The CAN in Automation (CiA) association has established the CAN FD Light Special Interest Group with the goal of serving sensor/actuator and LED networking in automotive applications. Led by Fred Rennig from ST Microelectronics, the group will develop a low-cost communication network based on ISO 11898-1 CAN FD, with the expectation of offering fast performance and the low-cost, robustness and reliability for which CAN is well known. Data fields of up to 64 bytes are envisaged, along with fixed bit-rates of e.g. 1 Mbit/s.
The initial proposal for CAN FD Light is to follow a primary/secondary network architecture, instead of the more complicated arbitration and error signalling that is characteristic of CAN FD. With no need for an accurate clock source at each node, as is typical with CAN FD, single monolithic devices without crystal oscillators could be used. This would lower cost and improve the reliability of the implementation, as crystal oscillators are sensitive to vibration and temperature.
Asked why CAN FD was chosen as the base protocol, rather than other sensor protocols, Holger Zeltwanger, CiA Managing Director said:
While HVAC systems may require less than 1 Mbit/s, for the deeply embedded networks in smart lamp devices that CAN FD Light targets, OEMs have requested 1 Mbit/s, though Zeltwanger says that “2 Mbit/s seems to be achievable”.
Zeltwanger foresees few challenges ahead for this standardisation effort:
Kent Lennartsson, Research Manager at Kvaser AB comments:
Notably, the CAN FD Light proposition fits with the concept of recentralising control systems around new generation microcontrollers and processors that are hypervisor-capable i.e. can support different software programs within the same hardware, such as the ARM Cortex-R52. In this way, a cloud of sensors might be networked via CAN FD Light and controlled by a remote ECU.