Texas Instruments | Using CAN Arbitration for Electrical Layer Testing | Application notes | Texas Instruments Using CAN Arbitration for Electrical Layer Testing Application notes

Texas Instruments Using CAN Arbitration for Electrical Layer Testing Application notes
8th international CAN Conference
Using CAN Arbitration for Electrical Layer Testing
Sam Broyles and Steve Corrigan, Texas Instruments, Inc.
The Controller Area Network (CAN) protocol incorporates a powerful means of seamlessly
preventing data corruption during message collision. This arbitration process and its
relationship to the electrical layer variables are explained. Techniques to force message
collision and test arbitration are demonstrated with strategies to leverage arbitration as a
quantitative benchmark in safety-critical systems. The benchmark is then applied to
several example systems and results provided for comparison.
The ability of a Controller Area Network to manage message collision provides a unique proving
ground for protocol compliance in any application. A means of determining a benchmark for a
system’s performance by measuring a network’s ability to execute proper arbitration is developed
in this example. It is demonstrated that while a CAN bus appears to be functioning normally,
many arbitration errors may be unnoticed by system operators.
Arbitration Basics
Since any CAN node may begin to transmit when the bus is free, two or more nodes may begin to
transmit simultaneously. Arbitration is the process by which these nodes battle for control of the
bus. Proper arbitration is critical to CAN performance because this is the mechanism that
guarantees that message collisions do not reduce bandwidth or cause messages to be lost.
Each data or remote frame begins with an identifier, which assigns the priority and content of the
message. As the identifier is broadcast, each transmitting node compares the value received on
the bus to the value being broadcast. The higher priority message during a collision has a
dominant bit earlier in the identifier. Therefore, if a transmitting node senses a dominant bit on
the bus in place of the recessive bit it transmitted, it interprets this as another message with
higher priority transmitting simultaneously. This node suspends transmission before the next bit
and automatically retransmits when the bus is idle.
The result of proper arbitration is that a high-priority message transmitted without interruption is
followed immediately by a low-priority message, unless of course, another high-priority message
attempts to broadcast immediately following the same message. Since no messages are lost or
corrupted in the collision, data and bandwidth are not compromised.
Electrical-Layer Variables (bit timing requirements)
Each CAN bit is divided into four segments (see Figure 1). The first segment, the synchronization
segment (SYNC_SEG), is the time that a recessive-to-dominant or dominant-to-recessive
transition is expected to occur. The second segment, the propagation time segment
(PROP_SEG), is designed to compensate for the physical delay times of the network as shown in
Figure 2, and should be twice the sum of the propagation delay of the bus, the input comparator
delay, and the output driver delay. The third and fourth segments, both phase buffer segments
(PHASE_SEG1 & PHASE_SEG2), are used for resynchronization. The bit value is sampled
immediately following PHASE_SEG1.
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
Nominal Bit Length
Sample Point
forces rising edge
in first segment
Compensates for
SEG1 may be lengthened
or SEG2 may be shortened
for resynchronization
Figure 1. Partitioning of the Bit Timing Segments
The bit rate may be changed by either changing the oscillator frequency, which is usually
restricted by the processor requirements, or by specifying the length of the bit segments in “time
quantum” and the prescaler value. The prescaler value is multiplied by the minimum time
quantum, which is the reciprocal of the system clock frequency, 1/fsys, to determine the length of a
working time quantum. Bit time may then be calculated as the sum of each bit segment, and the
bit rate may be calculated as the reciprocal of this sum.
Each node must perform a hard synchronization upon every recessive-to-dominant edge after a
bus idle or received start of frame. Hard synchronization is a restarting of the internal bit timing to
force the edge into the SYNC_SEG, where edges are expected to occur. Resynchronization is
performed on all other recessive-to-dominant edges of other received bits by lengthening or
shortening the PHASE_SEG1 or PHASE_SEG2 by one to four time quanta as specified by the
resynchronization jump width. If the difference between the edge causing resynchronization and
the SYNC_SEG exceeds the resynchronization jump width, the effective result is the same as a
hard synchronization.
CAN Network Errors
CAN protocol specifies five different types of network errors. A transmitting node detects a bit
error when it monitors a bit value different than it is transmitting; the reaction to this condition
varies with the nature of the error. A stuff error occurs when the bit-stuffing rule is violated – a bit
of opposite value must be inserted immediately following any series of five consecutive bits of the
same value in a message. A cyclic redundancy check (CRC) error occurs when a receiving node
receives a different CRC sequence than anticipated. (Note that all nodes independently calculate
the CRC sequence from the data field). A form error occurs when a field contains an illegal bit
value. Finally, an acknowledgement (ACK) error occurs when the transmitter does not monitor a
dominant bit in the ACK slot to signify that the message had been received properly by another
node as shown in Figure 2.
When a node detects a bus error, it transmits an error frame consisting of six dominant bits
followed by eight recessive bits. Multiple nodes transmitting an error frame will not cause a
problem because the first recessive bits will be overwritten. The result will remain six dominant
bits followed by eight recessive bits, and cause the bus to be safely reset before normal
communications recommence.
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
ACK #1
ACK #2
Figure 2. Round trip propagation delay of network measured as the delta of the ACK bit
response time between two nodes on opposite ends of the network.
The CAN protocol provides a means of fault confinement by requiring each node to maintain
separate receive and transmit error counters. Either counter will be incremented by 1 or 8,
depending on the type of error and conditions surrounding the error. The receive error counter is
incremented for errors during message reception, and the transmit error counter is incremented
for errors during message transmission (for further details, see reference 1). When either of
these counters exceed 127, the node is declared “error-passive,”which limits the node from
sending any further dominant error frames. When the transmitted error count exceeds 255, the
node is declared “bus-off,” which restricts the node from sending any further transmissions. The
receive and transmit error counters are also decremented by 1 each time a message is received
or transmitted without error, respectively. This allows a node to return from error-passive mode to
error-active mode (normal transmission mode) when both counters are less than 128. The node
may also return to error-active mode from bus-off mode after having received 128 occurrences of
11 consecutive recessive bits. Overall, a network maintains constant transmit and receive error
counters if it averages eight properly transmitted or received messages for each error that occurs
during transmission or reception, respectively.
Analysis of Network Errors
As shown in Figure 2, the oscilloscope is an invaluable tool for observing bus status. For these
experiments, a Tektronix 784D oscilloscope with Tektronix P6243 1 GHz single-ended probes are
used. With careful choice of message identifiers and data fields, messages can be visually
associated with the transmitting node. This guarantees observation of the participation of each
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
node during an experiment. Additionally, the transmission lines from the controllers to the
transceivers are useful for monitoring the participation of each node during arbitration and ACK.
A more detailed record of bus activity may be found in the status and control registers of the
processors that are reviewed after each experiment. Though the error counters are incremented
and decremented through a complex series of rules, it is sufficient to note error status and types
of errors that occur to assess bus performance under a set of experimental conditions.
To ensure that every message collision results in proper arbitration, it is required that the
processor be programmed to identify the order in which each message is received. This is
accomplished by checking for a receive message pending flag. For the purpose of these
experiments, the processor of node A in Figure 3 is programmed to stop program execution upon
the first event of improper arbitration.
Forcing Message Collision
While operating with a two-node bus, if one node enters the bus-off state, either the bus becomes
silent or the other node continues to retransmit until the reception of a proper ACK bit is received.
With three or more nodes, a proper ACK bit would be inserted by one of the remaining nodes and
the bus simply becomes silent. Either case represents an experimental condition that would not
be recommended for use in a final application.
With the three-node bus in Figure 3, message collision is forced by programming two nodes to
respond immediately to a message from the third node. In these experiments, node A is
programmed to send a data frame and wait for nodes B and C to respond. If nodes B and C
respond in the wrong order, arbitration is not properly negotiated and node A does not retransmit.
If nodes B and C do respond in the proper order, arbitration is properly negotiated and node A
retransmits its message.
10 m cable
13 cm cable
120 Ω
120 Ω
Figure 3. The Experimental 3-Node Bus
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
Figure 4 displays a signal capture of this system during arbitration. Channel 1 is the bus signal
while channel 2 displays the output signal of the node with the highest priority, node A. Channel
3 is the output signal of node B, which is the lowest priority message on the bus. Channel 4 is
the output signal of node C, which is the medium priority message. Notice that both nodes B and
C participate in the ACK bit of node A and begin transmitting together. However, after just a few
bits node B stops transmission until node C, the higher priority node, is finished.
ACK bit
ACK bit
ACK bit
Node C wins arbitration
Figure 4. Arbitration Display
Experimental Network Systems
Three bus systems were tested to study arbitration and system performance. In each system,
three nodes actively participate on the network. Node A sends the first data frame. Nodes B and
C attempted to respond simultaneously and negotiated bus control. All messages have 29-bit
identifiers and 8-byte data fields. Node A retransmits it’s data frame and begins the process
again if the dominant message from node C is received before the recessive message from node
B. This continues for one million cycles unless node A receives the messages in the wrong order
or until any of the nodes enters a bus-off state. The bus-off state is caused by exceeding the
allowable transmit or receive error counts. Each network is constructed with very inexpensive
120-ohm impedance twisted-pair AWG 24 cable with grounded shielding and 120 ohm
terminating resistors on either end.
The first network consists of the three active nodes with nodes A and B separated by 13-cm
cables and nodes C and B separated by 10 m of cable, as shown in Figure 3. The second
example in Figure 5 is a network with 27 dummy load-nodes added between node C and the
termination. The cable is increased between nodes B and C to 40 meters. The dummy nodes
are powered transceivers without CAN controllers or processors and serve only to load the bus
as if other nodes are present. All dummy load-nodes are mounted on a bank of test boards for
ease of wiring.
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
Load Bank of SN65HVD230 CAN Transceivers
EVM #2
EVM #3
120 Ω
120 Ω
EVM #1
EVM #30
EVM #3
EVM #1
Termination on
Node #1 and #30
EVM #2
Load Nodes
#4 thru #30
EVM #30
cable jumpers
Figure 5. The 30-Node CAN Bus
The third network consists of 60 nodes, as shown in Figure 6. One bank of 30 Texas Instruments
SN65HVD230 CAN transceivers and one bank of 30 Philips PCA82C250 transceivers are
separated by a 200-m cable. Each bank has transceivers mounted on test boards separated by
13 cm of cable. Nodes A and B are nearest the 200-m cable on the ‘230 bank, while node C is
nearest the 200-m cable on the ‘250 bank.
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
Load Bank of 30
SN65HVD230 CAN Transceivers
Load Bank of 30
PCA82C250 CAN Transceivers
EVM #2
120 Ω
120 Ω
EVM #30
EVM #1
EVM #1
EVM #30
EVM #1
EVM #30
Node #30
Figure 6. 60-Node CAN Bus
Analysis of Experimental Network Systems
The first example network (3 nodes) is tested at 500 kbps, 625 kbps, 800 kbps, 1 Mbps, 1.25
Mbps, and 2 Mbps. It performs flawlessly at 500 kbps and 625 kbps, but continually fails
arbitration at signaling rates of 800 kbps, 1 Mbps, and 1.250 Mbps, and is inoperable at 2 Mbps
due to a form error. The test is performed again by checking that each message is received
without checking the receive order. Now the network functions without error at signaling rates of
800 kbps, 1 Mbps, and 1.250 Mbps. Messages from nodes B and C are being received, but in
the wrong order – an arbitration compliance failure.
The 30-node network is tested with signaling rates of 250 kbps, 500 kbps, 625 kbps, 800 kbps,
and 1 Mbps. It performs flawlessly at 250, 500, and 625 kbps, but fails arbitration at 800 kbps. At
1 Mbps, the network fails arbitration after encountering bit errors, stuff-bit errors, and cyclicredundancy- check (CRC) errors. The network is re-tested by checking only that each message
is being received without checking the receive order. This results in successful operation at 800
kbps, but messages from nodes B and C are being received in the wrong order. The network
remained inoperative at 1Mbps after encountering form errors.
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
The 60-node network is tested with signaling rates of 31.25 kbps, 62.5 kbps, 125 kbps, and 250
kbps. It performs flawlessly at 31.25 kbps and 62.5 kbps, but fails arbitration at 125 kbps having
encountered form errors. It is inoperable at 250 kbps, encountering stuff bit errors. Again, the
network is re-tested by checking only that each message was being received without checking
the receive order, and the network remained inoperable at signaling rates of 125 kbps and above
due to stuff bit errors.
Figure 7 displays the experiment without checking for receive order to better understand the
failure mechanism of arbitration. This shows node B transmitting (CH3) its lower priority message
before node C (CH4) transmits its higher priority message without node C ever having competed
for bus dominance. For proper arbitration, it was required that both node B and C would attempt
to transmit following the message from node A (CH2).
Figure 7. CAN bus signal at 1 Mbps with 11-bit identifiers and 2-byte data fields.
Note that node C is transmitting a valid ACK bit synchronously with node B in response to node
A’s message, but is unable to participate in bus arbitration properly. In each of these
experiments, programs for nodes B and C are identical except for the message identifiers and
data fields. If a dominant bit is received on the last bit of the intermission field in the interframe
space between messages, it should be interpreted as the start of frame bit. Any other node also
intending to transmit a start of frame bit immediately following intermission should begin
transmitting the first bit of its identifier during the next bit so proper arbitration may commence. In
other words, if one node jumps the gun on the last bit of the interframe space, all other nodes
should accept this as a valid start of frame, synchronize, and transmit or receive as normal. Also,
any high-priority message is delayed until the end of a lower priority message if the lower priority
message began transmitting first. Therefore, node C must have attempted to transmit more than
a full bit length after the dominant start of frame bit from node B, even if node B began to transmit
a bit too early.
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
Node C’s inability to negotiate arbitration properly and consistently, whether from bit error during
message identifier or delayed message transmission, represents an experimental condition that
exceeds the network’s ability to be completely CAN protocol compliant. This condition can be
remedied by slowing the signaling rate, since the network is exceeding its maximum rate.
Summary of Experimental Systems Performance
The final assessment of each experimental network reflects the highest bit rate at which the
network maintains full CAN compliance with proper arbitration. The first network of 3 nodes and
10-m bus is fully compliant at 625 kbps. The second network with 30 nodes and 40-m bus is also
fully compliant at 625 kbps while the third network with 60 nodes and 200-m bus is fully compliant
at 125bps.
Conclusion: Evaluation of Arbitration as a Quantitative Benchmark
Clearly, testing for proper arbitration is a more stringent test than merely watching for bus errors
in normal operation and suggests that an application can appear to be quite functional although it
is unable to support proper arbitration. Arbitration problems are invisible to a user if no
operational errors are encountered in the application. Both the first and second experimental
networks operate without bus errors at much higher signaling rates than each network could
support with proper arbitration.
Proper arbitration is critical to CAN performance because this is the mechanism that guarantees
message collisions do not decrease bandwidth with multiple retransmission or loose messages.
The quantitative benchmark produced by this method of testing a network for proper arbitration
compliance is therefore defined as the maximum bus speed attainable on a safety critical network
with the full data security enabled by CAN protocol.
This method is compatible with any network topology. Special care should be taken in the
application of this method when selecting the location of nodes to monitor (A, B, or C) to ensure
that the C monitor node is the worst-case node. Node C should be positioned to maximize the
propagation delay, signal reflections, and other network conditions of the final application. This
assures the greatest difference between the response of nodes B and C to node A. Therefore,
networks with completely different topologies may be compared quantitatively with this arbitration
This method is also adaptable for use during normal network operation, and offers the ability to
check protocol compliance and provides confirmation that maximum data security is being
enforced in the application while the network continues normal operation. If even more security is
required for safety-critical systems, Time-Triggered CAN (TTCAN) has been developed by Bosch
to addresses concerns about low-priority messages occupying a bus when a very high-priority
message needs to be sent.
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002.
8th international CAN Conference
ISO-11898 Can Specification 2.0
Controller Area Network, Basics Protocols, Chips and Applications; Dr. Konrad Etschberger;
ISBN 3-00-007376-0
CAN Systems Engineering, From Theory to Practical Applications; Wolfhard Lawrenz, ISBN 0387-94939-9
Sam Broyles
Texas Instruments, Inc.
P.O. BOX 660199, MS 8701
Dallas, Texas 75266
phone # 214 480 3232
fax # 703 940 8113
Steve Corrigan
Texas Instruments, Inc.
P.O. BOX 660199, MS 8710
Dallas, Texas 75266
phone # 214 480 4743
fax # 214 480 3160
This paper was presented at the 8th international CAN Conference in Las Vegas on February 27, 2002. 10
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, modifications,
enhancements, improvements, and other changes to its products and services at any time and to discontinue
any product or service without notice. Customers should obtain the latest relevant information before placing
orders and should verify that such information is current and complete. All products are sold subject to TI’s terms
and conditions of sale supplied at the time of order acknowledgment.
TI warrants performance of its hardware products to the specifications applicable at the time of sale in
accordance with TI’s standard warranty. Testing and other quality control techniques are used to the extent TI
deems necessary to support this warranty. Except where mandated by government requirements, testing of all
parameters of each product is not necessarily performed.
TI assumes no liability for applications assistance or customer product design. Customers are responsible for
their products and applications using TI components. To minimize the risks associated with customer products
and applications, customers should provide adequate design and operating safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any TI patent right,
copyright, mask work right, or other TI intellectual property right relating to any combination, machine, or process
in which TI products or services are used. Information published by TI regarding third–party products or services
does not constitute a license from TI to use such products or services or a warranty or endorsement thereof.
Use of such information may require a license from a third party under the patents or other intellectual property
of the third party, or a license from TI under the patents or other intellectual property of TI.
Reproduction of information in TI data books or data sheets is permissible only if reproduction is without
alteration and is accompanied by all associated warranties, conditions, limitations, and notices. Reproduction
of this information with alteration is an unfair and deceptive business practice. TI is not responsible or liable for
such altered documentation.
Resale of TI products or services with statements different from or beyond the parameters stated by TI for that
product or service voids all express and any implied warranties for the associated TI product or service and
is an unfair and deceptive business practice. TI is not responsible or liable for any such statements.
Mailing Address:
Texas Instruments
Post Office Box 655303
Dallas, Texas 75265
Copyright  2003, Texas Instruments Incorporated
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF