CN0401

CAN FD Protocol

CAN was designed as a robust communication network to eliminate the need for point-to-point wiring between individual microcontrollers. Rather than connecting each point (or node in CAN terminology) that must share information with one another, each node is connected to a single twisted pair. This configuration reduces wiring cost and complexity. Unlike many other protocols, CAN is multiple master and the physical location of the node does not determine which node takes priority on the CAN bus. There is not a master/slave relationship as exists in SPI, I²C, or RS-485 where one node controls and services the network. Instead, it is the message ID, or programmed importance of the message, that determines the priority of a node to control the bus. A single node may send a message or various priority levels. The AN-1123 Application Note, Controller Area Network (CAN) Implementation Guide, provides additional details on this subject. 

CAN FD compliant controllers are backwards compatible with classical CAN and are tolerant of classical CAN frames. It is possible to have groups of nodes communicating with classical CAN while other nodes communicate with CAN FD on the same network. However, the vintage of existing CAN controllers must be considered when adding CAN FD nodes. Legacy classical CAN controllers may identify the arbitration phase of the CAN FD frame as an error and produce an error frame that overwrites the CAN FD frame. Such controllers make CAN FD communication impossible. These controllers need to be replaced, or limit future nodes to classical CAN. The controller in this circuit, the MCP2518FD, can operate in classical CAN or CAN FD modes. It is tolerant of CAN FD frames when operating in classical CAN mode. 

Fast Loop Delay and High Speed Data Rate  

A distinguishing feature of CAN is the arbitration phase at the start of each CAN frame. Nodes arbitrate between one another to determine message transmission priority by each transmitting a sequence of dominant and recessive bits. Each bit requires a certain amount of time to propagate throughout the network, and each node participating in arbitration must have enough time to respond. This means that during the arbitration phase, the maximum data rate is limited by the longest total signal propagation time between any two CAN controllers on the network. 

As illustrated in Figure 2, the signal path starts when the Node A CAN controller begins transmission. This signal first passes through the Node A transmitter, then propagates over the cables, then through the receiver of the furthest node, and finally reaches the furthest CAN controller. During the arbitration phase, the receiving node may also be transmitting, so the signal propagation delay from Node B to Node A must also be considered. The worst case of this propagation delay determines the absolute maximum data rate of the arbitration phase.

Propagation delay along the bus lines increases with the cable length and construction. Cable length is typically fixed by the physical locations where nodes need to be located and this portion of signal propagation delay becomes basically fixed. The propagation delay through the transceivers receive and transmit circuitry is called the loop delay. The ADM3055E has an industry leading maximum loop delay of only 150 ns, which allows a network designer to dedicate less of the bit time to the transceiver. These time savings can translate to higher arbitration data rates, longer bus cables, or longer bus signal settling time for added communication robustness at any arbitration data rate. More details on the total propagation delay and optimizing CAN networks can be found in the Analog Dialog article, “Configure Controller Area Network (CAN) Bit Timing to Optimize Performance.”

The maximum data rate in the data phase of the CAN FD frame, by contrast, is not determined by propagation delay but rather network signal quality. Reflections due to impedance mismatches and cable stubs are among the factors limiting the data rate in multiple node networks. As CAN FD multiple node networks are put in service, 2 Mbps data rates are a popular conservative choice. The ADM3055E isolated signal and power transceiver can operate at up to 12 Mbps data rates, which support very fast data transfers for point-to-point connections, and future proofs the multiple node network to adapt to faster data rate requirements in the future.

Standby Mode, and Remote Wake Up

The CAN FD controller and isolated CAN FD transceiver can be set to standby mode with commands issued by the development platform over the SPI bus. The CAN FD controller sets itself and the isolated CAN FD transceiver to standby mode. In standby mode, the transmit functionality of the transceiver is disabled and its output is set to a high impedance state. 

The transceiver can only be taken out of standby mode by the local CAN FD controller, however the transceiver does respond to the remote wake-up calls made by other nodes. The remote wake-up pattern is defined in ISO11898-2:2016 and in the ADM3055E data sheet. The pattern can be sent out in the arbitration field of a frame with no data, or it can be sent in the data field. It must only meet the timing requirements of the transceiver. When the remote wake up pattern is received, the RXD pin of the ADM3055E transceiver toggles in response to low speed data on the CAN FD bus. 

The state changes on the RXD pin is used to trigger an interrupt to the CAN FD controller. When the ADM3055E receives the remote wake-up pattern, it does not exit standby mode. It is then up to development to determine whether to respond, or to toggle the standby pin of the transceivers to discontinue reception of low speed data and return to standby mode until the remote wake-up pattern is again received. While in standby mode, the isolated auxiliary channel of the transceiver latches in the last state, and its integrated isoPower^® isolated dc-to-dc converter continues to operate, providing power to the bus side circuitry.

Isolation

Harsh environments, lengthy physical separation, and different power supply sources between nodes can and often do have different local ground potentials. The differing local ground potentials causes currents to flow through the ground wire causing common-mode offsets and noise. Isolating the physical bus lines breaks the ground loops and eliminates these problems. The ADM3055E breaks ground loops and carries system level safety certification for 5 kV rms signal and power isolation between the CAN FD node and CAN bus lines.

Switchable Termination

For best signal integrity, implement termination at both ends of a CAN bus. Switchable termination allows software to configure the location of the termination. Software control is useful for on-the-fly network reconfiguration of the CAN bus due to the addition or removal of nodes.

To keep the network reliability as high as possible, termination circuitry must not restrict the common-mode range. The termination circuitry must also not be affected by the common-mode range of the signal or, more specifically, the termination circuitry must remain off when set to off and remain on when set to on. To meet the required circuit characteristics, the termination circuitry on the EVAL-ADM3055E-ARDZ evaluation board floats with the transmitting node using very compact optically isolated SPST solid state relays (SSR). 

Controlling the relays from the auxiliary isolated channel of the ADM3055E means the relays do not bridge the isolation gap. Because the relays do not bridge the isolation gap, they are not required to serve a safety isolation function and can be chosen to have the smallest package possible to save printed circuit board (PCB) area.

The 120 Ω termination resistance can be accomplished with a single resistor. However, splitting the resistor into series 60 Ω resistors offers an inexpensive measure of electrostatic discharge (ESD) protection to both relay pins exposed to the CAN bus. Implementing the switchable termination circuit with a second SSR allows for the addition of a filter capacitor. This added capacitor acts with the split termination resistors to provide a low-pass filter, reducing common-mode noise on the CAN bus.

Silent Mode and Slope Control Mode

The CN-0401 circuit also supports bus baud rate discovery via a software configurable trial and error in conjunction with the silent mode of the transceiver. Silent mode disables the transmit channel of the transceiver and allows a CAN controller to produce error frames while attempting to synchronize to the bus data rate without interrupting bus traffic with those error frames.

The CN-0401 circuit provides access to the ADM3055E slope control mode. For low speed signaling, slope control decreases the slew rate of the CANH and CANL recessive to dominant transitions. Decreasing the slew rate minimizes ringing and electromagnetic interference (EMI) caused by fast edges. Do not use slope control mode for high speed signaling.

CAN FD Controller

The EVAL-ADM3055E-ARDZ uses the MCP2518FD external CAN FD controller. The controller and the ADM3055E transceiver are compatible with legacy CAN, CAN 2.0B, and CAN FD. Joining CAN or CAN 2.0B requires software changes. 

The standalone CAN controller and ADM3055E offer a convenient option for system designers to add one or more isolated CAN ports to existing designs by simply connecting to the ubiquitous SPI port. The MCP2518FD supports a maximum CAN FD data rate of 8 Mbps. See the MCP2518FD datasheet for further information.