Intermittent faults on a modern excavator usually live in the communication between modules rather than inside a component. A controller that loses contact with a sensor or with another module for a few milliseconds logs a code, and the machine carries on. The evidence is in the stored data, not in the present behavior: freeze-frame records the values at the moment the fault occurred, the fault counter says whether it has happened once or four hundred times, and the conditions recorded at the time — engine speed, temperature, load — say which circuit to inspect. Reproduce it with a wiggle test on a warm machine, and only then condemn a part.
What you will notice
| Symptom | What it usually points to |
|---|---|
| A warning that appears and clears on its own | A fault that self-clears is stored. The stored record is the only reliable evidence, because the live values usually look normal. |
| Several unrelated functions affected at the same time | Multiple unrelated symptoms point at a shared data link or shared ground rather than at several components failing together. |
| Fault only after an hour of work, never from cold | Heat changes resistance and clearances. A fault that needs temperature is usually a connection or a sensor, not a module. |
| Warning appears after rough ground or a bump | Mechanical movement is changing a connection. That is a harness or connector fault until proven otherwise. |
Most likely causes
| Cause | How to confirm it |
|---|---|
| Terminal tension in a connector | The cheapest fault in the system and the most common. Terminals lose their spring with age and vibration and lose contact intermittently. |
| Chafed or stretched harness | A harness that has been stretched around a repair, or chafed against a frame edge, breaks contact only in certain positions or after warming. |
| Poor ground connection | A corroded or loose ground reference makes a module reset or misread every signal at once, which looks like several faults rather than one. |
| Sensor failing with temperature | A sensor can pass cold and fail hot. The stored freeze-frame temperature at the moment of the fault is what gives this away. |
What the stored record tells you
| Stored item | How to read it |
|---|---|
| Active versus stored fault | Stored faults are history. A code stored hundreds of times with no active fault is a real intermittent, not a ghost. |
| Freeze-frame values | The readings at the moment the fault set. Compare them with the values now — the difference is the clue. |
| Fault counter | One occurrence is a glitch. A count that climbs with hours worked is a fault that is getting worse. |
| Occurrence conditions | Engine speed, temperature and load at the time of the fault narrow the search to a specific operating state. |
| Related codes | Codes that appear together usually share a supply, a ground or a data link. Read the group, not the single code. |
How to confirm it
- Read and write down every stored code before clearing anything
Write the codes, the counters and the freeze-frame values on paper or photograph the screen. A code cleared in a moment of impatience destroys the only evidence of a fault that may take a week to recur.
- Group the codes by what they share
Codes that share a module, a ground or a data link are one fault, not several. If the group has nothing in common except the machine, suspect the supply or the ground first, because that is what every one of them depends on.
- Do a wiggle test on a warm machine
With the engine at working temperature and the key on, flex each relevant harness section and connector by hand while watching live data or the warning lamp. A fault that appears when a harness moves has just told you where it is. This test is how most intermittent faults on working machines are actually found.
- Measure the ground first
Check the ground reference between the module and the frame with a meter on low resistance, then with a voltage drop test with the circuit loaded. Grounds cause more intermittent faults than modules do, and they cost nothing to check.
- Check terminal tension with a proper test
A meter can read a healthy voltage through a terminal that will not carry current. Load the circuit, or use the specified terminal gauge, before deciding a connector is sound.
- Replace the smallest justifiable part, then re-test hot
Replace the connector section or the sensor first, run the machine to temperature and re-read the counters. Rotating parts on an intermittent fault without a re-test is how a machine ends up with four new components and the same warning.
What to do about it
Connector or terminal repair: correct terminal tension and a new seal are usually the whole repair. Use the specified terminals, and re-seal the connector afterwards, because a connector opened and closed without a new seal is a water path.
Harness repair: repair a chafed section properly with a soldered or crimped joint of the right gauge and heat-shrink with adhesive, then re-route it with the correct clips and clearance. A repair that leaves the harness bearing on a frame edge will fail again within a season.
Ground repair: clean to bare metal, apply a protective compound and tighten to the specified torque. On machines working in wet or coastal conditions, re-check grounds at every major service.
Sensor replacement: only after the circuit is proven and the freeze-frame values support it. Guangzhou Shenyue Machinery Technology Co., Ltd. supplies sensors, switches and connectors from Guangzhou, holds 3,000+ part types on the shelf with 100,000+ part numbers available to order, and declares the grade of every line. The five brands for which the company holds authorized dealer status are Garrett, Mitsubishi Heavy Industries, TONGMYUNG, NIKURA and Marusan, and the Mitsubishi authorization covers MHI engine parts only. Kobelco is not among them, and Kobelco parts are supplied through original channels or as quality alternatives.
Parts you may need, by grade
| Part | Grades commonly quoted |
|---|---|
| Sensors and senders | Genuine; OEM-quality; aftermarket equivalent where the specification matches. |
| Connectors, terminals and seals | Genuine; OEM-quality; matched aftermarket connector kits with the correct terminal tension. |
| Harness sections and repairs | Genuine harness; OEM-quality section repair with specified terminals. |
| Controllers and monitors | Genuine; OEM-quality; remanufactured exchange with core return. |
Need the parts this covers?
Send the machine model, serial number and the part numbers you need. We confirm fitment before payment and state the grade of every line — genuine, OEM-quality or aftermarket.
Request a quoteFrequently asked
The code cleared itself. Does that mean nothing is wrong?
No. A self-clearing fault is stored, and a stored code with a rising counter is a genuine intermittent that will return. What cleared is the warning, not the fault. Read the stored data before clearing anything, because the freeze-frame values at the time of the fault are usually the only evidence you will get.
Why does several unrelated functions fail together?
Because they share something: a ground, a supply or a data link. Components rarely fail in groups on the same day. When they do appear to, look at what the affected circuits have in common and test that first, since one shared connection explains every symptom at once.
Should I replace the controller on an intermittent fault?
Only after the circuit has been proven. A controller is the most expensive part in the chain and the least likely cause. Connector tension, harness condition and ground quality explain the majority of intermittent faults, and each of those is checked in minutes with a meter.