CAN Basic Architectures - CAN System Engineering - page 10

Fig. 1.4 CAN error frame

The notification of a detected fault to all other participants is done by an Error
Frame , which is characterized—like the End Of Frame— by a deliberate breach of
the bit-stuffing rules. The structure of such an Error Frame is shown in Fig. 1.4 .
The Active Error Flag overwrites, with a series of six or more consecutive domi-
nant bits, all other current bus levels. This pattern is—by violating the bit-stuffing
rules—recognized by all other network nodes.
To prevent a persistent local disturbance of a CAN node or a group of CAN nodes
from permanently paralyzing the CAN bus with dominant Active Error Flags , the
affected CAN nodes—according to a specific algorithm—gradually withdraw from
CAN bus activity. After the first stage of the withdrawal from the CAN bus, a CAN
node may send only so-called Passive Error Flags (see Fig. 1.4 ). This algorithm is
described in detail below.
The fault confinement consists of three parts: error detection, error handling and
fault isolation.

1.2.4.1 Error Detection

The error management is able to identify five different types of errors:

• Bit - Error
A transmitted bit is not received with the same logical value with which it was sent.
Excluded are the Arbitration Field and the ACK Slot .
• Stuff - Error
More than five consecutive bits of the same level have been detected. Excluded are
End Of Frame and Interframe Space (IFS) .
• CRC - Error
The calculated CRC checksum does not match with the received CRC checksum.
• Form- Error
There has been a violation of the frame format, e.g. CRC Delimiter or ACK Delimi-
ter was not recognized as a recessive bit, or End Of Frame was disturbed.