TCAN1042V-Q1: Bus recessive level outside expected electrical characteristics

The schematic is:

R18, R19 are not populated. The system is not connected to any other node. Signal probed at the end of 1m wire of CANH, CANL

Shyama,

The schematic looks correct, there isn't any specific reason why the CANH and CANL signals would be attenuated. It looks like when there is CAN data transmitted, they are reaching the correct voltages for a dominant signal. I'd recommend populating both R18 and R19, but not populating these won't cause the CANH and CANL to be attenuated as you are seeing. What is the part number for the diode at D7? 

Can you also capture TXD and RXD along with the CANH and CANL waveforms on an oscilloscope? And in case there is some kind of damage to this device, is it possible to change to a new TCAN1042V and verify that the voltages are corrected?

Regards,

Eric Hackett 

Hi,

CDSOT23-T24CAN is the D7 part number. I'll check if behavior changes if we change CAN IC. Should we remove D7 and see if that is causing the recessive level to change?

The Power supply given to TCAN1042V is through REGIN (5.6V) and a diode to ensure the power supply doesn't cross 5.5V. This causes fluctuation in Vcc power supply of CAN IC from 5V to 4.6V due to change in diode drop when dominant bit occurs. Can this cause an issue?

Above is Vcc supply schematic

Above is Vcc (VS) -GND waveform 

Shyama,

Yes this can cause issues in the form of the CAN bus voltage dropping every time VCC drops due to a dominant bit. This is why we recommend a bulk and decoupling capacitance on VCC, to maintain a steady voltage level during dynamic current draw. That being said, this shouldn't cause CANH and CANL to be biased as low as they are in your system. It still seems like something is pulling them low, a low-impedance path to GND.

Removing the diode could be one process of elimination, but I would be surprised if this is the culprit.

Regards,

Eric Hackett 

We have seen that R18, R19 are populated to 60 ohm each. Sorry for previous miscommunication. But the resistors were burnt (Please suggest why they might be burnt). The values of the burnt resistors are: R18: 4.95 kohm, R19: 40.5kohm. We also saw CANH-GND resistance as 2.4kohm, CANL-GND as 24 kohm when the burnt resistor where still in the system. 

Once we depopulated the burnt resistors, the recessive bit is back to 2.5V for both CANH, CANL and the resistance from CANH-GND and CANL-GND are in Megaohms. 

Can R18, R19 create resistance between CANH-GND and CANL-GND? Did the IC get temporarily latched to ground by some partial resistance? Please suggest how a burnt resistor case can occur and how to avoid it in future. 

My bad, the supplied product is an old version and uses L9616, which has 2.5V generated and connected at the middle of split termination resistor. That in combination with bad termination resistor might have caused. We have only recently shifted to TCAN1042 for our battery management system product. Collecting information from field is creating some confusion. I hope you are clear now. 

It'll be great if you could recommend an ideal design for automotive level circuit design for CAN.

Shyama,

The resistors on the CAN bus are burnt because too much energy has been applied to them based on their rating. Typically 1206 resistors are used to survive automotive applications where battery shorts can occur. If the resistors are burnt out, they can cause shorts, or at least lower-impedance paths for current to travel, and measuring this on the CAN bus would show the overall resistance from CANH-GND and CANL-GND to be lower due to parallel resistance combination of the termination resistors and internal resistance of the CAN bus receiver. What kind of event caused the resistors to burn out is, again, likely some kind of voltage transient that caused a lower-impedance path to be created, which drove a lot of current across the CAN bus when driving a dominant signal, and ultimately damaged the resistors. But more often it's a short to battery case. Since the TCAN1042 has higher voltage protection inherently for the CAN bus pins, there will be more protection available.

The schematic you've shown for your CAN transceiver looks sufficient, except I would also recommend a common-mode choke on the CAN bus for further filtering. 

Regards,

Eric Hackett

Hi,

What do you mean by short to battery case? How can that cause an issue? Are we assuming some connection to battery case?

Shyama,

I meant the case where a short to battery occurs since these devices are typically used in Automotive systems. I was just giving an example of how a high-voltage transient could occur. Based on your observations, it seems like a high-voltage transient is the most likely cause of what you're observing in your waveforms.

Regards,

Eric Hackett