The diesel engine in a modern excavator, loader or dozer reports its problems as numbers. Those numbers travel over the machine’s J1939 network to the dash, the service tool and any telematics gateway listening in.
Read them right and the mechanic drives out with a meter, the right connector and a plan. Read half of the code and the mechanic drives out with a sensor the machine does not need. This guide covers the parts of a code, the common ones, a triage rule and what to record.
The anatomy of a code
Under SAE J1939-73, a diagnostic trouble code combines three parts:
- Suspect Parameter Number (SPN), 19 bits. What has the problem: a sensor, a component or a measured value. Nineteen bits leave room for more than half a million numbers.
- Failure Mode Identifier (FMI), 5 bits. How it failed. Five bits give 32 values, 0 to 31.
- Occurrence Count (OC), 7 bits. How many times the code has gone from inactive to active. Seven bits means it caps out rather than counting forever.
The common SPNs are standard across engine makers. Manufacturers can also assign their own numbers from a reserved range for proprietary parameters, so look up any unfamiliar SPN in the manufacturer’s service information before trusting a generic list.
Each code also carries lamp status: the red stop lamp, the amber warning lamp, the protect lamp and the malfunction indicator lamp. That is the engine maker’s own opinion of how urgent the code is. Active codes travel in a message called DM1; previously active codes are stored and read through DM2.
SPNs you will see often
These standard SPNs show up again and again on heavy equipment. Confirm the meaning against the engine manufacturer’s service information for your model.
| SPN | Parameter |
|---|---|
| 94 | Fuel delivery pressure |
| 97 | Water in fuel indicator |
| 100 | Engine oil pressure |
| 102 | Intake manifold (boost) pressure |
| 110 | Engine coolant temperature |
| 111 | Coolant level |
| 168 | Battery potential (system voltage) |
| 190 | Engine speed |
| 639 | J1939 network communication |
| 1761 | DEF tank level |
| 3251 | DPF differential pressure |
FMI: how it failed
The SPN tells you where to look. The FMI tells you what kind of problem you are chasing, and that decides who goes and what they carry.
| FMI | Meaning | Usually points to |
|---|---|---|
| 0, 1 | Value above or below the normal range, most severe | A real condition in the machine |
| 15 to 18 | Value above or below normal, least or moderately severe | A real condition, building |
| 2 | Data erratic, intermittent or incorrect | Sensor, connector or wiring |
| 3, 4 | Voltage above normal or shorted high; below normal or shorted low | Electrical circuit |
| 5, 6 | Current below normal or open circuit; above normal or grounded | Electrical circuit |
| 7 | Mechanical system not responding or out of adjustment | A mechanical part |
| 8, 10 | Abnormal frequency or pulse width; abnormal rate of change | Sensor signal or the part it measures |
| 9, 19 | Abnormal update rate; network data in error | Communication between controllers |
| 11 | Root cause not known | Start with the manual’s procedure |
| 12, 13 | Bad intelligent device; out of calibration | A controller or smart sensor |
| 14 | Special instructions | The manufacturer’s procedure |
The same SPN means very different days depending on the FMI. SPN 100 with FMI 1 says engine oil pressure is really below the safe range: the operator needs to shut down and someone needs to look. SPN 100 with FMI 4 says the oil pressure sensor circuit is shorted low: the engine may be fine, and the job is a meter and a harness, not an oil pump.
Triage rules for the shop desk
Put this on the wall next to whoever reads the alerts. Adapt it to your fleet, and confirm diagnostic and shutdown procedures against the manufacturer’s operation and maintenance manual and service information.
- Red stop lamp, or FMI 0 or 1 on oil pressure, coolant temperature or coolant level. Call the operator now. The machine stops safely and stays stopped until it is inspected.
- FMI 2 to 6. Electrical. Send a mechanic with a multimeter and connector kit. Check harness rub points, connectors near heat and water, and recent repairs nearby.
- FMI 15 to 18. A condition building. Schedule within the shift or the week, and pull the trend for that parameter.
- FMI 7. Mechanical. Check the part itself: a stuck valve, a sticking actuator, a worn linkage.
- Several unrelated codes at once, especially with SPN 168 or 639. Suspect power, ground or the network before any single sensor.
- A code going active and inactive with a climbing occurrence count. Intermittent. Look for a pattern before you replace anything.
Read the occurrence count
The occurrence count is easy to ignore, and it is often the most useful number in the code.
Say an excavator logs SPN 110 FMI 16 (coolant temperature above normal, moderately severe) with an occurrence count of 14 over two weeks. The trace shows every event between 2 p.m. and 5 p.m., and none in the morning. That is not a sensor. That is a cooling stack packed with dust meeting afternoon heat, and the fix is an hour with low-pressure air and a check of the fan and shroud, not a new thermostat.
A single occurrence with no repeat may be a one-off: a loose connector bumped during a walkaround, a low-voltage dip at cranking. Record it and watch it. A climbing count is the machine telling you the same thing over and over.
Aftertreatment codes need a clock
The EPA adopted Tier 4 emission standards for nonroad diesel engines in 2004, phased in from 2008 to 2015. Machines built under those rules carry diesel particulate filters, DEF dosing or both, and with them a family of codes that behave differently from a coolant alarm.
DEF level, DEF quality and DPF codes often start as warnings and escalate. Left alone, many engines reduce power to force a fix, which turns a ten-minute top-up into a stalled excavator and a waiting crew of trucks. Treat the first aftertreatment warning as a work order with a deadline: top up from a clean container, schedule the parked regeneration the manual calls for, or book the diagnosis. Our article on keeping DPF and DEF systems healthy goes deeper.
Use codes to plan ahead
Once a month, count codes by model and SPN across the fleet. Look for the few that keep repeating. Each one is a question worth an hour:
- The same electrical code on several machines of one model points at a harness route or connector the manufacturer may already have a bulletin for.
- Coolant temperature codes that cluster in summer point at cooler cleaning intervals, not parts.
- Water in fuel codes that follow one fuel truck point at the fuel, not the machines.
What to record before you clear
Clearing a code erases evidence. Before anyone clears anything, record:
- SPN, FMI, occurrence count and lamp status
- Engine hours when the code first appeared
- Machine unit number and serial number
- Job, location and what the machine was doing
- What the operator saw, heard or smelled
- A photo of the display or the service tool screen
- Any snapshot or freeze-frame data the controller stores
- What was done, which parts were replaced and by whom
This is the record that backs a warranty claim. A claim that shows the code, the hours, the date and the repair holds up far better than “it threw a code last month.” If warranty money matters to you, our warranty recovery service is built on exactly this record.
Common mistakes with fault codes
- Reading the SPN and ignoring the FMI. Half the code sends you to the wrong half of the problem.
- Replacing a sensor on a wiring fault. FMI 3 to 6 points at the circuit first.
- Clearing codes before recording them. The occurrence count and history go with them.
- Treating every code as urgent. If every alert goes to everyone, people stop reading alerts. Route by severity.
- Using a generic list for a proprietary SPN. Look it up in the manufacturer’s service information.
- Comparing brands by raw code. Some telematics feeds send their own codes rather than SPN and FMI, as our ISO 15143-3 guide explains. Convert them before you compare.
- Skipping service bulletins. A known issue with a published fix can look like a new mystery.
Where Gytix fits
Asset health monitoring reads fault codes from every machine, whether from an A1 gateway on the J1939 network or a manufacturer feed, and shows them in one format with lamp status, occurrence count and hours. Codes are routed by severity to the right phone, and each machine carries an asset health score built from its codes and trends. It is part of the Shop plan.
The AI maintenance agent reads the same codes alongside hours, inspection notes and your service manuals, then drafts a work order and parts list for a person to approve. It never closes a work order or orders parts on its own. For the vocabulary, the telematics glossary defines SPN, FMI, DM1 and the rest.
