding Further new members
machines) together with its customers. To digitize the
used welding technologies Some new members are original equipment manufacturers
and move towards the Industry (OEM). They are looking for CAN-connectable devices and
4.0, the company participates sub-systems featuring interoperability and partly exchange-
in the CiA SIG welding and ability. Some of these members are interested in dedicated
cutting. The same goal follows CiA application profiles, for example the CiA 417 series also
ESAB (Sweden). It provides known as CANopen Lift.
advanced welding equipment, Won Tech Won (South Korea) is a company working in
welding consumables, automa- the medical sector. Since establishment in 1999, it has been
tion, and digital solutions that researching and developing laser systems, for example, the
enable the everyday and de- Figure 3: PEK is a WON-PDT (photo dynamic therapy) laser for various cancer
manding welding work. The controller for several treatments.
company already uses CAN for ESAB power sources CiA has also assigned a CANopen vendor-ID to Swift
communication between the and is connected to them Home Lifts (Sweden) providing a lift concept for use in
welding equipment. via CAN (Source: ESAB) modern homes. The Swift Lite is a small house lift version and
the Swift Pro is the company’s residential lift with advanced
Remote control solutions design features, a dynamic touch display, and several options
for personalized settings.
Remote control is in some applications a requirement, Gunda Automation (Germany) is producing motion
especially in lifting devices, construction machines, etc. control solutions including stepper motor drives, servo
Additionally, there is an increasing demand on telematics drives, positioning systems, multi-axis systems, and linear
services for embedded CAN-based control systems. actuators. For instance, the Colibriservo drive is housed in
Cattron (Germany) offers remote control systems, a compact housing that is directly connected to the motor.
panels, machine stop systems, and telemetry monitoring The device is available in three performance groups covering
solutions with CAN(open) CC and J1939 connectivity. The torque ranges from 0,22 Nm to 20 Nm. The drive supports
products are dedicated for agriculture, industrial automation, the CiA 402 CANopen device profile for drives and motion
material handling, mining, mobile equipment, oil and controllers. t
gas, power and fluid solutions, rail operations, and water of
management.
Founded in 1988, IMET (Italy) is a designer and
Correction for “CANopen-certified by
manufacturer of industrial safety radio remote controls
used primarily in the construction, ecology, and concrete-
March 2024” article
processing sectors. The portfolio includes transmitters, In the March 2024 issue of the CAN Newsletter
receivers, and CAN-based wired remote control also magazine, we reported about the recent CANopen-
dedicated for explosive and other demanding environments. tested devices. It should be corrected that the XU
Since 2001, Almec (Italy) provides electronics and Endurance joystick series from the CiA member Sure
mechatronics including customized design, production, Grip Controls (Canada) are not CE-certified and not
and technical consulting. The products comprise radio SIL-2 compatible. of

Edge-AI boards and computers with CAN
interfaces
Figure 1: Artificial intelligence (AI) makes its way from the cloud to the edge (Source: Adobe Stock)
On the Embedded World 2024 trade show, several companies exhibited

edge-AI (artificial intelligence) products providing CAN connectivity. Such devices
allow computations to be done close to where data is actually collected, rather
than at a centralized cloud computing facility or an offsite data center.
T he Embedded World exhibition in Nuremberg

(Germany) is an important exhibition to get an idea of
the latest trends in embedded control technologies. Over
rejects. Dirt or scratches on the component, poor lighting
conditions in the work area, or highly reflective materials
such as copper make positioning difficult.
1100 exhibitors from almost 50 countries presented their Offering only CAN hardware interfaces on edge-AI
products, solutions, and innovations. "The resounding boards and computers is not convenient for system
response from the well over 32000 visitors from more than designers. In many application domains, the currently
80 countries emphasizes the significance of the Embedded applied CAN networks connecting sensors and actuators
World as the key meeting place for the embedded use standardized HLPs (higher-layer protocols) such
community," said Benedikt Weyerer, Executive Director of as CANopen, DeviceNet, Isobus, or J1939. The support
the trade show organizer. of HLPs would decrease the effort to adapt edge-AI
Several companies exhibited in Nuremberg so-called technology, especially for small and medium-sized
edge devices featuring processors deploying AI algorithms companies.
and AI models. These products with local AI computing On the Embedded World, many Taiwanese suppliers
capability are called edge-AI devices enabling real-time offered edge-AI products with CAN connectivity. According
data processing and analysis without constant reliance on to the International Trade Administration (TITA) and the
a cloud infrastructure. Taiwan External Trade Development Council (TAITRA),
In some AI applications, CAN-networked sensors feed Taiwanese companies manufacture over 90 percent of AI
Trade shows
the AI processor with data collected in the local (front-end) servers worldwide. This trend seems to go on for edge-AI
control systems. Application examples include in-vehicle products.
data for fleet management. Such vehicles could be road, Many of the edge-AI boards are equipped with Nvidia
off-highway, off-road (e.g. agriculture vehicles and forklifts), processors (see also article “AI evaluation boards with
rail, or automated-guided vehicles for intralogistics. Other CAN connectivity” in this issue). Diamond Systems offers,
examples are AI-powered machine or process control for example, carrier boards for the Jetson processor by
systems. An example is the AI-supported laser welding as Nvidia designed for industrial and military applications.
introduced in 2023 by Trumpf, a German machine builder. The Taiwanese company has married AI and the PCIe/104
To ensure that the weld seam is always in the right place, board format. The carrier boards provide dedicated I/O
the laser's sensor technology positions the weld geometry functionality, which includes one or two CAN CC interfaces.
precisely on the component – otherwise there is a risk of The PCIe/104 is the stackable version of the PCIe format.

The company provides additional PCIe/104 modules with the development and deployment of production-grade AI
Trade shows
two CAN CC (classic) ports. applications. Edge-AI products by Advantech come with
Advantech, another Taiwanese company, has ROS (robot operating system) support. ROS is a set of
launched AI acceleration modules and edge-AI computers software libraries and tools for robot applications, which
powered by processors from Nvidia and Hailo. Many of supports CANopen and some of its profiles.
these products come with two CAN CC interfaces. The The edge-AI computers by Neousys are based on
company is an Nvidia distributor of industrial PCs that Jetson processors by Nvidia. Some of them come with one
are certified with the AI Enterprise software platform for CAN CC port. They are mainly intended for camera-based
AI solutions. The Taiwanese products are water-proofed
and suitable for industrial and vehicle applications. Some
CAN XL on Embedded World housings are IP69K-rated. These products are suitable
for unmanned ground vehicles and drones. For example,
On the CiA booth, NXP, ST Microelectronics, and the SEMIL-2000 series, featuring IP69K protection, can
Vector demonstrated running CAN XL networks with operate in temperatures ranging from -40 °C to +70 °C.
attached oscilloscopes by Pico Technology showing It is designed for enabling autonomous applications in
the transmitted CAN XL data frames and the bit wave military UGVs (unmanned ground vehicles), mining trucks,
forms. Cast exhibited the CAN XL IP core by Fraunhofer or agricultural tractors. The FLYC-300 series based on the
on its booth and reported a high interest in CAN XL Jetson Orin NX processor comes with a built-in CAN CC
associated with the CANsec security protocol, which interface to interact with the flight controller.
is still under development within CiA. Peak-System Acrosser is another Taiwanese manufacturer of edge-
and DCD-Semi launched prototypes of their CAN XL AI computers, which comprise AI processors from Nvidia
core implementations on the trade show in Nuremberg, or Intel. Some of them are designed for AGVs (automated
Germany. guided vehicles), but they do not feature CAN connectivity.
But the Nvidia-powered computers provide even one
CAN FD interface.
The edge-AI computers by Aaeon (Taiwan) are
suitable for intralogistics devices such as AGVs and AMRs
(autonomous mobile robots). They use Intel processors
and many of them feature one CAN CC interface.
C&T Solution from Taiwan launched in Nuremberg a
series of edge-AI computers. Some of the family members
feature CAN CC connectivity. The real-time industrial PCs
are designed to withstand harsh-environment conditions.
Vecow located in Taiwan partners with Hailo AI
chipmaker. The company also offers products with AI
processors by Intel and Nvidia. The edge-AI computer
The Miriac module provides two CAN XL ports and up for rail vehicles applications comes with two CAN CC
to 16 CAN FD interfaces (Source: Microsys) channels. Other products like the EAC-2100 powered by a
Jetson Xavier processor come with one CAN CC interface.
Microsys launched the Miriac MPX-S32Z2 interface Forecr situated in Estonia has developed carrier
module with two CAN XL and 16 CAN FD ports based on boards for Nvidia processors. The company is an Nvidia
the S32Z2 micro-controller by NXP, currently available in partner as Diamond Systems. Their boards feature one or
engineering sample. The MCU features eight ARM R52 two CAN CC channels and several other connectivity ports
cores. The module measuring 82 mm x 50 mm does not including M.2 slots for expansion cards. The company
provide transceiver chips for CAN interfaces, they need offers also the Milbox-AGX ruggedized AI computer
to be added on a carrier board. supporting two CAN CC interfaces. It is equipped with
On the Embedded World trade show, Green Jetson processors.
Hills Software introduced RTOS (real-time operating Bressner, a German company, offers fan-less
system) software support for the Aurix MCU family by edge-AI computers with CAN CC interfaces. These
Infineon coming with CAN XL interfaces. The μ-velosity products use Nvidia Jetson Orin processors and measure
operating system is intended for SDV (software-defined 105 mm x 90 mm x 52 mm. This makes them suitable for
vehicle) electronic control units (ECU). The Aurix TC4x applications, in which space is limited.
micro-controller provides an upward migration path from
the previous TC3x family of ASIL-compliant automotive AI in healthcare
MCUs. The TC4x featuring CAN XL connectivity includes
a parallel processing unit (PPU) as well as several Edge-AI products are expected to be used in medical and
accelerators, in areas such as data routing, digital healthcare equipment. This seems to become a huge
signal processing, radar processing, and cryptographic market. However, there are risks and ethical impacts to
computing. According to the supplier, the launched be considered. The European Parliament has released
RTOS complies with the ISO 26262 ASIL-D safety level. a study on AI in healthcare. As in other AI applications,
hz there are two phases: First is the training phase, in which

Figure 2: Edge-AI computers are available in rugged
housings (Source: Neousys)
information is recorded, stored, and labeled. This includes

information from human doctors and scientific studies. In a
second step, the AI machine uses the intelligence gathered
and stored to understand the new data. In this phase, the CiA
edge-AI product uses inference to identify and categorize
the provided information, in order to provide a diagnosis. In
educational
such complex scenarios, the AI inference learning can be training
used to augment human decision making.
In many medical laboratory and healthcare equipment,
embedded CAN networks are applied. CAN-connected
sensors can feed the edge-AI products with dedicated
information and CAN-linked actuators can be controlled
by edge-AI computers. Advantech offers dedicated
edge-AI devices for this purpose as well as Seco. The
Get the CAN-related
SOM-Smarc-Genio700 board offered by Seco (Italy) is
powered by the Genio 700 processor from Mediatek (USA)
knowledge!
featuring multiple Arm-Cortex cores. This module comes
with a stand-alone CAN CC controller connected via SPI Online seminars Date Language
to the host processor. In addition, Seco has developed the CAN for newcomers 2024-06-18 English
Clea AI software platform. According to the company, it CANopen for newcomers 2024-06-19 English
is partnering with NXP and will support in the future the
eIQ processors. t
On-site seminar Date Language
hz
CAN and CANopen 2024-11-07 German
CAN interface module for sensors CiA technology days Date Language*
Chinese technology day 2024-06-12 Chinese
CiA member Microcontrol Chinese technology day 2024-06-14 Chinese
(Germany) has introduced *Depending on the speaker's language capability, the event is held in Chinese or English language.
the µCAN-sensor module
family featuring CAN CC CiA webinars Date Language
and CAN FD hardware. It CAN XL status and future 2024-06-06 English
can run locally CANopen (Source: Microcontrol)
Functional safety 2024-09-19 English
CC, CANopen FD, J1939-
21/71 protocol stacks or customer-specific higher-layer CiA profiles 2024-10-24 English
protocol software. These one-euro sized products are CANopen FD status and future 2024-12-12 English
available in customized board designs for implementation
directly into the sensor casing. CiA in-house seminars online
According to the supplier, the board features
sampling rates of up to 1 kHz. The one-channel
CiA engineers discuss your urgent CAN-related issues that
module can handle voltage signals in a range from
are currently of high interest with regard to your projects.
±10 V, current signals from 0(4) mA to 20 mA as well
as signals from strain gauges, thermocouples, and
Pt100 as well as Pt1000 sensor elements. Two-
For more details please contact
channel modules are available, too. hz CiA office: events@can-cia.org
www.can-cia.org
Real-time communication: Part III - Temporal
Engineering
dynamics of CAN-based systems
(Source: Embedded Systems Academy)

Within this article series, Olaf Pfeiffer from Embedded Systems Academy is
setting in perspective the timing requirements for different real-time capable
communication systems, such as CAN, CANopen, and real-time Ethernet.
I n this third article, the author examines the real-time

capabilities and limitations of CAN and CANopen in
greater detail. The Part I, published in the December 2023
no standardized solution that provides both. The CiA (CAN
in Automation) user’s group currently has multiple working
groups reviewing various aspects of both safe and secure
issue of the CAN Newsletter magazine, describes how to communication with CAN CC (classic), CAN FD, and CAN XL.
select the right real-time timeframe for certain applications.
The Part II, published in March 2024, shows the different Real-time capabilities of CAN
timeframes required by different applications and reviews
what this means for the communication system used. The CAN's real-time effectiveness is closely tied to its
last article "Part IV – From theory to practice: CANopen communication speed, and further affected by its priority-
source code configuration" shows which optimization based arbitration mechanism. Calculating CAN frame
options are typically available when working with CANopen transmission times is not a straightforward task; the time
source code, here, on example of Micro CANopen Plus from depends on both the number of data bytes and their content.
EmSA. This complexity arises because stuff bits may be added
The Controller Area Network (CAN) protocol serves as to a frame depending on its data. Therefore, the following
the foundation for numerous applications across a wide range determined values should be considered as approximations,
of industries, each with its own distinct real-time demands. providing a general sense of the scope at hand.
Prominent examples like CANopen and J1939 highlight the It's important to remember that your maximum bit rate
diverse adaptations of this protocol to meet specific needs. also depends on the physical topology of the cabling, and
It's important to note that the real-time requirements for these depending on your application, the total transfer required for
applications are not uniform across the board. While some a single control cycle might include two transmission paths:
applications require reaction times measured in milliseconds, one for input data to the control unit and another from the
many others operate effectively under more relaxed criteria. control unit to the outputs.
Factors such as physical constraints, network topologies, Though the author focuses on CAN CC here, most of the
and computational tasks play a crucial role in shaping these following considerations also apply to the CAN FD (flexible
requirements. As we explore tighter real-time constraints, data rate) and CAN XL variants. Both of these protocols
the complexity of communication configurations and code feature a dual bit rate mechanism, further enhancing their
handling increases. However, when real-time requirements data throughput capabilities. However, when discussing
are more relaxed, it opens up opportunities for simpler, more timing-related dynamics, most of the considerations outlined
streamlined system designs without sacrificing functionality here predominantly apply to the "nominal bitrate." This
or reliability. foundational bit rate essentially establishes the pace for
Although both safety and security have been addressed control information such as arbitration, acknowledgment,
with (by CANopen Safety and CANcrypt) there is currently and error signaling. For those using CAN FD and

CAN XL, it's crucial to be aware of the additional thousands of CAN frames per second. There is sooo much
complexities introduced by the "data phase bitrate," which room for communication that one can easily grasp that with
governs the transmission of the actual data. One of the key some well-defined parameters on how to use all this “space”,
concerns in these systems is determining the maximum one can very well design real-time capable systems.
duration a lengthy message might occupy the network
and how much longer this delay might be compared to the Table 1: Ballpark figures for transmission delays (Source:
longest CAN CC (classic) frame with 8 bytes. EmSA)
At its maximum speed of 1 Mbit/s, CAN CC allows for
the exchange of more than ten frames within a millisecond.
Conversely, at a modest rate of 125 Kbit/s, it averages around
one frame per millisecond. Beyond mere transmission times,
signals or frames can experience delays if higher-priority
communication is in the queue. To put it simply, the worst-
case transmission time would be the sum of the frame's own
transmission time and the delay expected from the longest
sequence of higher-priority traffic in the system. This assumes
that all communication happens on a single CAN network. If
signals need to be forwarded via bridges or gateways, delays
become longer and even more challenging to predict.
The system of message prioritization can be a double-
edged sword. However, there is a mitigating factor: by
strategically limiting the duration of sequential high-priority
traffic, even communications with the lowest priority can
be dispatched with minimal delay. This approach ensures
consistent and timely data exchange throughout the system. The next section of the table shows potential
Looking at CAN (and the FD and XL variants) by itself, transmission delays and depends on many factors.
it is clear that “as is” it is not deterministic. A single device Therefore, it is only a rough estimate for a specific use case,
producing high priority frames can block the communication you need to adapt it to your own use case. The first row
for all others. To make CAN deterministic, we need to ensure shows the delay even the highest priority will have, if the bus
a controlled frame triggering – when may which CAN ID be is currently in use (arbitration already started, transmitter is
used. To activate CAN's real-time capabilities, the following too late to join). Transmission has to wait, until the current
design goals should be considered. While these guidelines frame is completed. The second row shows an arbitration
may vary based on application specifics, they serve as a delay – if there are other devices also trying to transmit a
reliable starting point: frame, how long do we have to wait? Here is shown the delay
A) Aim to keep the overall busload at a level where even low- for five other frames currently pending for transmission and
priority frames have sufficient time to access the network. having a higher priority followed by a line of further delays, if
While the exact threshold can vary by application, the a throttle mechanism is used protecting from back-to-back
author’s initial recommendation is to stay below 75 % transmissions. Further on in this article will be reviewed what
busload (less if communication is purely change-of-state- can be done if the sum of delays shown is unacceptable in
based). your application.
B) Ensure that no individual node can generate an extended The last section of the table shows the potential
stream of consecutive messages. Some drivers offer a processing delays caused by executing various code on the
transmission "throttle" to limit the maximum transmission device handling the CAN communication. Here is assumed
rate. that a modern 32-bit MCU (micro-controller unit) with an
C) For those seeking finer control over transmission timing integrated CAN CC interface is used, running at 80 MHz
and sources, consider the SYNC mode of CANopen. This or faster. In such environments, the code execution directly
mode enables trigger messages, providing enhanced related to handling the CAN frames is typically marginal.
control over transmission schedules, allowing trigger Potential delays come from “what else is happening” on that
modes like those used by time triggered systems. MCU.
Translating this knowledge into real-world system
Engineering
Mastering temporal dynamics performance requires actionable strategies and

considerations. With the previously established benchmarks
After exploring the various use cases and their respective from Part I – seconds, 100 ms, 10 ms, and single milliseconds
temporal demands of CAN-based systems, you can – as the guideposts, let's review practical recommendations
imagine that matching CAN configurations to specific time for optimizing your CAN-based systems.
requirements is both an art and a science. Table 1 summarizes
some of the main numbers and factors to consider. The first Mastering CAN applications with response
section of the table gives a summary of the CAN timings and times beyond seconds
throughputs that you can expect at various bit rates – this is
all for CAN CC using 11-bit CAN identifiers. The fact is that In the domain of applications operating with delays
even at the lower bit rates we are still talking about potentially stretching into seconds or even minutes, designing

CAN-based systems to meet these response times is
Engineering
not particularly challenging. Interestingly, even a device

burdened by sub-optimal drivers or firmware might still be
suitable, as even sub-optimal drivers will even still perform
within 10 to 100 of milliseconds.
Figure 2: While many non-real-time operating systems

can still achieve a 100-ms response, it is advisable to lean
towards a real-time operating system (Source: EmSA)
can dominate the bus, posing risks of delays for lower-

priority messages. Effective strategies for avoidance or
control, such as limits on what each node can transmit per
timeframe or synchronized triggering, can be employed
to mitigate these bursts, ensuring more predictable and
Figure 1: For applications operating with delays stretching harmonious network communication, even as the system
into seconds, even a device burdened by sub-optimal scales.
drivers or firmware might still be suitable (Source: EmSA) While many non-real-time operating systems (OS)
can still achieve a 100-ms response, it is advisable to lean
However, when working with devices that rely on towards an RTOS (real-time operating system) in such
non-real-time operating systems, the challenge lies not scenarios (if an OS is required at all, many simple I/O
so much in countering communication delays, but rather devices typically do not have an OS at all). Using an RTOS
in upholding consistent performance and avoiding the aligns naturally with the demands of a 100-ms response
worst-case possible delay. Regular testing and thorough window. If a non-RTOS is chosen, rigorous and extended
monitoring are essential to ensure that these devices never testing becomes imperative to ensure the OS consistently
falter in allocating the necessary resources for seamless meets the desired response times under all conceivable
CAN communication. It is also crucial to proactively curb any operational circumstances.
potential system disruptions. Simple yet effective measures, Within this 100-ms response time framework, the
such as ensuring the absence of updates or other resource- software and firmware requirements remain relatively
draining operations during active communication periods, forgiving. Specific optimizations are often unnecessary;
can strengthen the system's responsiveness and reliability. even drivers or stack implementations deemed sub-
optimal in high-performance environments (for example
Mastering CAN applications with response not taking advantage of CAN controller hardware features
times of 100 ms for advanced filtering and buffering) can adequately serve
the purpose.
This domain is where the potential of CAN, in synergy with
higher-layer protocols like CANopen, truly comes into play. Mastering CAN applications with response
The CANopen PDO (process data object) communication times of 10 ms
mechanisms inherent in CANopen provide users with
flexible control, simplifying the configuration of message As we move into the 10-ms response time zone, precision
content and triggering. These PDOs facilitate real-time and control over every system component becomes
data exchange between nodes, optimizing communication essential. This is where detailed scrutiny of network data
efficiency. flow is essential.
At this response time, CAN-ID assignment and overall Time-triggered networks, optimized for hard real-
busload remain critical, but not overwhelmingly so, as they time applications, are often the preferred choice in such
are unlikely to cause delays approaching anywhere near demanding scenarios. The CANopen SYNC mode is an
100 ms. The system architecture should be designed such effective approach to mimic communication behavior
that even frames with the lowest priority have timely bus as used in time-triggered communication systems. By
access, ensuring their transmission within the stipulated utilizing SYNC triggering messages, it enables specific
timeframes. As we navigate this middle ground, it becomes nodes to transmit their associated PDO messages at
increasingly important to review potential high-priority precise moments, bringing predictability and consistency
frame bursts. Back-to-back high-priority transmissions to system communication.

While an RTOS might seem ideal for such tight timing The extreme precision required at this level means that
requirements, it comes with its set of challenges. An RTOS many system configurations may need to be hardcoded,
offers a range of configuration options, and managing potentially bypassing higher-layer protocol stacks like
these tasks requires careful coordination. Within the tight CANopen that would otherwise delay processing. This also
10-ms window, the process involves a sensor sending its helps avoid potential delays introduced by configuration
current data, a control device with an RTOS receiving and handling, ensuring maximum predictability. Every component,
processing this data, and then acting upon it. message, and byte transmitted on the network must be
However, simply implementing an RTOS does not judiciously managed. Given the stringent requirements, if a
guarantee the desired outcomes. Task prioritization 1-ms response time is a necessity for your application, it is
and configuration must align perfectly with the system's wise to review if other communication solutions beyond CAN
stringent timing requirements. Additionally, a detailed might be better suited to your needs. This domain requires
review of driver functionality, firmware, and stack significant commitment in terms of development time, testing,
structures is crucial. Potential issues, such as priority and optimization. If this endeavor is taken on, one should be
inversion where a low-priority message in the queue fully prepared for a time-consuming project journey.
might delay a higher-priority one, need to be addressed.
Highly optimized drivers can address priority inversion, Conclusion
but this may cause changes in transmission sequences.
A change in transmission sequences can be problematic The exploration of the temporal dynamics of CAN-based
for certain higher-layer protocols and needs to be reviewed systems has underscored the adaptability and capabilities of
carefully. the CAN protocol across various response time requirements.
For applications with response times extending beyond
Mastering CAN applications with response seconds, there is less emphasis on precise timing, and
times of 1 ms decisions regarding CAN-ID usage, higher-layer protocols
employed, or the operating system selected generally have a
Venturing into the 1-ms response time territory for CAN- less pronounced impact on performance.
based applications is akin to treading on the edge of the
protocol's capabilities. These applications truly push the
boundaries, requiring an unparalleled level of optimization
and attention to every detail.
At this threshold, conventional approaches and
tools often prove inadequate. Even some RTOSs, which
typically excel in managing real-time tasks, may struggle to
consistently adhere to this tight window. This necessitates
reliance on micro-controller-specific implementations,
where most tasks are handled directly within interrupt Figure 4: In conclusion, understanding the balance between
service routines, bypassing the typical layers of an application requirements, the inherent strengths of CAN,
RTOS. and the related temporal constraints is vital (Source: EmSA)
However, as we reach into tighter time constraints of

100 ms and 10 ms, system design considerations become of
greatest importance. These include total busload, message
priorities, and the strategic employment of functionalities like
CAN-open's SYNC mode. When navigating the demanding
1-ms response time domain, every element of the system
requires meticulous attention and may even prompt a re-
evaluation of the network system selected.
In conclusion, understanding the balance between
application requirements, the inherent strengths of CAN, and
the related temporal constraints is vital. It's this knowledge
that empowers CAN system designers to make informed
Engineering
decisions across diverse temporal scenarios. t
Author
Olaf Pfeiffer
Figure 3: Venturing into the 1-ms response time territory for EmSA (Embedded Systems Academy)
CAN-based applications is akin to treading on the edge of info@esacademy.com
the protocol's capabilities (Source: EmSA) www.esacademy.de

In-vehicle device temperature acquisition
Engineering
Hongke (China) is a company providing automotive technology solutions and

acting as a distributor of Peak-System and other CiA members, for example, in
China, Japan, and Korea.
W ith the advancement of the automotive industry,

the importance of monitoring and acquiring tem-
perature data from in-vehicle devices has in-
creased. Hongke provides a solution for monitoring
and acquiring in-vehicle device temperatures, utilizing a
thermocouple module to achieve accurate and real-time
temperature measurement and transmission, thereby
ensuring the safe and reliable operation of the vehicle
system.
The MU-Thermocouple1 CAN FD module from Peak-
System is a temperature measurement sensor based
on the thermoelectric effect. It converts the measured
temperature into electrical signals. It is widely used
in automotive, industrial, aerospace, and scientific
application fields. The module offers eight measurement
channels and supports both CAN CC (classic) as
well as CAN FD transmission. The device provides
different temperature range options with T, K, and J
thermocouple type models, depending on the specific
requirements.
Thermocouple modules in automotive

applications
To ensure the normal operation and safety of the vehicle

system in the automotive manufacturing process, it is Figure 1: Vehicle battery system temperature monitoring
necessary to monitor and collect real-time temperature data (Source: Hongke)
of various devices within the vehicle. The thermocouple
modules are applied in three main sections: Temperature monitoring of vehicle battery
1. Avoid damage from high engine temperatures: systems
Thermocouple modules can detect the temperature
of the engine, such as coolant temperature and oil For electric and hybrid vehicle manufacturers, monitoring
temperature, to avoid abnormal conditions such and updating the temperature of the battery pack in real
as overheating. This prevents engine damage and time is crucial to ensure the safety and performance of the
improves the internal performance parameters of the vehicle's battery system. To achieve this, the thermocou-
ECU (electronic control unit). ple module enables stable and rapid temperature moni-
2. Optimizing exhaust system performance: Replacing toring.
traditional temperature sensors with thermocouple By integrating the thermocouple module with CAN
modules not only prevents failures at high temperatures CC and CAN FD connectivity, the measured temperature
but also optimizes the performance of exhaust gas value can be packed into a CAN-based message, enabling
treatment devices (e.g., catalysts) and reduces exhaust real-time updates of the temperature change curve. This
emissions, thus ensuring the stable operation of the integration not only ensures the safety and stability of the
exhaust system. vehicle battery system but also provides manufacturers
3. Ensure safe operation of the braking system: with critical data for optimizing battery performance and
Thermocouple modules can be used to monitor the extending its lifespan.
temperature of brake discs and pads, ensuring the Thus, the electric and hybrid vehicle manufacturers
safety performance of the brake system at high speeds can monitor and maintain their battery systems, ensuring
or under high load conditions. peak performance and safety for their customers.

MU-Thermocouple1 CAN FD
Depending on the product version, eight connectors
for thermocouples of the types J, K, or T are available
for temperature measurement. Temperatures can
be captured in Celsius, Fahrenheit, or Kelvin and
processed with an individual scale and offset. The
CAN communication is done via a D-Sub connector.
The measuring unit supports CAN FD with data bit
rates up to 10 Mbit/s and is at the same time downward
compatible to CAN CC (classic). With the product,
temperature measurement can be integrated directly
into automotive test benches or industrial plants using
CAN FD communication.
A router for the conversion from CAN CC to
CAN FD is no longer necessary. The configuration of
data processing, CAN communication, and LED indication
is done with the company’s free Windows software. Several
devices can be operated and configured independently
on a single CAN network.
of (Source: Peak-System)
Sensor monitoring in extreme conditions Conclusion
The thermocouple modules enable real-time temperature

data acquisition and monitoring of key in-vehicle systems
and devices. It enhances the safety and stability of the
vehicle system and also offers flexible configuration
options to adapt to different environmental and application
requirements. By leveraging the capabilities of the
MU-Thermocouple1 CAN FD, engineers can monitor
vehicle sensor temperatures in extreme environments,
ensuring reliable performance and safety of the
vehicles. t
Figure 2: Simultaneous temperature measurement of eight

sensors (Source: Hongke)
Monitoring the temperature of vehicle sensors in hot, high-

pressure, and extremely cold environments is crucial for
ensuring their reliability and performance. Engineers face
significant challenges in such conditions, as temperatures
can vary widely and affect the sensor performance.
The MU-Thermocouple1 CAN FD addresses these
Engineering
challenges with an operating temperature range of -40 °C

to +85 °C, an aluminum casing, a measurement accuracy
of 0,2 %, and support of multiple parallel acquisition
channels. Thus, the module is capable of measuring the
temperature changes of multiple sensors simultaneously.
This allows engineers to make on-site adjustments Author
and controls, ensuring the stability and high quality of
the production process. By monitoring the temperature Riley Shao
of critical systems and devices in real-time, engineers can Hongke Technology
anticipate and mitigate potential issues, thus improving the shao.yue@hkaco.com
overall safety and reliability of the vehicle system. www.hkaco.com

The nonprofit CiA organization promotes CAN. CiA and its
members shape the future of CAN-based networking, by
developing and maintaining specifications and recommendations
for CAN CC (classic), CAN FD, and CAN XL.
Join the community!

u Access to all CiA specifications, already in work draft status
u Get CANopen vendor-IDs free-of-charge
u Develop partnerships with other CiA members
u Participate in plugfests and workshops
u Initiate and influence CiA specifications
u Get credits on CiA training and education events
u Get credits on CiA publications
u Get the CANopen CC (classic) conformance test tool
u Participate in joint marketing activities
u Get credits on CiA testing services
For more details please contact CiA

office at headquarters@can-cia.org
www.can-cia.org

CAN_DOC

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

CAN_DOC

Uploaded by

Document Information

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

CAN_DOC

Uploaded by

Copyright:

Available Formats

June 2024

Functionally-safe J1939 communication

Commercial road and off-highway vehicles and off-road

www.peak-system.com PEAK-System Technik GmbH

Commercial road and off-highway vehicles and off-road

Reiner Zitzmann (CEO)

© Copyright ◆ Which topics do you like to see in next issues?

CAN Newsletter 2/2024 3

Commercial road and off-highway vehicles as well as off-road construction

T his article discusses the two SAE J1939 standards for

4 CAN Newsletter 2/2024

Table 1: Coverage of communication errors by employed

The adaptive machine

◆ one sequence number, Short product Easy reconfiguration

◆ and the source address (SA) of the associated SDM.

B&R | A member of the ABB Group

◆ A 32-bit Sequence Number.

◆ The length of the data over which the CRC is calculated.

6 CAN Newsletter 2/2024

CAN Newsletter 2/2024 7

computationally more complex than that of any other

Figure 5: Format of Profile #3 assurance data specified in

(This article is based on the 18th international CAN

8 CAN Newsletter 2/2024

(Source: Adobe Stock)

N MEA 2000 is a higher-layer protocol based on CAN

CAN Newsletter 2/2024 9

◆ Details of CAN signal timing.

◆ Tests provided in addition to those required by the NMEA,

A common misunderstanding is that NMEA 2000 is simply

10 CAN Newsletter 2/2024

As connectivity in the world expands, so do we.

Join us on this journey

(Source: Adobe Stock)

T he history of logistics dates back to the first ancient

12 CAN Newsletter 2/2024

Highspeed FPGA Design

Software Driver Support

esd electronics gmbh

Figure 2: The E-Movekit is a complete CAN-connectable

For forklifts and AMRs with lifting functions, linear

14 CAN Newsletter 2/2024

T he S32N family enables automakers to evolve their

Figure 1: S32N processor is scalable across core vehicle

All S32N products integrate a CAN hub, an advanced

Core vehicle functionality

Using S32N processors enables automakers to centrally

CAN Newsletter 2/2024 15

Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52 Cortex-R52

1 MB L2 Cache 1 MB L2 Cache 1 MB L2 Cache 1 MB L2 Cache

Ethernet Acceleration Memory Communication Real-Time Subsystem (CSR)

Ethernet Subsystem (NETC3) 48 MB Platform SRAM with ECC Communication Manager

System MMU (ATU) Resource Isolation Architecture

Enhanced Direct Memory Access (lockstep)

Figure 3: Block diagram of the S32N55 processor (Source: NXP)

16 CAN Newsletter 2/2024

CAN SIC transceiver

CAN FD star network multi-node communication