CAN high-speed transmission

From the beginning, the CAN network used the high-speed transmission physical layer. It was limited to 1 Mbit/s. This physical layer was originally part of the ISO 11898:1993 standard. After the ISO 11898 document was splitted, it is now specified in ISO 11898-2. Transceivers implementing this standard may be powered with 5 V or 3,3 V. Usage of both kinds in the very same network is possible, but decreases the robustness significantly.

On the bus-lines CAN_H(igh) and CAN_L(ow), the transceiver provides a differential voltage. For recessive bits, both lines are driven nominally to 2,5 V. In case of a dominant state, CAN_H nominally measures 3,5 V and CAN_L 1,5 V. Of course, there are some tolerance ranges:

  • Dominant differential voltage range: 0,9 V to 2,0 V
  • Recessive differential voltage range: -1,0 V to 0,5 V

The voltage range from 0,5 V to 0,9 V is indifferent. Using Classical CAN nodes, the bit-rate is limited to 1 Mbit/s. Due to the used arbitration method, the maximum possible bit-rate depends on the network length. Theoretically, at 1 Mbit/s you can reach 40 m. Using real cables, connectors, and other physical layer components including transceiver chips, the achievable length is much shorter. It also depends on the location of the sample point of the bus-level measurement. Of course, when you reduce the bit-rate, you can use longer cables. However, after 1 km normally the voltage drop comes into the game. Dominant bits are not detected reliably anymore. This means that you need cables with a larger cross-section for longer networks. The cables used for CAN high-speed transmission should have a maximum delay of 5 ns/m.

It is recommended to use a line topology, terminated at both ends with 120-Ohm resistors matching the impedance of the cable. Sometimes, the termination resistor is split (2 x 60 Ohm) and an additional capacitor is connected to the ground.

Low-power mode and selective wake-up

The automotive industry is required to avoid empty batteries even after several weeks of standstill. Additionally, it has to save energy during operation. This is why high-speed transceivers with low-power functionality were introduced. They were originally standardized in ISO 11898-5. However, this standard was merged with the original ISO 11898-2 standard.

The low-power high-speed transceiver can be awakened by a dedicated bus signal. Than all nodes in the network are powered. If you want to awake just a subset of the connected nodes, you need to implement a selective wake-up procedure. This can be done by means of a dedicated CAN data frame. Such transceivers were originally standardized in ISO 11898-6. Also, this standard was merged to the new ISO 11898-2 standard.

With the selective wake-up function you can realize partial networking. You can switch-off all nodes that you need rarely or only under specific conditions. For example, the parking-related ECU can be in low-power mode when driving on the highway. Also, the power-window and power-mirror can be switched-off and just be powered when you want to use them. These are not time-critical applications.

CAN high-speed transceiver implementations

In the past, stand-alone transceivers compliant with ISO 11898-2 dominated the market. Nowadays, the automotive industry increasingly uses so-called SBCs (system base chips). Besides the CAN transceiver, they comprise other circuits. These include, for example, a combination of low drop out (LDO) voltage regulator(s), switching voltage regulator(s), and high-side switches. Other optional elements are reset generators, watchdogs, and wake-up logic. Of course, LIN transceivers may be added as well. Normally, an SPI is provided to communicate with the ECU's host controller.