Dear,
Product platform: S32G274, NON-AUTOSAR based development of the program has been tested on the same board is no problem. I used 3.0.2 RTD before, but now I use 4.0.1 RTD based on AUTOSAR. Because this problem occurred after 100ms of power-on, while LLCE initialization took place after 200ms, I think it has nothing to do with LLCE. It may be caused by port, which uses MCAL. So it shouldn't have anything to do with AUTOSAR.
I suspect that the Port driver mistakenly operated the level in the CANTX pin during initialization. I measured the TX pin of CAN:
PC_12/LLCE_CAN0_TX Pin: As shown in the following figure, the green pin is PC_12/LLCE_CAN0_TX. When powered on for about 120ms, it should be during Port initialization. PC12 is pulled down for 110ms and then pulled up.
No pull-down resistors are added to the hardware, which is a directly connected can transceiver. See Attachment 1 for the hardware schematic diagram.
There is no operation for PC_12/LLCE_CAN0_TX except initialization. In addition, this pin configuration is the input, the software can not modify the output level.
In addition, we have done comparative tests, and PC_12 is configured as ordinary input IO, there is no problem.
So I suspect there is a drop down inside the S32G or some other internal chip logic or Port module internal logic.
The following is the comparison test done. In the two tests, only the mode configuration of PC_12 pin is different, and other codes and hardware environments are identical. The software also blocks the initialization of LLCECAN, FLEXCAN, and all code.
PC_12 is configured as the LLCE_CAN0_TX power-on waveform, as shown in Figure 2
PC_12 is configured as a common IO power-on waveform, as shown in Figure 3
Thank You!
Best Regards!