TCAN1043-Q1: CANL short to GND

Hi Tuong,

It looks like the first figure here may be mislabeled. Before the fault is applied the blue signal is switching between 2.5V and ~1.0V which is consistent with CANL behavior and the pink signal switches between 2.5V and 4.0V which looks like CANH behavior. Can you confirm if this is the case? If so it looks like both waveforms show the same CANL short to ground test. 

What data rate is being run in yours system for this test case? The bus fault detection feature only works reliably for data rates of 400kbps or slower (bit time of tCBF = 2.5us). 
What are the bus conditions for this test? What are the termination values? What are the characteristics of the signal harness between communicating nodes?

Regards, 
Eric Schott

Hi Eric,

The first figure is correct, please note that CAN_L is -5V offset.

I retest as below, with C1 is nFault, C2 is CAN_H, C3 is CAN_L

Data rate i believe it's less than 500kbps

Test conditon: MCU => CAN => Kvaser => Laptop

Termination Reistor: 120R (60R for each line)

Hi Eric, 

could you please take a look at this issue it’s quite urgent . Thanks

Tuong,

Thanks for your patience, can you try different data rates to see if nFAULT toggles low? I would expect a slower data rate make it work.

The bus fault detection works by measuring the differential current between CANH and CANL. When the current passes a certain threshold based on the fault type, the nFAULT pin will go low. In this case, since CANL to GND isn't a particularly abnormal condition, while still being considered a fault, the current may not be passing the threshold expected. 

Is there a CMC on the board? Is it possible to share a schematic?

Regards,

Eric Hackett

Hi Eric,

Sorry I was missed this discuss during holiday. 

Regarding sch, please refer to photo. We do have CMC. 

Your concern about data rate, I cannot slowdown data rate as it from MCU. 

I also don't see problem with CANL shorted to GND, but my testing team they expect nFAULT should behaviors similar to CANH.

Looking for your feedback.

CMC should be move to the right of resistor 120 ohm

Eric Hackett could you please give feedback.

Hi Tuong,

In your schematic, the termination resistors are on the IC-side of the common mode choke. This will allow the initial current from the dominant drive state to appear nominal to the transceiver while the short is isolated from the circuit from the inductance of the CMC. If the dominant period were to last long enough (if the data rate was lower), the circuit would reach a steady state and the transceiver would be able to more reliably detect bus faults. Alternatively if the termination were placed on the bus-side of the CMC, the fault current would be more apparent to the transceiver earlier in the bit time. 

Is it possible to repeat your testing with the R92/R95 removed and the additional 120-ohms placed elsewhere in the system? 

Regards, 
Eric Schott

Hi Eric,

I carried out the test with two cases: case 1 is Terminator res not on DUT and case 2 is 120R on DUT

And below is result:

Looking for your feedback.

Thanks