TCAN1051HV: line termination and ESD protection for fault protected CAN transceiver

Carlos,

You're right that this is a tough set of requirements, since it would be difficult to find a TVS diode that could either tolerate the 500-ms overvoltage condition or stay out of breakdown until 65 V while still clamping at 70 V maximum. One approach you could evaluate, though, is AC-coupling the TVS diode so that fast transients (i.e., your ISO 10605 ESD pulse) are clamped without subjecting the TVS to low-frequency stresses like your load dump.

Another approach would be to insert a component between the TVS and your connector that could "open" the circuit in the event of a high-current surge. Here are some "transient blocking units" that you could take a look at: www.bourns.com/.../tbu-high-speed-protectors-hsps. This approach would also help solve the issue of having to size the termination resistors to handle worst-case surge/fault conditions (which usually results in using multiple large resistors).

To answer your question on TCAN1051HV - it is rated for ISO 10605 ESD pulses up to 8 kV (contact discharge) or 15 kV (air discharge), so some external protection is indeed needed for a 25-kV strike.

Regards,
Max

Carlos,

For some reason the image you pasted isn't displaying, but I'm pretty sure I understand what you are asking about.

For Case 1, CANL will still try to pull towards ground when the driver is dominant. It will not be able to pull it as low as during normal operation, though. There are two factors involved that limit the sink current in cases like this - the impedance of the CANL driver (about 30 Ohms to 50 Ohms in the dominant state) and its short-circuit current limit (100 mA). With a 24-V short on CANH the CANL input would be current-limited, meaning that you would only end up with about a 6-V difference between CANH and CANL (100 mA flowing through the 60-Ohm bus termination).

Note that under this condition the CANL pin would have about 18 V on it and be sinking 100 mA, resulting in significant power dissipation. This would likely heat up the device and trigger the over-temperature shutdown feature. This would make CANH and CANL high impedance until the device cools off, meaning that both CANH and CANL would again be close to 24 V (since no substantial currents will be flowing into or out of the CAN transceiver).

For Case 2, the CANH level will be stuck close to 24 V as well. There is a diode in the CANH output driver that would prevent higher bus voltages from coupling through onto the transceiver's power supply. The CANL output would be current-limited and again likely go into thermal shutdown given a long enough dominant state.

For Case 3, CANL will be close to ground as well. This would look basically like an open-drain driver (pull-down FET) connected to a 60-Ohm pull-down externally (i.e., the combination of the 60-Ohm bus termination and the short-circuit on CANH).

For Case 4, your assumption is basically right, but the CANH voltage may end up being slightly lower in practice due to the additional common-mode loading presented by the 60-Ohm pull-down (i.e., connection to the ground short through the bus termination).

It's really the "double fault" condition of 0 V on one end of the termination and VBAT on the other that drives selection of high-wattage resistors. For this kind of fault you don't get the benefit of the protection features of the CAN driver (like short-circuit current limiting and thermal shutdown).

Hope that all makes sense - let me know if you have questions.

Max

Hi Max,

Thanks a lot for your explanation. That what I was looking for.

Next step is to design the protection network, and TBU-CA series seems a good choice.

Thanks,

Carlos