0

I want to collect some CAN traffic from my car. I got two devices. One is Saleae Logic8. The other is Vector VN1630A. I think the CAN protocol is straightforward that we can listen the CAN traffic by connecting our device to CANH and CANL. So I connect the Saleae Logic8 to the pin 6 and 14 of OBD II port, which is HS-CANH and HS-CANL respectively. However, I got a CAN traffic which is repeatedly messages with a recessive level in ACK slot, as shown in the pics. The voltage showed in the pics is differential voltages by CANH minus CANL. enter image description here enter image description here enter image description here Let me clarify what the pic showed. The identifier showed which is '456' is decimal number, and its binary is '001 1100 1000', which is correct for the transmitted message. The red cross in the pic means the stuffed bit. Sorry for the confusion.

The same messages is repeatedly sent and seems like there is only one ECU on the bus. However, if I connect the professional CAN bus tool, Vector VN1630A, I can get the normal CAN traffic, many different messages. I then connect the Saleae device to my 2009 Mazda6 OBD pin 6 and 14, it also can detect normal CAN traffic.

Is there any difference in newer vehicles' OBD port? Or is there any request command I should sent firstly to get a stable, normal CAN traffic? Thanks for your help!

Joe
  • 59
  • 5
  • Your tool seem to decode the CAN frame wrong. There's no 14 bit id field. Also it doesn't seem to take stuffing in account? And where do you see the ACK bits recessive? This is CAN L so everything is inverted. I can't be bothered to decode this manually. – Lundin Mar 25 '19 at 07:46
  • 1
    It might be caused by connecting the logic analyser to non logic level pins - the CANH and CANL pins are not logic level pins, instead connect to a CAN transceiver's Tx and Rx pins. – Martin Mar 25 '19 at 09:38
  • @Lundin sorry for the confusion, the id showed in the pic is decimal number, not the hex. The red cross is the stuffed bit. – Joe Mar 25 '19 at 14:51
  • Yes I understood that much, but the id field is still not 14 bits wide. So the picture/decoding simply doesn't make any sense to me. – Lundin Mar 25 '19 at 14:59
  • @Martin 0-3V suggests logic levels actually. Still, that doesn't explain why the length of the various fields is bananas. – Lundin Mar 25 '19 at 15:02

0 Answers0