CAN-Bus Diagnostics on Yachts

CAN networks carry data between engines, instruments and control systems. This guide explains the hardware, electrical signals and message decoding, then sets out practical checks for configuration, wiring and communication faults.

What is a CAN bus?

CAN means Controller Area Network. It lets several devices exchange messages over shared wiring. An engine controller, for example, can send engine-speed data for other devices to use. Each connected device is called a node. [3][12]

On yachts, two common CAN-based systems are:

• NMEA 2000: connects compatible marine instruments and equipment. Devices join a main cable, called the backbone, through shorter drop cables. It runs at 250 kbit/s. [4][5]
• J1939: carries data for compatible engines, transmissions and gensets. A suitable gateway translates selected messages between J1939 and NMEA 2000. [2]

Proprietary networks use the equipment maker’s rules. This article covers conventional two-wire, high-speed CAN. Other CAN types, redundant systems and manufacturer-specific installations require their own instructions. [1]

The components and their jobs

These parts form the path from a sensor’s value to a display. Power feeds and the specified signal-reference connection must also appear on the installation drawing. [1][3][6]

Think of the CAN signal as a wave travelling along the cable. At each end, a terminator acts like a cushion, reducing the electrical “echo” that would otherwise bounce back and interfere with messages. A conventional high-speed bus uses a nominal 120-ohm termination at each physical end. [1]

A basic NMEA 2000 network: compatible devices connect by drops to one backbone, with a 120-ohm terminator at each end and a switched, fused nominal-12-V supply. Schematic only; follow the installed equipment manuals for cable limits, power-feed position and connections.

How the signals become data

A receiver measures the difference between CAN-H and CAN-L. The two signal states represent binary bits:

During a dominant bit, CAN-H rises while CAN-L falls. These are illustrative high-speed-CAN values measured against the specified reference; use the installed equipment’s limits for testing. CAN-L is a data wire, not the power-supply negative. [3][10][12]

The CAN controller groups the received bits into frames and checks them for communication errors. Each frame includes an identifier and, for a data frame, a data field. The receiving software uses the relevant protocol or manufacturer’s definitions to interpret those bytes as values such as engine speed or temperature. [12]

For example, displaying engine speed requires three things: the correct frame must arrive, the software must decode its data correctly, and the display must select the intended engine. This separates electrical faults from decoding and configuration problems.

Worked Example

J1939 engine speed at 1,000 rpm: In a J1939 EEC1 message (PGN 61444), engine speed (SPN 190) is represented by two bytes at 0.125 rpm per count. So 1,000 rpm corresponds to 8,000 counts (0x1F40); with the lower byte first, the speed bytes are 40 1F (01000000 00011111 in binary). A simplified frame sketch is ID 0x0CF004xx (the sender’s address varies) and data FF FF FF 40 1F FF FF FF. Here, the FF bytes are placeholders for other fields, not values to assume on a live network. This is a readable sketch, not the complete wire-level bitstream: the full frame also contains control and error-checking fields, and may include bit-stuffing. Verify the message definition, scaling, and byte order against the installed system’s documentation.

Practical configuration checks

Read and record the existing settings first. Before changing operational settings, agree a safe, authorised maintenance state and save the configuration for recovery. Use the equipment manual’s approved values.

An instance identifies one item among several of the same type, such as port and starboard engines. Maretron’s J2K100, for example, has separate settings for the incoming J1939 source address and outgoing NMEA 2000 engine instance. Follow the installed system’s mapping rather than assigning numbers by guesswork. [2][5][13]

NMEA 2000 uses a nominal-12-V network supply. Never connect its backbone directly to 24 V; multiple feeds require the specified isolation arrangement. Some devices have separate power feeds as well. [7]

Troubleshooting: choose a useful first test

Save alarms and timestamps before resetting equipment. Note recent wiring, software or configuration changes.

These patterns help choose the next measurement. Confirm the cause before replacing a device.

Before intrusive testing: agree the safe maintenance state with the responsible engineer and bridge team. Account for propulsion, steering, automatic starts, protective functions and redundancy. Use approved test points and the OEM procedure. Incorrect engine-network connections can disrupt controls. [2]

Testing power, wiring and messages

Power and connectors

Measure voltage at the affected device’s supply terminals and specified return, under the permitted operating condition. Compare it with the device’s operating range. Comparing readings at the device and distribution panel can help locate a voltage drop.

Check fuses, loose connectors, moisture, damaged contacts and cable strain. Confirm cable lengths against the installation rules: for example, the NMEA 2000 maximum individual drop is 6 m. Engine-CAN limits may differ. [7][8]

Resistance: power off

Shut down as agreed, isolate all relevant feeds, including separately powered controllers, and verify that voltage is absent. Keep the intended nodes and terminators connected unless the OEM instructs otherwise. Then measure resistance across CAN-H and CAN-L. [9]

Compare the result with the network drawing. Some electronic terminations disappear when power is removed, and an adaptor may add its own resistor. A 60-ohm result still leaves signal quality and terminator placement to check. [10]

Never measure resistance on an energised network or apply an insulation-resistance tester to connected CAN electronics.

Voltage and waveform: approved live test

A multimeter shows average signal voltage. An oscilloscope reveals individual transitions and problems such as ringing, where a signal oscillates after an edge. Compare the capture with OEM guidance or a healthy capture from the same design. [3][10]

Use a suitably rated differential probe or genuinely isolated inputs. An ordinary bench-scope ground clip may be connected to protective earth; attaching it to CAN-L can earth the data wire. Never defeat protective earth. Include chargers and USB connections when checking isolation, and use the documented signal reference rather than assuming the hull or shield is suitable. [11]

Messages: capture, then decode

For passive raw-CAN capture, configure an approved interface for genuine silent/listen-only mode before connection. It must suppress data, error frames and acknowledgements. Existing active nodes provide the acknowledgement confirming reception; a silent logger cannot do this for a lone transmitter. [12][13]

Some approved NMEA 2000 diagnostic gateways actively send management messages. Connect these under their documented procedure. Avoid experimental active auto-baud detection, frame injection or setting changes on operational control networks. [13][16]

Look for the expected messages and their timing. If they arrive, check decoder definitions, units and source selection. Save raw captures and settings. Bus-off means a CAN controller has stopped participating after excessive errors; an interface’s counters describe its own controller. Obtain remote ECU status through its approved diagnostics. [14][15]

Confirm the repair

Isolate sections only through an approved shutdown plan. Removing a device can also remove a terminator or interrupt other connections. Keep separate engine and instrument networks connected through their specified gateway.

After work, restore connections and settings, remove test equipment, and verify the affected readings, alarms and controls. Repeat the original fault condition only where safe. Record the cause, test locations, measurements and results so the next engineer can follow what was done.

CAN-Bus Diagnostics on Yachts - Cheat Sheet

Sources

Sources reviewed 24 October 2026. Manufacturer examples illustrate principles; use the manual for the installed equipment and revision.

1. NI, CAN Physical Layer and Termination Guide, updated 15 August 2024; high-speed and low-speed/fault-tolerant sections.
2. Maretron, J2K100 User’s Manual, revision 1.4, 2015; cover warning, §§2.4–2.5. A historical product example, not a current manual for every J2K100 revision.
3. Texas Instruments, Controller Area Network Physical Layer Requirements, SLLA270, January 2008; §§2–4 and physical-layer discussion.
4. NMEA, NMEA 2000 Standards, public overview; accessed 24 October 2026.
5. Kvaser, NMEA 2000 protocol overview, accessed 24 October 2026.
6. Maretron, J2K100 J1939 to NMEA 2000 gateway, product overview; accessed 24 October 2026.
7. Garmin, Power Connection Considerations, Technical Reference for Garmin NMEA 2000 Products, v1; accessed 24 October 2026.
8. Actisense, How to test an NMEA 2000 network, updated 13 February 2023.
9. Kvaser, How to test if your CAN termination works correctly, 2019; powered-down resistance test only.
10. Enovation Controls, CAN BUS Troubleshooting Guide, updated 1 July 2026; termination caveat and multimeter averaging.
11. Tektronix, The Three Facets of Floating Measurement Solutions, grounding, probes and isolation sections; accessed 24 October 2026.
12. Kvaser, The CAN Bus Protocol Tutorial, acknowledgement and protocol discussion; accessed 24 October 2026.
13. CSS Electronics, CL1000 CAN configuration documentation, firmware 05.88; silent-mode and auto-detect notes; accessed 24 October 2026.
14. Kvaser, CAN Error Handling, accessed 24 October 2026.
15. Kvaser, CANlib: Bus Errors, CANlib 5.52 documentation dated 31 May 2026.
16. Actisense, NGT-1 User Manual, issue 2.01, 2015, p.15, minimum PGN list. Product-specific example of management traffic.