The diagram above shows CAN-high on top and CAN-Low below. The signals represent 2x the door control device and 1x the central control unit in the Bus. If the voltage on CAN-high changes from 0 to approx. 5 volts, it drops simultaneously from 5 volts to 0 on CAN-Low. Indeed, single data blocks are visible, but there are no signal changes if, e.g., the driver shifts gears. For this the oscilloscope (Bosch FSA 600), in spite of a time deflection of 25 ms, is too slow. Normal oscilloscopes bring output onto the screen, e.g., when they register an excess of a certain voltage. Thus, the pursuit of a message is quite difficult.
When trying to find an error in the system, an oscilloscope is just used to detect Bus activity. Such measurements are also possible with the multimeter, because of its ability to distinguish between values of 0, 0-5 and 5 V. Picture 2 displays an interruption of the CAN_L wire. Two control units may still communicate with each other through the Low wire. The signal was taken off right there. The missing data block indicates that the Low signal is interrupted. The small deviation is the confirmation of the two other control units, because the data transfer still works properly in spite of the missing Low signal.
The two next figures simulate the interruption of CAN_H. For this, the signal transmission was firstly taken between both correctly linked control units and secondly from the high-disconnected control unit. In the first case, you recognize that there are two control units communicating with each other also through CAN_H. In the second case just one control unit still transmits on CAN_H and also confirms constantly the messages of both others which it has received through CAN_L and mass.
The following three figures display a CAN Bus with one totally immobilised wire. Now it is useless for any device. Nevertheless, the data transfer continues, of course. The varying representation of the intact data blocks should not confuse you, because a normal oscilloscope selects any data blocks. A multimeter would do almost the same job, because it would indicate a wire with data transfer of values of 2 - 3 volts.
Instead of an interruption a 300-ohm resistance is put in between. Again, the first figure represents the two fully intact control units, and the second figure represents the information beyond the disturbance point. Clearly there is a weakened signal on Low with a simultaneous full signal on High. This is also a proof that the confirmation originates in every case from the others, because their deviation is not weakened.
The ignition has been switched off; the Bus switches over into the Sleep mode. The Low signal shows the full deviation of 12 volts. By the way, a data block would have to be transmitted after a certain time, to prevent that the synchronization of the single control units would deviate completely from one another.