🎓 Lesson 20 D5

CAN Bus Noise Diagnosis Flowchart

CAN Bus noise diagnosis is like listening for static on a walkie-talkie to figure out why messages between farm machinery computers are getting garbled or lost.

🎯 Learning Objectives

  • Analyze CAN differential waveform distortions using oscilloscope measurements to identify noise sources
  • Explain how improper grounding topology introduces common-mode noise on CAN_H/CAN_L lines
  • Apply ISO 11898-2 termination and stub-length rules to diagnose signal integrity failures
  • Design a noise-resilient CAN segment by calculating maximum allowable stub length and verifying termination resistance

📖 Why This Matters

In modern agricultural machinery—like precision planters or autonomous harvesters—CAN Bus networks coordinate critical subsystems: engine control, GPS guidance, hydraulic valves, and sensor arrays. A single undiagnosed noise event can cause false fault codes, intermittent actuator failure, or even safety shutdowns mid-field. Grounding flaws account for >65% of field-reported CAN communication failures (John Deere Field Service Report, 2023). Mastering noise diagnosis isn’t just troubleshooting—it’s ensuring operational continuity, yield protection, and functional safety.

📘 Core Principles

CAN Bus relies on robust differential signaling (CAN_H and CAN_L) to reject common-mode noise—but only when implemented correctly. Noise enters via three primary paths: (1) conducted noise through shared power/ground conductors, (2) radiated EMI from high-frequency switching (e.g., PWM-driven hydraulics), and (3) ground potential differences between nodes caused by multipoint grounding or long chassis runs. The grounding architecture—especially star vs. daisy-chained ground references—directly determines whether noise appears as differential (destructive) or common-mode (often rejectable). Signal integrity further depends on controlled impedance (120 Ω characteristic), proper termination (120 Ω at each bus end), and stub lengths <0.3 m per node (per ISO 11898-2). Understanding the interplay between grounding topology, cable shielding, and transceiver common-mode rejection ratio (CMRR) is foundational to accurate diagnosis.

📐 Maximum Stub Length Calculation

Stub length directly impacts signal reflections and ringing. Exceeding ISO 11898-2’s recommended limit causes bit-time violations and dominant/recessive state misreads. This formula estimates the maximum electrically safe stub length based on bus bitrate and propagation delay.

Maximum Stub Length (ISO 11898-2)

L_{stub} ≤ 0.3 \, \text{m}

Mechanical maximum stub length permitted for reliable CAN communication in mobile machinery environments.

Variables:
SymbolNameUnitDescription
L_{stub} Stub length m Physical distance from CAN bus trunk to node connector
Typical Ranges:
Agricultural machinery (ISO 11898-2): 0.0 – 0.3 m
Heavy-duty off-road vehicles (SAE J1939): 0.0 – 0.3 m

💡 Worked Example

Problem: A John Deere S700 harvester uses a 500 kbps CAN FD network with twisted-pair cable (propagation delay = 5.1 ns/m). What is the maximum allowed stub length?
1. Step 1: Determine minimum bit time = 1 / bitrate = 1 / 500,000 = 2000 ns
2. Step 2: Apply ISO 11898-2 rule: stub delay ≤ 5% of bit time → 0.05 × 2000 ns = 100 ns
3. Step 3: Convert delay to length: 100 ns / 5.1 ns/m ≈ 19.6 m — but this is theoretical; ISO mandates ≤ 0.3 m mechanical stub length regardless of bitrate for robustness in mobile machinery
4. Step 4: Confirm real-world constraint: ISO 11898-2 Table 3 specifies max stub length = 0.3 m for all bitrates in automotive/agricultural applications
Answer: The maximum allowable stub length is 0.3 m. While calculation yields ~19.6 m, the standard enforces 0.3 m to prevent reflections under vibration, temperature cycling, and EMI exposure.

🏗️ Real-World Application

Case: A Case IH Axial-Flow 140 combine reported intermittent 'ECU Communication Lost' warnings during header raise. Diagnosis revealed clean CAN waveforms at the main display ECU but severe ringing and >2V common-mode offset at the header control module. Investigation showed: (1) header harness grounded locally to the header frame (not main chassis), creating a 120 mV ground potential difference; (2) unshielded 2.1 m stub to the header ECU; and (3) missing 120 Ω termination at the header node. Solution: Installed shielded twisted pair with drain-wire grounded *only* at the main chassis star point, shortened stub to 0.22 m, and added proper termination. CAN eye diagrams normalized, and error frames dropped from 42/hour to 0.

📋 Case Connection

📋 Case Study: Sprayer Boom Sensor Noise Reduction via Ground Plane Optimization

Erratic nozzle pulse width modulation (PWM) triggering causing inconsistent application rates at speeds >12 mph

📚 References