Fault tolerant CAN (ISO11898-3) is terminated differently than high speed CAN (ISO11898-2). Therefore the signal levels for dominant and especially recessive bits differ.
Newer CAN FD compatible transceivers can have baudrates up to at least 8 MBit/s, but none of their datasheets specify minimal baudrates or compliance with ISO11898-3.
Is it safe to assume they are backwards compatible or should I look for some kind of feature indication in the datasheet?
The only problem I can think of is the dominant timeout function of some transceivers. If the baud rate is too low, it might cause problems with longer strings of dominant bits.
Best Answer
Probably not in any sort of reliable way. As you note, they are terminated differently to give a bipolar differential signal. I have previously tested some Microchip transceivers to see if they can operate like this, but the performance was not good (if I recall correctly, the propagation delays and slew rates were mismatched by a large amount, and extremely dependent on the termination resistance). Additionally, fault tolerant CAN has a threshold voltage of something like 3.2V, while regular CAN is about 1V so your noise performance is going to be severely compromised.
However, the only real way is to test it out, but I doubt the manufacturer is going to guarantee any of the performance when operated in this mode, so it's probably not a good idea.
Also you lose all the fault tolerant (ability to lose one wire) benefits if you're not using proper low speed transceivers, so what's the point then?